NORTH AMERICAN TUNNELING 2004

09096-Prelims.qxd

2/18/04

5:25 PM

Page III

PROCEEDINGS OF THE NORTH AMERICAN TUNNELING CONFERENCE 2004, 17–22 APRIL 2004, ATLANTA, GEORGIA, USA

North American Tunneling 2004 Edited by

Levent Ozdemir Colorado School of Mines, Golden, Colorado, USA

A.A. BALKEMA PUBLISHERS LEIDEN / LONDON / NEW YORK / PHILADELPHIA / SINGAPORE

Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-Prelims.qxd

2/18/04

5:25 PM

Page IV

Copyright © 2004 Taylor & Francis Group plc, London, UK All rights reserved. No part of this publication or the information contained herein may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, by photocopying, recording or otherwise, without written prior permission from the publisher. Although all care is taken to ensure the integrity and quality of this publication and the information herein, no responsibility is assumed by the publishers nor the author for any damage to property or persons as a result of operation or use of this publication and/or the information contained herein. Published by: A.A. Balkema Publishers, a member of Taylor & Francis Group plc www.balkema.nl and www.tandf.co.uk For the complete set (book  CD ROM), ISBN 90 5809 669 6 CD ROM: ISBN 90 5809 670 X Printed in The Netherlands

Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-Prelims.qxd

2/18/04

5:25 PM

Page V

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Table of Contents

Foreword Levent Ozdemir

XI

Organization

XIII

Session 1 – Differing site conditions as applied to design/build contracts Track 1 – Project management Design review boards – current state of practice A. Elioff & W.W. Edgerton

5

Drawing from past experience to improve the management of future underground projects C. Laughton

15

The ECIS story J.W. Critchfield & B. Miya

21

Track 2 – Security of critical infrastructure and key national assets: use of underground space Internal blasting and impacts to tunnels Wern-ping (Nick) Chen

29

Track 3 – Mechanized tunneling Improvements of the capabilities of cutting tools and cutting systems R. Bauer

37

MTBM and small TBM experience with boulders S.W. Hunt & F.M. Mazhar

47

Joint orientations for TBM performance analysis using borehole geophysics to orient rock cores T. Tharpe, B. Crenshaw & J. Raymer

65

Slurry type shielded TBM for the alluvial strata excavation in downtown area W.R. Jee

73

Estimating ground loss from EPB tunneling in alluvial soils for ECIS project, Los Angeles T.R. Seeley

79

Some aspects of grouting technology for Manhattan tunnels M. Ryzhevskiy & P. Barraclough

87

Track 4 – Specialized urban construction Design and construction of an LRT tunnel in San Jose, CA P.J. Doig

V Copyright © 2004 Taylor & Francis Group plc, London, UK

95

09096-Prelims.qxd

2/18/04

5:25 PM

Page VI

Underpinning design and construction – Atlantic Avenue Station complex rehabilitation, New York, USA A. Grigoryan & L.G. Silano

101

Slurry walls accelerate shaft construction in rock in Los Angeles M.P. McKenna, K.K. So, M.A. Krulc & E. Itzig-Heine

109

Performance of Russia Wharf Buildings during tunneling H.S. Lacy, M.D. Boscardin & L.A. Becker

121

Blasting adjacent to high voltage duct banks K.R. Ott, D.A. Anderson & S.E. Haq

129

Subway rehabilitation – secant wall cofferdams and penetration of tunnel liner V. Tirolo & N. Hirsch

135

Overcoming the complex geotechnical challenges of urban construction T.J. Tuozzolo

143

Session 2 – Subsurface investigations and geotechnical report preparation for design/build projects Track 1 – Risk allocation Risk management in tunneling – occupational safety  health plans for drill and blast and tunnel boring machines A. Moergeli

153

Managing underground construction risks in New York N. Munfah, S. Zlatanic & P. Baraclough

163

Risk allocation in tunnel construction contracts W.R. Wildman

171

Getting back on-track: Exchange Place Station Improvements M.F. McNeilly, S.A. Leifer & G.F. Slattery

177

Influence of geologic conditions on excavation methodology E.C. Wang, L.M. Hsia, C.C. Chang & A.N. Shah

185

Track 2 – Owners opinion forum Discussion and panel talk sessions – no written papers

Track 3 – Non-mechanized construction Santiago’s Metro expands C.H. Mercado, G.S. Chamorro & K. Egger

195

Benchmark for the future: the largest SEM soft ground tunnels in the United States for the Beacon Hill Station in Seattle J. Laubbichler, T. Schwind & G. Urschitz

201

Application of the Press-In Method in East Side Access tunnel project J. Liu & V. Nasri

209

Shotcrete for tunnel final linings – design and construction considerations V. Gall, K. Zeidler, N. Munfah & D. Cerulli

215

Robotic shotcrete applications for mining and tunneling M. Rispin, C. Gause & T. Kurth

223

VI Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-Prelims.qxd

2/18/04

5:25 PM

Page VII

Development of the LaserShell method of tunneling C.M. Eddie & C. Neumann

231

Ground support design and analysis: Exchange Place Station Improvements M.R. Funkhouser & M.F. McNeilly

241

Track 4 – Ground modification for underground construction Cantilever frozen ground structure to support 18 m deep excavation D.K. Chang, P.W. Deming, H.S. Lacy & P.A. van Dijk

251

New chemical grouting materials and delivery equipment technologies G.N. Greenfield & A.C. Plaisted

259

Jet grout bottom seal for cut and cover tunnel T.M. Hurley

265

North airfield drainage improvement at Chicago O’Hare International Airport: soil stabilization using jet grouting D.A. Lewis & M.G. Taube

271

Ground freezing and spray concrete lining in the reconstruction of a collapsed tunnel S.J. Munks, P. Chamley & C. Eddie

277

Ground freezing for urban applications P.C. Schmall, D. Maishman, J.M. McCann & D.K. Mueller

285

Session 3 – Design/build contracting practices Track 1 – Predicting and controlling cost and schedule An economic approach to risk management for tunnels B. Altabba, H. Einstein & H. Caspe Top down construction of Ramp L, Value-Engineering Change Proposal for the Massachusetts Turnpike Authority, Contract CO9A4 W.D. Driscoll & G.A. Almeraris Contemporary methods of budget preparation B. Martin & S. Sadek

295

303 313

Geotechnical mapping methods utilized in the Chattahoochee Tunnel Project, Cobb County, Georgia, USA J. Reineke, J. Raymer, M. Feeney & K. Kilby

319

Value engineered design facilitates Grand and Bates Relief Sewer Tunnel Construction, St. Louis, MO J.R. Wheeler & N.E. Thomson

327

Track 2 – Show me the money Discussion and panel talk sessions – no written papers

Track 3 – Investigation, inspection and rehabilitation Monitoring excavations using 3D Laser Scanning and Digital Close-Range Photogrammetry T. Trupp, L. Liu & Y. Hashash

VII Copyright © 2004 Taylor & Francis Group plc, London, UK

337

09096-Prelims.qxd

2/18/04

5:25 PM

Page VIII

Durability and corrosion protection of support systems in soil and rock tunnels M.R. Jafari, V. Nasri & M. Wone Investigation of complex geologic conditions for the Second Avenue Subway tunnel alignment in New York City, New York C.P. Snee, M.A. Ponti & A.N. Shah

345

357

An automated structural monitoring system for the Federal Reserve Bank of Boston T.L. Weinmann & L. Edgers

363

A deep horizontal boring – technical and contractual issues J. Glastonbury, K. Ott, J. Freitas, B. Russell, M. Wooden, W. Meakin & J. Canale

373

Rehabilitation of the Big Walker Mountain Tunnel in Bristol, Virginia D. Kukreja & P. Moran

381

Corrosion evaluation of the Manhattan rocks and corrosion protection of the rock reinforcement system for subway tunnels M. Ryzhevskiy & M. Berman

389

Rehabilitation of the Amtrak Long Island City ventilation structures S.G. Price

395

Using seismic tomography and holography ground imaging to improve site investigations E.J. Kase & T.A. Ross

401

Track 4 – Machine mining – soft ground to hard rock to everything in between Conditions encountered in the construction of the Braintree-Weymouth Tunnel Project, Boston, Massachusetts D.W. Deere, J. Kantola & T. Davidson

411

The Manapouri Tailrace Tunnel No. 2 construction – a very large TBM tunnel in very strong rock D.W. Deere, S. Keis & C. Watts

421

South Austin Regional Waste Water Treatment Plant Interconnect Tunnel Project S. Cheema, K. Koeller, R. Pohren, G. Sherry & R. Webb

433

Tunneling through an operational oilfield and active faults on the ECIS Project, Los Angeles, CA, USA E. Keller & M. Crow

441

Rock tunneling at the Mill Creek project M. Schafer, B. Lukajic, R. Pintabona, M. Kritzer, T. Shively & R. Switalski

449

Construction of the Dougherty Valley Tunnel, San Ramon, California, USA G.S. Nagle & H. Thom

453

City of Los Angeles Northeast Interceptor Sewer Tunnel Z. Varley, R. Patel & J. McDonald

461

Session 4 – Design/build risk Track 1 – SEM/NATM practices/prescriptive specifications NATM and its practice in the US Wern-ping (Nick) Chen & H. Caspe

473

SEM/NATM design and contracting strategies J. Gildner & G.J. Urschitz

477

VIII Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-Prelims.qxd

2/18/04

5:25 PM

Page IX

Engineers, contractors, and soft-ground tunneling equipment W.H. Hansmire & J.E. Monsees

485

Track 2 – Transit oriented development – making the case for going underground Atlanta West Area Combined Sewer Overflow Storage Tunnel and Pumping Station R.C. Divito, W. Klecan & G.D. Barnes

497

Track 3 – Analysis and design Consideration on machine data and load in TBM excavation for tunnel support selection N. Isago, H. Mashimo, W. Akagi & H. Shiroma

507

A durability design for precast concrete segments for tunnel linings G. Bracher & D. Wrixon

515

Design and construction of the Lindbergh Terminal Station, Twin Cities, Minnesota E.E. Leagjeld, B.K. Nelson, C.R. Nelson, D.L. Petersen, R.L. Peterson & B.D. Wagener

521

Design and impact of the Beacon Hill Station exploratory shaft program C. Tattersall, T. Gregor & M.J. Lehnen

529

Comparison of the predicted behavior of the Manhattan TBM launch shaft with the observed data, East Side Access Project, New York V. Nasri, W.S. Lee & J. Rice

537

Drop shafts – selection principals J.F. Zurawski & E. Petrossian

545

Stability evaluation and numerical modeling Exchange Place Station Improvements J.F. Lupo & M.F. McNeilly

553

Track 4 – Conventional underground construction Tunnel and shaft construction for the Pingston Hydro Project B. Downing, Z. Vorvis, G. Rawlings & P. Kemp

561

Shoal creek raw water intake and pump station construction on Lake Lanier D. Ackerman, R. Wiek & R. Gutridge

571

Design and construction of shafts at the San Roque Project M. Funkhouser, R. Humphries, W. Warburton, J. Daly & E. O’Connor

575

Ten years’ experience using roadheaders to bore tunnels for the Bilbao Metro J. Madinaveitia

581

Rio Piedras Project, San Juan, Puerto Rico B. Fulcher, N. Kofoed, P. Madsen & M. Bartlett

589

Devil’s Slide Tunnels Y. Nien Wang, B. Hughes, H. Caspe & M. Amini

605

IX Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-Prelims.qxd

2/18/04

5:25 PM

Page XI

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Foreword

The theme of this North American Tunneling Conference (NAT 2004) is “Underground Construction – The Sensible Solution to Urban Problems”. This title reflects the increasing importance of locating facilities underground for enhanced security and function of urban areas and to build critical infrastructure for sustainable development. This conference includes papers covering a wide range of subjects dealing with nearly all aspects of underground construction, tunneling and effective utilization of underground space. The papers are grouped under four major tracks. Track 1 addresses the management of underground projects and includes presentations on project management, risk allocation and predicting and controlling cost and schedule. Track 2 includes presentations and panel discussions on issues related to security of critical infrastructure and key national assets, owner’s opinion forum, financing of underground projects and transit oriented development making the case for going underground. Track 3 addresses new advances in technology, including sessions on mechanical tunneling, non-mechanized construction, investigation, inspection and rehabilitation and analysis and design of underground structures. Track 4 covers trials, tribulations and triumphs in tunneling industry by presenting significant case histories. The sessions address specialized urban construction, conventional underground construction and machine mining in soft ground, hard rock and mixed-face conditions. I would like to express my appreciation to NAT 2004 organizing committee, track and technical program chairs, panel members and the authors for their contribution to the success of the conference. The continuing support of cooperating organizations, AMITOS, TAC, NUCA, NASTT and UTRC is also acknowledged. Levent Ozdemir Proceedings Editor

XI Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-Prelims.qxd

2/18/04

5:25 PM

Page XIII

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Organization

NAT 2004 EXECUTIVE COMMITTEE Raymond W. Henn, Conference Chair Lyman Henn, Inc., Denver, CO Refik Elibay, Vice Chair Jordan Jones & Goulding, Atlanta, GA Susan Nelson, Executive Director AUA, Minneapolis, MN William W. Edgerton, Chair, Track I Managing Underground Projects Jacobs Associates, San Francisco, CA Brenda M. Bohlke, Chair, Track II Public Policy and Underground Projects PB Consult, Herndon, VA Robert J.F. Goodfellow, Chair, Track III Advances in Technology URS Corp, Gaithersburg, MD Gary Almeraris, Chair, Track IV Case Studies: Trials, Tribulations and Triumphs of Tunneling Slattery/Skanska, Whitestone, NY George Yoggy, Exhibition Chair GCS LLC, Allentown, PA Thomas Clemens, Technical Tour Chair American Commercial, Louisville, KY Carin Mindel, Exhibition Manager AUA, Minneapolis, MN

SESSION CHAIRS Dan Dobbels, Haley & Aldrich Brian Fulcher, Kiewit Construction Company Michael Goode, Telford Consulting Michael Greenberg, NYC Department of Environmental Protection John Kaplin, Gilbane Building Company Gary Irwin, City of Portland Bureau of Engineering Laurene Mahan, PBConsult, Inc. Bill Mariucci, Kiewit Construction Joseph M. McCann, Freeze Wall Chris Mueller, URS Corporation Galen Nagle, URS Corp.

XIII Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-Prelims.qxd

2/18/04

5:25 PM

Page XIV

Levent Ozdemir, Colorado School of Mines Stephen C. Redmond, Frontier Kemper Constructors Tibor Rozgani, Colorado School of Mines Heiner Sander, ILF Vince Tirolo, Slattery/Skanska

AUA BOARD OF DIRECTORS Officers Raymond W. Henn, President Thomas F. Peyton, President-Elect George D. Yoggy, Past President Hugh S. Caspe, Treasurer Susan R. Nelson, Executive Director Directors Gary Almeraris Charles H. Atherton Brenda M. Bohlke Jack Brockway Thomas Clemens Joseph P. Gildner Michael Greenberg Hugh Lacy Robert A. Pond Gregory L. Raines Kirk Samuelson Don Zeni Designated Representatives Charles W. Daugherty Randall J. Essex D. Tom Iseley Levent Ozdemir

XIV Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-01.qxd

2/18/04

8:13 PM

Page 1

Session 1 Differing site conditions as applied to design/build contracts

Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-01.qxd

2/18/04

8:13 PM

Page 3

Session 1, Track 1 Project management

Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-01.qxd

2/18/04

8:13 PM

Page 5

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Design review boards – current state of practice Amanda Elioff Parsons Brinckerhoff, Los Angeles, California, USA

William W. Edgerton Jacobs Associates, San Francisco, California, USA

ABSTRACT: This paper summarizes the current state of practice on the use of design review boards, or consulting boards, as used primarily used during design of underground construction projects. It discusses the various types of boards in use and reviews the history, and then, using the results of an industry survey of both owners and consultants, it discusses the purpose and typical uses, provides examples of specific outcomes, and reviews the methods of selection and modes of operation. It summarizes the advantages and disadvantages, evaluates the use of the construction manager to provide design review, and provides recommendations for future users based upon the lessons learned to date.

1 INTRODUCTION

•

Owner agencies have used a number of different methods for evaluating or “verifying” the design of underground facilities before advertising for bids. These methods include: Independent Peer Review, Value Engineering, Boards of Consultants, and Technical Review Committees. For the purposes of this paper, we have developed the following definitions:

•

•

•

This paper focuses on the current state of practice of Boards of Consultants, and Technical Review Committees, and is not intended to evaluate the use of either IPPR or VE panels.

Independent Project Peer Review: An independent panel tasked with design review for some outside party such as financing agency, congressional committee, etc. This process typically includes an in-depth review of criteria, analysis, and calculations. Value Engineering: Formal evaluation of design documents that evaluates design and to some extent anticipated construction methods and is focused primarily on cost objectives. This process typically consists of a one-week workshop with participants specially-trained in value-engineering skills, and results in recommendations for design changes to reduce cost while maintaining objectives. Board of Consultants: A separate board or panel under contract to the owner agency to evaluate the design prepared by the design consultant. This review can be done at specific time periods as the design proceeds, and is intended to determine bigpicture design issues and does not typically review detailed analysis or calculations. Can also be referred to as a Technical Advisory Panel (TAP).

2 HISTORY OF DESIGN CONSULTING BOARDS Owners have relied upon individual consultants to supplement the prime designer for some time. (See Terzaghi (1958) which contains an excellent review of his personal experience serving as an individual consultant on design and construction projects.) Review boards have been employed on major complex public works projects dating at least to the early 19th century during the bridge building era. (Petroski, 1996). Similarly, boards have been assembled for large dam projects under construction by public agencies such as the Tennessee Valley Authority, the US Army Corps of Engineers (COE) and the Bureau of Reclamation. In recent years, major large subway projects, for example the Bay Area Rapid Transit (BART)

5 Copyright © 2004 Taylor & Francis Group plc, London, UK

Technical Review Committee: A Board that is formed as a part of the design team, to evaluate the design progress and solutions on a periodic basis. The type of review and evaluation is similar to that performed by a Board of Consultants, with the primary difference being that the Board’s client is the design firm, rather than the owner agency.

09096-01.qxd

2/18/04

8:13 PM

Page 6

have stated that “The purpose of the Board should be to provide an objective, balanced and impartial view of the overall design and construction progress on a project. The Board should not be used as a substitute for normal consulting services …” Hoek & Imrie (1995). This usually requires an “independent” board that reviews the work of others (e.g., designers) at pre-determined intervals. To achieve this purpose, the owner must keep the review as a separate function and not let the consulting board become a part of the design team, thus insuring its “independence” and the ability for the board to testify (if necessary) as to the design adequacy. (2) More recently, boards have been used to improve both the efficiency and accuracy of the work product, acting as a part of the design team. The work product is typically plans, specifications, other contract documents, and (sometimes) cost estimates. To fulfill this purpose the consulting board (or technical review committee) need not be “independent” but can contribute to the design process at any time, even continuously at certain key period. Since the board or panel is an integral member of the team, the members cannot be presumed to provide an “independent” review of the work products.

System, designed in the 1960s, WMATA (1970s) and Los Angeles Metro (1980s-present) have maintained boards in some form to advise on project design. Within the past 30 years there has been an increase in the use of consulting boards. This may be in part due to the increasing complexity and multi-disciplinary nature of large projects. It may also be due to increasing oversight of the use of public funds and the arguably increased level of litigation resulting from the construction of such large projects. To the extent that this litigation is founded upon the theory of inadequate or defective design documents, both owners and designers are motivated to minimize these problems. 3 SURVEY From April to October 2003, an industry survey was conducted which asked questions concerning (1) the history of the use of such boards, primarily in the underground industry, (2) the purpose and typical uses of these boards as they are currently constituted, (3) typical criteria for selection of board members, (4) various modes of operation, and (5) approximate cost. The survey also solicited feedback from the respondents as to the use of the construction manager providing such design review, the perceived advantages and disadvantages of boards, and recommendations for improvement in the future. The results of this survey are incorporated into this paper. The survey instrument itself is available from the authors upon request. Although respondents were promised confidentiality, the raw data itself, absent attribution, is also available for subsequent researchers upon request. The survey was sent to 95 people identified by the authors as either consultants or owners who have experience either employing or participating on boards. We received 48 replies, a 51 percent response rate. The respondents’ experience represents over 500 boards as users of the process (i.e., receiving advice, either as an owner, designer, or construction manager), and over 300 boards as board members (i.e., providing advice). A list of the projects from upon which the respondents have based many of their comments is included as an Appendix. Projects represented by the experience of the survey respondents include tunnels both in soft ground and rock, transit stations, underground powerhouses, wastewater treatment plants, large dams, highway projects, pipelines, microtunnels, and large diameter shafts.

Occasionally a funding agency will require a design review board, and when that is the case, the owner’s purpose is to fulfill the specific agency requirements. Examples of such agencies are the Federal Energy Regulatory Commission (FERC), and the Federal Highway Administration (FHWA). Other agencies provide internal review teams when there are significant specialty design issues or high-risk elements. In the case of the Federal Transit Administration (FTA), project management oversight consultants (PMOC’s or PMO’s) are be established to “help ensure that grantees [of federal funds] constructing major transit projects have the technical capability to carry out the projects’ design and construction according to accepted engineering principles,” GAO (2000). The PMO function was incorporated into FTA “new starts” projects after some quality, cost and construction management issues occurred in the 1970s and early 80s. The oversight function includes review and evaluation of various project processes to ensure: compliance with statutory, administrative, and regulatory requirements. The PMOC and the other members of the design team typically work together in the design phase, but although the typical operation of PMOC is similar to that of consulting boards, because the purpose is to assure compliance with specific funding agency requirements, it does not serve the same function as a design review board, and the owner may not rely upon it to fulfill the same purpose

4 PURPOSE AND TYPICAL USES Two primary reasons are given for creating a design review board: (1) To reassure upper level decision makers that the design solution is adequate. Previous commentators

6 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-01.qxd

2/18/04

8:13 PM

Page 7

Changes in concept: Examples included changes from an exploratory shaft to an exploratory tunnel, and changes in tunnel ventilation systems to implement European design methods in the United States. More effective methods: In one example, use of explosives was deemed too dangerous by the design team and owner, as neither believed it would be safe or would be acceptable to the public. The consulting board was able to convince the design team that blasting could be done safely, if designed and implemented properly, and this resulted in savings of significant time and money. Contract packaging/Contracting methods: Examples cited were recommendations for pre-qualification of bidders, changes in contract pricing methods (e.g., plugged prices), and owner purchased TBM’s to save schedule. In addition to these examples of specific outcomes, consulting boards have been instrumental in the development of and/or “blessing” the use of new or “never been done before” design and contracting approaches: Acceptance of unique or first time designs can be difficult for some owners, and conservative designers to accept. The boards’ recommendations for additional testing to verify design assumptions was cited in a few cases, such as for special seismic designs, high loading assumptions, and gas barriers.

Table 1. Technical issues addressed by board. Technical issue

Percent of respondents

Geotechnical Engineering Design Methods Estimating/Scheduling Constructability Contracting Methods Equipment Selection/Approval Risk Evaluation/Assessment Special Construction Techniques

93 67 54 89 48 39 67 65

for which a design review board is established. In addition, the selection of PMOC consultants is usually much different, and the criteria and skill of the participants varies significantly from that used on consulting boards. Consulting boards and panels are usually formed to provide advice and make recommendations on certain technical issues. The industry survey indicated the percentage of respondents who have used such boards for these technical issues as given in Table 1. In some cases design consulting boards continue to provide consultation during construction, and in such cases they evaluate construction issues such as verification of design intent, basis of design, and contractor performance. There is little evidence that such design boards are used to resolve disputes between the contracting parties, although in some cases they have advised the owner on pre-dispute technical issues. Consulting boards discussed in this paper are not Dispute Resolution Boards. (For further information on DRB’s, see Matyas et al. (1996)).

6 SELECTION OF BOARD MEMBERS Members are selected for most consulting and/or review boards on the basis of recommendations by the design team or others. In fact, 95% of the survey respondents indicated that this is the most common method. In some cases an RFP or letter of interest is sent to the industry, and the board members selected using this method. 20% of the respondents had used this method, but only 5% used this method exclusively; i.e., did not rely upon recommendations of the design team. The background of board members appears to be quite varied. Members of academia (university professors) are used frequently, as are contractors, construction managers, and other designers. The industry survey indicated the percentage of respondents who have selected members with the backgrounds as given in Table 2. The use of individual consultants with one specific technical specialty is the most popular; and the technical specialties are determined based upon the key issues on the individual project. For most underground projects, board member backgrounds include geology, geotechnical engineering, tunnel boring machine (TBM) design, ground support design, and other specialties as required. Also noted was the use of operations and maintenance staff, the program manager,

5 SPECIFIC OUTCOMES Survey respondents gave numerous examples of outcomes resulting from consulting board meetings. These can be categorized as changes in design approach, construction methods, concepts, and contracting packaging and/or methods: Changes in design approach: Numerous respondents reported alignment changes (higher or lower tunnel profile) to improve excavation conditions, reduce settlement, avoid hazardous material, and subsequently save cost. Other recommendations were additional explorations, to collect more geologic or groundwater data. Changes in excavation method: These included changes from excavation by Tunnel Boring Machine (TBM) to use of the Sequential Excavation Method (SEM), use of closed face and/or pressurized face TBMs in lieu of open face shields to control settlement or mitigate hazardous conditions, and recommendations for changes in excavation sequence.

7 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-01.qxd

2/18/04

8:13 PM

Page 8

Most consulting boards are asked to reduce their observations to a written report at the end of the meeting. The only exception to this policy appears to be when the owner’s primary consideration is to limit the cost of the board, and in such cases results are reported verbally to the design team. In 70% of the reported cases, both verbal and written comments are provided. Also, in addition to comments to the design team, 80% of the survey respondents indicated that the board has provided comments to owner-agency upper level staff at a meeting. This comports with one of the underlying purposes that is to provide independent review.

Table 2. Backgrounds of respondents. Background

Percent of respondents

Academia Construction Contractors Construction Managers Designers

77 64 60 81

retired government employees, and upper management representatives from other public agencies. Owners that continue board operation through construction sometimes change the makeup, deleting academia and designers, and adding ex-contractors and construction managers to better evaluate the construction issues. In at least one case, a board has had access to a separate group of specialists, “… individuals that are not involved in the design of the project but available to serve in an ad hoc capacity to the Board on an asneeded basis on specialty issues” (Shamma et al. (2003)).

8 COST Most boards are composed of senior-level people who operate on a consulting basis and are compensated by the hour. (For information concerning contract provisions for senior-level consultants, see Dunnicliff & Parker, 2002). The hourly rates are relatively high compared to those of the design team, but owners who have used consulting boards report that the limited use of the board’s time, in part because of the ability of most experienced consultants to quickly identify the root issues, results in total cost to the project that is quite low, comparatively speaking, for the value added. Reported cost ranges from 0.5% to 1.5% of the total design fee; less than 0.1% of the construction cost. The total cost of the design review board ranges from $30,000 to $300,000, although the total cost is quite variable depending upon the frequency of meetings, length of the design period, and number of project contracts.

7 MODES OF OPERATION Virtually all of the survey respondents have used consulting boards during the planning and design phases of the project. A surprising 70% have continued the use of these boards into the construction phase, although only 15% have used them in the post-construction phase, presumably to defend contract disputes with the contractors. During the planning phase, most consulting boards meet only one or two times, although some respondents indicated 10 to 12 meetings, and some have experienced quarterly meetings. (These responses were received from individuals who were part of large construction programs that included more than one construction contract.) During design, some boards meet once, and some up to 6–8 times; although most respondents indicated from 2–3 meetings. These meetings are typically held at pre-determined milestone times, such as 30%, 60% and 90% design completion. The type of work reviewed at early periods is typically quite different from that reviewed at later stages of completion. More importantly, the ability and willingness of the design team to accept recommendations at the later stages of design is limited. Typical meetings are from one to five days in length. Summary or relevant documents are usually provided to the board for review in advance, and at the beginning of the meeting the designer makes a short presentation setting forth the key issues and the status of work currently under development. In some cases a tour of the work site is provided, especially if local conditions are critical to the design solution.

9 ADVANTAGES AND DISADVANTAGES 9.1

8 Copyright © 2004 Taylor & Francis Group plc, London, UK

Advantages

There are advantages to using consulting boards, some of which have been previously mentioned. Advice from a senior advisory board provides an independent check on the design criteria, which is helpful because “Those involved in the design and construction of a project can often become so involved in the details of the work that they find it difficult to stand back and take an impartial view of alternate approaches” (Hoek & Imrie (1995)). This advice can also provide the owner with the support to make decisions and design changes when warranted. If completed early enough in the process, it can provide a level of credibility and a “stamp of approval” to the design solution, and also provides the owner with confidence in its designer. Survey respondents confirmed this summary and also provided a range of other advantages summarized in categories as follows:

09096-01.qxd

2/18/04

8:13 PM

Page 9

a format to brief the Agency directors, PMO and other invited guests to all or summary portions of the meeting. These representatives may otherwise have little direct contact with the designer for questions and answers about the design or for relating political issues to the design team. When used during construction, the board can provide the owner with “third party” advice on contract disputes or differing site conditions, which is helpful if there is disagreement between the designer and CM. Risk Management: The board’s assessment of project risk is not likely to be as “sugarcoated” or conservative as that provided by the design team, thus providing more value to the owner agency. Conversely, the board may point out over conservatism in design.

Overall review of project: Coming to the project with experienced, “fresh” sets of eyes, the board’s review and concurrence with the design approach and criteria developed bring additional confidence to owners and engineers. In the process of review, they may point out overlooked issues, and recommend new areas to look into or additional study, such as more geotechnical exploration. For programs with multiple design contracts and no program manager, the board can provide a level of consistency with the design criteria and other factors. Bring additional experience, perspective, and trust: Given the collective years of experience, and worldwide exposure, the board members are able to compare the project at hand to past experiences, i.e., lessons learned from many underground projects. These board members may have access to information about other projects well before it is published. In some cases, negative experiences are never published, thus making it difficult to apply these lessons learned to future projects without the input of people who possess the appropriate first hand knowledge and can relate it confidentially in a venue such as a consulting board meeting. The board can also provide input to specialties that are not present in the personnel on the design team. This generally results in better quality contract documents that are more consistent, constructible, and results in better bid prices. Advice from disinterested “outsiders” may be more acceptable to politicians and the public. One survey respondent attributed the following quote to Walter Douglas: “ you hire a consultant (1) because you face a difficult problem you have never faced so you hire someone who has or (2) you need an expert to say something you could say but it is more believable because of his or her reputation. Point number two is particularly applicable as politicians, agency boards of directors, and others in upper management may be more willing to accept “bad news” if the source is a group of renown experts rather than staff or the design consultant. Focus on Key issues: In the process of preparing for the board meetings, designers and owners, must assemble relevant information for preview and presentation to the board. To have an effective meeting, they must develop the key issues they would like the board to address. These periodic meetings assist the design team by providing a “time-out” from the day-to-day crash program of completing the design documents. This allows for a review of “where are we going” that is of benefit to all participants. Often, it is not until faced with an upcoming meeting that the questions/issues are well defined. Increase Communication: The format of a “roundtable” type discussion, over a day to several days promotes better interaction between the personalities involved and facilitates better understanding of all positions and issues. The board meetings also provide

9.2

9 Copyright © 2004 Taylor & Francis Group plc, London, UK

Disadvantages

While disadvantages of consulting boards were cited less often, survey respondents did report several related to time and cost, disagreement between members, overreliance on board opinions, potential for late changes, and problems associated with the composition of the board: Time and Cost: Time and cost to prepare for and conduct, and present conclusions from the meeting may be as much as a week or two. Not only the board members, but also the project’s top management and designers may be tied up for days at these meetings, impacting the design schedule. The logistics of gathering all board members together can also be somewhat time consuming, since such senior people typically have full calendars. The design schedule can be impacted not only by the time for preparation and meetings, but also by the time required to review and revise the design documents should the board recommend changes. This can be a financial burden for smaller projects. Reliance on the Board: The presence of a consulting board can affect the design team’s view of responsibility for design decisions. Some designers (and owners) may be tempted to use such boards as “cover”, thus allowing them to avoid accountability for their design solution. Potential for Late Changes: The nature of the periodic review can make designers feel they must defend their design because they don’t have time to change it and still meet schedule. This works against the principle of collaboration, and can lead to the designer and/or owner disregarding the board’s advice to stay on schedule. Group Dynamics/Board Composition: Several instances were reported where board members disagreed, were uncooperative, or conversely, were too willing to compromise, resulting in “design by committee syndrome.” Boards and panels, in evaluating all of the issues, may not be able to find any middle ground. As a result, reaching consensus can be difficult,

09096-01.qxd

2/18/04

8:13 PM

Page 10

•

and such consensus is important for owners whose primary purpose is to convince upper level management that the design is adequate. On the other hand, there is also the disadvantage of “Design by Committee” which is different than “achieving consensus.” Without an effort to identify the reasons for all recommendations, when presented with a choice between alternate approaches or actions, there is a tendency to do both, thus resulting in an over-conservative design solution. Not all board members are helpful, nor do they all understand their role: Some want to be the designer, some may want to manage the entire process, some merely want to obstruct the designer’s progress for competitive reasons, and some simply are not qualified to be on the board. Opinionated panel members can also be counter-productive leading to conflict and/or delay.

•

• •

10 USE OF THE CONSTRUCTION MANAGER FOR DESIGN REVIEW There have been some suggestions that the owner could use its construction manager (CM) to perform the design review functions that are sometimes done with a consulting board or technical review board. The arguments in favor of this approach include: 10.1

• • • • •

•

Arguments in favor of using the CM

The CM’s construction experience will benefit the design solution. A document review from the point of view of a construction specialist will identify inconsistencies in the documents. Better bids will result if the contractors know that the CM has been involved at an early stage. The CM may be better equipped to consider market conditions and the advantages/disadvantages of different contract size and packaging strategies. It’s a good way to get the CM up to speed on the design intent, and have them buy into the design, thus reducing future disagreements with the designer. Review can be continuous rather than at specific times, allowing more timely and therefore less expensive design modifications than would be possible by waiting for the next scheduled consulting board meeting.

10.2

10.3

Can both be used?

Many users feel that both methods should be used, with the consulting board used early in the process, and the CM used later to provide constructability input and a consistency check of the documents prior to advertising. Many respondents said that the most important element in both methods is the use of knowledgeable personnel. 11 RECOMMENDATIONS After evaluating the comments from industry representatives, we offer the following recommendations for improving the results and the success of consulting boards. These recommendations are summarized by category: Purpose/Use, Member Selection, and Operation.

Arguments against using the CM 11.1

On the other hand, there are several disadvantages of using the CM instead of a separate design consulting board to perform these functions:

•

The future construction phase services contract may bias the CM’s recommendations to what the owner wants to hear. A separate consulting board would provide more independence from the process, i.e., the board has no self-interest in the outcome. Bringing on the CM earlier will increase the owner’s costs, and could encourage a postponement of the CM procurement, thus delaying critical input into the design. In addition, once mobilized, it is difficult to cut back CM’s costs if the project schedule slips due to public resistance or financing difficulties. Separate review boards are usually lower cost, as a result of a board’s “spot” reviews of specific issues, as opposed to the CM’s continuous review as a part of the design process. The CM may be perceived as less technically qualified (i.e., credible) on key issues. Because the design is undefined, it is difficult to identify the key issues and thus select the appropriate CM staff early enough in the design process to make beneficial use of their input. Also, procuring the CM at such an early stage may result in reliance upon staff that is subsequently unavailable when the construction starts. Using the construction manager adds an extra layer of review that cause delay and confusion of responsibility, especially if the CM firm is also a designer.

•

The CM staff tends to be generalist in nature, and the industry expert who can provide the technical expertise to the specific design problem is typically an individual consultant, not part of a CM firm.

10 Copyright © 2004 Taylor & Francis Group plc, London, UK

Purpose/Use

Review panels are most effective when retained and used aggressively during the concept design phase. It is more difficult to change direction when panels are convened after the project design is well advanced. Consider having a two part process, consisting of an early board to address the overall

09096-01.qxd

•

•

•

•

•

8:13 PM

Page 11

•

approach, and a subsequent technical board with some construction experience to address technical and constructability issues. Define the purpose and scope of the board for each specific project. Decide whether it is to be “independent” or an “integral part” of the design team. Confirm the meeting frequency and use these meetings as milestones in the design schedule. Determine what output is required, and write the scope of work for the board to define all of the above. Make the individual consulting contracts compatible with the purpose. For “independent” boards, ensure that the board members have a separate contract with the owner agency, not through the design engineer. For boards that are expected to function as an “integral part” of the design team, contracting through the designer is acceptable and perhaps preferable from a risk and liability perspective. On major projects, use a standing Board of Consultants to achieve consistency across separate projects that are all part of a large program.

11.2

•

2/18/04

•

11.3

•

Member selection

•

Select consultants whose technical expertise match the specifics of the project. In order to achieve a balance of expertise, select members with a diversity of experience (i.e., contractors, other owner reps, peers, and academic folks). Include representatives from the end-user; e.g., maintenance and operations personnel. Include experienced constructability experts. Recently retired construction executives can be effective panel members. Formally interview prospective board members before appointment in order to determine availability, understanding of public projects, predetermined positions and whether the board would have an “open” mind to innovations. Identify potential conflicts of interest, which could include competing design firms, a prospective member with contractor clients who may be potential bidders, or previous representation of third parties who may be opposed to the project. The board should have no stake in the outcome of the project. If so it can lead to some questionable “advice” and conclusions. Appoint consultants who are supportive of each other, and in particular, ask “... if they are willing to work with specific other potential members” (Hoek & Imrie (1995)). If the designer is solely responsible for selection of the board members, there is anecdotal evidence that the same members appear repetitively on many different boards. Whether this practice is beneficial is subject to debate, however given an ongoing consultant-designer relationship, the owners’ interests may not be considered using this approach.

• • • • •

Operation

Plan for meetings sufficiently in advance so that they are well organized and to ensure that the board members have a good understanding of the project. If information is limited, the board may not be able to raise critical issues, and its effectiveness will be limited. “Failing to keep the Board advised of critical decisions or events” and “Meeting only when the project is in trouble and expecting the Board to somehow rectify the problem or protect the parties” are ways to misuse the Board (Shamma et al. (2003)). The owner and/or designer should make a brief presentation at the beginning of the meeting to establish ground rules and bring the board up to date on recently completed studies, investigations, etc. Ask the board to reply to specific needs on the project. Be specific about the type of review or recommendations requested. Allow time before the meeting for the board to review documents, and after the meeting to think about and document their recommendations. Develop the client/consultant relationship so that it is not only technical and professional, but also business-like. Develop a rapport between the owner, designer, and the board members. Do not permit the board to “direct” the design. This can happen with a very aggressive and assertive consultant on the board. Give consultants feedback on what worked and what did not. This allows for continuous improvement in the process. To avoid diminishing the Board’s effectiveness, the Owner should be careful not to tell the Board what the Owner wants to hear (Shamma et al. (2003))

12 CONCLUSION As one respondent aptly wrote, “Each large underground project has common elements and elements that are unique to the individual project. How one uses boards is a function of the project, funding sources, public and political involvement, and how to get the best thinking and advance it with real

11 Copyright © 2004 Taylor & Francis Group plc, London, UK

Consider member personality. “Personality is also critical, since an effective Board consists of individuals unafraid of stating their opinions but who, on the other hand, do not attempt to dominate with dogmatic or irrational behavior” (Hoek & Imrie (1995)). Avoid those whose history indicates a trend toward becoming the “savior of the project.” Pay particular attention to selection of a chairman, to organize and direct the board’s operations so that the board collaborates with the other members of the design team.

09096-01.qxd

2/18/04

8:13 PM

Page 12

News. (December 2002 and subsequent discussion: March 2003, June 2003, and September 2003.) Hoek, E., & Imrie, A.S., 1995. Guidelines to establish project consulting boards. International Water Power and Dam Construction. August 1995. Matyas, R.M., Mathews, A.A., Smith, R.J., & Sperry, P.E. 1996. Construction Dispute Review Board Manual, McGraw-Hill. Petroski, Henry 1996. Engineers of Dreams: Great Bridge Builders and the Spanning of America. Vintage Books. Shamma, J.E., Tempelis, D., & Nakamura, D. 2003. Board of Consultants – A Requirement for Hard Rock Tunneling Projects. Proceedings of the Rapid Excavation Tunneling Conference; Society for Mining, Metallurgy, and Exploration. Terzaghi, K. 1958. Consultants, Clients and Contractors. Journal of the Boston Society of Civil Engineers. January 1958. United States General Accounting Office, 2000. Mass Transit, Project Management Oversite Benefits and Future Funding Requirements, Report to Congressional Requesters.

world aspects.” By incorporating the advice and recommendations of expert consultants through the process of a formally established design review board, the design and construction of underground and heavycivil projects can be completed more effectively and with less project risk. However, establishing an efficient design review board, selecting the right members, and operating it successfully requires the active consideration of the purpose for which the boards’ recommendations are solicited. Owner agencies contemplating the use of a design review board should consider lessons learned from previous projects, and take an active part in the establishment, member selection, and operation of the consulting board. The efficient use of both consulting boards and construction managers can contribute positively to project success.

ACKNOWLEDGEMENTS APPENDIX – REPRESENTATIVE PROJECTS

The authors wish to acknowledge the following participants who have provided a significant amount of information or assistance in the development of this paper: Alistair Biggart Hugh Caspe Pete Douglass Herb Einstein Refik Elibay Joe Guertin Bill Hansmire Ray Henn Roger Ilsley Jon Kaneshiro Gregg Korbin Jim Lammie Jack Lemley Dan Meyer Lew Oriard Harvey Parker Ralph Peck Pete Petrofsky Tom Peyton Ed Plotkin Bill Quick Wolfgang Roth Tim Smirnoff Joe Sperry Fred Estep Kim Chan

Examples of projects, which have used Consulting Boards or Panels – some of these projects are in various stages of completion at this writing. The authors acknowledge that project names may not be accurate and are reported as provided in survey:

Ron Drake Paul Gilbert Joe Gildner Paul Gribbon Richard Harasick Geoff Hughes George Morschauser Priscilla Nelson Joe Pratt Martin Rubin John Shamma Lily Shraibati Ralph Tripani Al Wattson Lee Wimmer Howard Lum Tom Kuessel Judy Cochran Ed McSpedon Richard Proctor Gordon Revey Rube Samuels Gordon Smith Francis Fong John Ramage Birger Schmidt

Subways Tunnels Sound Transit Central Link LRT, Seattle, WA Bay Area Rapid Transit Project, (BART), CA Washington Metropolitan Area Transit Authority Los Angeles Metro Rail, CA Baltimore Metro, MD Eastside Access, New York, NY Buffalo LRT, Buffalo, NY Shepard Line, Toronto, Ontario, Canada San Diego LRT Extension, San Diego, CA Tri-Met Tunnels, Portland, OR Water and Sewer Tunnels City of L.A., Central Outfall Sewer Rehabilitation Chattahoochee Interceptor, Cobb County, GA Mercer Street Tunnel, Seattle, WA Narragansett Bay Comm., CSO, Providence, RI MWRA Inter-Island Tunnel, Boston, MA MetroWest Water Supply, Boston, MA Claremont Tunnel seismic upgrade, Oakland, CA North Dorchester CSO, Boston, MA East Central Interceptor Sewer (ECIS), L.A., CA North East Interceptor Sewer (NEIS), L.A., CA MWD, Inland Feeder System, Los Angeles, CA South Bay Ocean Outfall, San Diego, CA Milwaukee Water Pollution Abatement, WI Upper Diamond Fork Project, Provo, UT

REFERENCES Dunnicliff, J., & Parker, H.W., 2002. The Care and Feeding of Individual Consultants and Their Clients, Geotechnical

12 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-01.qxd

2/18/04

8:13 PM

Page 13

Glenwood Canyon Tunnel, CO Central Artery Tunnel, Boston, MA Doyle Drive, San Francisco, CA Whittier Access Tunnel, AK Malmo City Tunnel, Sweden Wolf Creek Pass, CO Cumberland Gap tunnels, TN and KY Pinglin Highway Tunnels, Taiwan

Stanley Canyon, CO Wasatch County Water Efficiency Project Nancy Creek Tunnel, Atlanta, GA Colombia Slough CSO, Portland BES, OR Brightwater Conveyance Tunnels, Seattle, WA Baumgartner Interceptor Tunnel, St. Louis, MO Detroit River Tunnel, MI SWOOP, San Francisco, CA

Other Underground Uses

Highway Tunnels

Superconducting Super Collider, Dallas, TX Positron Electron Project (PEP), Palo Alto, CA

Interstate H-3, HI Devils Slide Tunnel, CA

13 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-02.qxd

2/18/04

8:13 PM

Page 15

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Drawing from past experience to improve the management of future underground projects Chris Laughton Fermi National Accelerator Laboratory, Illinois, USA

ABSTRACT: The high-energy physics community is currently developing plans to build underground facilities as part of its continuing investigation into the fundamental nature of matter. The tunnels and caverns are being designed to house a new generation of particle accelerators and detectors. For these projects, the cost of constructing the underground facility will constitute a major portion of the total capital cost and project viability can be greatly enhanced by paying careful attention to design and construction practices. A review of recently completed underground physics facilities and related literature has been undertaken to identify some management principles that have proven successful in underground practice. Projects reviewed were constructed in the United States of America and Europe using both Design-Build and more traditional Engineer-Procure-Construct contract formats. Although the physics projects reviewed tend to place relatively strict tolerances on alignment, stability and dryness, their overall requirements are similar to those of other tunnels and it is hoped that some of the principles promoted in this paper will be of value to the owner of any underground project.

vertical shaft and numerous chambers and caverns up to 25 m in span. Sadly, the Superconducting Super Collider (SSC) project, the largest such project so far attempted, was terminated before tunnel construction was complete. This project, perhaps above all others referenced, stands as an excellent example of what can be achieved when good contracting practices tailored to underground construction are adopted. The physics community is now developing a new set of accelerator projects, including the Tera ElectronVolt Superconducting Linear Accelerator, the Next Linear Collider and the Very Large Hadron Collider. The scope of underground construction for these facilities will be larger than any so far undertaken. Rock tunnel housings as currently envisaged will range in length from approximately 50 to 250 km. In addition, a number of new proposals for detector-based underground experimental programs are being developed, notably relative to the study of beta and neutrino particles, at sites in Brazil, France, Japan, Russia and the USA. Effective management of underground design and construction is a critical focus of the planning process as these projects move forward. The goals of this planning are to deliver satisfactory facilities quickly at an affordable price (“better, faster, cheaper”).

1 INTRODUCTION Over the past twenty years the particle physics community has built a number of underground projects worldwide. Underground sites are preferred for many experiments because the groundmass overlying the facility acts to block the passage of particles and/ or radiation that could otherwise have a deleterious impact on the experiments and/or the surrounding environment. Underground accelerator-based projects constructed in this timeframe include the Super Proton Synchrotron, the Large Electron Positron and the Large Hadron Collider located at the European Particle Physics Laboratory, in Switzerland and France; various projects at the Deutches Elektronen-Synchrotron in Germany and the Stanford Linear AcceleratorCollider, Superconducting Super Collider Laboratory and Neutrinos at Main Injector (NuMI) projects in the USA. A number of underground detector sites have also been constructed in this same timeframe, notably including excavations made within existing mine boundaries at the Creighton, Homestake, Kamiokande, and Soudan mines or located adjacent to road tunnels within the Fréjus, Mont Blanc and Gran Sasso alpine massifs. The combined underground scope of these projects totals close to 100 km of tunnel, 10 km of

15 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-02.qxd

2/18/04

8:13 PM

Page 16

2 UNDERGROUND PLANNING

3 TUNNELING IS DIFFERENT

The main design and construction phases of a rock tunnel project are shown in the flowchart in Figure 1. The flowchart is based on that proposed by the International Tunneling Association and discussed by Lowe (1993). This chart outlines the basic steps in tunnel design and construction from alignment through construction. The flowchart omits reference to some key tasks, notably those associated with estimating and scheduling the work. However, the flowchart does provides a framework for the discussion that follows in which ten general principles are proposed to support an effective tunnel design and construction process.

Decisions made at the start of the project will have a great influence on project outcome. As far as a tunnel project is concerned, probably the most critical decisions that need to be addressed at the outset are related to preparing the owner for changes to his normal construction practices. The owner may need some convincing that “normal” business practices may not work so well underground. “First-time” tunnel owners, in particular, may see no particular benefit or need to change established ways of doing business and will need convincing that the changes are worth the effort, notably because

•

Site Investigation & Alignment

• •

Rock Mass Characterization

Of course, the underground project may go smoothly or encounter problems irrespective of whether an owner decides to take such precautions. However, such precautions are warranted in order to be responsive to the particular vagaries of the underground project. It will take more effort in the short-term, but will provide for more effective protection of the project over time. If the owner can be convinced of the value of these changes up front, the rest should be easy!

Excavation Methods & Means & Structural Elements

Detailed Design & Modeling Experience Estimation Bypass

Design & Safety Criteria Review

4 FAMILIARITY WITH LOCAL CONDITIONS An early understanding of the host rock mass conditions is a critical element in the design process. To evaluate a site’s suitability, regional and locationspecific geologic information will need to be gathered. Information should be collected on rock units, structural folds and faults, groundwater and in situ stress regimes. This geological information will need to be assimilated and interpreted at an early stage in design in order to characterize the rock mass along the alignment(s) and provide input for concept constructability and engineering analyses. Early acquisition and interpretation of this data is key in support of the design process. This data will help quickly eliminate showstopper situations and avoid much of the “wheel-spinning” (multiple layouts, designs and drafting work) that can occur during design and can consume a sizeable amount of a highly limited resource. At the earliest stage of design, shown in Figure 1, adequate site investigation data can generally be drawn

Accept/Reject

Risk Assessment & Contract Structure

Tunnel Construction

Field Observation

Stability

Figure 1. Tunnel design process flowchart.

16 Copyright © 2004 Taylor & Francis Group plc, London, UK

normal design and construction partner(s) may not be able to provide the types and breadth of support necessary for underground construction significant resources will need to be expended on site investigation and this work will need to start early the bid documents may need to be changed to address the added elements of risk that tunnel construction brings to contracting.

09096-02.qxd

2/18/04

8:13 PM

Page 17

from field visits and desk studies. In all but the remotest of areas, published matter can be found to support desk studies (e.g. topographic and geologic, land use mapping and related studies). The design team should also seek to supplement the public domain data sets with specific information on construction projects of a similar nature undertaken in the region. As underlined by Trautman and Kulhawy (1983) such information can most readily be tracked down with the help of a local “geo-practitioner” (geologist, engineering geologist, geological engineer). Such individuals will know where the data is and, more importantly, know how to access it. Their familiarity with local formations and involvement on other projects will prove invaluable to the team throughout design and construction. Every design team needs access to such a professional, particularly at the outset of the project when data acquisition and rock mass characterization skills are at a premium.

Rock Conditions Excavation Size Structural Behavior Excavation Shape

Excavation Support Initial Stresses

Figure 2. Factors influencing the structural behavior of a tunnel, after Sutcliffe et al. (1990).

with due regard to the constraints of the construction process results in a more practical design and ultimately provides for a more affordable and lower risk construction product. A more integrated design strategy that involves the contractor can also provide for a more innovative approach to tunneling (Songer and Molenaar, 1996) and help to lower risks associated with unreasonable end-user demands.

5 CONTRACTORS’ DESIGN INPUT By the time a basic rock characterization has been attained for a site, key underground end-user requirements will also need to have been established. These requirements will typically include a definition of the space and environmental needs of the operating systems as installed. In this regard, the physics end-user is likely to focus on issues such as foundation stability, dryness and alignment given that the success of their operations (accelerator and/or detector) will be highly dependent upon these aspects of the opening’s performance. However, before decisions are made and drawings developed defining alignment and crosssectional requirements, the end user should be made aware that some compromises might be needed if the facility is to be built economically. Absolutes in precision, stability and watertightness cannot be met easily in a natural, variable rock material and the needs of the experiment will need to be balanced against the practical constraints that the ground mass imposes. To reach the economic compromises discussed above, the requirements setter(s), the designer(s) and builders should ideally have an opportunity to discuss the factors that will impact tunnel behavior, as shown in Figure 2. Ironically, contractors, who undoubtedly have the best appreciation of the constraints of tunnel construction and are the ones who will ultimately price and build the facility, are often completely excluded from all stages of the design process. A way needs to be found, regardless of the contract format, to solicit the input of the tunnel builder in order to establish an understanding of the process and build-up confidence in the practicality of the design (Atkinson et al., 1997). A tunnel design developed

6 CASE HISTORY BENCHMARKING One basic question that needs to be addressed during design is that of precedent. Have similar tunnels been built before? And if they have, what was the outcome? Such questions usually emanate from the owner or their representative who are interested in understanding exactly what kind of situation they have gotten into! These are reasonable questions for which the owner should expect comprehensive answers. Underground projects with similar rock mass and construction methods and means should be researched and made available for the design team to review. Some papers and reports that have compiled tunnel project data bases include the United States National Committee on Tunneling Technology (USNCTT) (1984), Sinha (1986), Parkes (1988), the Association Française des Travaux en Souterrain (AFTES) (1994), and, Nelson et al. (1994). These databases are recommended as a resource for anyone seeking an objective evaluation of case histories, they describe mining performance and problems encountered over the length of the tunnel. In addition to the compiled data base material listed above, tunnel construction issues are often reported in a number of industry journals and in conference proceedings such as those of the Australian Tunneling Conference, International Tunneling Association, North American Tunneling Conference, Rapid Excavation and Tunneling Conference and Tunneling Symposium.

17 Copyright © 2004 Taylor & Francis Group plc, London, UK

Excavation Method

09096-02.qxd

2/18/04

8:13 PM

Page 18

Whatever the level of risk anticipated on the job it is important to find a mechanism that allows this risk to be objectively expressed and communicated to others. To manage such risks effectively the impacts of risk on cost and schedule are perhaps best expressed under a series of “what if ” scenarios. These scenarios are needed to complement the deterministic cost and schedule performance reporting systems and will serve to remind management that although underground problems are not shown as activities on the schedule the possibility of encountering them is real! Even the most thorough site investigation of the most uniform geologic conditions will not be able to completely define the scope of an underground construction contract. Some surprises from the natural material should always be anticipated along the way and an effort should be made to provide management with a clear expression of risk as an integral part of the normal reporting process.

The owner’s confidence in the viability of the “tunnel plan” will be improved if comparable case history data can be compiled and assimilated. The owner will be even more convinced if visits to similar sites can be organized. Examination of case studies also serves as a reality check on plans. A similar case whose outcomes are inconsistent with current projections may raise useful questions or may point to key parameters that differ between the projects. 7 INTEGRATED ENGINEERING In the title of their 1979 paper, Curtis and Rock frame the problem of working on structural linings underground as follows: “Tunnel Linings – Design?” This title is a simple acknowledgement that ground loading on a tunnel lining is difficult to predict even in the most homogeneous of groundmasses. This uncertainty can result in conservatism and/or complexity in design; for example, the use of thick cast-in-place linings to support an otherwise strong rock mass. The over-design of the final lining is difficult to avoid when loading conditions cannot be predicted with great certainty. Key to minimizing such overdesigns is a consideration of the ground’s ability to contribute to the long-term stability of the opening. To this end there is a need to better integrate the geotechnical engineer’s knowledge in to the structural engineer’s model. Such integration may allow greater opportunity for a discussion of the strengths of the rock mass and ultimately result in the streamlining or even elimination of a “permanent structure.”

9 CONTRACTING STRATEGIES Nowadays, design and build is commonly held to have distinct advantages over more traditional Engineering-Procurement-Construct contracting, but design and build will not always provide the best solution. Under the right circumstances, a design and build approach may save the owner time and money and offer the individual contractor the best opportunity to integrate the design needs of construction with their preferred methods and means. As Cording (1985) notes, “The separation of design and specifications from the contractor’s planning create unnecessary impediments and adds unnecessary costs to the project.” However, there are circumstances where the owner may wish to maintain greater active control of the underground project through its execution, notably where public interest is high and/or architectural features are an important part of the project. As pointedout by Boye and Eskensen (2003) the argument for design and build is weakened as public involvement in the permanent works design (geometry, layout, aesthetics) and complexity of the contract interfaces increases. As the needs for prescriptive language in design and construction is reduced, the case becomes stronger for leaving the contractor greater flexibility in his/her choice of methods and means within the framework of the design and build contract option.

8 RISK MANAGEMENT Risks associated with underground construction are notoriously difficult to describe and quantify and setting realistic expectations for scope, cost and schedule is always a major challenge. Risks underground are strongly influenced by a number of factors, including the diversity/complexity of the geology, the density of the site investigation coverage, the amount and relevance of compiled case history information, the flexibility of the selected mining methods and means and the skill-set of the construction team. Risk analyses should be performed at critical junctures during design and construction to ensure that risks are properly characterized. Risk analysis should be performed to identify the types of risk to which the project is exposed and provide for an estimate of their frequency of occurrence, and the severity of their impact, ultimately in terms of cost and schedule. Management should use such information to decide upon the type and extent of mitigation required for each type of risk event.

10 ORGANIZING FOR SUCCESS All of the issues discussed above, while important, are secondary when compared to the need for assembling and maintaining a good project team to manage the work. Care should be exercised in the selection of

18 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-02.qxd

2/18/04

8:13 PM

Page 19

the framework of discussion. Issues should be framed in such a way that participants are not asked to answer leading questions and attention should be made to ensure that individuals are not placed in positions where conflicts of interest might arise. The review process should encourage frank and open discussion between participants aimed at comprehensively addressing agenda topics and answering specific questions. Review outcomes should include a single attendee-reviewed document that faithfully records the topics discussed, findings and recommendations. Any review recommendations that require follow-up should be addressed and appropriate actions taken.

Department of Energy

M&O Contractor

Texas Commission

AE/CM Contractor(s)

Design Consultants

Technical Systems

Construction Contractors

Figure 3. Management organization for the SSCL, after USNCTT (1989).

12 LESSONS LEARNED Many of the decisions made during the course of a tunnel project are experience-driven. Despite improvements in rock mass modeling and the prediction of mining performance the industry is likely to remain heavily dependent on this “experience factor” for the foreseeable future. Within the industry there is an ongoing need to share and learn from our collective experiences, both good and bad. The industry, cannot afford to let every owner learn from his/her own mistakes. If past successes and failures go unreported opportunities for improved practices will be lost and the same common errors will continue to be repeated. A more concerted effort is needed to methodically analyze and openly discuss the underlying reasons for success and failure of tunnel jobs. Sharing these experiences would allow the tunneling protagonists the opportunity to get smarter more quickly and allow potential owners better insight in to the workings of the underground construction industry.

all team members whether searched and selected from in-house staff or out-sourced. At a minimum, candidate members should be expected to demonstrate a requisite level of individual and corporate competence, and work products should be provided that exemplify the candidate’s ability to fulfill projectspecific roles. Focus should be placed on judging the relevance of past experience (similar requirements, geology, methods and means, etc.). When there is inadequate expertise within the owner’s existing organization, responsibility for the management of the design and/or construction may be delegated, as shown in Figure 3. Here the SSCL Architect/Engineer and Construction Manager (AE/ CM) team was carefully selected following guidelines setout by the US National Committee on Tunneling Technology, Geotechnical Board (USNCTT, 1989). The selected AE/CM (Parsons Brinckerhoff and Morrison Knudsen) provided a dedicated team of experienced professionals to the SSCL project. The project was managed to cost and schedule up until its termination in the early 1990s.

13 CONCLUSIONS Digging a hole underground is not as simple as it sounds. Cost and risk are potentially much higher than they are for equivalent surface-based or cut and cover structures. Tunneling really does present the owner with a different set of construction challenges than he/she may be accustomed to dealing with. At the outset of the tunnel project, focus should be placed on educating the owner to the particular vagaries of the underground contract. As work commences attention should be paid to developing an early appreciation of the site in general and the rock mass in particular. During the design, focus should be placed on properly integrating the end-user and engineering needs of the facility with the construction preferences of the contractor.

11 THE VALUE OF REVIEWS Technical reviews are a common part of most large tunnel projects. They can be regarded as a distraction from the core project objectives, but if properly run they can provide valuable opportunities for improved communication and learning between project members and ultimately result in a better project. Reviews are most likely to be effective if the agenda is established ahead of time and if participants are invited based on their ability to address agenda items. In some instances, an individual may be nominated to play the role of “devils advocate” to encourage and broaden

19 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-02.qxd

2/18/04

8:13 PM

Page 20

Underground Transport, Future Developments in Technology, Economics, and Policy, Boston, MA, US., April, pp. 121–141. Lowe, P.T. (1993), “The Planning and Design of the Prospect to Pipehead Tunnel.” Proceedings, 8th Australian Tunneling Conference, Sydney, Australia, 24–26 August, pp. 21–27. National Research Council (1989), “Contracting Practices for the Underground Construction of the Superconducting Super Collider”, Washington DC, p. 99. Nelson, P.P., Al-Jalil, Y.A. and Laughton, C. (1994), “Tunnel Boring Machine Project Data Bases and Construction Simulation,” Geotechnical Engineering Center Report GR. 94-4 to the National Science Foundation, December. Parkes, D.B. (1988), “The Performance of Tunnel-Boring Machines in Rock,” Construction Industry Research and Information Association, Special Publication No. 62, p. 56. Songer, A.D. and Molenaar, K.R. (1996), “Selecting DesignBuild: Public and Private Sector Owner Attitudes.” Journal of Management in Engineering, November– December, pp. 47–53. Sutcliffe, M.L., Rogers, S.F., Whittaker, R.N. and Roberts, B.H. “Integrated Approach to Geotechnical Assessment of Rock Tunnel Stability and Performance.” Proceedings of Tunnel Construction ’90, London, UK, April 1990, pp. 145–153. Trautman, C.H. and Kulhawy, F.H. “Data sources for Engineering Geologic Studies.” Bulletin of Association of Engineering Geologists, Vol. XX, No. 4, pp. 439–454. US National Committee on Tunneling Technology (1984), “Geotechnical Site Investigations for Underground Projects.” National Research Council, Washington DC, National Academic Press.

For tunneling particular attention should be placed on establishing and updating expectations for costs and schedule performance. Regardless of the contracting strategies and the instruments chosen to mitigate and/or allocate risks, the owner will need to be regularly briefed on issues of project risk as tunnel projects are vulnerable to critical path delays. Reviews can be valuable tools for providing fresh technical and contractual insights to the management team. During construction, the contract will require active management in order to ensure that contract provisions are met and, that ground conditions are evaluated and timely decisions made as necessary. At the end of each tunnel job the process and outcome should be objectively reported so that any lessons learned can serve as a reference and guide for other owners and industry professionals alike. REFERENCES AFTES, Working Group No. 4 on Mechanization of the Excavation Process (1994), “Fiche Signalétiques de Chantiers Mécanisés, Recueil 94.” Atkinson, A., Cavilla, J. and Wells, J. “Securing the Contractor’s Contribution to Buildability in Design.” Project Report 27, CIRIA. London 1997. Boye, C. and Eskesen, S.D. (2003), “Specifying underground Works – the Challenge of Developing the Optimal Requirements.” Proceedings Underground Construction Conference, London September, 2003 pp. 509–520. Cording, E.J. (1985), “Constraints on Tunneling Technology,” Proceedings of the Conference on Tunneling and

20 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-03.qxd

2/18/04

8:14 PM

Page 21

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

The ECIS story J.W. Critchfield Parsons Brinckerhoff, Los Angeles, CA, USA

B. Miya Bureau of Engineering, City of Los Angeles, CA, USA

ABSTRACT: At $240 M, the East-Central Interceptor Sewer (ECIS) is the largest construction contract ever awarded by the Los Angeles Department of Public Works. Four Earth Pressure Balance (EPB) Tunnel Boring Machines, in an urban setting excavated the 18.4 km tunnel. Construction management issues that affected the completion cost and schedule involved construction access, design changes, permeation grouting, existing utilities, community issues, tunneling mishaps, and unforeseen conditions. These challenges were met and overcome by the combined efforts of the City and the Construction Contractor, to complete a vital infrastructure improvement, as mandated by the State of California.

1.2

1 INTRODUCTION 1.1

Ground conditions along the tunnel alignment include a thin surficial layer of fill overlying alluvial and marine sediments. There are three generally recognizable deposits, from the ground surface downward: Recent alluvium – inter-fingered layers of streamdeposited loose to dense silty and sandy soils with gravel, cobbles and boulders, including some local deposits of soft organic soils. Encountered in about 5 to 10% of the tunnel. Lakewood formation – alluvial and shallow marine deposits including layers, lenses and pockets of generally dense silty sands and sandy silts, with gravel, cobbles and boulders. Encountered in about 80% of the tunnel. San Pedro formation – deep marine deposits composed of hard silts and clays with zones of dense sand and gravel. Encountered in about 10 to 15% of the tunnel. Hydrogen sulfide and methane were encountered in these deposits. The tunnel alignment is within or near several oil fields. Contaminated soil and groundwater were encountered at some work sites. Several active and inactive faults are present along the tunnel alignment. The most significant is the Inglewood Fault, located in the vicinity of the Baldwin Hills, near the downstream end of the project. The regional groundwater table is generally 25 to 50 m below tunnel invert. However, water is above the

Project description

The North Outfall Sewer – East Central Interceptor Sewer Project (NOS-ECIS) is designed to divert wastewater from the middle portion of the existing 80 yearold NOS. This will provide increased capacity to handle wastewater flows and allow the NOS to be rehabilitated. The project is the first phase of an urgently needed program required to prevent sewer overflows during storm events. A Cease and Desist Order (CDO) mandated construction, under threat of heavy financial penalties, from the California Regional Water Quality Control Board. An 18.4 km-long tunnel was constructed from East Los Angeles, westward to Culver City, along the alignment shown in Figure 1. The depth to tunnel invert ranges from approximately from 10 to 30 m, with a maximum depth of about 110 m under the Blair Hills. The excavated diameter was 4.7 m. The finished inside diameter is 3.35 m. The alignment was divided into five tunnel drives. Other project elements include 7 shafts, 31 maintenance holes, 11 junction structures for future connections, 2 diversion structures, a 90 m-long siphon and a 250 m-long micro tunnel, and a connection to NORS and other sewer lines. Conduits for fiber optic cables are incorporated into the final tunnel lining.

21 Copyright © 2004 Taylor & Francis Group plc, London, UK

Ground conditions

09096-03.qxd

2/18/04

8:14 PM

Page 22

Figure 1. Project location.

tunnel crown at the east end of the project, beneath the Los Angeles River. Water is also perched within the various soil layers and trapped within fault zones. 1.3

Construction environment

The alignment is largely within public right-of-way, but does cross beneath several private properties and structures. The neighborhoods near the construction sites include residential and commercial areas with numerous schools, churches and other sensitive facilities. Accordingly, site access and work hours are restricted and limitations are imposed on construction noise and vibration. Hundreds of existing structures are present within the zone of influence above the tunnel excavation. The tunnel is directly under several buildings and major utilities, the Interstate 10 and 110 Freeways, the Los Angeles River Channel, the Metropolitan Transportation Authority (MTA) Blue Line, Union Pacific and Amtrak Railroads, and the new Alameda Corridor facilities. 1.4

Figure 2. Completed sewer pipe in tunnel.

tunnel lining consists of segmental pre-cast concrete rings, installed and grouted at the tail of machine. Sections of Precast Concrete Cylinder Pipe (PCCP), lined with polyvinyl chloride (PVC) as corrosion protection, were installed inside the tunnel to complete the sewer, as shown in Figure 2. Cellular concrete was used to fill the annulus between the carrier pipe and

Construction methods

EPB tunnel boring equipment was specified as the primary means to control ground loss during excavation, limit surface settlements, and prevent damaging existing structures. Four Lovat EPB tunnel boring machines (TBM) excavated five sections of tunnel. The initial

22 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-03.qxd

2/18/04

8:14 PM

Page 23

An Owner-Controlled Insurance Program (OCIP) and a Project Labor Agreement (PLA) were included in the construction contract. A local hiring goal was incorporated into the PLA.

the tunnel segments. Fiber optic conduits are embedded in the annular grout. Permeation grouting, with cement or chemicals, was specified to improve ground strength prior to excavation beneath critical structures. No compaction grouting or structural underpinning was expected to be necessary. A geotechnical instrumentation program was used to monitor ground movements around the tunnel and measure effects on existing structures. A preconstruction survey was also conducted to document the condition of existing structures. Conventional mining methods was also used. Staged excavation and support installation was needed to excavate beneath the siphon and to construct starter tunnels and breakouts for junction structures. The connection to the active NORS was specified to be accomplished almost entirely underground, due to work site restrictions at the downstream end of the project. Micro-tunneling technology was used to construct portions of the siphon and for a primary connecting sewer. Large shafts and junction structures were constructed within soldier pile and slurry wall shoring systems. Small diameter shafts and maintenance holes were be installed by drilling. 1.5

2 MANAGEMENT CHALLENGES Issues and events that affected the cost of construction and the schedule for completion are outlined below. 2.1

Property acquisition and construction access issues soon became apparent, once ECIS was underway. The City struggled to obtain dozens of underground easements, often from intransigent property owners. Shaft construction for the Siphon Inlet cut off driveways to four homes. Planned back alley access turned out to be impracticable and City real estate agents scrambled to arrange for construction of a temporary driveway across the private properties. Work areas and traffic control plans on city streets needed to be revised to accommodate larger than anticipated equipment. Arrangements for a construction site at the North Outfall Relief Sewer (NORS) connection in Culver City took over a year longer than expected, due to numerous third-party complications.

Construction contract

Bids were opened in November 2000. A low bid of $240 M was submitted by the Joint Venture of Kenny, Shea, Traylor & Frontier-Kemper. The Engineer had estimated construction costs at $255 M. The Board of Public Works established a construction budget of approximately $280 M, including the bid amount, plus $10 M for insurance and $30 M as a construction contingency. Funding is provided entirely by local sources. The Contract, awarded in January 2001, is the largest construction contract ever for the LA Board of Public Works. Notice-to-proceed was given in February 2001. The original contract duration was 1000 calendar days, giving a contract completion date in midNovember 2003, to meet a completion deadline imposed by the State of California. 1.6

2.2

Design changes

A major design change was necessary on the ECIS project. The original design called for cast-in-place concrete structures at the bottom of each maintenance hole, some of which have junction structures for future sewer connections. The Contractor used 3-D Computer modeling to demonstrate that cast-in-place construction was impracticable and proposed an alternative scheme using pre-fabricated concrete and steel pipe. Eight unneeded maintenance holes were deleted to offset extra costs. The configuration of a slurry wall shaft at the upstream end of the project had to modified to make it constructable on a small work area, immediately adjacent to operating railroads. Additional modifications were made to facilitate connection of the Northeast Interceptor Sewer (NEIS), by another contractor. The NORS connection work site needed extensive site development work. Additional temporary shoring was constructed that allowed recovery of a tunnel boring machine and the installation of carrier pipe, as part of a schedule recovery strategy. Odor control facilities were added to the contract after award. Additional chemical injection and air scrubbers were needed to control odor during construction of sewer connections. Interim carbon filters

Project management

Design work was performed by the City of Los Angeles Bureau of Engineering, with assistance from Parsons Brinckerhoff Quade & Douglas. Construction management duties were carried out by a combined team of Bureau of Engineering and Bureau of Contract Administration personnel, together with consultant staff led by a Joint Venture of Parsons Brinckerhoff Construction Services and Brown & Root Services.

23 Copyright © 2004 Taylor & Francis Group plc, London, UK

Construction access

09096-03.qxd

2/18/04

8:14 PM

Page 24

Figure 4. Maintenance hole/tunnel intersection.

to advertising the contract for construction. A oneblock long section of the large concrete box culvert had to be demolished and reconstructed at higher elevation. This entire effort was performed on a time and materials basis, at a total cost of $4.6 M, exceeding the contract allowance by $1.2 M. Relocation of underground utilities and overhead lines was a ubiquitous problem at maintenance hole sites. At one location, the presence of critical overhead electric and underground services combined with high-speed traffic control issues made it impracticable to use a large drill. The 3 m diameter maintenance hole had to be hand-excavated to a depth of 24 m.

Figure 3. Maintenance hole installation.

were added at three sites to serve until permanent Air Treatment Facilities can be constructed for the completed sewer system.

2.5

Community issues

The largest single extra cost was for permeation grouting work. Several unit price bid items were included in the contract, assuming that both cement and chemical grout would be injected from the surface and from within the tunnel. The contract also assigned responsibility for design of the grout program to a specialty subcontractor. The contractor design was completely different, including only chemical grout and working only from the surface. This resulted in some unused bid items, for cement and underground work, and massive quantity overruns on chemicals and surface work items. The total volume of grout used was 11,000,000 liters, roughly the same as the design. The total cost was $12.6 M, exceeding the bid price by $5.5 M. The engineer estimate had included $14 M for permeation grouting.

Respect for the community is of paramount importance to the Board of Public Works. Restrictions were placed on work hours, noise, vibration levels and traffic, to reduce the burden of construction activity on neighborhoods. Costs for mitigating construction noise, vibrations and disruption exceeded the contract allowance, mainly for the construction of additional noise barriers, and for additional traffic controls to maintain access to properties. Prevailing wages requirements of the Project Labor Agreement were misunderstood by the trucking subcontractor, in preparation of their bid. The City issued a $2.8 M change order to remedy the situation. The narrowly defined PLA local hiring goal of 40% proved impracticable. Using the PLA definition, local hiring peaked at 26%, but averaged 14% overall. The percentage of Los Angeles residents on the project was 30%.

2.4

2.6

2.3

Permeation grouting

Existing utilities

EPB tunneling was mandated to control the excavation, limit ground loss, and minimize surface settlement.

City surveyors discovered that a major LA County storm drain conflicted with the ECIS tunnel, just prior

24 Copyright © 2004 Taylor & Francis Group plc, London, UK

Tunneling mishaps

09096-03.qxd

2/18/04

8:14 PM

Page 25

Extra work was required at the NORS connection because the geometry of the existing structure was different than expected and the flow level within the operating sewer was higher due to operational changes since the design was completed. 3 PROJECT COMPLETION 3.1

The critical path of the schedule was directly effected early in the project. Archaeological work and the delay associated with real estate dealings for the temporary driveway at the siphon inlet added 99 calendar days to the contract duration. During the course of the work, additional delays began to accumulate. Major identifiable delays were associated with obtaining and developing the NORS work site and with repairing the tunnel mishaps. More insidious delays resulted from the cumulative and inter-related effects of design changes, and other extra work activities. In June 2003, the City and Contractor negotiated a “global schedule agreement”. Under terms of the agreement all schedule issues to-date were resolved by development of a revised schedule which will place ECIS in service by June 2004 and complete all contract work by August 2004. The agreement also provided for payment of $2.0 M and for release of a portion of funds retained from progress payments. Subsequently the City approached the State Board to request an extension of the CDO completion date. The State accepted the City’s explanation and justification for the delay based upon unforeseen conditions, and granted an extension consistent with the global agreement.

Figure 5. Chimney caused by ground loss during a tool change.

There was no more important technical objective on the project, even though the application for ECIS tunneling in dry sandy ground was somewhat unusual and controversial. The EPB tunneling machines were generally effective in achieving the goal of controlling ground loss and limiting surface settlements. However, lapses in application of the EPB techniques did result in over-excavation. The process of stopping the TBM to maintain the cutterhead sometimes resulted in unintended ground loss. Chimneys, as shown in Figure 5, formed above the TBM during a tool change stops along the Alameda Corridor, which had just opened to commercial service. An extensive program of Compaction Grouting was performed to repair the ground along a 300-meter section of the alignment. A short section of tunnel lining was deformed and some heaving of the railroad tracks resulted from the compaction grouting operations. 2.7

3.2

Construction cost

The approximate cost to complete the project, as estimated in November 2003, is summarized in Table 1, along with a breakdown of extra costs. At $259.2 M, the expected final completion cost exceeds the Engineer Estimate by 2% and is 8% more than the original bid price. The total cost is considered acceptable, at 96% of the originally assigned budget.

Unforeseen conditions

A “mono” (grinding stone) artifact was recovered during soldier pile drilling at the siphon inlet shaft. The State Historical Preservation Officer required the shaft excavation to proceed in lifts, under supervision of the Project Archeologist. An abandoned oil well casing was found within the work shaft at the upstream end of the project. A specialty contractor was needed to investigate and remove a portion of the old casing. The hand-excavated maintenance hole encountered two additional differing site conditions. An unmarked sewer line conflicted with the excavation and had to be relocated. Unexpected flowing ground conditions required dewatering and grouting work to complete the excavation.

3.3

Lessons learned

ECIS has been one of the most challenging projects ever undertaken by the Department of Public Works. The quality of the completed work is excellent. The City considers the project successful with respect to schedule and budget. Key lessons learned are summarized as follows: Construction access. Acquire easements and work sites prior to beginning construction. Allow

25 Copyright © 2004 Taylor & Francis Group plc, London, UK

Schedule

09096-03.qxd

2/18/04

8:14 PM

Page 26

Table 1. Cost to complete summary. Engineer estimate Original contract Bid Contingency funds (12%) Insurance Assigned budget

$255.0 M $240.3 M $29.7 M $10.0 M $280.0 M

Extra costs Design changes Construction access Permeation grouting Existing utilities Unforeseen conditions Community issues Tunnel mishaps Total extra cost Cost to complete

$2.0 M $1.8 M $5.5 M $1.6 M $2.9 M $3.2 M $1.9 M $18.9 M $259.2 M

Committee is helpful to elevate utility issues for resolution. Community issues. Provide adequate funding allowances to meet commitments to the community. Tunneling. EPB has been shown to be the tunneling method of choice and it will be specified on future projects in Los Angeles. Unforeseen conditions. Cooperation and perseverance between the Owner and Contractor, with help from the various stakeholders can overcome unforeseen conditions.

ACKNOWLEDGEMENTS The authors would like to acknowledge the contributions and support of the following individuals:

•

realistic space for construction equipment and operations. Design changes. Review designs thoroughly for constructability. Permeation grouting. Consider implications of bidding schemes to reduce the potential for unintended consequences. Existing utilities. Accurate as-built information about existing utilities is difficult to find. An Executive

• • •

26 Copyright © 2004 Taylor & Francis Group plc, London, UK

Commissioners Valerie Shaw and Ellen Stein, Los Angeles Board of Public Works. City Engineer Vitaly Troyan and Deputy City Engineer Tim Haug. Inspector of Public Works Stan Sysak and Bureau of Contract Administration Inspectors John Reamer, Chris Smith and George Stofila. Project Director Ted Budd, and Patrick Kenney, of the Kenny Shea Traylor Fontier-Kemper Joint Venture.

09096-04.qxd

2/18/04

8:15 PM

Page 27

Session 1, Track 2 Security of critical infrastructure and key national assets: use of underground space

Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-04.qxd

2/18/04

8:15 PM

Page 29

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Internal blasting and impacts to tunnels Wern-ping (Nick) Chen HNTB Corporation, Boston, MA, USA

ABSTRACT: During an external explosion event, the ground would shelter a nearby tunnel and the explosion impact to the tunnel may be minimum; however, if the explosion occurs in the tunnel, the implication from the explosion event may be enormous. Publications and researches for internal blast loading and its impact to tunnels are limited. The intents of this paper are to review general blast loading phenomenon and structure response to this type of loadings; to recommend load factors and dynamic material strength during a blast event for limit state design; to qualitatively estimate the impacts of internal blast loadings to tunnels; and to propose feasible hardening countermeasures and other security alternatives. Literature review was performed to explore current policies and design criteria for blast loadings on tunnels.

tunnels is seldom, it can happen. What are the sources that cause internal blast in a tunnel? It may be from trucks with chemical explosives, may be from the transportation of military vehicles, and may be from terrorist attacks. Next questions are: What is the special condition for a blast in a tunnel that is different from that of surface structures? What is the size of this blasting source? How is the blasting pressure determined? What is the response of tunnel structures from blast attack? What is the material behavior of structures in a blast event? What are exiting policies and criteria for design and prevention of blasting in tunnels? The purpose of this paper is to address these issues.

1 INTRODUCTION Traditionally, tunnels, as others underground structure, are very resilient to dynamic loadings, which include seismic events. The reason of this phenomenon is that once the ground is stabilized by tunnel supports during excavation, the redistribution of stresses in the ground occurs and eventually the ground reaches a new state of equilibrium and is self-supported. After this new state of equilibrium is reached, man-made ground supports are no longer needed in theory, since the inherent strength of the ground is mobilized. During an internal blasting event, with the composite effect from tunnel supports and the ground, the blasting impact to tunnel itself may be small. Local concrete spalling of tunnel lining is likely to occur, but the overall integrity of the tunnel remains. On the other hand, its damages to life and functional systems of the tunnel may be tremendous. This is especially critical for transportation tunnels. Consequences and damages to transportation tunnels from an internal blasting event can include:

• • • •

2 BLAST PHENOMENON A unique feature of blast loading in tunnels is its confinement. Overpressure builds up in tunnels, from a blast event, at different phases. First, the incident pressure reaches tunnel walls and generates reflected overpressure. Because of confinement, this reflected overpressure generates re-reflected overpressure. This process produces a series of blast waves of decaying amplitude. While this is happening, the second loading phase develops as the gaseous products of detonation independently causing a build-up of pressure, the gas pressure. The phenomenon is different from a free-air burst that is remote from any reflecting surface. The free-air blast is categorized as spherical airburst. In applying the spherical airburst to the hemispherical surface

Life safety issues from blasting overpressure, falling debris, fires, smokes, and flooding, Damages to transportation vehicles, Damages to ventilation systems, and Damages to tunnel structures, causing lining spalling and cracks, which may subsequently cause inundation to the tunnel if the tunnel is subaqueous and the inflow is excessive and can’t be stopped.

Tolerance for these consequences is null, but how do we detect and prevent it. Though blasting incident in

29 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-04.qxd

2/18/04

8:15 PM

Page 30

The most widely used blast wave scaling approach is developed independently by Hopkinson (1915) and Cranz (1926), which is commonly referred as the cubic root scaling.

burst, such as an explosive sitting directly on the ground or in a tunnel, an enhancement factor of 1.8 is generally recommended. The hemispherical blast pressure may also be obtained from blast chart based on tests. Figure 1 displays the pressure and time relationship after a blast, where negative pressure occurs after time t0. Figure 2 shows the Shock-reflection phenomenon in a region where , incident angle, is greater than 45° (Norris et al., 1959). 2.1

(d1/d2)  (W1/W2)1/3

(1)

(R1/R2)  (W1/W2)1/3

(2)

W1 and W2 are charge masses of charge diameters d1 and d2, respectively. Ranges at which a given overpressure is produced can thus be calculated from Equation (2), where R1 is the range at which a given overpressure is produced by W1 and R2 is the range at which the same overpressure is generated by W2. It is inferred that the charge weight is inversely proportional to the cubic of the standoff distance, R; therefore, the best way of mitigating a blast event is to increase the standoff distance. Spherical blast pressure, airburst, can be obtained from classic derivation or from blast curves generated by experimental testing such as those generated by the Departments of the Army, the Navy, and the Air Force (1990). Analytic results by Brode (1955) are listed below.

Blast loading

The industry standard to determine the magnitude of an explosive is in terms of its equivalent weight to TNT. Most explosive device used by terrorist attack in the US is a mixture of Ammonium Nitrate and Fuel Oil (ANFO), which is about 80% equivalency to TNT. Conversion factors for other explosives to TNT equivalent weights can be found in many other literatures, such as Conrath’s (1999).

Ps  6.7/(Z3)  1 bar for Ps  10 bar

(3)

Ps  0.975/(Z)  1.455/(Z2)  5.85/(Z3)  0.019 (4) bar for 0.1  Ps  10 bar Z is the scaled distance, given by Z  R/W1/3

Ps is the peak static overpressure. R is the distance from the center of a spherical charge in meters and W is the charge mass expressed in kilograms of TNT. Hemispherical blast pressure can be obtained by multiplying the results from the spherical results by the factor 1.8 or from blast curve generated by the Department of the Army, the Navy, and the Air Force (1990), as shown in Figure 3, where, Pr (psi) is the reflected overpressure, ta (ms) shock arrival time, td positive phase duration, Lw (ft) positive phase wave length U shock front velocity (ft/ms), and is and ir side-on and normally reflected impulse (psi-ms), respectively. Analytical hemispherical blast pressure can be derived based on works by Rankine and Hugoniot (1870) and Liepmann and Roshko (1957) as:

Figure 1. Pressure–time relationship after a blast.

Pr  2Ps [(7po  4Ps)/(7po  4Ps)]

(6)

where, po is the ambient air pressure. Published testing result for blasting pressure in confined space, such as tunnel, is not available. This blast pressure must consider shock wave re-reflection

Figure 2. Shock-reflection in a region where  is greater than 45° (Norris et al.).

30 Copyright © 2004 Taylor & Francis Group plc, London, UK

(5)

09096-04.qxd

2/18/04

8:15 PM

Page 31

•

phenomenon. Baker et al. (1983) has estimated this pressure as: PrT  1.75Pr1

(7)

PrT is the total peak pressure and Pr1 is the first reflected pressure as identified in Figure 3 or Equation (6).

2.2

2.3

•

Dynamic structure strength

The strain rate of a material will increase in a fast rate of loading condition. In such a condition, the mechanical properties of the material behave differently. Concrete and steel strengths are usually higher in a fast rate loading condition than at in a static loading condition. The factor by which the static stress is enhanced to calculate the dynamic stress is called the dynamic increase factor (DIF). Typical values are shown in Tables 1 to 3. Tables 1 to 3. Dynamic Increase Factor (DIF) for Design (Mays and Smith, 1995).

Structural response to blast loading

The positive duration, td, of a blast wave and the natural period of a structure determine the response characteristic of the structure. The structure shall be design in accordance with its response behavior as described below, where  is the natural frequency of the structure.

•

Dynamic response (0.4  td  40) – True dynamic loading only occurs when the plosive phase duration of a blast wave is equivalent to the natural period of a structure, and is seldom occurred in underground structures.

Quasi-static response (40  td ) – When the natural period of the structure is much less than the positive phase duration of a blast wave, the structure will be fully displaced before the decay of the blast load. Such loading is quasi-static or pressure loading condition, such as the gas pressure in a tunnel after blast. Impulse response (0.4  td ) – When the positive phase duration of a blast wave is much less than the natural period of a structure, the blast wave decays significantly before the structure has had time to respond. Most blast events in tunnels have this type of response, since tunnel structures, in combination with the ground, are massive.

Table 1. Concrete Type of stress

fdcu/fcu

Bending Shear Compression

1.25 1.00 1.15

Table 2. Reinforcing bars Type of stress

fdy/fy

fdu/fu

Bending Shear Compression

1.20 1.10 1.10

1.05 1.00 –

Table 3. Structural steel Type of stress

fdy/fy*

fdu/fu

Bending Shear Compression

1.20 1.20 1.10

1.05 1.05 –

* Minimum specified fy for grade 50 steel or less may be enhanced by the average strength increase factor of 1.10.

Figure 3.

where, fy – static yield stress, fdy – dynamic yield stress, fu – static tensile strength, fdu – dynamic tensile strength, fcu – static concrete compressive strength, and

Hemispherical blast curve (TM 5-1300, 1990).

31 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-04.qxd

2/18/04

8:15 PM

Page 32

Table 4. Threat parameters for tunnel design. Very High 2,000 lb

High 500 lb

Medium 100 lb

Low 50 lb

From 12,000 lb Truck

From 5,000 lb Truck

From 4,000 lb Car

From 4,000 lb Car

gas pressure build-up in blasting events; however, it does not protect personnel in the tunnel from injuries, since the injuries are directly associated with the initial overpressure and debris (or fragmentation if a bomb is cased). It is, therefore, concluded that detecting and preventing adverse explosives from entering tunnels is the primary countermeasure for blasting events in tunnels.

fdcu – dynamic concrete compressive strength. The modulus of elasticity for steel and concrete are insensitive to loading rates. Also, since blast loading is an ultimate event, its design load factor shall be set to unity. 2.4

4 POLICIES AND DESIGN CRITERIA An attempt was made to review existing policies and design criteria for tunnels under internal blasting events. Most of these documents are from military facility programs and policies triggered from tragic events between 1990s and earlier 2000s caused by terrorist attacks.

Threat parameters

Blast threat to tunnels is most likely from explosives or bombs in vehicles. Therefore, the maximum possible load from a blast will be a function to the size of the vehicles that carry the explosives. Table 4 is a recommendation for threat parameters for private-sector facilities (Conrath et al., 1999), which is suitable for tunnel design purpose as well. It also includes the likelihood of each threat occurrence and its associated vehicle size.

4.1

The section summarizes results from literature reviews of the latest available policies under blasting or terrorist attack events. These documents include:

•

3 COUNTERMEASURES Besides detecting and preventing blast threats from the outside of tunnels, physical countermeasures for tunnels under blasting events include:

• • • • • •

Structure hardening to improve structure resistance to blasting load, Provide shielding around tunnel, such as tube in tunnel and separating tunnels from direct exposure to blasting, Provide shielding around critical structural elements, including ventilation and fire fighting systems, Provide mechanisms in a tunnel to automatically detect and isolate blasting events and prevent their spreading (a blast proof automatic venting system would be required), Ground strengthen around tunnels by contact grouting and consolidation grouting, Provide external groundwater cut off mechanisms around, above, or in the tunnel by ground improvement, slurry walls, and internal automatic bulkheads.

•

•

Structural hardening is not cost-effective for blasting events in tunnels, since it is difficult and costly to design each element of a tunnel system to be blastproof. Even so, the life and safety issues for users in the tunnels can’t be guaranteed. For example, blast-proof automatic venting system is a countermeasure against

32 Copyright © 2004 Taylor & Francis Group plc, London, UK

Policies

Use of Underground Facilities to Protect Critical Infrastructures, Summary of a Workshop (Little et al., 1998) – It is a summary for workshop conducted to discuss the use of underground facilities for protection of critical infrastructures. This workshop discussed findings of the President’s Commission on Critical Infrastructure Protection (PCCIP) and key issues in going underground, but no policy issue was addressed. A Guide to Updating Highway Emergency Response Plans for Terrorist Incidents (AASHTO, 2002a) – This document addresses the existing state and DOT emergency management plans and practices, the standard view of the terrorist threat since 9/11, and a process guidance as to how state Departments of Transportation (DOTs) can update their emergency response plans. No specific policy is addressed in this document. A Guide to Highway Vulnerability Assessment for Critical Asset Identification and Protection (AASHTO, 2002b) – This guide was developed as a toll for state DOTs to (1) assess the vulnerabilities of their bridges, tunnels, roadways, and inspection and operation facilities, (2) develop countermeasures to deter, detect, and delay the consequences of threats, (3) estimate the capital and operating costs of such countermeasures, and (4) improve security operational planning for better protection against future acts of terrorism. This document addresses mostly surface structures. It does not mention specifics to tunnels or underground structures.

09096-04.qxd

•

2/18/04

8:15 PM

Page 33

Though not design criteria, several design manuals developed by federal agencies (mostly for military use) are helpful for general blast deigns. These manuals would be used to assist the development of specific criteria for blasting designs for civilian facilities and tunnels in the future. These manuals include (their distribution may be limited):

National Need Assessment for Ensuring Transportation Infrastructure Security (AASHTO, 2002c) – This document examines three key security planning program areas: (1) protecting critical mobility assets, (2) enhancing traffic management capabilities, and (3) improving state DOT emergency response capabilities. It estimates the total costs for the proposed initiatives, including capital investment and operations and maintenance expenses during the TEA-21 six-year reauthorization period. Annual cost for tunnel related security program and 54 critical tunnels are identified; however, emphasis of this document is still on bridges and surface facilities.

• •

From above reviews, it is clear that most of these documents provide guidelines and guidance in handling infrastructure security and threat identification and prevention, but not policies. Furthermore, most of these documents address on surface facilities, such as buildings, highways, and bridges. Document that directly addresses tunnel policies does not exist.

•

• 4.2

Design criteria

Civilian blasting design criteria for infrastructures does not exist either. Most design criteria for facilities are developed by US federal agencies for federal properties and most design manuals are derived by the US Department of Defense, Department of State, and General Services Administration for antiterrorism requirements for military, embassy, and federal facilities. The following sections review these criteria and design manuals. They could be used as guides in developing specific tunnel documents in the future. Most criteria that were reviewed are for federal surface facilities. Their applications to civilian infrastructures and tunnels are not direct and must be revised. These criteria include:

• •

• •

• •

GSA Security Criteria (GSA, 1997) – This document has been used for new facility designs and has been the basis of performance standards in retrofit analyses of existing buildings. ISC Security Design Criteria for New Federal Office Buildings and Major Modernization Projects, (ISC, 2001) – This document is fundamentally built from GSA Security Criteria. Its purpose is to adopt GSA criteria to suit all federal agencies. This document was review by NRC in 2003. Major comments by NRC are that though the intent of this document is performance based, its performance-based design process is unclear and explicit statement that mandates the use of the ISC criteria for all projects is missing. The intent of this document is clear, but its execution may be an issue since it is not mandated.

•

5 CONCLUSION Conclusions drawn from this paper are:

•

33 Copyright © 2004 Taylor & Francis Group plc, London, UK

Structures to Resist the Effects of Accidental Explosions (U.S. Departments of the Army, Navy, and Air Force, 1990). It is the mostly used publication by both military and civilian organizations. A Manual for the Prediction of Blast and Fragment Loadings on Structures, DOE/TIC-11268 (U.S. Department of Energy, 1992). This manual provides guidance for facilities subject to accidental explosions and aids in the assessment of the explosion-resistant capabilities of existing buildings. Protective Construction Design Manual, ESL-TR87-57 (Air Force Engineering and Services Center, 1989). This manual provides procedures for the analysis and design of protective structures exposed to the effects of conventional (non-nuclear) weapons. Fundamentals of Protective Design for Conventional Weapons, TM 5-855-1 (U.S. Department of the Army, 1986). This manual provides procedures for the design and analysis of protective structures subjected to the effects of conventional weapons. Design of Structures to Resist Nuclear Weapons Effects, Manual 42 (ASCE, 1985). This manual was prepared for civilian use, and has been widely distributed throughout the world. The Design and Analysis of Hardened Structures to Conventional Weapons Effects (DAHS CWE) (DNA, 1995). This new Joint Services manual, written by a team of more than 200 experts in conventional weapons and protective structures engineering. Security Engineering, TM 5-853 (U.S. Department of the Army, 1993). Terrorist Vehicle Bomb Survivability Manual (Naval Civil Engineering Laboratory, 1988). This manual contains information on vehicle barriers and blast survivability for buildings. Structural Design for Physical Security – State of the Practice Report (ASCE, 1995). This report is intended to be a comprehensive guide for civilian designers and planners who wish to incorporate physical security considerations into their designs or building retrofit efforts.

Blast wave propagation in tunnels is complicated. Blast pressure for design must take into

09096-04.qxd

• • • • • • •

2/18/04

8:15 PM

Page 34

consideration the shock wave re-reflection phenomenon. No blast testing result for tunnels is available. This paper presents a simplified procedure for blasting pressure in tunnels and provides recommended blast loads for tunnel designs. This paper provides countermeasures for tunnels under blast events; however, the best defense for blast events is early detecting and preventing adverse explosives into tunnels. Mandate security policy for civilian infrastructures does not exist in the US. Unified security policy for tunnels does not exist. It varies from state to state and from tunnel to tunnel. Mandate civilian blast design criteria for infrastructures do not exist. Blast design criteria and manual for tunnels does not exist. Blasting events in tunnels can happen. The need to address issues, polices, and design criteria for blasting in tunnels are immediate.

Baker, W.E., Cox, P.A., Westine, P.S., Kulesz, J.J. and Strehlow, R.A. (1983) “Explosion Hazards and Evaluation,” Elsevier. Brode, H.L. (1955) “Numerical Solution of Spherical Blast Waves,” J. App. Phys., Mo. 6, June. Cranz, C. (1926) “Lehrbuch Der Ballistik,” Springer, Berlin. Conrath, E.J., Krauthammer, T., Marchand, K.A. and Mlakar, P.F. (1999) “Structural Design for Physical Security – State of the Practice,” American Society of Civil Engineers. Department of the Army, the Navy, and the Air Force (1999) “Structures to Resist the Effects of Accidental Explosions,” Revision 1 (Department of the Army Technical Manual TM 5-1300, Department of the Navy Publication NAVFAC P-397, Department of the Air Force manual AFM 88-22), November. General Service Administration (1997) “GSA Security Criteria,” October. Hopkinson, B. (1915) British Ordance Board Minutes 13565. Interagency Security Committee (2001) “ISC Security Design Criteria for New Federal Office Buildings and Major Modernization Projects,” May. Little, R.G., Pattak, P.B. and Schroeder, W.A. (1998) “Use of Underground Facilities to Protect Critical Infrastructures, Summary of a Workshop,” National Academy Press. Mays, G.C. and Smith, P.D. (1995) “Blast Effects on Buildings,” Thomas Telford. Norris, C.H., Hansen, R.J., Holley, M.J., Biggs, J.M., Namyet, S. and Minami, J.K. (1959) “Structural Design for Dynamic loads,” McGraw-Hill Company, Inc. National Research Council (2003) “ISC Security Design Criteria for New Federal Office Buildings and Major Modernization Projects – A Review and Commentary,” The National Academies Press. Rankine, W.J.H. Phil. (1870) Trans, Roy, Soc., 160, pp 277–288.

REFERENCES American Association of State Highway and Transportation Office, in cooperation with the Federal Highway Administration (2002a) “A Guide to Updating Highway Emergency Response Plans for Terrorist Incidents,” May. American Association of State Highway and Transportation Office, in cooperation with the Federal Highway Administration (2002b) “A Guide to Highway Vulnerability Assessment for Critical Asset Identification and Protection,” May. American Association of State Highway and Transportation Office (2002c) “National Need Assessment for Ensuring Transportation Infrastructure Security,” October.

34 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-05.qxd

2/18/04

8:16 PM

Page 35

Session 1, Track 3 Mechanized tunneling

Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-05.qxd

2/18/04

8:16 PM

Page 37

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Improvements of the capabilities of cutting tools and cutting systems R. Bauer VOEST-Alpine Mining Tunneling & Construction

ABSTRACT: Newest improvements of the capabilities of cutting tools and cutting systems for hard rock conditions within the Icutroc research project. Underground excavations for infrastructural development and extraction of minerals in urban areas is increasingly becoming a necessity. Wherever conditions and circumstances permit mechanical rock excavation methods such as roadheaders, drum miners or tunnel boring machines are used. However in the past, certain situations (e.g highly abrasive material and/or high strength of the material to be cut) precluded the usage of certain mechanical excavation methods such as roadheaders. In such cases drill and blast was the only economical and practical alternative. With the development of Icutroc, an exiting new opportunity for cutting rock that provides numerous practical, logistical, environmental, and safety benefits, was introduced and is on the edge of making a big impact on the construction and mining world in the US. The continuous, mechanical cutting process provides an excellent opportunity for automatization with a high potential for various cost reductions. Furthermore it is very often the only viable solution in urban, congested areas where drill and blast is restricted or prohibited.

industrial initiator. Further partners were two customers (Thyssen Schachtbau and Somincor) as well as three research institutes (Seibersdorf Research Institute, Vienna, Armines CGES, Paris and the mining engineering department of the Montanuniversity Leoben, Austria)

1 PROJECT OBJECTIVE What does Icutroc mean and what implications will it have for the future of the North American tunnel and construction world? Icutroc is a corporate research and development project that was partly funded by the European community. Its original main target was to develop the necessary cutting tools to be able to apply higher cutting forces to economically cut higher rock strengths. When VOEST Alpine Bergtechnik, situated in Zeltweg, Austria took the initiative in 1995 to start a research and development program under the acronym “Icutroc” the goal was to extend the range of economic applications for the existing roadheader lines by moving into territories of harder and more abrasive rock types. The project objective was to achieve the required target with a type of roadheader that does not exceed 130 metric tons of operating weight and to stay within a range of 300 kW installed power on the cutterhead. The reason for these premises were that the maneuverability of the machine shouldn’t be sacrificed neither should the investment cost for this roadheader type exceed an acceptable, economic range. 1.1

1.2

Research partnership

The research program incorporated VOEST Alpine Bergtechnik and Sandvik Rock Tools as the main

37 Copyright © 2004 Taylor & Francis Group plc, London, UK

The Icutroc research approach

The Icutroc research approach was characterized by a combined development work addressing the necessary improvements of the cutting system and the machine system. Additionally Icutroc aimed to significantly improve the mechanical and wear properties of the cutting tools. In detail the whole project included: Development and refining of cutting systems and processes followed by simulations of their real world behavior by using computer-aided modeling and FEM-calculations. Better understanding of the interaction between rock/rock mass behavior and its influence on the cutting process. New concepts and material designs of cutting tools and new production technologies to manufacture these tools. Laboratory testing of these units. Testing and optimization of the newly developed system by civil engineering and mining end users under practical conditions.

09096-05.qxd

2/18/04

8:16 PM

Page 38

Figure 1. Icutroc’s project target for the development of a hard rock roadheader.

techniques in order to harmonize the factors of the cutting process with the characteristics of the cutting machine. Thereby taking into consideration the rock properties, geometric parameters of the cutting unit and the operating characteristics of the cutting system (cutting speed, sump in depth and machine stiffness). A finite element model of the complete machine has been set up for the simulation of the elastic behavior of the system. Concepts to reduce the elasticity of the cutting system were investigated to meet the required stiffness parameters for hard rock cutting. A substantial improvement of the overall system stiffness for the cutting action could be achieved by the boom stabilization system acting on the hydraulic boom cylinders. The boom stabilization system allows for a better adherence to the preset cutting depth as well as an improved compliance with a uniform swivel process. Together with the added benefit of reduced vibrations (the boom stabilization system also reduces the “bouncing” of the boom significantly) the shorter overall path length of picks further improves pick life. Development of a new rock mass rating specifically adapted for the new generation of roadheader technology, which increased the quality and reliability of performance prediction tremendously. In order to gain a revised RMR, two approaches were used first the theoretical net cutting rate based on cuttability of intact rock thereby reflecting the machine characteristics. Second the effective net cutting rate directly measured on site reflecting the actual operating conditions. Outcomes of this investigation were an exceptional correlation between NCReff/NCRtheor and the

Depending on the toughness and abrasivity, economically cutting of rock hardness up to 200 Mpa. Provide all conceptual prerequisites to implement updated control and data logging facilities, as they are required due to project conditions. The economic and environmental significance of this project is emphasized by the fact that the research work was funded by the European union within the Brite-Euram III program for Industrial and Material Technologies, managed by the European CommissionDG XII. 1.3

Development of a new cutting process

Using the newly developed VOEST Alpine cutterhead design software accounting for parameters such as optimized forces, cutting depth, cutting distances, slew and feed speed of the cutter boom, cutter head diameter, geological parameter, etc. the research project was able to design lacing layouts that resulted in the highest possible cutting efficiency. These cutting systems employing low speed cutting and providing greater power at the cutterhead in connection with the development of a new generation of cutting tools, were able to cope with the higher forces to be expected when cutting rock above 130 Mpa.

2 DEVELOPMENT OF AN ADAPTED MACHINE SYSTEM The research project included intensive modeling of the complete system using computer simulation

38 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-05.qxd

2/18/04

8:16 PM

Page 39

Figure 2. Specifically for Icutroc developed cutter heads designed with VOEST Alpine’s proprietary software. Rating of uniaxial compressive strength

Rating of block size Block size [m3]

Rating 20

UCS [MPa]

Rating

>0,6

1–5

15

0,3–0,6

16

5–25

12

0,1–0,3

10

25–50

7

0,06–0,1

8

50–100

4

0,03–0,06

5

100–200 >200

2 1

0,01–0,03 <0,01

3 1

Rating of joint conditions Surface

Aperture

Wall/Fill

Rating of orientation of joint set Rating

Influence on cuttability

Rating

rough

closed

hard, dry

30

very favorable

ⴚ12

slightly rough

<1 mm

hard, dry

20

favorable

ⴚ10

slightly rough

<1 mm

soft, dry

10

fair (and if block size < 0,03m3)

ⴚ5

smooth

1–5 mm

soft, damp

5

unfavorable

ⴚ3

very smooth

>5 mm

soft, damp to wet

0

very unfavorable

0

Figure 3. Revised rock mass rating approach for Icutroc roadheader technology.

RMRrev. No significant difference for different rock types and different roadheader types, and most important the higher effect of parting systems on cutting performance at low cutting speed.

The ratio between net cutting rate and effective net cutting rate was used as a measurement for the increase of performance including the influence of rock mass conditions.

39 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-05.qxd

2/18/04

8:16 PM

Page 40

10 Erzberg Premadio

9

Stillwater Athens

Pozzano

NCReff / NCRtheor

8 NCReff/NCRtheor = 46,537RMRrev-0,9877 R2 = 0,9072

7 6 5 4 3 2 1 0 0

10

20

30 RMR (revised)

40

50

60

Figure 4. Evaluation of rock mass influence on roadheader performance (slow cutting speed).

10 Balsareny Bogdanka Saudi Arabia Blumenthal

9

NCReff / NCRtheor

8 7

Wujek Prosper (AM 105) Prosper (AM 85)

Borynia Erzberg Szieroszowice

6 5

NCReff/NCRtheor = 9,4302RMRrev-0,5614 R2 = 0,8256

4 3 2 1 0 0

10

20

30 RMR (revised)

40

50

60

Figure 5. Evaluation of rock mass influence on roadheader performance (high cutting speed).

9

Trendline for High Cutting Speed (~3m/s)

NCReff/NCRtheor

8 7 6

Trendline for Low Cutting Speed (~1.4m/s)

5 4 3 2 1 0 0

10

20

30 RMRrev

40

Figure 6. Comparison between high (3 m/sec) and low (1.4 m/sec) cutting speed.

40 Copyright © 2004 Taylor & Francis Group plc, London, UK

50

60

09096-05.qxd

2/18/04

8:16 PM

Page 41

considerable less developed rock mass features (higher RMRrev) to the cutting process. It gets evident that low cutting speed provides more time to activate parting systems, whereby the higher available pick forces lead to an increased influence zone ahead of the pick tip. With increasing values for RMRrev the encountered difference in influence decreases.

After evaluation and utilization of the gained data from the ratio NCReff/NCRtheor VOEST Alpine was able to either modify existing correlation (e.g. RQD, RMR) or to establish a new rock mass classification system adapted for the new mechanized cutting systems. Why do we see such a tremendous effect between low cutting speed and increased cutting performance? First slow cutting speed means more available time for an activation and subsequent response to the existing parting system. Second higher pick forces resulting from a redesigned gearbox causing an extended stress zone in front of the pick crushing area. The third reason is improved design parameters for the cutter head (e.g. increased pick spacing assists activation of parting systems). Fourth there is a strong relationship between heat generation and cutting speed leading to rapid weakening and failure of the temperature sensitive tungsten carbide pick whereas low speed cutting and the improved cooling system reduce the heat generation significantly. With the chosen ordinate NCReff : NCRtheor the application of these diagrams for rock-mass-related performance prediction is provided. Of significant practical importance is the finding that a “stiff ” machine system applying low cutting speed (1.4 m/s) adds significant more rock mass contribution at

3 DEVELOPMENT OF A NEW COOLING AND DUST SUPPRESSION SYSTEM Heat generation on the pick tip was found to be the second biggest reason (second only to the results of extensive impact energy) for pick failure during the cutting process. In order to overcome this problem a new pick cooling system ensuring optimal cooling at acceptable water flow rates had been developed (the amount of necessary water could be reduced by approximately 25%). In combination with an intermittent sector controlled spraying system this innovation reduced the specific wear and tear on picks by 50%. 3.1

Cutting groove spraying system (VAB “Wet Head”)

A high-pressure water jet is pointed directly to the zone of the maximum dust creation. This system

Figure 7. The new pick cooling system results in a significant reduced wear and tear compared with any existing system. (Pick wear pattern during the cutting of sandstone with 160 Mpa and a Cerchar abrasivity  2,8).

41 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-05.qxd

2/18/04

8:16 PM

Page 42

Figure 8. The newly developed cutting groove spraying system guarantees an optimal dust reduction and heat suppression.

we are now able to compose the structure of cobalt tungsten, cobalt grains and cobalt perfectly – brick by brick – to a desired composition, grain size distribution and morphology which results in a pick without porosity or any other imperfections. Accompanied by FEM calculations, numerous lab tests and research on wear mechanism and thermophysical properties, the life expectance of the new cutting tools over former grades was more than doubled. Various real world site tests confirmed the results from theory and test rig.

results in a drastically reduced dust emission due to dust suppression directly at the source. Areas where the intermitted cutting trace spraying system is most beneficial are:

• • • •

Optimized respirable dust reduction due to the suppression at the dust source Minimum water consumption Minimal tool consumption due to efficient pick cooling Prevention of possible explosion due to cutting trace cooling to values below eventual ignition temperature.

Tests performed on a roadheader in underground conditions confirmed the effectiveness of the cutting groove spraying systems. The total dust concentration measured one meter behind the cutter head was found to be reduced by more then 50% with the spraying system activated during the cutting process. The total dust concentration at about 4 m behind the face was reduced from 143,6 mg/m3 at the reference system to 14,6 mg/m3 with the new, low speed cutting. 3.2

4 SUMMARY OF THE ICUTROC PROJECT AND ITS PRACTICAL IMPLICATIONS The most important outcome of Icutroc was the realization of a much more efficient and effective cutting process. Figure 10 shows the energy requirements of an Icutroc machine per unit (m3) of rock cut in comparison to an optimized existing system. It is clearly evident that Icutroc shows a significant reduction of the required specific energy using the new, low speed cutting process. The combination of higher pick forces at a low rotation speed and the stabilized and controlled guiding of the cutter arm resulted in a more economic

Development of new cutting tools

A complete new generation of cemented carbide grades – the so called “S-grades” – have been developed and patented by Sandvik. The concept means that

42 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-05.qxd

2/18/04

8:16 PM

Page 43

Figure 9. Newly developed cutting tools.

A combination of these outcomes was integrated into the latest generation of roadheaders within the product development project “Icutroc”.

handling of much harder material and a significant expansion of the expected cutting performance. Furthermore the low cutting speed reduced the tool consumption dramatically and led thereby to a more economic and efficient overall performance. Lower cutting speed resulted in significant lower temperatures at the pick tip, which reduced the weakening of the heat sensitive tungsten carbide pick tip, and increased pick lifetime accordingly. Another positive aspect of the lower cutting speed was the decreased impact impulse that does not exceed the mechanical strength of the pick tip anymore.

5 EXAMPLE OF AN ICUTROC APPLICATION IN THE REAL WORLD 5.1

Two VOEST Alpine ATM 105-IC roadheader extent the current line 2 from Henry Bourassa station to

43 Copyright © 2004 Taylor & Francis Group plc, London, UK

Operating results at the Metro Montreal

09096-05.qxd

2/18/04

8:16 PM

14 kWh/m3

Page 44

Approx. tool consumption (DP 8080)

6 kWh/m3

16 1 pick/m3

0.25 pick/m3

Speed energy kWh/m3

14 1

12

0.9

10

0.8

8

0.7 0.6

6

0.5

4

0.4

2

0.3 0.2

0 Reference system

New system

0.1 0 Reference system

Figure 10. Energy requirements per unit of cut rock.

Figure 12. Comparison of tool consumption between Icutroc and a traditional reference system, pick consumption was reduced to one fourth of the reference system (Material of about 160 Mpa and a cerchar abrasivity of 2.5).

Approx. cutting rate (incl. sump in and shearing) 28 m3/h

New system

35–50 m3/h

50

The material to be cut is predominantly limestone with an unconfined compressive strength between 46 Mpa and 140 Mpa. The average unconfined compressive strength over the whole range of material to be cut is 120 Mpa. The thickness of the fossiliferous and crystalline limestone as well as the shale insertions has varied within a wide range. Shale layers ranging from 0.1 inches to about 12 inches can practical be disregarded because of the scarcity of their occurrence. Additional the project faces massive limestone with very low to no bedding planes. Occasionally there are diabase dykes, with a diameter of up to 2 m and a hardness of 300 Mpa that need to be excavated as well. According to VOEST Alpine’s developed system a RMR of 30 was calculated which would enhance the net cutting rate calculated from the unconfined compressive strength test by a factor of 1,6 leading to an average cutting performance of 35 m3 and an average pick consumption of 0,1 pick/m3. Reality during one year of excavation proved even these very ambitious performance predictions a great underestimation. The red line indicates that the actual net cutting rate was ranging between 30 and 65 m3 per net cutting hour. That results in an average of 43,8 m3 per cutting hour, which is about 10–15% higher than predicted. The green line represents the pick consumption per m3 excavated rock, it can clearly be seen how

45 Cutting rate m3/h

40 35 30 25 20 15 10 5 0 Reference system

New system

Figure 11. Comparison of cutting rates between Icutroc and a reference system without the new technology (Material of about 160 Mpa and a cerchar abrasivity of 2.5).

Laval. The extension is 5.2 km and will be partially excavated by open cast. The last part of the metro extension will be the most challenging part of this metro excavation. This part involves the underground excavation of approximately 600 m of double lane tunnel running under a River, with an intersection to the single line tunnel. At the intersection area a maximum open span of over 16 m with a minimum 9 m rock coverage to the riverbed at a 34 m water head will require a sequential tunnel advance.

44 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-05.qxd

2/18/04

Sample #

8:16 PM

Page 45

BTS [MPa]

UCS [MPa]

Rock Type

CAI

from

to

mean

from

to

mean

from

to

mean

2001/086/01

Shale; C02 (TF-33-01, 88'3''– 90')

45,43

71,01

63,18

3,35

6,75

4,92

0,36

0,94

0,65

2001/086/02

Limestone – shaly, fossiliferous; C01 (TF-26-01, 82'2''– 84'3'')

63,55 134,78

102,26

4,81

10,89

6,81

0,55

1,15

0,82

2001/086/03

Limestone – slightly shaly, crystalline; C04 (TF-29-01, 96'8''– 98'6'')

88,09 117,05

105,87

5,11

9,81

7,50

0,29

0,84

0,62

2002/003/01

Limestone – fossiliferous, fine grained; C01 (TF-50, 33'5''– 34'3'')

93,12 117,10

102,14

4,58

9,28

7,37

0,44

0,75

0,63

2002/003/02

Limestone – crystalline, shaly; C01 (PF-2, 63'2''– 65')

58,08

92,84

78,25

4,12

5,49

4,98

0,59

0,82

0,68

2002/069/01

Diabase – dyke; C05 (TF-67, 19.25 –20.85m)

272,83 346,05

300,61

10,75

13,16

12,06

1,44

1,73

1,55

2002/069/02

Limestone – crystalline; C05 (TF-67, 57.0 – 58.2m)

63,77

82,62

70,20

3,31

7,91

5,92

0,68

1,00

0,82

0,45 0,40 0,35 0,30 0,25 0,20 0,15 0,10 0,05

Date

Daily NCR

Currently Average NCR

Daily SPC

Figure 14. Typical average cutting performance during one year at the job site.

45 Copyright © 2004 Taylor & Francis Group plc, London, UK

Currently Average SPC

30.04.03

27.04.03

24.04.03

21.04.03

18.04.03

15.04.03

12.04.03

09.04.03

06.04.03

03.04.03

31.03.03

28.03.03

25.03.03

22.03.03

19.03.03

16.03.03

13.03.03

10.03.03

07.03.03

04.03.03

01.03.03

26.02.03

23.02.03

20.02.03

17.02.03

14.02.03

11.02.03

08.02.03

05.02.03

02.02.03

30.01.03

27.01.03

24.01.03

21.01.03

18.01.03

0,00

Specific Pick Consumption [picks/solid m3]

95 90 85 80 75 70 65 60 55 50 45 40 35 30 25 20 15 10 5 0 15.01.03

Net Cutting Rate [solid m3/h]

Figure 13. Summary of the rock test results at Montreal.

09096-05.qxd

2/18/04

8:16 PM

Page 46

the occurrence of diabase influenced the pick consumption. On April 21st one dyke reached up to 2 m on one part of the face, causing a pick consumption of 0,6 picks/m3 during that day. However, in general the pick consumption arrived at an average of 0,064 picks per m3, which is about 30% below the predicted values. In terms of actual daily productivity the machine was averaging 434 m3 per 2 times 10-hour shifts, which translates to nearly 10 m of daily face advance. The best daily advance was close to 16 m. The biggest advantages of the mechanical excavation in Montreal are the positive environmental impact

and the compliance with a nearly perfect profile by a largely avoidance of any over cut. While drill and blast had ongoing vibration problems exceeding vibration limits with hundreds of complaints filed by the community the roadheader operator had no complains at all. It was also recognised that the mechanical excavation does not damage the immediate rock integrity, it results in ground stability improvements and does not open vertical joint systems for ground water inflow. Because of the delicate excavation under the river, close to an existing station in operation, the project owner actually demanded the use of a roadheader for any further excavation work.

46 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-06.qxd

2/18/04

8:17 PM

Page 47

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

MTBM and small TBM experience with boulders S.W. Hunt & F.M. Mazhar MWH

ABSTRACT: Boulders present a significant and potentially costly challenge when mining small (less than 3 m) diameter tunnels. This paper provides an overview of methods generally used to cope with boulders and summarizes experiences with boulders during forty tunneling cases. The results indicate that microtunneling with drag cutters and no face access significantly increases the risk of a stuck machine. The addition of roller cutters or tunneling with face access reduces the risk of getting stuck. Use of multiple features for handling boulders result in the least risk of getting stuck. When specifying or selecting tunnel boring machine features, the cost and schedule consequences of getting stuck should always be considered.

The results of a thorough desk study should provide the framework for the next step in boulder characterization: site investigation.

1 BOULDER CHALLENGES 1.1

Overview

Boulder occurrence is an extremely important consideration for design and construction of pipelines with microtunnel boring machines (MTBMs) and small diameter tunnel boring machines (TBMs). When large boulders or abundant quantities are anticipated, a rugged MTBM or TBM features such as face access, roller cutters, and a durable cutting wheel are keys to reducing the risk of getting stuck or experiencing expensive problems. 1.2

1.2.2 Site investigation Site investigation for tunnel projects with a boulder risk should be phased and focused. Unless substantial previous site investigation data is available, the site investigation should not be a single phase of routinely spaced borings. Instead, it should be phased with each phase designed to reduce uncertainties determined from the previous phase, starting with the desk study results. The most appropriate methods of site investigation depend on the geology and uncertainties. Hunt & Angulo (1999) discuss subsurface exploration methods for identification of boulders and cite most of the available pertinent references at that time. Since 1999, several additional papers involving site investigation for boulders have been published. Frank & Daniels (2000) discuss use of ground penetrating radar for boulder identification. Frank & Chapman (2001) describe how geologic setting, site reconnaissance, conventional borings, roto-sonic borings and large diameter (0.9–1.2 m) auger borings were used to characterize cobble and boulder conditions for the Big Walnut Augmentation/Rickerbacker Interceptor tunnel project in Columbus, Ohio. An important conclusion reached or implied by both Hunt & Angulo (1999) and Frank & Chapman (2001) is that a site investigation relying on one method, e.g. conventional borings, is unlikely to be sufficient unless there is significant previous local tunneling

Boulder characterization

Boulder characterization is essential to proactive management of tunneling risk in bouldery ground. A concise discussion of the main characteristics of boulder occurrence is given in Hunt & Angulo (1999). Boulder characterization generally requires both a thorough desk study and a focused site investigation program. 1.2.1 Desk study Boulder characterization should start with a thorough desk study that:

• • •

Determines the geologic setting and character of units and formations that may be encountered. Obtains available previous pertinent site investigation data in the project area. Finds and assesses available local tunneling case histories for cobble, boulder and abrasive ground conditions and impacts.

47 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-06.qxd

2/18/04

8:17 PM

Page 48

made of many types and materials, but can be generally grouped into two categories – roller cutters and drag cutters (Cigna & Ozdemir, 2000). Roller cutters are generally used for full-face hard rock, mixed-face and cobbly-bouldery ground and include:

experience in similar geologic conditions. Furthermore, a single phase of site investigation is also unlikely to be as successful or cost-effective as a phased investigation where during later phases increasingly more specific subsurface investigation and sampling methods are utilized. 1.3

• • • •

Boulder baselining

Boulder baselining is more than boulder characterization. The data and conditions found during the desk study and site investigation need to be converted into anticipated boulder frequencies, sizes and matrix conditions for the planned tunneling methods. Hunt & Angulo (1999) attempted to use probabilistic methods to relate boulder indications from borings to boulder encounters in tunnels, but did not have success with the method. Instead, they developed a semiempirical method that correlates boulder conditions encountered in borings and tunnels for similar local geologic units. Hunt (2002) described several case histories where the method was used since 1999 and concluded that “Boulder quantities can be reasonably estimated if boulder occurrence records from similar geologic units are correlated with indications of boulders from properly logged borings.” Other methods of quantifying boulder occurrence are discussed in Hunt & Angulo (1999). More recent papers have not presented specific methods for quantifying and baselining boulders, but have provided significant useful data on boulder conditions encountered in tunnels. DiPonio et al. (2003) discuss the occurrence of 2169 boulders within approximately 12.8 km of 2.1 and 2.3 m ID tunnels in the Detroit, Michigan area. Cronin & Coluccio (2003) discuss the occurrence of 34,300 boulders within approximately 2.5 km of 3.7 m ID tunnels in Portland, Oregon. Both papers provide useful data on frequency and size distribution of the boulders. 1.4

Single disk cutters Single and multiple mini-disc cutters Strawberry (button) roller cutters Multiple row carbide insert cutters

Examples are shown in Figure 1. Drag cutters are generally more efficient at mining fine-grained soil without cobbles, boulders and hard rock layers. Drag cutters include:

• • • • •

Chisel teeth Block scrapers Plow teeth Pick (bullet) teeth Blade cutters Examples are shown in Figure 2.

single disk cutter

double disk cutter

triple disk cutter with carbide inserts

carbide button cone cutter

two row carbide insert cutter

five row carbide insert cutter

Figure 1. Examples of roller cutters.

TBM features

TBM features substantially influence the potential impact of boulder occurrence. The features of most relevance include cutter types, face access, cutting wheel opening size, mucking system and cutting wheel torque and thrust. Other relevant features include: rock crusher type, cutting head armor (abrasion resistance), and ability to retract the cutting head 10–20 cm or more if to reestablish cutting wheel rotation. 1.4.1 Cutters Cutter type and configurations are extremely important TBM features that affect advance rate productivity, ability to excavate and fracture boulders, and extent of cutter and cutting wheel wear. Cutters are

chisel teeth on cutter arm

blade cutter

plow tooth

pick or bullet teeth

block scrapers

Figure 2. Examples of drag cutters.

48 Copyright © 2004 Taylor & Francis Group plc, London, UK

chisel tooth cutter

09096-06.qxd

2/18/04

8:17 PM

Page 49

hold boulders in place during fracturing. In soft or loose ground, boulders may become dislodged and roll around on the cutting wheel causing cuter damage and high face abrasion until fractured small enough for ingestion. If the ground is soft or loose enough, partially fractured boulders may pushed aside rather than ingested. The matrix shear strength issue has been discussed by a number of authors including: Navin et al. (1995), Becker (1995), Dowden, & Robinson (2001), and Goss (2002). Goss completed a PhD thesis on this subject and summarizes the issues and his results in the referenced 2002 paper. Goss concluded that the two most important parameters are the actual in-situ shear strength of the soil matrix and the shear or unconfined compressive strength of the boulders. Based on finite element modeling and case history evaluations, Goss suggested that: “…when the rockto-soil shear strength ratio is greater than 600:1, boulders cannot be broken by disc cutters.” Goss provided data on 18 case studies that generally support this conclusion. The case histories that are summarized subsequently in this paper suggest that disc and multiple row carbide insert cutters are essential to minimization of the risk of getting stuck on microtunneling and small TBM drives where there is no face access or access is problematic due to ground and groundwater conditions. This data will show that matrix shear strength was generally not the most important factor causing stuck TBMs. Although lower productivity and higher cutter wear were generally experienced when boulders were encountered in soils with low matrix shear strength, rugged TBMs with roller cutters were generally able to advance through bouldery ground even if the matrix soil was soft or loose.

A combination of roller and drag cutters are often used to maximize mining flexibility and TBM performance. Drag cutters are generally preferred for tunneling in mostly fine-grained soils because of higher cutting efficiency, however, drag cutters are vulnerable to rapid wear and breakage when mining through abrasive cobbly and bouldery ground. Furthermore, drag cutters are generally not capable of fracturing large boulders with unconfined compressive strengths over approximately 100 Mpa. In order to help minimize drag cutter wear Dowden and Robinson (2001) recommended that disc cutters be included whenever larger numbers of hard boulders are anticipated. A decision to include roller cutters for mining in cobbly or bouldery ground should not only depend on TBM productivity and cutter wear, but also on the risk and consequences of getting stuck. The risk of getting stuck depends on boulder size, cobble and boulder concentrations and distribution (nested vs. scattered), rock strength, TBM diameter, cutting wheel opening size, rock crusher capability, face access and method of face pressurization (slurry shield, earth pressure balance by pressure relieving gates, earth pressure balance by screw auger). Nishitake (1987) discusses roller cutter types and configurations and presents the results of tests on cutting effectiveness. Specific recommendations for earth pressure balance machine features in bouldery ground are given. Friant and Ozdemir (1994) discuss the development of mini-disc cutters for microtunneling and small diameter TBM tunneling and compare their effectiveness to other drag and roller cutters. A study of cutter efficiency at the Colorado School of Mines (Ozdemir, 1995) resulted in a conclusion by Ozdemir that: “In summary, the mini-disc cutter offers many significant advantages over any other type of cutting tolls currently used on microtunneling machines. These advantages include very high cutting efficiency: high penetration rates; low machine thrust, torque and power requirements; excavation capability in any type of soil and hard rock:* low initial cost: low replacement costs; ease of replacement and the elimination of the need for a cutter shop; true-rolling feature (meaning reduced torque and power requirements compared to button or multi-kerf cutters); greater lifetime and drive lengths compared to carbide cutters; and significantly reduced fines, meaning less slurry cleanup requirements.”

2 CASE HISTORIES 2.1

1.4.2 Disc cutter effectiveness in soil Cutter configuration and choice of both roller and drag cutters will significantly impact success at mining through bouldery ground. One concern with use of disc and multiple row carbide insert cutters is the soil matrix shear strength or density and its ability to

49 Copyright © 2004 Taylor & Francis Group plc, London, UK

Data

A search of tunneling case history articles, papers and project files resulted in 40 cases from 36 projects for this study. Extracted data for these 40 cases are listed in Table 1 (sheets 1a-1h). The primary references for the cited data are listed in the bottom row. These references may have additional information of interest. Some of the references lack all the desired pertinent data. In many cases boulder quantities and sizes were not thoroughly reported, often because documentation of boulder quantities and sizes is not practical (Hunt, 2002). Where boulder quantities and sizes were estimated by the Authors based on an interpretation of reported information, this data is preceded with an asterisk (*).

09096-06.qxd

2/18/04

8:17 PM

Page 50

Table 1a. Case histories 1a–4. Case no.

1a

1b

2

3

4

Location

Clearview Snohomish River Undercrossing, 1st Attempt – Snohomish, WA SJP Alluvial sand, gravel and cobbles

Clearview Snohomish River Undercrossing, 2nd Attempt – Snohomish, WA SJP Alluvial sand, gravel and cobbles

Tolt Pipeline No. 2, Snoqualmie River – Seattle, WA SJP Sand, gravel, and cobbles

Tolt Pipeline No. 2, Bear Creek – Seattle, WA

Swamp Creek – Snohomish County, WA

SJP Silty clay till, sand and gravel outwash

RCP Glaciolacustrin silty clay to beach sand

Loose to dense

Loose to dense

Loose to dense

Hard to dense

21.3

21.3

24.4

5.5

Hard to med. dense 7.6

22.9–33.5

22.9–33.5

25.9

10.7

10.1

339.9 339.9

339.9 339.9

652.3 213.4

113.4 113.7

315.5 169.2

177.4 1575

339.9 1575

213.4 2286

113.4 1905

108.2 1219

*8

*17

*400

50

15

1 *0.30%

*2 *0.10%

* 80 *2.50%

10 0.90%

4 0.20%

1016

*508

914

508

1219

65% Igneous and metamorphic

32% Igneous and metamorphic

27% Igneous and metamorphic

100% Igneous and metamorphic

STBM Isecki Unclemole None Block scraper drag cutters

STBM Lovat MTS

40% Igneous and metamorphic 234.4 STBM Soltau RVS 800 STS None Triple disc roller cutters and chisel teeth Cutters and cone crusher

STBM Soltau RVS 600 None Five row carbide insert roller and strawberry cone cutters Cutters and cone crusher

STBM Isecki TCC 1000 Yes Block scraper drag cutters

Jacked pipe type Primary soil type (with varying quantities and sizes of boulders) Soil matrix Consistency/density Maximum groundwater head, m Approx. tunnel depth, m Tunnel length, m Selected drive length, m Length achieved, m Excavated diameter (ED), mm No. of boulders in drive No. of large boulders Est’d % boulders (by Vol) Max. boulder Size mm Boulder size/ED, % Rock types Rock Qu, MPa TBM type TBM make Face access TBM cutters

Fracture method

Cone crusher

TBM advance Impact Cutter, cutterhead Damage

STBM stuck and abandoned Bits broken and worn

Remedial Measures

Sank new shaft, re-mined w/new STBM

Risk of getting stuck Consequence of stuck References

Very high $3 million Staheli, 2003

None Multi-row carbide insert roller cutters and chisel teeth Cutters and cone crusher None Considerable cutterhead surface wear NA

Drag teeth badly worn, disc cutters ok NA

Minor

Low $3 million Staheli, 2003

Low $1 million Molvik et al., 2000

Low $1 million Molvik et al., 2000; Beieler et al, 2003

* Estimated from best available data – reliability uncertain.

50 Copyright © 2004 Taylor & Francis Group plc, London, UK

Cone crusher EPBM w PRG stuck at 355 ft Major-one cutter arm torn off Recovery tunnel mined to finish drive, remove EPBM Very high $400,000 Genzlinger, 1995

09096-06.qxd

2/18/04

8:17 PM

Page 51

Table 1b. Case histories 5–7. Case no.

5

6

7a

7b

7c

Location

Mercer St. Railroad Crossing TunnelsSeattle, WA RCP Glaciolacustrine silty clay

NEHLA undershore crossing, Keahole Point, Hawaii SJP Coarse sand and gravel with cobbles

Folsom east IIB Test A Sacramento, California RCP Alluvial sand, gravel and cobbles

Folsom east IIB Test B Sacramento, California RCP Alluvial sand, gravel and cobbles

Folsom east IIB Sewer Sacramento, California RCP (2 pass) Alluvial sand, gravel and cobbles

Hard

Med. dense

Loose to dense

Loose to dense

Loose to dense

12.2

13.7

0.0

0.0

0.0

15.2

13.7

10.7

10.7

24.4

213.4 106.7

353.6 176.8

304.8 152.4

304.8 152.4

906.8 431.6

106.7 3048

176.8 1746

61.0 1778

26.2 1778

431.6 3048

2

*10

*32

*72

*1900

1 *0.03%

*2 *0.10%

*7 *4.00%

*5 *3.00%

*300 *2.00%

*762

*610

889

610

889

25% Igneous and metamorphic

35% Basalt

50% Hard igneous

34% Hard igneous

29% Hard igneous

138–270 STBM Soltau RVS 800 AS None Multi-row carbide insert roller cutters Button bits and cone crusher

186–268 OFRW Akkerman

186–269 STBM Akkerman

Yes Chisel teeth

None Chisel and bullet teeth

Passed through head

Cone crusher

TBM advance Impact

STBM jammed/ stuck 6 times

STBM stuck, replaced by OFRW TBM

Cutter, cutterhead Damage

Minor

Remedial measures

Cutting wheel freed by retracting head 10 cm High $100,000 Smith, 1995: Miller, 1995a

OFRW stuck, replaced by another OFRW TBM Severe wear of drag cutters, cutterhead TBM removed in rescue shaft, drive ended High $100,000 Staheli et al., 1999

186–276 OFRW Lovat 121 PJ/RL -8600 Yes Chisel, block scraper and carbide bullet teeth Jackhammer for occasional v. large boulders Periodic cutter replacement

Jacked pipe type Primary soil type (with varying quantities and sizes of boulders) Soil matrix consistency/density Maximum groundwater Head, m Approx. tunnel depth, m Tunnel length, m Selected drive Length, m Length achieved, m Excavated diameter (ED), mm No. of boulders in drive No. of large boulders Est’d % boulders (by Vol) Max. boulder size mm Boulder size/ED, % Rock types Rock Qu, MPa TBM type TBM make

EPBM-PRG Lovat M-102

Face access TBM cutters

None Chisel teeth

Fracture method

Hyd. Splitting, most passed through head,

Risk of getting stuck Consequence of stuck References

Very low $500,000 Genzlinger, 1995

51 Copyright © 2004 Taylor & Francis Group plc, London, UK

Severe wear of drag cutters, cutterhead STBM removed in rescue shaft, drive ended Very high $100,000 Staheli et al., 2000

Severe wear of cutters, cutterhead

Very low $500,000 Castro et al., 2001

09096-06.qxd

2/18/04

8:17 PM

Page 52

Table 1c. Case histories 8–12. Case no.

8

9

10

11

12

Location

Carmichael Water Crossing of American River – Carmichael, CA RCP Alluvial sand and gravel with cobbles

CCWA Santa Ynez River Crossing, Mission Veijo, CA

Peralta Blvd Sanitary Sewer, Fremont, CA PCP Mixed clay, gravel and cobbles

North Mission Valley Interceptor, San Diego, CA

SJP Claystone, silty sand and cobbles

Dense

Hard to dense

Pacific Coast Highway Sewer, Phase II, Santa Monica, CA RCPP Sand and gravel with cobbles and some clay seams Soft to stiff

Stiff

Dense to hard

15.2

11.6

3.0

27.4

16.5–23.5

6.7

4.9–6.1

6.7

219.5 219.5

365.8 248.4

1,036.3 365.8

914.4 128.0

1,402.1 85.3

219.5 1524

248.4 1549

365.8 1930

128.0 635

21.3 780

*10

*5

*200

*16

*100

*1

*1

0

0

*20

*0.10%

*0.10%

*0.50%

*1.00%

*1.00%

*457

1549

381

*457

457

30% Hard igneous 186–269 STBM Soltau RVS 600 A-S None Block scraper and carbide bullet teeth

100% Igneous 241.3 STBM Soltau RVS 600 A5 None Multi-row carbide insert roller cutters and button cones Cutters and cone crusher

20% Igneous

72% Hard igneous

STBM Herrenknecht AVN 1500T Yes Chisel teeth and disc cutters

STBM Akkerman

59% Igneous 69 STBM Isecki TCC600

None Chisel and bullet teeth

None Block scraper drag cutters

Cutters and cone crusher

Cone crusher

Cone crusher STBM stuck 3 drives

Jacked pipe type Primary soil type (with varying quantities and sizes of boulders) Soil matrix consistency/density Maximum groundwater head, m Approx. tunnel Depth, m Tunnel length, m Selected drive Length, m Length achieved, m Excavated diameter (ED), mm No. of boulders in drive No. of large boulders Est’d % boulders (by Vol) Max. boulder Size mm Boulder size/ED, % Rock types Rock Qu, MPa TBM type TBM make Face access TBM cutters

Fracture method

Cone crusher

TBM advance impact

VCP Cemented silty sand, silty clay and cobbles

3.7

Cutter, cutterhead damage Remedial measures

Moderate

Minor

Minor

Steering and grade difficulties from boulders Moderate

NA

NA

NA

NA

Risk of getting stuck Consequence of stuck References

High

Low

Low

High

2 recovery shafts, 1 recovery trench Very high

}

Over $500,000

$100,001

$100,000

$100,000

Castro et al., 2001

Miller, 1996

Rush, 2000; Rush 2002

Miller, 1997a

Miller, 1997b

52 Copyright © 2004 Taylor & Francis Group plc, London, UK

Moderate

09096-06.qxd

2/18/04

8:17 PM

Page 53

Table 1d. Case histories 13–16. Case no.

13

14

15a

15b

16

Location

Deicing Fluid Line, Denver Int. Airport – Denver, CO SJP Hard silty clay and layered soft shale

CSO Separation RNC 5318 – Omaha, NE

Lincoln Way Drive 1 -Ames. IA

Lincoln Way Drive 2 -Ames. IA

MN-320 Sewer Minneapolis, Minnesota

RCP Alluvial sand and gravel

RCP Silty clay till w sandy gravel outwash

RCP Silty clay till w sandy gravel outwash

FMCP Silty clay till, gravelly sand outwash

Hard

Loose-medium dense 7.6

Very stiff to hard 9.1

Very stiff to hard 9.1

Very stiff 4.6

9.8

12.2

3.7–12.2

3.7–12.2

10.4–12.2

316.4 241.7

231.6 231.6

929.0 111.6

929.0 132.3

431.3 253.0

241.7 1092

231.6 2159

103.9 787

132.3 787

140.8 1524

*10

*10

*10

17

16

0 *0.05%

*5 *1.00%

*2 *0.70%

3 *1.00%

6 2.00%

356

762

610

610

1372

33% Igneous

35% Igneous

77% Granite, other igneous

77% Granite, other igneous

90% Gabbro, granite

Rock Qu, MPa TBM type TBM make

OFRW Akkerman

OFRW Akkerman 720

Yes Chisel teeth

Yes Chisel teeth

Fracture method

Pneumatic hand tools 10 hour delay

Blasting

Cone crusher

Stuck on large boulder at 10 ft advance

Stuck at 341 ft

Major cutter wear and broken teeth Recovery trench, added disc cutters Very high $100,000 Schumacher & Ellis, 1997; Najafi & Varma, 1996

STBM Herrenknechht AVN600 None Double-row carbide insert and disk cutters and plow teeth Cutters and cone crusher STBM w rock head successful – stuck at launch using head with drag cutters Major cutter wear and broken teeth Pulled back TBM, changed to head with disc cutters Low $100,000 Schumacher & Ellis, 1997; Najafi & Varma, 1996

STBM Akkerman

Face access TBM cutters

STBM Herrenknechht AVN600 None Chisel teeth

Jacked pipe type Primary soil type (with varying quantities and sizes of boulders) Soil Matrix consistency/density Maximum groundwater head, m Approx. tunnel depth, m Tunnel length, m Selected drive length, m Length achieved, m Excavated diameter (ED), mm No. of boulders in drive No. of large boulders Est’d % boulders (by Vol) Max. boulder size mm Boulder size/ED, % Rock types

TBM advance impact

Cutter, cutterhead damage

Minor

Minor

Remedial measures

Man access to face to split, dislodge boulders Low $5,000 Coss, 1993

Man access to face to blast boulders Moderate $10,000 King et al., 1997

Risk of getting stuck Consequence of stuck References

53 Copyright © 2004 Taylor & Francis Group plc, London, UK

None Chisel and bullet teeth

Cone crusher STBM stuck twice, 2 rescue shafts Half of teeth gone, rest very worn Cuttingwheel replaced once, TBM replaced Very high $3 million Hunt, 2003

09096-06.qxd

2/18/04

8:17 PM

Page 54

Table 1e. Case histories 17–21. Case no.

17

18

19

20

21

Location

Menomonee River Water Pipeline – Marrinette, Wisconsin SJP Silty clay till w outwash pockets

Oklahoma Ave. Relief Sewer – Milwaukee, WI

Ramsey Ave. Relief Sewer – Milwaukee, WI

Miller 37th & State Area MIS – Milwaukee

S. Penn. Ave. Relief Sewer – Oak Creek, WI

RCP Silty clay till w pockets of outwash sand Very stiff

12.2

7.6

12.2

RCP Silty clay till and cobbly sand and gravel outwash Very stiff – hard to very dense 3.0

RCP Silty clay till w thick layers of outwash

Very stiff

RCP Silty clay till w outwash pockets & boulder clay Very stiff

10.4–15.2

10.7–14.9

6.1–19.8

15.2–21.3

6.1–12.1

268.2 268.2

874.8 874.8

941.2 201.5

390.4 405.4

2,497.8 179.8

268.2 1422

874.8 1537

195.1 997

405.4 2642

179.8 1397

7

151

7

346

160

2

71

1

60

21

0.10%

0.40%

0.40%

1.60%

2.50%

762

762

610

1067

914

54% Hard igneous

50% Igneous erratics and dolomite

40% Igneous erratics and dolomite

STBM Soltau RVS 600 None Multi-row carbide insert roller cutters and chisel teeth Disc cutters and cone crusher

OFRW Decker Yes Chisel teeth

61% Gabbro and dolomite 206.8 STBM Isecki Unclemole None Block scraper drag cutters

65% Igneous erratics and dolomite 206.8 OFRW Decker Yes Chisel teeth

Cone crusher

Blasted boulders 60 times Hand mined from cut wheel twice to replace cutters

Jacked pipe type Primary soil type (with varying quantities and sizes of boulders) Soil matrix consistency/density Maximum groundwater Head, m Approx. tunnel depth, m Tunnel length, m Selected drive length, m Length achieved, m Excavated diameter (ED), mm No. of boulders in drive No. of large boulders Est’d % boulders (by Vol) Max. boulder size mm Boulder size/ED, % Rock types Rock Qu, MPa TBM type TBM make Face access TBM cutters

Fracture method

Blasting

OFRW Modifed Decker Yes Chisel teeth

Very stiff to hard or dense 13.7

Severe – wear and broken drag teeth NA

Very low

Moderate – worn, broken drag teeth Recovery tunnel mined to finish drive and remove STBM High

Blasted boulders 177 times Slower production and delays for blasting boulders Moderate – worn, broken drag teeth NA

Low

Low

$1 million

$100,000

$600,000

$100,000

$10,000

Vadnais, 2002

Hunt, 2002

Hunt, 1999

Hunt, 2002

Hunt, 1999

TBM advance Impact

Stuck at 640 ft

Cutter, cutterhead Damage

Minor cutter wear

Minor cutter wear

Remedial measures

NA

NA

Risk of getting stuck Consequence of stuck Reference nos.

Low

54 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-06.qxd

2/18/04

8:17 PM

Page 55

Table 1f. Case histories 22–26. Case no.

22

23

24

25

26

Location

Oak Creek Southwest Relief Sewer – Oak Creek, WI RCP Silty clay till with cobble and boulder lag zone Very stiff to hard or dense 15.2

Elgin Interceptor – Elgin, IL

Northeast Interceptor Sewer – Libertyville, IL

Weller Creek Sewer Project – Arlington Hts, IL

Evanston CSO Phase IV – Evanston, IL

RCP Silty clay till and outwash sand and gravel Very stiff to hard or dense 12.2

FMCP Silty clay till w sandy gravel outwash Very stiff to hard 10.7

RCP Silty clay to clayey silt till

RCP Silty clay to clayey silt till

Very stiff to hard 12.2

Very stiff to hard 9.1

16.8

6.1–14.3

10.7–14.0

15.2

11.6–12.2

1,524.0 1,524.0

1,313.1 405.4

929.0 152.4

3,116.0 393.2

335.3 182.9

1,524.0 1397

405.4 1981

152.4 1397

393.2 3353

182.9 2159

156

45

*10

*20

*5

80

20

*2

*5

*2

1.60%

1.60%

*0.18%

*0.04%

*0.20%

1219

1016

*914

*762

*1067

87% Igneous erratics and dolomite

51% Igneous erratics and dolomite

65% Igneous erratics and dolomite

23% Igneous erratics and dolomite

49% Igneous erratics and dolomite

Rock Qu, MPa TBM type TBM make

OFRW Decker

OFRW Lovat

OFRW Lovat

Face access TBM cutters

Yes Chisel teeth

Yes Chisel teeth

OFRW Akkerman 720C Yes Chisel teeth

Fracture method

Blated boulders 80 times

Blasting

STBM Herrenknechht AVN 1200 None Chisel teeth and block scraper drag cutters Cutters and cone crusher

TBM advance impact

Slower production and delays for blasting boulders Moderate – worn, broken drag teeth NA

Stopped 24 times to blast large boulders

Cone crusher

Moderate – worn, broken drag teeth NA

Low

Moderate

Moderate – worn, broken drag teeth Relief shafts bored, large boulders below MTBM High

$50,000

$100,000

Hunt, 2002

Hunt, 1999

Jacked pipe type Primary soil type (with varying quantities and sizes of boulders) Soil matrix consistency/density Maximum groundwater head, m Approx. tunnel depth, m Tunnel length, m Selected drive length, m Length achieved, m Excavated diameter (ED), mm No. of boulders in drive No. of large boulders Est’d % boulders (by Vol) Max. boulder size mm Boulder size/ED, % Rock types

Cutter, cutterhead damage Remedial measures

Risk of getting stuck Consequence of stuck Reference nos.

55 Copyright © 2004 Taylor & Francis Group plc, London, UK

Yes Chisel teeth

Most past through, blasted large boulders

Hydraulic split large boulders

Minor cutter wear

Minor cutter wear

NA

NA

Low

Low

$25,000

$100,000

$100,000

Rickert et al., 1999, Westcon.net

Hunt, 1999

Miller, 1995b

09096-06.qxd

2/18/04

8:17 PM

Page 56

Table 1g. Case histories 27–31. Case no.

27

28

29

30

31

Location

Marathon Oil Storm Sewer – Indianapolis, IN

Downriver DRSTS Contract No. 4 – Wayne Co. MI

PSE&G Power Tunnel, Jersey City, NJ

Chelsea River Crossing – Boston, MA

Jacked pipe type Primary soil type (with varying quantities and sizes of boulders) Soil matrix consistency/density Maximum groundwater head, m Approx. tunnel depth, m Tunnel length, m Selected drive length, m Length achieved, m Excavated diameter (ED), mm No. of boulders in drive No. of large boulders Est’d % boulders (by Vol) Max. boulder size mm Boulder size/ED, % Rock types

FMCP Hard silty clay till w sandy gravel outwash

RCP Silty clay till and hardpan till with cobbles

SJP Clayey silt and sand and rubble fill

SJP Silty clay till and alluvial silty sand

New Neponset Valley Force Main – Canton, MA RCPP Sand, silty sand, some gravel, cobbles, peat

Hard to very hard or very dense 3.7

Very stiff to very hard 15.2

Stiff

Dense

9.1

Hard to very dense 24.4

6.7–7.6

18.3

10.7–11.3

7.3–10.4

8.2

684.3 239.9

4,572.0 4,572.0

283.5 94.5

299.9 299.9

430.1 430.1

239.9 1758

4,572.0 2642

94.5 965

299.9 1961

430.1 2108

8

995

*5

32

5

3

283

*1

6

0

0.17%

0.20%

*0.50%

*0.15%

0.01%

610

1219

*508

914

457

35% Granite, other igneous

46% Igneous erratics and dolomite

53% Igneous, Metamorphic

47% Granite, other igneous

22% Granite, other igneous

OFRW Akkerman

OFRW Lovat M-120

STBM Soltau RVS 600

EPBM -PRG Lovat M77

Face access TBM cutters

Yes Chisel and bullet teeth

Yes Chisel and bullet teeth

STBM Soltau RVS 250A/S No Chisel teeth

Yes Chisel and bullet teeth

Fracture method

Hydraulic split large boulders

Blasted boulders 10 times – most past through

Disc cutters and cone crusher

Yes Multi-row carbide insert cutters, bullet and chisel teeth Disc cutters and cone crusher

Minor NA

Moderate – worn, broken drag teeth NA

Minor cutter wear NA

Minor cutter wear NA

Low

Low

Minor cutter wear Relief shaft needed to remove Hand timber piles High

Low

Low

$100,000

$100,000

$100,000

$1 million

$100,000

Garrett, 1992

DiPonio, et al., 2003 Miller, 1994

Tarkoy, 2001

Boscardin, 1997

Rock Qu, MPa TBM type TBM make

TBM advance impact Cutter, cutterhead Damage Remedial measures

Risk of getting stuck Consequence of stuck Reference nos.

56 Copyright © 2004 Taylor & Francis Group plc, London, UK

4.6

All boulders passed through

09096-06.qxd

2/18/04

8:17 PM

Page 57

Table 1h. Case histories 32–36. Case no.

32

33

34

35

36

Location

Kelvin Valley Wastewater, Scotland UK

Sudden Valley Sewer, Lancashire UK

East Dock Relief Sewer, Dundee, Scotland UK

Carronvale Sewer, Scotland UK

Jacked pipe type Primary soil type (with varying quantities and sizes of boulders) Soil matrix consistency/ density Maximum groundwater head, m Approx. tunnel depth, m Tunnel length, m Selected drive length, m Length achieved, m Excavated diameter (ED), mm No. of boulders in drive No. of large boulders Est’d % boulders (by Vol) Max. boulder size mm Boulder size/ED, % Rock types

SJP Silty sand w cobbles, schist and sandstone

FCMP Silty sand and gravel w cobbles and clayey zones

GRP Silty clay, sand and gravel over sandstone, basalt

RCP Fill, clay, silt and fine sand

Neva River Undercrossing – St. Petersburg, Russia RCP Silty fine sand and gravel, some clayey zones

Hard

Dense

Stiff to dense

6.0

8.0

Soft to stiff and loose to medium dense 4.0

20.1

9.1–12.2

10.0

10.0

5.0

25.0

438.3 178.9

570.0 570.0

930.0 301.0

132.0 132.0

774.0 774.0

178.9 1956

570.0 1460

264.0 2200

132.0 900

774.0 2540

*5

*100

*20

*9

*25

0

0

*5

1

*5

*0.02%

*0.50%

*0.50%

*0.50%

*0.03%

406

330

762

610

508

21% Schist, other

23% Igneous

35%

68%

20% Granite, other igneous

STBM Wirth-Soltau BH1920/1600 Yes Triple disc cutters and chisel teeth

EPBM Markham OKMS

STBM Herrenknecht AVN 1600 Yes Disc cutters and chisel teeth

EPBM Howden-716

STBM Herrenknechht AVN 2000D Yes Double disc cutters, chisel and plow teeth Disc cutters and cone crusher

Rock Qu, MPa TBM type TBM make Face access TBM cutters

Fracture method

Yes Disc cutters, bullet and chisel teeth

Disc cutters and crusher arms

Disc cutters

Cutter, cutterhead damage Remedial measures

Minor cutter wear NA

Minor cutter wear NA

Risk of getting stuck Consequence of stuck Reference nos.

Low

Yes Disc cutters and chisel teeth

Minor cutter wear NA

NA

Low

Disc cutters and cone crusher Stuck at 866 m in nested cobble and boulder backfill Minor cutter wear Relief shaft sunk to remove nested cobbles and boulders Moderate

Moderate

Medium

$500,000

$100,000

$100,000

$500,000

$1 million

Gehlen & Huhn, 2002

Jones, 1990

Fleet & Owen, 2002

Clarke, 1990

Wallis, 2002

TBM advance impact

57 Copyright © 2004 Taylor & Francis Group plc, London, UK

Plenum rock crusher

Loose to dense or stiff

Minor

09096-06.qxd

2/18/04

8:17 PM

Page 58

The following abbreviations were used in the table for jacked pipe type: SJP RCP RCPP VCP FMCP PCP GRP

all cases 30% stuck = 12/40

Steel jacking pipe Reinforced concrete pipe Reinforced concrete pressure pipe Vitrified clay pipe Fiberglass polymer composite pipe Polymer concrete pipe Glass resin pipe

drag cutters 37% stuck = 10/27 roller or roller and drag cutters 15% stuck = 2/13 drag cutters

The following abbreviations were used in Table 1 for tunnel boring machine (TBM) type: STBM OFRW EPBM EPBM-PRG

67% stuck = 8/12

Slurry tunnel boring machine Open face rotary wheel machine Earth pressure balance machine w screw Earth pressure balance machine with pressure relieving gate (PRG)

20% stuck = 2/10

roller or roller and drag cutters

drag cutters 13% stuck = 2/15 roller or roller and drag cutters

2.2

Analysis of case history data

0% stuck = 0/3

Table 1 summarizes data from 40 microtunneling and small diameter tunneling cases where boulders were encountered. The following paragraphs provide an analysis of the data.

Figure 3.

OFRWs, EBPMs 11% stuck = 2/18

Distribution of stuck TBMs.

pressure relieving gates (PRGs) and face access. None of the TBMS for these cases became stuck. Of the 27 cases involving tunneling with only drag cutters, 13 cases involved open-face rotary wheel machines (OFRW) with face access. Two of the OFRW machines or 13 percent became stuck (Cases 7a and 14). In the Case 7a, the drag cutters and cutting wheel of the OFRW were so badly damaged by cobbles and boulders upon becoming stuck that the TBM was removed from the project (Staheli et al., 1999). In Case 14, a very large boulder was encountered. The stuck OFRW and boulder were removed with a recovery trench rather than by splitting or blasting the boulder. Subsequently, roller disc cutters were added to the cutting head for the remaining tunnel drives. In the other 11 OFRW cases, face access allowed any large boulders that would not pass through the cutting wheel and mucking system to be blasted or split ahead of the cutting wheel. In general, the ground at these headings had sufficient stand-up time for face access without air pressure. In some cases, grouting or localized dewatering was required for face access. While blasting or splitting boulders through face access generally prevented the TBMs from getting stuck, but it did not prevent cutter damage. In some cases, No. 20 for example, chisel cutters were so badly worn and broken by cobbles and boulders during each of two 340–350 m long drives that hand-mining in front of the cutting wheel was required to replace chisel cutters before pipe jacking could proceed. Fortunately, increased side friction from set-up during cutter repairs was not sufficient to cause stuck drives.

2.2.1 Stuck STBMs For the case histories evaluated in this paper, the tunnel boring machines were stuck (unable to advance without intervention) a total of 12 times in 40 cases (30 per cent overall stuck rate). Figure 3 shows a graphical summary of the results for stuck TBMs. Slurry shield tunnel borings machines (STBMs) were stuck during 10 of the 40 cases (a 25 percent stuck rate) and during 10 of the 22 cases where STBMs were used (45 percent stuck rate). Of the 10 stuck cases, 8 involved STBMs with drag cutters only. The other two cases were STBMs that had some roller cutters. The STBM for one of the stuck cases had only roller cutters (Case 6). This STBM was jammed by cobbles and boulders six times, however, cutting wheel retraction capability of about 10 cm allowed the head to be freed each time resuming roller cutter fracturing and allowing the drives to be finished without other intervention (Miller, 1995a). The other stuck STBM had combined drag and roller cutters (Case 34). This STBM encountered a fill with nested cobbles and boulders that required a relief shaft to remove the obstructions. Twenty-seven of the 40 cases involved tunneling with drag cutters only. Of these 27 drag cutter only cases, 12 cases involved tunneling with STBMs. The STBMs with only drag cutters became stuck during 8 of these 12 cases – a 67 percent stuck rate. Of the 27 cases involving tunneling with only drag cutters, 2 cases were mined with EPBMs having

58 Copyright © 2004 Taylor & Francis Group plc, London, UK

STBMs 45% stuck = 10/22

09096-06.qxd

2/18/04

8:17 PM

Page 59

Thirteen of the 40 cases in Table 1 involved tunneling with roller cutters only or a combination of roller and drag cutters. Ten of these 13 cases involved STBMs with combined roller and drag cutters or roller cutters only. Two of these machines or 20 percent became stuck (Cases 6 and 34). In Case 6, which was previously discussed, the STBM became jammed six times and was freed each time by retracting the head about 10 cm to re-establish cutting wheel rotation. The cutters and rock crushers then successfully fractured the obstructing boulders and no other intervention was required. In Case 34, the STBM encountered a nested cobble and boulder fill zone that required a relief shaft to remove the obstructions. Three of the 13 cases with combined roller and drag cutters involved EPBMs (with screw augers). None of these machines became stuck. In summary, STBMs with only drag cutters became stuck 67 percent of the cases (8 of 12), while STBMs with combined drag and roller cutters or only roller cutters became stuck for 20 percent of the cases (2 of 10). TBMs with face access (OFRW machines and EPBMs with pressure relieving gates) became stuck 13 percent of the cases (2 of 15). EPBMs with combined roller and drag bits did not get stuck.

large boulder cases 43% stuck = 12/28 drag cutters 50% stuck = 10/20 roller or roller and drag cutters 15% stuck =2/8 drag cutters 80% stuck = 8/10

29% stuck = 2/7

STBMs 59% stuck = 10/17

drag cutters 20% stuck = 2/10 roller or roller and drag cutters

OFRWs, EBPMs 18% stuck = 2/11

0% stuck = 0/1 Figure 4. Distribution of stuck TBMs with maximum boulder size 33% of excavated diameter.

2.2.2 Boulder size factor Maximum boulder sizes encountered as reported in the references or as estimated based on information provided are listed in Table 1. Assuming that cobbles and boulders smaller than 33 percent of the excavated diameter can generally pass through the cutting wheel and be crushed or pass through a conveyor mucking system without further fracturing, then boulders larger than 33 percent would require fracturing, pushing aside or removal by relief shaft or tunnel intervention in order for a TBM to proceed. For some TBMs, particularly EPBMs with a screw auger, the digestible boulder size is smaller than 33 percent of the excavated diameter and may be as low as 5 to 10 percent of the excavated diameter. A total of 28 of 40 or 70 percent of the cases in Table 1 encountered boulders larger than 33 percent of the excavated diameter. In 43 percent (12 of 28) of the large boulder cases, the TBM became stuck (Figure 4). Of these 28 large boulder cases, 10 involved STBMs with drag cutters and 80 percent (8 of 10) of them were stuck. Ten of the 28 large boulder cases involved OFRWs with drag cutters and 20 percent (2 of 10) of them were stuck. Seven of the 28 large boulder cases involved STBMs with only roller or roller and drag cutters and 29 percent (2 of 7) of them were stuck. One of the 28 large boulder cases involved EPBMs and it was not stuck. A comparison of the results presented in Figure 4 to those in Figure 3 shows that the frequency of stuck TBMs is greater when the maximum boulder size

encountered is greater than approximately one third of the excavated diameter. 2.2.3 Boulder volume (frequency) Table 1 includes estimated boulder frequencies and volumes as a percentage of the excavated tunnel volume. The listed total boulder and large boulder (greater than 50 cm in size) quantities were based on available project data or estimated based on reported information on boulder encounters. The most uncertain data is preceded by an asterisk (*). Boulder volumes were estimated using an Excel spreadsheet that based on a normal distribution of common sizes using an estimated mean size, limited maximum size and assumed standard deviation. Boulder volumes were computed as 0.7 D3 where D is the average diameter for sub-rounded boulders (Hunt & Angulo, 1999). This method worked well for several cases where better boulder size data was available (e.g. Cases 18–23), but should only be considered a crude estimate of boulder volume. A total of 23 of 40 or 58 percent of the cases in Table 1 encountered estimated boulder volumes that were equal to or more than 0.4 percent of the excavated tunnel volume. The 0.4 percent division was arbitrarily based on experience indicating that boulder impacts were generally more significant when this relative volume was reached or exceeded. In 30 percent

59 Copyright © 2004 Taylor & Francis Group plc, London, UK

roller or roller and drag cutters

09096-06.qxd

2/18/04

8:17 PM

Page 60

feature that allowed the cutting wheel to be retracted about 10 cm to allow cutting wheel and roller cutter action to resume. Several methods of removing boulder obstructions are discussed in Hunt, 2002. Staheli and Hermanson, 1997 in a paper entitled “What to Do When your Head Gets Stuck” discuss some additional alternatives for dealing with boulder obstructions. In most cases of stuck TBMs without face access, the remedial alternatives are limited and often result in a recovery shaft or tunnel, particularly if severe cutter damage has occurred from the boulder encounters prior to getting stuck. If face access is available, more remedial options exist for boulder obstruction removal. The least risky method, if the ground and groundwater conditions allow it, would be open-face mining with an OFRW machine or other type of shield or hand-mining operation. Combined drag and roller cutters would help minimize cutter wear and damage. Boulder blasting or hydraulic splitting can generally be achieved in a cost-effective manner if ready face access exists in ground with suitable standup time. Having face access makes use of drag cutters less risky because the larger boulders can be accessed and fractured before excessive, damaging grinding occurs. While this method generally prevents the TBM from becoming stuck, it does not prevent problems. The cobbly and bouldery ground in Case 7a so severely damaged the drag cutters and cutting arm that the OFRW machine was removed and mining with it was discontinued. In Case 20, the cobbly and bouldery ground caused such severe wear to chisel drag cutters that a cutter repair chamber had to mined in front of the TBM on each of both 200 m long drives even though 60 boulder obstructions were blasted. If an EPBM is utilized due to poor standup time or variable ground conditions, the most flexibility for boulder obstruction removal would result if semipressurized mode mining with a pressure relieving gate is allowed. This method was successful for Case 31 where a small number of boulders were encountered. Similar results were experienced in Milwaukee on the Northshore 9 Collector System tunnel (Goss, 2002; Budd & Cooney, 1991). However, if cobble and boulder quantities are high, flood doors and pressure relieving gates may be destroyed or severely damaged (Castro et al., 2001; Cronin & Coluccio, 2003). The data in Table 1 indicate that for 8 of 12 stuck TBM cases, a recovery shaft or tunnel was mined to remove the TBM. In 6 of these 8 cases, the stuck TBM was removed and replaced by a different TBM or cutting wheel or tunneling was discontinued. For the other two cases, the TBM resumed mining after cutter and cutting wheel repairs. Boulder obstruction removal shafts were completed for 2 of the 12 stuck TBM cases (Cases 24 and

cases w boulder vol. 0.4% exc. vol. 30% stuck = 7/23 drag cutters 35% stuck = 6/17 roller or roller and drag cutters 17% stuck =1/6 drag cutters 67% stuck = 6/9

25% stuck = 1/4

roller or roller and drag cutters

STBMs 54% stuck = 7/13

drag cutters 25% stuck = 2/8 roller or roller and drag cutters

OFRWs, EBPMs 20% stuck = 2/10

0% stuck = 0/2 Figure 5. Distribution of stuck TBMs when estimated boulder volume 0.4 percent of excavated volume.

(7 of 23) of the larger boulder volume cases, the TBM became stuck (Figure 5). Of these 23 larger boulder volume cases, 9 involved STBMs with drag cutters and 67 percent (6 of 9) of them were stuck. Two of the 23 larger boulder volume cases involved OFRWs with drag cutters and 24 percent (2 of 8) of them were stuck. Four of the 23 larger boulder volume cases involved STBMs with only roller or roller and drag cutters and 25 percent (1 of 4) of them were stuck. Two of the 23 larger boulder cases involved EPBMs with drag and roller cutters and neither of them became stuck. The results for stuck TBMs related to larger boulder volume (or more total boulders) are similar to those for maximum boulder size. However, of significance is that 5 of 17 or 29 percent of the stuck TBMs occurred when estimated boulder volumes (frequencies) were small. Four of 5 or 80% of these cases involved STBMs only with drag cutters. 2.3

Remedial measures if TBM gets stuck

If a TBM becomes stuck because of cobble and boulder obstructions, a relief or recovery shaft or tunnel may be required to remove the obstruction and repair or remove the TBM, unless the TBM has face access. Case 6 (Miller, 1995a) describe how a STBM with roller cutters, but without face access was able to unjam itself and avoid being stuck by use of a TBM

60 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-06.qxd

2/18/04

8:17 PM

Page 61

stuck on this project also exceeded US 3 million and resulted in a completion delay of nearly one year. For Case 19, the original TBM (a STBM with drag cutters) was removed within a hand-mined relief tunnel. Problems with groundwater control and flowing ground required significant compensation and permeation grouting resulting in over three months delay and US $600,000 in contractor expenses. After the boulder obstruction was removed, the contractor found that site friction setup had caused the pipe string to also become stuck.

34). Tunneling was then completed without TBM removal. In one case (No. 14), an open-face TBM was stuck on a very large unanticipated boulder that resulted in significant delays before permitting and preparation for blasting could be achieved. 2.4

Cost and schedule consequences of getting stuck

Specifications and planning should be focused on minimizing the chances of a stuck TBM, particularly if the consequences are severe. The consequences of getting stuck should be evaluated on every project. The delay and cost consequences of getting stuck can be very high when one or more of the following constraints would prevent or complicate access for boulder or TBM removal by a recovery shaft or tunnel:

• • • • • •

3 CONCLUSIONS 3.1

Buildings, railroads, restricted access roadways, utilities or other facilities. Environmental impact restrictions below water courses, wetlands or other protected areas. Presences of contaminated ground or groundwater, particularly if hazardous waste is involved. Tunnels deeper than about 15 m. Large potential damages for delay of completion. Boulder conditions are too severe to continue with TBM requiring another TBM to be mobilized and launched.

This study of microtunnel and small diameter TBM encounters with boulders clearly shows that microtunnel boring machines equipped with drag cutters and no face access (Figure 6) have a high risk of becoming stuck when boulders are encountered. The risk of becoming stuck increases as the maximum boulder size or frequency increases. The risk of a STBM or EPBM becoming stuck decreases if roller cutters replace or are combined with drag cutters (Figure 7). Microtunnel boring machines that were equipped with roller or a combination of roller and drag cutters became stuck approximately one third to one fourth less often.

A row listing approximate or roughly estimated cost consequences from getting stuck or potentially getting stuck is provided in Table 1. Many if not most of the estimated cost consequences are probably higher than listed. More accurate information on cost consequences of becoming stuck was available for Cases 1, 4, 16 and 19. For Case 1a, the original TBM (a STBM with drag cutters) and over 150 m of jacked pipe had to be abandoned and replaced with a new launch shaft and STBM with roller and drag cutters (Case 1b). The additional cost was over US $3 million and approximately one year of delay. For Case 4, a STBM with drag cutters became stuck below a multi-lane interstate highway. A recovery tunnel was excavated to remove the badly damaged TBM and finish the drive. The additional costs were reported over US $400,000. For Case 16, the cutters and cutting wheel of the original TBM (a STBM with drag cutters) was totally replaced after becoming stuck the first time. After becoming stuck the second time, the TBM was removed and sent back to the manufacturer. It was replaced with an OFRW machine having face access for boulder blasting, but also which required expensive dewatering to provide suitable standup time and minimize settlement damage. The cost of becoming

Figure 6. STBM with drag cutters and no face access.

61 Copyright © 2004 Taylor & Francis Group plc, London, UK

The risk of getting stuck on boulders is highest with a TBM equipped with only drag cutters and no face access

09096-06.qxd

2/18/04

8:17 PM

Page 62

3.3

Soil matrix shear strength is important, but boulders in weak soil may not result in a stuck TBM

The ratio of rock shear strength to soil matrix shear strength is an important factor, but having a ratio greater than 600:1 may not result in a stuck TBM, particularly if the cutting wheel is well armored to resist abrasion and some strawberry cone or multi-row carbide insert cutters are included on the cutting wheel. The STBMs and EPBMS used for Case Nos. 1b, 2, 10, 35 and 36 had combined roller and drag cutters and were able to handle boulders without becoming stuck even though soft or loose soil zones were reported that probably resulted in shear strength ratios less than 600:1. Even if the TBM does not get stuck, which is important, the risk of cutter damage and much slower progress than desired may result when boulders are encountered in a soft or loose matrix than if the boulders are embedded in stiffer or denser ground that is capable of holding boulders for effective fracturing. Goss, 2002 listed four case studies where disc cutters failed to fracture boulders as desired. If a more rugged TBM with roller cutters is selected for handling boulders in weak ground, the potential costs of higher cutter and cutting wheel wear, slower productivity and higher risk of getting stuck should be carefully compared to the costs and risks of other tunneling alternatives such as open-face tunneling with face stability provided by compressed air, grouting or dewatering. Figure 7. Rugged STBM with drag cutters and roller cutters.

3.2

3.4

Multiple or redundant boulder handling capabilities should result in the least risk of getting stuck

The potential consequences of getting stuck should be carefully considered when specifying or selecting the TBM. In many cases the cost of mobilizing a more rugged TBM with more boulder handling features is well worth the reduction in risk of getting stuck that results. In some situations the cost of getting stuck exceeds several million dollars.

Bouldery ground results in a risk of getting stuck that can be significantly reduced by careful selection of TBM components, but remains a risk that cannot generally be eliminated. The least risk of becoming stuck results if multiple, redundant TBM features are provided so that more than one method can be used if the primary method of handling boulders does not work sufficiently. For example, a STBM with drag cutters and no face access allows few if any alternatives other than a recovery shaft or tunnel if it becomes stuck. If roller or roller and drag cutters are used, the STBM has capability of fracturing boulders larger than those ingestible and being further fractured by a cone crusher. If capability for cutting wheel retraction is added, jammed cobbles and boulders are more likely to be handled without outside intervention. If an option exists for face access or back-loading cutters for manual boulder fracturing when needed or cutter changing, the risk of getting stuck would be even less.

REFERENCES Abramson, L., Cochran, J., Handewith, H. & MacBriar. 2002. Predicted and actual risks in construction of the Mercer Street Tunnel. In Ozdemer, L. (ed). Proceedings Of The North American Tunneling 2002: 211–218. Rotterdam: Balkema. Becker, C. 1995. The Choice Between EPB- and Slurry Shields: Selection Criteria by Practical Examples. In Williamson, G.E. & Fowring I.M. (eds). Proceedings, 1995 Rapid Excavation and Tunneling Conference. Chapter 31, 479–492. Littleton CO: SME. Beieler, R., Gonzales, D. & Molvik, D. 2003. City of Seattle – Tolt Pipeline No. 2 Bear Creek and Snoqualimie River

62 Copyright © 2004 Taylor & Francis Group plc, London, UK

The consequences of getting stuck should be carefully considered

09096-06.qxd

2/18/04

8:17 PM

Page 63

Microtunnels. Proceedings of North American No-Dig 2003, NASTT, Las Vegas, April, Paper A-2-03. Bennett, D. & Wallin, M.S. 2003. American River Crossings: Then and Now. Proceedings of North American No-Dig 2003, NASTT, Las Vegas, April, Paper D-2-02. Boscardin, M., Wooten, R.L. & Taylor, J.M. 1997. Pipe Jacking to Avoid Contaminated Groundwater Conditions. In Proceedings of Trenchless Pipeline Projects – Practical Applications:135–141. Boston, MA. June 8–11, 1997. ASCE Pipeline Division. Budd, T.H. & Cooney, A.M. 1991. Milwaukee’s North Shore 9 Collector System – A Case History. In Wightman, W.D. & McCarry, D.C. (eds). Proceedings, 1991 Rapid Excavation and Tunneling Conference: 349–378. Society for Mining Metallurgy and Exploration, Littleton, CO. Castro, R., Webb, R. & Nonnweiler, J. 2001. Tunneling Through Cobbles in Sacramento, California. In Hansmire, W.H. & Gowring, I.M. (eds). Proceedings 2001 Rapid Excavation and Tunneling Conference: 907–918. Littleton, Colorado: SME. Cigla, M. & Ozdemir, L. 2000. Computer Modeling For Improved Production of Mechanical Excavators. In Proceedings of Society for Mining, Metallurgy and Exploration (SME) Annual Meeting, Salt Lake City, UT, February 2000. Clarke, I. 1990. Carronvale Sewer Project. No-Dig International. April 1990: 17–19. Coss, T.R. 1993. Pascal Tunnels Under New Denver Airport. Trenchless Technology, May/June. 1993:41–42. Cronin, H.E. & Coluccio, J.J. 2003. The True Cost of Boulders in a Soft Ground Tunnel. 2003. In Robinson, R.A. & Marquardt, J.M. (eds), Proceedings 2003 Rapid Excavation and Tunneling Conference. Littleton, Colorado: SME: 535–539. DiPonio, D.D., Manning, F.B. & Alberts, J.B. 2003. An Encounter with Bolulders During Soft Ground Tunneling in Wayne County, Michigan: A Case History. In Robinson, R.A. & Marquardt, J.M. (eds), Proceedings 2003 Rapid Excavation and Tunneling Conference. Littleton, Colorado: SME: 522–534. Dowden, P.B. & Robinson, R.A. Coping with Boulders in Soft Ground Tunneling. 2001. In Hansmire, W.H. & Gowring, I.M. (eds). Proceedings 2001 Rapid Excavation and Tunneling Conference: 961–977. Littleton, Colorado: SME. Ellis, M. 2003. Northeast Interceptor Sewer – Libertyville, Illinois. Westcon.net. Fleet, J. & Owen, D. 2002. Difficult Drives in Dundee. World Tunneling, 2002:216–218. Frank, G. & Daniels, J. 2000. The Use of Borehole Ground Penetrating Radar in Determining the Risk Associated With Boulder Occurrence. In Ozdemir, L. (ed). Proceedings Of The North American Tunneling 2000: 427–436. Rotterdam: Balkema. Frank, G. & Chapman, D. 2002. “Geotechnical Investigations for Tunneling in Glacial Soils,” In Hansmire, W.H. & Gowring, I.M. (eds). Proceedings 2001 Rapid Excavation and Tunneling Conference: 309–32, Littleton, Colorado: SME. Friant, J.E. & Ozdemir, L. 1994. Development of the High Thrust Mine-Disc Cutter for Microtunneling Applications. No-Dig Engineering. June 1994: 12–15.

Garret, R. 1992. Refined Solutions at Indianapolis. North American Tunneling Supplement to World Tunneling. May 1992: N27-N30. Gehlen, H. & Hunn, C. 2002. Microtunneling in the Scottish Highlands. Trenchless Technology International. Aug 2002: I-16 – I-17. Genzlinger, D.D. 1995. Teamwork Overcomes Tunneling Difficulties. Trenchless Technology. Feb. 1995: 34–36. Goss, CM. 2002. “Predicting Boulder Cutting in Soft Ground Tunneling,” In Ozdemer, L. (ed), Proceedings Of The North American Tunneling 2002: p37–46. Rotterdam: Balkema. Hunt, S.W. 1996. Evaluation of Represented and Encountered Subsurface Conditions to Determine Merit of Differing Site Conditions Claims – Elgin Interceptor. STS Consultants Report to Michels Pipeline Construction Company. November 20, 1996. Hunt, S.W. & Angulo, M. 1999. Identifying and Baselining Boulders for Underground Construction. In Fernandez, G. & Bauer (eds), Geo-Engineering for Underground Facilities: 255–270. Reston, Virginia: ASCE. Hunt, S.W., Bate, T.R. & Persaud, R.J. 2001. Design Issues For Construction of a Rerouted MIS Through Bouldery, Gasoline Contaminated Ground, In Proceedings of the 2001 – A Collection Systems Odyssey Conference. Session 6. Alexandria, VA: Water Environment Federation, Inc. Hunt, S.W. 2002. Compensation for Boulder Obstructions. In Ozdemir, L. (ed), Proceedings Of The North American Tunneling 2002: 23–36. Rotterdam: Balkema. Hunt, S.W. 2003. MWH Files: MN-320 Project, Minneapolis, Minnesota. Jones, M. 1990. Sudden Valley Sewer Project. No-Dig International. April 1990: 22–24. King, J., Najafi M. & Varma, V. 1997, Pipe Jacking Operation Completed in Flowing Ground. Trenchless Technology, Sept. 1997: 88–89. Mazhar, F. 1995. Flood Control and Combined Sewer Overflows. Harza Engineering Company Project Profile. 1p. Miller, P. 1994a, Microtunneling Delivers Transmission Crossings. Trenchless Technology, Aug. 1994: 36–37. Miller, P. 1994b. Pipe Jacking Delivers Nearly Two-Mile CSO Sewer. Trenchless Technology, Dec. 1994: 26–27. Miller, P. 1995a. Nova Battles Waves for Seawater Recovery Tunnels. Trenchless Technology, Oct. 1995: 26–28. Miller, P. 1995b. L.J. Keefe Conquers Squeezing Clays for Relief Sewer. Trenchless Technology, Dec. 1995. Miller, P.J. 1996. West Coast Microtunneling Finds Niche. Trenchless Technology, Mar. 1996: 40–43. Miller, P. 1997a. MT Project Uses Pipe Array. Trenchless Technology, Feb. 1997: 28–29. Miller, P. 1997b. San Diego Project Marks Technologies and Teamwork. Trenchless Technology, Oct. 1997: 22–24. Molvik, D., Breeds, C.D., Gonzales, D. & Fulton, O. 2003. Tolt Pipeline Under-Crossing of the Snoqualmie River. In Robinson, R.A. & Marquardt, J.M. (eds), Proceedings 2003 Rapid Excavation and Tunneling Conference. Littleton, Colorado: SME: 396–403. Najafi, M. & Varma, V. 1996, Two Firsts for Iowa – Microtunneling and RCPP. Trenchless Technology, Dec. 1996:36–37. Navin, S.J., Kaneshiro, J.Y., Stout, L.J. & Korbin, G.E. 1995. The South Bay Tunnel Outfall Project, San Diego,

63 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-06.qxd

2/18/04

8:17 PM

Page 64

California. In Williamson, G.E. & Gowring, I.M. (eds). Proceedings 1995 Rapid Excavation and Tunneling Conference: 629–644. Littleton, Colorado: SME. Nishitake, S. 1987. Earth Pressure Balanced Shield Machine to Cope with Boulders. In Jacobs, J.M. & Hendricks R.S.. (eds). Proceedings, 1987 Rapid Excavation and Tunneling Conference. Chapter 35,. 552–572. Littleton CO: SME. Ozdemir, L. 1995. Comparison of Cutting Efficiencies of Single-Disc, Multi-Disc an Carbide Cutters for Microtunneling Applications. No-Dig Engineering. March 1995. 18–23. Rickert, W.R. & Galantha, M.A. 1999. Northeast Interceptor Meets Future While Respecting the Environment. Public Works. April 1999: 50–53. Rush, J.W. 2000. West Coast Microtunneling. Trenchless Technology. Apr. 2000: 28–31. Rush, J.W. 2002. Microtunneling Key to California Earthquake Repair Project. TBM Tunnel Business Magazine. Aug. 2002: 22–23. Rush, J.W. 2002. Northwest Boring Completes WorldClass Microtunnel. Trenchless Technology. Oct. 2002: 28–31.

Schumacher, M. & Ellis, M. 1997. Conquering Glacial Till in Ames, Iowa, Proceedings of No-Dig ’97: 455–461. NASTT, Seattle, April 1997. Session 4B-3. Smith, M. 1995. NEHLA Undershore Crossing. North American Tunneling. June 1995: N16-N20. Staheli, K., Bennett, D., Maggi, M.A., Watson, M.B. &. Corwin, B.J. 1999. Folsom East 2 Construction Proving Project: Field Evaluation of Alternative Methods in Cobbles and Boulders. In Fernandez, G. & Bauer (eds). Geo-Engineering for Underground Facilities: 720–730. Reston, Virginia: ASCE. Staheli, K. & Duyvestyn, G. 2003. Snohomish River Crossing: Bring on the Boulders, Success on the Second Attempt. Proceedings of North American No-Dig 2003, NASTT, Las Vegas, April, Paper B-4-03. Tarkoy, P.J. 2001, Challenges & Successes in MicroTunneling on the Chelsea River Crossing. Proceedings of 5th International Microtunneling Symposium – BAUMA 2001. 16p. Vadnais, P. 2002. Personal Discussion with Steve Hunt on Marinette Water Main River Crossing Project. Wallis, S. 2002. Remotely controlled passage under the Neva. World Tunneling. Feb. 2002: 25–27.

64 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-07.qxd

2/18/04

8:18 PM

Page 65

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Joint orientations for TBM performance analysis using borehole geophysics to orient rock cores T. Tharpe, B. Crenshaw, and J. Raymer Jordan, Jones & Goulding, Inc., Norcross, Georgia

ABSTRACT: A borehole televiewer was used to orient rock core for joint analysis as part of the Atlanta CSO Tunnel Geotechnical Investigation. The Atlanta CSO tunnels are about 30 feet in diameter and will be bored through medium grade gneiss and schist. The televiewer is a wireline geophysical tool that uses sonic waves to map the traces of individual joints around the inside of the borehole wall. From these traces, the televiewer software calculates the strike, dip, and aperture of each joint. The televiewer was used in 51 core holes averaging 300 feet in depth, and was much less expensive, faster, and more accurate than using oriented core. The televiewer provided joint characteristics, an acoustic velocity log to indicate areas of weathered or blocky ground where core recovery is typically poor, and a graphical picture of the borehole that can be used to orient the core for more detailed analysis. Each joint in the televiewer data was correlated to the core and classified in terms of RMR and Q parameters. Machine breaks and core damage were easily recognized because they occurred only in the core, but were not read by the televiewer in the borehole wall. A stereoplot of the joint data was made for each borehole and for the project as a whole. These stereoplots were used for three purposes: (1) kinematic wedge analysis and support design; (2) classifying the ground into different baseline types; and (3) estimating the potential benefits of the fractures on TBM performance using the Norwegian Fracturing Factor criteria.

During this process, all joints were described in terms of type, fit, roughness and planarity, and alteration. Orientation and geotechnical property data gathered during the rock core fracture analysis was used for numerous aspects of the Atlanta West Area CSO Project. Orientation data was used to generate stereonets for the Geotechnical Data Report (GDR) as well as stereonets used in various analyses conducted for the purposes of design and baselining. Geotechnical property data was used to aid in prediction of ground conditions along the tunnel alignment.

1 INTRODUCTION A digital acoustic televiewer (DATV) was utilized as a part of the City of Atlanta Combined Sewer Overflow (CSO) Tunnel geotechnical investigation. The geotechnical investigation for the CSO Project included geological mapping, 77 core borings, geotechnical analysis of the core, laboratory testing of the rock properties, and DATV logging. Televiewer logging was conducted in 51 of the 77 completed core borings. The DATV is a geophysical tool that provides highresolution data that can be used to determine dip direction and dip angle of planar features intersecting a borehole (Keys, 1990). Both travel time and amplitude of the acoustic signal are recorded and displayed in real time. The data is analyzed, and orientation of downhole features is recorded. The orientation of the joints delineated from the televiewer data was transferred to corresponding joints in the rock core. The rock core was then oriented to North based on identified joint azimuths, and a North line was scribed on the core. Fractures not oriented from televiewer data (low angle fractures, healed fractures, microfractures, etc) were identified and oriented from the geophysically derived northern orientation.

2 PROJECT DESCRIPTION Like many older cities of its size, Atlanta’s sewer system consists of areas where stormwater and sanitary sewer flows (wastewater), are collected in the same pipe, known as a combined sewer system; and areas where they are collected in separate pipes, known as a separated system. Approximately 85 percent of Atlanta’s system is separated, mainly in residential areas that have developed within the last 75 years. The remaining 15 percent of the system, the older section that serves the core of the City, consists of a

65 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-07.qxd

2/18/04

8:18 PM

Page 66

commonly weathered away, leaving zones of broken, water-bearing rock that are more easily eroded to form topographic valleys and draws. Gouge and shattered rock are present in some of the lineaments. Also found in the Atlanta region are large, subhorizontal stress relief fractures (Cressler et al., 1983). These fractures may be visualized as a low arch in cross section with the largest opening occurring at the midpoint. Fractures are typically open and are very transmissive. For this reason, they may produce a large flow of water when encountered in wells or tunnels. A flushflow of approximately 2,000 gpm was encountered in the Atlanta Three Rivers Tunnel when a stress relief fracture was intercepted. Geophysical information was useful in recognizing and characterizing these features along the West Area CSO alignment because these fractures are not observable at the surface.

combined sewer system. It was constructed in the late 1800s through 1920 to carry both stormwater and wastewater to a treatment facility, where treatment would occur before the water was discharged into the Chattahoochee River. During dry weather, all flows are conveyed to the treatment facility. When rainfall occurs, flows are still conveyed to the treatment facility. However, during larger rainfall periods, sewer capacity is sometimes exceeded, resulting in portions of the flow being diverted from the wastewater treatment facility to a combined sewer treatment facility that provides a lesser degree of treatment (screening and disinfection). This condition is known as a Combined Sewer Overflow (CSO). The CSO Storage and Treatment System plan involves capturing, storing and conveying CSOs. The overflows are stored in large underground tunnels in bedrock. The captured CSO volume is conveyed to a separate treatment system for removal of suspended solids and other pollutants, undergoing disinfection before discharge to receiving waters. The plan has two components: the construction of the new West Area CSO Storage Tunnel and the East Area CSO Storage Tank. The West Area CSO Project consists of two large diameter tunnels, one smaller diameter connecting tunnel, three intakes, one pump station, one overflow shaft and tunnel, and four additional construction shafts. The two large diameter tunnels are the North Avenue Tunnel and the Clear Creek Tunnel. Both of the large diameter tunnels are about 27 feet in excavated diameter and about 24 feet in finished diameter. Both tunnels will be excavated at an average depth of 200 feet below land surface. The North Avenue Tunnel is approximately 23,333 feet long, and the Clear Creek Tunnel is approximately 20,783 feet long.

4 GEOPHYSICAL INSTRUMENTATION An acoustic televiewer provides a digital record of the location, character, and orientation of any features in the casing or borehole wall that alter the reflectivity of the acoustic signal. These include diameter and shape of the borehole, drilling or lithology induced rugosity, differences in rock hardness, and structural features such as bedding, fractures, and solution openings. The acoustic televiewer provides a magnetically oriented, 360-degree, image of the acoustic reflectivity of the borehole wall (Keys, 1990). Because the collected data is spatially oriented, it can be used to calculate the dip azimuth and dip direction of planar featured that intersect the borehole. 4.1

The digital acoustic televiewer tool used for this study utilizes an acoustic transducer that operates at a frequency of .5-MHz. The acoustic transducer, which functions as a transmitter and receiver, rotates at 12 revolutions/second and digitizes 256 data samples per revolution, which allows for the collection of highresolution data. The high frequency induced by the acoustic transducer is reflected from the borehole wall and is received by the instrument. An internal flux-gate magnetometer is triggered each time the acoustic transducer rotates past magnetic north. The signal from the magnetometer is transmitted to the recording equipment at land surface (Keys, 1990).

3 BEDROCK DESCRIPTION The West Area CSO Tunnel is located in the Piedmont region of the southeastern United States. The ground along the West Area CSO Tunnel generally consists of medium-grade metamorphic rocks that have been intruded by granitic rocks in some places. The bedrock in the project area has undergone intense deformation, weathering, erosion, and some regional uplifting. The bedrock is overlain by approximately 15 to 120 feet of soil and partially weathered rock. Lineaments are a common feature in parts of the Piedmont, including the area of the Atlanta West Area CSO Tunnel. Lineaments in the Piedmont are typically long, narrow topographic valleys and draws. In many cases, these lineaments represent surface expressions of subsurface features, such as fracture zones. These fracture zones may become cemented with minerals at depth. At shallower depths, these cements are

4.2

Field methods

Upon completion of the coring process, the resulting borehole is logged with geophysical instruments. Before DATV logging is conducted, it is necessary to verify that the tool’s internal magnetometer is triggering

66 Copyright © 2004 Taylor & Francis Group plc, London, UK

Instrument function

09096-07.qxd

2/18/04

8:18 PM

Page 67

be achieved if the rock coring rig vacated the drilling site, followed by the immediate geophysical logging of the core hole, and then followed by the grouting of the borehole by a second crew dedicated only to borehole grouting. This procedure may not be possible for all projects, and stand-by charges, from drilling contractors, may be incurred during the logging of the borehole. When initiating a logging program, site access issues must be considered. For the West Area CSO project, a two-wheel drive vehicle was utilized for logging services, and the project was conducted in an urban environment, which allowed for easy access to nearly all of the drilling locations. However, some locations could not be accessed due to the terrain and vehicle limitations. Data collected in geologic settings where the geology is very homogeneous, and where regional dip is very shallow, is difficult to use for rock core orientation. In homogenous, shallow-regional-dip geologic settings, there may be very few distinctive planar features that intersect a given borehole. In order for the rock core orientation process to be successful, a number of distinctive features, which alter the acoustic reflectivity of the borehole wall, must intersect the borehole. In sections of featureless core, the orientation process is not possible unless strong geologic layering is present. If the geologic features can be delineated through data analysis, then orientation can be accomplished in very competent rock that is not fractured. In the geologic setting in which the City of Atlanta CSO project took place, gneissic banding in the bedrock is often very prevalent. In instances where competent bedrock, devoid of fracturing, was encountered, gneissic banding was used to orient sections of rock core. During the course of this project, it was determined that it is difficult to use shallow dipping planar features to orient rock core. When delineating planar features using the televiewer data, the margin of error for strike and dip determination increases with features that have a dip of less than 5°. Although shallow dipping features can be delineated with the DATV, it is problematic to transfer this data to the rock core and should be avoided. Additionally, when examining both the DATV data and the rock core, steeply dipping planar features are more easily and quickly identified than are shallow dipping features. Thus, the rock core orientation process is simplified by first using distinct, steeply dipping fractures (dip 20°) to establish the north axis on the core, followed by the manual measurement of shallow dipping fractures that exist in the core. The collection and interpretation of DATV data can be complicated by a number of factors. Of these factors, incorrect gain settings and insufficient tool centralization within the borehole severely impact the quality and usefulness of the collected data. The resolution of the collected data is determined by the gain setting. If the gain setting is too low then the acoustic

Figure 1. Schematic diagram for logging setup used during West Area CSO Project investigation.

on magnetic north, and an electronic calibration file is downloaded to the tool before logging commences. After it has been verified that the tool is functioning properly, the televiewer is advanced down the borehole with an electric drawworks utilizing a fourconductor wire line (Figure 1). The wire line is secured to the geophysical instrument by a watertight locking cable head. Once the tool has reached the terminal depth of the borehole, the logging procedure begins. The borehole is logged from the bottom of the core hole up to land surface at approximately 5 feet/minute. While logging, the data is recorded and displayed in real-time using a laptop computer. 5 CONSIDERATIONS FOR DATV LOGGING In part, the geologic setting in which a project or study area lies will determine the usefulness of DATV data for the purposes of orienting rock core. A geologic setting that is conducive to DATV logging is needed. Excellent televiewer data can be collected in a competent, fractured bedrock environment such as that of the Piedmont of Georgia. Televiewer data collected in the Piedmont of Georgia often shows distinctive, open, weathered and steeply dipping fractures that are easily identifiable in the core. The logging speed for the DATV tool is approximately 5 feet/minute. Depending on the investigation depth for a given geotechnical study, it could take two to three hours to complete a round of DATV logging. Time was a consideration for the West Area CSO project, thus, it was determined that a substantial time savings could

67 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-07.qxd

2/18/04

8:18 PM

Page 68

formation to remain mostly intact. Thus, the televiewer can be used to determine fracture orientation in highly fractured and contorted areas where orientation data cannot be determined from rock core samples. Also, drilling induced mechanical breaks and post drilling core damage are easy to identify because these features are not present in the televiewer data. Thus, data quality from the analyses of the rock core is improved because mechanical and post drilling core breaks are more easily discarded from the data set.

signal will be too weak to accurately detect features within the borehole. Conversely, if the gain setting is too high, the borehole will be flooded with acoustic signal, and data resolution will be greatly decreased. Poor tool centralization will also decrease data resolution. The acoustic signal must be introduced perpendicular to the borehole wall. If the televiewer is not centered within the borehole, travel time of the signal is altered because the acoustic transducer will either be closer to, or farther away from, the borehole wall. The DATV is a complex geophysical tool. A highly experienced equipment operator, or geophysical contractor, is needed to perform the logging services. Geophysical equipment is composed of complex electronic components. If a problem with the equipment should arise, a great deal of time could be spent trying to diagnose and repair the source of the malfunction. Additionally, it is possible for an inexperienced operator to collect data, which at first glance seems reasonable, but may prove to be useless once the data analysis process is initiated. An experienced geophysicist is also needed not only for data integrity, but for data interpretation as well. In order to use orientation data, derived from geophysics, to orient rock core, it is necessary to have a skilled geophysicist complete a thorough interpretation of the collected data. The geophysicist must be accustomed to analyzing geophysical data collected in complex geologic settings. It is also helpful if the geophysicist has knowledge pertaining to tunneling or underground construction and is able to recognize and delineate subsurface planar features that are important in tunnel construction.

7 CORE ORIENTATION USING GEOPHYSICAL DATA To collect data on healed fractures, to describe joint characteristics, and to verify geophysical data, rock core was oriented using features from geophysical logs. The methodology used to orient rock core is outlined in the following sections. 7.1

Where a is arc length (in the same units as r), r is the radius of the rock core, and is the azimuth of the down dip expression of the feature in degrees. For the Atlanta West Area CSO Tunnel, HQ core was used with a diameter  61.1 mm (r  30.55 mm). The calculated arc length (a) is the distance from the down dip expression of the feature to North measured counter clockwise around the core when the core is viewed from the top (Figure 2). A core trough was used to orient individual core runs (core runs were 10 feet for the project) and the North orientation was marked for the length of the run with a magic marker.

6 POSITIVE ASPECTS OF DATV USE During the drilling process, rock cores were removed from the inner collection tube, placed in wooden core boxes, and then labeled. The boxes were then transported to a warehouse facility where the core could be analyzed. When transferring the core to the boxes, and during transport of the core, the samples may be disturbed. During transport, core samples obtained in heavily contorted zones had a tendency to crumble and take on the appearance of gravel. These contorted zones are often characterized by fracture sets with a dip of 20°. Joint orientations within these zones were often highly varied. Utilizing DATV, it is possible to determine fracture orientation in highly contorted zones, which are often weathered, or in zones where geologic unloading or exfoliation has occurred. Additionally, the orientation of planar features within these zones can be determined because the televiewer is viewing the contorted zones in situ. The drilling process somewhat disturbs the investigated area, however, lithostatic and hydrostatic pressure enables the

7.2

Rock core orientation

The orientation process began by identifying a distinct feature within the cored interval to be oriented (the bottom 160 feet of the borehole). A distinct feature is defined as a joint that can be identified definitively within the core based on televiewer data. Distinct joints include, but are not limited to, the following:

•

68 Copyright © 2004 Taylor & Francis Group plc, London, UK

Method overview

To orient the core, features recorded on the televiewer geophysical log were identified on the rock core. The down dip azimuth from the televiewer log was then used to orient the down dip expression of features identified on the rock core. The orientation of the down dip expression of the feature was used to orient the core to North. The core was oriented to North using a diameter tape measure and the following conversion:

A single joint within a zone of otherwise featureless rock core

09096-07.qxd

• • • •

2/18/04

8:18 PM

Page 69

Each identified feature was assigned an identification number. The rock core was marked with the identification number and televiewer data. Recorded televiewer data included the dip, dip azimuth, up dip depth, and down dip depth. After identification and verification of initial orientations, the process was carried out again on adjacent runs. For every run, an attempt was made to identify and orient at least one feature. North orientations between identified features and adjacent oriented runs were then compared. If no feature could be identified in a run, the North orientation was carried on from an adjacent oriented run by aligning the ends of runs.

A joint that is in close proximity to a unique feature, such as a series of partially penetrating joints, a crushed or soil zone, vugs, etc. Cross cutting joints A joint which terminates into another joint A series of closely spaced, similarly oriented joints.

After identification of a distinct feature, the down dip expression of the feature was identified and the dip azimuth was used to mark North on the rock core, as described previously. Care was taken to align rock core pieces based on the fit of fractures and breaks. When first beginning to orient a borehole, a series of closely spaced joints was the preferred distinct feature. Identification of a number of closely spaced joints allowed several features to be oriented and north to be compared between them. After aligning the core in the core trough, a line was marked for the length of the run along the 0° (north) azimuth (Figure 3).

8 APPLICATION OF ORIENTED DATA Oriented core data was used in various analyses for the Atlanta West Area CSO Project. For design purposes, orientation data was used to conduct wedge analyses for design of support methods to be used in tunnel segments. Joint descriptions collected during joint orientation were used to calculated Q and RMR values and used to predict ground conditions along the tunnel alignment. Additionally, core orientation data was used to make stereonets for all oriented boreholes along the tunnel alignment, showing adjacent tunnel azimuth, for the Geotechnical Data Report (GDR). Orientation data was also used in baseline data interpretations. An analysis to determine the fracturing factor (ks-tot) was done based on the work of Bruland (1998). A detailed description of the use of oriented core data in the determination of ks-tot values is presented in the following sections.

Figure 2. Three-dimensional view of rock core demonstrating how a dip azimuth from geophysical data is transferred to the rock core and North is established. An example calculation is provided.

8.1

Fracture factor (ks-tot) is an index value used to estimate the benefit that fractures provide to Tunnel Boring Machine (TBM) performance. The value is determined based on fracture spacing (St class) and the orientation of fractures to the tunnel axis (). The methodology for calculating fracture factor was developed by NTNU and is described by Bruland (1998). The method is empirically based and was developed using post-excavation tunnel mapping. The method has been used successfully for calculating fracture factor values from rock for this and other tunnels in the Atlanta area (Dollinger, 2002). All variables needed to calculate fracture factors for the Atlanta West Area CSO Tunnel were determined during the core orientation and analysis process. These variables included fracture orientation and the location of the fracture along the borehole. The Atlanta West Area CSO Tunnel will consist of two intersecting tunnel alignments, the North Avenue Alignment and the Clear Creek Alignment. These two alignments will be bid separately, therefore, separate fracture factor

Figure 3. Picture showing fracture that is being used to orient rock core. The down dip expression has been identified (indicated by arrow on core pointing to left), the fracture has been numbered and geophysical data recorded, and the appropriate rotation has been calculated to orient the core to North. A geologist marks the North orientation along the entire core run along the core trough.

69 Copyright © 2004 Taylor & Francis Group plc, London, UK

Fracture factor calculation

09096-07.qxd

2/18/04

8:18 PM

Page 70

analyses were conducted for each alignment. The following sections will refer specifically to the North Avenue Alignment; the analysis of the Clear Creek Alignment was carried out using the same methodology. 8.2

Determination of fracture sets

Fractures sets were determined using a stereonet created using oriented rock core data along the North Avenue Alignment. These boreholes were located to sample random, non-biased subsurface conditions along the alignment. This method insured fractures represented typical conditions along the tunnel alignment. The stereonet included all fractures that were open or healed with weak cementing minerals, such as soft zeolites. All fractures were weighted by the secant of the dip to the core to account for the orientation bias of vertical boreholes. A stereonet of fracture orientations obtained through the analysis of geophysical data was compared to the stereonet created from oriented core analysis (Figure 4). These stereonets compare favorably when analyzing fracture sets. Variation in fracture set orientations is generally low, within approximately 10–15° for dip azimuths and dips. Fracture set intensity varies due to the criteria used to determine which fractures would be represented. Geophysical data (Figure 4a) represents only open fractures and fracture Set 1, foliation, is more strongly represented. Core data (Figure (4b) considers both open and weakly healed fractures, it includes more high angle fractures which are generally filled, and therefore fracture Sets 2 and 3 are more strongly represented. A visual analysis of the stereonet was used to group fractures into major fracture sets. Two of the major fracture sets identified along the North Avenue Alignment were further divided into fracture sub-sets (Figure 4b). This was necessary because of the highly undulatory nature of the fracture sets along the alignment. Boundary limits in terms of dip azimuth and dip were determined for each set and an analysis was done to separate fractures into Sets 1, 2, and 3. All fractures which fell outside of these boundary conditions were considered to be random, not affecting the ks-tot value, and were not considered further. Sets 1 and 2 were further divided into sub-sets based on boundary dip azimuth and dip conditions. Each fracture within a major fracture set was included in a fracture sub-set. Sub-division of the major fracture sets was necessary to account for the effects of undulations in fracture factor calculations. 8.3

Figure 4. Stereonets of fractures along North Avenue Alignment. a) Open fractures identified from interpretation of geophysical logs. b) Open and weakly healed fractures identified from core orientation. Major joint set boundaries are represented by bold dashed lines and are denoted by numbers. Sub-set boundaries ate denoted by dotted lines and denoted by lower case letters following number designation of major fracture set.

calculating the spacing along the rock core between two adjacent fractures of a set. This distance was then multiplied by the cosine of the average dip of the major fracture set to derive the length of an orthogonal line between two parallel planes. The length of this line is the effective spacing between two fractures.

Calculation of fracture spacing (St classes)

Once fracture sets were established, the spacing between all fractures of a major fracture set within a borehole was calculated. This was done by first

70 Copyright © 2004 Taylor & Francis Group plc, London, UK

Fracture Class (St) O O-I II II III IV

8:18 PM

Page 71

NTNU (cm)

CSO Tunnel spacing (cm)

massive 160 80 40 20 10 5

greater than 200 200 to 100 100 to 50 50 to 25 25 to 12.5 12.5 to 6.3 6.3 or less

IV 4.0

Figure 5. Fracture class classifications used the West Area CSO Tunnel. Adapted from Bruland (1998).

3.0 III-IV

2.0 III II-III 1.0

0.0

The effective fracture spacings were divided into spacing groups corresponding with NTNU St classes (Figure 5). All effective spacing values in each St group were summed and divided by the total effective distance of core evaluated to calculate a relative percentage of rock mass that contains each St class.

8.4

II I 0 0

10

20

30 40 50 60 alpha angle ()

70

80 90

Figure 6. Diagram for estimating fracturing factor (ks) from alpha angle () and fracture spacing (St class). Adapted from Bruland (1998).

The ks values for each major fracture set were used to calculate the ks-tot value for each tunnel segment using the following equation:

Calculation of fracture factor (ks-tot)

Both alignments of the Atlanta West Area CSO Tunnel contain numerous curves, and therefore, various segments of the alignment have different azimuths. Since fracturing factor values are dependent on the relationship between fracture orientations and the tunnel axis, alignments were divided into straight line segments between points of intersections of curves or between a point of intersection and a shaft. The North Avenue Alignment contained ten straight line segments. The ks-tot values were calculated for each straight line segment. To calculate ks-tot, a fracture set fracture factor (ks) value must be determined for each major fracture set. First, an alpha angle was calculated using average fracture orientation data for each fracture sub-set (or major fracture set if no sub-sets were designated) and each tunnel segment azimuth. These alpha angles were used to obtain a raw fracture factor (ks-raw) value for each NTNU St class (Figure 6). The ks-raw values did not take into account the relative abundance of fracture spacing (St classes) within the rock mass or the relative abundance of fracture sub-sets within major fracture sets. The ks-raw values were first adjusted to reflect the relative abundance of the St classes in the rock mass and then summed to obtain a ks-sub value for each tunnel segment. The ks-sub values were adjusted to reflect the relative abundance of the fracture sub-set in the major fracture set and summed to obtain the ks value of the fracture set in each tunnel segment. This step was not necessary for major fracture sets with no fracture sub-sets.

as described by Bruland (1998), where n is the number of fracture sets. The ks-tot values for each tunnel segment were provided in the baseline documents for use by the contractor and were used by the engineer for making TBM performance assessments for use by the owner. 9 SUMMARY Geophysical data was successfully used to orient rock core for the Atlanta West Area CSO Project. A strong correlation between the location of joint groupings was observed for stereonets produced from orientation data derived from geophysical methods and orientations of joints measured from the core. This indicates that features were correctly identified between the geophysical logs and the core, because if the incorrect joints were identified, then the joint groupings on the stereonet derived from core data would have been rotated differently then those from geophysical analysis. The successful implementation of geophysics as an orienting tool allowed for the acquisition of a large amount of orientation data at a reasonable cost to the project. Additionally, joint orientations obtained from geophysics supplemented the core orientations in zones that were too heavily weathered or fractured to measure manually. These orientations would not have been obtainable using traditional orientation methods. Orientations within

71 Copyright © 2004 Taylor & Francis Group plc, London, UK

Fissure Class

2/18/04

Fracturing Factor (ks)

09096-07.qxd

09096-07.qxd

2/18/04

8:18 PM

Page 72

orient core, but also served as a secondary verification of the subsurface conditions that exist along the tunnel alignment in addition to the rock core. This benefit is not associated with any of the traditional orientation techniques. The cost effectiveness of this technique over tradition orientation methods as well as the benefits associated with this technique merit its consideration for many tunneling projects.

these zones are of great benefit to understanding the ground conditions in and around them. These zones are particularly important during tunnel investigations because they represent non-typical conditions that have the greatest potential to negatively impact the project. The large amount of data collected from oriented core allowed numerous, thorough analyses to be conducted for the purpose of tunnel design and contract baselining. Stability analyses were conducted for many regions of the tunnel based on fracture characteristics within each region. From these analyses, regions of the tunnel where wedge failures may occur were identified. Additionally, stability analyses and computer modeling were used to design and test tunnel support. Ground conditions were characterized based on joint characteristics associated with each joint through the calculation of Q and RMR values. Lastly, oriented joint data was used to calculate fracture factors for each segment of the West Area CSO Tunnel. These values quantify the benefit of fractures to TBM performance. The use of geophysical televiewer data proved to be a viable, cost-efficient core orientation option for the CSO Tunnel. Not only did it provide a means to

REFERENCES Bruland, Amund, 1998, Hard rock tunnel boring: Advance rate and cutter wear: Trondheim: Norges teknisknaturvitenska-pelige universitet. Cressler, CW, Thurmond, CJ, and Hester, WG, 1983, Ground water in the greater Atlanta region, Georgia: Information Circular 63, Georgia Geologic Survey. Dollinger, Gerald L, Raymer, and John H, 2002, Rock mass conditions as baseline values for TBM performance evaluation: in North American Tunneling 2002, Ozdemir (ed), Swets & Zeitlinger, Lisse. Keys, W, Scott, 1990, Techniques of Water-Resources Investigations, U.S. Geological Survey, Book 2, Chapter E-2.

72 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-09.qxd

2/18/04

11:51 AM

Page 73

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Slurry type shielded TBM for the alluvial strata excavation in downtown area Wang Ruel Jee Colorado School of Mines, Golden, CO

ABSTRACT: The soil and rock types relevant of construction lot 9 of Seoul subway line number nine of Yeoi Island area show a distinct difference between the eastern and western part of the alignment especially in the eastern part. The tunnel is expected to be located entirely in alluvial deposits, predominantly in sandy gravels with low strength of alluvial soils with presence of groundwater. For the western part, the geological conditions are considerable more complex, varying between hard rocks, residual soils and alluvium, with the higher proportions in weathered rocks. The alluvial deposits probably dominated by sandy gravels including some boulders are expected to be favorable for the application of a slurry shielded TBM with the higher underground water levels. For a more detailed assessment, reliable grain size distribution curves of the alluvial deposits along the tunnel alignment are investigated to select the optimal selection of separation plant and suitable advance rate of the slurry type shielded TBM. Along the whole alignment closed mode tunneling with active face support will be required, due to the low strength of the alluvial soil and the presence of groundwater. There are two basic options for the machine type to be applied: A Slurry type shielded TBM or Earth pressure balanced shielded TBM. For specific details concerning the technical machine operation and the safety precautions of the tunnel construction on this alluvial deposits are described in this paper.

1.1 Slurry type shielded TBM

1 INTRODUCTION OF THE SHIELDED TBM FOR SHALLOW TUNNEL CONSTRUCTION

Tunneling with slurry type shielded TBMs has been proven to be a safe excavation method causing low settlement in all kinds of loose ground with higher ground water condition. This method has another advantage more easy to handle the boulder problems during the tunnel excavation. The tunnel face is supported by a

Fundamentally, the main aim of geotechnical investigations is to identify the ground conditions for a proposed underground structure. Based on the results of the geotechnical investigation, designers can make decision on the selection of optimal construction methods, and also get the suitable idea of the underground structures as well as to prepare the geomechanical parameters for the numerical analysis of the safety precautions. Since the construction plan of Seoul subway line number 9–9 was established in the shallow alluvial deposits where the subway tunnel’s permanent stability should be guaranteed during the construction periods and also the operation periods continuously, and so it was hard to select the suitable tunneling methods in such poor ground conditions. It was considered that for the entire given alignment, closed-mode tunneling with active face support shall be required, due to the low strength of the alluvial soils and the presence of high level groundwater. There are two machine types to be used in such a poor ground conditions, a slurry type shielded TBM (Fig.1), and an earth pressure balance shielded TBM (EPB), which are discussed below.

Figure 1. Typical layout of the slurry type shielded TBM.

73 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-09.qxd

2/18/04

11:51 AM

Page 74

Hydroshield

Slurry Shield

Figure 3. Separation plant of the Slurry type shielded TBM (Busan, Korea). EPB Shield

Coarse gravel Medium gravel

1 10-1 10-2 10-3

slurry, generally, a suspension consisting of water and bentonite or clay. Through this filter cake, the pressurized suspension in the excavation chamber balances the earth and water pressure. The excavated material mixes with the suspension fluid and it will be pumped to a surface separation plant. Recycled suspension fluid will be fed again into the front working chamber of the shielded TBM. In comparison to other systems, economical usage of this system is mainly determined by the separation efficiency, speed, and suspension requirement, and the permeability of tunneling ground which are directly connected with the advance rate of this tunneling machine and how could they reduce the noise through the separation works in downtown area is the critical problem in urban construction. Slurry shield operates most efficiently in non-cohesive, watersaturated soil where the particle size distribution ranges between coarse silt/fine sand and coarse gravel. However, for operation in fine cohesive soils, it is difficult to perform the special technical installation of the separation plant with complex and expensive operation with higher energy consumption (Fig. 3). The optional installation of hydraulic rock crusher in front of the suction inlet will break boulders (Maximum 500 mm diameter) down to size corresponding to the diameter of the slurry discharge pipes (Diameter: 100–150 mm) (Fig. 4).

Fine gravel

10-4

Coarse sand

10-5 10-6

Fine sand Sandy, silty clay Silt

10-7 10-8 10-9 10-10

Clay

10-11 -12

10

Figure 4. Optimal particle sizes for the better implementation of Slurry & EPB shield.

utilize the material as excavated by the cutting wheel serves as a support medium. The support pressure is mainly influenced by the following two processes (Fig. 2).

• •

The forward thrust of the TBM determines, with cutter head rotation, the volume of excavated material, which is forced into the working chamber by advance of the TBM. The rotation of the screw conveyor determines the removal rate of excavated material from the working chamber.

The excavated material in the screw conveyor is forced to form a “Plug” which acts as a seal against the pressure from the working chamber, and allows the excavated material to be discharged at atmospheric pressure onto a conventional belt conveyor. Transport of the excavated material through the tunnel can be by belt conveyor, track-bound vehicles, dump trucks or, solid handling pumps to the surface.

1.2 Earth pressure balance shielded TBM Apart from the high separation costs and environmental hazards involved, the confined space in a big city like Tokyo, Seoul, EPB machines were developed as a substitute system against the slurry type TBM. Operation principle of EPB is very simple to

74 Copyright © 2004 Taylor & Francis Group plc, London, UK

10

Permeability Factor K [m/s]

Cobbles

Figure 2. Schematic figures showing the principles of the shielded TBM.

09096-09.qxd

2/18/04

11:51 AM

Page 75

Table 1. Comparison of two different shield machine types.

Face support

Earth pressure support

Liquid support

•

Dull regulation and control characteristic of the support pressure, influenced by many parameters Increased risk in soils with high permeability; face support in coarse grain size (sandy gravel with cobbles up to  250 mm, k  1*102 to 1*104 m/s) requires a lot of experience in EPB tunneling Increased blow-out risk in case of low overburden (For example, km 15  500: 1.0 m to pipeline, km 15  520: 2.5 m to surface, km 15  900: 0.7 m to foundations of Seoul Subway Line No.9) High expenditure for soil conditioning especially in gravel

•

•

•

•

• •

Exact and quickly responsive support pressure regulation in unstable face support situations (e.g. due to high permeability) Increased safety against settlements in permeable soils (sands, gravel) and at low overburden due to the liquid face support – lowered blow-out risk Increased safety in case of low distance bridge foundations

Ram loads

•

Increased ram loads due to noncompressible nature of the support medium, possible damage to the lining

•

Moderate ram loads due to cushion effect of slurry (and compressed air)

Material discharge

•

Difficult pressure suppression in highly permeable soils (k  1 102 to 1 104 m/s)

•

Pressure reduction over pipe length

Wear

•

Ground contact of tools and chamber structures on all sides, high wear in particular on screw conveyor and cutter head Uncertainties concerning long-lasting functionality of disks, increased number of interrupts for tool inspection/ replacement expected

• •

Reduced wear on cutter head, tools and chamber structures Wear in tubes and pumps exists but can be controlled

• Time consumptive procedure (clearing the working chamber from soil) • Difficulties in generating a membrane if

•

High stability due to betonies filter cake and air pressure

•

•

For Hydro-/Mixshield: submerged wall gate valve allows work under atmospheric conditions, full face support still maintained

•

Chamber access (face support with pressurized air)

tunnel face is unstable, possible difficulties in restart

Service and repair of screw conveyor/ in suction area

Entire evacuation of working chamber necessary (in particular for works on front spiral), high expenditure for service under atmospheric conditions

To assist in forming the “Plug” the excavated material may require conditioning. The conditioning agents frequently include bentonite, long chain polymers, and foam. Usually the cutterhead of EPB machines has a more closed design as compared to slurry type machines, and voids in the front plate allow for passage of excavated material into the working chamber. As for the boulder handling, the EPB machines cannot be fitted with a rock crusher because that the excavated material has to be mixed within the working chamber to produce a homogeneous support medium, therefore, boulders have to be broken exclusively by roller discs. In order to replace the cutter tools, the working chamber must be emptied from soils and filled with compressed air in the same manner as a slurry type machines. EPB machines are

most in cohesive clayey-silty and silty-sandy soils, which allow the formation of an appropriate earth mixture in the shield chamber with any water conditions. 1.3 Hydroshield TBM The principle of hydroshield is a modified slurry type shield system. The most important design elements of this machine are the separation of the excavation chamber by the submerged wall and the support pressure at the fluid supported tunnel face is regulated. A submerged wall with a gate at the invert of the TBM separates the front compartment from a second chamber, where a compressed air cushion maintains and controls the fluid pressure. Therefore, face support

75 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-09.qxd

2/18/04

11:51 AM

Page 76

settlement during the TBM operation. Actually, the advantage of given face stabilization a TBM based on the hydroshield principle is favored over a slurry type shielded TBM. Therefore, the designer recommended a hydroshield (Slurry) TBM with stone crusher and a cutterhead, allowing to change of cutting tools. In general, the design of TBM is the responsibility of the machine manufacturer according to the given geological conditions. The cutter head of the TBM must be equipped with suitable tools for the whole range of expected soil types. For this project, the capability to cut hard rock must be provided for tunneling in the western section of the project area. Further, in the alluvium tunneling, provisions should be included for breaking up and removal of boulders embedded in the ground matrix. It is the contractors responsibility to identify the ground conditions and has to meet all requirements of the technical specifications. The control of ground movements shall be maintained to minimize ground surface settlement by the loss of ground and potential damage to all adjacent structures and services. The machine characteristics shall comply with the following technical data.

and fluid circulation are decoupled, and so it is possible to compensate for eventual volume imbalance quicker and more accuracy than an ordinary slurry type TBM. The gate in the submerged wall can be closed with a hydraulic sliding door, then, it is available to empty the second chamber for maintenance work on the stone crusher or for clearance of the suction inlet, while the excavation face is still supported by fluid in the front compartment. 1.4 Mixshield TBM The mixshield is a proprietary name used by Herrenknecht in Germany for a shield TBM to be adapted to variable ground conditions. The mixshield provides the possibility to operate the TBM alternatively in EPB modes or in slurry mode. Table 1 shows the two different shield machine types regard to practical machine operation and tunneling safety. 2 OPTIMAL SHIELD MACHINE SELECTION FOR THE ALLUVIAL ISLAND The ground conditions relevant of Yeoi-island section show a distinct geological difference between the eastern and western part of the alignment. In the eastern part between km 15  860 and km 17  020, the subway tunnel is expected to be located entirely in alluvial deposits, predominantly in sandy gravels. For the western part between km 14  956 and km 15  860, the geological conditions are considerably more complex, varying between hard rocks, weathered rocks, residual soils, and alluvium, with the higher proportions in weathered rocks. The alluvial deposits probably dominated by non-cohesive sandy gravels under water saturation are expected to be favorable for the application of a slurry shield. For a more detailed assessment reliable grain size distribution curves of the alluvial deposits are necessary to deem the selection of machine types. Weathered rock and residual soil contain a higher percentage of clayey material and are expected to be more suitable for EPB shield excavation. However, through the excavation of core stones, boulders, and hard rocks machine will produce chips instead of an earth mixture. In addition, the use of plastifying additives must be considered carefully with regard to environmental protection regulations for free muck disposal. The present decision is to employ only one machine for the entire TBM tunneling, because of the consumptive construction time and the total tunnel length, only 4 km. Through these tunneling conditions, slurry shielded type TBM is highly recommended to choose a machine with liquid face support to be able to tunnel through the alluvial layers safely and with minimum

3 TREATMENTS OF EXISTING BOULDERS DURING TUNNELING The presence of cobble, boulder, and corestones can present significant difficulties during the shield tunnel operation in alluvium deposits. Design the shielded TBM should be based on the information presented in geotechnical investigation report. The TBM should have the capability to handle boulder and cobbles. Normally, modern TBM should be provided with a robust cutterhead with double bladed cutter disks, which can break boulders and remove the debris. The cutter head and the mucking system should design to operate in the presence of this type of materials (Fig. 5). During the tunnel excavation by shielded TBM, it is essential to monitor and record continuously all operation parameters, especially such parameters that may affect ground movements and settlements. As far as bigger boulders, cobbles, and corestones are concerned, they may be broken by twin disk cutters, but sometimes these boulders will be the main reason of delay in the construction schedule. A slurry type machine shall be equipped with a stone crusher optionally, which can crush boulders up to a size of 500 mm in diameter. Practically, some of the boulders are not crushed by the cutterhead because the soft soil matrix allows the boulder to move, in such a case, the TBM machine must be stopped and prepared for intervention inside the excavation chamber. If the corestones are located at the front of cutterhead in a

76 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-09.qxd

2/18/04

11:51 AM

Page 77

to prevent water inflow. The working miner can enter into the chamber through the airlock to work at the face front of the cutterhead. The boulder can be broken up with hydraulic breakers or drilled with hand held drilling machines, and then crushed by blasting or expanding mortar. After the crushing the boulders, the TBM can resume the normal boring operation. 4 GEOPHYSICAL PROBE SYSTEM FOR THE BETTER TUNNEL OPERATION No matter how they investigate the ground conditions at the design stage, it is hard to avoid the unforeseen geological conditions during construction work because of variations in ground conditions, the variation of ground materials will be much easier to predict with a forward looking geophysical technique. A possible method is seismic prospecting in soft ground that is currently used in EPB and slurry shield machine like TSP by Amberg consultant Switzerland. The Sonic soft ground probing system by Herrenknecht uses acoustic reflection to record contrasts in physical properties within soil formations. A special coded acoustic signal travels into the ground after being emitted by a sonic transmitter, installed on one of the cutting wheel arms. These methods allow detecting and localizing irregular bodies and obstacles within a range of about 50 m ahead of the tunnel face in maximum. However, accuracy and classification results of these surveys are still far to reach for the practical purposes at the tunneling sites. For the improve of this kind of blind shielded TBM, boulder detection technology shall be developed for the better machine operation to reduce the risk of adaptation of Shield tunneling technology.

Figure 5. Boulders should be handled properly during tunnel excavation. Table 2. Optimal technical specification of the slurry type shielded TBM Shield type

Slurry/Hydroshield

Design static pressure Cutter head type

3 bar Option 1: cutter head with spokes and rim Option 2: closed cutter head with minimum passages of 200 mm Drag picks and twin roller disk cutters Left and right 3500–4000 Nm 10000 hours 10000–12000 kN 5000 kN 2000 mm

Tools Direction of rotation Torque Main bearing lifetime Maximum thrust force Emergency thrust force Push ram stroke (for 1500 mm segments) Slurry pump capacity Tail shield seals Tail shield seal grease lines Man lock type Man lock operating pressure Tail shield articulation Tail shield grout injection lines Tail shield grout injection sensor Max. advance rate Possible ring assembly time

1200 m3/h 3 rows of brushes plus steel plates outside 8 No., 4 No. per chamber

5 RESULTS This subway tunnel project shows an example for a railway tunnel construction in soft alluvial deposits with higher ground water levels in a congested urban area. Although the prevailing ground conditions do not indicate corestones and boulders along this planned alignment, designer shall choose the tunnel excavation method by considering the poor ground conditions to increase the safety during the construction period. It is proven that the ground water could be controlled by slurry type shielded TBM. In this project area, most of the overburden depth ranges from 10 m to 20 m with low earth pressures. By adopting the proper construction sequence, geophysical probe system is proposed for detecting the boulders ahead of the tunnel working face and to prepare the suitable boulder handling to keep the given construction time schedule. Finally, this slurry type shielded TBM is suggested as an optimal tunnel excavation method in

Double chamber 3 bar Yes 4 No. with cleaning facilities 4 No. 60 mm/min 20 min

shallow depth, it is possible to crush them by the direct drilling work from the surface, and so on. To gain access to the excavation chamber, it will first be emptied, keeping the required pressure by means of compressed air in order to support the tunnel face and

77 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-09.qxd

2/18/04

11:51 AM

Page 78

picks, Ph.D Thesis, The University of New South Wales, Sydney, Australia, May, 1992. Maidle, B., Herrenknecht. M. & Anheuser. L., Mechanised Shield Tunnelling, 1995. Belling, W. & Eisenbach, R., Schwierigkeiten und Stillstande beim Shildvortrieb mit Flussigkeitgesestutzter Ortbrust und Uberwindung der Storfaktoren Durch Einsatz Eines Neuen, Veranderten Schildes, Forschung und Praxis 33, 1990. Fong, M.L., Bednarz, S.L, Boyce, G.M. & Irwin, G.L. History and explortation redefine Portland’s West Side CSO Tunnel alignment. North American Tunneling 2002, Seattle Washington USA.

such a soft, poor alluvial deposits, including the possibility of boulders with water saturated conditions. This will also help overcome the noise and vibration problems at nearby buildings with low settlement by the strong slurry pressures in downtown area of Seoul. REFERENCES Muirwood, A.M. The circular tunnel in elastic ground. Geotechnique 25, No.1, pp 115–127, 1975. Jee, W.R. An assessment of the cutting ability and dust generation of Polycrystalline diamond compact insert

78 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-10.qxd

2/18/04

11:51 AM

Page 79

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Estimating ground loss from EPB tunneling in alluvial soils for ECIS project, Los Angeles Theo Robert Seeley City of Los Angeles, Department of Public Works, Bureau of Engineering, Los Angeles, CA, USA

ABSTRACT: Urban environments require a higher degree of ground loss control than other areas. To accomplish this the City of Los Angeles required pressurized face support for tunneling in alluvial soils above the water table on the East Central Interceptor Sewer tunnel. To monitor the ground loss the City’s Geotechnical Engineering Division along with the Contract Administration and Soils Laboratory gathered data from the tunneling operation to estimate the ground loss. This paper will present the various methods used to gather data as the operation progressed, and methods used to analyze the data to determine if the contractor was in compliance with the specifications. Specifically to determine expected bulk volumes of muck for the various soil types encountered. This was then used to determine where ground treatment would be attempted. Ground treatment after tunneling was performed at various recommended locations. The success of the ground treatment provided feedback on how accurate the methods of analyses were. This case study will present the various methods of gathering the data, the methods of analyzing the data, and the accuracy of the analyses made from the data.

mode. The specifications allowed only 19 mm of settlement at the ground surface. The tunnel alignment was then broken into four units. The ECIS alignment with its four units is presented below in Figure 1.

1 INTRODUCTION Historical background

The East Central Interceptor Sewer (ECIS) consists of approximately 18.46 km (11 miles) of new sewer to add additional capacity to the City of Los Angeles existing North Outfall Sewer (NOS). The NOS was constructed in the 1920s and is now flowing at full capacity. Although the NOS is strictly a sanitary sewer at times of heavy rains it has overflowed by raising maintenance hole covers in the low lying sections of the city. Various routes and construction methods were considered for nearly a decade prior to construction of ECIS. The primary concerns were to build a new trunk line across the Los Angeles basin that would gravity flow with the least amount of disruption to the citizens. As late as 1997 a cut-and-cover option was considered. This was dismissed as too disruptive to the citizens. Past tunnel projects including the new metro rail system have also been disruptive due to excessive settlement. The primary goal of the ECIS tunnel would have to be the control of settlement. The City of Los Angeles enlisted the help of a Technical Advisory Panel to develop a plan to minimize the ground settlement. Their recommendation to accomplish this was to require that the tunneling be done using new tunnel boring machines (TBM) that would be operated in earth pressure balance (EPB)

1.2

The west to east alignment of ECIS crosses the Los Angeles Basin from the tunnel’s lowest point on the

Hollywood

From San Fernando Valley

101

5

it

Un

1

Unit 2

Unit 3

Downtown L.A.

t4

Blair Hills

Uni

110

NORS

Culver City LAX

Baldwin Hills

International Airport

To Hyperion Treatment Plant

0

Figure 1. Project location.

79 Copyright © 2004 Taylor & Francis Group plc, London, UK

Geology and alignment

2

4

6

Los Angeles River

1.1

8

710

10 km

09096-10.qxd

2/18/04

11:52 AM

Page 80

Figure 2. Geologic cross-section.

To the east of the hills, the alignment traverses the nearly flat topography of the Ballona Gap and the flood plain of the Los Angeles River established prior to 1825. Uplift along the Newport-Inglewood Fault zone resulted in ponding and the formation of swamps and buried peat deposits in the low lands just east of the Baldwin Hills. Interbedded peat and silty clay deposits have been observed to a depth of 12 m in the area (City of Los Angeles, 1995).

west side of the Baldwin Hills to its highest point on the east bank of the Los Angeles River. Unit 1 passes under the Baldwin Hills and then turns north into a low-lying flood plain. Unit 2 crosses the low-lying flood plain running upstream along the former historic Los Angeles River alignment from west to east. Unit 3 continues east under a slightly steeper portion of the flood plain and finally turns north for the last 540 m. Unit 4 begins north and quickly turns east again for about 3 km and then turns north along the west side of the Los Angeles River channel. Finally the last 400 m turns east passing under the concrete lined section of the Los Angeles River to its terminus at the Mission and Jesse shaft. The surface feature of the Los Angeles Basin covers an area 75 by 30 km wide. It is primarily a lowland coastal plain that slopes gradually southward and westward toward the Pacific Ocean. The plain is interrupted by a series of small hills created by the uplift of the Newport-Inglewood fault. The fault trends northsouth across the west side of the coastal plane. The uplift of the Newport-Inglewood Fault created the Baldwin Hills on the west side of the fault. This fault zone is also responsible for creating the structural traps for the oil fields found along the NewportInglewood Fault zone (Yerkes et al., 1965).

1.3

1.4

Unit 2

Unit 2 tunneling encountered a wider variety of geologic conditions that varied from buried peat deposits to very dense sand and gravel deposits. The middle 1.6 km

80 Copyright © 2004 Taylor & Francis Group plc, London, UK

Unit 1

Unit 1 geologic conditions encountered by the tunnel machine were favorable in terms of controlling ground loss. For all but the northern most 300 m, the tunneling encountered the very stiff to hard clays and silts of the San Pedro Formation that form the Baldwin Hills. The 300 m north of the San Pedro Formation consisted of very dense silty sand and very hard clayey silt.

09096-10.qxd

2/18/04

11:52 AM

Page 81

Unit 3 East, it was refurbished and reinserted into the Grand Avenue Shaft to mine Unit 3 West.

of this 4.8 km reach was in a mixed face condition. The upper half of the face was in younger alluvium that included the peat deposits, while the lower half was in the very dense granular deposits of the Lakewood Formation. The shallowest ground cover condition was approximately 5 m and coincided with the mixed face condition. Also, the groundwater table (GWT) fluctuated in this area over time. The GWT at Maintenance Hole 7, near the middle of Unit 2, varied from 6.5 m to 8.5 m below the ground surface. The tunnel at that location is 7 m below the ground surface. 1.5

1.8

At the request of the contractor, the TBM design was modified to move the auger screw up to allow for a larger main bearing on the cutter head. The contractor also had the TBMs modified to accept a muck ring and pressure relief gates. The Geotechnical Baseline Report (GBR) discussed a similar modification to an EPB machine used for the North Outfall Replacement Sewer. The final conclusion of the GBR on that issue was that the pressure relief gate does not maintain pressure in the chamber and therefore there is no earth pressure support. The first TBM started east out of the Unit 3 Grand Avenue Shaft Site with the muck ring modification. After about 100 m of mining the contractor chose to remove the muck ring primarily due to ground loss problems. Muck rings were not used again on the ECIS project.

Unit 3

Unit 3 geology was considered to be relatively favorable with the tunnel alignment completely in the dense granular soil deposits of the Lakewood Formation and above the water table. The cover ranged from 10 m at the west end to 25 m at the east terminus. The ground surface sloped uniformly up stream at a rate of 0.26% steeper than the sewer. 1.6

Unit 4

Unit 4 the tunnel alignment was also completely in the dense granular soils of the Lakewood Formation. Near its eastern terminus it is below the groundwater table which is locally higher in the vicinity of the Los Angeles River. Near the middle of Unit 4 the ground cover decreases to 10 m of Lakewood Formation overlain by 7 m of younger alluvium and fill soils. 1.7

Modifications to TBM

2 CONSTRUCTION MONITORING 2.1

Construction overview

The contract was awarded January 5, 2001, to Kenny, Shea, Traylor, Frontier-Kemper J.V. The construction started February 14, 2001. First tunneling occurred out of the Grand Avenue shaft site on December 14, 2001 with the last hole through on September 26, 2003 at the west end of Unit 4. The project was mined with four identical Lovat TBMs with a cut diameter of 4.714 m. The TBMs met the City’s requirement for new site customized Slurry or EPB-TBMs capable of working above and below the GWT. The tunnel segments have an inside diameter of 4.150 m and are 200 mm thick. The straight segments were 1.524 m long and curved segments for a 150 m radius were 1.375 m long. Tunneling began out of the east side of the Grand Avenue Shaft Site excavating Unit 3 East. The second TBM started out of the Siphon Shaft at the west end of Unit 2 mining eastward. The third TBM started mining Unit 4 from the east end proceeding west and south toward Unit 3. The fourth TBM started mining Unit 1 heading west and south from the outlet side of the Siphon Shaft. After the first TBM completed

2.2

Analyzing data

The data obtained by these five sources was then analyzed and sent to the Construction Management (CM) Team. When the analyzed data indicated that excessive ground loss was occurring, the CM Team notified the contractor so they could modify their mining operation, as they deemed necessary.

81 Copyright © 2004 Taylor & Francis Group plc, London, UK

Sources of information

The primary goal of the ECIS project was to build a new sewer with the minimum amount of disruption to the public. Therefore, construction monitoring was a high priority. This was done in five different ways. First, the ground surface along the alignment was monitored continuously by full time survey crews. Second, the contract required that the contractor install 268 Multiple Point Borehole Extensometers MPBX for the City to monitor settlement at depth. Third, the City’s inspection division monitored the mining operation continuously and estimated the muck volume for each shove. Fourth, the City’s soil laboratory sent technicians to obtain samples of the muck as it was being mined by the four TBMs. And fifth, the contract required that the TBMs be equipped with automatic recording system for data collection. The data was transmitted to the surface for both the City’s and the contractor’s use.

09096-10.qxd

2/18/04

11:52 AM

Page 82

tunnel elevation. However, this did not always give reasonable values. The typical spacing of boring information was on the order of one boring every 50 m to 100 m. This is because driven samples in very dense granular soil tend to disturb the sample and give a lower dry density. Also, in the space between borings the mined soil would typically change consistency several times. However, the boring logs did show that soil in the Lakewood Formation was consistently very dense. To resolve this problem GED requested the compaction tests to determine the laboratory maximum dry density. Then, based upon the fact that the mined soil in Units 3, 4, and the east end of Unit 2 are very dense, a relatively high percentage of the laboratory Maximum Dry Density was found to give reasonable results. The percentages used were 92% initially and later we found that 95% relative compaction gave more reasonable results.

3 SOIL SAMPLING 3.1

Goals of soil sampling

The goals of soil sampling were to provide: 1. Samples of the condition muck to test the quality of conditioning. 2. Samples for determine the wet and dry density of the conditioned muck. 3. Samples for estimating the in-place density of the mined soil. 4. Database of mined soils for future comparison and create a permanent record of the materials mined, since EPB mining precludes any regular mapping or sampling from the face. 3.2

Sampling methods used

The primary test used to determine the muck conditioning was the slump cone test. This utilized the same cone test procedure used for concrete testing. For determining the wet density, samples of the muck were poured into 150 mm diameter 300 mm high cylinders with a minimum amount of compaction effort. To simulate the compactive effort of muck falling off the end of the conveyor belt into the muck car, the technician filled the cylinder in three lifts. The side of the cylinder was tapped lightly after each lift. Excess soil from the final lift was then carefully trimmed and the cylinders were then capped. Full capped cylinders were then weighed in a field lab. Given the known volume of the cylinder, the weight was then converted to determined wet density. A representative sample was then taken from the cylinder to determine the moisture content and calculate the dry density of the muck in the tunnel prior to transporting to the shaft. This was important because the inspectors estimated, to the nearest cubic meter, the volume of mined muck at the heading. When the muck was saturated granular soil, it was observed that the muck would consolidate during transportation. Therefore, the technicians would only add a small amount of compactive effort to simulate muck falling from the conveyor belt to the muck car. The muck for a typical shove would fill three to four muck cars. Muck cars were designed to hold 10 m3 when filled to the brim. Samples were taken from the second or third car. This was to avoid sampling muck that was on the conveyor belt or in the plenum when the machine was not mining. 3.3

3.4

During the initial stages of mining, the City of Los Angeles, Division of Standards, under the direction of GED, started taking samples of the muck as it was being loaded into the muck cars from the conveyor belt. This sampling operation was moved to the surface as the tunnel length increased. The majority of the sampling was done at the surface as the muck was being removed from the shafts. Near the end of mining Unit 4, which was the longest drive, GED requested sampling the muck at the heading and then take a second sample from the same muck car as it was being unloaded at the ground surface. The Unit 4 shaft site is located at Station 18460. The results of this study are presented on Figures 3 and 4 where moisture content and bulking factor are compared. The bulking factor is a ratio of the in-place density to the muck density. Therefore, two tests are required for each data point, and six of the seven sets of points are within 1.5%. An attempt was also made to compare the slump of the muck at the heading and at the shaft; however, all samples in this study exhibited zero slump. On one occasion earlier in the project the slump was taken in the tunnel of Unit 2 at Station 5092.0 and again at the surface. The elapsed time between samples was one half-hour. The slump at the heading was 12 mm and at the shaft 3 mm. The one sample at Station 13340 shows that a higher moisture content and lower bulking factor was measured at the heading of the tunnel. This sample was classified as GP-GM. Its counter part that was sampled at the shaft from the same muck car was classified as SM silty sand. This variability in soil type appears to be the reason for the difference in test results. The other sample pairs did not have this variability.

Determing in place density

Near the beginning of the project, boring log data was used for the dry density of the in-place soil at the

82 Copyright © 2004 Taylor & Francis Group plc, London, UK

Checking soil sampling procedures

09096-10.qxd

2/18/04

11:52 AM

Page 83

a ground loss was to manually drill up from the crown of the tunnel approximately 3 m and fill voids above the tunnel. This method had a short window of time. Within a matter of hours the mining operation would build enough rings to pass the location where the miners could stand on top of the trailing gear and hand drill up. Also, the grout was pumped under pressure using the pump that was near the front of the trailing gear. Therefore, a quick qualitative test was needed to decide if the muck counts indicated lost ground or just bulked up soil. Observations early on were that the clean SP sands had a higher probability of creating chimney-type cavities. They were also found to have a lower bulking factor. The calculation for the bulking factor required drying the soil back to optimum moisture and running a laboratory maximum dry density test. This typically took a half a day or more. Therefore, it was decided to experiment with the Sand Equivalent test to help determine if there was a problem or not. The California Test 217, Method of Test for Sand Equivalent was developed as a field test primarily for cut and cover pipe-laying operations. Its primary use in construction is to determine if proposed pipebedding material is acceptable. What makes pipebedding material acceptable is its ability to flow into the cavity under a pipe laid in a trench when a small amount of water is added. Ironically, this ability to flow into open spaces is what made the SP sands so difficult to mine without loosing ground. When the screw conveyor was taking soil out of the plenum faster than the advance of the TBM, then the clean SP sands would drain into the TBM like sand in an hourglass. The sands of the Lakewood formation were typically very dense, and dry with virtually no cementation. The vibration of the TBM was enough to loosen the dry sand and cause it to flow. The only thing preventing ground loss was the EPB pressure. Our observation was that if the SE was over 30% and the EPB pressure was low or fluctuating, then ground loss was likely to occur. A comparison of the bulking factors and SE has been attempted to find a correlation. The results were somewhat erratic, but in general, a SE over 30% indicated the soil would have a low bulking factor and was more likely to run if not supported. Soils with a SE below 30% have a longer stand up time and higher bulking factor.

Figure 3. Moisture content comparison.

Figure 4. Bulking factor comparison.

4 LABORATORY TESTS 4.1

Development of tests

Initially, the laboratory testing consisted of the moisture content (ASTM D2216-98) and density of the muck along with grain-size analysis. After about 100 m of mining, the laboratory maximum dry density (ASTM D1557-91) was added. Conventional concrete slump tests (ASTM C143/C143M-00) as described in EFNARC (2001) were also run on the muck as it was being excavated. These were performed both at the heading and at the shaft site. Atterberg limit tests were run on the cohesive soils. The primary problem with most of these tests was the time it takes to run them. By the end of the project we had added visual classification of dry, pasty, or wet.

5 ANALYSES FROM TEST RESULTS 5.1

4.2

Sand equivalent tests

The use of the sand equivalent test came out of a need for fast turnaround on the laboratory tests. The problem was that the time to react to a lost ground condition is very short. The primary method used to correct

83 Copyright © 2004 Taylor & Francis Group plc, London, UK

Using sand equivalent tests

The primary use of the SE test was to help the inspector in the field decide if ground loss was occurring or not. For example, if the muck volume per shove stayed the same but the SE increased from 20 to 45, then it was very likely that the bulking factor had decreased.

09096-10.qxd

2/18/04

11:52 AM

Page 84

5.3

Therefore, the only reason for the muck volume to stay the same is the TBM was taking in more ground than it was before. At that point, the inspector would require more ground loss check grouting and ask the operator to modify their mining operation. When the laboratory had completed the tests necessary for calculating the bulking factor, that data was relied upon instead of the SE test. 5.2

The sampling and laboratory test results from the mixed face condition encountered in Unit 2 were not satisfactory. The primary problem is that mixing two soils of different gradation creates a new soil that does not resemble the two parent soils and the laboratory analyses are useless. 6 OVER EXCAVATION CHARTS

Bulking factor calculations

The over excavation charts were developed to take into account all of the pertinent lab and field data to determine the amount of soil mined per shove. As the laboratory data started coming in and a method of estimating the bulking factor was developed, it became clear that analyses could be made to calculate the ground loss. This was done on a spreadsheet and then graphed on a bar chart. The spreadsheet takes into account the inspector’s estimated volume of muck, in-place density, density of the muck, theoretical grout volume and actual annulus grout volume. Using these factors against the theoretical cut volume it was possible to develop the bar charts that indicated the amount of ground loss per shove. The results of these bar charts had to be moderated against the fact that there was one sample for every 9 to12 rings in the Lakewood Formation. There was some variability in the muck density. This typically showed up as abrupt changes in the bar charts. Using a moving average solved this. The bar charts presented on the next page used a moving average of three muck density values. The in-place density was taken from averaging laboratory maximum dry densities over a range where the mined soil strata was similar. Examples of over excavation charts are included as Figures 6 and 7. These were chosen to demonstrate that the volume and location of cavities could be estimated. Figure 6 shows a drop in the over excavation volume between 11706 and 11715. This area was pressure grouted from the spring line of the tunnel up to several meters above the tunnel for a Maintenance Hole. The Maintenance Hole was eliminated from the project after the grouting was completed. Between Station 11790 and 11800 there was an attempt to fill cavities simply by drilling a 200 mm diameter hole and filling it with grout. This drill and fill repair method was not satisfactory to the City. However, GED observed that 11 of the 14 borings drilled at that location either hit grout from a previous boring drilled and filled the year before, or found cavities. The previous boring took 9 m3 of slurry to fill. This drill and fill method was not successful because the borings caved about half way down to the tunnel crown. The total amount of grout pumped in the 14 holes was 19 m3. So the total fill to date is 28 m3 and more compaction grout will be required. From the chart the last

The bulking factor analysis that was developed by GED for the ECIS project and first reported by Crow and Holzhauser (2003) is simply a ratio of the dry density of the muck as it is being loaded into the muck cars from the conveyor belt to its in-place dry density prior to being mined. Figure 5 shows how this was derived. The analysis is simple; the main problem, as discussed above, is obtaining the correct data to put in the equation. The dry density obtained from typical soil sampling methods is lower than the in-place density. This is due to sampling disturbance. Then, there is the problem of simulating the compactive effort of muck falling from the conveyor belt into the muck car. The laboratory results from this project indicated that the cohesive soils are the difficult soils to estimate the compactive effort for muck falling into a muck car. However, the cohesive soils have a long standup time and are usually not a problem for ground loss. The granular soils, that are a problem for ground loss, gave more consistent results. This along with the fact that the bulking factor for granular soils is typically only 10% to 20%. This smaller range in bulking factor means that the analysis will usually be close to the true bulking factor. Analysis of the expected excavated soil volume by use of the Bulking factor BEPB In situ

Excavated Air

Air Water

VA,e

VA,i VW,i ,GW,i

Water

VW,e, GW,e

Solids

VS,e, GS,e

Ve

Vi Solids

VS,i ,GS,i

Note: V : Volume [m3] Index i: in situ G : Weight [kN] Index e: excavated γd : Dry unit weight [kN/m3]

Bulking Factor due to EPB-tunneling: Dry unit weight of soil in situ: Dry unit weight of excavated soil: WithVG,i = VG,e use (2) and (3) in (1): Note:

Excavated soil volume includes bulking due to boring process and addition of conditioning agents

BEPB =

Ve - Vi Vi

(1)

γd,i

=

GS,i Vi

(2)

=

GS,e Ve

(3)

γd,e

BEPB

Limitations on calculating the bulking factor

γd,i - 1 [-] = γ d,e

(4)

92% of the max lab dry unit weight was taken as the dry unit weight of soil in s

Figure 5. Bulking factor analysis.

84 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-10.qxd

2/18/04

11:52 AM

Page 85

7 MONTHLY TUNNEL SUMMARY CHARTS About one third of the way through the tunneling a monthly set of charts was developed and delivered to the construction management team. These charts were produced in 100-m sets. Figure 8 as shown on the previous page, is an example of these charts. The purpose of these Tunnel Summary graphs is to summarize the quality of the tunneling. They were summaries of the extensometer data, ground survey data, inspector’s non-compliance and their related job orders, and the estimated ground loss from the over-excavation charts. The charts were color coded to indicate where there was no major ground loss

FACE STATION

Figure 7. Over excavation chart 118 to 119.

Figure 8. Monthly tunnel summary charts.

85

118+98.93

118+92.81

118+86.69

118+80.57

118+74.45

118+68.33

118+62.21

118+56.09

118+49.97

118+44.46

118+37.73

118+31.61

118+25.49

118+19.37

118+13.25

118+07.13

FACE STATION

Figure 6. Over excavation chart 117 to 118.

Copyright © 2004 Taylor & Francis Group plc, London, UK

ECIS UNIT 3E STATION 118+00 TO 119+00 20.0 17.0 14.0 11.0 8.0 5.0 2.0 -1.0

118+01.01

OVER EXCAVATION [m3]

117+91.83

117+85.90

117+79.78

117+73.66

117+67.54

117+61.42

117+55.30

117+49.15

117+43.03

117+36.91

117+30.79

117+24.67

117+18.55

117+12.43

117+06.31

ECIS UNIT 3E STATION 117+00 TO 118+00 17.0 14.0 11.0 8.0 5.0 2.0 -1.0

117+00.19

OVER EXCAVATION [m3]

7 shoves took in about 38 m3. In this same area a reference hole was drilled 6 m north of the tunnel centerline. It did not cave or encounter cavities. The reference hole only took the theoretical volume of grout to fill. Compaction grouting was performed in 6 holes drilled along the centerline between Stations 11856 to 11863.5. The total for all 6 grout holes was 12 m3. However, prior to compaction grouting exploratory borings were drilled at 11859 and 11861 and they each took 5 m3 of grout. So a total of 22 m3 were pumped or poured into an area five shoves long for an average of 44 m3 per shove. In this area the ground loss was calculated to be about 10 m3 per shove.

09096-10.qxd

2/18/04

11:52 AM

Page 86

However, in Units 3 and 4 the dense granular soil of the Lakewood Formation proved to be more difficult to mine. 5. Preparing Tunnel Summary charts on a monthly basis provided a good reference to keep track of the tunneling, and the repair of the few areas where ground loss occurred. 6. Sampling the muck at the shaft site gave reasonable results.

problems, apparent problems, problems that were severe enough to result in a non-compliance and areas where the contractor had completely filled the cavities created by the ground loss. These were colored green, orange, red, and yellow, respectively. The charts also showed the amount of cover relative to the tunnel size. Graphics were also developed to show where various types of grout were used to fill cavities and there estimated elevations and volumes. Below the tunnel cross-section the survey data and the MPBX results were plotted. Where the surface settlement exceeded the allowable 19 mm or the bottom anchor of the MPBX exceeded 32 mm, they were plotted in red. The primary problem with the MPBX and survey data is the spacing between points. This is where the over-excavation charts were used to define the limits of the problem areas. These areas have since become known as the Red Zones. The City of Los Angeles is presently negotiating with the contractor to have the cavities filled. It should be noted that the majority of these charts were colored green meaning that the majority of the tunneling was satisfactory.

ACKNOWLEDGMENTS The Author wishes thank the other members of the Geotechnical Engineering Division who helped develop the charts and graphs used on the ECIS Project, and the members of the Division of Standards Soil Laboratory that obtained and tested the soil. Without this data the development of the methods to estimate ground loss would not have been possible.

REFERENCES 8 CONCLUSIONS

Crow, M.R. & Holzhauser, J. 2003. Performance of Four EPB-TBMs Above and Below the Groundwater Table On the ECIS Project Los Angeles, CA, USA. Proceedings of Rapid Excavation and Tunneling Conference, Society for Mining, Metallurgy, and Exploration, Littleton, Colorado, USA p. 905–931 EFNARC. 2001. Specification and Guidelines for the specialist products for Soft Ground Tunneling. ENFARC, Aldershot, UK Yerkes, R.F., McCulloh, T.H., Schoelhamer, J.E. & Vedder, J.G. 1965. Geology of the Los Angeles Basin, California – An Introduction, U.S. Geological Survey Professional Paper 420-A

1. The volume of ground loss and its limits can be estimated. 2. The bulking factor calculations give reasonable results except in split face condition. 3. Sand Equivalent test is useful in tunneling as a quick test to check on changes in the material that is being mined. 4. During the design of ECIS, Unit 2 was believed to be the area where ground loss would be a problem due to the shallow cover and mixed face condition.

86 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-11.qxd

2/18/04

11:53 AM

Page 87

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Some aspects of grouting technology for Manhattan tunnels M. Ryzhevskiy STV Incorporated, New York, USA

P. Barraclough Parsons Brinckerhoff Quade & Douglas, Inc., New York, USA

ABSTRACT: Conventional methods of tunneling in hard but jointed and heavily sheared rock formations, particularly those containing groundwater, may require additional pre-excavation rock mass reinforcement or improvement methods. These methods can include pre- and post-excavation grouting for both purposes of rock stabilization and controlling groundwater inflow into the excavation. Published geological data for the Manhattan area indicates that rock mass quality is generally good to very good, although it is locally complex, containing heavily faulted and sheared zones of high permeability, which are expected to be unstable. Driving hard rock TBM’s through these zones can pose significant problems and pre- and post-grouting methods should be considered in order to minimize the risk of rock mass failure and water inundation. The methodology and procedures for pre- and post-grouting during the excavation of TBM driven tunnels have been developed and are described within this article.

phase defines the regional structure of Manhattan with the axial plain striking N35°E and generally dipping toward the south-southwest at approximately 10° to 15°. Published information identifies the presence of four major joint sets within the Manhattan Schist. However further to recent and extensive ground investigation and interpretation the structure appears to be more complex than historical data suggests, revealing significant localized faulting, shearing, alteration and folding. The existing shear zones have been identified and categorized as major and minor. Major shear zones occur on a scale of approximately 10 ft to 100 ft (3 m to 30 m) and are characterized by the original rock being sheared, brecciated, and rehealed in a mylonite matrix. The fractures are often coated with secondary minerals. The boundary between the brecciated and the undamaged rock mass is distinctive, with this zone including clusters of open infilled joints and secondary mineralization. Minor shear zones occur on a onefoot (0.3 m) scale with their associated clusters of infilled, stained, mineralized joints and slickensides being apparent on a 10-foot (3 m) scale. Manhattan Island is bound by the East River to the east and the Hudson River to the west and is slightly above sea level. Groundwater levels measured in borehole standpipes range from 15 feet to 60 feet (5 m to 20 m) below street level. The quality of the

1 INTRODUCTION Manhattan Island is the heart of New York City and is one of the most urbanized cities, housing some of the most expensive real estate in the US and the world. Manhattan imposes serious construction considerations owing to its high density of buildings including historic residential districts and hi-rise commercial and residential properties, often with deep basements. In addition to the buildings is a highly developed infrastructure with many existing tunnels and other underground structures. In Manhattan it is therefore very important that during construction, systems are provided to mitigate risk and impact caused by tunneling. 2 GEOLOGY & HYDROGEOLOGY Manhattan is underlain by Proterozoic and Ordovician metamorphic rocks, locally known as the Manhattan Prong. These metamorphic rocks are characterized by three lithologies comprising schist, gneiss and marble, although the greater part of Manhattan is dominated by the more erosion resistant schist and gneiss. The rocks of the Manhattan Prong have been subjected to multiple tectonic episodes including folding, faulting and intrusion, resulting in an intensely folded and locally sheared rock mass. The prominent fold

87 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-11.qxd

2/18/04

11:53 AM

Page 88

of ground conditions. The principle of this method is to replace the water or air contained in the ground (pores, voids, cracks, joints) with a liquid material, which sets to a solid during short interval of time. There is an abundance of choices for the liquid but most commonly used are cement mixes, sodium silicate or organic resins. In tunneling there are two major purposes for grouting: to create a barrier against groundwater flow and to increase shear strength of the rock mass in order to maintain stability of the excavation, reduce settlement and ground movement. Shallow tunnels are often grouted from the surface, so that grouting and excavation procedures can go on simultaneously. Grout patterns would normally be rows of boreholes forming squares, with the spacing between boreholes determined by local conditions. When conditions preclude working from the surface, and when tunnels are very deep, grouting can be done from the tunnel face. The thickness of grouted area surrounding the tunnel varies and depends on the ground conditions and the purpose of the grouting (ground strength and/or permeability).

rock mass controls the permeability of the Manhattan Schist and generally the permeability of the undisturbed and unweathered rock is very low. In faulted and sheared zones the permeability is considerably higher as the network of fractures behaves as conduits for the groundwater. The permeability of the discontinuities is influenced by several factors including roughness, tightness and presence of joint infill. The coefficient of permeability has been derived from in-situ packer tests and typically varies between 105 to 107 cm/sec. However in shear zones the permeability ranged between 104 to 106 cm/sec and it is anticipated that widely spaced, open, steeply dipping fractures may transmit groundwater at greater rates than indicated by packer tests. This is especially true during excavation of new underground openings where flowing groundwater entering the excavation can wash out the clay infill from slickensided joints and cataclasite within shear zones, which can result in the rock mass permeability being in the order of 102 cm/sec. 3 CONSTRUCTION CONSIDERATIONS Construction experience in New York City and the findings of geotechnical programs indicate that the rock mass is generally good quality, stable and generates moderately low amounts of water. Only shear zones are expected to be unstable and to be sources of significant groundwater flow. Typically excavation of any underground structure in similar conditions with drill and blast or TBM methods will require initial support, such as rock bolts/dowels, shotcrete and occasionally steel ribs or lattice girders. The initial rock support systems are designed to prevent failure of blocks and loosened rock mass from the crown and sidewalls of the tunnels. These support systems in the hard rock formations usually take the form of fully grouted dowels and resin anchored rock bolts in combination with steel reinforced shotcrete. Often fiber-reinforced shotcrete is used as opposed to steel wire mesh. Tunneling in unstable rock formations where stand-up time of the excavation is limited due to densely distributed discontinuities, shear zones, and water saturated zones, will require additional special pre-excavation reinforcement/improvement methods to increase the rock mass quality, thus avoiding rock instability and controlling groundwater into the tunnel. In cases where the rock mass conditions dictate the need for pre-excavation rock stabilization, pre-excavation grouting techniques can be utilized.

5 PRE-EXCAVATION GROUTING Specific procedures for pre-excavation grouting from inside the tunnel space can be developed for any tunnel configuration, shape, diameter and excavation method. The philosophy of the pre-excavation grouting is to the increase rock mass properties or to seal a limited area ahead of the face and around the tunnel, using a grout of suitable strength, low permeability to water and of high durability. In general the pre-excavation grouting method involves filling all (or as much as possible) fissures, cracks and voids for a distance of a least 50–80 feet (10–25 m) ahead the tunnel face and 180° above the springline or 360° all around the tunnel. After finishing one round of pre-excavation grouting, 70–80% of the grouted length can be excavated. Subsequent to excavation of the improved rock mass, probing of the rock ahead of the TBM face will ascertain whether additional pre-excavation grouting is necessary. This cycle may be repeated as long as required, depending on geological conditions. To achieve a “dry” tunnel, postexcavation grouting may be required later in addition. Access to the tunnel face with the drill and injection equipment is very important for effective preexcavation grouting. For tunnels excavated with drill and blast method, access to the tunnel face is not often a problem. TBM’s are rarely designed to facilitate injection drilling close to the face, due to the lack of access. However, modern TBM’s are capable of performing pre-excavation grouting over or through their shield and sometimes through the cutterhead, although the latter is a very sophisticated method.

4 GROUTING METHODS Grouting is an established and common technique in modern tunneling. Grouting can be used in most types

88 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-11.qxd

2/18/04

11:53 AM

Page 89

Figure 1. Fault zone prediction ahead of the TBM.

Figure 2. 1st round of pre-excavation grouting.

When pre-excavation grouting is to be adopted for a project, it is important to identify the area and/or volume of ground that needs to be treated and where it is with respect to the tunnel face (Fig. 1). For this purpose probe holes must initially be drilled ahead of the cutterhead to establish the nature of the rock mass and groundwater conditions (location, flow rates and pressure). Tunneling in intensely fractured and crushed rocks, where groundwater inflow is minimal will only require an improvement of the rock mass properties for maintaining stability. This is accomplished by pre-excavation grouting over the TBM shield up to 180° above springline. The angle at which the rig can drill the holes into the tunnel wall will govern how far ahead of the TBM the grout can be injected. The usual set-up for this drill equipment has to allow for drilling holes at a minimum of 5–10° inclined to the tunnel axis. If the holes are drilled at angles greater than this, then the drill holes will be too far away from the tunnel perimeter, greatly reducing the efficiency of the grouting. If the inflow of the groundwater from the probe holes exceeds a predetermined threshold, it will be required to drill grout holes 360° all around the TBM to achieve control of the groundwater. Thus the anticipated flow of groundwater to be encountered plays a major role in the selection of equipment and its configuration. In the cases where the amount of the groundwater can be tolerated during the construction period and where the rock mass will only require increasing its stability to prevent the rock falls behind the shield of the TBM, pre-excavation grouting would be specified above of the TBM cutterhead only (Figs 2 & 3). In these circumstances the following action should be executed: probing ahead of the TBM, pre-excavation grouting of the potentially unstable area ahead and above of the TBM shield and post-excavation grouting of the invert section to control groundwater inflow in to the tunnel.

Figure 3. Subsequent round of pre-excavation grouting.

The procedure will start with drilling a probe hole near the anticipated unstable rock formation or fault zone ahead of the tunnel face, which should be drilled up to 100 feet (30 m) in length at the 12 o’clock position. If this probe hole indicates any potential problem, such as the presence of weak rock (higher drilling rate), loss of flush water or high groundwater inflow through the drilled probe hole, then two more probe holes should also be drilled at the 9 and 3 o’clock positions to verify the findings of the first probe hole. If the additional probe holes indicate poorer rock mass quality ahead, then an appraisal will be required to judge the extent of pre-excavation grouting. Different grouting trigger levels should be adopted for specific projects. The following pre-excavation grouting trigger levels may be used for assessment whilst probing:

•

89 Copyright © 2004 Taylor & Francis Group plc, London, UK

Sudden loss of drilling water; over 50% (an abrupt change in the amount of water returning to the surface or face usually signifies that the drill has reached a highly permeable horizon)

09096-11.qxd

•

•

2/18/04

11:53 AM

Page 90

Sustained (over half an hour) groundwater inflow through any probe hole drilled ahead of the tunnel face, over the 25 gallons/minute or fi gallons/ minute/feet (100 liters/minute or 3 liters/minute/ meter) Total groundwater inflow from all drilled holes during an half an hour period exceeded 50 gallons/ minute (200 liters/minute).

If pre-excavation grouting is necessary, based on trigger levels, the grout holes should be drilled above the tunnel springline as shown on the Fig. 2, Section A. The grout holes should be drilled at least 50 feet (15 m) ahead of the TBM cutterhead through the TBM shield at an inclination between 5°–10° to the tunnel axis. The distance between the grout holes is to be determined according to local rock properties and its grout penetration value. Previously gained experience in similar rock conditions suggests that the maximum distance between the grout holes of 5 feet (1.5 m) should be used. Based on former experience in Manhattan rocks, cementitious grouts are recommended as primary injection materials. Cement can enter fissures as small as 0.3 mm and therefore has a comparable penetration/infiltration capacity as silicate or acrylic resins, but makes a much more durable, more economical and more environmentally friendly solution. The main components of cementitious grout are water and generally Type I or II Portland cement and can be altered by using other cement types, such as Type III (high early strength), Type IV (low heat of reaction), or Type V (resistance to chemical attack). By varying the water to cement ratio it is simple to change the grout’s bleeding rate, subsequently altering its plasticity and ultimate strength. Mixing the main components with additives, such as bentonite, sodium silicate, dispersants, retarders, and accelerators, will also change the grouts properties. Microfine cement is an alternative version of cement that overcomes the difficulty of using Portland cement grouts in low permeability ground.

Figure 4. Post-excavation grouting.

face, once the TBM has passed (Fig. 4). Post excavation grouting alone is usually wasteful as it is costly and often unsuccessful, but in combination with pre-excavation grouting it is very effective. Using cementitious pre- and post-excavation grouting improves the water tightness of the rock mass, achieving lower water flow rates into tunnels typically between 0.1 to 2.5 gallons (0.5 to 10 liters) per minute per 300 feet (100 m) of the tunnel. In cases where the seepage rates into tunnels are required to be more stringent, governed by the sensitivity of the environment and construction, chemical grouting (organic resin) can be utilized as an alternative to cement. 7 CONCLUSION Tunneling under Manhattan generally does not pose significant stability problems as the majority of the rock mass is good quality, therefore the methods detailed within this article are not always required. Published information identifies the existence of several locally complex and significant fault and shear zones in which the rock mass can be expected to be unstable and to yield high rates of groundwater. When such ground conditions are anticipated these techniques described herein can be considered. Among existing pre- and post-excavation rock mass improvement/reinforcement methods available, grouting is very effective in combination with TBM driven tunnels.

6 POST-EXCAVATION GROUTING Most tunneling projects cannot tolerate large volumes of groundwater during construction of the final structure, therefore groundwater ingress is required to be limited, which can be done by using post-excavation grouting in addition to pre-excavation grouting. Clearly, pre-excavation grouting all around the tunnel is more effective than post-excavation grouting alone, but it will significantly slow TBM advance rates. For this reason, when ground conditions permit, it is favored to perform pre-excavation grouting of the rock ahead and above the springline and then post-excavation grouting of the invert area, from behind the tunnel

REFERENCES Merguerian, C. 2002. Brittle Faults of the Queens Tunnel Complex, NYC Water Tunnel #3. In G.N. Hanson, Ninth Annual Conference on Geology of Long Island and

90 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-11.qxd

2/18/04

11:53 AM

Page 91

metropolitan New York. 20 April 2002, Sate University of New York at Stony Brook, NY, Long Island Geologists Program with Abstracts, 116 p. Snee, C., Sarkar, S., Benslimane, A., Stewart, C., and Osborne, C. 2003. Rock Mass Characterization for the Manhattan East Side Access Project. In P. Cullgan et al (ed.), Soil Rock America 2003 (The 12th Panamerican Conference for Soil Mechanics & Geotechnical

Engineering and the 39th US Rock Mechanics Symp.), June 22–25, 2003, Vol. 1, pp. 129–136. Ryzhevskiy, M. 1987. The main principles of the new technologies for construction tunnels in unstable rock formations. Energetic Construction. Moscow, N7. Ryzhevskiy, M. 1988. The advance experience of the chemical ground improvement by jet grouting. VPTI Transstroj. Moscow.

91 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-11.qxd

2/18/04

11:53 AM

Page 87

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Some aspects of grouting technology for Manhattan tunnels M. Ryzhevskiy STV Incorporated, New York, USA

P. Barraclough Parsons Brinckerhoff Quade & Douglas, Inc., New York, USA

ABSTRACT: Conventional methods of tunneling in hard but jointed and heavily sheared rock formations, particularly those containing groundwater, may require additional pre-excavation rock mass reinforcement or improvement methods. These methods can include pre- and post-excavation grouting for both purposes of rock stabilization and controlling groundwater inflow into the excavation. Published geological data for the Manhattan area indicates that rock mass quality is generally good to very good, although it is locally complex, containing heavily faulted and sheared zones of high permeability, which are expected to be unstable. Driving hard rock TBM’s through these zones can pose significant problems and pre- and post-grouting methods should be considered in order to minimize the risk of rock mass failure and water inundation. The methodology and procedures for pre- and post-grouting during the excavation of TBM driven tunnels have been developed and are described within this article.

phase defines the regional structure of Manhattan with the axial plain striking N35°E and generally dipping toward the south-southwest at approximately 10° to 15°. Published information identifies the presence of four major joint sets within the Manhattan Schist. However further to recent and extensive ground investigation and interpretation the structure appears to be more complex than historical data suggests, revealing significant localized faulting, shearing, alteration and folding. The existing shear zones have been identified and categorized as major and minor. Major shear zones occur on a scale of approximately 10 ft to 100 ft (3 m to 30 m) and are characterized by the original rock being sheared, brecciated, and rehealed in a mylonite matrix. The fractures are often coated with secondary minerals. The boundary between the brecciated and the undamaged rock mass is distinctive, with this zone including clusters of open infilled joints and secondary mineralization. Minor shear zones occur on a onefoot (0.3 m) scale with their associated clusters of infilled, stained, mineralized joints and slickensides being apparent on a 10-foot (3 m) scale. Manhattan Island is bound by the East River to the east and the Hudson River to the west and is slightly above sea level. Groundwater levels measured in borehole standpipes range from 15 feet to 60 feet (5 m to 20 m) below street level. The quality of the

1 INTRODUCTION Manhattan Island is the heart of New York City and is one of the most urbanized cities, housing some of the most expensive real estate in the US and the world. Manhattan imposes serious construction considerations owing to its high density of buildings including historic residential districts and hi-rise commercial and residential properties, often with deep basements. In addition to the buildings is a highly developed infrastructure with many existing tunnels and other underground structures. In Manhattan it is therefore very important that during construction, systems are provided to mitigate risk and impact caused by tunneling. 2 GEOLOGY & HYDROGEOLOGY Manhattan is underlain by Proterozoic and Ordovician metamorphic rocks, locally known as the Manhattan Prong. These metamorphic rocks are characterized by three lithologies comprising schist, gneiss and marble, although the greater part of Manhattan is dominated by the more erosion resistant schist and gneiss. The rocks of the Manhattan Prong have been subjected to multiple tectonic episodes including folding, faulting and intrusion, resulting in an intensely folded and locally sheared rock mass. The prominent fold

87 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-11.qxd

2/18/04

11:53 AM

Page 88

of ground conditions. The principle of this method is to replace the water or air contained in the ground (pores, voids, cracks, joints) with a liquid material, which sets to a solid during short interval of time. There is an abundance of choices for the liquid but most commonly used are cement mixes, sodium silicate or organic resins. In tunneling there are two major purposes for grouting: to create a barrier against groundwater flow and to increase shear strength of the rock mass in order to maintain stability of the excavation, reduce settlement and ground movement. Shallow tunnels are often grouted from the surface, so that grouting and excavation procedures can go on simultaneously. Grout patterns would normally be rows of boreholes forming squares, with the spacing between boreholes determined by local conditions. When conditions preclude working from the surface, and when tunnels are very deep, grouting can be done from the tunnel face. The thickness of grouted area surrounding the tunnel varies and depends on the ground conditions and the purpose of the grouting (ground strength and/or permeability).

rock mass controls the permeability of the Manhattan Schist and generally the permeability of the undisturbed and unweathered rock is very low. In faulted and sheared zones the permeability is considerably higher as the network of fractures behaves as conduits for the groundwater. The permeability of the discontinuities is influenced by several factors including roughness, tightness and presence of joint infill. The coefficient of permeability has been derived from in-situ packer tests and typically varies between 105 to 107 cm/sec. However in shear zones the permeability ranged between 104 to 106 cm/sec and it is anticipated that widely spaced, open, steeply dipping fractures may transmit groundwater at greater rates than indicated by packer tests. This is especially true during excavation of new underground openings where flowing groundwater entering the excavation can wash out the clay infill from slickensided joints and cataclasite within shear zones, which can result in the rock mass permeability being in the order of 102 cm/sec. 3 CONSTRUCTION CONSIDERATIONS Construction experience in New York City and the findings of geotechnical programs indicate that the rock mass is generally good quality, stable and generates moderately low amounts of water. Only shear zones are expected to be unstable and to be sources of significant groundwater flow. Typically excavation of any underground structure in similar conditions with drill and blast or TBM methods will require initial support, such as rock bolts/dowels, shotcrete and occasionally steel ribs or lattice girders. The initial rock support systems are designed to prevent failure of blocks and loosened rock mass from the crown and sidewalls of the tunnels. These support systems in the hard rock formations usually take the form of fully grouted dowels and resin anchored rock bolts in combination with steel reinforced shotcrete. Often fiber-reinforced shotcrete is used as opposed to steel wire mesh. Tunneling in unstable rock formations where stand-up time of the excavation is limited due to densely distributed discontinuities, shear zones, and water saturated zones, will require additional special pre-excavation reinforcement/improvement methods to increase the rock mass quality, thus avoiding rock instability and controlling groundwater into the tunnel. In cases where the rock mass conditions dictate the need for pre-excavation rock stabilization, pre-excavation grouting techniques can be utilized.

5 PRE-EXCAVATION GROUTING Specific procedures for pre-excavation grouting from inside the tunnel space can be developed for any tunnel configuration, shape, diameter and excavation method. The philosophy of the pre-excavation grouting is to the increase rock mass properties or to seal a limited area ahead of the face and around the tunnel, using a grout of suitable strength, low permeability to water and of high durability. In general the pre-excavation grouting method involves filling all (or as much as possible) fissures, cracks and voids for a distance of a least 50–80 feet (10–25 m) ahead the tunnel face and 180° above the springline or 360° all around the tunnel. After finishing one round of pre-excavation grouting, 70–80% of the grouted length can be excavated. Subsequent to excavation of the improved rock mass, probing of the rock ahead of the TBM face will ascertain whether additional pre-excavation grouting is necessary. This cycle may be repeated as long as required, depending on geological conditions. To achieve a “dry” tunnel, postexcavation grouting may be required later in addition. Access to the tunnel face with the drill and injection equipment is very important for effective preexcavation grouting. For tunnels excavated with drill and blast method, access to the tunnel face is not often a problem. TBM’s are rarely designed to facilitate injection drilling close to the face, due to the lack of access. However, modern TBM’s are capable of performing pre-excavation grouting over or through their shield and sometimes through the cutterhead, although the latter is a very sophisticated method.

4 GROUTING METHODS Grouting is an established and common technique in modern tunneling. Grouting can be used in most types

88 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-11.qxd

2/18/04

11:53 AM

Page 89

Figure 1. Fault zone prediction ahead of the TBM.

Figure 2. 1st round of pre-excavation grouting.

When pre-excavation grouting is to be adopted for a project, it is important to identify the area and/or volume of ground that needs to be treated and where it is with respect to the tunnel face (Fig. 1). For this purpose probe holes must initially be drilled ahead of the cutterhead to establish the nature of the rock mass and groundwater conditions (location, flow rates and pressure). Tunneling in intensely fractured and crushed rocks, where groundwater inflow is minimal will only require an improvement of the rock mass properties for maintaining stability. This is accomplished by pre-excavation grouting over the TBM shield up to 180° above springline. The angle at which the rig can drill the holes into the tunnel wall will govern how far ahead of the TBM the grout can be injected. The usual set-up for this drill equipment has to allow for drilling holes at a minimum of 5–10° inclined to the tunnel axis. If the holes are drilled at angles greater than this, then the drill holes will be too far away from the tunnel perimeter, greatly reducing the efficiency of the grouting. If the inflow of the groundwater from the probe holes exceeds a predetermined threshold, it will be required to drill grout holes 360° all around the TBM to achieve control of the groundwater. Thus the anticipated flow of groundwater to be encountered plays a major role in the selection of equipment and its configuration. In the cases where the amount of the groundwater can be tolerated during the construction period and where the rock mass will only require increasing its stability to prevent the rock falls behind the shield of the TBM, pre-excavation grouting would be specified above of the TBM cutterhead only (Figs 2 & 3). In these circumstances the following action should be executed: probing ahead of the TBM, pre-excavation grouting of the potentially unstable area ahead and above of the TBM shield and post-excavation grouting of the invert section to control groundwater inflow in to the tunnel.

Figure 3. Subsequent round of pre-excavation grouting.

The procedure will start with drilling a probe hole near the anticipated unstable rock formation or fault zone ahead of the tunnel face, which should be drilled up to 100 feet (30 m) in length at the 12 o’clock position. If this probe hole indicates any potential problem, such as the presence of weak rock (higher drilling rate), loss of flush water or high groundwater inflow through the drilled probe hole, then two more probe holes should also be drilled at the 9 and 3 o’clock positions to verify the findings of the first probe hole. If the additional probe holes indicate poorer rock mass quality ahead, then an appraisal will be required to judge the extent of pre-excavation grouting. Different grouting trigger levels should be adopted for specific projects. The following pre-excavation grouting trigger levels may be used for assessment whilst probing:

•

89 Copyright © 2004 Taylor & Francis Group plc, London, UK

Sudden loss of drilling water; over 50% (an abrupt change in the amount of water returning to the surface or face usually signifies that the drill has reached a highly permeable horizon)

09096-11.qxd

•

•

2/18/04

11:53 AM

Page 90

Sustained (over half an hour) groundwater inflow through any probe hole drilled ahead of the tunnel face, over the 25 gallons/minute or fi gallons/ minute/feet (100 liters/minute or 3 liters/minute/ meter) Total groundwater inflow from all drilled holes during an half an hour period exceeded 50 gallons/ minute (200 liters/minute).

If pre-excavation grouting is necessary, based on trigger levels, the grout holes should be drilled above the tunnel springline as shown on the Fig. 2, Section A. The grout holes should be drilled at least 50 feet (15 m) ahead of the TBM cutterhead through the TBM shield at an inclination between 5°–10° to the tunnel axis. The distance between the grout holes is to be determined according to local rock properties and its grout penetration value. Previously gained experience in similar rock conditions suggests that the maximum distance between the grout holes of 5 feet (1.5 m) should be used. Based on former experience in Manhattan rocks, cementitious grouts are recommended as primary injection materials. Cement can enter fissures as small as 0.3 mm and therefore has a comparable penetration/infiltration capacity as silicate or acrylic resins, but makes a much more durable, more economical and more environmentally friendly solution. The main components of cementitious grout are water and generally Type I or II Portland cement and can be altered by using other cement types, such as Type III (high early strength), Type IV (low heat of reaction), or Type V (resistance to chemical attack). By varying the water to cement ratio it is simple to change the grout’s bleeding rate, subsequently altering its plasticity and ultimate strength. Mixing the main components with additives, such as bentonite, sodium silicate, dispersants, retarders, and accelerators, will also change the grouts properties. Microfine cement is an alternative version of cement that overcomes the difficulty of using Portland cement grouts in low permeability ground.

Figure 4. Post-excavation grouting.

face, once the TBM has passed (Fig. 4). Post excavation grouting alone is usually wasteful as it is costly and often unsuccessful, but in combination with pre-excavation grouting it is very effective. Using cementitious pre- and post-excavation grouting improves the water tightness of the rock mass, achieving lower water flow rates into tunnels typically between 0.1 to 2.5 gallons (0.5 to 10 liters) per minute per 300 feet (100 m) of the tunnel. In cases where the seepage rates into tunnels are required to be more stringent, governed by the sensitivity of the environment and construction, chemical grouting (organic resin) can be utilized as an alternative to cement. 7 CONCLUSION Tunneling under Manhattan generally does not pose significant stability problems as the majority of the rock mass is good quality, therefore the methods detailed within this article are not always required. Published information identifies the existence of several locally complex and significant fault and shear zones in which the rock mass can be expected to be unstable and to yield high rates of groundwater. When such ground conditions are anticipated these techniques described herein can be considered. Among existing pre- and post-excavation rock mass improvement/reinforcement methods available, grouting is very effective in combination with TBM driven tunnels.

6 POST-EXCAVATION GROUTING Most tunneling projects cannot tolerate large volumes of groundwater during construction of the final structure, therefore groundwater ingress is required to be limited, which can be done by using post-excavation grouting in addition to pre-excavation grouting. Clearly, pre-excavation grouting all around the tunnel is more effective than post-excavation grouting alone, but it will significantly slow TBM advance rates. For this reason, when ground conditions permit, it is favored to perform pre-excavation grouting of the rock ahead and above the springline and then post-excavation grouting of the invert area, from behind the tunnel

REFERENCES Merguerian, C. 2002. Brittle Faults of the Queens Tunnel Complex, NYC Water Tunnel #3. In G.N. Hanson, Ninth Annual Conference on Geology of Long Island and

90 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-11.qxd

2/18/04

11:53 AM

Page 91

metropolitan New York. 20 April 2002, Sate University of New York at Stony Brook, NY, Long Island Geologists Program with Abstracts, 116 p. Snee, C., Sarkar, S., Benslimane, A., Stewart, C., and Osborne, C. 2003. Rock Mass Characterization for the Manhattan East Side Access Project. In P. Cullgan et al (ed.), Soil Rock America 2003 (The 12th Panamerican Conference for Soil Mechanics & Geotechnical

Engineering and the 39th US Rock Mechanics Symp.), June 22–25, 2003, Vol. 1, pp. 129–136. Ryzhevskiy, M. 1987. The main principles of the new technologies for construction tunnels in unstable rock formations. Energetic Construction. Moscow, N7. Ryzhevskiy, M. 1988. The advance experience of the chemical ground improvement by jet grouting. VPTI Transstroj. Moscow.

91 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-12.qxd

2/18/04

11:54 AM

Page 93

Session 1, Track 4 Specialized urban construction

Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-12.qxd

2/18/04

11:54 AM

Page 95

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Design and construction of an LRT tunnel in San Jose, CA P.J. Doig Hatch Mott MacDonald, Pleasanton, California

ABSTRACT: The Santa Clara Valley Transportation Authority (VTA) is constructing significant extensions to its Light Rail Transit (LRT) system in the metropolitan San Jose area. These include the Vasona line that will extend from downtown San Jose to the suburb of Campbell. At busy Diridon station, trains will travel for 480 m in a subterranean section, comprising tightly-curved tunnel and straight approach ramps. The tunnel passes under streets, a bus terminal, rail tracks and platforms, and is very close to the historic station building. The project also includes an extension to the existing pedestrian tunnel, that will connect the station to a new LRT station. The tunnels were built by cut-and-cover methods. The contract for the tunnel project commenced in February, 2001 and was completed in June, 2003. This paper covers the execution of the project, from the latter stages of design to completion of the LRT tunnel structure.

the light rail system. This may yet have an effect on the scale and operation of the Vasona line. A second funding initiative, passed in 2000, extended the sales tax for a further 30 years beyond 2006. This was intended to fund a number of projects, but primarily the extension of BART from Fremont to San Jose. VTA has been investigating ways of utilizing this projected revenue to support its current expenditures, so as to limit cutbacks in service. However, should the economy not rebound, VTA will be faced with difficult choices for its future that may well have a substantial impact on the scope of transportation services offered in the San Jose area.

1 INTRODUCTION San Jose is situated at the southern tip of the San Francisco Bay in northern California. San Jose is the major city in Santa Clara county which has a population in excess of 1.7 million. The San Jose area is more loosely known as Silicon Valley, being the home of hitech icons such as Apple, Intel and Cisco Systems. Per capita income and house prices in the area are amongst the highest in the US. Many people commute from distant communities and traffic congestion is chronic. Public transportation has a high level of public support, which is reflected in numerous funding initiatives. The Vasona LRT extension arose out of an initiative approved by voters in 1996. Known as the Measure B Transportation Improvement Program, the initiative authorized a half-cent sales tax in Santa Clara county, expiring in 2006. It was anticipated that the tax would generate around $1.6 billion in revenue that would be used to fund a specific package of countywide transportation improvement projects. Management of the sales tax is by the Santa Clara County Board of Supervisors. The Santa Clara Valley Transportation Authority (VTA) is the implementing agency for the 1996 Measure B Transportation Improvement Program. VTA is an independent special district responsible primarily for bus and light rail transit (LRT) service. The recent downturn in the economy, particularly the in hi-tech sector, has had a significant effect on sales tax revenue. VTA has been obliged to scale back its operations and curtail plans for future development of

2 VASONA LINE OVERVIEW The LRT system is currently around 50 km in length, with two lines arranged in a rough T-shaped configuration. New lines under construction will add a further 20 km. Phase I of the new Vasona line will be 8 km long, running from downtown San Jose to Campbell in the southwest. The line will have nine stations and is expected to carry 8000 to 9000 riders per day. The LRT system is largely at grade with tracks located in the central divide of city streets. Trains are powered from an overhead contact system. The initial segment of the Vasona line from the existing Guadalupe line to just east of the San Jose Diridon main line station will be over city streets. At the station itself, the line will be in tunnel. Between San Jose Diridon Station and

95 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-12.qxd

2/18/04

11:54 AM

Page 96

Vasona Junction, the line will follow the former Union Pacific Railroad (UPRR) right-of-way, now owned by VTA. The capital cost of Phase 1 of the Vasona line is budgeted at $321 million, of which around $53 million is expected to come from Federal funding. Construction of the Vasona line commenced in early 2001 and is anticipated to be complete in late 2004. Revenue operation is expected to commence in Spring, 2005. 3 SAN JOSE DIRIDON TUNNELS SEGMENT The San Jose Diridon Tunnels project, designated C345, involved the construction of some 490 m of trackway below grade. This section of the line comprises 120 m of approach ramp at the western end, and 95 m of approach ramp at the eastern end. In between, there is 275 m of tightly-curved, double box tunnel. The tunnel is necessary to negotiate the railyard and other facilities at the historic Diridon Station. The station is owned by the Peninsular Corridor Joint Powers Board (JPB) which also operates the Caltrain commuter service. The station serves three local commuter services, Amtrak long-haul passenger service, and UPRR freight trains. The station is also the maintenance facility for much of the rolling stock. The station yard has twelve tracks, five of which share three platforms. One of those tracks is also used by freight traffic. The remaining tracks are for maintenance and storage. The tunnel project involved an intricate series of staging operations in the railyard to allow tracks to be taken out of service so that construction could proceed. Groups of tracks were isolated and removed for a number of months at a time. This allowed a trench to be excavated from surface to accommodate the LRT tunnel. A second, smaller trench was excavated for the extension of the existing pedestrian tunnel. In three locations, the track was supported on a temporary bridge and the trench excavated beneath it.

Figure 1. Construction within the rail yard.

to concerns from JPB about potential effects on the station. The main station building, and other elements including the platform canopies, are designated historic structures and must be preserved. The design therefore proceeded on the basis of a cut and cover tunnel, skirting the station building. Discussions with JPB led to the development of a staging plan for removal and/or bridging of the tracks during construction. Originally, it was thought that most of the tracks would need to be kept in service and therefore have to be placed on bridges. However, JPB reviewed their operational requirements and decided they could function with four through tracks. Discontinuous tracks were to be kept open at the south end of the yard so that they could be used for storage. Figure 1 illustrates the proximity of the west channel construction to the rail yard. VTA’s designer analyzed the storage requirement and this was factored into the staging plan. The plan allowed for the work to be completed in three distinct stages. This was to be accomplished with the installation of three shooflies, one permanent crossover, and a temporary bridge under the freight track. Signaling work associated with the track modifications was not included in the final design. JPB reserved the signaling design to themselves and this was only made available after the construction contract had been let. The contract design also included the temporary shoring system for the open-cut excavation. The choice of a temporary shoring system is normally left to the contractor. However, JPB’s requirement that they approve the system mitigated against leaving this item open in the bid documents. It was known that JPB had a preference for soil mixing due to its limited impact on the operation of adjacent rail lines, and success with the method elsewhere on the system. Also, JPB does not allow tie-backs on their property because of a perception that the system might adversely affect subsoil conditions. Driven piles were specifically prohibited

4 DESIGN The designer for the C345 project was the General Design Consultant (GDC). This comprises Parsons Brinkerhoff Quade and Douglas, in association with MK Centennial and Korve Engineering. Design of the tunnel structures under subcontract to the GDC, was undertaken by San Francisco-based, Biggs, Cardosa Associates, with shoring design by EQE of Oakland. Detailed design was completed in September, 2000. As part of the review process, the design submissions were routinely reviewed by the rail companies. In the early stages, consideration was given to traversing the railyard in conventional, bored tunnel. However, this was rejected due mainly to the cost, and

96 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-12.qxd

2/18/04

11:54 AM

Page 97

because of noise restrictions. Therefore, soil-mix walls with internal bracing became the specified support system. The system was to comprise of a low-strength, reinforced soil-mix wall installed within the ground on either side of the trench prior to excavation commencing. The wall was made up of contiguous bored columns, with every second column reinforced with a steel beam. The generic soil profile for the LRT tunnel indicated fine-grained sediments, coarse-grained sand, and fine gravel to a depth of 17 m. Stiff clay extended below that depth. The groundwater table was at a depth of 4 m, although the design assumed it was 2 m higher. In order to allow for construction in the dry, the soil-mix wall was designed to terminate in the clay. Before excavation could proceed below the water table, the walls had to be closed on all sides. This meant that temporary bulkheads had to be installed across the alignment in some places to give a closed cell. The soil-mix columns were designed to be 75 cm in diameter, positioned at 60 cm centers. This gave a nominal 60 cm thick wall. The reinforcing beams were mostly W610 82 section, and were required to extend around 7 m below the bottom of the excavation. Ground surface was at approximately 27 m, and base of excavation at 20 m. Maximum design pressure on the walls was 115 kPa. The design generally included for two levels of struts, one at surface and one at mid-depth. Top level walings were specified as W460 113, and lower level were W460 177. Struts were 46 cm diameter pipes with 0.65 cm wall thickness. Top level struts had to be preloaded to 300 kN and lower level struts to 600 kN. Given the sensitive nature of the station terminal building and other structures in the immediate locale, it was decided to mandate a settlement monitoring program. This required installation of inclinometers outside the shoring wall, and establishment of elevation points on the adjacent buildings. Groundwater monitoring wells were also to be installed. These all had to be checked on a regular basis, with a requirement that action to be taken in the event of excessive movement.

Figure 2. Drilling soil-mix columns.

manager, South Bay Transit Associates (SBTA). SBTA is a joint venture of Hatch Mott MacDonald and URS, formerly O’Brien Kreitzburg. Vali Cooper & Associates and Booz Allen also provided staff to the project as subconsultants to SBTA. 6 CONSTRUCTION 6.1

5 PROCUREMENT The contract documents were issued for bid in September, 2000. Bids were opened in November, 2000. Three bids were submitted, ranging from $23 to $26 million. The low bid was submitted by Condon-Johnson & Associates (CJA) of Oakland, California. The contract was awarded in February, 2001 with overall completion required in February, 2003. Construction management of the project was by VTA utilizing its own staff and staff drawn from the program

97 Copyright © 2004 Taylor & Francis Group plc, London, UK

Soil mixing

Preliminary work on the project was sufficiently complete to allow soil mixing to commence in April, 2001. CJA used the Geo-Jet® process developed by Verne L. Schellhorne of Aerial Industrial, Inc. CJA is the exclusive licensee for use of Geo-Jet® in the states of California, Oregon, Washington, Utah and Nevada. Figure 2 shows the process in operation. The Geo-Jet® process is similar to other soil-mixing techniques, with some unique features. The process generally involves cutting a column with a single auger soil-cement processor. Cement slurry is ejected at high velocity through nozzles in the auger to create a high shear mix of soil cuttings and cement slurry. The process of forming the in-situ soil-cement columns is monitored by a computer. The specified strength of the finished soil-mix was 1.4 MPa at 28 days. A test program was instituted at the beginning of the project utilizing production columns. This confirmed a ratio of around 0.2:1, dry Portland cement to in-situ soil. For most of the C345 project, the soil-mixing rig was a Link Belt 218 crawler crane with 32 m of leads mounted to the boom. Reinforcing I-beams were inserted with an ABI pile-driver, on a Sennebogen carrier. Later in the project, an ABI rig was also used for soil-mixing. Grout was produced at a mobile plant. This was generally located within 50 m of the soilmixing rig, although it could be much further away as conditions dictated. The plant had storage for up to 160 000 kg of cement.

09096-12.qxd

2/18/04

11:54 AM

Page 98

mitigating delays caused by the factors mentioned earlier.

The soil-mix operation started with the excavation of a starter trench, around 2 m in depth. This was to capture the material that is pushed out of the top of the column by the mixing operation. A length of wall would be identified and the unreinforced intermediate columns installed first, at 120 cm centers. The following day, the reinforced columns were installed between the intermediate columns. As well as providing walls for the open-cut excavation, soil-mix columns were installed for bulkheads. These could be reinforced where they had a structural purpose, or unreinforced where they were for water control. In addition, at one location where the structural bulkhead had to be supported with diagonal bracing, soil-mix columns were installed en masse to act as a buttress. The buttress columns were designed such that a higher strength of 2.1 MPa was required for the soil-cement. Generally, soil-mixing proceeded without major incident, and progress was satisfactory. Production averaged 256 m of columns per 8 hour shift, with the best shift reaching 665 m of columns. The soil-mixing was not continuous due the project having to be done in stages. Access to some stages was delayed due to issues with third parties, notably the railroad companies, utility owners, and private property owners and businesses. Also, there were frequent encounters with obstructions, including hazardous materials, utilities, rail, and the apparent remnants of a grouting program. These various factors combined to delay the contract for around four months. The work was also delayed by difficulties the contractor experienced with alignment of the columns. In the initial stage of the construction, it was found that some sections of soil-mix wall had impinged on the structure, with the occasional column toeing well into the excavation. Nowhere was this sufficient to require complete removal of the column beam. However, some beams had to be cut back and other measures had to be taken, which included some major rework. The contractor addressed this problem in later stages by moving the wall out another 7.5 cm from the structure. This proved to be successful in achieving the required clearances. 6.2

6.3

Once the soil-mix walls had been installed in a particular stage, the excavation was commenced. Sufficient depth of trench was excavated to accommodate the top level of bracing. The top end of the channels and the pedestrian tunnel had only one level of bracing. The deeper areas had two levels. Excavation was generally carried out with a dozer in the cut, pushing to an excavator located on the surface. Once the soilmix walls had been exposed, they were trimmed to the face of the beams. Thereafter, a 7.5 cm-thick layer of shotcrete was applied. The shotcrete was intended primarily to provide a smooth backing for the waterproofing membrane that would be installed later. However, in one location, it had a structural purpose, serving as a key against uplift of the base slab. After the shotcrete had been applied, bracing was installed. This comprised longitudinal walings spanning a number of beams, normally around four or five, on both sides of the excavation. Two transverse struts were then placed between the walings. The walings were H-beams and the struts large pipe sections. All steelwork was pre-fabricated off-site. The struts were then preloaded to a given load, packed and welded in place. Excavation was then continued down to the second bracing level and the procedure repeated. Thereafter, excavation was completed to sub-grade. Upon completion of the excavation, a 10 cm-thick mudslab was placed on the subgrade. A waterproofing membrane was then installed over the top of the slab and up the walls. A second protective mudslab was then poured over the waterproofing so that invert slab construction could begin. The contract documents specified Preprufe® 300 for blind-side application, and Bituthene® 3000 for exposed-side application. Both materials are of sandwich construction and include a layer of HDPE. Joints have to be taped and penetrations repaired. Due to what he perceived as difficulty installing and maintaining the product intact, the contractor offered Paraseal® as an alternative. Paraseal® comprises a butyl membrane with bentonite prills impregnated on the interior surface. The sheets are nailed to the wall, and require only lap joints. The material is intended to be self-healing. At joints and holes, any water entering from outside is expected to encounter bentonite which expands and closes the hole or joint. VTA accepted the alternative and Paraseal® was installed. Experience with the Paraseal was not entirely satisfactory. Whether this related to deficiencies in the system, problems with installation, or construction effects was not entirely resolved. Wherever the material was

Track staging

The general intent of the staging plan contained in the contract documents was generally followed. However, there were significant modifications to the detail. The contractor elected to install two additional temporary bridges, which reduced the requirement for shooflies and improved access. VTA, with the concurrence of JPB, made stage 3 available prior to the completion of stages 1 and 2. Other constraints on work procedures and sequencing were relaxed. These measures had the effect of opening a second front and

98 Copyright © 2004 Taylor & Francis Group plc, London, UK

Tunnel excavation

09096-12.qxd

2/18/04

11:54 AM

Page 99

Figure 3. Interior of tunnel.

Figure 4. West channel with pedestrian tunnel portal at right.

exposed for any period, it was susceptible to degradation by rain or drainage water. This was particularly evident at joints between stages. Also there were problems with properly terminating the material at grade, on the exterior of the channels. The result was that the finished structure leaked. A chemical grout injection program was conducted during the spring of 2003 and this appeared to substantially reduce infiltration. However, this has been during the dry season and it remains to be seen how effective the system is when the rains commence. The contractor has to warrant the material for ten years and so VTA may direct more remedial measures if leakage is a concern. Figure 3 illustrates the generally dry condition of much of the tunnel as at October, 2003.

along with conduits for other systems such as communications and CCTV. Additionally, there was a dry fire line provided with external connections to hydrants and internal stand-pipes. All conduits and pipes were cast into the concrete so that there was no potential for service interruption in the event of a fire or derailment in the tunnel. The tunnel construction contract did not include installation of the LRT track. This was done under a follow-on contract. The tunnel construction included forming shallow troughs in the invert, complete with threaded rebar inserts. The troughs were used later as the foundation for low plinths poured by the track contractor. Rails were affixed directly to the plinths with purpose-built hardware. The tunnel was constructed with vertical and horizontal curves, including a spiral section. This made the track installation process more difficult and some rework was necessary. Track installation was completed in November, 2003.

6.4

Tunnel construction

The concrete construction was relatively routine. Pour lengths were limited to 18 m. Mechanical rebar couplers were used at construction joints between stages. The exterior walls were formed with a series of full height panels, mounted on wheels. They were moved by hand along the tunnel invert, as the shoring made it difficult to pick and land the panels with a crane. The walls in the channels incorporated architectural features and these required particular care when matching form panels. The roof slab was formed with a proprietary shoring system involving interlocking towers. Concrete was delivered into the forms from a pump on surface. Invert and roof slabs, and cantilever walls in the channels, were required to reach specified strength before shoring could be removed. All exposed surfaces received a Class I finish, and the channel walls were treated with anti-graffiti paint. The tunnel included combined system ductbanks at the base of each exterior wall which doubled as emergency walkways. Lighting was provided in the tunnel,

7 CONCLUSION The San Jose Diridon Tunnels project was completed in June, 2003. The project had to overcome numerous challenges presented by underground obstructions and third party issues. The time for completion was extended by four months as a result. However, the original date for access to the tunnel for the track contractor was met, ensuring that the overall project remained on schedule. Despite these issues, the final cost of the contract was only 8% over the original price, and wellwithin VTA’s budget. As Figure 4 shows, only installation of the overhead contact system remains for the tunnel section to be ready for testing. This will begin early in 2004. It is anticipated that test trains will run on the Vasona line in the Fall of 2004, with revenue service scheduled to begin in April, 2005.

99 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-13.qxd

2/18/04

11:54 AM

Page 101

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Underpinning design and construction – Atlantic Avenue Station complex rehabilitation, New York, USA A. Grigoryan & L.G. Silano Parsons Brinckerhoff Quade & Douglas, Inc., New York, New York, USA

ABSTRACT: The Atlantic Avenue Station Complex in Brooklyn comprises three New York City Transit (NYCT) subway stations and the Long Island Rail Road (LIRR) Flatbush Avenue commuter rail terminal. The Atlantic Avenue Station on the IRT subway line is undergoing a major structural rehabilitation. This paper addresses underpinning methods and procedures, as well as the construction of new passageway sections and a new lower-level concourse under operating tracks without interruptions to transit or station operations and minimal impacts to street traffic.

1 INTRODUCTION The Atlantic Avenue Station on the Eastern Parkway Interborough Rapid Transit (IRT) Line in Brooklyn

opened in 1908 and is located under Flatbush Avenue near the intersection of Atlantic Avenue (Figure 1). In addition to providing Eastern Parkway Line NYCT service (Nos. 2, 3, 4 & 5 trains), the station provides

Figure 1. Project location plan.

101 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-13.qxd

2/18/04

11:54 AM

Page 102

Figure 2. Existing structure plan.

connections to the Pacific Street Station (B, M, N & R trains) on the Fourth Avenue Brooklyn–Manhattan Transit (BMT) Line, the Atlantic Avenue Station (D & Q trains), and the LIRR Flatbush Avenue Terminal. The station services over 65,000 passengers daily. A connecting passageway runs under the tracks of this station, skewed across and near the center of the IRT platforms. This passageway, the vital link of this complex with its multiple connections to other subway and commuter rail lines, is being modified and replaced by three new station elements: A new west passageway section, a reconfigured east passageway section, and a new lower-level concourse connecting east and west passageway sections. Meanwhile, the station structure had deteriorated to a state of disrepair as a result of intensive use and irregular or neglected maintenance. Water infiltration had damaged many structural components and the facility was not adequate to handle present and projected future increased passenger flow demands. Coordinating major project tasks involves working closely with the capital construction and operational departments of two major transit agencies (NYCT

and LIRR), with the city’s Department of Transportation for street traffic issues, with several other city agencies for utility relocations, and with numerous subconsultants. 2 EXISTING STRUCTURE The existing subway station utilizes typical NYCT framing, with columns spaced at 4.6 meters (Figure 2). Most of the existing steel framing members are built-up sections, comprised of web plates, angles, channels and cover plates, connected by 22-millimeterdiameter rivets. The majority of existing columns are supported on individual spread footings, measuring 1.5 meters 2.7 meters and extending 0.6 meters below the IRT invert slab (Figure 3). Columns are located along six column lines with roof beams running along each column line, located approximately 2.1 meters below street level. The depth of roof beams within the construction limits is 0.76 meters. The existing reinforced concrete roof structure is supported by roof beams or by roof beams and exterior station walls.

102 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-13.qxd

2/18/04

11:54 AM

Page 103

Figure 3. Existing structure section.

The invert slab of the existing passageway is a nominally reinforced concrete slab on grade. The structural thickness of the invert slab is approximately 0.2 meters. Four subway tracks, generally consisting of wood half-ties and rail assemblies, are positioned directly on the 0.3-meter-thick track slab on grade. The existing columns, platforms, and tracks are supported on the roof of the existing passageway by transfer grillage beams and girders, spanning the passageway at a skew. The existing structure is sealed against water infiltration using a waterproofing membrane. The IRT Atlantic Avenue Station has three passenger platforms: Two side platforms service northbound and southbound local tracks, while a center island platform serves two express tracks. Two stairs lead from the passageway to each of the platforms. In addition, there is an old control house located adjacent to the exterior wall of the southbound local platform. This structure, which has historic value, was temporarily relocated to a place near the construction site. As part of this project, the control house was repaired, rehabilitated, and returned to its original location. A subconsultant along with a specialty contractor handled this task.

3 PROPOSED MODIFICATIONS The existing connecting passageway is reconfigured and widened in accordance with the Atlantic Avenue Master Plan Study, and is remediated by three new station elements (Figure 4):

•

A new west passageway section

• •

The connecting passageway is reconfigured to allow the station to comply with Americans with Disabilities Act (ADA) accessibility provisions. The west section of the connecting passageway is reconstructed to eliminate stairs by installing an ADAaccessible ramp as well as a new ADA elevator. This elevator provides service to the local southbound platform. Meanwhile, the east section of the connecting passageway is reconfigured to provide ADA elevator access to the northbound local platform along with access to the Atlantic Avenue Brighton Line station. A new lower-level concourse is also built to provide ADA elevator access to the express island platform. A primary function of the concourse is to provide ADA elevator access to the express platform. This new passageway/concourse configuration provides the opportunity to expand the section of passageway bounded by stairs to the southbound local and island platforms, thereby alleviating congestion and circulation problems at the existing connecting passageway and providing more stairs to those platforms. The new concourse, spanning between the local southbound platform and express island platform, has a total of five stairways: Two to the local southbound platform and three to the express island platform. In order to increase the level of service for the section of the passageway east of the new concourse, the passageway is widened near a currently overcrowded platform stair. The area is also reconfigured to install an elevator to the northbound platform.

103 Copyright © 2004 Taylor & Francis Group plc, London, UK

A reconfigured east passageway section A new lower-level concourse connecting the west and east passageway sections

09096-13.qxd

2/18/04

11:54 AM

Page 104

Figure 4. Modified structure plan – platform level.

Figure 5. Modified structure section.

4 DESIGN CONCEPT In order to accommodate extensive passageway modifications and new concourse construction, our firm developed several design concepts. Steel was selected

as the material that would provide both the required strength and flexibility for the necessary construction staging, while reinforced concrete was selected for designing the invert slab, walls, and stairs (Figure 5). Due to architectural considerations as well as passenger

104 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-13.qxd

2/18/04

11:54 AM

Page 105

Figure 6. Modified structure longitudinal section.

flow considerations, the plan arrangement of columns in the modified passageway and the new concourse did not follow the grid of existing columns on the platform level above. This presented major structural challenges – both from a design strength and construction staging perspective. Typical existing column loads range from 135 to 180 metric tons. Transfer girders were designed to support existing columns at the platform and track levels (Figure 6). Due to modifications and new construction, the lower-level columns carry axial loads ranging from 155 to 725 metric tons. Some of the columns are also subjected to bending moments due to unbalanced loading from framing girders and beams. Base plates for the columns measure in plan view up to 0.76 by 0.84 meters, with thicknesses reaching 95 millimeters. Most of the steel framing uses ASTM A36/A36M Grade 250 steel (Fy  250 Mpa), with some of the plate girders composed of ASTM A572/A572 Grade 345 steel (Fy  345 Mpa). The sidewalls of the new concourse, modified passageway, and stairs were designed conservatively as cantilever retaining structures, constructed integrally with the invert slab. One of the main reasons for the conservative design involved a constructibility issue: It would have been more expensive to provide temporary supports during all staging procedures until the walls were completed and platforms reconstructed to provide lateral support at top of walls. The invert slab, meanwhile, is designed as a mat foundation. In comparison to the existing framing, the new framing represents a major structural change. The

original columns extend only from the underside of the platform or track bed up to the roof girders. Loads are applied to these columns only on top, at roof level. The new columns, however, extend from the invert slab of the lower-level through the track and platform levels up to the roof level, as seen in Figure 5. Loads are applied to the new columns from various directions at two or three levels. At track level, floor beams are framed into new columns or into transfer girders. Transfer girders carry track and platform loads and transfer them to lower-level columns. Our firm designed special seats to accommodate the connection of existing columns to transfer girders, shown in Figure 6. These seats provide flexibility during construction and fabrication, with most information verified in the field prior to fabrication. This design solution allowed the contractor to fabricate girders and columns to specified dimensions, before gaining access to the bottom of existing columns to measure their exact elevation. Due to IRT northbound local platform modifications, the demarcation line between two stations – the Atlantic Avenue IRT Station and the Flatbush Avenue LIRR Terminal – was moved east from its original location. This presented another structural challenge as the existing platforms of these two stations had different elevations and slopes. Within the construction limits, the difference ranged from approximately 0.15 to 0.6 meters. The existing grade separation between the stations also had a full height chain link fence extending from the platform level up to the underside of the roof girders.

105 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-13.qxd

2/18/04

11:54 AM

Page 106

Figure 7. Typical cross section at Track No. 6.

No columns could be located in the passageway below the new demarcation line to support the platform framing. As a result, there were only two column lines on either side of the demarcation line available to support the platform framing at this location: Column line “A” of the IRT northbound local platform and column line “K” of the LIRR Track No. 6 platform. Transfer girders run along these column lines, as do variable length platform support beams, spanning between column lines A and K. This challenge was resolved by introducing a custom-designed kinked welded plate girder. The plate girders, spaced at 1.5-meter centers along the northbound platform, support platform glass paver panels which, in turn, limit girder flange width to 0.15 meters. Welded plate girders are supported by longitudinal platform girders on column line A at one end and by LIRR Track No. 6 transfer girders at the other end (Figure 7). 5 CONSTRUCTION METHODS The public interest, safety, and convenience were emphasized at every stage of project design and

construction. Since the clear width of the existing passageway was only 4.6 meters, staging procedures were established to maintain the existing opening at all times throughout each phase of construction. Furthermore, criteria were established to maintain access to the stairways providing access from the passageway to the three subway platforms. As there were two existing stairs at each platform, temporary and/or permanent stairs needed to be constructed before the existing stairs could be closed, removed, or modified during construction. Considering the complexity of construction, and numerous operational limiting factors, the client requested the development of two feasible construction methods as part of the design package. Our firm refers to these methods as:

• •

The major difference between the two methods is that the drift method offers minimum interference with existing train and passenger operations, as well as minimum visual impacts during construction.

106 Copyright © 2004 Taylor & Francis Group plc, London, UK

Suggested Construction Method “A” (Drift Method) Suggested Construction Method “B” (Pile Support Method)

09096-13.qxd

2/18/04

11:54 AM

Page 107

Figure 8. Construction method “A”.

Figure 9. Construction method “B”.

Occurring primarily in locations not visible to the public, the drift method makes limited use of heavy equipment, with construction performed in various stages in tight, confined spaces below existing platforms and tracks (Figure 8). The pile support method, on the other hand, offers more flexibility regarding equipment usage. However,

it has a greater visual impact on passengers, as many temporary roof support piles and beams are visible during construction. Although most construction progresses behind shielding, the general public is fully aware of the ongoing work (Figure 9). Both methods required a construction shaft to provide access for the excavation, removal of spoil, and construction of the

107 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-13.qxd

2/18/04

11:54 AM

Page 108

lower concourse. The suggested work sequence was organized in three construction phases following a general preparatory phase. This preparatory phase consisted of work necessary to prepare the existing structure for actual excavation and underpinning efforts required for construction of the lower concourse. During this phase, the conditions of structures and utilities were inspected and verified; the control house was relocated to a storage location; utilities were relocated or protected; and shaft excavation was completed. The three general construction phases for both suggested construction methods A and B included:

• • •

Phase I: The area north of the existing passageway, comprising excavation, underpinning, demolition, and construction of temporary and final support framing. Phase II: The existing passageway area and the southern portion of the lower concourse, consisting of excavation, underpinning, demolition, and construction of temporary and final support framing Phase III: All finish work within the lower concourse

Major construction phase limitations and techniques included:

• • •

• •

Maintaining the 4.6-meter width of the existing passageway beneath the IRT station for any temporary routing of passengers during construction phasing. Maintaining two stairways from the passageway up to each platform. The capacities of these replacement stairways needed to equal the capacity of existing stairways. Limiting major construction requiring contractor access to transit trackways, to one trackway and the adjacent platform. This work required general orders (GO’s) for train diversions that were only available on weeknights or weekends. Field welding to existing steel members was generally not allowed. Constructing temporary platforms and stairways, barricades, and overhead shields to protect passengers and workers during demolition, construction, and phasing.

The contractor elected to use Method B, modified somewhat to reflect his method of underpinning and procedures, as well as to respond to actual field conditions and ever-evolving operational limitations. The contractor also modified the pile types, pile layout, and track support details, as shown in Figure 6.

The underpinning of the existing structure – roof with full live load of street traffic above, fully operational platforms, and four subway tracks with uninterrupted service – was one of the most challenging and complicated issues encountered during the project. Of course, public safety was also a major issue. For several months, the entire subway structure needed to be underpinned, and major structural elements, such as columns, foundations, stairs, platforms and tracks, needed to be rebuilt. The client requested that we check and scrutinize every detail of the contractor’s proposed changes and modifications, necessitating the submittal of sets of calculations with every new detail. Considering it our highest priority to be continuously on alert for any logistical or structural conflicts, or for any possible flaws in calculations, details or procedures, we scrutinized all proposed modifications and procedures to assure public safety and structural integrity. As a result of our efforts, some of the most critical shop drawings were repeatedly returned to the contractor with numerous questions, comments, and clarifications before finally being approved and accepted for construction. This effort was aided by an extensive research effort involving client archives performed prior to the beginning of design in 1998. The acquired information was essential in addressing the many necessary limitations imposed on the design, underpinning methods and procedures, and construction techniques and staging. This is a prime example of the successful cooperative efforts typically exhibited during this complex project among the client, contractor, and consultant.

ACKNOWLEDGMENTS Clients New York City Transit (NYCT), New York Long Island Rail Road (LIRR), New York Structural Designer Parsons Brinckerhoff Quade & Douglas, Inc., New York Project Architect di Domenico & Partners, LLP Contractor Schiavone Construction Company, New York Santop Construction Company, New York

108 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-14.qxd

2/18/04

11:55 AM

Page 109

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Slurry walls accelerate shaft construction in rock in Los Angeles M.P. McKenna & K.K. So Jacobs Associates, Los Angeles, CA, USA

M.A. Krulc Traylor Brothers, Los Angeles, CA, USA

E. Itzig-Heine Ed Heine Construction Services, Leesburg, VA, USA

ABSTRACT: This paper details the challenges associated with the design and construction of a “figureeight” shaped, or dual cell shaft through soft ground and sedimentary rock for the Northeast Interceptor Sewer project in Los Angeles. The Humboldt Street Shaft was excavated as two cells, using a combination of both reinforced and un-reinforced concrete diaphragm walls. The Contractor chose to construct the diaphragm panels to full depth, using a Hydrofraise rather than sinking them only to the top of rock, thus eliminating the need for conventional rock support with shotcrete and ribs or dowels. This method of construction is unusual for sedimentary rock. The two cells varied in diameter and excavated depth, as each served a different purpose. The 21-m-diameter cell was excavated to a depth of 41 m to support tunneling operations and to allow construction of a junction drop structure and maintenance hole. The 12.5-m-diameter cell was only excavated to a depth of 19 m, allowing the construction of a stub-out connection to a future sewer. Other notable aspects of shaft construction included the use of rock anchors through the partition wall below the shallow cell and the use of weep holes through the shaft walls below the top of rock.

1 INTRODUCTION 1.1

Project description

The City of Los Angeles, Department of Public Works, Bureau of Engineering is presently undertaking two major construction projects to provide relief and redundancy for the aging North Outfall Sewer (NOS). These two projects are the Northeast Interceptor Sewer (NEIS) and the North Outfall Sewer – East Central Interceptor Sewer (NOS-ECIS). At the time this paper was written, the joint venture formed by Kenny, Shea, Traylor, and Frontier-Kemper (KSTFK) had completed tunneling for NOS-ECIS project. Meanwhile, a separate joint venture formed by Traylor, Shea, Frontier, and Kenny (TSFK) is currently mining the NEIS tunnel. An overall vicinity plan for both projects is shown in Figure 1. The NEIS project involves the construction of an 8.5-km-long, 2.4-m-inside-diameter (ID) sewer pipeline in a 4.0-m-diameter excavated tunnel, three drop structures, and seven special maintenance holes. The project must be completed by November 30,

2004 in order to comply with a Cease and Desist Order (CDO) deadline imposed by the Regional Water Quality Control Board. When complete, NEIS will convey flows from existing sewers and the future Eagle Rock Interceptor Sewer southward to NOS-ECIS. 1.2

NEIS will extend from a junction structure with NOS-ECIS at the intersection of Mission Road and Jesse Street, northward to the intersection of Division Street and San Fernando Road, in the Glassell Park area of Los Angeles, on an alignment roughly parallel to the east bank of the Los Angeles River. A project alignment map is provided in Figure 2. 1.3

Site-specific description

This paper focuses on the Humboldt Street work shaft, one of three work shafts being constructed for the NEIS project. The Humboldt Shaft is located on

109 Copyright © 2004 Taylor & Francis Group plc, London, UK

Alignment

09096-14.qxd

2/18/04

11:55 AM

Page 110

Figure 1. Vicinity map.

Figure 2. Alignment map.

the site of a former warehouse structure, near the intersection of Humboldt Street and San Fernando Road (see Figure 3). The shaft is 41 m deep on one side and 19 m on the other. The design team planned the shaft to serve three functions, as:

•

•

The design specified that after the tunnels are mined and the carrier pipe is installed in each reach, a combined drop-and-junction structure with associated

drive shaft for the middle-reach earth pressure balance tunnel boring machine (EPBM), tunneling towards the Richmond Shaft;

•

110 Copyright © 2004 Taylor & Francis Group plc, London, UK

receiving shaft for the upper-reach rock tunnel boring machine (TBM) tunneling from the Division Street Shaft; work shaft for the stub-out connection tunnel for future tie-in to NOS.

09096-14.qxd

2/18/04

11:55 AM

Page 111

Figure 3. Site plan.

Figure 4. Geologic profile.

maintenance holes would be constructed in the deep shaft.

(GED) in the project’s Geotechnical Data Report (GDR):

1.4

•

Geology

Three major geologic units are present at the Humboldt Shaft (see Figure 4). The following is a description of each unit, as described by the Los Angeles Bureau of Engineering’s Geotechnical Engineering Division

•

111 Copyright © 2004 Taylor & Francis Group plc, London, UK

Artificial Fill – Variable in soil type along the alignment, ranging from clayey silt to angular gravel and sand. Recent Alluvium, (Qal) – Fluvial and alluvial deposits (channel deposits, point bar deposits, and flood plain deposits) that have been deposited within

09096-14.qxd

•

•

2/18/04

11:55 AM

Page 112

the past 10,000 years (Holocene age). It consists predominantly of cohesionless silty sands, poorly graded to well-graded sands with gravel, and sands with silt and gravel. Old Alluvium, (Qoal) – These deposits are generally considered to have formed between 10,000 and 700,000 years ago (upper Pleistocene age). Brown fine gravel with fine to coarse sand, containing scattered sand with gravel layers and scattered organic fragments in a clay/silt matrix. Puente Formation, Unit 2 (Tp2) – The lower unit of the Puente Formation, an interbedded siltstone, claystone, and sandstone of Miocene age. The Puente Formation is divided into Tp1 and the Tp2 for the NEIS project. The major difference is that the beds of the Tp2 are thicker and notably stronger than the thinner beds of the overlying Tp1.

For the Humboldt Shaft the GBR indicated that up to 1 m of artificial fill could be expected, underlain by 8 to 9 m of medium dense to very dense recent alluvium, a thin layer (0 to 1 m thick) of older alluvium, then Tp2 to the bottom of the excavation. The groundwater table lies at a depth of about 9 m. None of the project borings around the Humboldt Shaft encountered gas, liquid oil, or tar within the Tp2 or alluvial soils. However, natural hydrocarbons were found in several locations along the alignment. Oil and

in this part of the Los Angeles Basin originates in the petroliferous Tp2 and propagates up along the bedding planes through seams of sand and silty sand. Therefore, Cal/OSHA classified this shaft as “potentially gassy” during shaft excavation. 2 CONTRACT REQUIREMENTS 2.1

The Contract Documents prohibit the Contractor from dewatering outside the limits of the Humboldt Shaft excavation. The reason for this restriction is to prevent migration of potential groundwater contamination and to minimize disruption to the natural groundwater flow. Therefore, the design team selected reinforced concrete diaphragm walls (slurry walls) to support the excavation through the alluvium and to prevent lowering of the groundwater table outside of the excavation. The conceptual design of the Humboldt Shaft shown in the Contract Documents is roughly circular in shape, with an adjoining shallow rectangular cell to the south. The conceptual design included 12 wall panels to approximate a ring and seven to enclose the shallow shaft for the NOS diversion structure (see Figure 5), tied together at the surface with a reinforced concrete cap beam. The design team determined the

Figure 5. Plan view of conceptual shaft.

112 Copyright © 2004 Taylor & Francis Group plc, London, UK

Slurry walls

09096-14.qxd

2/18/04

11:55 AM

Page 113

minimum shaft diameter of a circular shaft at this site to be 21 m in the alluvium and 19 m in rock. These dimensions were chosen to accommodate the permanent structures to be constructed within the shaft, as well as to minimize the amount of rock excavated. The design of the circular cell assumes the walls act as a compression ring, carrying the load by thrust in the panels in the ring’s plane, with no internal bracing required. The design of the rectangular shaft included internal steel bracing and additional reinforcing steel in the wall panels to resist bending stresses. Since slurry walls are very rigid and generally do not allow significant ground movement, the lateral earth pressure loading criteria in the contract documents are a triangular distribution based on averaging the active and at-rest earth pressure coefficients (Ka and Ko respectively). In addition to the triangular distribution, the design criteria included an apparent earth pressure envelope based on the same average K value, which was to be used only for the internally braced, rectangular cell. 2.2 Rock reinforcement The geotechnical exploration program indicated that the Puente Formation is a very weak to moderately strong rock, with most unconfined compression test values falling below 5 MPa. The designers felt the rock strength was adequate to resist the compressive stresses due to hydrostatic and horizontal rock pressures in the rock mass around the circular shaft

opening. However, to ensure a ring of intact rock would carry this load in compression, where joint sets and inclined bedding planes are present, additional rock support analyses were performed. These analyses assumed joint orientations and joint strengths developed from data contained in the GDR and GBR, as well as shaft geometry and locations of contacts between rock and soil. The designers calculated an apparent uniform rock loading, based on the force required to resist the movement of a wedge of intact rock sliding along the most prominent joint sets and/or bedding planes. The pressure diagrams in the Contract Documents included a uniform rock pressure of 67 kPa as a minimum requirement for the Contractor to design the rock support. The Contract Documents also required the Contractor to install strip drains with weep holes between the rock surface and the shotcrete to drain water-bearing joints around potentially unstable wedges that intersect the shaft walls. For this reason, the rock loading minimum design criteria did not include hydrostatic pressure for the design of rock reinforcement. The conceptual design for rock reinforcement shown in Figure 6 included two alternatives for ground support rock:

• •

Figure 6 Section view of conceptual shaft.

113 Copyright © 2004 Taylor & Francis Group plc, London, UK

W8 ribs at 1.8 m vertical spacing, with 150 mm of steel-fiber-reinforced shotcrete. Rock bolts at about 1.8 m 1.8 m spacing, 8.5 m long with 150 mm of fiber-reinforced shotcrete.

09096-14.qxd

2/18/04

11:55 AM

Page 114

These alternatives served as a basis for initial support, with provisions in the contract indicating that modification to these designs may be required, depending on conditions observed in the field. 2.3

Geotechnical instrumentation

The contract required the Contractor to install the following three sets of geotechnical instruments around and within the Humboldt Shaft:

• • •

three inclinometers (shown as ▲ on Figure 5); four piezometers (shown as ■ on Figure 5); twelve horizontal multiple-point borehole extensometers.

The inclinometers and piezometers are considered typical, minimum instrumentation for a shaft of this size and depth. The horizontal multiple-point borehole extensometers are specified for the portion of the shaft in rock. Their primary purpose is to measure lateral ground movement and warn of potentially large block movements. If the maximum lateral movement of 25 mm were exceeded, additional rock anchors or steel ribs would be installed to arrest ground movements and maintain stability of the rock mass.

3 CONTRACTOR’S REVISED DESIGN 3.1

Shaft geometry

The conceptual design consisted of two shafts adjacent to one another. The small, shallow shaft consisted

of slurry walls with waler and strut supports. It was intended that the shallow shaft would carry lateral loads by flexure, which necessitated walers and struts for support. The large, deep shaft was comprised of two different support types. In the alluvium and fill, the deep shaft would carry lateral loads by hoop compression. In the Puente Formation, rock anchors and fiber-reinforced shotcrete would carry the loads directly, or ring beams could be used in hoop compression. Using the variety of support systems as described above would have added time and complexity to the Contractor’s operations. Therefore, the Contractor elected to use a “figure-eight” or dual-cell slurry wall shaft (shown in Figures 7 and 8), similar to the concept used for the Richmond Shaft. The small cell is approximately 19 m deep, and the deep cell is approximately 39 m deep. Ed (Itzig) Heine P.E., and Steve Blumenbaum of Alpha Corporation designed the dual-cell shaft for this joint venture. 3.2

The radial walls were designed by the hoop stress method and were considered to be unreinforced compression members. Only the circular band of concrete inscribed within the limits of the slurry wall panels was considered effective in compression. The Hydrofraise (French for hydro-mill) construction method chosen by the Contractor resulted in average chord lengths of 2.4 m. The contractor-proposed chord lengths were much shorter and therefore more efficient in hoop compression than anticipated in the conceptual design.

Figure 7. Plan view of contractor’s revised geometry.

114 Copyright © 2004 Taylor & Francis Group plc, London, UK

Radial walls

09096-14.qxd

2/18/04

11:55 AM

Page 115

The short chord lengths afforded the opportunity omit a cap beam. When long chord lengths are used, a reinforced concrete cap beam is often used at the top of the shaft to provide continuity and resistance to deformation. But with the close approximation of a circle provided by the shorter panels, a cap beam is not required. Steel reinforcement was not needed for the purpose of resisting lateral loads. However, contingency reinforcing was installed, in case the panels were not installed within the specified tolerances. In that event, the panels could span vertically to remedial ring beams or walers. In the Puente Formation, drain holes were provided in the slurry walls to relieve any groundwater pressures. This was mostly precautionary, given the relative impermeability of the formation. 3.3

Center wall

The center wall was designed for different load conditions depending on depth. Where it is a common wall between the two cells, the center wall was designed for compressive horizontal loads coming from the radial loads in the two cells. It was also designed to accommodate a 1.5 m differential soil loading between elevations in each cell. Reinforcement in this area was designed to limit buckling. At elevations below the bottom of the shallow shaft, the center wall behaves differently. It is subjected to lateral earth loads as well as compressive loads coming from the radial wall of the deep cell. In this area, the wall spans vertically, and reinforcement is used for flexural strength. This

load condition controls the design of the center wall reinforcing. The wall spans between 9 m long rock anchors, which are installed on 2 m 4.5 m centers. Reinforcing cages were only terminated at the top of rock in the circular portion of the shaft. At all elevations, the ends of the radial walls were poured integrally with the end panels of the center wall, and reinforcement was provided across the center/radial wall joint. In this way, shear transfer across the joint is ensured. 3.4

Groundwater considerations

In the fill and alluvium, the slurry walls were designed for the combination of earth and hydrostatic pressures. In the Puente Formation, the slurry walls were designed only to support wedges of rock. In rock, the slurry walls confine the rock mass and the ground to support the horizontal rock and hydrostatic pressures present deeper in the rock mass. Hydrostatic pressures behind the slurry wall and in water-bearing joints around potentially unstable wedges intersecting shaft walls are relieved through the use of weep holes drilled horizontally through the slurry walls. Weep holes were not drilled through the walls above the rock to prevent dewatering of the overlying alluvium. The slurry in the alluvium provides a water barrier that the designers did not want to compromise. 3.5

Summary of advantages

There are several advantages to using the dual-cell, full-depth, slurry wall shaft instead of the conceptual design. First, the construction methods were simplified and shortened by using one excavation method. Second, Hydrofraise construction allowed the designer to eliminate the cap beam. The shorter chord lengths also minimized the need for panel reinforcement. Third, replacing the rectangular small shaft with a circular one minimized the need for reinforcement and eliminated the need for walers and struts in the small shaft. Fourth, no setback was required to change from slurry wall to rock anchor support. This not only reduced the footprint of the large shaft, but it also eliminated the need for a cast-in-place ring beam at the transition from slurry wall to rock support.

4 CONSTRUCTION MEANS AND METHODS 4.1

Figure 8. Section view of contractor’s shaft.

The TSFK Joint Venture selected Soletanche Inc. as their slurry wall subcontractor and chose the Hydrofraise

115 Copyright © 2004 Taylor & Francis Group plc, London, UK

Advantages of hydrofraise method for slurry wall construction

09096-14.qxd

2/18/04

11:55 AM

Page 116

excavation method for several reasons, which are described below: 1. Schedule Advantages – Time restrictions placed on the Contractor by the CDO required that shaft construction be expedited. 2. Achieving Tight Tolerances at Depth – The realtime data supplied to the operator from the fraise allows precise alignment of each panel which assures tight vertical joints at depth. 3. Versatility of the Hydrofraise – The Hydrofraise is able to excavate through all types of materials with minimal modification to the cutting tools. 4. Minimal Impact on Environment – The Hydrofraise method imposes minimal impact on the surrounding environment, which was a necessity in the densely populated vicinity of the shaft site. The operating principle of the Hydrofraise is similar to that of a slurry shield TBM, in which the excavated opening is supported by a pressurized suspension that balances the earth and water pressure of the excavation. In most cases, this suspension is a bentonite and water slurry. The slurry acts not only as a support fluid, but also as a transport medium. The ground excavated by the cutting tool is mixed with the support fluid slurry near the excavation face, where it can then be pumped to the surface. A separation plant, usually on the surface, then separates the support fluid from the ground, and the fluid is again pumped to the excavation face. Fresh bentonite can be added as slurry properties dictate. A typical equipment spread for the Hydrofraise method is comprised of a modified crane, a fraise cutting tool, a bentonite slurry mixing and storage facility, a separation/de-sanding plant, and one or two support cranes. The specific layout of the Hydrofraise used in this project can be seen in Figure 9. 4.1.1 Schedule advantages The Hydrofraise can shorten the construction schedule because of its ability to continuously excavate. The tool is lowered under its own weight into a pre-built concrete guide wall, with the cutting wheels turning. It continues excavating until it reaches the desired depth. In conventional clamshell excavation, the continual raising, lowering, and dumping cycles are time consuming. However, as with any sophisticated piece of equipment, the Hydrofraise is susceptible to mechanical and electrical downtime, whereas the clamshell can be kept running with a minimum of specialized maintenance, tools, and equipment. However, on the NEIS project, the Hydrofraise experienced minimum downtime and was therefore able to keep the project on schedule. Slurry wall panel construction at the Humboldt Shaft lasted 49 days, with the crew working two 10-hour shifts. The approximate area of the slurry wall panels is 3,381 m2, in elevation.

Figure 9. “The Hydrofraise evolution II” by Soletanche, Inc.

4.1.2 Achieving tight tolerances at depth One of the major reasons the Contractor chose the Hydrofraise method was due to its precise excavation control. The design assumptions of shaft geometry are dependent upon the construction tolerances that the equipment can achieve. The cutting tool is equipped with inclinometers and tilt meters that are linked to a computer screen in the operator’s cab. The operator is able to read the information provided by the instrumentation in real-time and make corrections as needed. Several features are available to the operator for steering purposes. First, each of the two rotating cutting wheels can be run at variable speeds to correct for left and right misalignment. Second, the entire cutting head can tilt up to 1.5° in the vertical plane of excavation to correct for front and back misalignment. Such precise control over tool guidance enabled the Hydrofraise to excavate panels on the NEIS project within a tolerance of 0.3%, which equates to a variance of only 125 mm from the designed vertical alignment, over a depth of 42 m. 4.1.3 Versatility of the Hydrofraise The Hydrofraise cutting tool can quickly and easily adapt to changes in ground conditions with modifications to the cutters. Both of the cutting wheels can be removed and replaced in a single shift, which allows the Hydrofraise to perform in almost any ground

116 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-14.qxd

2/18/04

11:55 AM

Page 117

Figure 10. Photo of broken teeth.

Figure 11. Excavation in rock.

condition. As an example, at the Richmond Shaft site, which was the first of three slurry walls constructed for NEIS, the Hydrofraise was equipped with selfcleaning cutting paddles to deal with the soft claystone and mudstone present in the area. However, when harder sandstone was encountered at the Humboldt site, the slurry wall Sub-Contractor quickly replaced the paddles with carbide tipped picks. The carbide picks performed well; the only problem was chipping of the carbide tips in the hardest layers of the Tp2 as shown in Figure 10. It was the versatility of the Hydrofraise that enabled the Contractor to further accelerate the schedule by extending the slurry walls through the Puente Formation and eliminating rock bolting and shotcreting from the shaft excavation activities.

concrete of the secondary panels to bond to, thus producing a strong and relatively watertight joint. The concrete was poured using dual tremie pipes in the primary and follow-up panels and a single tremie pipe in the secondary panels.

4.2

Panel construction sequence

The circular shape of the shaft was approximated with short chords because the Hydrofraise is limited to excavating rectangular shaped sections. The chord length for the Humboldt slurry wall ranged from 1.804 m to 2.448 m. The wall was constructed in 43 “bites,” with each bite being one pass of the cutting tool. The wall was also constructed in 19 “panels,” which were either: a primary panel comprised of three bites, a secondary panel comprised of one bite, or a follow-up panel of five bites. A secondary panel separated each primary panel and the follow-up panels were used to create the joint between the two cells. Each primary and follow-up panel was excavated and concreted first, before the secondary panels were excavated. Tight joints between the primary and secondary panels were constructed by spacing the primary panels so that the Hydrofraise cut into the previously poured concrete of the primary panels, while it was excavating the secondary panels. The cutting of the primary panels produced a rough surface for the

4.3

Because the shaft was originally classified as “gassy” by Cal/OSHA, the Contractor chose to drill a test hole prior to excavation of the shaft for the purpose of drawing gas samples. It was hoped that the Cal/OSHA would reclassify the shaft based on favorable results from these samples, and thus allow the use of conventional equipment for the shaft excavation. The shaft was indeed reclassified to “potentially gassy” with special conditions, based on the gas samples, and the Contractor was allowed to proceed with conventional equipment. The first 5.2 m of the shaft, which consisted mostly of artificially backfilled sand and alluvial sand and clay, was excavated from the surface by a Caterpillar 325 excavator. For the next 4 m of excavation, the CAT 325 excavator was placed in the shaft where it then loaded two 3.8 m3 circular muck skips, which were hoisted on a single line by a 125-ton-capacity American 9260 crane, as shown in Figure 11. The crane was previously factory modified for deep tunnel operations. This 4 m of excavation consisted mostly of alluvial sand and clay. At an approximate 7.6 m depth, the soil became sticky and produced a strong hydrocarbon odor. A chemical analysis of the excavated material revealed that the soil contained a high concentration of natural petroleum hydrocarbon, which is not uncommon in the Los Angeles Basin. The soil was classified as “Contaminated” and was removed and dealt with by the Contractor’s environmental subcontractor.

117 Copyright © 2004 Taylor & Francis Group plc, London, UK

Shaft excavation

09096-14.qxd

2/18/04

11:56 AM

Page 118

The Tp2 Formation was encountered at a depth of 9.1 m, at which point the CAT 325 was no longer able to freely excavate the material with a bucket. Consequently, a 53 kN hoe-ram was attached to the CAT 325 to break the hard layers of the Tp2,, while a CAT 312 excavator was put into service to load the broken material into muck skips. The production rate of each piece of equipment was quite evenly matched so that the two excavators could follow each other around the shaft, one breaking material and one loading material. The efficiency of the operation led to a production rate of about 550 m3 per shift, which equated to approximately 1.2 m (in depth) per shift in the large cell. The CAT 312 excavator was utilized in the small cell, since the shaft was not large enough to accommodate the CAT 325. With aggressive bucket teeth, the CAT 312 was able to excavate nearly all the material down to a depth of 26 m unassisted. Where it was not able to dig, a smaller hoe-ram attached to a Case 580 loader assisted the operation by breaking the harder material. With these production rates, the shaft was sunk in 39 working days, utilizing two 8-hour shifts per day. 4.4

Figure 12. Drilling weep holes in the slurry wall.

Rock anchor installation

The rock anchor scheme in the straight center wall of the shaft consisted of four rows of anchors with five anchors per row and a vertical and horizontal spacing of 4 m and 2 m, respectively. The design specified 9 m long, 35-mm-diameter, 1,030 MPa Dywidag Threadbar anchors, which were to be installed in a 70 mm hole and encapsulated in cementitious grout. The Contractor chose to use a Gardner-Denver PR 123 rock drill mounted to an ATD 3800 Air Trac drill carrier (as seen in Figures 12 and 13) for the drilling in the relatively soft rock of the Tp2 Formation, 10-m-long, 100 mm diameter holes could be drilled in a matter of minutes. Shorter holes drilled as weep holes yielded some water immediately after drilling and periodically during a relatively dry rainy season in 2002. Contract Specifications stated that shaft excavation was not allowed more than 1 m below a row of anchors until each bolt was pull tested. In order to gain high early strength and a quick turnaround on the pull test, a prepackaged non-shrink rock anchor grout formulated by Euclid was initially selected for the cementitious encapsulation. It was to be batched and pumped by a Hany IC 310 colloidal mixer. A prepackaged product was selected with the hope that it would reduce batching times and improve quality assurance of the grout, since the Contractor could not afford to halt shaft sinking production in order to reinstall a failed anchor. However, after a number of unsuccessful attempts to mix and pump the prepackaged product, it became apparent that the Euclid material was not compatible with the Hany equipment at the water-to-cement

Figure 13. Drilling for rock anchor installation.

ratio the Contractor needed to achieve. It was decided that a switch would be made to a traditional cementand-water grout mix. Master Builders MEYCO Fix Flowcable was added to the mix to reduce water requirements, increase pumpability, and provide nonshrink properties. After this change was made, the Hany mixer and pump performed flawlessly. Each anchor was pull tested approximately 20 hrs after installation, with all anchors passing, except the last one.

118 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-14.qxd

2/18/04

11:56 AM

Page 119

execution of the NEIS contract. Jacobs Associates provided design services for the Department of Public Works, Bureau of Engineering and prepared the minimum design criteria used by the Contractor to design the Humboldt Shaft. Parsons Brinckerhoff Construction Services/Brown & Root Services (J.V.) and Jacobs Associates continue to work with Baron Miya and Rajni Patel of the Bureau of Engineering as an integrated team managing the construction of the NEIS project. Chris Smith and Richard Calvo of the Bureau of Contracts Administration supervised inspection of the work and ensured that the slurry walls were constructed in accordance with the demanding project specifications. Figure 14. Weep holes drilled through the slurry wall.

REFERENCES

The anchor failure is believed to have been caused by a transient water flow which washed grout out of the hole. Another anchor was installed immediately and passed pull testing. 5 CLOSING REMARKS Excavation of the Humboldt Shaft ended on March 28, 2003. The shaft excavation was never on the project’s critical path. This can be attributed to the successfully planned and implemented slurry wall operation devised by the TSFK joint venture and their subcontractor, Soletanche Inc. The joint venture kept the project on schedule through shaft construction, despite the rigorous demands dictated by the CDO.

McKenna, M. P., Traylor, D. A., Tarralle, B. and Itzig-Heine, E. 2003. Design and Construction of a Deep, Dual-Cell, Slurry Wall Shaft in Soft Ground, Proceedings of the Rapid Excavation and Tunneling Conference, 368–382. City of Los Angeles Department of Public Works, 2001. Geotechnical Baseline Report: Northeast Interceptor Sewer (NEIS. City of Los Angeles Department of Public Works, Bureau of Engineering. City of Los Angeles Department of Public Works, 2001. Geotechnical Data Report: Northeast Interceptor Sewer (NEIS. City of Los Angeles Department of Public Works, Bureau of Engineering.

ACKNOWLEDGEMENTS The authors would like to recognize some of the key firms and individuals responsible for the design and

119 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-15.qxd

2/18/04

11:56 AM

Page 121

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Performance of Russia Wharf Buildings during tunneling Hugh S. Lacy Mueser Rutledge Consulting Engineers, New York, NY, USA

Marco D. Boscardin Boscardin Consulting Engineers, Inc., Amherst, MA, USA

Leslie A. Becker Massachusetts Bay Transportation Authority, Boston, MA, USA

ABSTRACT: Twin transit tunnels were mined through the timber pile foundation system that supports two historic buildings while maintaining the serviceability and occupancy on all seven floors. Ground stabilization using soil freezing was employed to permit tunneling via NATM methods. This paper discusses and compares the anticipated and actual performance of the buildings.

1 INTRODUCTION MBTA’s Silver Line Phase II in Boston (formerly known as the South Boston Piers Transitway Project) includes construction of a transit tunnel below two, 100-year old buildings. The tunnel under the building is a binocular-shaped structure about 8.5 m high by 13 m wide, constructed using NATM techniques. The tunnel passes directly under both buildings with about 4.6 to 7.6 m feet of cover. Lacy et al. (2000) describes how the buildings were supported while tunnel construction occurred below. This paper will focus on the response of one of these buildings, the Graphic Arts Building, the construction activities and how adverse impacts on the building were mitigated. At the time this paper was prepared, the tunnels were nearly 80% complete. 1.1

The buildings

The three Russia Wharf buildings are shown in Figure 1. The buildings are seven-story, historic structures with steel frames and brick and granite masonry facades, circa 1897. The northwest corner of the third building, the Tufts Building, is located immediately adjacent to the tunnel. The buildings have single-story basements used for parking, and the brick exterior bearing walls and steel interior columns are supported on large granite block pile caps and timber piles. The floor system in each building consists of relatively flat arch masonry and concrete barrel vaults spanning

Figure 1. Russia Wharf Buildings, facing east.

between steel beams that frame into the columns. Typical column spacings are 4 m to 4.6 m . The building use is light commercial and tenants include a ship gallery, architectural firms, development companies, a copy/ printing shop, and a restaurant. The tunnels extend diagonally below the Russia Building (foreground), an atrium between the buildings

121 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-15.qxd

2/18/04

11:57 AM

Page 122

RUSSIA WHARF

EXISTING SEAWALL (LOCATION AND DEPTH UNKNOWN)

ELEVATION, FEET (PROJECT)

EDGE OF WOODEN WHART

Figure 2. Site plan.

M.H.W. 110 EL.104.91 M.L.W. FILL 100 EL.95.10 ORGANIC SOIL 90 Fort Point channel 80 70 Tunnel Alignment 60 CLAY 50 GLACIAL TILL 40 30 20 10 0 BEDROCK -10 -20 93+00 94+00 95+00 96+00 97+00 98+00 99+00 100+00 STATION IN FEET

Figure 4. Soil profile along tunnel alignment.

Figure 3. Section through Graphic Arts Building.

and then below the Graphic Arts Building (middle building). The prime concerns relative to the buildings during tunnel construction include safety of the tenants and users of the buildings, protecting the historic fabric of the structures, and maintaining a facility that continues to serve the functions of the tenants without interruption. 1.2

Transit tunnels

The tunnel alignment is shown on Figure 2. The tunnel profile extends deeper to the east as it approaches Fort Point Channel. Figure 3 is a section through the tunnels at the Graphic Arts Building. 2 DESIGN 2.1

Subsurface conditions

The project site is located at the edge of the Boston’s Shawmut peninsula and adjacent to the Fort Point Channel. The buildings are constructed on an area that is filled land created by several episodes of filling of the working waterfront and mudflats during the 19th century. The soil profile along the tunnel alignment is shown in Figure 4 and in general consists of fill over organic silts and clays, over a silty marine clay, over a

dense to very dense silty, sandy glacial till. The fill ranges from 1.5 m (5 ft) to 4.6 m (15 ft) thick generally located above the crown of the tunnel. The fill includes granular and silty soils (primarily glacial till origins) excavated from the uplands to the west and miscellaneous debris ranging from timbers, boulders, cut stone, bricks etc. The organic silts and clays are discontinuous and range from 0 m to more than 3 m (10 f t) thick and are of marine origin (formerly in the mudflats and channel bottom area). The tunnel crown runs along and generally below the organic/marine clay interface. The marine clay is Boston Blue Clay and ranges from 3 m (10 ft) to more than 9 m (30 ft) thick along the tunnel alignment under the buildings. The upper 3 m of the clay is a stiff crust and the remainder of the clay is of medium stiff consistency. For most of the alignment, the tunnel face is primarily in the clay. Below the clay is a very dense silty, sandy glacial till. The invert of the tunnel is in the glacial till at the western end of the tunnel. Water levels are nominally at the level of the adjacent Fort Point Channel. 2.2

The tunnels were planned in a “binocular” shape to minimize total tunnel width yet permit staged construction with reduced impact on the buildings. Figure 3 illustrates the relative position of the two tunnels with the top bench of the tunnel to the right mined first using lattice girders and shotcrete for the initial lining. The secondary lining consisted of castin-place invert and vertical dividing wall and fiber reinforced shotcrete for the remaining curved sections. 2.3

Support of buildings

The Russia Building was underpinned prior to tunneling using high capacity mini-piles installed from within

122 Copyright © 2004 Taylor & Francis Group plc, London, UK

Tunnel construction method

09096-15.qxd

2/18/04

11:57 AM

Page 123

Figure 7. Tunneling procedure.

the load from a row of severed timber piles to arch around the narrow strip being mined at the tunnel face as illustrated in Figure 7.

Figure 5. Russia Wharf Building underpinning plan.

2.4

Figure 6. Temporary column support at Graphic Arts Building.

the building basement. The mini-piles were installed along the edges of and in the center wall between the two tunnels. A system of distribution beams transferred the column and wall loads to the mini-piles as shown in Figure 5. Columns near the tunnels were re-leveled using temporary cribbing and jacks. The Graphic Arts Building was temporarily underpinned and supported on a pad of artificially frozen ground prior to tunneling. As the tunnels were being mined, the timber piles were cut off in the tunnel face and the pile stub in the tunnel roof was embedded in the lining, transferring support of the column above to the tunnel. Temporary cribbing and jacks at the column base (Figure 6) were used to minimize building movement. The frozen ground provided pile cap support causing

123 Copyright © 2004 Taylor & Francis Group plc, London, UK

Groundwater control

The water level in the fill, organic soils and clay is nominally at the level of the adjacent Fort Point Channel (FPC) and approximately 4.6 m to 7.6 m above the crown of the tunnel. The measured groundwater levels exhibit a tidal influence of 0.3 m to 1.0 m depending on distance from the channel. During test pitting in the basements of the buildings, it was noted that gaps were present under the basement slab and that water rushed into the test pits during high tide. In addition, highly conductive voids related to the site filling in the 19th century were expected to be present around timbers and cribbing. Due to the shallow nature of the tunnel crown, the weak and loose condition of the fill and organic soils at or immediately above the tunnel crown and the close proximity of a large source of water (the FPC), positive means of ground water control/cutoff were needed to tunnel safely. This was provided by constructing frozen ground cutoff walls on either side of the tunnel alignment and mass ground freezing over top of the tunnel. In addition, jet grouting was used to form closure between the frozen ground mass and frozen cutoff walls and the existing slurry walls at the west end of the tunnel. During the freezing, the contractor experienced difficulty in reaching target temperatures in the Atrium cutoff wall suggesting the possible presence of voids in the area. This condition was addressed by injecting, at low pressures, limited mobility cement/bentonite grout through a line of cased grout holes to fill larger voids. This was followed by permeation grouting using a thin microfine cement grout to fill the smaller, remaining voids. Ground water control methods were

09096-15.qxd

2/18/04

11:57 AM

Page 124

effective so only limited sumping of water from inside the tunnel was performed. 2.5

Ground stabilization

Ground stabilization consisted of artificial ground freezing above the tunnel spring line below the buildings and grouted spilling across the tunnel roof combined with local dewatering where the tunnel was advanced below the atrium and non-building areas. Tunneling was advanced in 760 mm increments. Ground freezing causes an increase in shear strength through a phase change of soil moisture to ice. Colder frozen ground results in higher frozen strength. Frozen soil temperature requirements ranged from 10°C to 15°C, and were determined based on analysis of stresses in the ground around the mined opening. Freezing of clays, silts and to a lesser extent silty sands causes soil expansion due to the expansion of water during phase change. This had the potential to cause both lateral and vertical ground movement particularly where a large area is being stabilized. This is in contrast to a single line of freeze pipes in a circular pattern for stabilizing the sides of a deep shaft where much smaller ground movement is generally observed. Initial horizontal ground movement occurs during the period where the ground is freezing around freeze pipes and spreading to merge with that of adjacent freeze pipes. The initiation of ground freezing was planned in two stages to minimize heaving of the pile caps. The first stage was to circulate brine in the pipes closest to the pile caps to permit lateral movement resulting in ground heave between the pile caps where the basement floor had been removed. The second stage turned on the flow of chilled brine to freeze pipes further from the pile caps in areas where ground heave would have less impact on the building foundations. Following initial completion of the frozen mat, temperatures were gradually lowered to meet the strength requirements (Figure 8). During this period, additional ground heave occurred due, in part, to increased growth of the frozen zone downward and horizontally at the edges of the frozen pad. Prior to the start of tunneling below the Graphic Arts Building, contours of ground temperature above the tunnel alignment (Figure 9) were evaluated to determine if soil strengths met or exceeded the design requirements. 2.6

Mitigation of movements

The tunnel design incorporated temporary and permanent underpinning, as well as, ground-freezing to support the buildings and control their movements during and after tunnel construction which included cutting out and removing the timber piles and transferring the loads to the tunnel lining or the permanent underpinning. To evaluate the combined impacts

Figure 8. Freeze pipes in place.

Figure 9. Ground temperature contours.

including the tunneling, ground displacements due to the tunneling were estimated using an ABAQUS finite element model (Dr. G. Sauer Corporation, 1999). The effects of the ground freezing on soil properties (Mueser Rutledge Consulting Engineers, 1998) were included in the modeling. The tunneling-related ground displacements were then considered in combination with the open cut-related ground displacements to estimate building distortions and potential for damage. The conclusions of the combined impacts evaluation were that the tunnel construction could maintain the building response in the range typically associated with very slight to slight cosmetic damage. Due to the building foundation loads and the very shallow nature of the tunnel, the tunnel construction included temporary underpinning (Figure 6) to provide additional support to building foundations during the tunneling. The underpinning system also included provisions for adjusting column elevations in the freezing zones to mitigate the effects of potential ground movements due to ground freezing. Based upon input and requirements from the Building Owner, the Historic Commission, the Project Conservator, and the MBTA, the Design Team established threshold and limiting values for heave/settlement and angular distortion of the buildings. Threshold heave/settlements and angular distortions

124 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-15.qxd

2/18/04

11:57 AM

Page 125

3 CONSTRUCTION

10

3.1

Temperature, deg C

5

0

-5

-10

-15

-20

Jul-03

Aug-03

Jun-03

Apr-03

May-03

Mar-03

Jan-03

Feb-03

Dec-02

Oct-02

Nov-02

Sep-02

Jul-02

Aug-02

Jun-02

May-02

-25

Date

03/14/04

12/15/03

09/16/03

06/18/03

03/20/03

12/20/02

09/21/02

Column Lowered 8-21-02 Column Lowered 11-22-02

06/23/02

03/25/02

12/25/01

09/26/01

06/28/01

03/30/01

12/30/00

10/01/00

0.03 0.02 0.01 0 -0.01 -0.02 -0.03 -0.04 -0.05 -0.06 -0.07 -0.08 -0.09 -0.1

Column Lowered 7-3-02

Figure 10. Ground temperature history at CPC-06 (Graphic Arts Building).

Change in Elevation (ft)

Building movement and building column adjustments prior to tunneling

Date

Figure 11. Ground movement with time columns C-9 and D-9.

set in the design documents were 6 mm and 1/700, respectively. Where the frozen zone extended close to sensitive structures such as the building heating plant or where the frozen soil groundwater cutoff wall extended close to areas that could be impacted by ground freezing expansion, ground warming pipes were used to prevent expansion. The Contractor also used jet grouting in place of the frozen groundwater cutoff wall to avoid a buildup of lateral load on slurry walls. Following a period of sustained ground freezing to bring temperatures down to required levels and experiencing a high rate of column heave in the area of ground freezing, a program of cycling the freezing system was employed to reduce the rate of heave. This maintenance mode typically involved turning off part of the freezing system for two weeks followed by reactivation for one week as shown on Figure 10 for one of the nearly 50 temperature monitors. This caused a slowly rising average temperature trend which was maintained below required levels. The impact on column heave was dramatic as shown on Figure 11.

During construction, the elevation of each column in the structures was monitored at least weekly, and in local areas more frequently during periods of active tunneling within 15 m of the tunnel face. These data were transmitted to the engineer immediately, and the contractor and engineer each reviewed the data and compared it to settlement and angular distortion limits agreed upon with the Building Owner. Based on the data and observations of the building response, decisions were made regarding when and how much to adjust column elevations to keep the angular distortion within the agreed upon limits. Prior to the start of tunneling in October 2002, the underpinning had been installed and the freezing operations for the cutoff wall and mass freeze over the tunnel had achieved target temperatures. At this time, the measured ground displacements in response those construction activities ranged from 0 mm to a heave of approximately 56 mm. Figure 12 illustrates contours of total column heave without column adjustment. An example of column adjustment is shown on Figure 11. During the pre-tunneling construction period, adjustments to the building column elevations were made periodically to maintain building angular distortions at 1/400 or less after consultation and agreement with the building owner. This equates to a differential vertical displacement of about 10 mm or less between adjacent columns. Note: typical measured vertical column displacements during underpinning activities were in the range of about 5 mm to 10 mm. In some areas, the ability to adjust the columns was limited and the angular distortion limits were further relaxed, but in no case were angular distortions less than 1/200 permitted to develop. Most column pairs sustained angular distortions of 1/700 or less. Columns at the edges of the freezing zone generally required the most frequent adjustment. Column adjustments were generally performed during periods of low occupancy of the building (e.g. 5 a.m.) and were monitored by representatives of the MBTA, the Engineer, and the Building Owner. During this period, column elevation adjustments were necessary, relatively frequently, sometimes a couple of times a month. The building response was consistent with this level of angular distortion with cracks observed in the 1 mm or less range. 3.2

Building movements that developed during the time period that tunneling occurred have two

125 Copyright © 2004 Taylor & Francis Group plc, London, UK

Building movement and column adjustments during and following tunneling

09096-15.qxd

2/18/04

11:57 AM

Page 126

components: displacements (settlements) related directly to the tunnel excavation and displacements (heave or settlement) related to the maintenance freezing of the ground above the tunnel. After mining of the outbound side of the tunnel was complete (approx. March 2003), cumulative ground displacements from the start of MBTA construction ranged from about 6 mm of settlement to 46 mm of heave. Measured displacements directly related to tunnel excavation were settlements generally in the range of 5 mm to 8 mm. Although adjustments were made to column elevations during the tunnel construction, the adjustments were made in response to the ground freezinginduced heave and not the tunnel settlements. Again column displacements and angular distortions were monitored and adjustments were made when the data indicated angular distortions approaching agreed upon limits. Again, most column pairs sustained angular distortions of 1/700 or less. However, one column pair at the edge of the freezing zone sustained angular distortions approaching 1/200 and underpinning and adjustment of a column that had not previously been underpinned was performed to correct the condition. Building response was consistent with the distortions. Observed cracking was generally in the 1 mm or less range, even at the newly underpinned column, with only a couple of instances where preexisting cracks opened more. One particular case was in a stairwell at the edge of the frozen zone where the configuration of the stairwell and entry doors on each floor served to concentrate movements and open an existing crack by 5 mm to 6 mm. During this phase of the work, column adjustments were performed at a frequency of less than 1 per 2 to 3 months. After the mining of the inbound tunnel through the Graphic Arts Building was completed (approx. October 2003), cumulative ground displacements from the start of MBTA construction ranged from about 15 mm of settlement to 41 mm of heave. Measured displacements directly related to tunnel excavation again were settlements generally in the range of 5 mm to 8 mm. During this phase of the work, the ground freeze energy cycling was tuned sufficiently so that no column adjustments were performed. Cycling of ground freezing started in early 2003 (Figure 10) caused settlement of column pile caps (Figure 11) as illustrated in a comparison of contours in Figures 12 and 13. 3.3

Tunnel deformation during and following tunneling

Following completion of the outbound tunnel initial lining interior monitoring of control points mounted on the lining indicated both north and south tunnel walls moving slightly to the south at a rate of 0.1 to 0.7 mm per day with no significant tendency of convergence or expanding of the sidewalls. The southward

Figures 12 and 13. Contours of pile cap heave from 9/02 and 10/03, respectively, (in feet).

126 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-15.qxd

2/18/04

11:57 AM

Page 127

movement appears to be the result of the more predominant mass of frozen ground to the north that continued to expand. Total movement was not significant.

(5) Successfully limiting building distortion required extensive monitoring results to adjust building columns in a timely fashion. This required a high level of collaboration between all parties.

4 CONCLUSIONS ACKNOWLEDGEMENTS This highly unusual and challenging method of constructing transit tunnels below historical buildings resulted from the need to maintain the fully occupied buildings in operation during construction. Successful completion of construction of this project permits the following conclusions: (1) Ground freezing can successfully be employed to stabilize ground below pile caps permitting mining of tunnels, severing of piles and resupport of the building on the tunnel lining. (2) Ground movement during freezing including the creation of forces necessary to lift 7-story buildings are a function of ground conditions, groundwater control during freezing, confinement, frozen ground temperature, and details of the freezing operation as implemented and maintained by the contractor and can not easily be predicted. Measured values can exceed estimated values. During formation of the mass freeze relatively modest ground heave occurred. During subsequent lowering of frozen ground temperatures, the rate of column pile cap heave increased markedly, due in part to secondary freeze affects. (3) Use of cycling of the freezing units to maintain the required temperatures without causing expansion of the frozen ground area was very effective in limiting additional building distortion. The measured building pile cap settlement during and following tunneling appears to be related primarily to cycling of the freeze system rather than tunnel deformation. (4) The method of releveling columns at their base used for this project successfully limited building distortion to acceptable values. The building sustained relatively little distress. The most noticeable cracks were where the frozen ground cut-off walls passed beneath exterior walls where there was no provision to relevel these massive walls.

The authors wish to acknowledge several firms and individuals for the information that form the basis for this paper, including: The Massachusetts Bay Transportation Authority – Project Owner (Mr. D. Ryan, Mr. E. Karpinski, Jr., and Mr. T. Bretto); DMJM+ HARRIS – Prime Design Consultant; Dr. G. Sauer Corporation – Tunnel Design Consultant; Modern Continental Construction Company – General Contractor (Mr. T. Hennings and Mr. R. Cotes); Layne Christensen Company – Ground Freezing Subcontractor (Mr. J. Sopko); Mueser Rutledge Consulting Engineers – Ground Freezing and Underpinning Consultant (Mr. F. Arland, Mr. T. Popoff, Dr. D. Chang, Mr. P. Deming); and GEI Consultants, Inc. – Geotechnical Consultant (Ms. K. Wood).

REFERENCES Dr. G. Sauer Corp., 1999. Memoranda to GEI Consultants, Inc. regarding ABAQUS FEM analyses for Russia Building and Graphic Arts Building, MBTA Contract E02CN15, South Boston Piers, Russia Wharf and Fort Point Channel Tunnel. Boston, MA. GEI Consultants, Inc., 1999. Estimate of combined impacts of CA/T (C17A1 & C17A3) and MBTA Transitway construction on Russia, Graphic Arts, and Tufts buildings, MBTA Contract E02CN15, South Boston Piers, Russia Wharf and Fort Point Channel Tunnel. Boston, MA. GZA Geoenvironmental, Inc., 1998. C17A3/17A1 combined impacts. Memorandum prepared for FST/HNTB, Joint Venture, Central Artery (I-93)/Tunnel (I-90) Tunnel Project. Boston, MA. Mueser Rutledge Consulting Engineers, 1998. Geotechnical analysis of tunnel stability using artificially frozen soil, South Boston Piers Transitway, Section CC03A. Russia Wharf report prepared for Frederic R. Harris.

127 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-16.qxd

2/18/04

11:58 AM

Page 129

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Blasting adjacent to high voltage duct banks K.R. Ott, D.A. Anderson & S.E. Haq Parsons Brinckerhoff Quade & Douglas, Inc., New York, USA

ABSTRACT: Development of appropriate specifications and criteria related to blasting is a necessity for a project to proceed effectively. A vibration limit of 5 cm/sec is often applied to close in blasting, where it is often not appropriate. Rock excavation was required for construction of an elevator pit in an underground station that will be part of New Jersey Transit’s Hudson-Bergen Light Rail Transit System. The new underground station is being constructed within an existing railroad tunnel, the Weehawken Tunnel, which was originally constructed in 1881 as a dual-track freight line. New Jersey Transit purchased the tunnel to incorporate it into the HudsonBergen Light Rail Transit System, which will link the line along the west shore of the Hudson River, through the Palisades Ridge, to areas north and west of the Hudson River in New Jersey. The elevator pit at the stationshaft intersection was excavated as close as 0.70 m from existing in-service 230 kV oilostatic (pressurized oil conduit) electric duct banks using controlled blasting techniques. The two duct banks run through the tunnel along the invert-sidewall corners of the tunnel. Blasting operations were monitored with dynamic foil strain gages, high-frequency geophones, and standard blasting seismographs. Our recommended criterion of 50 cm/sec as a safe maximum for vibration levels and a strain limit of 600 microstrains as a conservative value for protecting the structures from damage were successful. Peak particle velocities of up to 25 cm/sec and strains up to 150 microstrains did not cause damage.

1 INTRODUCTION Development of appropriate specifications and criteria related to blasting is a necessity for a project to proceed effectively. A vibration limit of 5 cm/sec is often applied in close-in blasting, where it is often not appropriate. It can be hard to convince a layperson that a vibration criterion for threshold damage to residential structures doesn’t apply to a duct bank, a pylon, a dam, or a tunnel wall. Damage to credibility can occur with no damage to structures. The best approach is to make sure the specifications are written clearly in the first place, and communicated to concerned parties. Sometimes, measurements other than standard seismographic data are needed to provide assurance that proper protective measures are being taken. We will discuss development of specifications for a project in which an existing structure was in close proximity to proposed blasting, and how the job was completed. Rock excavation was required for construction of an elevator pit in an underground station that will be part of New Jersey Transit’s Hudson-Bergen Light Rail Transit System. The new underground station is being constructed within an existing railroad tunnel, the Weehawken Tunnel, which was originally constructed

in 1881 as a dual-track freight line. New Jersey Transit purchased the tunnel to incorporate it into the Hudson-Bergen Light Rail Transit System, which will link the line along the west shore of the Hudson River, through the Palisades Ridge, to areas north and west of the Hudson River in New Jersey. The major tunnel construction components include the following: 1. Demolition and reconstruction of two portal structures, 2. Construction of enlarged and retrofitted running tunnels with dual track ways from the portals to an underground station, and 3. Construction of an underground station, which includes a large, single shaft for rider access from the ground surface and ventilation and other utilities. The elevator pit at the station-shaft intersection was to be excavated as close as 0.70 m from existing in-service 230 kV oilostatic (pressurized oil conduit) electric duct banks using controlled blasting techniques. The two Public Service Electric & Gas Company (PSE&G) duct banks run through the tunnel along the invert-sidewall corners of the tunnel (Figure 1). Each existing 230 kV line consists of a 230 kV cable installed in a 22 cm pressurized oil pipe. The pressurized steel pipe and two 12.7 cm PVC conduits are encased in

129 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-16.qxd

2/18/04

11:58 AM

Page 130

Figure 1. Weehawken Tunnel cross section with PSE&G duct banks.

Jurassic igneous sill that extends as a ridge for more than 64 km along the west bank of the Hudson River in a northeasterly direction from Jersey City through Hoboken and Weehawken to Haverstraw, New York. Near vertical jointing patterns are characteristic of the Palisades cliff-face and generally extend throughout the body of the sill. The bedrock of the Palisades is a gray to dark gray fine-to medium-grained diabase, and is characterized as strong to very strong, close to moderately fractured, unweathered to slightly weathered. Unconfined compressive strength tests on rock samples varied from 165 to 248 MPa.

3 PREPARATION

Figure 2. Elevation view of the elevator pit and duct bank locations along the tunnel.

unreinforced concrete to form a duct bank in each invert corner. The duct banks abut directly against the rock tunnel sidewalls and are typically cast on top of ballast material on the tunnel invert, although in some places the duct banks rest directly on rock. Construction plans called for both duct banks to remain in the tunnel, but to be relocated from their present location in the invert corners to a pair of new trenches to be cut into the tunnel invert, below future light rail tracks. Prior to relocating the existing duct banks, the elevator pit was excavated (Figure 2). The tunnel construction work is being performed for New Jersey Transit (NJ TRANSIT) by a joint venture of Frontier-Kemper Constructors, Inc., Shea Construction, and Beton und Monierbrau (FKSB). Parsons, Brinckerhoff, Quade and Douglas, Inc. (PBQD) of New York is the design engineer for NJ TRANSIT.

Construction specifications outlined blasting limitations, blast design requirements, and blast monitoring requirements for blasting adjacent to the PSE&G duct banks. During the design phase, a maximum strain criterion was developed to reduce the potential for damage to the duct bank from blast vibrations. A strain limit was used as a means to avoid overstressing the steel conduit of the oilostatic line, which in turn maintains continuous service. Although vibration criteria have been developed for pressurized pipelines (Westine et al., 1978; Dowding, 1996), these are typically buried in soil, and the strain calculations are related to the soil-pipe interaction. We were looking for criteria for a concrete beam, which is the enclosure that stiffens and supports the oilostatic line. Although an exact analog was not available, we knew that strain limits had been used on other close in blasting projects; such as the Folsom Dam project (Revey and Scott, 1999) and a Lock and Dam project in Minneapolis (Tart et al., 1980). Based upon the spall limits in the Tart et al. study, a limit of 600 microstrains was stated in the contract specifications as a conservative value for protecting the structures from damage. Obtaining strain measurements are time-consuming and expensive. Therefore, in addition to the strain measurements, peak particle velocity (PPV) measurements were required to be recorded at the same location, to develop a relationship between strain and PPV. This relationship would be used to assess predicted strain levels based on PPV measurements during future production blasting. However, during the planning of the blasting program, FKSB requested that a vibration criterion be established prior to any blasting for ease of determining allowable charge weights. We used the plane-strain conversion formula (Dowding, 1996): (1)

2 GEOLOGY The rock type along the tunnel consists mainly of the Palisades Diabase. The Palisades Diabase is an early

. where  strain (in microstrains), u  PPV, and c  seismic velocity of the transmitting medium

130 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-16.qxd

2/18/04

11:58 AM

Page 131

(concrete) (Note that the PPV and seismic velocity need to be in the same units.) If we assume the seismic compressional wave velocity of concrete is about 3000 m/sec, a strain criterion of 600 microstrains yields a peak particle velocity of about 175 cm/sec. Meetings with PSE&G and the design and construction team were held prior to start-up of blasting. PSE&G stated concerns about blast induced damage to the duct banks due to the high PPV criterion. They wanted a more conservative value, 5 cm/sec. The lower criterion is in the New Jersey Administration Code and is applicable and appropriate for protection of residential structures and other above ground buildings. With the lower criterion, blasting would not have been feasible. After explaining the source of the 5 cm/sec criterion to PSE&G, and listening to their concerns, we recommended an intermediate criterion of 50 cm/sec as a safe maximum for vibration levels in the duct bank. (Due to advances in blast monitoring equipment since the original contract specification was prepared in 1998, use of a PPV to monitor the duct banks was possible.) Charge weights could then be established, using the Oriard high-confinement equation (ISEE, 1998): (2) where Ds  scaled distance (D/W0.5) Blast design requirements included the following: 1. Smoothwall blasting techniques were specified to minimize overbreak and damage to the final rock surface, 2. Loading density was restricted for perimeter (smoothwall) holes; 0 to 0.6 kg/m, and buffer holes were allowed up to 1.0 kg/m, 3. Burden to spacing ratio of 1.3 to 1.5, and 4. Line drilling along the perimeter of the elevator pit (hole spacing equal to two to four times the hole diameter). We judged that vibration was in fact not the most likely factor that could adversely affect the duct banks. Mass movement due to possible back-break from the confined elevator shaft would potentially have a greater effect on movement of the duct bank. Therefore, we required the line drilling, smoothwall holes, and a blast sequence that provided adequate relief along the walls closest to the duct bank. The contractor developed and submitted a blast plan, which was reviewed by PBQD and PBQD’s blasting consultant, Mr. Gordon Revey. After several reviews and coordination meetings, a plan was accepted and blasting commenced.

1

Figure 3. Plan view of blast monitoring instruments.

The final aspect of the contract requirements included an instrumentation program to monitor the duct banks during blasting. Seismographs were used to monitor PPV and strain gage arrays were used to monitor the strain in the duct bank. The layout of the instruments is presented in Figure 3. An array of 6 seismographs was used for vibration monitoring. Instantel BlastMate II and BlastMate III seismographs were used for monitoring, with the BlastMate III equipment set up opposite of the center of the blast (one on each duct bank). The BlastMate III seismographs were equipped with high frequency geophones, which could record vibration levels up to 250 cm/sec, with a frequency range of 28 Hz to 1 kHz. Strain gage arrays consisted of two sets of two fiber optic strain gages mounted on top of the duct banks. Each set of strain gages were aligned perpendicular and parallel with the tunnel alignment. 4 BLASTING To reassure PSE&G that blasting near the duct banks was not going to impact service, two small test blasts were detonated and vibrations were monitored. These two shots were placed in the tunnel invert near the center of the tunnel about 3.3 m from the duct banks. The first test shot consisted of a 1 m deep hole with a 0.5 kg charge (Austin Powder Co. Ex Gel 40%) in the hole and 0.6 m of 6 mm stone stemming. The shot was matted and initiated non-electrically. The second shot was a 0.9 kg charge in a 1 m deep hole, with 0.45 m of stemming. Both shots were monitored with the seismographs and strain gages mounted on the duct banks. The results were favorable and are discussed in the section below. Following the small test blasts, the next shots were for excavation of the elevator pit. Figure 4 shows the planned sequence of shots for the elevator pit blasting. Shots 1, 2, and 3 were performed as planned; however, shots 4 and 5 were charged and shot simultaneously.

131 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-16.qxd

2/18/04

11:58 AM

Page 132

Figure 5 presents the loading of the holes for the initial elevator pit shot. This shot included 15 loaded holes and a circular pattern of 6 burn holes to provide relief for the shot. The holes were approximately 2.4 m deep of which 0.6 m was subdrilling. As shown in Figure 5, the holes closest to the duct banks were loaded as buffer holes, with 1 kg of explosive. The production holes were loaded with 2.3 kg of explosives. All loaded holes were stemmed about 1 m with 6 mm stone. Each hole was separately delayed with delay periods of

5 3

1

2

4

Figure 4. Plan view of layout of elevator pit shots.

25 milliseconds to reduce the total kilograms per delay, which subsequently minimized vibration levels. The burden to spacing ratio was approximately 1.1. Due to the confined nature of the first shot, it was anticipated to be the most critical shot and would likely produce the highest vibration levels on the duct banks. To protect the duct banks from impact by flyrock generated by blasting, the contractor covered the shot with blast mats. Shot no. 2 in the elevator pit consisted of 23 loaded holes. The outer most holes were lightly loaded smoothwall holes intended to break the rock along the excavation limits and were delayed to fire last. Production holes were fired first in the sequence. The holes were separately delayed and sequenced in an alternating pattern. Delay periods were a minimum of 8 milliseconds apart to avoid a cumulative effect on vibrations. Shot no. 2 had the benefit of relief as it was delayed to move rock into the area of shot no. 1. Similarly, shot no. 3 had the same advantage, which significantly reduced vibration levels in the duct banks. Shot nos. 4 and 5 were combined into one shot. The distance from the duct bank to the loaded holes ranged from 0.83 to 1.19 m. The limits of the elevator pit were line drilled using 30 cm spacing. Shot nos. 4 and 5 also had relief to the center of the pit, which would reduce vibration levels. 5 RESULTS Actual peak particle and strain levels measured in the field were lower than the anticipated or predicted levels. Table 1 presents a summary of the blasting results. After the first shot at the elevator pit, the blast mats hit and damaged the strain gages. Later shots were not monitored with strain gages, as they could not be replaced in time due to the blasting schedule. It was decided that the next blasts were not as critical as the initial blast and that monitoring using only the seismographs was sufficient. The first shot at the elevator pit produced the highest vibration levels as expected, due to the confined nature of the shot. However, they were much lower

50

Figure 5.

Shot no. 1 – hole layout and delay sequence. Table 1. Summary of blast monitoring readings.

Shot no.

Distance to duct bank (m)

Kilograms of explosive

Scaled distance

Predicted max. PPV (cm/sec)

Measured PPV (cm/sec)

Measured strain (microstrain)

Test shot 1 Test shot 2 1 (Buffer hole) 1 (Production hole) 2 (Production hole) 4/5 (Buffer hole) 4/5 (Smoothwall hole)

3.35 3.35 2.01 2.77 2.01 1.25 0.83

0.5 0.9 1.0 2.3 2.3 1.0 0.7

4.7 3.5 2.0 1.8 1.3 1.2 1.0

38 60 147 171 284 312 449

3.4 2.8 25 25 6.4 5.2 5.2

1.15 0.98 147 147 NA NA NA

NA  Not available.

132 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-16.qxd

2/18/04

11:58 AM

Page 133

than anticipated. Subsequent shots rendered low vibration levels even when blasting within about 1 m from the duct banks. The favorable results are due to: 1. Relief provided by previous shots, 2. Using one delay per hole, 3. Line drilling along the perimeter of the excavation, and 4. The duct bank was not directly founded on the rock in the vicinity of the blasting. The predicted vibration levels were conservatively based on the high confinement equation by Oriard due to the sensitivity of the duct banks. However, blasts were less confined due to the layout and sequencing of the shots and the vibration levels are thus lower than predicted by this equation. Even with the standard equation, vibration levels typically vary by about an order of magnitude at a given scaled distance. 6 CONCLUSIONS The criterion for strain/vibration levels for typical residential type structures cannot be applied blindly in all circumstances. Monitoring devices that could get damaged during the course of the work should be provided with standby replacements on site. Reasonable yet prudent, substantiated strain and vibration limits can be used effectively to protect sensitive structures from close-in blasting. Careful planning and communication between all parties allowed the work to move forward with satisfactory results to all parties. NJ TRANSIT will get the station constructed, PSE&G’s duct banks were not impacted, and the contractor completed the work in a timely fashion.

ACKNOWLEDGEMENTS The authors would like to thank NJ TRANSIT for allowing the publication of this paper. In addition, the authors would like to thank Jose Morales and Michael Babin for their assistance with the preparation of this paper.

REFERENCES Dowding, C., 1996, Construction Vibrations, Prentice-Hall, Inc., Upper Saddle River, NJ. ISEE, 1998, Blaster’s Handbook, 17th Edition, International Society of Explosives Engineers, Cleveland, Ohio. Parsons Brinckerhoff Quade and Douglas, Inc., January 2001, Geotechnical Report, Design Unit N-30, Weehawken, Union City & North Bergen, NJ, Hudson-Bergen Light Rail Transit System. Parsons Brinckerhoff Quade and Douglas, Inc., January 2001, Geotechnical Design Summary Report, Design Unit N-30, Weehawken Tunnel and Bergenline Avenue Station, Hudson-Bergen Light Rail Transit System. Parsons Brinckerhoff Quade and Douglas, Inc., December 2001, Engineering Specifications (Conformed), Design Unit N-30, Weehawken Tunnel and Bergenline Avenue Station, Hudson-Bergen Light Rail Transit System. Revey, G.F. and Scott, G.A., 1999, Blasting Tunnel through Folsom Dam, ISEE Conference, Nashville, TN. Tart, R.J., Oriard, L.L. and Plump, J.H., 1980, Blast Damage Criteria for a Massive Concrete Structure, in Minimizing Detrimental Construction Vibrations, ASCE, NY. Westine, P.S., Esparza, E.D., and Wenzel, A.B., 1978, Analysis and Testing of Pipe Response to Buried Explosive Detonation, Report L51378, American Gas Association, Arlington, VA.

133 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-17.qxd

2/18/04

11:58 AM

Page 135

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Subway rehabilitation – secant wall cofferdams and penetration of tunnel liner Vincent Tirolo, Jr. & Norman Hirsch Slattery Skanska Inc

ABSTRACT: In order to improve passenger access a new escalator, elevator and reconstructed mezzanine were constructed at New York City Transit’s (NYCT) Lexington Avenue/Third Avenue IND Station. The station is located on East 53rd Street between Lexington Avenue and Third Avenues in Manhattan. The station is a hub that not only provides subway access to the east side of Manhattan but also is a transfer point to other subway routes. The subway platform is about 22 meters below the street grade. The platform was constructed between two mined tunnel bores. The tunnels and station were constructed in 1928. The existing station is in an area with a three tube concrete liner. During construction, the street was decked from building line to building line using a longitudinal steel stringer decking system supporting both the precast concrete roadway and sidewalks. After excavating to the new mezzanine level 6 m below street grade, two separate secant pile cofferdams were constructed. The secant piles were drilled to the top of rock and/or the top of the existing tunnel concrete liner. The secant wall was constructed using Bauer BG-18 and BG-22 drill rigs. The secant wall cofferdams provided a watertight excavation and their rigidity minimized ground movements during excavation. After the secant wall cofferdams were completed, the process of penetrating the tunnel arch began. First temporary struts and columns were erected and loads transferred by hydraulic jacks from the existing tunnel arch to these structural elements. This allowed the tunnel arch to be penetrated for the new escalator and elevator. Demolition of the arch was accomplished without blasting using pneumatic and hydraulic tools and splitters. Deformations were monitored during jacking and penetrations of the roof arches at both the escalator and elevator using a Bassett Convergence System. Maximum deflections were less than 7 mm and averaged about 4 mm. The new escalator and elevator work was performed with only minor disruptions to station operations.

1 INTRODUCTION The purpose of this project was to connect the two existing mezzanines and to provide additional elevator and escalator service to the existing platform. New York City Transit’s (NYCT) Lexington Avenue/Third Avenue IND Station is located on East 53rd Street between Lexington Avenue and Third Avenues in Manhattan. Construction was performed by a joint venture of Slattery Skanska Inc and Gottlieb Skanska Inc (the JV). The station consists of two separate mezzanine structures, one near Lexington Avenue and the other near Third Avenue. The existing mezzanines are not connected. They are shallow structures that were constructed by cut-and-cover construction methods. The roof is about 1.5 meters below street grade. The subway platform is about 22 meters below the street grade. The platform was constructed between two tunnel bores mined through rock and mixed face. The tunnels were extensions from the East 53rd Street Tunnel that was

mined under the East River east of the station. The tunnels and station were constructed in the late 1920s. The station is a hub that not only provides subway access to the eastside of Manhattan but also is a transfer point to other subway routes. 2 SITE CONDITIONS 2.1

The ground at the site can be divided into five distinct strata. The upper stratum is a man made miscellaneous Fill consisting of fine to coarse sand, gravel, cobbles, boulders, brick, concrete, cinders and silt. The Fill depth varies from 2 to 12 m. Underlying the Fill is a loose to very compact fine Sand. The thickness of Sand varies from one foot at the east end of the site to 9 m at the west end. Below the sand is a 1 to 4.5 m stratum of Glacial Till. The Till consists of a matrix of very compact fine to coarse sand enclosing coarser materials

135 Copyright © 2004 Taylor & Francis Group plc, London, UK

Geotechnical

09096-17.qxd

2/18/04

11:58 AM

Page 136

Figure 1. Existing station tunnel liner.

including rock fragments, cobbles and boulders. Underlying the Till is a discontinuous stratum of Decomposed Rock. Below the decomposed bedrock is Manhattan Schist and Granite Gneiss. RQD values ranged from zero to 82 percent. Some of the cores included cement grout that had penetrated the rock during tunnel construction. Groundwater is approximately 11 meters below existing street grade. The presence of boulders in both the Fill stratum and immediately above bedrock had a major impact on the progress of the work. Another major influence was the undulating top of rock surface. During construction steel sets, rock bolts timbers and concrete walls were uncovered. 2.2

Buildings

Major commercial and hospital structures are on the perimeter of the project. The major buildings surrounding the site include the 59 story headquarters of CitiGroup at the northwest side of the site; 599 Lexington Avenue, a 50 story building at the southwest side of the site, and the 11 story Memorial SloanKettering Cancer Center out-patient facility at the southeast end of the site. Other buildings on East 53rd Street vary in height from 8 to 11 stories. Fortunately for construction, all of these buildings are supported by deep foundations to rock. However, underpinning piers were installed at the 8 story Meyer’s Garage and the 11 story Memorial Sloan-Kettering Cancer Center to protect those buildings’ basement slabs. 2.3

Utilities

All underground projects constructed within the streets of New York City encounter a myriad of utilities. This project was no exception. The site includes major sewer and water lines, including their service connections; and electrical, power and telecommunications duct banks. The sewer and water lines were temporarily

relocated during construction. The new mezzanine roof precluded restoring the sewer lines within the roadway. Therefore the sewers were permanently relocated under the sidewalks. The other utilities were hung below the temporary roadway decking system. The supported utilities had to be moved during secant wall construction if they interfered with the drilling operations. 2.4

3 CONSTRUCTION APPROACH A Contractor’s primary concerns are safety and schedule. A safe project completed on schedule is generally a profitable project. Therefore, when we take an initial look at means and methods we focus on schedule and safety. An example of one of the schedule issues on the Lexington Avenue Station Rehabilitation Project, was the support of the roadway and sidewalk decking system. The suggested support system involved the installation of approximately twenty-seven 54.3 Mg pipe piles. Pile load testing would be required to confirm a minimum factor of safety of 2. The preparation of shop drawings and the fabrication of steel for a decking system are a time consuming process that could not await the results of a pile-testing program. If for any reason the pile tests indicated 108.6 Mg piles were not successful, it would be impossible to modify the decking installation without a serious delay in the project

136 Copyright © 2004 Taylor & Francis Group plc, London, UK

Existing subway structure

As shown in Figure 1, the existing subway structure is a triple barrel unreinforced concrete arch. The 4.2 m wide center arch contains two longitudinal built up steel girders supported by steel columns placed 4.6 m center to center. The arch is designed for full soil and rock overburden pressure. The soil and rock is considered fully drained.

09096-17.qxd

2/18/04

11:58 AM

Page 137

Figure 2. Project cross-section.

schedule. Also, pile penetrations of the new mezzanine roof would result in formwork and waterproofing problems. Finally, there was also the unknown impact of the piles on the rock block geometry above the tunnel roof. Another issue that impacts both safety and schedule was the uncertainties associated with the variable rock surface and rock quality. The secant wall must be seated on rock to provide a groundwater cut-off. In additional to analyzing the data contained in the MRCE’s geotechnical report, the JV took 10 additional borings to further define the rock surface. These borings indicated that the top of rock surface dropped off significantly from east to west. In fact the top of rock surface dropped below the existing tunnel roof at the west end of the escalator secant wall SOE. Based on our previous experience, investing our own money to obtain additional geotechnical information almost always reduces risk and improves the performance of our temporary structures. The general layout of the revised system is shown in Figure 2. Blasting was not permitted on the project. Rock was split with jackhammers and hydraulic rock splitters. 4 DECKING SYSTEM Construction of the new mezzanine structure and the relocation of utilities required that the decking system

extend for the full width of the roadway (three 3.3 m wide lanes) and both 4 m sidewalks (building line to building line). Since the roadway must remain open during construction, the decking is installed in stages. The decking system should also minimize requirements for utility relocation. For these reasons and to minimize the potential adverse impacts on the existing tunnel and mezzanine construction discussed in the previous section, the JV elected to install a raised decking system, utilizing a longitudinal beam and header system, which would be supported on shallow spread footings. Precast decking panels are installed between the flanges of the longitudinal decking beams. Top of the W760 decking beams were set at the crown of the existing roadway. The longitudinal decking beams, coped at their ends, were supported by six transverse W360 header beams. Thus the total depth of the decking system is limited to about 760 mm. The shallower the penetration of the decking system into the existing roadway, the less the possibility of intersecting existing utilities. Supporting utilities in-place is always preferred to utility relocation which is a time consuming “schedule breaking” activity. To maintain the project schedule, an individual box sheeted pit was constructed for each footing. The footings were then poured prior to the arrival on site of the decking steel. Prior to placing the footing concrete, the subgrade was inspected to verify its New York City Building Code (NYCBC)

137 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-17.qxd

2/18/04

11:58 AM

Page 138

classification. In general the soils at subgrade were classified as compact fine to coarse sands, NYCBC 7–65 or 8–65 soils. After the footings were poured, the decking columns were installed within the pits. When designing a decking system it is important to consider construction access. The header beams were then installed at night during a complete roadway shutdown. There should be enough room between the decking support columns for equipment to operate efficiently. The JV developed a monitoring system to check the performance of the decking and footings during construction. A jacking system was also designed and installed to allow the header beams to be jacked up if footing settlements exceeded 25 mm. During the course of the project, the jacking system was implemented on two occasions. On both these occasions footing settlement was related to loss of ground that occurred during the drilling of an adjacent secant pile. 5 SUPPORT OF EXCAVATION – SECANT WALL Two secant pile cofferdams were required for the Lexington Avenue Station Rehabilitation Project. The secant wall cofferdam for the new elevator was 5.6 m square and consisted of 32 secant piles. The new escalator cofferdam, was rectangular, 28.8 m by 7.4 m, and consisted of 100 secant piles. The secant walls are a combination of all cast-in-place concrete primary piles, which are installed first, and secondary piles that are drilled between primary piles. The secondary piles were first drilled to the top of rock and a steel wide flange “core” beam was then placed into the pile before concreting. The temporary casing was withdrawn in 1.5 m sections as the concrete was placed. A total of sixty-six core beams were installed. The overlap between two adjacent drilled shafts is 152 mm. The secant piles were installed from the level of the new mezzanine structure, about 6 m below the top of the decking system, to the top of rock. 100 mm OD steel pipes were attached to the core beams to permit postgrouting of the secant pile tip/top of rock interface. The core beams in the secant walls were W310X158 A572M Grade 345 MPa members. The SOE system consists of the secant walls and three levels of wales and struts. The JV designed the details of the secant wall. Underpinning and Foundation Inc. (U&F) constructed the secant wall under a subcontract to the JV. The secant wall was constructed using crawler mounted hydraulic rotary Bauer BG-22 and BG-18 drill rigs. The mast of the Bauer BG-18 drill rig was modified for low headroom installations. The Bauer drill rig is equipped with a casing drive adapter (turntable) which rotates and advances a watertight 750 mm diameter double wall casing. The casing

Figure 3. BG-22 drill rig on roadway decking.

is advanced by attaching new 1500 mm long casing segments. Conical bolts located around the circumference of the casing connect the segments. The lead casing is equipped with a cutting shoe. Hardened metal inserts attached to the cutting shoes allow for rotary drilling through soils, soft rock and secant pile concrete. Soil inside the casing was removed by conventional earth drilling tools. In dry ground, soil was removed with an auger. In loose wet soil, a bucket auger was used. This auger has a lockable revolving bottom gate. Rock augers were used whenever boulders or the top of decomposed rock were encountered. These excavation tools are designed to work within the 750 mm double wall casing. The installation sequence for the secant wall cofferdams on Lexington Avenue Station Rehabilitation Project included:

•

•

•

138 Copyright © 2004 Taylor & Francis Group plc, London, UK

Installation of decking. The decking system was designed to not only support normal roadway traffic but also the Bauer BG-22 rig that would be used for drilling secant piles from the decking level (see Figure 3). Where the location of a secant pile interfered with the decking steel, the Bauer BG-18 rig installed these secant piles from beneath the decking. Since the BG-22 is a more powerful machine than the BG-18, we attempted to minimize the number of secant pile installations from below the decking. It was also necessary to design the decking system so that the BG-18 could be lowered down through the decking. Excavate down to the new mezzanine level. Layout controls, benchmarks, and install guide wall template. Proper installation of the guide template is critical to the success of the wall. It is also an expensive item because of the precise formwork required. Multiple use of this formwork is important to reduce costs. Remove or relocate all overhead obstructions interfering with normal movement of BG-18. These “obstructions” often were utilities that had been previously hung from the decking system. These

09096-17.qxd

2/18/04

11:58 AM

Page 139

Figure 4. Actual vs. theoretical rock line.

• • • • •

often had to be moved again to accommodate movements of the BG-18 and sometimes the BG-22 drill rig. Install the primary piles first. As the casing is advanced excavate within the casing using an auger or bucket, as soil conditions require. Confirm casing is down to competent rock before placing concrete. Concrete was placed by the tremie method or by chutes. Install the secondary piles; install core beams and place concrete Grout rock secant wall interface as required.

The average progress was one secant pile installed per drill rig per 8 to 10 hours shift. Two major impediments to construction were the presence of obstructions, e.g. timbers, concrete wall, boulders, and the irregular rock surface. Figure 4 illustrates the top of rock profile based on the boring shown in the Geotechnical Report and the additional borings taken by the JV versus the actual top of rock from the secant wall as-built data. The two primary purposes of the secant wall structure were first to act as a “bath tub” structure to allow excavation within its perimeter without external dewatering and second to provide a “rigid” cofferdam to limit settlement of adjacent structures. The water tightness of the wall is primarily a function of the secant wall construction. The functioning of the secant wall as a SOE is primarily dependent on the details of design. A secant wall design uses the same methodology as the design of a soldier pile and lagging wall. The

spacing between the core wide flange beams in the secondary secant piles is approximately 1219 mm. The thickness of the secant wall at its narrowest dimension is about 381 mm. In our design we did not assume any composite action between the core beams and the concrete. The concrete can either be designed as a concrete lagging, e.g. simple span for bending and shear, or as an inscribed arch. The inscribed arch is similar to the NYCT jack arch concept. In either case no reinforcement was necessary. The secant wall is then designed for both down stage and up stage conditions. For the Lexington Avenue Station Rehabilitation Project, we used three levels of wales and struts. The upper level was located about 1.8 m from the top of the secant wall. The span between the upper and second level of struts was 3 m. The third level was 1.8 m below the second. For the final down stage condition, we assumed the core beams were partially socketed into rock. The maximum depth of excavation below the mezzanine level was 10 m. The total deflection of the wall was minimal.

6 INSTRUMENTATION AND MONITORING After the secant walls were completed the next stage in the work was the penetration of the existing tunnel for the new elevator and the new escalator. Prior to this work, monitoring instrumentation was placed at the arch penetration locations. The instrumentation consisted of a Bassett Convergence System (BCS) to monitoring deformation of the concrete tunnel and

139 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-17.qxd

2/18/04

11:58 AM

Page 140

Figure 7. Uncovered steel sets at tunnel roof penetration.

Figure 5. Typical array deformation.

through rock was by drill and blast. The heading was normally advanced using heading and bench methods. A typical round was 2.5 to 3 m long. Progress was typically 3 to 4.5 m for three 8-hour shifts. Mined tunnels entirely in rock were normally lined with unreinforced concrete. Occasionally, in subaqueous sections mined in poor rock, the permanent tunnel liner was a combination of a primary cast iron liner and a secondary concrete liner. The permanent tunnel liner at the Lexington-Third Avenue Station is a triple barrel unreinforced arch with longitudinal built-up girders along the platform. 7.2 Figure 6. Array trend plot.

strain gages to monitor loads on the temporary struts during penetration of the middle tunnel arch. The BCS reference pins were arranged in six arrays, three at the elevator and three at the escalator. Figures 5 and 6 respectively illustrate typical array deformation data and trend charts. The maximum displacement at the springline was 7 mm and 6 mm in the crown. Average deformations were considerably lower (3–4 mm). It was difficult to protect the reference pin arrays during construction. In general jacking loads into the temporary columns and struts caused the major movements. After jacking and lock-off, movements during demolition of the arch were negligible. 7 EXISTING STATION AND TUNNEL 7.1

Tunnel construction in the 1920s

The original Lexington-Third Avenue Station was constructed as part of New York City Transit’s Route 104 Section 2 in the late 1920s. The contractor was Patrick McGovern Inc. Mined tunneling in the 1920s

As shown in Figure 4, the top of rock at the east end of the secant wall cofferdam is high while at the west end the top of rock drops below the crown of the tunnel. Therefore the easterly portion of the LexingtonThird Avenue Station was constructed as a rock tunnel, the middle portion in mixed face and the west end of the station was constructed as an open cut. The open cut was probably also used as a construction shaft for the tunnel. Figure 7 shows the temporary steel sets and timbers uncovered when the top of the subway tunnel was exposed. No steel sets were encountered at the tunnel penetration for the elevator (east end of secant wall cofferdam). Temporary 360 mm deep steel “cap” beams, typically less than 1 m center to center, were supported on the permanent built up girders, and were exposed along the entire length of the escalator penetration (west end of secant wall cofferdam). We have concluded that the mixed face portion of the station was constructed using multiple drift. We hypothesize that two drifts were excavated over EB and WB track sections using steel caps and posts. The drifts were advanced by forepoling. The invert and the longitudinal built up girders were erected in these drifts. The roof load was transferred from the interior steel posts to the build up girders and the posts

140 Copyright © 2004 Taylor & Francis Group plc, London, UK

Mixed face tunneling

09096-17.qxd

2/18/04

11:58 AM

Page 141

removed. The center drift was then constructed and the platform constructed. This methodology is similar to the illustrations in Szechy (1973) for station construction in Europe. West of the secant cofferdam the rock surface rises rapidly and the contractor was able to resume free air rock tunnel. Therefore, it is likely that the open cut section of the station was used as a work shaft. Figure 1 is a cross-section of the station’s three arch tunnel liner. 8 PENETRATION OF TUNNEL ROOF The major concern in penetrating the roof arch was the redirecting of the theoretical thrust in the roof arch around the opening. The system used to accomplish this was the sequential placement of temporary struts and strut jacking. The procedure involved first saw cutting and chipping pockets on either side of the middle (platform) arch. These pockets were approximately 1 m center to center. Immediately after a pocket was cut a strut was installed as shown in Figure 8. The struts were W310X179 members. After these initial struts were installed, the concrete between the pockets was removed and the remaining struts installed (see Figure 9). Two of these “temporary struts” became permanent. At the escalator penetration, 13 temporary struts were initially installed. Five struts became permanent. Because of the longer length of the escalator penetration compared to the elevator (10.5 m vs. 3.5 m) a secondary level of 127 mm diameter 9.5 mm wall pipe struts spaced 2.4 mm center to center were installed above the primary wide flange struts. There was also a concern that the longitudinal platform girders would carry additional vertical load when the platform arch was removed. These girders were supported by columns 4.572 m center to center. However during construction temporary columns were installed 1.524 m between the existing columns. The loads from the original arch and columns were transferred to the temporary columns and struts by jacking. The maximum vertical load in the temporary columns was limited to 900 kN and 670 kN in the temporary struts. The jack force in the pipe struts was limited to 290 kN. Figure 10 shows the penetration of a portion of the tunnel arch with all temporary struts installed. Fabricated double channels were used to permanently support the opening after the temporary struts were removed. These channels were 1 m deep with 100 mm flange plates and a 150 mm web plate. At the escalator they were over 8 m long. Placement of these members was difficult with an operating station. The channels were either lowered through the decking system or carried into the site with work trains. Figure 11 shown typical channel sections prior to lowering them through the decking.

Figure 8. Arch pockets and temporary columns.

Figure 9. Sockets expanded for intermediate struts.

Figure 10. Arch penetration.

9 CONCLUSIONS The escalator and elevator shafts successfully penetrated the roof without incident and with minimal

141 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-17.qxd

2/18/04

11:58 AM

Page 142

A raised roadway decking coupled with longitudinal decking stringers minimized interferences with existing utilities. ACKNOWLEDGEMENTS

Figure 11. Channel sections.

disruption to the public. Monitoring indicated small movements (generally less than 4 mm). The use of temporary struts and columns proved to be an efficient means of maintaining the existing middle concrete arch. The secant wall was successfully placed, within an urban environment, under roadway decking. The bottom of the secant wall was exposed during penetration of the tunnel liner to install the elevator shaft and half of the escalator. Except for two locations, all secant piles were founded on competent rock (RQD 50%). At the two locations not founded on competent rock, the secant piles were located on large boulders immediately above bedrock. At these locations, chemical grouting was used to seal the area.

We would like to acknowledge the assistance of those organizations without whose help this work could not have been accomplished. New York City Transit first, for giving us the opportunity to work on this project and later in assisting us in all our efforts. Next we would like to acknowledge Daniel Frankfurt, P.C. the prime consultant on the project and their geotechnical subconsultant, Mueser Rutledge Consulting Engineers, and Dr. Carl Costantino, the NYCT’s Geotechnical Consultant, for their fairness and open-mindedness. We wish to acknowledge our secant pile subcontractor Underpinning and Foundation Inc. Finally we would like to acknowledge the members of the JV team, both in the field and the engineering department, that have made this project both an engineering and financial success. REFERENCES Goodman, R. & Shi, G-H. 1985. Block Theory and its Application to Rock Engineering, Prentice-Hall. Mueser Rutledge Consulting Engineers, 1999. Geotechnical Report, Lexington Avenue Station Rehabilitation, New York. Szechy, Karoly. 1973. The Art of Tunnelling, Akademiai, Budapest.

142 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-18.qxd

2/18/04

11:59 AM

Page 143

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Overcoming the complex geotechnical challenges of urban construction T.J. Tuozzolo Moretrench Geotec, a Division of Moretrench American Corporation, Rockaway, New Jersey, USA

ABSTRACT: New construction within the confines of heavily developed urban settings can present designers and geotechnical contractors with complex challenges, particularly when the construction is to be developed on a constricted site. Such was the case for a Dental School addition to the University of Medicine and Dentistry in Newark, New Jersey. Earth retention and underpinning design for a site bounded by busy main thoroughfares and adjacent buildings was further complicated by the presence of adjacent underground utilities, low-headroom conditions and unanticipated soil conditions. The geotechnical contractor’s ultimate design/build solution, which required a proactive, on-site response, involved four distinct techniques – soil nailing, minipiles, conventional pit underpinning, and a drilled-in-place pipe soldier pile cofferdam – all implemented under a single contract. Extensive monitoring and testing was conducted to verify design assumptions and to provide data for future research.

1 INTRODUCTION When new construction is planned in already heavily developed urban areas, several critical factors need to be considered prior to commencement. One of the most critical factors is how the excavation for the proposed structure will be accomplished. Since most urban environments preclude open excavations, much thought must be given to the earth support system required to retain the subsurface materials during excavation for the new foundation. If the excavation will be made next to adjacent structures, considerations for underpinning these structures must also be addressed, as must the impact active utilities will have during excavation, and the limitation of the building and property boundaries. Such complexities are inherent in many of today’s urban construction projects. When project requirements, site conditions and restrictions render a conventional approach to earth retention and underpinning impractical, if not impossible, specialty geotechnical techniques can offer a viable and economical alternative. 2 CASE STUDY Expansion of the Dental School at the University of Medicine & Dentistry of New Jersey (UMDNJ) in Newark, NJ, included construction of a five-story, steel structure with a one-story, below-grade basement.

The site was bounded on one side by the existing Dental School and on the opposite side by 12th Avenue, a major access road for emergency vehicles en route to the adjacent University Hospital. Parallel, underground, electric and telephone duct banks, each 1.5 m by 0.6 m, ran along 12th Avenue. The telephone bank lay nearest the site, 1.5 m below existing grade and within millimeters of the proposed building. The proposed structure, approximately 85 m by 37 m in plan, entailed excavations to a depth of 8.2 m below grade. In order to facilitate these excavations, a temporary earth retention system was required for approximately 81 linear meters along 12th Avenue and for 38 linear meters along the west side of the site. Also, a permanent retention and underpinning system was required for 18.3 linear meters on the south side of the project along the existing 3-story Dental School where the new building would be constructed on-line and tied in. Preliminary excavation retention system concepts included driven steel sheeting and soldier beams and lagging along 12th Avenue and the west side, as well as a permanent, on-line, soil nail wall along the existing Dental School. However, the proximity of the duct banks along 12th Avenue precluded the sheeting and soldier beam options, since both systems would encroach into the new building and would require the structure to be moved and redesigned. In addition, the soil borings encountered rock at or just above the proposed subgrade elevation, which could impede the

143 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-18.qxd

2/18/04

11:59 AM

Page 144

driving of soldier piles or steel sheets and likely necessitate a more expensive, drilled-in soldier pile system. 2.1

Engineered solution

The pre-construction geotechnical investigation indicated that the subsurface conditions consisted of 1.2 m to 1.8 m of fill, underlain by natural silts and silty sands containing varying amounts of gravel, cobbles and boulders. Weathered sandstone bedrock was encountered at or above the subgrade elevation of the proposed structure. After evaluating the soil conditions, space limitations and the schedule, the geotechnical contractor offered a system that entailed a temporary soil nail wall along 12th Avenue and the west side of the site, and a combination of drilled-in minipiles and soil nailing along the existing Dental School. The minipiles would be drilled through the existing column footings to underpin the building and transfer the loads below the proposed subgrade. Soil nailing would be used to support and retain the soil between the existing columns. 2.2

Figure 1. Cross-section showing soil nail wall installation.

12th Avenue and west side

The soil nail wall along 12th Avenue and on the west side of the site entailed three to five levels of soil nails installed at 15 degrees from the horizontal on a 1.5 m grid pattern. A 75 mm thick layer of shotcrete, reinforced with wire mesh, was used to tie the system together and retain the soil between the nails. Construction began with the excavating contractor making an initial, sloped precut to locate the top of the telephone duct bank and remove the existing fill soils which did not exhibit good ‘standup’ time. The precut also allowed the first level of soil nails to be installed at a 15-degree angle below the duct bank, and did not require any redesign or relocation of the proposed structure. Once the first level was installed, and the soil nails and shotcrete had cured, the process was repeated and the remaining lifts were installed until subgrade was reached (Fig. 1). A view of the soil nail wall along 12th Avenue is presented in Figure 2. The wall was designed to withstand the temporary lateral and traffic surcharge loads. The system was also designed to withstand the surcharge loads associated with a large crane scheduled to be placed on the sidewalk directly behind the soil nail wall to aid in construction of the new addition. Since the soil nail wall for this part of the project was installed on-line, the new foundation wall was constructed using a onesided form and poured directly against the soil nail wall. A drainage composite and waterproofing membrane were placed between the soil nail wall and the new concrete wall. This 935 m2 system allowed the proposed foundation to be built ahead of schedule and without the need for any design modifications.

Figure 2. View of soil nail wall along 12th Avenue.

2.3

144 Copyright © 2004 Taylor & Francis Group plc, London, UK

Existing dental school

In order to facilitate the 8.2 m deep excavation along the existing Dental School, a combination of minipile underpinning and soil nail earth retention was designed and offered by the geotechnical contractor. The minipiles were installed through the existing column footings to underpin the footings and transfer the column loads below the newly proposed subgrade and thus eliminate any surcharge load induced on the new soil nail wall. The next step was to build the soil nail wall in lifts, finishing at the new building subgrade elevation, 8.2 m below grade. During construction of the permanent soil nail wall, unanticipated conditions were encountered during excavation of the first 1.5 m lift. A loose, cohesionless fill with poor standup time was encountered directly below the existing first-floor slab. The soil was so loose and uncompacted that voids in excess of 25.4 mm were present directly below the existing slab. To assess the actual conditions, the geotechnical

09096-18.qxd

2/18/04

11:59 AM

Page 145

Figure 3. New concrete underpinning pier tied into existing pedestal.

contractor installed a trial section of shotcrete and test nails to determine if soil nailing would still be viable. The shotcrete did not adhere to the loose fill, causing sloughing of the material and “cave-in” of the test section. Additionally, the test nails did not achieve the bond values that the design assumed, since the shear strength of the soil proved to be very low. Taking all of these considerations into account, the geotechnical contractor recommended the soil nailing system should not be installed, and offered a redesigned system. 2.3.1 Redesign at existing dental school Since the soil nail system could not be used as the permanent, earth-retention system, other methods were examined. The new system would need to:

• • •

be able to retain loose soils up to 8.2 m below the existing floor slab, be installed on-line so the new structure could abut and tie into it, and finally, be able to eliminate any lateral loads that would be induced on the existing concrete pedestals during excavation.

After examining many different options, from driven soldier piles and lagging to sheet piling and minipiles, the geotechnical contractor concluded that the only practical system was to construct concrete underpinning piers by the pit method, with treated timber lagging installed to retain the loose soil between the piers. As excavation progressed, two levels of tieback anchors would be installed to resist the permanent lateral loads. Although the minimum depth for the new basement was at 4.8 m below grade, the depth of excavation

had been designed to extend to 8.2 m below grade in order to bear the new foundation on the sandstone bedrock. In conjunction with the redesign for the earth retention system, the geotechnical contractor further proposed using a minipile foundation to support the new column loads rather than the originally proposed spread footing foundations. The minipiles would be installed from an elevation 4.8 m below grade and terminated approximately 6 m into the rock. The minipiles would be tied into pile caps at the 4.8 m elevation, thus eliminating the need to excavate to 8.2 m and significantly reducing the depth of excavation and the associated costs to the owner. 2.3.2 Production work The 0.9 m by 1.5 m underpinning pits were installed on 1.8 m to 2.4 m centers, spacings that directly related to the spacing of the existing concrete pedestals. Once the underpinning pits were formed, reinforcing dowels were installed to tie the existing pedestals to the new underpinning pier and eliminate any lateral load on the pedestals (Fig. 3). The pits were then filled with 20 MPa concrete, forming the new underpinning piers. Following completion of the piers, excavation proceeded in 1.5 m lifts as the treated timber lagging was placed. At approximately 1.5 m below grade, the first level of permanent tieback anchors was installed through the new piers to resist the lateral loads induced on both the piers themselves and the existing concrete pedestals into which they were tied. After the second-tier tiebacks were installed, tested and locked off, excavation and lagging installation continued to the new subgrade elevation, 4.8 m below the existing slab (Fig. 4).

145 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-18.qxd

2/18/04

11:59 AM

Page 146

Figure 4. Redesigned earth support and underpinning.

3 INSTRUMENTATION AND MONITORING Figure 5. Pullout test detail.

In order to evaluate design assumptions and monitor the construction, a comprehensive testing program was implemented, consisting of field pullout tests, proof testing, surveying, and strain monitoring. 3.1

Verification testing

At least one pullout test was performed on each soil nail row (lift) to verify that adhesion values used in the design program were obtained in the field. The pullout test was performed incrementally, generally to a factor of at least two times the design value on a shorter sacrificial nail (Fig. 5). In addition, a minimum of five percent of the production nails were proof tested to 1.33 times the design load to measure the bond and creep values. 3.2

Figure 6. Strain gage instrumentation detail.

Survey and monitoring

Survey monitoring was implemented to measure horizontal and vertical displacements. The deflection of the wall was generally 0.2 percent of the cut height, which is within the range of the anticipated deformation for wall construction in these soils (Byrne et al., 1996). 3.3

Strain monitoring

In addition to the normal monitoring, a strainmonitoring program was developed and implemented for the temporary soil nail wall to compare the in situ parameters of a multi-tier soil nail earth retention system with that of the original design parameters, information that would be valuable in the future. Two pairs of vibrating, wire-strain gages were installed on each of three, predetermined soil nails in a vertical plane in order to determine the average strain (Fig. 6). The nails selected for instrumentation were located on tiers 2 through 4 of the 5-tier system.

Strain readings were taken on a weekly basis for a period of two months from initial installation. Readings were taken more frequently immediately following nail installation. Readings were also taken prior to and after each new 1.5 m lift. Long-term monitoring was not feasible. However, data trends shown in Figure 7 indicate that in situ field stresses were similar to or less than the theoretical stresses calculated using the CALTRANS Design Program, SNAIL (Caltrans, 1991). As shown, the nail stress increases as the next cut is made to a new lift. Results in the upper nail are closest to the theoretical design stresses, more than likely because it was monitored longer than the lower nails. Of interest also is how the nail stress spiked in the strain gages installed 1.5 m behind the wall immediately after the crane used to set the steel structure was brought on site and erected approximately 2.5 to 3.5 m behind the soil

146 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-18.qxd

2/18/04

11:59 AM

Page 147

Figure 7. Composite in situ stress plot. Table 1. Summary of strain gage stresses.

Strain Gage Number

Bar Stress (MPa)

Average Stress (MPa)

B47-15-U B47-15-L B47-15-U B47-5-L C47-15-U C47-15-L C47-5-U C47-5-L D24-15-U D24-15-L D24-5-U D24-5-L

112.00 131.00 112.00 119.00 55.00 50.00 N/A 97.00 25.00 N/A 48.00 45.00

122.00 122.00 115.00 115.00 52.00 52.00 97 97 25 25 46.00 46.00

Theoretical SNAILZwin Stress Output (MPa) 167 167 130 130 139 139 118 118 109 109 97 97

chilled water system to provide cooling was to be installed. To facilitate this installation, an excavation with a plan dimension of 4 m by 4.9 m was required to a depth of approximately 10 m below existing grade. The difficulties of performing such a confined and deep excavation were compounded by the fact that the work had to be accomplished directly below an existing corridor, limiting the overhead clearance to 3.7 meters. Due to space limitations, the excavation could not be open cut; therefore, a method of earth retention was required. 4.1

nail wall. As previously mentioned, the geotechnical contractor’s design took this loading into account. A summary of the stresses is presented in Table 1. 4 NEW CHILLER WATER AREAWAY During the completion of the earth retention and underpinning work, the owner and construction manager were faced with yet another challenge. In order to service the new dental school addition, a new,

147 Copyright © 2004 Taylor & Francis Group plc, London, UK

Earth retention design and installation

Given this complicated condition, the geotechnical contractor had to provide a method of earth retention that could support the soil during excavation and construction of the confined areaway, and also be installed within the limited-access and overhead-clearance constraints. Working closely with the excavating contractor, the geotechnical contractor developed a plan and a design for the earth retention that entailed the installation of low-headroom, small-diameter minipiles to act as soldier piles, timber lagging, and an internal waler and bracing system. Construction began with the installation of 245 mm diameter minipiles to a depth of approximately 12 m below existing grade. The minipiles were drilled below the proposed subgrade of the chiller areaway and socketed into rock. Since the headroom was limited to

09096-18.qxd

2/18/04

11:59 AM

Page 148

Figure 8. Chiller water areaway earth retention system.

3.7 meters, the minipiles had to be installed in onemeter lengths. Once all the minipiles were installed and filled with grout, excavation began. Since the excavation was performed in such a confined location, most of the soil was removed by hand and lifted out of the areaway by conveyors. As work proceeded, an internal, three-level, ring waler system was installed around the perimeter of the earth retention system and abutted into the existing building. Large plates were bolted into the existing building wall in order to accept the ring walers. Timber lagging was bolted to the face of the drilled minipiles to retain the soil. The completed earth retention system is shown in Figure 8. This innovative approach not only allowed the system to be built, but work was performed within budget and in an accelerated time. 5 CONCLUSIONS New construction within urban settings can present complex challenges. Site conditions that could

potentially impact the execution of the excavation support and underpinning work should therefore be fully evaluated prior to the design stage. Conventional techniques such as soldier beams and lagging or sheetpiling are typical options. However, on sites where these options are not viable, specialty geotechnical techniques offer effective alternatives. Geotechnical contractors can play a vital role in the planning stage of complex, urban construction projects by offering owners, designers and general contractors experience-based input into overcoming design and production challenges.

REFERENCES Byrne et al. 1996. Manual for Design and Construction Monitoring of Soil Nail Walls. FHWA-SA-96-069, Federal Highway Administration, Washington, DC. CALTRANS, 1991. A User’s Manual for the SNAIL Program, Version 2.02. California DOT, Division of Technology, Material & Research.

148 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-19.qxd

2/18/04

12:00 PM

Page 149

Session 2 Subsurface investigations and geotechnical report preparation for design/build projects

Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-19.qxd

2/18/04

12:00 PM

Page 151

Session 2, Track 1 Risk allocation

Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-19.qxd

2/18/04

12:00 PM

Page 153

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Risk management in tunneling – occupational safety  health plans for drill and blast and tunnel boring machines A. Moergeli moergeli  moergeli consulting engineering, CH Schmerikon, Switzerland

ABSTRACT: Control your critical success factors by applying generally accepted best risk management practices. Establish, implement, update and document a thorough and comprehensive Risk Management System where required by law, project challenges, owner’s specifications and/or where best risk management practices are not available.

Examples for requirements:

1 WHAT CAN A RISK MANAGEMENT SYSTEM DO FOR YOU?

•

A Risk Management System (RMS) must increase your productivity – from your customer’s perspective as well – or it is useless … A scientific risk management is a relatively new approach:

• • • •

Many managers (of big companies) are talking about it these days. Most people remain pretty confused when it comes to day-to-day application. Very few people really have personal experience. There is an opportunity in combining ongoing US underground activities with previous project risk management experience, providing owners, engineers and contractors with real added value in successfully pursuing their goals.

2 WHEN SHOULD YOU USE A RISK MANAGEMENT SYSTEM?

3 HOW DOES RISK MANAGEMENT WORK? Control your critical success factors by: 1. Apply generally accepted best risk management practices. 2. Establish a thorough and comprehensive Risk Management System (RMS) where required by law, project challenges, owner’s specifications and/ or the absence of best risk management practices.

4 APPLY GENERALLY ACCEPTED BEST RISK MANAGEMENT PRACTICES

Where required

• • • •

•

by law by project challenges by owner’s specifications and/or by the absence of best risk management practices.

Control your critical success factors by establishing a pro-active Quality Assurance/Quality Management System by applying generally accepted best risk management practices.

You may choose to control your critical success factors by establishing a thorough and comprehensive RMS.

153 Copyright © 2004 Taylor & Francis Group plc, London, UK

US Military Standard 882D requires a RMS of every party within the acquisition process. Swiss Law requires all companies operating in a high risk environment (e.g. contractors in the construction industry) to perform a risk assessment.

1. Use an interdisciplinary team. Tunneling large, sophisticated and very complex projects, in partly still unknown geology and on the background of a highly sensitive environment, has to be well

09096-19.qxd

2/18/04

12:00 PM

Page 154

thought through. A well built-up interdisciplinary team is required to deliver best results. Include the owner, the designer, the construction management, all special sub-consultants etc. as well as representatives of operations and maintenance. Lead the team by an experienced Risk Manager with a substantial background in quality management and safety  health. 2. Leave the ground risk to the owner. 3. Perform a very thorough, very detailed site investigation. Be prepared to invest anything from 1% to a maximum of about 10% (as a general rule) of your Total Cost of Ownership (TCO). Be prepared to core drill at least about 0.6 times the length of your tunnel when trying to avoid dealing with successful contractor’s claims (Waggoner, Daugherty, 1985). 4. Work out a professional Geotechnical Baseline Report (GBR) at least before requesting contractor’s bids. Hand out your GBR to all bidders (Essex, 1997). 5. Plan for contingencies. 6. Establish, run and maintain (before, during and after construction) an adequate control system. Enable real-time on-site interpretation and decision-making. 7. Use a partnering approach. 8. Require Escrow Bid Documents (EBD). 9. Prepare for Alternative Dispute Resolution (ADR). Avoid settling claims before courts as this often proves to be a very expensive process. 10. Emphasize a complete documentation. Keep all available records updated and carefully stored after the project’s completion. 11. Have the contractor to enroll a project-specific Quality Management System (PQM) on site. Fully integrate Safety  Health (OSH) and environmental issues into your PQM. 12. Whenever feasible and practical, test and validate your control of all critical success factors on a small-scale first (e.g. by a pilot tunnel, a test shaft, etc).

5 HOW DOES A RISK MANAGEMENT SYSTEM WORK? A four-folded approach is recommended by the author (m  m/am): 1. Start with a Preliminary Hazard Analysis (PHA). 2. For hazards with a high potential of harm and where generally accepted best risk management practices are not available (risks in Risk-Zone 1), apply a full and thorough Risk Assessment (RA). 3. Use a Fault Tree Analysis (FTA) to verify your preventive action as an option.

Figure 1. Example of a preliminary hazard portfolio.

1 Start Risk Assessment 2 Define system’s limits

3 Identify hazards

4 Estimate risks

5 Assess risks

6 Tolerable residual risks?

7 Documentation

8 Sign list of residual risks

9 Periodic update

No

10 Risk Management

Figure 2. Example of a risk assessment procedure.

4. Establish, implement, deploy, maintain and document System Safety (SS) for risk mitigation. Obviously the recommended risk management procedures do not only apply to the construction phases of a project but – sometimes even more important, at least to the owner – to the overall system life expectancy as well. None of the enlisted procedures are new; some have even more than a fifty-year’s track record. What is really new is the

• • •

154 Copyright © 2004 Taylor & Francis Group plc, London, UK

Yes

combination sequence and extrapolation from safety  health and machinery safety systems to build up a comprehensive RMS.

09096-19.qxd

2/18/04

12:00 PM

Page 155

7 HOW DOES A RISK ASSESSMENT WORK? A risk assessment matrix, adopted to meet your requirements, let you decide which risks you can (or have to) accept. US Military Standard 882D can deliver a starting point.

8 HOW DO YOU MITIGATE YOUR RISKS? 1 Start Risk Assessment 2 Define system’s limits

3 Identify hazards

4 Estimate risks

Yes

5 Assess risk

6 Tolerable residual risks?

No

Yes

7 Documentation

8 Sign list of residual risks

12 New hazards?

9 Periodic update

No

10 Substitution?

Yes

Figure 3. Example of a risk matrix.

11 Susbstitution of hazardous procedures/ materials etc

No

6 HOW DOES A PRELIMINARY HAZARD ANALYSIS WORK?

13 Safe system/ new strategy ?

A Preliminary Hazard Analysis acts as a risk-focused filter to allocate your restricted resources (to perform thorough and scientific risk assessments) to the most urgent and important issues. It regularly follows a three step procedure:

Mitigation iteration

Yes

14 Change of strategy, Safety System (S)

No

No

16 Technical measures?

15 Yes Risk mitigation ok?

Yes

17 Technical measures (T)

18 Yes Risk mitigation ok? No

No

a) Set up a comprehensive hazard inventory (a preliminary, qualitative risk identification) b) Provide evidence of available best risk management practices c) Fill in your identified risks in the appropriate Risk-Zone of your hazard portfolio.

Yes

19 New system’s limits?

No

20 Measures to change human behavior patterns (O/P)

21 Yes Risk mitigation ok? No

Figure 4. Example of a risk management procedure.

You will end up with a fast and easy to understand graphic overview (a risk map). You may then concentrate your efforts to perform a thorough risk assessment of all risks remaining in Risk-Zone 1. But before taking off into that timely and costly task make sure that your interdisciplinary team really confirms your hazard portfolio’s allocations.

9 HOW DO YOU APPLY A RISK MANAGEMENT SYSTEM TO TUNNELING? Any underground construction is considered to be a high risk environment. This is for several good reasons:

•

155 Copyright © 2004 Taylor & Francis Group plc, London, UK

Rock/ground always remains unpredictable

09096-19.qxd

2/18/04

12:00 PM

Page 156

Start

1 (Integrate in) Mission Statement

2 Build up/ optimise organization 3 Training

Hazards with high Yes damage potential?

No

4 Available safety rules?

Yes

No

5.1a Risk Assessment/ Management 5.1b Work out a Safety Concept

Figure 7. Gotthard Base Tunnel: Geology.

5.2 Build up a Safety Program 4 Available safety rules?

No

5.3 Work out safety rules

Yes 5.4 Support company’s superiors

6 Plan+implement measures 7 Emergency planning 8 Ensure employee’s contribution 9 Build up a Health Program 10 Controlling + Audits => CIP*

Emergencies/ absences * CIP = Continuous Improvement Process

Figure 8. Gotthard Base Tunnel: Scheme of tunnel system.

Figure 5. Swiss Safety  Health 10 Points System.

Figure 6. Gotthard Base Tunnel: Construction Program.

Figure 9. Gotthard Base Tunnel: TBM Cross Section.

•

• • • •

• •

Unforeseen water in big quantities can always be a big, crucial factor anywhere anytime Available space is very limited Heavy weight, high energy transport activities

156 Copyright © 2004 Taylor & Francis Group plc, London, UK

Darkness presumes, light is rare High construction noise High temperatures, high moisture Dealing with explosives, high voltage

09096-19.qxd

2/18/04

12:00 PM

Page 157

Figure 13. Current excavation status at Amsteg.

Figure 10. Gotthard Base Tunnel: State of excavation.

Owner Project Location Lot Designer & Client Representative Contractor

AlpTransit Gotthard AG Gotthard Base Tunnel Amsteg, Canton Uri/Switzerland 252, Tunnel Amsteg ING GBTN (Engineering Joint Venture Gotthard North) Joint venture Amsteg, Lot 252, Gotthard Base Tunnel North (AGN) Tunnel length 2 tunnels by ca. 11’350 m Excavation cross section Ca. 71 m2 Construction method (Drill & Blast +) TBM Tunnel construction costs Ca CHF 627 M* Ca. USD 471 M** Construction time frame 02/2002–07/2007 Project status (11/2003) TBM operation just started*** Contractor’s support for OSH Author’s mandate

Figure 14. Control Center at the portal.

* Excl. VAT (Value Added Tax) ** 1 USD (US Dollar) ≈ CHF (Swiss Franc) 1.33 (11/2003) *** For more information please log on to the owner’s website http://www.alptransit.ch and the contractor’s website http://www.agn-amsteg.ch - thank you.

Figure 11. Selected project data Lot 252, Tunnel Amsteg.

Figure 15. Control Center: Continuous monitoring.

•

Fresh air is very limited, etc.

In addition, any underground activity normally brings with it

• •

High public profile, high capital investments Work schedules around the clock.

All these puzzles bring about owners requests for at least some kind of proven risk management procedures. 10 HOW DO YOU MANAGE SAFETY  HEALTH IN TUNNELING?

Figure 12. Gotthard Base Tunnel: Tunnel Boring Machines (TBM).

Swiss Law requires the implementation of a (Occupational) Safety  Health (OSH) 10 Points System.

157 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-19.qxd

2/18/04

12:00 PM

Page 158

Figure 19. Site Safety Officer (SS0) at work. Figure 16. HK-TBM S-229. Owner

Luzern Stans Engelbergbahn LSE Steilrampe Tunnel Engelberg Project Grafenort (near Luzern)/ Location Switzerland North + South Lot Designer & Bucher + Dillier Client Representative Ingenieurunternehmung AG + IG LSE Joint Venture Tunnel Engelberg Contractor (ATE) Ca. 4'040 m Tunnel length Excavation cross section Ca. 30 m2 Construction method Drill & Blast Tunnel construction costs Ca. CHF 72 M* Ca. USD 54 M** Construction time frame 05/2001 – ca 12/2005 Project status (11/2003) Under Construction*** Author’s mandate Contractor’s support for OSH

* Excl. VAT. ** 1 USD ≈ CHF 1.33 (11/2003). *** For more information please log on to the owner’s website http://www.lse-bahn.ch - thank you.

Figure 17. Risk assessment workshop.

Figure 20. Selected project data Engelberg Tunnel.

Figure 18. Excerpt from SSO’s routine inspection report.

11 WHERE HAS A RISK MANAGEMENT SYSTEM IN TUNNELING ALREADY BEEN USED? A random selection of several (larger) tunneling projects in Switzerland (CH), in which the author is currently involved, may illustrate the benefits of an RMS approach for controlling the construction processes and their inherent risks.

Figure 21. Annual OSH System Check at the face.

158 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-19.qxd

2/18/04

12:00 PM

Page 159

Figure 25. AlpTransit Gotthard AG (ATG).

Figure 22. Water ingress at the face.

Figure 26. Arbeitsgemeinschaft AMSTEG, Los 252, Gotthard-Basistunnel Nord (AGN).

Figure 23. Water at the portal.

Figure 27. Murer/Strabag.

Figure 24. Probing ahead from the face.

Figure 28. Herrenknecht.

For further information please feel free to visit the owner’s website http://www.alptransit.ch and the author’s website (the author’s contribution to the AUA NAT02 conference)

• •

159 Copyright © 2004 Taylor & Francis Group plc, London, UK

http://www.moergeli.com/d1doc10e.htm http://www.moergeli.com/d1doc11e.htm Thank you.

09096-19.qxd

2/18/04

12:00 PM

Page 160

11.3

Engelberg Tunnel/Switzerland

Early on, the owner took some strategic decisions with regard to risk management:

• •

Figure 29. rowa.

Systematically probe ahead for gas detection. Require the construction contractor to establish, implement, keep current and document an Integral Safety Plan.

The contractor contributed by establishing and implementing an Integral Safety Plan following suva’s (Swiss National Accident Insurance Fund) suggestions. Annual OSH System Checks keep it updated. So far the contractor dealt successfully with more than his fair share of unexpected ground conditions. 12 CONCLUSIONS

Figure 30. Swietelsky.

Control your critical success factors by: 1. Apply generally accepted best risk management practices. 2. Establish a thorough and comprehensive Risk Management System (RMS) where required by law, project challenges, owner’s specifications and/or where best risk management practices are not available.

Figure 31. Amberg.

11.1

AlpTransit, Gotthard Base Tunnel/Switzerland

Early on, the owner took some strategic decisions with regard to risk management:

• • • • • •

Where possible, cross the alpine mountains through the most favorable geology. Where possible, cross difficult ground/rock perpendicular to strata as short as possible. Build two tunnels. Provide cross passages about every 312 m . Establish, implement, keep current and document a comprehensive owner’s RMS in all phases of the project. Require the consultants, site supervisors and main construction contractors to run their own RMS, keep it updated and document it at least twice a year.

11.2

Gotthard Base Tunnel, Lot 252, Tunnel Amsteg/Switzerland

The contractor contributed by establishing, implementing, periodically updating and documenting a project-specific OSH solution:

A RMS provides a unique opportunity in combining ongoing US underground activities with previous project experience, providing owners, engineers and contractors with real added value in successfully pursuing their goals:

• • • • •

Identify all known hazards graphically on a onepage-sheet early on. Allocate limited resources for Risk Assessments where inevitable (Risk-Zone 1). Mitigate your risks systematically by System Safety. Add value to the owner’s project by minimizing costs, time and third party impacts. Provide evidence of legal compliance.

To perform an RMS, a four-folded approach is recommended by the author (m  m/am): 1. Start with a Preliminary Hazard Analysis. 2. For hazards in Risk-Zone 1, apply a full and thorough Risk ssessment. 3. As an option verify your planned preventive action by a Fault Tree Analysis. 4. Establish, implement, deploy, maintain and document System Safety to mitigate your risks. ACKNOWLEDGEMENTS

• • •

Preliminary Hazard Analysis Risk assessments in team workshops Implementation of System Safety

The author thanks and acknowledges the very competent support and kind provision of plans, schemes and

160 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-19.qxd

2/18/04

12:00 PM

Page 161

pictures by: Last but not least, my thanks goes to Mrs. S. Tschupp for her always very competent support in improving my use of the English language. Without their big help this document would not have been possible. The biggest thanks goes to all crews on site, safely coping with the unforeseeable as their daily routine. Every day they move into places where no human being has ever been before. Always just one small step for a man, but a giant leap for mankind … The author’s apologies go to the readers for any inconvenience dealing with small print, reduced tables and pictures. An update of the paper can be offered through our presentation at the AUA Conference. The original paper will be available for download on http://www.moergeli.com/dldocuebersichte.htm after the AUA North American Tunneling 2004 Conference, April 17–22, Hyatt Regency Hotel Peachtree Plaza, Atlanta, GA-USA.

REFERENCES

AlpTransit Gotthard AG (ATG), Zentralstrasse 5, CH-6003 Luzern/Switzerland (http://www.alptransit.ch). “Geologic Site Investigations for Tunnels”, USNC/TT Study, Eugene B. Waggoner, Charles W. Daugherty, Underground Space, Vol. 9, pp. 109–119, 1985. “Geotechnical Baseline Reports for Underground Construction, Guidelines and Practices”, Randall J. Essex, American Society of Civil Engineers, 1997, ISBN 07844-0249-3. Herrenknecht AG Tunneling Systems, Schlehenweg 2, D77963 Schwanau/Germany (http://www.herrenknecht.de). Joint venture Amsteg (German: Arbeitsgemeinschaft Amsteg), Lot 252, Gotthard Base Tunnel North (AGN), Grund, CH-6474 Amsteg/Switzerland (Murer AG/Strabag AG) (http://www.agn-amsteg.ch). Joint venture Tunnel Engelberg (German: ARGE Tunnel Engelberg): (ATE), CH-6388 Grafenort/Switzerland (Achermann AG – Swietelsky Bau Tunnelbau Gesellschaft m.b.H). Rowa Tunnelling Logistics AG, Leuholz 15, CH-8855 Wangen SZ/Switzerland (http://www.rowa-ag.ch). System Safety Scrapbook, P. L. Clemens, 2002, Sverdrup Technology, Inc. Swietelsky Bau Tunnelbau Gesellschaft m.b.H., EduardAst-Str. 1, A-8073 Feldkirchen/Graz (http://www.swietelsky.ch). US Military Standard 882D; 10 February 2000, US Department of Defense.

AIB, Amberg Consulting Engineers Ltd. (AIB), Trockenloostr. 21, CH-8105 Regensdorf-Watt/ Switzerland (http://www.amberg.ch).

161 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-20.qxd

2/18/04

12:01 PM

Page 163

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Managing underground construction risks in New York Nasri Munfah STV Inc.

Sanja Zlatanic & Paul Baraclough Parsons Brinckerhoff

ABSTRACT: New York Metropolitan Transportation Authority (MTA) has embarked on the construction of the largest transportation project ever undertaken in New York City. The East Side Access Project (ESA), a $6.3B project, will connect the Long Island Rail Road (LIRR) to the landmark Grand Central Terminal (GCT) on the East Side of Manhattan. In Queens, the connection will require the construction of five soft ground tunnels, using pressurized face TBM. The tunnels will be constructed under active railroad yard and the main line of the LIRR. In addition, cut and cover tunnels and underpinning of existing transit lines will be required. In Manhattan, the construction will require rock tunnels totaling over 7300 m long, and a series of rock caverns connecting to two station caverns to accommodate eight tracks with four platforms in a stacked configuration. The station caverns are approximately 20 m wide 22 m high 400 m long each and are situated about 10 m beneath the historic Grand Central Terminal (GCT) and high-rise buildings. The tunnels pass under existing subway tunnels, and a variety of historic and sensitive buildings in one of the world most expensive real estates area. In addition two underground ventilation and traction power facilities and eleven shafts of varying sizes and functions will be constructed. Several construction techniques will be used to excavate the tunnels, caverns, and shafts including tunnel boring machines, roadheaders, raise bores, and controlled blasting. Such major undertaking involves significant construction and commercial risks that must be dealt with. This paper provides a status report of the project focusing on the technical challenges in Manhattan, the construction risks and the identified measures to deal with such risks.

1 EAST SIDE ACCESS PROJECT GOALS AND DESCRIPTION 1.1

Project goals

The East Side Access project is vital for the general growth of New York Metropolitan Area by increasing the transportation facilities in the region and improving the daily commute to over 300,000 people. Presently, the Long Island Rail Road operates 36 trains per hour during the morning and evening rush periods to and from Penn Station in the West Side of Manhattan carrying about 240,000 commuters each way. Approximately half of these commuters final destination is the East Side of Manhattan. They travel within New York City using transit lines or surface transportation to reach their final destinations. This imposes added congestion on already overcrowded transit lines and

surface streets. In addition, Penn Station, which serves two other railroads, is at its capacity. The East Side Access project will improve the mobility in the region, will improve commuting for the LIRR passengers, and will relieve congestion in Penn Station. LIRR is planning to operate 24 additional trains per hour during the morning and evening rush periods to Grand Central Terminal on the East Side of Manhattan. This will increase its service by about 109,000 passengers. In addition, the project will provide a single seat ride to most riders to their final destination and will reduce their overall travel time by about 30 minutes each day. Overall, the project will improve mobility in the region and will stimulate the economic growth in New York. Recognizing these needs dates back to the 1960’s when the Metropolitan Transportation Authority (MTA), the

163 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-20.qxd

2/18/04

12:01 PM

Page 164

parent company of the LIRR, developed a plan to provide an East Midtown Terminal for the LIRR. In the 1960’s and 1970’s the MTA constructed 2560 m (8400 ft) tunnel under the East River. The tunnel extends from 63rd Street and 2nd Avenue in Manhattan under the East River to 29th Street and 41st Avenue in Queens. A four-track (two over two stacked) tunnel constructed by TBM, drill-and-blast, immersed tube and cut-and-cover tunneling methods. The Metropolitan Transit Authority (MTA)-New York City Transit (NYCT) operates a rapid transit line through the two upper level tracks. Due to fiscal constraints the two lower level tracks remained dormant until now and will be used by the LIRR. The project was revived in the 1990’s. An Environmental Impact Statement and Preliminary Engineering were completed in year 2000 and submitted to the Federal Transportation Administration. The selected alternative is a connection of the LIRR Main Line and the Port Washington Branch to a new deep underground terminal under the historic Grand Central Terminal. The project is budgeted to cost $6.3 Billion and to be completed by year 2012. See figure 1. 1.2

Figure 1. Overall plan.

Project alignment

The alignment in Queens extends about 1800 m (6000 ft) from the existing bulkhead near 29th Street and 41st Avenue across Northern Boulevard and connects to the LIRR’s Main Line and the Port Washington Branch. The existing two track tunnel branches into four track configuration which in turn connects to the Main line at the Harold interlocking complex East of 43rd Street passing across and underneath Yard A and Sunnyside Yard. In order to provide midday storage capacity, lead tracks will be provided to Yard A which will be reconstructed to provide a mid-day storage. Figure 2 illustrates the alignment in Queens. The alignment in Manhattan starts at the corner of 2nd Avenue and 63rd Street where the existing tunnels are terminated about 50 m (140 ft) underground and extends south-westerly to Park Avenue turning south. The tunnels start as single-track tunnels then separate through two wye caverns into four singletrack tunnels. The outer tunnels converge under the inner tunnels. At 51st Street the tracks converge via two crossover caverns situated over each other. They then separate again through two bi-level wye caverns into four two-level structures that tie into the end walls of the station caverns. The alignment passes under several underground transit lines and structures including the Metro North Railroad (MNR) Park Avenue Tunnel, the Lexington subway line, the 60th and the 53rd Street subway lines, and the 42nd subway shuttle. The LIRR Station will be situated under the historic Grand Central Station and it will consist of two caverns having eight tracks (two over two in

Figure 2. Alignment in Queens.

each cavern) with four island platforms. The caverns are situated between 44th and 48th Streets approximately 52 m (170 ft) below street surface. The cavern dimensions are 18.3 m (60 ft) wide by 23.8 m (78 ft) high and they are approximately 400 m (1200 ft) long and they are located 30.5 m (100 ft) on centres along the centreline of Park Avenue and the operating underground Metro North Railroad tracks. The lower level of the existing Grand Central Terminal will function as a large concourse providing for circulation, waiting areas, station functions, and retail spaces. South of the station eight tunnels extend south passing under the landmark New York Central Building, the 59-storey MetLife building, and the historical main hall of Grand Central Terminal. Each pair of tunnels merge via two three level caverns to provide four tail track tunnels extending to 38th Street. See figure 3. 1.3

164 Copyright © 2004 Taylor & Francis Group plc, London, UK

Project geology

Geotechnical considerations are the key elements of the success of the Grand Central Connection project. Most of the construction is underground and geotechnical issues will impact almost all design and construction activities.

09096-20.qxd

2/18/04

12:01 PM

Page 165

Figure 3. Alignment in Manhattan.

The excavations in Queens will be predominantly in soil. Some excavations in the Ravenswood Gneiss will be encountered in the cut-and-cover area as well as the initial portion of the bored tunnels where they are in a mixed face condition. The soil deposits in Queens represent a complex glacial depositional environment upon which a post-glacial tidal marsh environment and manmade fill deposits were superimposed. The soils consist of fill of heterogeneous mixture of coarse to fine silts and sands with building rubble and a possibly high concentration of rock fragments; glacial outwash deposits, predominantly granular soils, deposited by glacial meltwaters, consisting of clean coarse to fine sands to silty sands. Ice-rafted boulders of various sizes are anticipated in this deposit. Glacial lake deposits consisting of stratified silts, sand and clays deposited in quiet water is another major formation. Ice-rafted boulders may also be encountered in this deposit. Glacial till consisting of dense mixture of sand, silt, clay, cobbles and boulders, and nested boulders, is major layer along the tunnels in Queens. Also found along the alignment in certain areas compressible soils consisting of peat and partially decayed vegetation underlying the fill within the limits of Sunnyside Yard, along the course of a former creek. Most of the construction in Manhattan will be in a rock formation known as the Manhattan Schist. The rocks underlying Manhattan belongs to the New England Upland, and is locally known as the Manhattan Prong. This rock comprises three lithologically distinct sequences of a metamorphic assemblage of Proterozoic to lower Paleozoic age consisting of schist, gneiss and marble. The main rock types recovered from borings along the Manhattan alignment are metamorphic, dominated by schist and gneiss. The essential minerals are muscovite, biotite, quartz and feldspar (plagioclase, microcline and orthoclase). Garnet is the principal accessory mineral occurring as fine disseminated crystals and clusters. The schist frequently grades into a granofels (a fine to medium grained, equigranular metamorphic rock, in which there is no discernible foliation or banding). Amphibolite, consisting mainly of hornblende, plagioclase feldspar, quartz and biotite, is occasionally intercalated with the schist and gneiss, and lies parallel to the foliation. Also occurring within

the rock are frequent “igneous” layers normally parallel to the foliation that vary from medium-grained granite to coarse-grained pegmatite. The tectonic history of the rock has left the Manhattan Schist fractured and dislocated. Four dominant joint sets for the rock mass have been confirmed by the geotechnical investigation and exposed rock wall mapping. However, dip directions and dip angles vary widely across the Manhattan alignment. Manhattan is slightly raised above the sea level, and is bounded by the East River to the east and the Hudson River to the west. The area is heavily urbanized so infiltration of rainfall is low. More intense conductive fracturing occurs at locations of buried streams. These fractures are conduits for the groundwater with much greater hydraulic conductivity than other fractures in the undisturbed rock mass. The groundwater levels measured in observation wells range from 4.5 m (15 ft) below the street level along Park Avenue to less than 1.5 m (5 ft) below the invert of the existing lower level of Grand Central Terminal. The permeability was determined from in-situ packer tests and varies from 107 m/sec to 104 m/sec. 2 CONSTRUCTION OF THE MANHATTAN SEGMENT Building the ESA project will involve construction of a number of tunnels using a variety of construction methods. Selection of an appropriate tunneling method for each section of the alignment from a variety of tunneling techniques available in current construction practice will have a major impact on the successful completion of the project. Technical, practical, operational and economical factors affect the tunnel selection. In selecting a tunneling method a number of factors are evaluated including: ground conditions, groundwater, geometry and site constraints, impact on adjacent structures and utilities, environmental concerns, local construction practice, permitting and community acceptance, cost, schedule, contracting forms, and construction risks. In Manhattan, the construction will require rock tunnels totaling over 7300 m (24,000 ft), and a series of rock caverns (single level wyes, three-level wyes, crossovers) connecting to two station caverns. In addition, two underground ventilation facilities and traction power substations along with utility shafts, ventilation shafts and several vertical circulation shafts will be constructed. To meet the construction schedule constraints the Manhattan segment was divided into three major civil underground tunneling contracts: CM009 Manhattan Tunnels, CM012 GCT Caverns, and CM013 Ventilation Facilities. In addition a variety of preparatory contracts and finishing contracts are provided to complete the project.

165 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-20.qxd

2.1

2/18/04

12:01 PM

Page 166

Mechanized tunneling

The first tunneling contract (Contract CM009) will be let as an early contract to build most of the singletrack tunnels in Manhattan: two tunnels branching into four tunnels excavated by two tunnel boring machines (TBMs). The TBM will be 6.55 m (21-6) in diameter to accommodate an internal tunnel diameter of 5.95 m (19- 6). See figure 4. The tunnels in this contract will be provided with the initial support only consisting of rock dowels, rock bolts, and steel ribs as needed. Future underground structures include the East 55th Street ventilation plant, the East 51st Street crossover structures, the station approach wye caverns, the GCT caverns, the tail track wye caverns and the East 38th Street vent structure, will be enlarged in future contracts. This Contract also includes the construction of an underground TBM assembly chamber and two re-assembly chambers in the location of future track wyes. The TBM assembly and reassembly chambers will be constructed using controlled blasting techniques. The contract also incorporates vertical raise bores 3 m in diameter (10-0) in five future vertical shaft locations and four inclined raise bores in future station escalator shafts. A preparatory contract CM016 will be awarded earlier to excavate the approach tunnels of about 100 m (330 ft) to the assembly chamber using roadheader. In this contract the performance of the roadheader will be assessed for its suitability technically and commercially to excavate the Manhattan schist for potential use in future contracts in lieu of the drill and blast method. If successful the roadheader use will significantly mitigate issues related to excavation overbreak, noise and vibration, and impact on adjacent facilities and the public. The TBM tunnels will be excavated using one of three initial support systems referred to as support classes using the observational approach. These support systems are designed based upon anticipated ground conditions and ground behavior. The support

Figure 4. Typical single track tunnel.

systems are designed to arrest potential movement of rock blocks and wedges and to prevent loosening of the surrounding rock mass. Support elements are installed concurrent with the TBM excavation. In drill and blast or roadheader excavations, initial ground support elements include rock bolts, dowels and a reinforced shotcrete lining. The shotcrete lining will consist of a reinforced shotcrete layer 100 mm (4 inches) to 250 mm (10 inches) in thickness, depending on ground conditions, tunnel size and support class. Shotcrete reinforcement will include welded wire fabric or lattice girders. A secondary, final cast-inplace concrete lining will be provided for long-term support under a subsequent contract. 2.2

The major underground contract in Manhattan is CM012 GCT Caverns. This contract completes the civil underground work in the previously constructed tunnels. Several single level and triple level caverns will be built in this contract to accommodate track crossovers, switches, and railroad system facility spaces. In addition cross passages, vertical circulation and utility shafts, service tunnels, inclined escalator shafts, and other underground structures will be built as part of this contract. The final liner, duct bench, and other embedment for the entire structures and tunnels will also be constructed in this contract. Furthermore, rehabilitation of the existing tunnels will be done in this contract. The main feature of this contract is the twin station caverns situated between 44th and 48th Streets, with their invert about 40 m (131 ft) below grade. The caverns are spaced approximately 30 m (100 ft) on centers horizontally beneath Park Avenue, a 35-storey high landmark building and a portion of the Met Life tower. The caverns are approximately 18 m wide 20 m high 360 m long each (59 ft 66 ft 98 ft) and are situated about 10 m (33 ft) beneath the historic GCT. See figure 5: GCT caverns.

Figure 5. GCT caverns.

166 Copyright © 2004 Taylor & Francis Group plc, London, UK

Cavern construction

09096-20.qxd

2/18/04

12:01 PM

Page 167

Each cavern will have four tracks, two over two with center-island platforms. Each cavern has three levels, a lower train room, a mid level mezzanine, and an upper train room. The two caverns are interconnected with four cross passages leading to four escalator banks, two elevator banks, and emergency stairs connecting the caverns with the concourse level in the existing Grand Central Terminal. The inclined escalator shaft at 45th Street houses five escalators and is approximately 11 m (36 ft) wide by 7 m (23 ft) high. The other escalator shafts house four escalators each and are approximately 9 m (30 ft) wide by 6 m (20 ft) high. In addition, several utility and ventilation shafts and tunnels will be constructed to provide the lifelines to the station. The excavation of the rock in this contract will be done by drill and blast using controlled blasting techniques. Alternatively, if the roadheader has proven its ability to excavate the Manhattan schist economically, it can be used to excavate the top headings. The initial liner will be rock dowels, rock bolts, and lattice girders and shotcrete. The final liner will be cast in place concrete liner varying in thickness from 300 mm (12 inches) for the running tunnels to 750 mm (30 inches) for the GCT caverns. The tunnels and caverns will be waterproofed utilizing an open waterproofing system consisting of a continuous PVC membrane and a layer of drainage fabric. The waterproofing system will be designed as a drained, open system with perforated sidewall drainpipes for collection of groundwater. The architectural finishes and the electrical and mechanical works will be part of a subsequent finishing contract. 2.3

Ventilation facilities

Three underground ventilation structures will be built as a separate contract (CM013) to provide the ventilation needs in the tunnels during emergency and congested modes. The equipment of the ventilation fans and associated mechanical, electrical and control systems will be part of a project wide contract. The ventilation facilities will be located in the 38th Street, the 50th Street and the 55th Street. The latter also houses traction power substation. The ventilation facilities vary in size and equipment. The 38th Street facility consists of two fans 200,000 cfm each placed horizontally in an underground cavern and connected via a ventilation tunnel to a shaft to the surface placed in a parking lot. The shaft is 5 m 5 m (15 ft 15 ft) and is 42.5 m (129 ft) deep. The construction of the shaft will be top down using drill and blast technique. The 50th Street facility consists of four fans 150,000 cfm each placed horizontally in the mid-level of the station’s northern interlocking structures. The two sets of fans are connected via cross flues and dampers to enable ventilating any of the eight tunnels

at that location. The fans are connected to the surface via ventilation tunnel and a vertical shaft. The construction of the shaft will be top down using drill and blast. The 55th Street ventilation facility consists of four vertically placed ventilation fans 200,000 cfm each and a two-unit traction power substation. The facility is a two level underground structure approximately 12.4 m 11.7 m (40-7 38-6) and 80 m (263-6) long. The upper level consists of a plenum and mechanical electrical rooms while the lower level houses the traction power substation. The facility will be placed in the 55th Street bed and is connected with the tunnels via a vertical shaft 14.5 m 14.5 m (44 ft 44 ft) and 44 m (134 ft) deep. The shaft connects with tunnel via plenum cavern. The construction of the surface facility and the shaft will be done by the cut and cover method under street decking using drill and blast. 3 RISK MANAGEMENT The Manhattan segment will be built under and adjacent to several operating transit lines including the 63rd Street line, the Lexington Avenue line, the IND 53rd Street lines, MNR’s Park Avenue tunnels, Park Avenue viaduct, the 42nd Street No. 7 line, and the Time Square Shuttle, and the Park Avenue Tunnel. The project alignment lies beneath mid-town Manhattan, the most densely populated urban business area in the region and certainly of extreme value to the City. Due to the large number of sensitive and critical structures above and adjacent to the planned tunnel and shaft excavations and construction, it is imperative that major risks and associated mitigation measures be identified during the early stages of the project. At the early stages of the design development, the project team prepared a risk management plan. The plan emphasized that risk management is not a one-time task, but rather a continuous process throughout the life of the project to identify, track and manage risks in the planning, design, and construction stages. Generally, the process consists of four elements:

• • • •

Each area of the project, during the development of the respective construction packages, was examined and analyzed to identify potential risks. The risks identified were evaluated for their probability of occurrence and a risk response and monitoring plan was developed. It was recognized that the level of risk sharing is a major factor in deciding the type of procurement practice to be implemented. Although several potential

167 Copyright © 2004 Taylor & Francis Group plc, London, UK

Risk Identification Risk Quantification Risk Response Development Risk Response Monitoring during design development and construction.

09096-20.qxd

2/18/04

12:01 PM

Page 168

procurement approaches were evaluated it was determined that due to the high risk and exposure of this project that the conventional design-bid-build process be used for the underground construction contracts. This decision was made because the design-bid-build process provides strong owner control and better quality; it delineates clearly the roles and responsibilities between the owner, and the contractor and provides an equitable sharing of the risks. The MTA provides complete identification of the geotechnical and tunneling issues and resolutions to the environmental issues. It also provides a complete design and obtains all approvals, while the contractor executes the work with the best means and methods. Using the conventional contracting practice and to equitably share and manage risks, several provisions were provided in the contract form. These include pre-qualification of contractors, complete geotechnical disclosure, the implementation of a Dispute Review Board (DRB), the use of differing site condition clause, the use of unit prices and contingent bid items, the provision of value engineering, the provision of owner controlled insurance program, and a partnering program. 3.1

Pre-qualification of contractors

The major underground contracts required technical and financial pre-qualifications of the potential bidders to ensure their ability to perform the work effectively, economically, and to a high quality. The pre-qualification would include the company’s or the joint venture’s technical ability to perform the work. Items to be evaluated include: approach, experience with similar projects, with similar ground, and similar proposed methodology. Pre-qualification of the key staff, such as the contractor project manager, the field manager or superintendent, the TBM operator, the TBM maintenance person, etc., is critical for a successful project. In addition to the bidder’s financial ability to obtain bonding and its solvency, its history of completing projects on time and within budget is a factor in the bidder financial qualifications. Pre-qualification could also extend to major subcontractors and major suppliers such as the TBM manufacturer, liner supplier, etc. 3.2

Full geotechnical disclosure

Experience has shown that full disclosure of geotechnical information would reduce the risk to both the owner and the contractor and thus the project cost. Therefore it was important for the MTA to invest in a comprehensive geotechnical program including an extensive boring program, exposed rock face mapping, laboratory testing, and in-situ testing. The information is included in the contract documents in the form

of Geotechnical Data Report, Geotechnical Design Summary Report, Geotechnical Interpretive Report, and/or Geotechnical Baseline Report. The GBR establishes quantitative values for selected conditions anticipated to have great impact on construction. These values are established through technical interpretation of the data and commercial considerations of risk allocation and sharing. The advantages of this report are ease of administration of contractual clauses, unambiguous determination of entitlement, clear basis of contractor’s bid, and clear allocation of risk between owner and contractor. The intent of the disclosure of geotechnical information and the use of the Geotechnical Baseline Report is to allocate and share underground construction risks between the owner and the contractor. 3.3

In the DRB process, a board of independent, experienced, and impartial members is selected to hear and address disputes. Generally the board consists of three members, one representing the owner, one representing the contractor, and the third, who will act as the chair of the board, selected by the other two members. The board provides recommendations to resolve disputes that participants are unable to make. It was found that this process has resulted in lower bids, better communication and less acrimony at the job site, fewer claims, and more timely and cost effective resolutions. For the ESA project, the DRB process will be limited to underground related issues. It will involve formal and informal processes in which position statements and expert reports are made and presented to the board. The board will meet and visit the construction site regularly to familiarize themselves with the project issues. Board findings and recommendations will be made; however, acceptance of the DRB findings is not binding or admissible in a legal process. The MTA is committed to a fair and equitable resolution and committed to support the process and does not want the DRB process to be just a step in a legal process, rather it will be the basis of negotiating a fair and equitable resolution. 3.4

Differing site condition

The Differing Site Condition clause is used on the ESA Project as a measure of allocation of risk between the owner and the contractor relative to the ground condition. In exchange for lower initial bids, the owner bears some portion of the risk of subsurface condition. Bid contingencies in the low bid are paid by the owner whether adverse conditions are encountered or not. On the other hand, DSCs are paid only if they are encountered. Two categories of DSCs are identified in the contract documents: Category 1 governs when subsurface

168 Copyright © 2004 Taylor & Francis Group plc, London, UK

Dispute review board

09096-20.qxd

2/18/04

12:01 PM

Page 169

conditions differ from those indicated in the contract. This is based solely on what is stated in the contract, including geotechnical data or geotechnical interpretation, or geotechnical baselines included in the contract documents. Category 2 applies when conditions which were not known to the contractor at the time of contracting differ from those normally encountered in the area. It is generally related to unusual conditions and not based on contract documents. It is important to note that to recover on the DSC clause; the contractor must show impact on cost and time and must show causality. 3.5

Partnering

The MTA incorporated in all the contracts a formal partnering provision. The goal of this process is to minimize disputes and to prevent them from escalating in time and value by resolving them at the lowest possible level in the project organization. It attempts to establish a win-win attitude between the project participants, including the owner, the contractor, the engineer, and the construction manager. This process encourages dialogue among the various participants and relies on reasonable people to resolve disagreements reasonably. It seeks to eliminate adversarial posturing and positioning that often develop when disputes and claims arise. Through this process a series of dialogues and interactions are developed whereby the team members are encouraged to work out differences for the best interests of the project. When an issue is not resolved at the lowest level, it is brought up to a higher level for resolution avoiding legal proceedings. 3.6

•

•

Value engineering

To stimulate innovative approaches within the limitations of the contractual requirements, the MTA opted to include a value engineering clause in the contract. Efficiency is achieved by relaxing the design criteria where not critical or meeting the intent of the design more efficiently via creative approaches. The saving achieved by value engineering is shared between the owner and the contractor. It is important to assess the potential effects of differing site conditions on the design as modified by the value engineering. 3.7

Environmental issues

The selection of the tunneling method in Manhattan takes into consideration not only the technical challenges and the construction risks, but also environmental concerns and public acceptance. Therefore several unique approaches were developed for implementation of this project:

•

Access Shaft: Locating the TBM access shaft in Manhattan is difficult because of the extensive

169 Copyright © 2004 Taylor & Francis Group plc, London, UK

development near the site and the potential for shaft construction to disrupt the local community and businesses. Extensive utility relocations will be required and significant impact on traffic will occur in addition to the noise, dust and disruption. Therefore, it was decided to construct the Manhattan shaft in Queens. All materials will be brought into Manhattan via the shaft in Queens and through the existing tunnels. Similarly, the muck will be removed using conveyors or hoppers through the existing tunnels to Queens where it will be hauled away by rail or trucks. Construction of the Queens shaft proved to be a feasible, economical and environmentally friendly solution that minimizes construction risks on adjoining developments and the public. Minimizing Impact on Overlying Buildings: To assess possible impact of the tunnels and caverns excavation on the existing high-rise buildings and underground structures such as transit lines, stations, and utilities, settlement analyses were performed at critical sections along the alignment. In addition, a sophisticated system of monitoring was designed using inclinometers, extensometers, liquid level, and global optical survey. The results of these analyses showed insignificant elastic surface settlements in the range of a few millimeters. The analysis included considerations of sequential excavation followed by immediate installation of initial support. Settlements (tunnel and surface), convergence, and stress and strain measurements will be closely monitored and response measures will be implemented quickly to avert adverse impact if encountered. Noise and Vibration: Determination of the likely levels of vibration, noise, and ground borne noise during construction is necessary to provide contractors with guidelines for the explosive charge and delay configuration during blasting, selection of construction equipment for drilling, mucking and mechanical excavation, and to provide for appropriate construction sequencing. To estimate impacts of construction, design criteria were developed for peak particle velocity (PPV), extensive data collection and interpretation were made, and analyses based on the collected information were performed. The results were used to recommend measures for structures protection and instrumentation and monitoring. As part of public participation program the results of the analysis will be shared with affected third parties. The Contractor will perform test blasts before production blasting starts for each construction contract. This test blast program will confirm the applicability of the resulting vibration regression equations. Furthermore, the test blast program will aid in determining necessary adjustments to the blasting procedures, round length, and delay. Construction sequencing will be further adjusted to comply with the established criteria. Ground

09096-20.qxd

•

2/18/04

12:01 PM

Page 170

borne noise impact was assessed. Factors that influence ground borne noise include local geology, building foundation construction, number of floors above street level, and resonance of upper floors. During drilling, TBM, and raise bore excavation, the estimated ground borne noise level at sensitive receptors is estimated. The data will be used to monitor construction performance and a basis to provide good relationship with the public and affected neighbors. Coordination with the Public and Stakeholders: An important element in a successful major construction program is public opinion and the approval of citizen groups and public civic associations especially in a well-established place such as Manhattan. It was recognized the importance of obtaining consensus from the public and the facility users early in the project development stage. It is important to identify their concerns and address them. Concerns are usually related to construction noise and vibration, working hours, disruption to businesses and daily routine activities, emission of dust and pollutants, and the potential increase in pests and rodents. It is important to listen to the people’s concerns and address them early in the project developments. It is important to obtain buy-in of construction methodologies and approaches by the public in the early stages of the project and is critical for its successful completion. An extensive public outreach program was established and implemented during the design development to address the people’s concerns. This program will continue during construction to assure the people that agreements reached are being implemented and also to address new concerns that might develop after the start of construction. Coordination with the affected third parties has been an important aspect of the project. Public outreach, dissemination of timely and accurate information about the project’s different construction phases

and the impact of construction are part of the overall program and are being successfully implemented. 4 CONCLUSION The East Side Access Project, the largest transport ation project ever undertaken in New York City, will improve the overall New York metropolitan area transportation system and will stimulate economic growth. Scheduled for completion by 2012, early preparatory contracts have been under way and the award of first tunneling contract is scheduled to be early 2004. To be successful the project has been planned using state of the art design concepts, advanced construction approaches, suitable provisions for risk management and sharing, and careful attention to environmental issues.

REFERENCES Della Posta, M. and Zlatanic, S. “Manhattan Segment of The East Side Access Project: Design Evolution” RETC Proceeding 2001 Munfah, N. “Connecting Long Island Rail Road to Grand Central Terminal in Midtown Manhattan” RETC Proceeding 2001 Munfah, N. “Contracting Practices for Underground Construction” Proceeding Underground Construction British Tunneling Society 2003 Munfah, N., Zlatanic, S. and Stehlik, E. “Connecting a Commuter Railroad to Historic Terminal” Proceeding ITA Prague, 2003 Munfah, N. “The Manhattan Connection” Tunnel and Tunneling, 2001 Munfah, N. “The Application of the DRB on the East Side Access Project”, Presentation in the RETC New Orleans, 2003 Wone, M. “Rock Tunneling Challenges in Manhattan” Proceeding ITA 2003

170 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-21.qxd

2/18/04

12:29 PM

Page 171

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Risk allocation in tunnel construction contracts William R. Wildman Sutherland Asbill & Brennan LLP, Atlanta, Georgia, USA

ABSTRACT: One way to minimize claims on a tunnel construction project is to develop a fair contract that allocates risks to the party best able to manage that risk. This paper will explore various provisions found in tunnel construction contracts that allocate risks among the parties, including the differing site conditions clause; delay damages provisions that limit contractors to their extended job site conditions cost and owners to liquidated damages; indemnification for site accidents and insurance provisions, including wrap up policies covering all the parties to a tunnel project. This paper will then briefly touch on design/build as an alternative project delivery method for allocating risks on a tunnel project. The paper concludes with the general hypothesis that a contract which fairly allocates risks is a contract which minimizes the risk of litigaton.

1 THE DIFFERING SITE CONDITIONS CLAUSE A tunnel project involves great uncertainty and potential for catastrophic cost overruns. Under the common law, the contractor would bear the risk of additional expenses due to unforeseen conditions. Indeed, even if the owner decided to voluntarily modify the contract and pay the contractor the additional expenses, the pre-existing duty rule made it very difficult to do so because the new promise would be unsupported by consideration. The general rule is that no matter how onerous the burden, once a contractor has promised to perform, any added expenses caused by unforeseen conditions is allocated to the contractor. Although the common law allocated the risk of unforeseen conditions to the contractor, public bodies engaged in large civil construction projects, including tunnel projects, became mired in endless litigation over claims regarding unforeseen conditions or claims that the plans and specifications inadequately described the sub-surface conditions. Moreover, these government bodies also faced the problem of bankrupting contractors who could not afford to complete the work based on the agreed upon fixed price in the contract. A differing site conditions provision now almost uniformly appears in fixed fee contracts. [The American Institute of Architects (AIA), the Federal Acquisition Regulations (FAR), and the Engineers Joint Contract Documents Committee (EJCDC) agreements all have differing site conditions clauses.] An example of such a differing site conditions clause was

included in the Milwaukee Metropolitan Sewerage District (“MMSD”) Northshore Inceptor-Phase 1A 30-foot diameter main tunnel and access shafts contract. This 28,000 foot long, 300 foot deep, 30-foot diameter tunnel was designed to capture combined sewer overflows during storm events and then eventually pump the combined sewer overflow to a sewage treatment plant for later disposal into Lake Michigan. The differing site conditions clause provided: “A. The Contractor shall promptly, and before such conditions are disturbed, notify the Owner by written notice of: Subsurface or latent physical conditions at the site differing materially from those indicated in this Contract, or Unknown physical conditions at the site, of an unusual nature, differing materially from those ordinarily encountered and generally recognized as inherent in the work of the character provided for in this Contract. B. The Owner shall promptly investigate the conditions. If he finds that conditions materially differ and will cause an increase or decrease in the Contractor’s cost or the time required to perform any part of the work under this Contract, Owner shall, after receipt of the Contractor’s written statement under “D” below, make an equitable adjustment and modify the Contract in writing. C. No claim of the contractor under this Article shall be allowed unless the Contractor has given the notice required in paragraph A of this Article. However, the Owner may extend the time prescribed in paragraph A of this Article.

171 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-21.qxd

2/18/04

12:29 PM

Page 172

D. If the Contractor intends to assert a claim for an equitable adjustment under this article, he must, within thirty (30) days after the owners determination as to whether a Differing Site Condition exists, submit a written statement setting forth the nature and monetary extent of such claim. The Owner may extend the thirty (30) day period. E. No claim by the Contractor for an equitable adjustment hereunder shall be allowed if asserted after final payment under this contract.” This differing site conditions clause can be characterized as dealing with a “Type I” and “Type II” Condition. Under a Type I Condition, the contractor must show that it “encountered subsurface or latent physical conditions at the site that differed materially from those expressly or impliedly indicated in the contract.” To prove a Type II Condition, the contractor must prove that it encountered “an unknown physical condition at the site, of an unusual nature, differing materially from those ordinarily encountered and generally recognized as inherent in the work of the character provided for in the contract.” Even with the differing site conditions clause, the contractor continues to bear the common law risk of the usual and expected conditions at a construction site; however, the contractor no longer must gamble that the unexpected might occur. In the Milwaukee tunnel project, the contractor encountered conditions that were materially different from those represented in the contract documents. When large water inflows and poor rock support were encountered through major reaches of the tunnel construction, the engineer granted differing site conditions status to the problems and compensated the contractor on a time and material basis over and above the contractor’s lump sum price. Although the contractor’s original contract price was for $46,000,000.00, the Milwaukee Metropolitan Sewer District ended up paying the contractor $166,000,000.00 to complete the North Shore Interceptor Tunnel. 1.1

Justifications for the differing site conditions clause

[For a very detailed examination of these justifications, see Hazel Glen Beh, Allocating the Risk of the Unforeseen, Subsurface and Latent Conditions in Construction Contracts: Is There Room for the Common Law? 46 U. Kan. L. Rev. 115 (1997).] A differing site conditions clause in a tunnel construction contract will encourage bidders to submit their lowest bid rather than build cushions into their bids for contingencies that may never occur. The differing site conditions clause should save the government money over time because it allows the contractor to remove its contingency from its bid and the owner avoids overpayment on the majority of projects and is

required to pay for differing site conditions only when they occur. The differing site conditions clause should also minimize claims. Unlike the common law, a construction contract containing a differing site condition clause requires the owner to negotiate a new price for the unanticipated work. Although this is the noble goal of the differing site conditions clause, in practice, while the owner must adjust the contract if a differing site condition exists, the owner can still dispute the existence of a valid differing site condition claim or the amount of the adjustment requested by the contractor. Still, under the common law, the contractor was faced with financial ruin. The contractor would necessarily have to litigate if the differing site condition was significant. It is therefore more likely that the differing site conditions clause has reduced the adversarial nature of tunnel construction projects. The differing site conditions clause also reduces the likelihood that contractors will go out of business if they encounter a differing site condition. Absorbing the costs of unforeseen conditions protects the industries upon which the large owner depends. Additionally, if contractors have to face the costs of unforeseen conditions in tunnel projects, they may choose not to bid on such high risk projects, and in the long run there will be fewer competitors performing this type of specialized work. The different site conditions provision also keeps down the cost of preparing bids and doing business. With such a clause the contractor is not forced to do his own extensive and expensive soil boring tests. This also encourages the owner to provide the contractor with as much information about the site as possible so that all the bidders are bidding on a level playing field. Rather than having multiple soil testing programs ongoing (and hence having that cost of that soil testing program incorporated into the contractor’s bid), the government can spend the money once on its own program and keep the cost of all bids down. Another benefit of the differing site conditions clause is that it permits the contractor to recover for the additional cost of completing the contract when a true differing site condition is encountered rather than cutting costs and otherwise inappropriately attempting to make up for the potential loss. Thus, the differing site conditions provision promotes direct recovery rather than indirect and inefficient recoupment of costs. The government body planning a tunnel project usually has substantial knowledge or the ability to acquire such knowledge about the nature of the latent conditions and risks involved at the site. Moreover, the government usually has greater ability and time to conduct site exploration and investigation than does a contractor who must confine its inspection to a brief pre-bid period. The government body planning a tunnel

172 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-21.qxd

2/18/04

12:29 PM

Page 173

project also has the ability to make a decision about the extent of investigation desired versus the amount of risk it wants to assume for unforeseen conditions. The public owner can then budget for the contingencies just as the contractor did at the common law. With the differing site conditions clause, the public owner has an incentive to share its knowledge about latent conditions with the contractor so that it gets the best and most accurate bids from the contractors. Under the common law, however, the owner has little incentive to be candid concerning the site because he would be rewarded by a contractor’s inaccurate low bid. The typical differing site conditions clause does not eradicate all problems associated with site investigations. Since a contractor performing tunnel work must still conduct some reasonable site inspection and can only make claims if the conditions materially vary from those typically encountered on such projects, public bodies still face the potential for significant litigation over whether a contractor did do an adequate site investigation and whether the conditions encountered were indeed unusual. In the long run, however, the typical unforeseen conditions clause will benefit the public by allocating the risk of such conditions to the party best able to manage it, the owner. 1.2

Risk sharing versus risk allocation

Under the common law, the risk of unforeseen conditions in a tunnel contract was allocated to the contractor. The unforeseen conditions clause found in most tunnel contracts today, however, allocates the risk of unforeseen conditions back to the owner. One commentator has proposed that the risk be shared by making sure that the contractor is paid for the additional work performed but that he does not reap a profit from the unexpected condition. [See Beh, supra.] By eliminating recovery for any profit on differing site conditions work, especially the Type II kind where neither party can foresee that the condition would be different from the type of condition normally encountered, the contractor gives up the benefit it receives for an unforeseen condition (the profit) without suffering a loss, while the owner pays the cost of the unexpected condition without rewarding the contractor. Both parties bear part of the financial risk of the unforeseen condition. Although this proposal may reduce a contractor’s incentive to submit an inordinately low bid with the hope that he will make it up through change orders, it is perhaps naïve to believe that contractors will bid that much more accurately knowing that they will only be paid for the cost of doing their work and not any profit, since it is very difficult to determine exactly how contractors calculate their profit. Moreover, the owner might still litigate whether or not the condition is truly atypical, because the bulk of the costs paid are

for direct costs rather than any additional profit, and it might be in the owner’s best interest to challenge the contractor’s entitlement to the claim. On balance, however, it would be worthwhile to experiment with this proposal to see if it truly saves the public money. 2 DELAY DAMAGES Apart from the high cost of dealing with the extra work associated with encountering differing site conditions in tunnel projects, the public owner is often faced with significant delay claims from the contractor. These claims typically involve extended job site general condition costs (e.g., additional supervision, jobsite trailer, utilities, and other costs), acceleration costs in the form of overtime labor charges, unabsorbed home office overhead claims, and lost bonding capacity claims. Some public bodies have responded to these claims by including “no damages for delay” clauses in their contracts. Thus, apart from the direct costs associated with overcoming the unforeseen subsurface conditions, the public owner may require the contractor to give up his delay damage claim and settle for only a time extension. The “no damage for delay” clause has resulted in a significant amount of litigation throughout the country. [Bates & Rogers Const. Corp. v. North Shore Sanitary Dist., 92 Ill. App. 3d 90, 414 N.E.2d 1274 (1981) (enforcing the clause); Corrino Civettes Const. Corp. v. City of New York, 67 N.Y.2d 297, 502 N.Y.S.2d 681, 493 S.E.2d 905 (1986) (finding an exception to the clause).] Although most courts do enforce the clause, others have not. A more reasonable approach is to reimburse the contractor for his direct jobsite overhead costs associated with the delay and to deny recovery for any of the more subjective or “soft” type of delay damages. If a contractor is delayed by additional work ordered by the owner or he encounters unforeseen subsurface conditions, the contractor should be able to recover what he can demonstrate as his extended job site general conditions. Some owners go one step further and agree up front what the daily overhead rate will be in the event the contractor is delayed by an unforeseen condition. The owner must be sure that he does not pay a higher daily rate than what he would pay if the contractor were forced to prove what his actual jobsite general condition costs were at the time of the delay. For example, a daily rate agreed to up front might reflect higher jobsite general conditions incurred at the front end of a project versus the lower overhead generally incurred as the project is winding down. If the delay is encountered toward the end of the project, the owner might pay a higher daily rate than he would otherwise pay if the contractor were forced to produce

173 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-21.qxd

2/18/04

12:29 PM

Page 174

records showing his actual jobsite general conditions costs at the time of the delay. Conversely, the owner may wish to avoid disputes over the contractor’s actual cost and simply agree to a daily rate up front. The owner in a tunnel project will undoubtedly want to allocate the risk of contractor delay to the contractor and require him to pay his damages if the contractor fails to complete the project by the substantial completion date. The public owner must decide whether to apply a liquidated damages provision or simply make a claim for the actual damages sustained because of the contractor’s delay. Most contractors are unwilling to sign construction contracts which expose the contractor to unlimited damages for delay. In tunnel construction contracts, it is also difficult for the owner to prove its actual delay damages because such damages for a public body are often speculative. The following liquidated damage provision is therefore a good choice for addressing contractor delay in a tunnel project: “The parties agree, by initialing here where indicated (owner) ______ (contractor) _______ that it would be extremely difficult and impracticable under the presently known and anticipated facts and circumstances to ascertain and fix the actual damages the owner would incur should contractor delay in achieving substantial completion by the date set forth hereof. Accordingly, the parties agree that if the contractor fails to achieve substantial completion within such time, then the owner’s exclusive remedy for such failure shall be to recover from the contractor the sum of $_______ for each calendar day substantial completion is so delayed by contractor.” The owner must insure that the amount of the daily rate for liquidated damages reasonably compensates him for any delays while at the same time bearing a rational relationship to the amount of actual damages sustained lest it be deemed a penalty and hence unenforceable. 3 SITE ACCIDENTS Perhaps one of the greatest risks on any tunnel project is the risk of worker injury or death. Most standard form construction contracts, including public contracts for tunnel projects, require the contractor to be responsible for means and methods, including all safety programs on site. The problem arises as to whether or not other parties involved in the construction process might be responsible for site safety accidents. Specifically, the design engineer might be responsible if he plays a more active role in the day to day operations of the tunnel project. The problem becomes acute when injured workers are prohibited from suing

their employers because of workers compensation statutory immunity. The injured worker or his estate then sues the engineer and the owner. An additional threat of liability is the Occupational Safety and Health Act (“OSHA”). [29 U.S.C. §651, et. seq, (2002).] Although most contracts will usually provide immunity to design professionals against third-party claims, design professionals need to be careful about potential OSHA violations. In one famous case, the Occupational Safety and Health Review Commission (the “Commission”) held that CH2M Hill Central, Inc. (“Hill”) was liable for violations of federal construction standards that apply to employers “engaged on construction work.” [See CH2M Hill Central, Inc., 1997 USAACL Lexis 34 (No. 89-1712, April 21, 1997.] The case arose out of the deaths of three workers from a methane gas explosion while working on the tunnel project in Milwaukee referenced above. In 1977, the MMSD undertook a $22 billion construction program calling for eighty miles of sewer tunnels. The MMSD contracted with a variety of companies including Hill, which served as the lead engineering consulting firm for the project. Hill’s contract with the MMSD included a provision that stated that visits to the construction site by Hill would not relieve the contractor of its obligation for, among other things, “all safety precautions.” In May 1988, a subcontractor requested a clarification from Hill regarding whether certain types of electrical equipment were approved for excavation due to a methane gas build up. After discussing the request with the MMSD, Hill provided its subcontractor with further explanation regarding electrical equipment. An explosion resulted when the electrical equipment ignited the methane. The OSHA commissioner issued citations to Hill as a result of this accident claiming that Hill violated construction industry specific standards promulgated by the Commission. The Commission established a new test holding engineering and architectural firms liable for the OSHA construction industry standards where they (1) have broad responsibilities in relation to construction activities, including both contractual and de facto authority relating to the work of trade contractors, and (2) are directly and substantially engaged in activities that are integrally connected with safety issues, notwithstanding contract language expressly disclaiming safety. [Id at 51, 53.] Although the Seventh Circuit Court of Appeals ultimately found that Hill could not be held accountable for OSHA violations because Hill did not “exercise substantial supervising over actual construction,” it did acknowledge that a design professional could be liable under OSHA under different circumstances. [CH2M Hill, Inc. v. Herman, 192 F.3d 711, 724 (7th Cir. 1999).] This again underlines the importance of

174 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-21.qxd

2/18/04

12:29 PM

Page 175

careful contract drafting to allocate the risk of site safety to the contractor and encourage design professionals to avoid any involvement in construction safety to avoid liability under OSHA standards. One way to insure that the contractor will cover and defend against such claims is to require the contractor to sign a broad form indemnification agreement in the contract. An example of a broad form indemnity that would cover such claims is as follows: “To the fullest extent permitted by law, the contractor shall indemnify, defend and hold harmless the owner, the owner’s representative, and the engineer, and their agents and employees, from and against all claims, liens, damages, losses and expenses, including reasonable attorneys’ fees, attorneys and expert fees, arising out of or resulting from the performance of the work, the existence of this contract, or the presence on the jobsite of the contractor, its subcontractors, their agents, servants or employees, or the ownership or possession of any property or material by said parties on the jobsite, regardless of whether or not it is caused in part by any of the parties indemnified hereunder. In all claims against the owner or the engineer or any of their agents or employees by any employee of the contractor, any subcontractor, anyone directly or indirectly employed by any of them, or anyone for whose acts any of them may be liable, the indemnification obligation hereunder shall not be limited in any way by any limitation on the amount or type of damages, compensation or benefits payable by or for the contractor, or any subcontractor under workers or workman’s compensation acts, disability benefits, or other employee benefits acts.” While such an indemnification clause (and the clear assignment of site safety responsibility in the contract) will not eliminate worker injury claims against the owner and his engineer, it should reduce them and provide the owner and his engineer with strong arguments that the contractor is responsible for defending against such claims. 4 PROJECT INSURANCE One way to control risks is to insure against them. The owner will typically require the contractor to carry comprehensive general liability insurance protecting the contractor, his employees, his subcontractors, the public, the interests of the owner and engineer against bodily injury and property damage. The policy should also cover contractual, strict and negligence type liability. It is also a good idea to require a umbrella excess liability policy. The general liability policy limits will of course be related to the size of the project. In most tunnel

projects, the limits should be at least $10 million, if not more. The typical general liability policy will exclude pollution, faulty workmanship, damages to selfperformed work, and professional services. The owner will usually provide the builders risk insurance coverage for the work included in the contract documents. This policy covers the property of the owner, and the liability of the owner for property of others, consisting of all real property in the course of construction, alteration or repair by the owner. The contractor must also purchase worker’s compensation and employers’ liability insurance for the statutory limits prescribed. On large tunnel construction projects, it may indeed be best for the owner to purchase a so called wrap-up owner controlled insurance program. The owner can cover all firms working at the jobsite and purchase workers compensation, general and umbrella liability, professional liability and builders risk. These types of insurance programs can also cover professional design services. Given the complexity of tunnel construction projects and the enormous risk of claims, it makes sense to implement such a wrap-up policy to minimize the potential litigation. If a single insurance company covers all types of claims that might arise on a tunnel project, it is more likely that the claim will be paid by the insurance company rather than result in endless litigation with the parties involved in the claim. 5 PROJECT DELIVERY METHOD Most tunnel projects will follow the traditional method of project delivery. The public owner retains an engineering firm to design the tunnel and then prepares bid documents for sealed competitive bidding. The lowest responsible bidder is selected and then the general contractor builds the project, hopefully in accordance with the contract documents. Many public owners believe that this is the “only” way to build a tunnel project since the engineer will or should be loyal to the owner’s interest in seeing to it that the contractor delivers the project on time, within budget and in accordance with the contract documents. Some government bodies might be prohibited from using alternative project delivery methods, but more states are permitting alternative project delivery methods as long as competitive bidding is used. Unfortunately, today’s design community is risk averse due to the uncertainty of subsurface construction and will often be at odds with its owner/client should the contractor make a claim that the contract documents contained errors and omissions. The owner must often defend its engineer in order to resist a claim by its contractor, only later to determine that the

175 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-21.qxd

2/18/04

12:29 PM

Page 176

contractor’s claim had merit and be forced to litigate with its engineer. Under the Spearin Doctrine [United States v. Spearin, 248 U.S. 132, 39 S.Ct. 59, 63 L. Ed. 166 (1918)], the owner impliedly warrants the sufficiency of the plans and specifications. An owner is responsible to the contractor for errors and omissions in the plans and specifications even if those errors and omissions do not constitute negligence. Thus, the public owner engaged in a tunnel construction project can find that it is responsible to its contractor for plan errors, but otherwise unable to recover from its design engineer unless he proves that the design engineer was negligent. Negligence is often difficult to prove since the owner is required to get another engineer to testify that the design engineer’s performance was below the standard of care for engineering work performed under like or similar conditions. One way to minimize this risk is to hire a design/builder to design and build the tunnel. Under the design/build project delivery method, the owner hires a design/builder for single point responsibility for the design and construction of the project. Unless the design/builder can show that the owner changed the scope of the project, the design/builder must absorb the cost of any errors and omissions in the plans and specifications. Thus, the design/build project delivery method may indeed reduce litigation over design errors.

the owner since the owner is usually in the best position to evaluate the risk and absorb the costs of such unforeseen conditions. A modified proposal of risk sharing, which deprives the contractor of his profit for Type II unforeseen conditions, may be reasonable. It is also fair to share the risk of owner-caused delays by limiting contractors to the recovery of their jobsite general conditions. Owners should also be limited to the recovery of liquidated as opposed to actual damages for contractor delays on the project. The contractor is clearly in the best position to control and absorb the risk of site injuries or death. An owner is well served to require the contractor to sign a broad form indemnification provision which protects not only the owner but his design professional from site safety claims. The wise public owner would do well to purchase an owner-controlled insurance policy that covers all of the parties involved in the tunnel construction project, including the design professionals, so as to minimize the potential for litigation over claims. Finally, although unusual, the public owner embarking on a tunnel construction project should consider whether to use a design/build project delivery method as opposed to the traditional design, bid, build method. By assigning single point responsibility to the design/builder, the owner may protect itself against a significant number of design error claims by the contractor.

6 CONCLUSION A contract that fairly allocates risk will necessarily minimize the risk of litigation. The risk of unforeseen subsurface conditions should typically be allocated to

176 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-22.qxd

2/18/04

12:30 PM

Page 177

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Getting back on-track: Exchange Place Station Improvements M.F. McNeilly Golder Associates Inc., Newark, New Jersey, USA

S.A. Leifer & G.F. Slattery The Port Authority of NY & NJ, Newark and Jersey City, New Jersey, USA

ABSTRACT: The September 11, 2001 attacks on NYC tragically resulted in the collapse of the World Trade Center (WTC) twin towers. In addition to the horrific lose of life and property damage, the collapse of the towers also destroyed the Port Authority Trans-Hudson (PATH) commuter rail station located within the foot print of the WTC site, flooded the two connecting Hudson River tunnels, and closed the Exchange Place Station in New Jersey. Immediately following these events, the Port Authority of NY & NJ (PANYNJ) recognized that restoration of PATH service to lower Manhattan would be vital to recovery efforts and the regional economy, and set in motion plans to rebuild the WTC Station and restore service by November 2003. This paper focuses on the Exchange Place Station Improvements Project, which required design and construction of underground rock excavations with spans upwards of 18 m (60 ft), rock cover as low as 7.4 m (25 ft), and buildings overlying planned excavations. This project also represents the first use of road-header type excavation equipment and the first application of steel fiber reinforced shotcrete (SFRS) linings for ground support on a mass-transit project in the NYC Metropolitan Region.

1 INTRODUCTION As a result of the September 11, 2001 terror attacks on the WTC towers, nearly 2,800 lives were lost and approximately 2.79 million m2 (30 million ft2) of commercial office space was either damaged or destroyed. The collapse of the towers also destroyed the Port Authority Trans-Hudson (PATH) commuter rail station, which occupied the lowermost levels of the WTC site, flooded the twin Hudson River tunnels carrying PATH’s downtown service, and forced the closing of the Exchange Place Station, the next station on the line. Loss of commuter rail service between the WTC and exchange Place Stations severed a vital transportation link between New Jersey and lower Manhattan affecting nearly 67,000 daily commuters. Commuters that previously disembarked at the WTC Station were forced to seek alternate routes including new/additional downtown ferry services and/or train/bus services to midtown Manhattan with connections to the NYC subway system. Immediately following the September 11, 2001 terror attacks, PANYNJ recognized that restoration of PATH service to lower Manhattan would be vital to recovery efforts and the regional economy, and needed to be completed as-soon-as-possible. Hence, PANYNJ

approved an ambitious program to restore PATH service to both the Exchange Place Station and the former WTC Station by July 2003 and November 2003, respectively. To achieve these fixed completion dates, it was apparent, from the beginning, that design and construction activities would have to be “fast-tracked”, and required the formation of somewhat unconventional partnerships between PANYNJ, contractors and numerous design consultants to expedite the project and mitigate project risks. This paper focuses on improvements to the Exchange Place Station, which required design and construction of underground rock excavations with spans upwards of 18 m (60 ft), rock cover as low as 7.4 m (25 ft), and buildings directly overlying planned excavation limits. This project also represents the first use of road-header type excavation equipment and the first application of steel fiber reinforced shotcrete (SFRS) linings for ground support on a mass-transit project in the NYC Metropolitan Region. 2 STATION IMPROVEMENTS Following the events of September 11, 2001, the Exchange Place Station was forced “out-of-service”

177 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-22.qxd

2/18/04

12:30 PM

Page 178

Figure 1. Existing tunnels and crossover tunnel plan.

because its configuration was originally designed to accommodate only “through” station and not “terminal” station traffic conditions. Hence, PANYNJ elected to construct a series of six (6) new track crossovers between the five (5) existing tunnels west of the station. In addition, PANYNJ elected to extend the station platforms about 46 m (150 ft) west to accommodate its planned 10-car capacity expansion program. See Figure 1.

3 HISTORY OF STATION Construction of the PATH system, by the Hudson Tunnel Railroad Company, started in 1874. Train service commenced in 1908, but the entire system was not operational until 1911. Exchange Place Station is located in Jersey City, New Jersey on the banks of the Hudson River, and is the first station in New Jersey on PATH’s downtown service to lower Manhattan. Originally, the twin Hudson River tubes comprising PATH’s downtown service connected the Exchange Place Station to the Hudson Terminal Station located beneath the former Hudson Terminal Building in lower Manhattan. However, the Hudson Terminal Building was demolished during the 1960s to make way for the WTC development, and the Hudson Terminal Station was abandoned in favor of the former WTC Station. Exchange Place Station and its connecting tunnels were constructed using drill-and-blast techniques, and recent field observations indicate that unsupported excavations were the preferred method of construction. Upon completion of the original rock excavations, the station and its connecting tunnels were lined with unreinforced concrete of variable thickness. See Figure 2.

Figure 2. Tunnels F to H to L Crossovers Circa 1907.

In general, the station and its connecting tunnel structures remained unaltered since commencing operations during the early 1900s. However, the station did undergo a major modernization sometime around 1986 to add/upgrade a new head house, ventilation towers, escalators and elevators from street-to-station levels. 4 CHALLENGES To achieve the project’s fixed completion date of July 2003, it was apparent that design and construction efforts had to be “fast-tracked” to re-open the station on-schedule, as-promised. This meant that tunnel excavation work had to commence by April 2002 and be completed by November 2002 to allow the other design disciplines adequate time to complete their parts of the project. This required PANYNJ to solicit and procure construction contracts prior to finalizing the project design.

178 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-22.qxd

2/18/04

12:30 PM

Page 179

Therefore, PANYNJ elected to implement a unique, unconventional contracting strategy to balance its associated project risks with the Contractors by putting the entire scope-of-work for Phase I of its Downtown Restoration Program (DRP) including the WTC temporary station, restoration of the two Hudson River tunnels, as well as the Exchange Place work under a single “net-cost-plus-fixed-fee” Contract. Contractors bidding on this contract were only provided general overviews of the scope-of-work, but were apprised of the project’s mandated schedule milestones. On February 1, 2002, this contract was awarded to a joint venture of Yonkers Contracting Co., Inc. of Yonkers, New York; Tully Construction Co., Inc. of Flushing, New York; and A.J. Pegno Construction Corp. of College Point, New York. Design and construction of the Exchange Place Station Improvements Project had to overcome the following technical challenges:

developed at these review meetings is as follows:

•

Based on the findings and recommendations, tunnel excavation drawings and technical specifications were generated and issued to the Contractor. In addition, these documents were revised, modified, expanded and re-issued to the Contractor, as necessary, under a series of design bulletins, as design and construction efforts advanced.

• • • • • •

Excavation of large underground rock caverns with spans upwards of 18 m (60 ft); Shallow rock cover as low as 7.4 m (25 ft); Localized areas of poor quality rock; Lower than expected rock mass strengths; Presence of multi-story buildings adjacent to and directly above planned excavation limits; narrowness of existing tunnel structures; and Limited/restricted site access.

5 PROJECT APPROACH From the start, the project’s highest priority was placed on development of construction alternatives and generation of Contract Documents. To this end, PANYNJ divided the project into nine (9) separate work order packages, based on the various design disciplines, and accepted an approach whereby documents were prepared and issued to the Contractor based on preliminary level designs with final designs advanced and completed concurrent with construction. In particular, the project’s requisite tunnel excavation activities fell under Work Order Package #3-EP. Contract Documents for this package were generated based on a series of geotechnical engineering design review meetings, which were convened to discuss and establish acceptable design and construction concepts based on a collective pooling of knowledge, experience and precedent. These design review meetings focused on ground support requirements, sequences of construction, methods of excavation and removal of muck materials, and alternate methods of site access. Consensus opinions among the meeting participants were formed and preliminary design recommendation established. A summary of the findings and preliminary recommendations

• • • • • •

6 DESIGN INVESTIGATIONS As part of the project’s design efforts, a series of geotechnical engineering design investigations were undertaken and completed to define subsurface soil and rock characteristics, establish relevant material properties and determine thicknesses and strengths of existing concrete tunnel linings. To define subsurface conditions, thirty-eight (38) vertical and/or inclined borings were drilled from both street-to-tunnel and tunnel-to-street levels. These borings were drilled immediately adjacent to and/or between the existing tunnels, and rock cores were collected using orientated and non-orientated drilling techniques. In addition, subsurface rock mapping was undertaken within each excavation heading, as tunnel excavations advanced. To determine thicknesses of existing concrete tunnel linings, one-hundred-two (102) cores were drilled through both tunnel sidewalls and crowns, and the concrete cores were collected for further strength testing. In addition, each corehole was videoed over its entire length. To establish relevant material properties, a laboratory testing program was conducted, including twentyfour (24) unconfined compressive strength (UCS) tests and sixteen (16) direct shear tests on selected rock core samples. In addition, sixteen (16) UCS tests were completed on selected concrete liner core samples.

179 Copyright © 2004 Taylor & Francis Group plc, London, UK

Excavation and support of the new crossovers would be feasible and constructible, but it was acknowledged that highly controlled excavation procedures and sequences would be required; Long-term ground support could be effectively provided using a combination of shotcrete linings, lattice girders and rock bolts; Portions of abandoned tunnels must be infilled with concrete to reduce excavated span widths; Rock cover thicknesses must be maximized to the greatest extent possible; Narrowness of existing tunnels will limit the type, size and amount of equipment that could work in the tunnels; and Removal of excavated materials will be difficult and may control production rates.

09096-22.qxd

2/18/04

12:30 PM

Page 180

7 SUBSURFACE AND TUNNEL CONDITIONS

•

Based on the findings and observations from the completed geotechnical engineering design investigation, subsurface and existing tunnel conditions can be characterized as follows:

•

• • • • • •

•

Concrete liner sidewall thicknesses varied from 0.3 to 1.5 m (1 to 5 ft), and were generally found to be “tight”; Concrete liner thicknesses at tunnel crowns varied from 0.25 to 0.5 m (0.75 to 1.5 ft), and voids upwards of 0.25 m (9 in) were observed; Overburden soils consisted of man-made fill, siltyclay, sand, silt and gravel deposits with a total overburden thicknesses that ranged from 4.8 to 9.1 m (16 to 30 ft); Groundwater was observed in the overburden at a depth of 2.5 to 3.5 m (8 to 12 feet) below ground surface; Manhattan Schist bedrock underlies the site. Average rock cover thicknesses were about 9 m (30 ft); Rock mass foliation was the observed dominant structural feature. Foliation dip angles ranged from sub-horizontal (less than 5 degrees) to 40 degrees, and dip directions varied from northeast to northwest; and Steeply dipping (vertical and sub-vertical) joints were uncommon and no measurable dip angles or directions were recorded from the collected rock cores. However, a series of steep joints were observed and mapped at tunnel level, as tunnel excavations advanced.

8 MATERIAL PROPERTIES 8.1

Overburden

Overburden materials were assumed to be cohesionless with the following properties: (a) friction angle of 30 degrees; (b) unit weight of 1,600 to 1,900 kg/m3 (100 to 120 pcf); (c) Poisson ratio of 0.3; and (d) modulus of elasticity of 150 MPa (21.7 ksi). 8.2

Bedrock

Manhattan Schist was encountered in all core borings, and consisted of light grey to dark grey, banded gneiss, schistose gneiss and/or schist intruded by pegmatite sills and dikes scattered throughout the rock mass. Material properties were established based on inspection of collected cores, geomechanical logging and laboratory testing. Relevant design values are as follows:

•

Total Core Recoveries (TCR) ranged from 50 to 100%, average 98%;

• • •

• •

8.3

Existing tunnel linings

Existing unreinforced concrete liners were assumed to have the following properties: (a) unit weight of 2,200 to 2,400 kg/m3 (140 to 150 pcf); (b) UCS of 21 to 31 MPa (3 to 4.5 ksi); and (c) modulus of elasticity of 2,068 to 2,758 MPa (300 to 400 ksi). Material properties for new concrete backfills used in the new crossover tunnel design were developed using American Concrete Institute (ACI) empirical equations. 8.4

Proposed ground support

Stability analyses incorporated galvanized, No. 9, Grade 75, resin grouted rock bolts, and all rock bolts were prestressed to 40% of the bar yield strength. Rock bolt lengths and spacing varied from 2.4 to 4.6 m (8 to 15 ft) and 1.5 to 1.2 m (5 to 4 ft), respectively, as excavated span lengths increased from 9 to 18 m (30 to 60 ft). In addition, new tunnel linings consisting of 34.5 MPa (5 ksi), high early strength steel fiber reinforced shotcrete (SFRS) having a thickness of 0.15 to 0.28 m (6 to 11 in) were constructed. Pre-fabricated steel lattice girders spaced on 1.5 m (5 ft) centers were also embedded within the shotcrete linings of excavated spans greater than 9 m (30 ft). Stability analyses assumed unit shear strengths of 1.5 MPa (219 psi) for SFRS materials, which was

180 Copyright © 2004 Taylor & Francis Group plc, London, UK

Rock Quality Designation (RQD) values ranged from 0 to 100%, average 86%, but typical RQD values varied between 50 and 100%; Rock mass unit weight, UCS and modulus of elasticity values ranged between 2,403 and 2,723 kg/m3 (150 and 170 pcf), 13.8 and 34.5 MPa (2 and 5 ksi) and 1,379 and 2,068 MPa (300 and 400 ksi), respectively; Rock Mass Ratings (RMR), (Bieniawski, 1976), ranged between 33 to 64; Rock Tunneling Quality Indices, Q-ratings, (Barton et al., 1974; Bieniawski, 1989) ranged between 2.7 and 15; Rock mass strength parameters were derived using the Hoek-Brown strength criteria (Hoek et al., 1998), and design values were as follows: a) “m” ranged between 1.31 and 3.95; and b) “s” ranged from 0.0018 and 0.0056; Spacing of foliation joints ranged between 0.2 and 0.6 m (0.5 and 2 ft), and these joints were assumed to have zero cohesion and friction angles of 20 to 23 degrees; and Spacing of steep (vertical, sub-vertical) joints crossing foliation were assumed to be 3 m (10 ft) or greater, and these joints were assumed to have zero cohesion and friction angles of 50 degrees.

09096-22.qxd

2/18/04

12:30 PM

Page 181

derived from correlations relating compressive and shear strength for plain and fiber reinforced concrete and shotcrete (ACI, 1984 and 1988; Fernandez et al., 1979; Mahar et al., 1975). These analyses also assumed composite unit shear strength of SFRS with embedded lattice girders equal to 2.7 MPa (391 psi), which was established in a similar manner except steel cross sectional areas were added to the SFRS cross-section. 9 DESIGN ANALYSES AND EVALUATIONS Design efforts included a combination of historical precedent evaluations and analyses to assess stress conditions around existing and new tunnel structures and evaluate the effectiveness of proposed rock reinforcement elements in two-dimensions for selected design cross-sections. 9.1

Historical precedent evaluations

Design of large underground rock excavations can be based largely on precedent by extending prior experience and construction observations from similar case histories to assess expected performances of specified ground supports. However, it should be acknowledged that reliance on precedent data must be used with caution. One such design reference (Cording et al., 1971) provides a compilation of case histories for underground rock excavations, and this reference presents relationships between excavation sizes (span and height) and installed rock bolt lengths and support pressures. Using this reference, rock bolt lengths and support pressures to excavated span ratios of 0.2 to 0.3 were found to be reasonable when compared to the case history data. Therefore, rock support pressures of 0.07 to 0.14 MPa (10 to 20 psi) and rock bolt lengths of 3.6 to 4.6 m (12 to 15 ft) were selected for spans of 15 to 18 m (50 to 60 ft). 9.2

UNWEDGE analyses

The computer program UNWEDGE (Rocscience, 2002, Version 2.37) was used to evaluate and define ground support requirements for each new crossover tunnel. This program provided three-dimensional visualizations of kinematically capable rock wedges, and calculated design factors-of-safety assuming the rock wedges are infinitely stiff, homogeneous masses acted upon by gravity, friction and applied internal support. Calculations were performed considering rock mass discontinuity and tunnel orientations for each new crossover, and results were used to evaluate required rock bolt lengths and spacing to achieve specified short-term (construction) and long-term

design criteria. In addition, sensitivity analyses were conducted to assess relative importance of variations in design parameters. Based on these UNWEDGE results, it was determined that calculated factors-of-safety are sensitive to excavated span lengths, strongly controlled by presence and location of steeply dipping discontinuities relative to excavation orientations and sensitive to applied hydrostatic pressures. In addition, it was determined that short-term (construction) design conditions were more critical than long-term design conditions. 9.3

The two-dimensional software program Phase2 (Rocscience, 2002, Version 5.0) was used to evaluate states-of-stress and stress field changes around existing and new crossover tunnels by modeling the rock mass as a continuum. The purpose of these models was to evaluate the potential for stress induced failures in both excavated crown and rib pillars. In addition, these models were used to assess stress changes within existing tunnel concrete liners. The modeling approach for these Phase2 stress analyses consisted of developing models to approximate state-of-stress conditions prior to excavation of the new crossover tunnels. After these initial conditions were developed, sequential excavations and/or construction of new concrete backfills were introduced into the model to simulate planned construction sequences, and rock bolts were added in areas requiring ground support. Results from these Phase2 stress analyses indicate that imposed incremental stress changes, due to the project’s crossover tunnel excavations, are relatively small when compared to in-situ rock strengths. In addition, compressive stresses that develop in adjacent rib pillars are moderate, but are well supported by adjacent concrete backfills and/or excavation support. 9.4

UDEC modeling

The two-dimensional computer code UDEC (Itasca, 1998, Version 3.0) was used on select design cross-sections to evaluate the stability of large excavated spans with shallow rock cover in a highly jointed and foliated rock mass. Conditions evaluated by these UDEC models consisted of the development of a 17 m (55 ft) wide span. Rock mass jointing consisted of foliation partings spaced 0.6 m (2 ft) apart and steep (vertical to subvertical) joints across foliation spaced 3 to 4.5 m (10 to 15 ft) apart. Based on these UDEC results, it was determined that the new crossover tunnels will be stable and the recommended rock bolt support will provide adequate short-term (construction) support.

181 Copyright © 2004 Taylor & Francis Group plc, London, UK

Phase2 modeling

09096-22.qxd

2/18/04

12:30 PM

Page 182

ROCKBOLTS INSTALLED AS EXCAVATIONS ADVANCE

PRE- SUPPORT ROCKBOLTS

SFRS LINING WITH EMBEDED LATTICE GIRDRS EXISTING CONCRETE LINING

Figure 3. Recommended ground support system.

9.5

Recommended ground support

Based on the results and findings of completed precedent evaluations, numerical modeling and stability analyses, the project’s recommended ground support included a combination of prestressed rock bolts, steel lattice girders and SFRS linings. See Figure 3 for typical design section showing recommended ground support.

Figure 4. Typical tunnel conditions and rock drill mounted on bobcat excavator.

10 CONSTRUCTION OBSERVATIONS Large projects such as the Exchange Place Improvements Project require integrated approaches to design and construction for them to be successful, and these efforts are complex enough when Contract Documents are prepared based on final designs. Given the unique nature of this project, it was apparent that unconventional methods were necessary to overcome what appeared to be “Herculean” schedule constraints. 10.1

Project coordination

To achieve the mandated scheduled milestones, an elevated, heightened degree of communication was required between the various parties. As the project progressed, all parties worked well together to achieve a common goal. As design efforts advanced, the need to modify previously issued Contract Documents became apparent and revised documents were re-issued under a series of design bulletins. The Contractor was made aware of each design bulletin before issuance, so its input could be incorporated to expedite construction. 10.2

Confined tunnel conditions

Inside clear dimensions of the existing tunnels were on the order of 4.3 m (14 ft), which limited the Contractor’s ability to utilize standard construction equipment. Hence, the Contractor was forced to think creatively, and adapt/modify available equipment to complete its work. See Figure 4 for photo of tunnel conditions and equipment used to install the specified ground support elements.

Figure 5. AM-50 road-header on track “F”.

10.3

182 Copyright © 2004 Taylor & Francis Group plc, London, UK

Tunnel excavations

At the project’s onset, it was envisioned that planned tunnel excavation activities would be undertaken and completed using controlled rock blasting techniques. However, the Contractor was unable to adequately control excavated perimeters, which resulted in excessive rock over-breakage and potential schedule slippages. Therefore, alternate methods of rock excavation were investigated, and road-header type equipment was considered to minimize rock over-breakage and accelerate the project’s tunnel excavation activities. See Figure 5. It should be noted that road-headers had never been used to excavate the local Manhattan Schist bedrock before consideration on this project. In addition, the Contractor and other industry professionals familiar with the project expressed reservations that roadheaders might not be well suited to excavate the rock.

09096-22.qxd

2/18/04

12:30 PM

Page 183

However, laboratory test results on collected rock cores indicated that intact rock UCS values typically ranged from 27.6 to 34.5 MPa (4 to 5 ksi). In addition, manufactures’ data for medium-sized road-headers, which could fit inside the existing tunnels, indicated that this equipment should be capable of excavating rock with UCS strengths in the range of 50 to 75 MPa (7 to 11 ksi). Therefore, PANYNJ and the Contractor elected to undertake a pilot program to assess the suitability of such equipment. This pilot program determined that road-header type equipment could excavate the local Manhattan schist rock. However, the machine used in the pilot program was found to be too small and experienced frequent breakdowns. Hence, the Contractor initially acquired a heavier ABM-330 Alpine Miner to start production excavation work, and later acquired two (2) additional VoestAlpine road-headers (AM-50 and AM-75) to complete the required excavations. As excavations advanced, the various road-headers used to complete the required tunnel excavations were found to be quite effective at excavating the rock, and production rates increased substantially over the previous drill-and-blast methodology, and the Contractor started to recover time on its schedule. Rock excavation rates were no longer hyper-critical to the project’s schedule. However, these roadheaders created new, additional challenges that had to be overcome, such as: (a) ventilation and water control issues; (b) greater volume of muck being generated than could be removed; and (c) need for closely controlled and carefully coordinated material handling sequences and procedures to reach the uppermost cavern crown limits. 10.4

Material handling and removal

The site was constrained by access from either end of the tunnels with the Contractor working simultaneously in five parallel tunnel structures. The Contractor was also confronted with PATH’s on-going commuter rail operations west of the site and the other two DRP projects (Hudson River Tunnel Rehabilitation and WTC Station Projects) east of the site. Consideration was given to constructing temporary access shafts from street-to-tunnel level to provided additional access to tunnel excavation work areas. In addition, combinations of belt and flight conveyor systems were considered to remove excavated muck materials through the existing station head house and onto barges in the Hudson River. However, both of these alternate access and material handling options were rejected in favor of more conventional rail haul methods utilizing only work trains operating out of a maintenance yard about 2.4 km (1.5 miles) west of the station.

Figure 6. Application of shotcrete.

This required coordination of work train movements with PATH’s on-going commuter rail operations, and necessitated restricting train movements to off-peak commuter hours. In addition, work trains could only use two of the five tunnels, which limited the number and frequency of work trains. 10.5

183 Copyright © 2004 Taylor & Francis Group plc, London, UK

Installation of ground support

Prior to start of tunnel excavation work, the Contractor installed pre-support rock bolts and completed contact grouting behind existing liners to provide continuous contact between the lining to remain and the rock. Presupport rock reinforcement consisted of 2.4 m (8 ft) long galvanized, No. 9, Grade 75, resin grouted rock bolts spaced on 1.5 m (5 ft) centers. The rock bolts were also pre-tensioned to 40% of the bar yield strength. As tunnel excavation advanced, the Contractor installed the specified rock reinforcement, which was similar to the pre-support rock bolts. Rock bolt lengths and spacing varied from 2.4 to 4.6 m (8 to 15 ft) and 1.5 to 1.2 m (5 to 4 ft), respectively, as excavated span lengths increased from 9 to 18 m (30 to 60 ft). This project included the first application of a steel fiber reinforced shotcrete (SFRS) tunnel liner system on a mass transit system project in the NYC metropolitan region. Nominal shotcrete thicknesses varied from 0.15 to 0.28 m (6 to 11 in), depending on excavated span lengths. In addition, SFRS linings were used in combination with pre-fabricated steel lattice girders spaced on 1.5 m (5 ft) centers and rock bolts installed perpendicular to the excavated rock face. SFRS linings were applied using “wet-mix” techniques, and steel fibers were incorporated into the mix design at the concrete batch plant to offset needs for welded wire fabric. This approach saved time during construction. SFRS materials were delivered using drop pipes from street-to-tunnel levels, and applied using nozzlemen and assistants operating from aerial man-lift equipment. See Figure 6.

09096-22.qxd

2/18/04

12:30 PM

Page 184

ACKNOWLEDGEMENTS The authors wish to acknowledge that the success of this project was not attributed to any one agency, company or individual, but rather the product of a combined partnership between the Owner, the Contractor and the numerous design consultants working on the project, and all parties involved should be proud to have assisted in re-opening of the Exchange Place Station on-schedule. The authors would also like to thank Bill Fetters and George Yoggi for providing selected images used in this paper.

REFERENCES Figure 7. Completed tracks F to H to L crossovers.

11 SUMMARY Design and construction of the Exchange Place Improvements Project required the development of six (6) new crossover tunnels having spans upwards of 18 m (60 ft) in width. This project also faced numerous contractual, technical and management challenges, which directly impacted the project’s overall critical path. Tunnel designs were advance and Contract Documents prepared on a “fast-tracked” basis (concurrent with construction), and these documents included controlled tunnel excavation sequences and procedures and installation details of pre-support and final ground support elements. To achieve the mandated schedule completion date, unique, unconventional partnerships were established between the PANYNJ, the Contractor and the various design consultants. This project was an example of how genuine partnership and cooperation between the various parties can accomplish a task some viewed as insurmountable. On June 29, 2003, the Exchange Place Station reopened to commuter service, and the new WTC Station is scheduled, as of this writing, to re-open in November 2003. Completing this project in two (2) years was a major accomplishment. See Figure 7 for view of completed tunnel crossovers.

ACI Committee 506, 1984, State-of-the-Art Report on Fiber Reinforced Shotcrete, ACI 506.1R-84, American Concrete Institute. ACI Committee 544, 1988, Design Considerations for Steel Fiber Reinforced Concrete, ACI Structural Journal, September–October, 1988, pp. 563–579. Barton, N.R., Lien, R., and Lunde, J., 1974, Engineering Classification of Rock Masses for the Design of Tunnel Support, Rock Mech., Vol. 6 No. 4, pp. 189–239. Bieniawski, Z.T., 1989, Engineering Rock Mass Classification, New York, John Wiley & Sons. Bieniawski, Z.T., 1976, The Geomechanics Classification in Rock Engineering Design, Proc. 4th Int. Congress on Rock Mech., ISRM Montreax, Vol. 2, pp. 41–48. Cording, E.J., Hendron, A.J., and Deere, D.U., 1971, Rock Engineering for Underground Caverns, Symposium on Underground Rock Chambers, ASCE. Fernandez, G.D., Cording, E.J., Mahar, J.W., and Van Sint Jan, M.L., 1979, Thin Shotcrete Linings in Loosening Rock, Rapid Excavation and Tunneling Conference, Vol. 1, pp. 790–813. Hoek, E., Kaiser, P.K., and Bawden, W.F., 1998, Support of Underground Excavations in Hard Rock, A.A. Balkema, Rotterdam. Itasca 1998, UDEC Users Manual, Itasca Consulting Group, Minnesota. Mahar, J.W., Parker, H.W., and Wuellner, W.W., 1975, Shotcrete Practice in Underground Construction Report No. FRA-OR&D 75–90, Washington, D.C., Federal Railroad Administration. Rocscience, 2002, UNWEDGE Users Manual, Rocscience, Inc., Toronto. Rocscience, 2002, Phase2 Users Manual, Rocscience, Inc., Toronto.

184 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-23.qxd

2/18/04

12:30 PM

Page 185

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Influence of geologic conditions on excavation methodology E.C. Wang, L.M. Hsia & C.C. Chang Parsons Brinckerhoff Quade & Douglas, Inc., New York

A.N. Shah MTA-NYCT, New York

ABSTRACT: The existing MTA-New York City Transit’s (MTA-NYCT) No. 7 Subway was built in the early 1900s and currently terminates at Times Square near 41st Street and Seventh Avenue. As part of the City’s redevelopment of the west side of Manhattan, the MTA-NYCT’s No. 7 Subway extension project will add approximately 1.5-miles of twin tunnel alignment to the current line. The proposed alignment will extend from the existing Times Square Station west beneath 41st Street then turn south under 11th Avenue and terminate between 24th and 25th Streets. The project will include two new stations, a two-track line station near 41st Street and 10th Avenue and a three-track terminal station at 34th Street and 11th Avenue. The alignment lies in close proximity to several major underground structures. The tunnels and station caverns will be constructed in the Manhattan Schist, which has been extensively intruded by pegmatitic and aplitic veins along and across the foliations. The lithology of the bedrock as well as the orientation and condition of the rock mass discontinuities are critical to the design of the tunnel boring machines. This paper will focus on the identification and influence of geology on the selection of excavation methodologies that minimize impact to the adjacent underground structures.

1 INTRODUCTION The proposed No. 7 Subway Line Extension is configured as a two-track subway extending approximately 1.5 miles of twin tunnel alignment, primarily in hard bedrock underlying New York City, westward from existing No. 7 subway tunnel at the Times Square Station between 7th and 8th Avenues below the Port Authority Bus Terminal Bus Ramp beneath 41st Street and turning with a 650-foot radius south to follow 11th Avenue. The alignment proceeds southward along 11th Avenue and terminates between 24th and 25th Streets. The alignment lies in close proximity to the following overlying structures: Eighth Avenue Subway, Lincoln Tunnel and Amtrak’s Empire Line and Hudson River Tunnels. Two future station locations are planned at 41st & 10th Avenue and 34th & 11th Avenue (see Figure 1). Subsurface geotechnical investigation is on-going to support the design and the EIS efforts. 2 GEOLOGIC SETTING The following represents the current working knowledge of local geology, combined with field observations

from investigation borings and subsequent review of recorded rock core samples. The project area forms a part of the Manhattan Prong of New England Upland Province. The hard crystalline metamorphic bedrocks form the ridges and valleys within this province. The project alignment mainly occupies the portion of western Manhattan along 41st Street from 8th Avenue to 11th Avenue and from 41st Street to 26th Street along 11th Avenue. A valley along 41st Street from 8th Avenue has been observed along the subsurface stream and further in the south between 30th and 27th Street, another valley is observed along another subsurface stream. The orientation of both these valleys is WNW-ESE.

2.1

185 Copyright © 2004 Taylor & Francis Group plc, London, UK

Overburden

The thickness of overburden deposits varies substantially. Based on the geotechnical investigation borings, relatively thin soil cover was encountered along 41st Street and generally increased both eastward toward 11th Avenue and southward towards 25th Street. The overburden consists of glacial and postglacial soils, combined with recent manmade fill.

09096-23.qxd

2/18/04

12:30 PM

Page 186

Figure 2. The Egbert L. Viele Map of 1874.

Figure 1. Site location plan.

Generally, the soils include glacial till, modified glacial drift, sand and gravels, some glacio-lacustrine silts and clays, and manmade land. The locations of old stream channels, exposed rock and marshland before the recent human intervention are recorded in the Egbert L. Viele Map of 1874 (see Figure 2). Review of these historic data reveals an apparent ancient creek which once flowed in a westerly direction, crossing 9th through 11th Avenues between 39th Street and 42nd Street, as well as between 29th and 32nd Streets; thus crossing and paralleling the future tunnel alignment. High water inflows are typically associated with the stream location; indicating zones of weakness in the rock mass (i.e., fault zones, joints, and fractures). 2.2

Bedrock geology

The geology of New York City is complex and has been studied and well documented in numerous publications. However, some of the observations made during the on-going subsurface investigation are noted herein. The crystalline rocks of New York City are divided into two major units separated by Cameron thrust, a regional NE-SW striking structural feature whose surface dips eastward. The rocks west of this line are called New York City Group or Manhattan Formation and the rocks east of this thrust fault are known as

Hartland Formation (Hutchinson River Group). The Cameron thrust fault extends from Connecticut through the east side of East River (between Roosevelt Island and Western Queens, through Staten Island and further south into central and southern New Jersey. This regional structural feature has also been called a “Suture”. The second prominent regional structural feature is the Manhattanville (125th Street) fault within Manhattan Formation striking WNW-ESE with low to moderate dip. A number of fractures, joints and faults (major as well as minor) have been observed along this orientation in the area. The project area comprises the youngest group of Manhattan Formation i.e., the Manhattan Schists. Although commonly referred to as schists, these crystalline schists vary in composition from quartzose schists, quartz-felspar schists, quartz-garnet mica schists, biotite schists, hornblende schists to muscovite schists. Numerous pre-, post- to late-kinematic (during Cameron thrust activity) pegmatitic intrusions of varying sizes (80 feet wide to a fraction of an inch) have been emplaced within these schists along and across the foliations and along fractures and joints in the area (see Photo 1). In certain places, mainly around midtown area of Manhattan, these intrusions have locally elevated the metamorphic grade and modified the texture of these schists which almost resemble aplitic, gneissic to granitic rocks. Serpentinite, talc and chlorite schists rocks were encountered in the borings between 27th and 28th

186 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-23.qxd

2/18/04

12:30 PM

Page 187

Photo 2. Deformation D3 – Regional foliation (S2) in Schist N10°E/S190°W. Minor folds shown by Schists.

Photo 1. Schist/pegmatite contact along foliation.

Street on 11th Avenue. These rocks seem to have been intruded along the fractures and fault planes during Late Taconic to Early Acadian orogeny. They are highly sheared and are seen as talcose schists, which display a distinct pulverized nature. Structurally, the rocks of Manhattan Formation have undergone multi deformational events. The imprints of these events can be seen from various discontinuities observed in the rock cores. The most prominent fold episode during the deformational event D3 has developed the main regional macroscopic antiformal (F2) structure in Manhattan Island, which has modified the cleavage/foliation (S1) generated by isoclinal recumbent folding (F1) during the second phase of deformation (D2). The orientations of cleavage/foliation (S2) during the deformational event D3 trend NNW to NE and dips at low to high angles to either W/NW to SW or E/NE to SE, according to the orientation of respective attitude of bed rocks due to the F2 folding. The orientations of cleavage/foliation and associated fractures and joints are classified according to deformational events which are described below: (a) D1 – the first foliation/bedding formed – (S0): orientation is obliterated. (b) D2 – the foliation/cleavage trends E/W to ENE/ WSW, highly obliterated, can only be seen near the fold cores (F1). (c) D3 – represents the main foliation/cleavage in the area trending NNW/SSE to NE/SW due to large

scale antiformal folding (F2), dipping either to the East or West (see Photo 2). (d) D4 – developed large scale shearing and fracturing due to the Cameron thrust faulting. Though this fault is not encountered in the area, the effect is seen in the rocks by shearing and chloritization (see Photo 3). (e) D5 – the rocks show moderate to steep plunging folds (F3). Foliations trend WNW ESE with the development of S3 cleavages (see Photo 4). These cleavages due to the extreme stresses resulted into slickensides, joints, fractures and faults on all scales (see Photo 5). This event marks the most important fault/fracture system in the area and is associated with the Manhattanville (125th Street) fault, which had been developed in this deformational phase (Middle Devonian–Acadian Orogeny). The resulting orientation and condition of these joint sets were evaluated with respect to the opening orientation and geometry using the Unwedge software program for stability analyses of potential wedge formations within the rock mass. These data are also being used in the final tunnel liner design. 2.3

Manhattan schist is generally not saturated, as groundwater is held in the open joints within a generally tight rock mass. Groundwater flow is controlled by the random interconnection of these more open joints as observed in deep excavation in and around the New York City. Groundwater levels measured in the soil

187 Copyright © 2004 Taylor & Francis Group plc, London, UK

Geohydrology

09096-23.qxd

2/18/04

12:30 PM

Page 188

Photo 5. Slickenside and stretched quartz veins along S2 & S3 cleavages.

Photo 3. Chloritization along foliation.

The network of fractures in bedrock will control the groundwater conditions in the rock mass. The permeability of the discontinuities will be influenced by several factors including the proximity of adjacent surfaces, alteration processes that have been removed or placed minerals on fracture surfaces, and joint wall material that has been fragmented or crushed by faulting or shearing. These geological processes can increase or decrease the permeability of individual joints. Water levels were measured in the observation wells installed in selected completed boreholes. The groundwater levels in the overburden do not appear to be sensitive to seasonal variation. Groundwater inflow during construction of running tunnel, station and crossover caverns, and construction shafts are being estimated using Packer test data. Photo 4. F3 folds in schists.

2.4 tend to follow the bedrock elevation, suggesting that they are probably perched on top of the bedrock, where rock forms a ridge and within the soil above the rock in bedrock valleys. The sources of groundwater recharge in Manhattan are leaking sewers, drains and water lines, and the adjacent East River and Hudson River. Though the recharge in the bedrock mass is unlikely to be from precipitation filtration due to the relatively impermeable nature of the city streets and buildings over land, there is evidence that the natural groundwater flow may be lateral following the old buried stream beds.

During the subsurface geotechnical investigation, which is still continuing, the following geologic features have been noted: (a) Between 25th and 29th Streets, there is a depression in bedrock surface such that the tunnel has minimum cover, and bedrock may be breached. (b) Between 25th and 29th Streets, there is an apparent shear zone, which contains serpentinite, talc schist and chlorite schist. (c) Talc schist and chlorite schist appears to be much weaker than the Manhattan schist.

188 Copyright © 2004 Taylor & Francis Group plc, London, UK

Distinct geologic features

09096-23.qxd

2/18/04

12:30 PM

Page 189

(d) Substantial soft compressible organic deposits underlying the fill encountered along 11th Avenue south of 28th Street.

3 EXISTING STRUCTURES IN THE VICINITY OF THE ALIGNMENT No. 7 Subway Line Extension construction impacts on the existing structures along the proposed tunnel alignment must be minimized. Nearby infrastructure includes NYCT structures, Port Authority Bus Terminal, Lincoln Tunnel East Approach tubes, Amtrak Empire Line and Hudson River Tunnels, the 11th Ave. Viaduct, the 8th Ave. Subway A, C, and E lines, as well as privately owned properties above or adjacent to the proposed alignment. The vertical separation from these underground structures was one of the important considerations in selecting the final tunnel alignment. Construction of the existing structures dates from the early turn of the 20th century to the 1950s, with expansions completed during the 1970s. The considerable amount of blasting that was necessary to excavate bedrock during construction of these overlying structures is expected to have altered the condition of the rock mass, and hence is being ascertained in the subsurface investigation program for the tunnel support design. The existing structures which may be impacted are as follows: (a) New York City Transit (NYCT) Facilities: 8th Ave. Subway A, C, and E lines – The active subway structure beneath 8th Avenue, consisting of an upper mezzanine level over four sets of tracks, which will be modified to accommodate the connection to the future tunnel extension. The proposed tunnels will be below these facilities. (b) The 41st Street Bus Ramp – This bus ramp is actually an underpass below 41st Street between Dyer Avenue and the west face of the Port Authority Bus Terminal Extension. This ramp provides underground access for buses entering the Terminal, and accommodates overhead ventilation system. (c) Lincoln Tunnel East Approach – This heavily traveled approach has three tunnels running in east-west direction along 38th and 39th Streets. Each tunnel tube has two traffic lanes. The proposed twin TBM tunnels will pass one tunnel diameter under these tubes. (d) Amtrak Empire Line – The Amtrak Empire Line originates from Penn Station, and continues as a tunnel extending to the west and southwest direction underneath 8th, 9th, 10th and 11th Avenues. At the proposed future 34th Street Station the tunnel turns abruptly north, crossing under 11th Avenue twice, and then proceeds beneath 35th

Street prior to day-lighting north of 37th Street. This 18-ft by 18-ft reinforced concrete box tunnel is founded on bedrock, and is above the proposed tunnel and the 34th Street station cavern. (e) Amtrak Hudson River Tunnel – The existing Hudson River Tunnel continues eastward from the Hudson River below Pier 62, underneath the Long Island Rail Road West Side Yards, and terminates at Penn Station. In the vicinity of 11th Avenue, the Hudson River Tunnel is consisted of a 19-ft span cast-in-place concrete twin tunnel with an elliptical brick arch. The proposed tunnel alignment is about two tunnel diameter above the proposed running tunnel, but is close to the proposed future 34th Station Street cavern. Although minimum one tunnel diameter separation was achievable for the running tunnel portion, such criteria will not be possible for the mined tunnel portion. Sections of the tunnel alignment where geology or nearby structures favor drill-and-blast excavation techniques will require specifications for controlled blasting methods to reduce overbreak and to minimize the vibration impact on existing structures. A comprehensive pre-construction condition assessment of all existing buildings and structures located within the influence zone of the new alignment will be conducted to establish baseline conditions. During construction geotechnical instrumentation will be installed to monitor ground movement at the streetlevel and within existing railroad and transit tunnels. The instrumentation data will be used to evaluate preconstruction assessments. 4 KEY ISSUES The key issues resulting from construction of the No. 7 Subway Line Extension project include the following. 4.1

The composition of ground (rock type) being excavated greatly affects the level of vibration from TBM’s. Mitigation of TBM vibration is generally through public information, reducing night-time operation or periods of long activity where possible. The mechanical cutting heads on Tunnel Boring Machines (TBM) typically generate ground vibration at frequencies between 20 to 80 Hz. However, occasionally, TBM’s have been reported to cause very low frequency motion (5 Hz) and it is believed that “machine-shaking” while mining through blocky or faulted ground caused the low-frequency ground motion. The intensity of low-frequency motion attenuates slower than motion occurring in higher frequencies, thus measured PPV values would be accordingly higher at the ground surface.

189 Copyright © 2004 Taylor & Francis Group plc, London, UK

Tunnel Boring Machines

09096-23.qxd

2/18/04

12:30 PM

Page 190

Vibration from operating TBM machines is influenced by distance, structural rock conditions, physical rock properties, ground water and depth of overburden. From the Chattahoochie Tunnel Project near Atlanta, vibration levels created by an 18.4-ft Robbins machine in granodiorite ranged from 0.02 to 0.08 in/sec. In this case the average tunnel depth was approx 200-ft below ground surface. When monitoring data was compared to overburden depths it was observed that PPV readings were inversely proportional to the depth of overburden. This may be explained by the well-known phenomenon whereby high-frequency attenuations occur more rapidly in loose deposits and energy loss due to reflection at the overburden-rock interface. Since the majority of Manhattan buildings are founded directly on bedrock and the depth of TBM tunnels are less than 200-ft, it may be assumed that the predicted PPV values will occur in the high-range of curves based on historical data. 4.2

Rock drilling

Hydraulic and pneumatic rock drills typically generate vibration occurring within a narrow band of frequencies ranging between 70 and 125 Hz. During quiet periods of the day, the resulting drill noise may sound like a hammer drill operating across the street. Throughout the alignment and at different times of the day, the ambient noise level will fluctuate; therefore it would be much more practical to limit groundborne noise increases to ambient levels plus 5 dBA. For the rock cover in Manhattan (generally 40 to 200ft), it is unlikely that ground-borne drill noise will increase ambient noise levels by more than 5 dBA. However, during quiet periods of the day it is likely that residents on ground floors or within basement areas will hear ground-borne drill noise. Also, despite occurring at very low intensity, this steady state noise will be “tonal”, meaning it occurs in a very narrow frequency spectrum (70 to 90 Hz). Public exposure to tonal noise generates more negative response compared to broadband noise. Inhabitants of buildings overlying the tunnels will hear and/or feel noise and vibration caused by TBM mining and rock drilling. Although, TBM induced vibrations are unlikely to damage structures; the persistent vibration and noise of TBM excavation may become a public annoyance, especially during quiet hours. 5 EXCAVATION METHODOLOGY The following description of potential excavation methodology along the alignment is based on the subsurface profile developed from the geotechnical

investigation program with the intent of minimizing construction impacts to overlying structures. Excavation of the running tunnels commencing from the Site A TBM Launch Shaft located along 11th Avenue between 25th and 26th Streets, and proceeding northward under 11th Avenue, turning eastward below 41st Street and terminating east of Ninth Avenue for retrieval from Shaft L located within the 10th Avenue Station footprint. Assembly and subsequent launching of the TBMs will require the following drill-and-blast structures: adit connecting off-line shaft to the tunnel alignment, assembly/launch chamber, back-shunt tunnel extending approximately 200ft south of the chamber, and TBM starter tunnels. Protection of nearby utilities along 11th Avenue may demand a rigid shaft excavation support system within overburden due to poor subsurface conditions. As such, Contract Specifications prohibit dewatering outside the excavation walls, in order to prevent reduction in pore water pressure, and initiating consolidation process in the compressible stratum underlying the fill. 5.1

Rock tunneling methods typically include: (a) Mechanical boring by Tunnel Boring Machine (TBM). (b) Drill-and-blast excavation/mining. (c) Mechanical excavation/ mining by machine (road header). All three tunneling options may be required to complete the tunnels, passages and caverns for the No. 7 Line Extension. 5.1.1 Mechanical boring by TBM Generally designed to perform all or most of the following functions: (a) Excavation of tunnel profile. (b) Control in the tunnel face. (c) Temporary support of the working area between the face and installed permanent lining by way of a steel shield usually around the full perimeter of the tunnel. (d) Installation of the permanent support and lining suitable for the final function of the tunnel. Three main types of TBMs, which are distinguished by the excavation and support method: (a) Open face machine – A road header or backhoe type excavator to excavate the face, which remains exposed and unsupported for long periods. Generally only used in soft rock or stable cohesive ground such as stiff clays and incorporate a shield to provide temporary support behind

190 Copyright © 2004 Taylor & Francis Group plc, London, UK

Tunneling methods in rock

09096-23.qxd

2/18/04

12:30 PM

Page 191

the face area. This type of machine is unsuited for the hard rock. (b) Full-face machine – Incorporates a full cutter-head, which cuts the full face area of the tunnel but does not necessarily provide support to the face. The cutter-head includes excavation systems of rotating discs or spokes with cutters. The excavated material falls through the face discs apertures to the inside of the cutter-head where it passes via chutes to the conveyor systems for removal. This TBM is designed for stable hard rock requiring minimal rock reinforcement. (c) Closed face machines – similar to full-face machines but possess the added capability to control and support the tunnel face. They are used to bore through weak, unstable and typically saturated ground. The Earth Pressure Balance Machine (EPBM) and Slurry TBM are examples of closed face machines. Of the above, a full-face hard rock TBM is expected to handle the majority of the main running tunnels excavation. The two main types of hard rock tunneling machines are identified by their means of propulsion:

•

•

Gripper-type TBM – such machines incorporate grippers, which are expanded against the tunnel rock walls at the rear of the machine providing the required reaction for the TBM cutter-head thrust at the face. Hence, the rock wall and the pillar between adjacent tunnels must be of sufficient strength to support the grippers. The gripper machines normally only feature a short slotted mechanical roof at the tunnel crown for protection of the operating crew. The slots within the mechanical roof provide exposure to the rock for rock bolt installation. Also, the machines may incorporate an erector system for placing segmental linings for permanent support. Typically however, they are independent from the permanent lining operation. In stable rock conditions, where tunnel lining is installed separately, impressive advance rates have been recorded. Shield-type TBM – Use a full shield to provide full perimeter support behind the cutter-head. Propulsion of the TBM is provided by rams thrusting off the last ring of the installed permanent segmental lining, which is placed under protection of the shield at the rear of the machine. Since the erection of the permanent segmental lining is required for the operation of these machines, construction rates are generally lower than the gripper-type machines. But, installation of the temporary or primary support is unnecessary for shield type machines. These machines are suited for a fractured hard rock environment, which might otherwise require considerable temporary support.

Selection of the most suitable machine will depend on the quality and fracture state of the rock mass, in addition to the type of permanent tunnel support selected. If competent rock possessing good stand-up time is encountered, then a gripper TBM will be preferred. Rock bolts would provide temporary support of tunnel walls, with optional shotcrete, as necessary. In this case, a cast-in-place (CIP) permanent lining could then be placed independent of the tunneling operations – either during tunneling or upon completion of the excavation. High progress rates would be expected under these conditions. For highly fractured rock mass a shield-type machine may be preferred to avoid installation of excessive temporary support. TBMs incorporating the segmental lining erector require a relatively long train of equipment. This requirement impacts the length of the TBM launch chamber required, as well as affecting the minimum turning radius of the machine. A larger construction laydown area may be required to accommodate stockpiles of liner segments near the shaft. 5.1.2 Drill-and-blast Drill-and-blast techniques are anticipated for cavern enlargements of initial twin TBM tunnel sections, as well as excavation of cross-passages and adits/ chambers associated with TBM launching and retrieval operations. Modern controlled blasting methods attempt to minimize blast vibration and noise. Reducing the blast disturbance of the surrounding ground is not only required to avoid disturbance to existing structures and building occupants within the immediate construction zone, but is also necessary for preserving the inherent strength of the surrounding ground. Blasting loosens the rock mass immediately surrounding the blast area. Excessive blast impact on the ground surrounding the tunnel could lead to loss of rock strength and reduction of stand-up time as well as excessive overbreak, resulting in increased support requirements to stabilize the underground opening. Smooth blasting techniques, line drilling, reduced excavation size and excavation round length, sequential excavation all provide means of limiting blast vibration and disturbance to adjacent ground and structures. 5.1.3 Mechanical excavator Hard rock applications typically restricted to fractured rock as well as circumstances, which preclude alternate techniques such as drill-and-blast. For extremely hard and very strong rock, in situations where blasting is not permitted expansive chemical products may be used to split rock between pre-drilled holes. Several types of mechanical excavators include: (a) Road header excavator – over the past decade, the cutting capacity of road headers have improved

191 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-23.qxd

2/18/04

12:30 PM

Page 192

substantially. However, the successful and economic performance of road headers in the strong Manhattan Schist will be governed by the degree of rock mass jointing and fracturing. Road headers may be used for sections of tunnel located close to the top of rock where weathering and loosened rock mass conditions are anticipated. Road headers may also be favored compared to blasting in sections of tunnel near to existing structures or utilities. In addition, the construction sequencing of the tunnels may require road header use to prevent damaging previously excavated tunnel sections. Road headers may be needed to mine the tunnels beneath the Port Authority Bus Terminal (PABT) underpass where relatively shallow rock cover exists. They may also provide an effective means of excavating sections of the relatively complex network of interlocking tunnels. (b) Hoe Rams – Hoe rams mounted on excavators will be restricted to relatively weak and/or highly fractured rock. Hoe rams will not be utilized for a full-scale tunnel excavation but may be favored for trimming tight spots and profiling work.

and its design of cutters, grippers, power requirement, and exploration equipment at heading. Serpentinite, talc and chlorite schist may further influence the design and construction due to their considerably distinct properties relative to the host rocks: Manhattan schist and pegmatite.

7 LIMITATIONS Any views or opinions presented in this paper are solely those of the authors and do not necessarily represent those of their companies, their employers or their subsidiaries. ACKNOWLEDGEMENTS The authors gratefully acknowledge P. McGrade, S. Singh, and M. Naik (MTA-NYCT) for their permission and D. Donatelli and P. Das (of PB Team) for their cooperation and support to publish this paper.

BIBLIOGRAPHY 6 CONCLUSION Geology is critical to the planning of the project tunnel alignment. The geologic conditions influence the planning and design process, from selection of alignment, to initial excavation support, tunnel final lining, through to the long-term operations. The subsurface conditions may be altered from the previous underground construction in the vicinity of the tunnel alignment. The lithology and the structural discontinuities as well as existing structures located within the tunneling influence zone will influence the selection of TBM

Baskerville, C.A., 1982. The foundation geology of New York City: Geological Society of America. Reviews in Engineering Geology. Vol. 5, pp. 95–117. Fluhr, T.W., 1941. The Geology of the Lincoln Tunnel Part 4: Journal of Rocks and Minerals Association. Vol. 16, No. 7, pp. 235–239. Isachsen, Y.W., 1980. Continental Collisions and Ancient Volcanoes, the Geology 9 of Southeastern New York: New York State Geological Survey, Educational leaflet, Vol. 24, pp. 1–15. Shah, A.N. et al. Geological Hazards in the Consideration of Design and Construction Activities of the New York City Area, Environmental and Engineering Geoscience, Vol. IV, No. 4, Winter1998, pp. 524–533.

192 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-24.qxd

2/18/04

12:31 PM

Page 193

Session 2, Track 3 Non-mechanized construction

Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-24.qxd

2/18/04

12:31 PM

Page 195

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Santiago’s Metro expands C.H. Mercado Metro de Santiago, Santiago, Chile

G.S. Chamorro Empresa de Ingeneria Ingendesa S.A., Santiago, Chile

K. Egger Dr. G. Sauer Corporation, New York, United States

ABSTRACT: Metro de Santiago has planned an ambitious array of new projects for the period of 2000 through 2005, which will almost double the reach of Santiago’s subway by late 2005 and bring service to an additional 1.8 million people. Metro is currently operating three rail lines with a total length of about 40 km, of which 23 km are below grade. New projects include the extension of two existing lines and construction of two new lines, which will increase the number of stations from currently 48 to 80. The subway operated by Metro S.A. is the backbone of the Urban Transportation Plan for Santiago (UTPS) which includes the extension of Line 2 to the north and to the south and Line 5 to the west, and foresees construction of Line 4 and Line 4A. All new or extended lines will be connected to bus transfer stations along the Metro system and, in addition, Line 5 will interface with a suburban railway line at a new intermodal terminal. 1 HISTORY Construction at Metro Santiago started in the mid 1970s with underground portions mainly in cut and cover boxes. By the 1980s, increasing public objection to the surface disruption forced Metro to investigate less intrusive alternatives, such as mined tunnels. The subsequent development from cut and cover construction to mined running tunnels and mined stations has to be described as rapid. In 1993 the design of the initial Line 5 Extension included the first mined tunnel experiment, a 2.0 km long running tunnel portion under Bustamante Park. The tunnel design was produced by Ingendesa, a Chilean engineering firm with tunnel expertise gained from designing hydro electric power plants. The experiment proved to be successful however, at the second Line 5 Extension in 1997, a 2.8 km continuation to the west the stakes for Metro were much higher. This time the alignment was planned beneath a heavily frequented street and adjacent to the city’s 300 year old cathedral and other historic buildings. To minimize the risk Metro required the Chilean design firm Cade Idepe to utilize foreign tunnel expertise. This expertise was provided by Geoconsult of Austria, which developed a comprehensive design according to the principles of the New Austrian Tunneling Method (NATM) for the running tunnels. While the

running tunnels for this Line 5 Extension were mined, cut and cover construction was still foreseen at the stations. With improved NATM techniques it even became viable and cost effective to mine stations which further reduced the environmental impact and added alignment flexibility. Therefore, based on a preliminary design by the Cade Idepe/Geoconsult team detailed mined tunnel designs for running tunnels and station tunnels for the extensions of Line 2 and Line 5 were developed in 2001. The mined approach was also chosen for the underground portions of the new Line 4 and a second extension of Line 2 to the north. 2 EXTENSION OF EXISTING LINES Currently Metro Santiago operates three lines, Line 1, Line 2 and Line 5. Line 4 currently under construction will be part of the expansion plan to be completed by the end of 2005. Line 3, planned to run parallel to the south of Line 1, was part of an earlier urban transportation plan which since has been revised due to the rapid development of the city. The extension of Line 5 to the west of Santiago covers an underground length of 1.9 km along Catedral Street and will feature two new stations. Quinta Normal Station, the western intermodal terminal will allow

195 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-24.qxd

2/18/04

12:31 PM

Page 196

Figure 1. Metro system map.

passengers to transfer to buses and a suburban railway line which runs in a north to south direction below Matucana Street. The neighborhoods along this line extension feature historic buildings with architectural and heritage value. Construction at Line 5 Extension commenced in May of 2001 and the completion of the work is expected by the end of 2003, with service to start by March 2004. The initial extension of the existing Line 2 to the north has an underground length of 2.2 km and two new stations. Running below Recoleta Street, the first extension crosses under the Mapocho River and Costanera Norte, a major urban highway under construction. Aside from preliminary work at shafts and access tunnels, construction for this line extension began in May 2001 and will be completed by mid 2004 with service expected to start in August of 2004. A second extension of Line 2 to the north is currently under design and will add 5.1 km of underground rail and five stations to the system. The new section extends Line 2 to Avenida Américo Vespucio, a major ring road around Santiago and construction is planned to start in late spring of 2004. Besides the addition to the north, Line 2 will also be extended 2.3 km to the south, near the merging point of Gran Avenida José Miguel Carrera and the southern

Figure 2. Area of Quinta normal station.

portion of Avenida Américo Vespucio. Two new stations will be constructed along the southern extension of Line 2. La Cisterna Station, the south terminal station, will be the link to Line 4, Metro’s future expansion line currently under construction. Construction for the extension began in January 2002 and is expected to be completed by December 2004.

196 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-24.qxd

2/18/04

12:31 PM

Page 197

The preliminary and detailed designs for the line extensions were produced by the design team of Ingendesa, Arze, Reciné y Asociados Ingenieros Consultores (ARA), two local engineering firms, and the Dr. G. Sauer Corporation (DSC) of Herndon, Virginia, as their expert foreign tunnel consultant, and the Cade Idepe/Geoconsult team. 3 NEW LINES The planned two new lines for Metro’s system will connect ten communities in the east and south of Greater Santiago through a total length of 33 km of underground, at grade and elevated rail. Line 4, which extends from the existing Tobalaba Station on Line 1 to Puente Alto in the south, includes a 7.9 km long tunnel section from Tobalaba Station to Rotonda Grecia. The 6.5 km section from Rotonda Grecia to Avenida Vicuña Mackenna will be constructed at grade along the median of Avenida Américo Vespucio and a 0.6 km long over-under tunnel section. From thereon the line will be constructed to Puente Alto mostly elevated along the alignment of Avenida Vicuña Mackenna with some at-grade and underground portions. The new Line 4 will also extend westward at Station Vicuna Mackenna to end at the termination of the southern extension of Line 2, La Cisterna Station. This branch, to be called Line 4A, will run at grade along the median of Avenida Américo Vespucio with a length of 7.9 km. Detailed designs for the underground portions of Line 4 were completed in July 2003. Three teams, Arcadis Geotecnica/Bureau de Projectos, Cade Idepe/ Geoconsult and Ingendesa/ARA/DSC were involved in the designs of mined running tunnels and underground stations.

4.2

Running tunnels

Metro’s running tunnels are single tube, double track tunnels with an average cross section size between 60 m2 and 65 m2. Tunnel size, stable ground conditions and an overburden of approximately 9 to 11 m allow for a full face excavation with a substantial earth wedge used for face stabilization as well as working platform during profiling, lattice girder and wire mesh installation and shotcrete application. The average excavation round length is 1.0 m with primary support comprising of 200 mm initial shotcrete with lattice girder and wire mesh reinforcement followed by a secondary layer of shotcrete, 150 mm thick, reinforced with wire mesh or rebar. Using data and experience gained on the Line 5 Extension the running tunnel linings have been reduced from a thickness of 500 mm to the proposed 350 mm to be used on the new line and line extensions, a savings of 30%. The shotcrete used is typically dry shotcrete, however, wet shotcrete has been used at the Line 2 North Extension and is currently in use at Line 4. Steel fibers in lieu of wire mesh or rebar reinforcement have been considered at Line 4. 4.3

Station tunnels

With a cross section size of up to 150 m2 and an overburden of as low as 7 m the new station tunnels on the

4 DESIGN AND CONSTRUCTION 4.1

similar properties as found in the second Mapocho Deposit. The groundwater table is higher at approximately 20 m depth. The sediments deposited by erosive streams from the Cordillera de Los Andes consist mainly of clay and silt with low to moderate plasticity and sand lenses of variable sizes. The material is partially saturated (30% to 75%) with the groundwater table within 20 m of the surface. This deposit is interlocked with the ripio in a saw-tooth pattern along a contact line in north to south orientation. About 4.0 km of the underground portion of Line 4 will be located in this formation.

Geology

Metro’s existing and new lines lie in quaternary sediments of gravel, the so called “Grava de Santiago” or “Ripio de Santiago” (Ripio) and locally contain deposits of over-consolidated clays. In the north those sediments originated from the Mapocho River, while in the south the sediment’s origin is the Maipo River. In the east sediments were deposited by erosive streams from ravines of the Cordillera de Los Andes. In the Mapocho Deposit a superior 4.5 to 6.5 m thick stratum of fluvial origin is followed by a stratum of fluvial-glacial deposits with similar gradation except the presence of plastic fines and a somewhat bigger compactness and cohesion. The groundwater level is variable, but generally located at approximately 80 m depth. The Maipo Deposit is of fluvial origin and with

Figure 3. Running tunnel during track work.

197 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-24.qxd

2/18/04

12:31 PM

Page 198

Santiago Metro produce some impressive underground structures. This impression is enhanced by perpendicular intersections with access tunnels of similar size. The design approaches to these massive caverns differ for each design team at the line extensions and also require an approach adapted to the clay dominated deposits at Line 4’s Plaza Egaña Station and Los Orientales Station. There the final design team, Cade Idepe/Geoconsult has developed a binocular configuration for the 140 m long station tunnels with a center pilot tunnel in which permanent central support columns are cast prior to the excavation of the station tunnels on each side. On the extensions of Line 2 and Line 5 in the good quality ripio above the ground water table, the excavation of the station tunnels follows a top heading and bench/invert sequence with two side drift tunnels and a central gallery. At the El Parrón Station on the Line 2 South Extension the Ingendesa/ARA/DSC design team applied experience gained on the Line 5 Extension and modified the design to a single sidewall drift sequence. This approach was also applied to the design of Las Mercedes Station at Line 4 and was taken over by the Arcadis Geotecnica/Bureau de Projectos design team for their station detailed designs on the new line. The access tunnels at the line extensions were designed by Cade Idepe/Geoconsult as part of their preliminary design work with a full span top heading and bench/invert sequence. The wide span however requires pre-support using grouted pipe spiling and an earth wedge providing face support. Besides the grouted pipe spiling on the full span top heading excavation and grouted pre-spiling, using self drilling bolts, for break-out situations from an access tunnel into a station tunnel, no significant pre-support is necessary in the ripio when appropriate excavation and support sequences are used. During excavation according to the designed excavation sequence, station tunnels just like the running tunnels generally receive a primary lining consisting of lattice girders, wire mesh and shotcrete of 300 mm thickness. This initial support is followed by a secondary, 200 mm thick shotcrete lining, to arrive at a combined thickness of 500 mm for the Line 2 South Extension stations. No integrated waterproofing systems have been considered by Metro so far for either running tunnel or station tunnel designs. 4.4

Figure 4. Double side wall drift station excavation.

Geotechnical instrumentation monitoring

Geotechnical instrumentation of the running tunnels comprises mainly of five point in-tunnel monitoring cross sections and seven point surface settlement sections at the same station. The in-tunnel convergence points and roof leveling points are read in three dimensions according to a specified reading schedule which considers the actual location of excavation face and

Figure 5. Single side wall drift station excavation.

198 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-24.qxd

2/18/04

12:31 PM

Page 199

work progress. The monitoring cross sections are laid out at 25 m intervals along the alignment, which with today’s extensive tunneling experience in the ripio is based on liability considerations rather than on geotechnical necessity. Monitoring points are also placed on sensitive structures within the zone of influence along the alignment. At station tunnels the monitoring point configuration within a monitoring cross section varies based on excavation approach. In addition to deformation monitoring, strains and stresses on and in the shotcrete linings are measured using strain gages, ground pressure cells and concrete pressure cells. As part of an optimization effort of the whole instrumentation program, the use of sliding micrometers, extensometers and inclinometers was drastically reduced, once sufficient information and data regarding the behavior of the ripio was available. This effort has reduced the cost for instrumentation and monitoring to approximately $250,000 per km running tunnel. The readings are taken by an independent geotechnical consultant, who transmits the processed data to the design engineers, construction supervision and Metro’s project management team for further analysis. The processed data is then compared to threshold values and trigger values established by the engineers during the design. These values, although set slightly different by the individual design teams are based on mining sequence, geology and surface developments. The set limits of approximately 15 mm for surface deformation above the station tunnel center line or 12 mm for a roof leveling point on the station tunnel center line have so far not been exceeded during the ongoing line extension works.

4.5

Metro’s administrative set-up for construction contracts is predominantly design-bid-build. The design work generally starts with a preliminary design, which for reasons of time savings includes preparatory work such as the detailed designs for shafts and access tunnels. Detailed designers pick up at an approximate 30% level and remain involved throughout the construction process by furnishing construction supervision, which includes a monitoring engineer responsible for the interpretation of gathered and processed geotechnical instrumentation data. Based on the monitoring engineer’s interpretation of the data and in coordination with the design team field modifications to the design can be made, if required. Another task of the construction

Figure 7. Quinta Normal Station excavation.

Figure 6. Surface settlements at Quinta Normal Station.

199 Copyright © 2004 Taylor & Francis Group plc, London, UK

Execution of work and supervision

09096-24.qxd

2/18/04

12:31 PM

Page 200

supervision is to verify design compliance and to respond to contractor’s questions regarding the design. To supervise the actual construction activities Metro awards construction management and inspection contracts for the individual line sections. The construction management and inspection work includes quality assurance tasks, enforcement of safety requirements as well as tracking of construction progress and construction cost. The technical documentation and coordination with the independent geotechnical consultant regarding access to the tunnel for monitoring purposes are also responsibilities of the construction management and inspection team. To date construction of Metro sections is mainly carried out by local contractors and Chilean subsidiaries of large foreign companies, such as Sacyr of Spain and Mendes Junior of Brazil. The quick adaptation by local contractors from cut and cover construction to mined tunneling methods is encouraging and current excavation rates of 3.0 to 4.0 m per 24 hour day in the running tunnels are remarkable. The ground with its extended unsupported stand-up time could allow excavation rounds beyond the current 1.0 m to expedite the construction. However, the use of specialized currently not utilized tunnel equipment, such as excavators with articulating arms, man-lifts and a more efficient mucking operation would increase productivity without compromising the current low level of risk. The use of fully mechanized excavation equipment was considered for the Line 4 construction and several studies were commissioned by Metro to investigate the technical and economical feasibility of Tunnel Boring Machines (TBM) and Earth Pressure Balance Machines (EPBM) for the clay zones and water bearing zones near the San Carlos Canal in the east and the Maipo River in the south. All studies concluded that while technically feasible there is no cost or schedule advantage using fully mechanized excavation equipment on the 7.9 km long underground stretch from the existing Station Tobalaba to Station Rotonda Grecia or elsewhere along Line 4. In fact the studies concluded that there is a considerable lower risk to the schedule using NATM. 5 CONSTRUCTION COST One of the reasons for the success of NATM on Santiago’s Metro was the cost savings realized after its introduction. While the first tunnel section under Bustamante Park proved that mined tunnels are technically feasible and caused less disruption to the city, the construction cost were approximately 20% higher compared with cut and cover structures, not considering costs for utility relocation and expropriation. Today, after further improvements to the NATM designs, the

construction cost average about $6,500 per linear meter of running tunnel and about $27,500 per linear meter of station tunnel. At present, the cost of the new Line 5 Extension is about 40% less compared to the Line 5 Extension built in the late 1990s. The reasons for the reduction in cost are more favorable ground conditions, lower rise buildings along the alignment and most important the conversion of expensive and disruptive cut and cover stations into mined stations. Metro hopes to reduce this already low tunneling cost even further during the construction of the new Line 4 by introducing value engineering as a tool for the contractors to optimize the designs. 6 CONCLUSION Not enough can be said about the fast development of Santiago’s Metro from cut and cover construction in the middle of an almost 5 million people metropolis to mined tunnels at reduced cost. Within approximately 10 years the local design and construction community with support of foreign experts has learned to develop and execute complex NATM designs for large underground spaces. There is no argument that the favorable ground conditions and a low water table have helped engineers and contractors in their swift learning process. However, one has to laude Metro’s balanced approach between their desire to reduce construction cost and their willingness to take risk. While the replacement of running tunnels with cut and cover boxes by mined tunnels in the early 1990s was a successful first step, Metro’s managers understood that with foreign tunnel expertise even bigger steps could be made. Thus in 1997 foreign NATM expertise was brought in for the design of the running tunnels during the second Line 5 Extension. This proved to be just the beginning of a prolonged success for the sequential excavation method on Santiago’s Metro, where today all underground work is carried out using this method. Advantages such as lower cost and lower risk gave NATM also the edge over mechanized excavation methods, which were considered for the new Line 4 construction. With room for even further improvement of productivity through efficient construction equipment, one can only look forward to future line extensions and new lines in Santiago. REFERENCES Wallis, S., 2003a. “Evolving NATM for Santiago’s Metro”, T&T International, March 2003. Wallis, S., 2003b. “Metro’s evolution”, T&T International, April 2003. Mercado, C., 2003. “Tecnologia sobre Rieles”, Revista Bit, September 2003.

200 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-25.qxd

2/18/04

12:32 PM

Page 201

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Benchmark for the future: the largest SEM soft ground tunnels in the United States for the Beacon Hill Station in Seattle J. Laubbichler, T. Schwind & G. Urschitz Dr. G. Sauer Corporation

ABSTRACT: Sound Transit’s Beacon Hill Light Rail Station in Seattle comprises the largest SEM soft ground tunnels in the United States to date and will be constructed in highly variable glacial soils with multiple groundwater horizons. The design is based on SEM principles, experience from previous projects, engineering judgment and numerical analyses, and provides for the flexible application of various pre-defined support measures (SEM Toolbox items) to cope with variable ground conditions, assure the most economic construction and minimize risks. A Test Shaft Program with the purpose of gaining additional geological information and confirming design assumptions was carried out. Findings in regard to ground behavior and groundwater were implemented in the design. To reduce construction risks, the owner decided to retain the SEM designers to provide SEM supervision and construction support services.

1 INTRODUCTION 1.1

Project overview

The Beacon Hill Tunnels and Station are part of the 14 mile initial segment of the Sound Transit Central Link Light Rail Line that will establish a high capacity commuter connection from downtown Seattle to Tacoma. The 4,300 foot running tunnel under Beacon Hill will be mined by Earth Pressure Balance Machine (EPB), while the deep mined Station will be constructed using slurry walls and the New Austrian Tunneling Method (NATM), referred to as Sequential Excavation Method (SEM) for this project. Figure 1. Station arrangement.

1.2

Beacon Hill Station arrangement

From the Station Headhouse, a 181 ft deep, 46 ft inner diameter Main Shaft will be constructed that will house four high speed elevators, emergency staircases, ventilation shafts and mechanical and electrical equipment. A 26 ft inner diameter Ancillary Shaft will accommodate another set of emergency staircases and ventilation shafts. From the Main Shaft, the 41 ft wide Concourse Cross Adit will provide passenger and emergency access to the Platform Tunnels. These are 380 ft long by 32 ft wide and were designed to accommodate the platforms, artwork and architectural finishes, and the light rail tracks. Two Cross Adits will connect the Platform Tunnels, and Ventilation Tunnels

will provide air flow in normal operation and for emergencies. 1.3

The Hatch Mott McDonald/Jacobs (HMMJ) Joint Venture is the lead designer for the Beacon Hill Tunnels and Station, the architectural design is carried out by Otak. The Dr. G. Sauer Corporation (DSC) provides the SEM design as a subconsultant for the Concourse Cross Adit, the Platform Tunnels and the Platform Cross Adits; Shafts, Ventilation Tunnels and Running Tunnels are designed by HMMJ.

201 Copyright © 2004 Taylor & Francis Group plc, London, UK

Design responsibilities

09096-25.qxd

2/18/04

12:32 PM

Page 202

2 GEOLOGY 2.1

Geologic setting

The Puget Sound Area is characterized by a complex mixture of glacial and non-glacial soils that have been deposited, consolidated, eroded and reworked by multiple major glaciations during the Pleistocene Epoch and numerous seismic events. Beacon Hill is an approximately 300-ft high ridge that is composed of holocene, vashon and pre-vashon deposits. An extensive subsurface exploration program has been conducted by Shannon & Wilson, Inc. of Seattle. In the course of this program, more than 70 investigation borings were drilled using Hollow Stem Auger, Mud Rotary, Triple-tube Rotary Core and Sonic Core techniques. Laboratory testing of the recovered soil samples was carried out and engineering properties were derived accordingly. The exploration program showed that most of the Beacon Hill Station will be excavated within in glacial, overconsolidated, partly fractured or slickensided clays and tills. Intermittent sand and silt layers will be present with multiple perched groundwater horizons. The Seattle Bremerton Fault zone is expected to be the cause for some of the inconsistencies, inclinations and fractures observed during the geotechnical investigation. 2.2

Ground classification and ground behavior

For design purposes, the soils were grouped into classes according to their engineering parameters and anticipated ground behavior during tunneling: Class 1: Loose to dense granular deposits This soil type consists of poorly graded sand and gravelly sand; it will be encountered when excavating the Headhouses and not be of concern for tunneling.

Class 2: Soft to Very Stiff Clay and Silt This soil type comprises normally consolidated clays, and silty clays and clayey silts; it will be encountered when excavating the Headhouses and not be of concern for tunneling. Class 3: Till and Till Like Deposits Heterogeneous mixtures of gravel, sand, and silt or clay; they will be encountered in the station shafts and in sections of the station tunnels. These soils have a compressive strength similar to very soft rock are expected to stand vertically in an excavation. Water bearing sand and silt lenses may cause local instabilities, unless properly treated. Class 4: Very Dense Sand and Gravel This soil type consists of poorly graded sand, gravelly sand and sandy gravel; it will be encountered in pockets and relatively thin layers in the excavation of the station tunnels and will likely be water bearing. This material has little to no cohesion and will show flowing behavior if charged with water or running behavior if allowed to dry out. Dewatering, pre-treatment and special consideration will be required when this material is encountered. Class 5: Very Dense Silt and Fine Sand This soil type consists of silty fine sand to sandy silt; it will be encountered over a substantial portion of the Main Shaft excavation and parts of the Concourse Cross Adit. Under hydrostatic pressure, this material will show flowing behavior. If drained, it is expected to stand well in small to medium sized openings with little face support. Class 6: Very Stiff to Hard Clay This soil type consists of overconsolidated silty clay or clayey silt, with some fine sand; it will be encountered

Figure 2. Geologic profile of station.

202 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-25.qxd

2/18/04

12:32 PM

Page 203

in the Main Shaft and most of the station tunnel excavation will be in this material. Slickensided and fractured zones will be encountered during tunnel construction; to take this into account in the design, a further subdivision into Clay A, Clay B and Clay C was considered necessary. Due to the hard consistency and cohesive nature of this material, it will stand well at excavation faces and will be relatively easy to excavate with a tunnel excavator. In the slickensided and fractured zones, spalling, raveling and wedge failures in the tunnel face and heading may occur if not properly presupported. Water bearing sand and silt lenses may cause local instabilities, unless properly treated. The high variability of the geology in the area of the future station poses the main design challenge. Special considerations and flexibility in the design are necessary to address this issue. During construction, a high degree of experience, alertness, and the proper tools to react appropriately to changing ground conditions are needed. Sound Transit therefore decided to extend the services of the design team to provide SEM supervision and construction support services. In order to get a better understanding of the soil strata and the ground behavior during an SEM type excavation, it was decided to construct an Exploratory Test Shaft and Test Adits within the boundaries of the future Main Shaft. A brief description of the program and the implications for the Station Design are provided in section 5.

3 LARGE SOFT GROUND SEM TUNNELS 3.1

General considerations

The design philosophy of SEM has been described and documented in depth in numerous publications. The original concept was adapted to be suitable for soft ground tunneling and first used in the Frankfurt Clay in 1968. Since then, means and methods have been developed further and a substantial number of large soft ground tunnels have been constructed in Europe, some of them in adverse ground conditions with shallow overburden. In the United States, soft ground tunnels of the size required for the Beacon Hill Station break new ground. Some of the key elements for large SEM tunnels in soft ground are: 1. Ovoid cross sections with rounded inverts and domed excavation faces to prevent stress concentrations. 2. Ring Closure within 1.5 times the tunnel diameter to prevent loosening of the surrounding ground and excess settlements.

3. Timely installation of sealing shotcrete/flashcrete and the initial shotcrete lining to prevent deterioration and loosening of the soils. 4. Subdivision of the faces into smaller drifts and adjustment of round lengths to be able to control and stabilize the excavation. 5. Utilization of the appropriate ground support, face support, pre-support and ground improvement measures. 6. Monitoring of the structure during construction to assure stability and verify design assumptions. 7. The ability to make adjustments in the field to deal with actual ground conditions encountered. 8. Experienced Construction Management, Site Supervision and Quality Control to ensure safety and efficiency. SEM tunnel design has to take these factors into account and relies heavily on engineering judgment and experience from previous projects, but also on advanced Finite Element Modeling Tools to determine the appropriate excavation sequences and support measures. 3.2

When there is some continuity in the geologic strata and ground conditions can be reasonably anticipated for certain reaches, different ground support classes can be predefined. These contain the excavation sequence and the required support measures, i.e. shotcrete thickness, number of spiles, soil nails, etc. However, when highly variable geology is encountered, ground types and ground behavior change within several feet and mixed face conditions are encountered over large portions of the tunnel alignment, a different concept needs to be developed and deployed which is described in the following. By using the “SEM Toolbox” approach, a conservative baseline scenario is defined, an excavation sequence is prescribed and standard support measures – e.g. shotcrete, wire mesh and lattice girders – are defined. Depending on the ground conditions encountered, additional support measures (“Toolbox Items”) are used on an as needed basis to ensure stability of the tunnel face and the surrounding ground. These include:

•

•

203 Copyright © 2004 Taylor & Francis Group plc, London, UK

Coping with variable ground conditions – the “SEM Toolbox” approach

Pre Support Measures – Rebar Spiling – Grouted Pipe Spiling – Metal Sheets – Grouted Barrel Vault/Pipe Arch Face Stabilization Measures – Face Stabilization Wedge – Pocket Excavation

09096-25.qxd

• •

2/18/04

12:32 PM

Page 204

– Reduction of Round Length – Face Bolts Ground Improvement Measures – Gravity and Vacuum Dewatering – Permeation Grouting, Fracture Jet Grouting Annular Support – Additional Shotcrete – Soil Nails – Temporary Invert

access/egress considerations, space requirements for mechanical/electrical equipment and the geometry of the junction to the Platform Tunnels. Grouting,

For estimating purposes, expected location and quantities of Toolbox Items are provided. This approach provides a high degree of flexibility during construction and makes it possible to control virtually all kinds of ground conditions, thereby greatly reducing the risks of SEM construction. However, it requires that contractors are familiar with the utilization of the mentioned support measures. Experienced site supervision is essential to ensure that the appropriate measures are taken in a timely manner. The Standard support measures are paid for on a linear foot basis for each tunnel, while the SEM Toolbox items are separate line items and paid for on a unit price basis.

4.1.1 Geology The Concourse Cross Adit will be constructed primarily in Very Stiff to Hard Clay and Till and Till like Deposits, with intermittent, cohesionless pockets of Silt and Fine Sand that may contain pressurized groundwater. Layers of Silt and Fine Sand and Very Dense Sand and Gravel are located at or near the crown of the excavation.

The cross section of the Concourse Cross Adit, the largest tunnel of the Beacon Hill Station, was developed according to architectural requirements, emergency

4.1.2 Design Due to the large size of the opening and the difficult ground conditions especially in the crown, excavation will be carried out using the dual side wall drift method. Grouted with a double packer system under high pressure (1000 psi), the Barrel Vault will provide presupport over the whole length of the tunnel and be used to improve the Very Dense Sand and Gravel. The maximum specified advance length is 3 ft 4 in., and the maximum separation between the two side wall drifts in longitudinal direction is two rounds. The stability assessment for the excavation sequence and the in-place structure of the Concourse Cross Adit Tunnels was performed using two three dimensional finite element models and the finite element program ABAQUS. The first model includes the Main Shaft, the breakout from the Main Shaft, the sequential construction of the Concourse Cross Adits and the headwall. The second model is used to assess the breakout from the Concourse Cross Adit into the Platform Tunnels.

Figure 3. Concourse Cross Adit – typical cross section.

Figure 4. Dual side wall drifts.

4 DESIGN OF THE BEACON HILL STATION TUNNELS 4.1

Concourse Cross Adit

204 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-25.qxd

2/18/04

12:32 PM

Page 205

Figure 6. ABAQUS model 2.

Figure 5. ABAQUS model 1.

The decision to utilize two models instead of one was made in order to limit model size and therefore keep running times for the finite element code within acceptable limits to facilitate an effective and flexible design process. The soils in the FE models were modeled using Mohr-Coulomb failure criteria (friction angles from 27° to 40°, cohesion from 0 to 48 kPa), the shotcrete and concrete for the primary and final linings were modeled as linearly elastic materials. The construction sequence for the Concourse Cross Adit was modeled by completing top heading construction of the side wall drifts, followed by bench and invert. The top heading, bench and invert excavation sequence of the center drift was modeled in the subsequent steps. As all anticipated construction stages were modeled in the FE analyses, the numerical results were used to assess the stability of the excavation and the excavation face as well as the structural performance of the tunnel linings.

4.1.3 Lining design For the section forces determined in the FE analysis, the structural design for the tunnels was performed to meet the requirements of ACI 318. It could be shown, that a 14 in. thick shotcrete lining (fcu  5000 psi) is capable of providing the required support for the tunnel structure. Due to stress concentrations around the openings in the Concourse Cross Adit at the junction with the Platform tunnels, a local thickening of the primary lining of 17 in., was required to avoid additional bar reinforcement. The final lining is designed for the assumption that the primary lining loses 90% of its stiffness in the course of time. Additionally, the full hydrostatic load is assumed to act on the final lining of the tunnel structures. It could be shown that a 14 in. steel fiber reinforced concrete lining (fcu  5000 psi, fiber content 70 lbs/yd3) is sufficient to withstand all the occurring loads. Additional reinforcement is only provided in the junction areas and the connection areas to the headwalls. 4.2

The cross section geometry for the Platform Tunnels was developed according to architectural requirements and train clearance. An additional requirement is the possibility of walking the TBM through the Platform Tunnel for the completion of the east section of the running tunnels. 4.2.1 Geology The Platform Tunnels will be constructed primarily in Very Stiff to Hard Clay and Till and Till like Deposits, with intermittent, cohesionless pockets of Silt and Fine sand that may contain pressurized groundwater. Layers of Silt and Fine Sand and Very Dense Sand and Gravel are expected to be located at or near the crown of the excavation in one section of the tunnel, and dry sand (“hour glass sand”) can be expected in

205 Copyright © 2004 Taylor & Francis Group plc, London, UK

Platform Tunnels

09096-25.qxd

2/18/04

12:32 PM

Page 206

Figure 7. Platform Tunnel.

Figure 8. Single side wall drift.

the invert of the Platform Tunnels in one area of the excavation.

is assumed to act on the final lining of the tunnel structures. It could be shown that a 12 in. steel fiber reinforced concrete lining (fcu  5000 psi, fiber content 70 lbs/yd3) is sufficient to withstand all the occurring loads. Additional reinforcement is only provided in the junction areas and the headwalls.

4.2.2 Design As the tunnel cross section for the Platform Tunnels is somewhat smaller than the Concourse Cross Adit, it was assessed that those tunnels can be constructed utilizing the single side wall drift method. The stability assessment for the specified construction sequence and the primary and final shotcrete and concrete structure was performed using the second three dimensional FE model. The Side Wall Drift for the Platform Tunnels is constructed using a top heading bench and invert excavation sequence; the remainder of the tunnel is excavated in the same fashion following the completed side wall drift with a minimum distance of 30. As all anticipated construction stages are modeled in the FE analyses, the analytical results were used to assess the stability of the excavation and the excavation face as well as the structural performance of the tunnel linings. 4.2.3 Lining design For the section forces determined in the FE analysis, the structural design for the tunnels was performed to meet the requirements of ACI 318. It could be shown, that a 14 in. thick shotcrete lining (fcu  5000 psi) is capable of providing the required initial support for the tunnel structure. As stress concentrations in the ground in the vicinity of the Concourse Cross Adit Tunnels could be observed, a localized thickening of the initial shotcrete lining was required in the junction area. The final lining is designed for the assumption that the primary lining loses 90% of its stiffness in the course of time. Additionally, the full hydrostatic load

4.3

For the construction of each of the SEM tunnels, prescriptive excavation sequences were developed. These contain breakout sequences, advance lengths, sizes of openings, distances to ring closure and distances between side wall drifts. In conjunction with the excavation, the standard support measures, i.e. flashcrete, wire mesh, lattice girders and shotcrete are defined. It is specified that the standard support elements for any round have to be complete prior to commencing the next excavation round in the sequence. To reduce the uncertainty about ground conditions ahead of the face, the systematic drilling of 35 ft long horizontal exploratory probe drill holes every 6 excavation rounds is specified. The results of the exploratory drilling and the assessment of ground conditions at the tunnel face will be used in the field to determine if there is a need for ground improvement or additional support measures. If so, the appropriate SEM Toolbox Items for the conditions encountered can be utilized to ensure the safety of the tunneling operation. To assist the contractor in choosing the appropriate support measure, requirements for the application of a particular item were defined in the GBR and the Special Provisions. For the preparation of the bid documents, baseline quantities for each Toolbox Item were defined according to the anticipated geologic conditions.

206 Copyright © 2004 Taylor & Francis Group plc, London, UK

Excavation and support design and utilization of NATM/SEM Toolbox

09096-25.qxd

2/18/04

12:32 PM

Page 207

5 MONITORING

7 BEACON HILL TEST SHAFT

During SEM Tunneling, monitoring, recording and interpreting deformations and stresses of the initial lining is essential to ensure construction safety and to verify the results of the design assumptions. For monitoring during construction of the mined Station, a comprehensive instrumentation scheme has been developed. Convergence Bolt Arrays will be used to monitor absolute and relative deformations. Concrete pressure cells will record the normal stresses in the tunnel lining, while earth pressure cells will be used to record the ground loads that are transferred to the tunnel lining. In addition, a surface monitoring program will utilize surface settlement points, inclinometers and extensometers to provide complete information about ground movements during the excavation.

7.1

6 WATERPROOFING The Beacon Hill Station is designed as a “tanked” structure, meaning it will be equipped with a waterproofing system to make it completely watertight. In addition, a Sectioning System is foreseen that will provide remedial repair options in case of leaks. The waterproofing system is installed between the initial shotcrete lining and the final lining and consists of the following elements:

•

•

•

Geotextile A non woven polypropylene geotextile is fastened to the initial shotcrete lining with PVC disks. It is designed to protect the waterproofing layer from sharp projections of the initial lining surface. Waterproofing Membrane The waterproofing layer is the actual sealing element of the system, designed to keep groundwater from the interior of the tunnel. It consists of flexible membrane sheets welded together to form a continuous, impervious layer. This geomembrane is made of a polymeric material, like polyvinylchloride (PVC). Its material properties allow it to adapt to the irregularities of the initial tunnel lining and it is designed to permanently withstand biological and chemical deterioration due to aggressive groundwater. Furthermore, it is fire retardant to minimize safety hazards during construction and operation. Sectioning System The waterproofing system is divided into sections by the means of water barriers. Should a leak occur at a certain location, only one relatively small section is affected, which can be repaired by grouting through preinstalled grout pipes.

In the course of the design process, the designers entertained the idea of a Test Shaft, its purpose being a more thorough understanding of the complex geology and the evaluation of the performance of the SEM construction method. The engineering team designed a 148 ft deep, 18 ft diameter SEM shaft and two Test Adits in different geologic strata within the foot print of the future Beacon Hill Station Main Shaft. By means of the Test Shaft and Adits, ground behavior of the various geologic strata, especially of the water bearing sands/silts considered most critical for tunneling, and of the hard clays, where most of the tunneling work will be performed, should be closely monitored and assessed. The value of a Test Shaft was determined to be the additional knowledge about geology and ground behavior and the resulting design optimization. Following an assessment of the experiences during construction and the results of the monitoring program, the original assumptions and the resulting design were to be confirmed or modified as required. 7.2

Test Shaft construction

The construction of the test shaft took place between April 2003 and September 2003. Deviating from the original intent to construct a shaft using the sequential excavation method (SEM), the contractor decided to excavate the first approximately 50 ft using an auger drill and subsequently install a reinforced shotcrete lining. Once ground conditions worsened and the excavation method using an auger drill could not be further utilized, SEM using a mechanized excavator and a reinforced shotcrete lining was used as prescribed in the Test Shaft design for the following approximately 60 ft. However, schedule delays and cost overruns, mainly due to more complex ground conditions and additionally employed dewatering measures necessitated the decision to terminate the construction of the Test Shaft before excavating the Test Adits. The test shaft was completed to its intended depth using a 6 ft diameter steel cased boring. 7.3

Test Shaft findings

Generally, the encountered ground conditions were well suited for SEM construction. For most of the depth the ground remained stable for the full depth of each excavation round (up to 6 ft) and for a considerable length of time (up to 5 hours and more). However, as permeable geologic layers (sands and silty sands) that had not been adequately dewatered, either by means of deep wells or vacuum well points, were encountered,

207 Copyright © 2004 Taylor & Francis Group plc, London, UK

Test Shaft purpose and design

09096-25.qxd

2/18/04

12:32 PM

Page 208

the sides of the shaft excavations locally displayed instability and the contractor could not install the required pre-support and support measures in the required time frame. During the Test Shaft excavation, it could be observed, that the geologic conditions in the Beacon Hill Test Shaft area were more diverse and variable than originally anticipated. Furthermore, it became apparent that for conditions where groundwater control is critical for the success of the construction method, the usual approach of a contractor designed dewatering system is problematic and should be replaced with an owner designed dewatering system. It has to be emphasized that the employment of a contractor capable of utilizing SEM principles and SEM Toolbox items in the required manner and time is paramount for the successful implementation of SEM design. Following the findings of the Test Shaft excavation, risk considerations lead to several design changes as outlined below. The value Test Shaft program was confirmed, as the design changes prior to the bid phase will avoid claims based on inaccurate design assumptions and save more than the costs incurred.

8 REDESIGN As the findings from the Test Shaft created concern among the designers of the shafts regarding the appropriateness of the SEM excavation in this geology, it was decided to redesign the SEM shafts and replace them with slurry wall shafts. For the construction of the tunnels, the SEM approach was maintained. The numerical analysis of the Concourse Cross Adit had to be rerun to take the changed geometry and stress regime of the slurry wall into account and the reinforcement and lining thicknesses were adjusted accordingly; the breakout sequences from the shafts were redesigned. Schemes for dewatering from the surface and from within the tunnels were added to the design package. In addition, more stringent requirements for exploratory probe drilling during construction were established and the application of SEM support elements was shown in more detail. Jet Grouting from the surface for the Ventilation Tunnels and to a limited extent for the Platform Tunnels was added. Finally, the anticipated ground conditions were adjusted in the GBR and new baseline

quantities and distributions for the SEM Toolbox items were established.

9 CONCLUSION The challenges posed by the geology and the station arrangement were systematically analyzed and addressed in the production of the design package. Concourse and Platform Tunnels, with widths of 45 ft and 34 ft respectively required the development of sequences able to cope with soft and potentially running conditions, but also very stiff and heavily slickensided soils. The Test Shaft program greatly reduced the uncertainty about the ground conditions and gave the design team the opportunity to evaluate and adjust the design approach. This will pay off as the information gained and implemented in the design package will give the contractors a better basis to bid the project. The decision to extend the designer’s services into the construction phase and to task him with the supervision of the SEM works will greatly minimize the construction risks and assure that the design intent is conveyed through construction. The design and subsequently the construction of the large Beacon Hill Station tunnels will serve as a Benchmark for the future of soft ground SEM tunneling in the United States.

REFERENCES Duddek, H. & Städing, A. 1990. Tunneling in Soft Ground and Sedimentary Rock for High Speed Double Track Railway Lines in Germany, Tunnelling and Underground Space Technology, Vol. 5, No. 3, pp. 257–263 Maidl, B. 1995 Handbuch des Tunnel und Stollenbaus. Essen: Glückauf Pacher, F. & Sauer, G. 1989. Grosse Querschnitte in nicht standfestem Gebirge. Wien: Springer Verlag Sauer, G. 2003. Ground Support and its Toolbox, ASCE Conference May 6 & 7, 2003. New York City Shannon & Wilson 2002. Geotechnical Data Report. Seattle Tunnels & Tunneling International Dec. 2002. Test Shaft to start at Beacon Hill Station: 51 Hatch Mott McDonald Jacobs 2003. Design Report, Seattle Hatch Mott McDonald Jacobs 2003. Geotechnical Baseline Report. Seattle Hatch Mott McDonald Jacobs 2003. Test Shaft Report. Seattle World Tunneling Oct. 2003. Beacon Hill Tunnel Project and Test Shaft: 314–315

208 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-26.qxd

2/18/04

12:33 PM

Page 209

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Application of the Press-In Method in East Side Access tunnel project Jinyuan Liu & Verya Nasri STV Incorporated, New York

ABSTRACT: The East Side Access (ESA) Project in New York is one of the largest tunneling projects undertaken by the US railroad. The project is divided into two major segments namely Manhattan and Queens. In this study, the potential application of an innovative Japanese construction technique, the Press-In Method is evaluated for the ESA project. This system will be used as the support of excavation for the four emergency exits and approach structures in the Queens segment. The deepest emergency exit for two adjacent parallel tunnels has an excavation depth of 85 ft and a water head of 68 ft. Soil within the excavation depth consists of typically coarse to fine, cohesionless glacial material, well sorted to well-graded, interspersed with cobbles and boulders. Tubular pile and the crush piler are selected because of the high water pressure and difficult subsoil conditions. The nonstaging method that continuously presses-in piles to construct a wall and utilizes the top of this pile wall as the platform for the equipments is selected to meet the stringent requirement of the site. The construction and design methods are presented in this paper, while the waterproofing, corrosion, and supporting details will be addressed. Based on the result of this study, the Press-In Method is a feasible and costefficient system for ESA project comparing to slurry wall.

1 INTRODUCTION 1.1

study of a new construction technique to be used in Queens part. More information about ESA project can be found in Nasri et al (2003).

East side access project

The East Side Access (ESA) Project in New York is one of the largest tunneling projects undertaken by the US railroad (Fig. 1). Its budget is around $5.2 billion. The project is divided into two major segments namely Manhattan and Queens. This paper presents a feasibility

Figure 1. Plan view of East Side Access project.

1.2

The aerial view of Queens bored tunnel is shown in Fig. 2. There will be four emergency egress shafts to

Figure 2. Aerial view of bored tunnels in Queens.

209 Copyright © 2004 Taylor & Francis Group plc, London, UK

Emergency shaft and approach structures

09096-26.qxd

2/18/04

12:33 PM

Page 210

be constructed and connected to tunnels. The deepest shaft that would be used for two overlapping tunnels will be discussed in this article. The bottom of this shaft will be approximately 85 ft below the existing ground surface. The plan dimension of this shaft will be an ellipse with two axial lengths of 44 and 58 ft, respectively. The approach structure is located on the north side of Harold Interlocking, immediately west of the 39th Street Bridge. The approach structure brings the track from the ends of the bored tunnels to the surface. It will be approximately 900 ft long and 25 ft wide. 1.3

Geotechnical condition

Per ESA Geotechnical Design Summary Report, the project site in Queens Segment is underlain by Ordovician/Cambrian age metamorphic bedrock, which is covered by Pleistocene glacial and interglacial deposits, and by postglacial materials. The various subsurface strata encountered in Queens site are not uniformly spread across the area. The soil stratum encountered at this site is a Fill with a varying thickness up to 10 ft, followed by a 20 to 45 ft thick Mixed Glacial Deposits (Stratum 2), and a 25–90 ft thick Glacial Till (Stratum 5). Bedrock is encountered at the deepest boring at a depth of 120 ft near emergency shaft and 40 ft at approach structure area. Water table is 68 ft above the bottom of excavation for the interested shaft. Water table is below the bottom of excavation in approach structure area. Stratum 2 consists of loose to dense coarse to fine sands with silts, gravels, cobblers, and boulders. Stratum 5 consists of dense to very dense sand with silts and gravels. Typical SPT N value for emergency shaft and approach structures are shown in Fig. 3, which ranged from 50 to greater than 100. 25

SPT N 50 75 100

0

0

0

20

10

40

20 Depth, Feet

Depth, Feet

0

60

SPT N 50 75 100

2.1

Press-In piles

The Press-In Method is a 30-year-old construction technique. Recently is has been further developed by research collaboration between Giken Seisakusho Co., Ltd. and Cambridge University Geotechnical Engineering Group Since 1994 (ENR 2001, Bedian 2002, and White 2002). The press-in principle is to utilize reaction force derived from fully installed piles and hydraulically press-in subsequent piles. The Silent Piler works on top of the reaction piles and self-moves to the next position gripping the pile being pressed-in. Technical details of the press-in mechanism are illustrated below in Fig. 4. In practical terms, the Silent Piler grips previously installed piles with hydraulic jaws. The next pile is hydraulically gripped by the Chuck at proper pressingin point and jacked into the ground with a static load generated by the main hydraulic rams. The Silent Piler derives reaction force from skin friction and interlock resistance of the previously installed reaction piles, which surpasses the press-in resistance during piling. Installing the pile to the designed depth by accurate hydraulic control, the Piler repeats the same press-in procedure until the last pile is put into the ground. Since the piles are pressed-in, the Silent Piler does not cause any damage to the environment including neighboring structures and local residents through noise and vibration. The Press-In Method operates at only 69 db of noise allowing pile installation in areas where environmental disruption is strictly precluded. Because reaction force is used as its basic mode of operation, the self-weight of the piling machine is rather unimportant. Moreover, the superiority of the principle enjoys great advantages with the integrated GRB System, which allows transporting of material, crane system and pressing-in to be systematically carried out from the top of fully installed piles utilizing a minimum of right of way.

30

80

40

100

50

120

25

2 PRESS-IN METHOD

Bed rock

60 Bed rock

(a)

(b)

Figure 3. Soil profile for (a) emergency shaft and (b) approach structure.

Figure 4. Schematic of Press-In Method.

210 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-26.qxd

2/18/04

12:33 PM

Page 211

3 DESIGN OF PRESS-IN METHOD As for the emergency shaft, water intends to inflow to the shaft due to the unbalanced water table. Ground modification is needed to reduce the inflow and provide a dry working environment. A jet grout plug or soil replacement plug of treated ground can be created prior to commencement of excavation within the shaft to prevent the inflow and balancing the uplift force (Coomber 1985). The bedrock is at 120 ft below the ground, which is about 35 ft deeper than the bottom of excavation. The pile will be driven through the soft ground and socketed into the bedrock. The length of tubular pipe for emergency egress shaft will be approximately 120 ft. Since the bottom of excavation is below water table, the tubular pile with P-P interlocking will be used in this shaft. As for the approach structure, tubular pile will also be used. The design length of piles will vary based on the retaining heights. 3.1

Press-In pile material

One of steel tubular pile manufacturers is Nippon Steel Corporation in Japan. Material can be procured and fabricated from American local markets. This material has a high elastic rigidity and high bearing capacity and can be constructed in soft ground. It has been used as both temporary structures and permanent retaining structures. Tubular piles provide bigger moments of inertia comparing to Z-type and U-type sheet piles with similar weight. Under the same loading condition, steel tubular piles can reduce the lateral deformation of the structure. With the introducing of inner bracing, the maximum deflection in emergency egress shafts can be reduced substantially to ensure the safety of nearby bridge pier and in-service tracks. 3.2

Physical restraints

There are many environmental constraints from nearby buildings and in-service railways. At the proposed optimum location for the emergency egress shaft at Honeywell Street Bridge, there is a dense network of surface tracks, utilities, catenaries, buildings and slopes. The design of the emergency egress shafts should minimize the impacts and relocations of these restraints. The compact equipment and its ability in limiting deformation of nearby structures makes the Press-In Method an excellent solution to these restraints. The main components of Giken’s Press-In Method consist of a piler and a power pack (Fig. 5). The PP260 piler proposed by Giken America Corporation is approximately 16 ft long, 7 ft wide, and 16 ft high, which is used for tubular piles with diameter of 31 to 36 in, The power pack for this piler is 14 ft long, 6 ft wide and about 8 ft high.

Figure 5. Components of Press-In Method equipment.

The compact and lightweight Silent Piler limit workspace to just the area ultimately required and it minimizes the effects on the environment. Under normal working conditions, the Silent Piler can operate with one crane to pitch piles. When a pile being pressed-in is sufficiently stable, the Silent Piler releases the clamps from the reaction piles and uses the pile to raise itself and travel forward. This “self-moving” system eliminates the need for support by a crane during the piling operation. In other words, even where a site requires a large jib radius for pitching, a relatively lightweight crane can be used. 3.3

For the approach structure, the press-in pile can be used for the permanent structures. Corrosion control should be considered. The corrosion control for steel tubular sheet piles is the same as that for any other type of steel piling. This is accomplished either by adding an extra pile wall thickness, which is technically called corrosion allowance, at the time of material production at the mill to make up for possible corrosion, or by spreading an adequate coating over the steel pipe surface. Nippon Steel Corporation has developed a number of corrosion-resistant steel and coated steel sheets using covering materials. As for the corrosion allowance, a corrosion rate of 0.03 mm/year can be selected for corrosion in soil above water level and 0.02 mm/year for the part in soil below water table. These rates can be refined in the detailed analysis.

4 CONSTRUCTION 4.1

Alignment

Alignment of the wall depends on the design. In general, straight line, circle line and their combination are often used. As the interlocking junctions can be welded at the designed place, tubular pipes can provide the curved alignment as well as the straight line, shown in Fig. 6 (Nippon Steel Corporation). For the case with curve and straight combination, the bracing

211 Copyright © 2004 Taylor & Francis Group plc, London, UK

Corrosion control

09096-26.qxd

2/18/04

12:33 PM

Page 212

Figure 6. The alignment of tubular pipe in a curved wall.

Figure 7. Mortar filling for waterproof.

and struts are normally used. In the case of the emergency shaft wall, braced supports will be used. 4.2

Waterproofing

Environmental site investigation finds the following minor environmental effects: soil with semi volatile organic compounds, petroleum-impacted soil, and groundwater with chlorinated volatile organic compounds near the emergency egress shafts. Waterproofing has to be guaranteed in order to not trigger the movement of the plumes. Steel tubular sheet piles are provided with interlocking junctions that have enough room to fill mortar concrete for water sealing of the junctions. There are two types of connections, P-T and P-P type interlocking connections. From waterproof point of view, pipe-pipe (PP) type will be used for its better waterproofing performance. The tubular sheet pile interlock should be fabricated to about 10 ft above the water table. The principle of waterproofing is to empty the interlocking connection by water jet and air lifting after piling work and then filling the room with mortar as shown in Fig. 7. Other than the retaining wall, the steel tubular sheet piles are often used as wall type foundation in water areas. That is because the steel pipe type sheet pile can serve as temporary water sealed cofferdam as well as the wall type foundation. Concrete work can be done in dry conditions inside the cofferdam of the steel tubular sheet pile wall. For instance, the maximum cut-off level was 89 ft in Tama River ventilation tower construction project in Japan. The construction completed successfully with only a few oozing out zones. 4.3

Misalignment

Based on the site condition in the ESA project, tubular sheet pile wall would be an excellent option for the shafts and approach structures. Giken’s equipment has been used successfully adjacent to live rail traffic in such a safe manner that construction was preformed without traffic interruption during peak traffic

conditions. This will provide the cost savings from having to work during strict working times, usually during the weekends at night. The pile installation is guided by a laser beam resulting in a remarkable alignment tolerance. The tolerance was less than 3 mm (1/8) in Long Island Expressway project. Giken’s system exhibits no perceived vibration during installation, which was proven by the U.S.A.C.E., New Orleans, so settlement to the adjacent ground is very limited. Settlement at the immediate adjacent area is generally limited to less than 1 inch in very dense gravel. Survey data recorded to date indicates that there was little to no settlement adjacent to the piles. Based on the soil conditions at the ESA project, the maximum settlement would be expected to be within 1/2 inch adjacent to the piles. More detailed analysis is needed to evaluate settlement of the bridge pier. 4.4

Due to the pile length of 120 ft and the soil conditions, a crush piler from Giken America Corporation is recommended for installing the tubular pile. This crush piler, which has been successfully used in similar soil, consists of a Silent Piler equipped with an integral augering method to advance the piles to the anticipated tip elevations. This system advances the tubular pile as far as it can through pressing techniques, at which time the Silent Piler can no longer press the pile any deeper, a continual flight auger is pitched and lowered on the pile. The auger is seated through a conventional crane on the pile with a series of hydraulic clamps. Once the auger is properly seated on the piler, and the head is lowered to the tip of the pile, the auger head is engaged. By having the auger fixed on the top of the pile, the auger head pulls the pile into the ground. This is done to the final tip elevation. The spoils generated from the augering are carried up a continual flight auger to the top of the pile where it is delivered to a chute, which drops the spoils into a spoil bucket. The spoils bucket is held in place with

212 Copyright © 2004 Taylor & Francis Group plc, London, UK

Tubular pile installation

09096-26.qxd

2/18/04

12:33 PM

Page 213

the crane. Once the bucket is full, it is lowered to the ground surface, when it is emptied into a stockpile, or directly backfilled into previously pressed piles. Once the pile is at sufficient depth, the auger motor is reversed and extracted with the crane and placed in a holder until the next pile. For a pile with a length of 120 ft, if a conventional crush system were used, it would require a continual auger system of over 120 ft in length and a crane with a minimum boom clearance of over 180 ft. A crush system newly developed by Giken will be proposed for this project, which will be similar to conventional crush system, however a mast leader system is used to allow for smaller spliced sections. This reduces the crane boom length to that of a medium sized hydraulic truck crane. This system is equipped with an auger lead system on which the auger motor is attached, similar to conventional augering rigs. This system, however, does not use a telescopic Kelly bar, sections are attached to the auger as the pile is advanced further in the ground. This system would allow for 40 ft long sections to be driven without requirement of a large crane. The 40 ft sections will be loaded with an auger in a horizontal position, prior to being hoisted to the piler. Once the sections have been inserted to the leader system and advanced for the next section to be attached to the auger. The pile section will be welded in the field. This system will be more suitable for the ESA project due to the presence of the catenary and compact construction space and will have a slightly better production rate in comparison with the conventional crush system. 4.5

Figure 8.

Inner brace system for tubular pipe.

Support system

In order to reduce the required modular of the wall and the lateral deformation of soil, brace system will be implemented in the excavation for emergency shafts. Tubular sheet piles can be fully equipped for additional bracing, struts, tiebacks, Wales, etc. Typically less bracing is required due to the large section properties of the tubular piles. Additional aesthetic facing can be fixed to the piles as either pre-cast units, or cast-in place. Studs will be required to be welded to the piles if the facing is cast in place. As to the inner support at the construction stage, steel braces and struts of H-Shaped will be used for the tubular pipe and sheet pile wall. In the tubular pile case, the space between the wall and brace should be filled with cast-in-situ concrete to distribute the pressure evenly to the pipe and avoid excessive deformation of the steel pipes (Fig. 8). As for the inner support used as the permanent structure, in general, reinforced concrete type braces and struts, or reinforced concrete slab are used. In these cases, shear connectors and steel tension bars are welded on the surface of the tubular piles. These bars

Figure 9. Brace system for emergency shaft.

are designed to transfer external loads between the reinforced concrete slab and tubular pile wall. The brace system for emergency shaft is schematically shown in Fig. 9. 4.6

213 Copyright © 2004 Taylor & Francis Group plc, London, UK

Cost efficiency

As for the emergency shaft, the unit cost for Giken system will be around $60 per SF, which does not include the cost of inner bracings and Wales. This is almost half the cost of slurry wall. The estimate includes Giken SCP-260 Crush Piler system and GRB system along with a 50-Ton clamp crane and pile runner. The cost estimate is based on a daily rental of Giken system with an assumed production rate from projects in similar soil conditions. The cost also includes the material including the shipping from Nippon Steel Corporation. Actual material price will likely vary depending upon the final design and material selected. Tax may vary depending on local tax rates.

09096-26.qxd

2/18/04

12:33 PM

Page 214

The second project was Long Island Expressway project, where a 35 ft cantilever wall was constructed on a steep slope with similar soil conditions to the ESA project (Fig. 11). 6 CONCLUSIONS Press-In Method is an advanced piling technique. Based on the analyses regarding the design and construction, press-in pile method can be used in the ESA project for supporting the construction of emergency shaft and approach structure. It is also cost efficient comparing to the slurry wall system.

Figure 10. Renovation project near railway in Japan.

ACKNOWLEDGEMENTS The authors are grateful for the helps of Mr. John Santos and Mr. Michael Carter of Giken American Corporation and Dr. Kazushige Tokuno, Mr. Shigeki Terasaki, and Mr. Takeshi Katayama of Nippon Steel Corporation. We also appreciate the editorial helps from Victor Shey and Ahmed Firoz.

REFERENCE Figure 11. Long Island Expressway extension project.

The cost estimate for the approach structures is based on the cantilever tubular wall. The embedded depth was determined from Plaxis analyses. The length of piles varies from 15 to 61 ft. The unit cost is similar to that in the emergency shaft, which is not economical because of high cantilever height near TBM reception pits. The cost can be reduced by introducing tie back system in the final design. Material cost can also be reduced by sourcing local market.

5 SUCCESSFUL PROJECTS Two successful case histories using press-in pile are introduced below. The first project, shown in Fig.10, was for a railway project in Aomori, Japan. It demonstrates the environmental friendly characteristics of Giken Press-in piling system, where the clearance was only 309 mm. Press-In Method will provide an excellent solution due to the physical restraints in the ESA, including live railway service.

PB/STV, ESA geotechnical design summary report for Queens segment, 2000. White, D. 2002. An investigation into behavior of pressed-in piles. PhD dissertation, Cambridge University, England. Coomber, D.B. 1985. “Groundwater Control by Jet Grouting.” Proc. 21st Reg. Conf., Eng. Group of Geolog. Soc., Sheffield, 485–498. Bruce, D.A., Boley, D.L. and Gallavresi, F. (1987). “New Development in Ground Reinforcement and Treatment for Tunneling.” 1987 RETC Proceeding, 2, 811–835. Bedian, M. 2002. “ ‘Value engineering’ in United States of America.” The 9th int. conf. on piling and deep foundations, Nice, France. Engineering News Record, 2001. “Japanese system quietly breaks ground on highway job.” July 9, 2001, 16. Giken American Corporation, Construction revolution guide, http://www.giken-smp.com/. Japanese Association of Steel Pipe Piles 2002, Steel Tubular pile foundation: Design and Construction, Nippon Steel Corporation. Nasri, V., Jafari, R. and Wone, M. 2003. East Side Access Project in New York, hard rock and soft ground tunneling. 12th PanAmer. Conf. SMGE, MIT Cambridge, MA, June 22–25. Nippon Steel Corporation 1987, Nippon steel’s steel pipe pile construction methods. Nippon Steel Corporation 1988, Steel pipe piles.

214 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-27.qxd

2/18/04

12:34 PM

Page 215

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Shotcrete for tunnel final linings – design and construction considerations V. Gall & K. Zeidler Gall Zeidler Consultants, LLC

N. Munfah Parsons Brinckerhoff Quade & Douglas, Inc.

D. Cerulli Parsons Brinckerhoff Construction Services

ABSTRACT: The use of shotcrete for tunnel final linings has gained increased popularity on a national and international basis. The high quality of the shotcrete material, flexibility in application and workability, as well as the ability to adapt to complex tunnel geometries have contributed to this popularity. When evaluating if shotcrete should be utilized as the final tunnel lining, several aspects should be carefully evaluated to determine the final product’s quality and durability, as well as cost and construction schedule implications for a given tunnel configuration. Among others, geometric complexity, tunnel length and size, staging of a multi-layered application, finish requirements and type of waterproofing will play a major role in the decision. This paper establishes and discusses aspects and criteria that should be considered in the evaluation process for, or against, a tunnel final shotcrete lining. This discussion is supported using recent case histories, in particular the Pedestrian Walkback Tunnel at Washington Dulles International Airport in Dulles, Virginia, and the Weehawken Tunnels in New Jersey for New Jersey Transit to demonstrate the decision process.

1 SHOTCRETE As reported in many documents, the material shotcrete has undergone significant developments during the past decade. Improvements of the material as well as the application method have been achieved. Intensive research in the material quality led to a better understanding of the interaction between the various constituents of a shotcrete mix, to the development of a series of new admixtures and better quality control of cement types. In particular, the use of wet mix techniques, the development of new low/non alkali accelerators, water content reducing admixtures and continuous cement quality resulted in improved final shotcrete quality. But also the use of fiber reinforcement and high-end concrete pumps and guns have furthered the shotcrete quality. The new materials have allowed better slump control, which did not only contribute to a more steady flow with the new pumps and therefore continuous shotcrete application, but much more to a more controlled and uniform compaction and, consequently, shotcrete density. The reduction of the W/C ratio, now

enabled by the use of plasticizers and partial replacement of cement, dramatically reduced the overall pore volume and, hence, improved the durability of shotcrete. With the help of the admixtures, the quantity of rebound was reduced to acceptable values, eliminating one economic disadvantage of shotcrete. With today’s shotcrete mix designs and application equipment, high final strengths of up to approximately 70 MPa (10,000 psi) are achieved in standard applications. Together with the use of shotcrete as permanent support material, requirements for the surface quality became more demanding. The improved workability, smaller aggregate grain sizes and better hydration heat control (cracks) enabled the contractors to satisfy these requirements. Trowel finished shotcrete surfaces (Varley 1998, Eddy & Neumann 2003) or architectural ornamental finishes (Gall et al 1998) are examples for shotcrete finishes achieved on past projects. The compressive strength of sprayed concrete is only an indirect indicator for the shotcrete durability. Durability and water tightness are intimately interconnected. Crack development and dispersion control

215 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-27.qxd

2/18/04

12:34 PM

Page 216

Figure 1. Trowel finished shotcrete lining.

and the volume limitation of the effective pores reduce the permeability of shotcrete. Values of 1012 m/sec, desired minimum values for sufficiently water tight and durable concrete, are achieved or even surpassed. If concrete is exposed to groundwater and no water flow exists across the concrete section, water absorption is of greater concern than permeability. The control of the volume of permeable pores within the concrete section and limitation to a maximum value of 14 to 17%, as recommended by various documents, is achievable in standard shotcrete applications. Fibers are not only used to better the behavior of shotcrete during fire, but also to increase the ductility of shotcrete and shrinkage crack control and dispersion. Above improvements combined with the inherent flexibility of shotcrete application resulted in a high acceptance of shotcrete within the industry and authorities. Shotcrete can be compared to high quality castin-place concrete and, in some fields, even proved to have superior characteristics. 2 LINING DESIGN PHILOSOPHIES During the history of tunnel lining designs, different lining philosophies have been developed. Dependent on the assumption, whether or not the initial lining will have sufficient quality and durability under the project specific conditions, the initial shotcrete lining has been taken into account for the long-term support, or has been considered sacrificial. In the latter, a secondary lining had to carry all expected ground and groundwater loads in the long term. The different water tightness criteria implemented at various projects under specific project conditions led to diverse waterproofing solutions, including the use of shotcrete for water tight linings, or the installation of membrane waterproofing systems sandwiched between initial and secondary lining.

In Europe, various authorities developed their preferences with respect to tunnel waterproofing systems. For example, most of the railroad and metro authorities in Germany and Austria tend to utilize shotcrete/ concrete to control the desired degree of tunnel water tightness, while the road and highway authorities prefer membrane waterproofing systems. The decision whether or not to use and be able to achieve a water tight concrete/shotcrete is also driven by the project specific environmental conditions, such as hydrostatic pressure conditions, chemical attack potential of the groundwater, and construction complexity. In some projects, the shotcrete initial lining has been considered sufficiently durable to withstand the longterm loads over the design life. The designers of several access shafts and stub tunnels for the upgrade project of London Electricity’s power supply network (London, UK) have opted to use the sprayed concrete lining, which was placed after excavation, for the long term support of these structures (Field et al 2000) as the so called Single Pass Lining. Specially detailed construction joints and high quality shotcrete were required to meet the client’s water tightness criteria. Damp patches were acceptable. The lining design thickness was considered appropriate to provide sufficient long-term stability, even when a certain portion of the shotcrete lining exposed to ground and groundwater will degrade. Similar to the classical two-pass lining systems with water tight cast-in-place concrete secondary linings, sprayed concrete has been used in lieu of cast-in-place concrete. At the Jubilee Line Extension, Contract C104 – London Bridge Station (London, UK), the complex geometry and alignment of the ventilation tunnels and the step-plate-junction housing a track bifurcation instigated the contractor to install a shotcrete lining on the inside of the initial lining (Varley 1998). The design was based on the assumption that the initial lining would deteriorate over the years and would lose its support capacity. The secondary lining has to carry all ground and hydrostatic loads expected to act during the design life. The water tightness criteria, where damp patches were permitted, were met by a high quality, steel fiber reinforced shotcrete and specially designed construction joints. A finishing layer of plain, small size aggregate shotcrete was applied to cover the steel fiber reinforced shotcrete. To meet the smoothness criteria for the ventilation tunnels, the finishing layer received a trowel finish. Similar principles have been applied at the ventilation chambers for DART’s City Place Station Project in Dallas, TX (Ugarte et al 1996). An early application of composite shotcrete linings was the lining system installed at the Heathrow Airport Transfer Baggage System Tunnel (Arnold & Neumann 1995). The shotcrete initial tunnel support was designed to provide the long-term ground support, while a

216 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-27.qxd

2/18/04

12:34 PM

Page 217

secondary layer will provide support to the hydrostatic loads. Both shells are interlocked by a rough, prepared joint surface and cross reinforcement and are expected to act as a composite structure with load sharing between the shells, effectively forming a single shell lining. Water tightness criteria, a dry tunnel had to be supplied, were achieved by high quality shotcrete and the continuous secondary layer of approx. 100 mm (4 in) thickness. Requirements for the composite function of the shotcrete layers and the shotcrete product itself have been identified by, among others, Kusterle & Lukas (1990) and Kupfer (1990). The more traditional two-pass lining system, combined with a membrane waterproofing system, is currently being applied at the Russia Wharf Segment in Boston, MA for MBTA’s Silverline Extension (Zachary 2003). There, the initial shotcrete lining is expected to deteriorate over time under the onerous environmental project conditions. A secondary shotcrete lining is being installed to provide long-term support to full overburden ground loads, surcharge and hydrostatic loads. A full-round membrane waterproofing system completely wraps the twin tunnels to provide a dry tunnel environment and to protect the secondary lining from potentially adverse groundwater affects. High quality shotcrete is used for the long-term support. Similar principles have been applied at WMATA’s Contract B10, Washington, DC for the construction of the double cross over and ventilation chambers in the mid 1980’s. Detailed design and practical considerations are described below based on a similar application at the Pedestrian Walkback Tunnel (PWT) at Washington Dulles International Airport (Hirsch et al 2003) and the Weehawken Tunnel project, in Weehawken, New Jersey (Ott & Jacobs 2003). These also include aspects of a layered shotcrete lining application. The PWT is approximately 240 m (800 ft.) long with a springline diameter of ca. 12 m (42 ft.) and features a double lining system, whereas a continuous PVC waterproofing membrane separates the initial and final linings. The Weehawken Tunnel involves the re-construction (enlargement) of a 1,269 m (4,156 ft) long, existing railroad tunnel into a two-track light rail tunnel with an underground station and a large passenger access and ventilation shaft. The widening of the tunnel to the station structure comprises a widening from an 8.4 m (28 ft) wide tunnel to an 18 m (60 ft) wide station tunnel structure to both sides of the future center platform station. Based on a Value Engineering Change Proposal submitted by the contractor, this transition, designed in a step plate junction configuration per contract, will be carried out using shotcrete for the arch final lining in a bifurcation as shown in plan in Figure 2. Another concept of lining design is currently being applied at the King’s Cross Station Redevelopment

Figure 2. Concrete vs. Final shotcrete lining geometry in plan and longitudinal section (schematic).

Project, London, UK (Cox et al 2003). The complex geometrical and alignment conditions, as well as the multiple tunnel junctions and intersections proved castin-place concrete secondary lining an uneconomical solution. Hence, the lining system will comprise a steel fiber reinforced shotcrete initial lining, a full round membrane waterproofing system (for completely dry tunnels) and a steel fiber reinforced shotcrete secondary lining. Rebar or welded wire fabric reinforcement may be required around tunnel junctions. Due to the rather benign environment offered by the surrounding London Clay and the groundwater contained in it, it has been decided to take some benefit from the initial shotcrete lining for the long-term support. The initial lining is not expected to completely deteriorate and lose its support capabilities. This is made possible in part by new shotcrete technologies, producing highdensity shotcrete, steel fiber reinforcement and a better understanding of the ground and groundwater impact on sprayed concrete. Part of the initial lining is expected to deteriorate over time, while the remaining portion will contribute to the ground support in conjunction with the secondary lining. Due to a requirement by the owner, all steel reinforcement forming parts of the permanent tunnel support must be located inside the membrane waterproofing system. Therefore, no benefit can be taken from any steel reinforcement located within the initial lining. The initial lining is taken into account as mass concrete material that will contribute to the support in confinement. The shotcrete secondary lining will, protected by the waterproofing system, provide the long

217 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-27.qxd

2/18/04

12:34 PM

Page 218

term support for the hydrostatic loads and the remaining part of ground and surcharge loads. The waterproofing membrane, sandwiched between the initial and secondary lining, is expected to permit radial load transfer only with negligible shear transfer between the linings.

3 GENERAL APPLICATION CRITERIA Shotcrete final linings are typically utilized where one or more of the following conditions are encountered:

•

• •

The tunnels are relatively short in length and the cross section is relatively large and therefore investment in formwork is not warranted, i.e. tunnels of less than 150–250 m (400–600 ft) in length and larger than about 8–12 m (25–35 ft) in springline diameter. The access is difficult and staging of formwork installation and concrete delivery is problematic. The tunnel geometry is complex and customized formwork would be required. Tunnel intersections, as well as bifurcations qualify in this area. Bifurcations are associated with tunnel widenings and would otherwise be constructed in the form of a step plate junction configuration and increase cost of excavated material (see Figure 2).

If the above conditions characterize a tunnel structure then a shotcrete final lining is likely to provide for flexibility in production, schedule advantages, savings in formwork and possibly savings in excavation. Therefore, a detailed shotcrete final lining cost analysis is warranted.

4 FINAL LINING EQUIVALENCY CONSIDERATIONS 4.1

Structural calculations

Structural calculations for final shotcrete linings follow the same principles and are based on the same structural codes as concrete linings. With current high shotcrete product quality and knowledge of application procedures, shotcrete is internationally viewed as concrete applied by different placement means. Due to the application process however, the reinforcement may, and in most cases will, be different in a shotcrete application. Whereas in a regular concrete section two layers of rebars at a wide spacing are sufficient, the shotcrete section will utilize welded wire fabric for better embedment within the shotcrete and to facilitate the shotcrete application. Where the loading conditions for the lining are well established, the same loadings are used in a structural calculation to arrive at reinforcement needs. Alternatively, equivalency considerations

may be applied, equating the given concrete section and its reinforcement to a proposed new section with a different reinforcement arrangement. The PWT shotcrete final lining reinforcement needs were a result of equivalency considerations, i.e. the reinforced shotcrete lining had to provide the same capacity as the castin-place concrete lining. An exception was the complex three-dimensional section between the mechanical room tunnel and the main tunnel where additional reinforcement beams were installed at the intersection along the groin lines (Figure 4). When considering the application of a final shotcrete lining, the following aspects should be addressed prior to acceptance and execution in the field. 4.2

In principle, there is no structural difference between a sprayed or cast-in-place concrete lining. However, when the sprayed lining is applied in multiple layers with distinct time intervals, which include installation of reinforcing steel, the bond between the different layers has to be adequate to qualify as a monolithic member in the structural sense. Limitations and requirements are therefore imposed on application sequencing, curing techniques, cleaning of surfaces and adapted concrete technology (Hoehn 1999). Keeping the time lag between shotcrete applications short aids this process. For verification, minimum tensile and shear strengths between the layers (in the joint) shall therefore be achieved. For example and to assess the requirements for these values at the PWT project, finite element calculations were carried out that considered a representative three-layer composite system with two joint surfaces in the final lining section (see Figure 3). The model investigated the capacity of the 30 cm (12 inch) layered shotcrete final lining for the long-term condition, when the initial support is assumed to be deteriorated and overburden and live loads are imposed onto the final shotcrete lining. From this model, minimum tensile and shear strength requirements in the joints were derived to be 0.69 MPa (100 psi) and 1.38 MPa (200 psi) respectively. Hoehn, 1999 for example calls for minimum values for strength for both tension and shear of 1.5 MPa (217.5 psi). Kusterle and Lukas, 1990 rather report ranges of values to account for statistical characteristics of sampling and testing. A review of these ranges, combined with the fact that the literature reports 1.5 MPa for tensile strength as a “universal number” and the availability of detailed calculations led to the conclusion that the above minimum values for tensile and shear were plausible. 4.3

Testing

Testing requirements for a final lining shotcrete resemble very much those of an initial shotcrete lining,

218 Copyright © 2004 Taylor & Francis Group plc, London, UK

Multi-layered vs. Monolithic

09096-27.qxd

2/18/04

12:34 PM

Page 219

Figure 3. FE model of shotcrete lining layers.

Figure 5. Final lining shotcrete application at PWT.

Figure 4. Shotcrete final lining installation at PWT intersection.

however with modified requirements, in particular to test for the bond capacity of the layered shotcrete. The shotcrete mix design is often developed based on historical data available from the initial lining application. At the PWT project pre- and during production testing requirements involved testing of tensile and direct shear tests on samples taken from test panels sprayed according to application and curing conditions resembling the site application, considering that the full thickness of the final shotcrete lining was to be achieved in panels not to exceed 10 m (30 ft) in length. Tensile strength was tested according to ACI 506R, whereas the shear tests were carried out according to Michigan DOT’s shear test. Minimum test requirements were as per the above, 0.69 MPa (100 psi) for tensile and 1.38 MPa (200 psi) for shear

strength. During pre-construction, testing time intervals between applications of 24-hours and 72-hours were tried and led to strength developments yielding a minimum of 2 MPa (290 psi) in tensile strength and 4.70 MPa (680 psi) in shear after ten days. During construction, a total of four tests with two samples each were required for the entire tunnel, again time lag and application to simulate application and site conditions. The minimum tensile strength developed at three days was recorded as 0.8 MPa (116 psi), with an average of 1.47 MPa (213 psi). The minimum shear strength at three days was 5.03 MPa (730 psi), with an average of 6.83 MPa (990 psi). Therefore, test results showed that the minimum bonding requirements of the composite final shotcrete layer were well achieved by the selected construction process. Application of the shotcrete final lining is shown in Figure 5. 4.4

The use of a dedicated waterproofing layer between the initial and final shotcrete linings creates a debonding effect. The degree of de-bonding depends on the type of waterproofing selected. In particular when using a loosely laid, continuous, flexible membrane type waterproofing (PVC) for complete water tightness (Gall 2000), special attention has to be given to membrane attachment, reinforcement installation and

219 Copyright © 2004 Taylor & Francis Group plc, London, UK

Waterproofing and contact grouting

09096-27.qxd

2/18/04

12:34 PM

Page 220

to contact grouting. A frequent use of attachment disks will achieve a tighter fit of the membrane to the initial shotcrete lining and reduce the amount of void space otherwise created by sagging membrane sections. For the spraying of shotcrete against the membrane, a carrying layer of welded wire fabric will be required. Spacers may be used between the welded wire fabric and the membrane to push the membrane further against the initial shotcrete lining. Despite these measures, a void space will exist between the membrane and the initial shotcrete lining. For proper contact between the initial and final shotcrete linings, systematic contact grouting is essential. This contact grouting, unlike the one in roof sections in cast-in-place final lining installations, is not limited to roof sections only, but a radial and more frequent distribution of grouting ports and pipes around the lining perimeter should be considered for this purpose. By injecting low viscosity cementitious grouts between final shotcrete lining and the membrane will assure a tight contact between the initial and final lining. Where water barriers have been utilized for the purpose of enhanced membrane repair (compartmentalization) a re-injectable grouting hose should be installed in the centerline of the barrier, between the ribs. Injection of grout through this hose will assure a tight embedment and contact between the ribs and shotcrete, and thus prevent leakage water to migrate across water barrier ribs. 4.5

Surface finish

There are various aspects of surface finish requirements that strongly depend on the tunnel’s intended use. These include, but are not limited to, reflectivity (in vehicular tunnels), ease of maintenance (washable), smoothness (in ventilation tunnels), appearance (general), and frost resistance (exposure to cold climates). For all of the special applications solutions exist and include screeding and trowel finishing, use of special mix shotcrete, and very fine aggregates for the finishing layer, yielding surface finishes that, by appearance and function, very well compete with the cast-in-place concrete. However, such surface finishes are often not required and omission of special finishes provides for further economy. At the PWT, for example, an internal architectural finish will be used. Therefore only limited requirements for the surface were established for ease of maintenance and facilitate installation of embedments and a flatness/smoothness criterion, which called for a deviation of not more than 2.5 cm (1 inch) in 1.5 m (5 ft.), was established. 4.6

Fire resistance

Recent fire incidents, in particular in European tunnels, have initiated numerous investigations in adequate

fire testing and the improvement of the fire resistance of concrete and sprayed concrete. One prime element contributing to spalling and subsequent section thickness loss has been identified: The free water contained within the concrete section leads, when evaporating due to rapidly increased temperatures, to explosive spalling of the concrete. Tests have proven that the addition of microfilament fibers to the shotcrete mix significantly improves the fire resistance of shotcrete. The fibers melt under the influence of heat and provide escape channels for the vapor, allowing the pressure to dissipate (Tatnall 2002). A detailed review of fire resistance needs at the Weehawken Tunnel led to the application of 1.9 kg/m3 (3 lbs/cy) of microfilament fibers for the inner 10 cm (4 inch) of the shotcrete final lining in transition sections.

4.7

• • • • • • •

Execution of Work (Installation of Reinforcement, Sequence of Operations, Spray Sections, Time Lag) Survey Control and Survey Method Mix Design and Specifications QA/QC Procedures and Forms (“Pour Cards”) Testing (Type and Frequency) Qualifications of Personnel Grouting Procedures

5 SUMMARY AND CONCLUSION Based on general trends in the application of shotcrete for final linings and as demonstrated on recent case histories, it is apparent that shotcrete presents a viable alternative to traditional cast-in-place concrete. The product shotcrete fulfills cast-in-place concrete requirements, or sometimes can even surpass those. Design and engineering, as well as application procedures, can be planned such as to lay the basis for a high quality product. However, excellence is needed in the application itself. Skilled nozzlemen have to ensure a high degree of workmanship through formalized training, experience and quality assurance during application.

220 Copyright © 2004 Taylor & Francis Group plc, London, UK

Method statement/application procedures

Probably the most important factor that will influence the quality of the shotcrete application is workmanship. While the skill of the shotcrete applying nozzlemen (by hand or robot) is at the core of this workmanship, it is important to address all aspects of the shotcreting process in a method statement. This method statement becomes the basis for the application procedures, the applicator’s and the supervision’s Quality Assurance/ Quality Control (QA/QC) program. Minimum requirements to be addressed in the method statement are as follows:

09096-27.qxd

2/18/04

12:34 PM

Page 221

ACKNOWLEDGEMENT The authors would like to acknowledge several firms and individuals for the information that forms the basis for the two projects discussed in detail. The Pedestrian Walkback Tunnel: Metropolitan Washington Airports Authority (Project Owner), Parsons Management Consultants (Construction Manager). The Weehawken Tunnel: New Jersey Transit Authority (Project Owner, Mr. Anthony Murtah, Mr. Ty Dickerson), Parsons Brinckerhoff Quade and Douglas (Prime Design Consultant), and Frontier-Kemper/ Shea/BuM Joint Venture (Contractor, Mr. Vincent Sambrato, Mr. Leon Jacobs). REFERENCES Arnold, J. & Neumann, Ch. 1995: Umsetzung eines innovativen NÖT-Konzeptes im Zuge eines “Know-howTransfers”. Felsbau 13 (1995), No.6, 459–563. Cox, R., Dulake, Ch. & Eddie, C. 2003: Complex redesign for London link. Tunnels and Tunnelling International, Vol. 35, No. 4, April 2003, 50–52. Eddie, C. & Neumann, Ch. 2003: LaserShell leads the way for SCL tunnels. Tunnels and Tunnelling International, Vol. 35, No. 6, June 2003, 38–42. Field, G., Legge, N. & Liew, B.S. 2000: Optimizing Shaft Design and Construction Using Sprayed Concrete. Our World in Concrete & Structures, Proc. 25th Anniversary Conference, Singapore. Gall, V., Zeidler, K., Predis, T. & Walter, J. 1998: Rehabilitation concepts for brick lined tunnels in urban areas. Tunnels and Metropolises, Proc. World Tunnel Congress Sao Paulo, Vol. 1, 539–546, Rotterdam.

Gall, V. 2000: Three Pillars for an Effective Waterproofing System. Proceedings, North American Tunneling 2000, Boston, Massachusetts, June 6–11, 2000. Hirsch, D., Moran, P. & Patel, A. 2003: Tunneling Under Washington Dulles International Airport. Proceedings, Rapid Excavation and Tunneling Conference 2003, 648–656. Hoehn, K. 1999: The Single-Shell Shotcrete Method Applied at Two Tunneling Sites – Concrete Technology and Economic Viability. Proceedings, Spritzbetontechnologie ’99, BMI 1/99, 255–270. Kupfer, H. & Kupfer, H. 1990: Statical Behavior and Bond Performance of the Layers of a Single Permanent Tunnel Lining, Proceedings, Spritzbetontechnologie ’90, 11f. Kusterle, W. & Lukas, W. 1990: High-Grade Shotcrete for the Single Permanent Shotcrete Lining Method, Proceedings, Spritzbetontechnologie ’90, 29–40. Ott, K. & Jacobs, L. 2003: Design and construction of the Weehawken Tunnel and Bergenline Avenue Station. Proceedings, RETC 2003, 936–946. Schreyer, J. 1999: Constructive and Economical Suggestions for the Lining of Single Shell Tunnels. Proceedings. Spritzbetontechnologie ’99, BMI 1/99, 271–281. Schwarz, J. 1999: Structural Design and Quality Assurance of the Joint between Outer and Inner Layer when Using the Single Shell Shotcrete Lining Method. Proceedings. Spritzbetontechnologie ’99, BMI 1/99, 237–240. Tatnall, P. C., Shotcrete in Fires: Effects of Fibers on Explosive Spalling. Shotcrete, Vol. 4, No. 4, Fall 2002, 10–12. Ugarte, E., Gall, V. & Sauer, G. 1996: Instrumentation and its Implications – DART Section NC- 1B, City Place Station, Dallas, TX. Proceedings, North American Tunneling ’96, April 21–24, 1996. Varley, N. 1998: Concrete tunnel linings at London Bridge. Concrete, Feb. 1998, 13f. Zachary, W. 2003: The Cold War: Boston’s Uncommon Dig. AUA News, Vol. 18, #3, 9–11.

221 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-28.qxd

2/18/04

12:35 PM

Page 223

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Robotic shotcrete applications for mining and tunneling Michael Rispin Underground Construction and Allentown Equipment Master Builders, Inc.

Chris Gause Underground Construction Master Builders, Inc.

Thomas Kurth Meyco Equipment

ABSTRACT: While shotcrete has evolved as a means and method for ground control, so too have the demands for faster and safer placement. Spraying manipulators, or robots as they are commonly referred to, have become the rule rather than the exception both in mining and larger tunnel projects. Even after the capital investment of a robotic shotcrete machine, the benefits can be measured and returns achieved by:

• • • •

Increased production Reduction of rebound Higher quality shotcrete in-situ Improved safety for shotcrete crews.

This paper discusses state-of the-art mechanized shotcrete machines and provides case histories describing the benefits in mining and tunneling.

1 INTRODUCTION If a Robotic Applicator is mentioned in connection with mining or tunneling and sprayed concrete, what is basically meant is an apparatus used to hold and control a spraying nozzle. Why should this be necessary when a man can do the same work? Tunneling and mining development are intrinsically hazardous forms of construction, when sprayed concrete is used as initial temporary support after blasting, using a mechanical arm to extend into an unsupported area is a great enhancement to personnel safety. A spraying manipulator is a hydro-mechanical, remote-controlled spraying unit for mechanizing and automating the application of sprayed concrete. It is suitable for use anywhere substantial quantities of wet or dry shotcrete will be applied, and offers significant advantages in construction applications where conditions are such that manpower might be exposed to potentially unstable, unsupported ground, rebound or dust. Mounted on various kinds of carrier vehicles and able to achieve a reach of up to 14.5 m (47 ft), a

robotic applicator will save the cost and time of erecting scaffolding, where due to the very size of the working area it would otherwise be needed. As this paper will show, there are many combinations and permutations of configurations of robotic applicators in use today around the mining and tunneling worlds. 2 A HISTORY OF ROBOTIC APPLICATORS Thirty years ago, the first manipulators really were just nozzle holders. Over the next twenty years, innumerable variations appeared in all parts of the world based upon cranes, drill jumbos and lifts with a device enabling the nozzle to be attached. These assemblies were not designed for quick and nimble nozzle and arm movements, so efficient placing of quality shotcrete with a smooth finish on difficult substrate was extremely difficult, if not impossible. Specialized spraying manipulators began to appear in the early 1980s, by which time sprayed concrete had become an acceptable form of construction (if only

223 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-28.qxd

2/18/04

12:35 PM

Page 224

by the dry method). The most suitable of these dedicated units had features that are still standard today: spraying heads with universal movements, eight fields of motion, and the “parallel-lance” with great extension possibilities. These arms were generally mounted onto existing carrier vehicles; used trucks and excavators were favorites. Features such as automation of movements of lance and nozzle holder were integrated into remote controls to help the nozzleman to produce a better spraying pattern. By the mid 1990s, with the proliferation of large scale wet-spraying, the spraying manipulator had firmly established itself as a piece of equipment to be found on almost all large construction projects where sprayed concrete was used as temporary or permanent support. But the demands made upon the manipulators had changed: the “bar” had been raised! A mechanical device to hold a nozzle was no longer enough. New standards and economic constraints demanded more speed and efficiency in placement. This meant that a manipulator had to be able to hold and control a nozzle and hose with diameters of up to 80 mm to enable the full capacity of the shotcrete pump to be used, in order to save time and therefore money. This required not just robustness, but also operational dexterity to allow large amounts of concrete to be placed quickly and accurately, typically impossible with a converted placing boom. Remote controls also developed, from hydraulic levers to electric operation with cables, and later radio remote control became a standard option. This period in time also saw the development of the autonomous spraymobile. These vehicles were trimmed from top to bottom with all the equipment necessary and with one aim in mind: quality sprayed concrete. Manipulators became very diverse and specialized, as construction was customised to be exactly suited to application conditions, be it for large civil projects, tight mining tunnels, shafts or even integrated into a TBM. The present day sees the demand for more quality and accountability in both results and the application process on site. Even more automation is required. There is only one way these attributes can be assimilated into a robotic applicator and that is through the use of computer technology. By 2000 the first computer-controlled robots had appeared. Able to be programmed to spray an area automatically and keep records of the work, this advance opened up vast new possibilities in improving tunneling and mining safety, economy and efficiency. Computer control eliminated the need for a nozzleman to work continuously close to the “danger” area. The required finished surface accuracy increased as the machine, coupled with the computer through laser measuring technology, worked much more precisely than a human. The fatigue and skill factor variables were removed from the equation.

Automation holds great advantages. In deep mines, for example, long travelling times and short shifts can be replaced with full employment of resources by a nozzleman who sits safely on the surface controlling processes through his MMI (ManMachine-Interface). Simpler units can be equipped with “teach-in” features that repeat various patterns. Work in hostile environments, such as a uranium mine, can now be tackled with much less risk. The future will belong to these types of robotic applicators, but there will also be place for the dedicated standard hydro-mechanical manipulator. 3 ROBOTIC SPRAYING VS. HAND SPRAYING The benefits of mechanized shotcrete application can be evaluated by three categories: 1. Increased production 2. Higher quality shotcrete in-situ 3. Improved worker safety. 3.1

A multitude of reasons exists which allows increased production with the use of a shotcrete robot, most of which are due to the elimination of the human fatigue factor. The predominant reasons are as follows:

•

•

•

224 Copyright © 2004 Taylor & Francis Group plc, London, UK

Increased production

Increased concrete hose diameter – Some spraymobiles are equipped with 102 mm (4.0 in) hoses. Hand nozzling will typically use a 51 mm (2.0 in) hose diameter. The weight of shotcrete in the 102 mm (4.0 inch) line is equal to 18.3 kg/m (12.3 lbs/ft). When you multiply this by 1–2 m (3–6 ft) of hose length often being supported by the nozzleman, combined with the compressed air supply, any person would quickly become exhausted. Fatigue also carries over to pumping rates or pump output. As the shotcrete output is increased, the nozzleman must also resist the increase in line surge that comes from temporary interruption of pumping while the swing tube changes cylinders and begins the next stroke. The nozzleman in a sense must act as shock absorber. In addition to pump surges, the compressed air (5–7 m3/min for hand spraying, whereas robotic spraying involves 10–14 m3/min with 7 bar pressure) delivered to the nozzle body also applies a backward pressure that must be compensated for by the nozzleman. This additional fatigue factor is of course eliminated with mechanized spraying equipment. With the human fatigue factor eliminated, shotcrete volumes can increase dramatically. Hand nozzling volumes can range from 7–9 m3/hr (9–12 yd3/hr), while mechanized spraying can easily reach volumes

09096-28.qxd

2/18/04

12:35 PM

Page 225

of 20 m3/hr (26 yd3/hr). This is particularly beneficial in larger diameter tunnels, stations, galleries or when shotcreting for the final lining is being utilized. 3.2

Higher quality shotcrete in-situ

There are combinations of capabilities with mechanized spraying that allow shotcrete to be placed with improved in-situ properties. Some of these are: 1. Dedicated maximum air volume for optimum compaction 2. Lance mounting is automatically held parallel to the axis of the tunnel 3. New robotic manipulating capabilities also allow for automated nozzle adjustments to be made to maintain proper standoff distance as well as nozzle angle to the substrate. 3.3

Improved worker safety

The contributions to a safer working environment via robotic spraying are clear. With use of a remote control, crews are able to remain in supported areas while letting the reach of the spraymobile apply shotcrete in the newly excavated areas. In areas that require a combination of rock bolts and shotcrete, the bolting crews can take advantage of working in a supported environment where an initial layer of shotcrete has been sprayed for temporary support.

4 STATE-OF-THE-ART ROBOTIC APPLICATORS In producing top quality sprayed concrete, the best manipulator is still only one component of a system. The complete system is imperative if the manipulator is to be used to its full potential. On large construction sites such as tunnels, it is imperative that the spraying set-up is installed and ready to start performing within minutes of the heading being ready for it. As soon as the spraying operation is finished, the equipment has to be removed so that the next work cycle can begin. Furthermore, it is a common trend to execute different jobs simultaneously, which demands complete, self-contained equipment. For example, because a central air supply is seldom large enough to supply all site demands, the complete mobile therefore carries its own compressor. 4.1

Meyco Potenza

The Potenza is one of the better examples of a complete mobile unit for the spraying of concrete. This type of spraymobile has been setting the standard for sprayed concrete in tunnels and other areas of application

using the wet-mix shotcrete method. They have become commonplace on many of the world’s most important sites where sprayed concrete must be applied in large quantities without compromising quality. The standard components of the complete mobile unit are:

• • • • • • • • • • • • • • •

225 Copyright © 2004 Taylor & Francis Group plc, London, UK

Robojet spraying manipulator Potenza sprayed concrete pump for wet-mix process Integrated Dosa TDC accelerator dosing unit MEYCO Data for compiling operating and performance information Central power unit Chassis, 4 wheel drive and steer, with stabilisers Cable reel Air compressor Nozzle system Liquid accelerator tank Water storage tank Working lights Water pump High pressure water cleaner Release oil pump.

09096-28.qxd

4.2

2/18/04

12:35 PM

Page 226

Allentown MBS-02E

The MBS-02E features most of the sprayed concrete capabilities outlined for the Potenza but is built as a smaller and more robust package designed for the smaller mining headings and the rigors of the North American underground mining environment. Most importantly, while its primary purpose is to spray concrete, it is also capable of quickly and efficiently travelling the underground tunnels and ramps in a typical mine in order to be efficient in its utilization in multiple headings. The spraying manipulator, designated Meyco Minima, boasts a folding boom that retracts for tramming, yet is unfoldable in a 3 m 3 m heading (10 ft 10 ft), and offers a maximum spraying range of 9 m in height (29.5 ft), 7 m lateral (23 ft), and 8 m forward (26.2 ft). Due to frequently encountered, unexpected conditions in a mine, the spraymobile is also equipped with a shotcrete pump and hopper assembly that can be hydraulically positioned at various heights to adjust to any type of feed, typically from a transmixer, even when parked on uneven surfaces.

enable movement, dexterity and ease of handling for the nozzleman. A large ring, as shown above, would be similar to equipment used in the Lötschberg Transalpino Project in Steg and Raron.

4.4

4.3

TBM ring construction sprayer

A TBM can also be viewed as a carrier vehicle. Manipulators on TBMs should be, of course, part of an integrated system designed with the quality of end product, i.e. sprayed concrete, in mind. TBMs vary greatly in their individual construction, dictated by the geological environment where they will be employed. This in turn influences the design of the manipulator. Space is at a premium and the logistics are difficult so a manipulator must have the greatest possible movement but not clutter up the already crowded back-up rig. MEYCO has manufactured both ring construction type manipulators and centrally placed lance units according to local requirements. These units are always tailor made, but should contain all the basic principles and components to

226 Copyright © 2004 Taylor & Francis Group plc, London, UK

Shaft robo

09096-28.qxd

2/18/04

12:35 PM

Page 227

Shaft manipulators actually have a lot in common with ring constructions made for TBMs. The big difference is obviously the angle and direction in which the carrier vehicle, in this case a Galloway stage, is either lowered or raised within a vertical excavation. Again, the manipulator must be an integral part of a coordinated sprayed concrete system. Depending on the diameter of the shaft, a centrally mounted lance or a ring running around the stage would be used. 4.5

Meyco Oruga

The Rama is a range of manipulators manufactured by MEYCO. Their common features are that they can be mounted on almost any type of carrier vehicle or mounting stage. They are all of robust and simple construction and they vary in that each model has a different maximum spraying range, derived from their physical dimensions. They all have a spraying head with two hydraulic oscillating motors with nutation device transmitting the required wobble movement to the spraying nozzle; adjustable speeds allowing optimum nozzle positioning. 5 CASE STUDY – MINING – KIDD CREEK

The Oruga is small and compact when driving around on its tracked carrier. It also has a reach of up to 8 metres and reliable stability when spraying. The Rama 4 manipulator operation is through electric remote control. They are ideal for slope protection and are compact enough to work within a TBM back-up rig. 4.6

Meyco Rama

The Kidd Creek Mine is located in Timmins, Ontario, Canada, where copper–zinc–silver deposits were discovered in 1963. Owned by Falconbridge Limited, it was put into production in 1965 with an open pit mine, which was excavated from 1965–1977. Subsequently, the ore body has been mined through three separate shafts known as the No.1, No.2, and No.3 mines. For years, Kidd Creek used bolt and screen construction for primary ground support. Dry shotcrete was used as secondary reinforcement and for repair where needed. At the end of the 1990s the company began searching for a better, faster, safer ground support method. In approaching the search for a new ground support protocol, the challenge was to develop a system that would be safe and economically feasible to apply in a complex and deep mining environment, and would be accepted by workers in the mine. They needed a viable new ground support method that would reduce exposure to unsafe working conditions, and meet stringent government and company regulations. Early in the process, the workers were focused simply on finding a better, faster way of applying dry shotcrete – they were not considering wet shotcrete. The mine had tried wet shotcrete in the early 1980s and it was not a success, so they were reluctant to explore that alternative. In 1999 the mine explored

227 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-28.qxd

2/18/04

12:35 PM

Page 228

new technologies to help them meet their goal. They recommended an innovative wet shotcrete system using the latest automatic delivery and spray equipment as a solution that would meet all of the Kidd Creek goals – increased productivity, enhanced safety and improved structural performance. In July of 2000 the Falconbridge Board of Directors approved the development of a new mine on the Kidd Creek site. Known as Mine D (Deep), it will extend the mine from a depth of 2070 M (6800 ft.) to 3050 M (10,000 ft.). When completed it will be the deepest base metal mine in the world, and it requires a significant infrastructure, including a new internal underground shaft, hoisting facilities, and approximately 15 kilometers (9 miles) of development. Started in 2001, the project is estimated to take four years to complete and approximately 100,000 m3 (130,000 yd3) of construction concrete and 60,000 m3 (78,000) of shotcrete will be used. As a result of the success of the SFRWS in field trials and subsequent use, Kidd Creek Mine management decided to use the system in the construction of Mine D. The mine commissioned a state-of-the-art on-site batch plant with capacity to feed two 200 mm (8) diameter cased boreholes to depths of 1400 m (4600 ft.) and 1460 m (4800 ft.) respectively. Five wet shotcrete spray mobiles and seven transmixers were acquired to meet the needs of the mine. Tenders were let and the mine chose the MSV shotcrete sprayer as supplied from MBT’s Allentown Equipment manufacturer. The MSV was designed especially to handle the underground environment. It features a robust carrier and utilizes some of the most effective and efficient drive components on the market. It has an overall tramming length of only (7.3 m) 24 ft. and a height of (228 cm) 90 inches. The sprayer was not only capable of higher and safer tramming speeds, it was also able to cover numerous headings in one shift. As of late summer 2003, the Mine D project had reached a level of 2438 m (8000  ft), with more than 14,000 m3 (18,200 yd3) of shotcrete applied using the MBT–MSV spraying units. An indicator of the improved safety is the fact that since fully implementing SFRWS as primary ground support in early 2002, there has not been one loose related injury. 6 CASE STUDY – CIVIL – BERGEN TUNNEL Economics based upon productivity will vary based upon the tunnel size, mining cycle and purpose of shotcrete application. A comparison between productivity of mechanized spraying vs. hand spraying can be extracted from the Bergen Tunnel Rehabilitation Project, North Bergen, NJ and the Cameron Run Tunnel Rehabilitation Project, Alexandria, Va.

During rehabilitation of the Cameron Run Tunnel, the contractor, Merco, Inc., hand sprayed using a Reed B30 concrete pump. The Reed B30 has a theoretical output of 22.8 m3/hr (30.0 yd3/hr.). Typical actual volumes applied were 3.6 m3 /hr.(6.0 yd3/hr.) or 30.4 m3/shift (40.0 yd3/shift). On the Bergen Tunnel Rehabilitation Project, Merco/Obayashi, JV utilized a self-contained robotic shotcreting machine known as the Meyco Potenza (described above). The Potenza is equipped with a Suprema shotcrete pump with a theoretical output of 20.0 m3/hr (26.0 yd3/hr). With the use of the Potenza shotcrete robot, the shotcrete volumes on the Bergen Tunnel project reached an hourly average of 14.0 m3, (18.2 yd3/hr) with the best day (two shifts) being 168.0 m3, (218.4 yd3). In addition to the increased output, the shotcrete crew size for robotic spraying was reduced to 3 men vs. 5 men for hand spraying. Although specific dollar values were not applied to the cost reduction in comparison, the increased output and shotcrete manpower reduction made some obvious contributions to an in-place cost savings.

7 THE FUTURE 7.1

228 Copyright © 2004 Taylor & Francis Group plc, London, UK

Logica

Meyco Logica is a relatively new machine, based on the well known kinematic principles of the Robojet, and has been developed in cooperation with industry and academia. This manipulator with 8 degrees of freedom has a new automatic and human oriented control system. The new tool enables an operator to manipulate the spraying jet in various modes, from purely manual to semi-automatic and fully automatic, within selected underground areas. It is also able to measure the tunnel profile with a laser scanner. In one of the modes the operator uses a six directional joystick (Space Joystick). The calculation of the kinematics is done by the control system. A laser scanner sensor measures heading geometry and this information is used to control automatically the standoff distance and the angle of the spraying jet. The aim of this control is not to automate the whole job of spraying but to simplify the task and enable the operator to use the robot as an intelligent tool, and to work in an efficient way with a high level of quality. With a correct angle of application and constant spraying standoff distance, a remarkable reduction in rebound and therefore savings in cost is achieved. Further, if the heading profile is measured after spraying as well, the system will relay information on the thickness of the applied shotcrete layer, which up to today was only possible with core drilling and measurement. If an exact final shape of the heading profile is required, the control system is being developed

09096-28.qxd

2/18/04

12:35 PM

Page 229

The Westerschelde Project was an example of a successful project using 2 Logicas to spray a 50 mm lining of passive fire protection mortar with tolerance 4 . The total tunnel length was 2 6 km. In the Netherlands, Meyco supplied equipment used for the application of fire protection mortar to the Groene Hart tunnel project. To fulfill the standards to spray apply a defined, constant and homogenous layer of a passive fire protection mortar, preference was given to a jobsite tailored solution. The entire spraying equipment was placed on gantries allowing trucks supplying the TBM to pass underneath. Mechanical engineers designed and built a spraying nozzle mounted on a wagon travelling on a guide rail along the tunnel. The whole construction moves on a ring beam along the tunnel wall. All these movements can be conducted with “teach-in” functions to allow automatic spraying within a defined area. After a 4 m longitudinal length is sprayed, the gantry will be moved and the next spraying phase can be repeated. By the time the whole set up was commissioned a spraying accuracy of 2 mm by 35 mm thickness was being achieved! 8 CONCLUSION Sprayed concrete is an economic, efficient, and versatile means of ground support for modern mining and tunneling operations. As we learn more about the benefits that shotcrete technology can bring to our underground industries, its use will proliferate. Robotic applicators have already proven to be a useful, and sometimes indispensable tool, in the application of shotcrete. The advances chronicled above have been built on systematically developed experience, and as each case study is completed and analyzed, further developments and efficiencies will ensue. The new frontier is automated shotcrete application, with a very high degree of applied thickness control. While the technology is here today, it needs to be employed on a larger scale where its benefits will be brought to bear for tunnel owners and contractors, and mine operators across the globe.

currently to manage the robot to spray to these defined limits automatically. The system shows that increased productivity in shotcrete application is doubtlessly possible without increasing danger to personnel or without huge increases in cost.

REFERENCES Mergentime, S.: personal communications. Melbye, T., Dimmock, R. and Garshol, K.: “Sprayed Concrete for Rock Support”, 2001. Master Builders article: “Shotcrete Developments at Kidd Creek Mine.”

229 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-29.qxd

2/18/04

12:36 PM

Page 231

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Development of the LaserShell method of tunneling C.M. Eddie Morgan Est, Tunneling Division, Rugby, UK

C. Neumann Beton-und Monierbau, Innsbruck, Austria

ABSTRACT: This paper describes the development of a new sprayed concrete tunneling process known as LaserShell™. The system has been developed by Morgan Est and Beton-und Monierbau to improve the safety, quality and efficiency of underground works using sprayed concrete and is currently being used on major road, rail and sewerage tunnels in the UK. The profile of both the excavation and spraying operations is controlled using an innovative real time survey system known as TunnelBeamer™. The developments of the TunnelBeamer™ and LaserShell™ technologies are described in the paper. The paper also describes the development of a permanent, ultra high quality steel fiber reinforced sprayed concrete mix which has enabled lattice girders and bar and mesh reinforcement to be effectively eliminated from the sprayed concrete tunneling process. In recent years a concerted effort has been made by the UK tunneling industry to put in place robust risk mitigation measures to further improve the safety of tunnel workers at the face. Exposure to unsupported (or inadequately supported) ground is an area of undoubted risk, which needed to be addressed in relation to sprayed concrete lining methods (SCL). Although the UK Health and Safety Executive (HSE) regards tunneling with sprayed concrete lining as an effective and viable method of construction, the residual risk of exposure to unsupported ground is considered unacceptable. A recent major project utilizing a sprayed concrete lining was the North Downs Tunnel for the Channel Tunnel Rail Link constructed by Eurolink (Morgan Est, Dumez GTM and Beton-und Monierbau). Following a serious incident involving a fall of material from a tunnel roof, Morgan Est made an undertaking to the UK Health and Safety Executive, that developments would be made for future projects whereby tunnel workers would not be required to enter an area of inadequately supported ground. This required therefore that sufficient support, as determined by analysis, had to be in place prior to entry.

lattice girders to provide profile control of the lining. Lattice girders also provide a mechanism for securing mesh during the application of the primary lining (Figure 1). It is the installation of the girders and mesh together with profile checks, which place the tunnel workers in the exposed vault to an unacceptable risk. To eliminate this risk, a method of controlling the lining shape, thickness and position remotely has

1 REQUIREMENTS Traditionally, sprayed concrete lined tunnels in soils; unstable ground or shallow tunnels require the use of

Figure 1. Mesh and Lattice Girder Installation on the North Downs Tunnel part of the Channel Tunnel Rail Link, UK.

231 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-29.qxd

2/18/04

12:36 PM

Page 232

Figure 4. Single Shot TunnelBeamer™ instrument fitted to a tunnel excavator.

Figure 2. Single Shot TunnelBeamer™ instrument.

Figure 5. Single Shot TunnelBeamer™ instrument fitted to a tunnel excavator.

Figure 3. Multi Shot TunnelBeamer™ instrument.

been developed by Morgan Est, Tunneling Division and Beton-und Monierbau. The brief was to develop a user friendly system of tunneling (methods and equipment) to enable the real time control of profiles of both the excavation and sprayed concrete lining without the use of lattice girders or mesh. The tunneling system was also required to provide comprehensive documentation of “as-built” work for quality and certification reasons.

With the removal of the lattice girders, the excavation and spraying operations have no existing orientation line or physical profile control mechanism. The “TunnelBeamer™” system consists of either a single laser (Figure 2) or a number of lasers (Figure 3) grouped together to act as a distometer which are directed at the excavation or sprayed concrete lining faces as required. The information from these lasers is linked continuously to a tunnel computer (situated in the tunnel), which contains the 3D-tunnel geometry information and which produces a comparison to the theoretical position. This comparison information is continuously displayed on a monitor in the operators cab (Figure 6). The TunnelBeamer™ instrument does not need the level and stable platform necessary for standard survey

232 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-29.qxd

2/18/04

12:36 PM

Page 233

Figure 6. Driver’s continuous display.

Figure 8. Driver’s continuous display screen. Figure 7. TunnelBeamer™ system integration.

instruments and has been designed to withstand heat, dust and vibration levels commensurate with being fitted directly to excavation or spraying equipment. A servo theodolite is used to locate the TunnelBeamer™ in three dimensions to allow the tunnel computer to relate the TunnelBeamer™ information to the theoretical tunnel alignment/profile (Figure 7). The system utilises existing tried and tested hardware and software components. The TunnelBeamer™ system has been developed by Morgan Est and Beton-und Monierbau around a construction method known as LaserShell™. This method employs an inclined face excavation for increased stability and improved safety for tunnel workers. To minimise the number of construction joints and improve productivity, for tunnels up to 6.5 m diameter it is proposed that LaserShell™ will be constructed full face. For the larger tunnels, only the crown or pilot excavation would be undertaken using full face LaserShell™ techniques. The LaserShell™ method can be used for the construction of “One-Pass”, “Composite” or “Traditional” sprayed concrete tunnel linings, although clearly the One-Pass approach delivers maximum economy. In tunnels adopting the One-Pass philosophy, the initial 75 mm layer is considered to be sacrificial and is therefore not considered in the permanent load case.

To provide a robust face and vault support measures at all times, whilst allowing access to clean and prepare the invert prior to the construction of the structural lining, the LaserShell™ construction sequence can be visualised as follows (Figures 9–15). 2 BENEFITS As previously stated, the main purpose of this development programme was to eliminate the risk to tunnel workers in an exposed ground situation. However there are many other benefits to be derived from such systems. Speed of advance and improved ring closure times, in conjunction with an inclined face significantly reduce surface settlement. Removal of lattice girders and the replacement of mesh with High Carbon Steel Fibre substantially improve the quality and durability of sprayed concrete linings by eliminating shadowing. Systematic capture of profile data relating to both the excavation and the sprayed concrete lining gives absolute confidence with respect to lining shape, thickness and position (Figures 16 & 17). Compared to traditional SCL construction methods, an assessment on cost and time has shown that savings of up to 50% will be achieved in certain applications. This can be demonstrated by the fact that the often three-stage excavation process of

233 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-29.qxd

2/18/04

12:36 PM

Page 234

Figure 9. LaserShell™ stage 1 (excavation commences).

Figure 10. LaserShell™ stage 2 (excavate top 70% of face).

Figure 11. LaserShell™ stage 3 (spray 75 mm initial layer on top 70% of face).

234 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-29.qxd

2/18/04

12:36 PM

Page 235

Figure 12. LaserShell™ stage 4 (excavate invert and carefully clean).

Figure 13. LaserShell™ stage 5 (spray 75 mm initial layer in invert).

Figure 14. LaserShell™ stage 6 (spray structural primary lining 360 degrees).

235 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-29.qxd

2/18/04

12:36 PM

Page 236

Figure 15. LaserShell™ stage 7 (extend temporary fill in invert and repeat stages 1 to 6).

Figure 18. Crown, Bench and Invert, three-stage excavation process. Figure 16. Profile data for the finish layer (sprayed concrete layers and excavation profile data similar).

Figure 19. LaserShell™ excavation commencement development trials). Figure 17. Visualization of the Multi TunnelBeamer™ instrument during spraying.

(Pre-

Shot

crown, bench, and invert (Figure 18) can be replaced with a one-stage excavation process (Figure 19). For larger tunnels or tunnels in unstable conditions where some sub-division of the face is considered to be

prudent, the savings relate to the elimination of the lattice girders and mesh. To meet the requirements of LaserShell™, Morgan Est, Tunneling Division, Research and Development (in co-operation with Beton-und Monierbau and Prof. Dr. Wolfgang Kusterle of Innsbruck University) has recently completed extensive pre-commencement

236 Copyright © 2004 Taylor & Francis Group plc, London, UK

process.

09096-29.qxd

2/18/04

12:36 PM

Page 237

Figure 22. Morgan Est R&D facility, Permeability testing. Figure 20. Morgan Est R&D facility, Core bath.

Figure 21. Morgan Est R&D facility, Bump table testing.

development trials for an ultra high quality permanent sprayed concrete mix. The purpose of the trials was to develop a sprayed concrete mix and application methods that would meet all the requirements of a single pass, permanent sprayed concrete lining system. The key objectives of the trials was to prove the structural integrity from application (15 minutes strength) up to 120 years, whilst also providing a mix with sufficient workability retention to enable efficient application in real tunnel environments.

Detailed and onerous performance criteria in respect of strength gain; flexural toughness, permeability, bond characteristics between layers and durability were set prior to the commencement of trials and benchmarked against comparable high quality cast in place structural concrete (Figures 20, 21 and 22). Early age strength development relative to workability retention times was a key factor and despite claims of all the admixture suppliers, the trials failed to support the view that the sprayed concrete could be heavily retarded without significant loss of early age strength. Following extensive laboratory testing, a total of 24 field mixes were tested. A combination of hand spraying and robot spraying techniques were used to replicate real construction conditions. Overhead spraying was undertaken on a purpose built frame incorporating a 5 m radius to simulate construction of a large diameter tunnel (Figure 23). In addition to the measurement of performance against the criteria set prior to the trial, an extensive investigation into the effects of high early age loading on immature sprayed concrete samples was performed (Figure 24). Samples were subjected to loading from 15 minutes to simulate. Utilisation Factors of between 30% and nearly 100% [Utilisation factor is the ratio of induced compressive stress to strength and is time dependent]. Even where samples had been subjected to stress states nearly equivalent to failure for up to 40 hours, no loss of integrity was measured. In order to prove the flexural toughness of the MF24 mix, a series of tests were performed by the British Building Research Establishment (BRE).

237 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-29.qxd

2/18/04

12:36 PM

Page 238

Figure 23. Overhead Spray Trials at Morgan Est R&D facility.

Figure 24. Early age testing of Sprayed Concrete.

Early testing with Dramix standard sprayed concrete fibres identified potentially brittle failure in certain circumstances due to the high strength and bond of the MF24 mix. Following consultation with Bekaert’s Engineers, a decision was taken to use High Carbon fibres. Tests on beams reinforced with High Carbon fibre (even when sprayed in separate layers) have shown excellent post crack toughness (Figure 25).

Figure 25. Testing of a layered MF24 Mix Beam Reinforced with High Carbon Steel.

control samples of in situ concrete and samples deliberately overdosed with accelerator. The control samples were taken from a mass concrete block cast and vibrated in a shutter. The sprayed concrete samples were recovered from panels sprayed overhead and the samples were cured in air up to 28 days. After 28 days, the samples were divided into three curing environments; under water at 20 degrees Celsius, in air at 65% Relative Humidity at 20 degrees Celsius and in cycles, one week in water at 40 degrees Celsius and 3 weeks in air at 40 degrees Celsius. Testing was performed at 1 month, 6 months and 1 year to determine compressive strength, Modulus of elasticity and Porosity. In addition, at 1 year, thin samples were taken and examined under a scanning electron microscope (SEM) to look for signs of deleterious behaviour. On completion of the trials, no adverse strength or stiffness values or trends have been recorded on any sample and the SEM/Petrographic analyses have shown no deleterious processes in any samples. In summary, it has been demonstrated that the sprayed concrete, even when overdosed with accelerator, is stable and can be categorised as highly durable.

4 PROOF OF CONCEPT TUNNEL 3 DURABILITY TESTING In order to understand the differences (if any) between sprayed concrete and traditional in situ concrete, three types of sample were prepared by Morgan Est at their facility in Rugby, UK. These samples were then carefully transported to Innsbruck University for curing and testing. The samples of sprayed concrete with the designed accelerator dosage levels were tested in addition to

In order to demonstrate the safety and accuracy of the system, a full scale proof of concept tunnel was undertaken. A 4.5 m diameter tunnel with one diameter of clay cover was constructed using the LaserShell™ method. The tunnel was extensively monitored and tested during construction and valuable data was recovered relating to soil and tunnel displacements, profile control and concrete quality. An attempt was also made to

238 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-29.qxd

2/18/04

12:36 PM

Page 239

Figure 26. Proof of concept tunnel showing DIBIT survey to confirm accuracy of TunnelBeamer™.

Figure 27. Sprayed concrete accreditation course – Overhead Robot Spraying.

measure stresses within the lining, but as with many previous attempts to capture stress data, results proved disappointing. To demonstrate the accuracy of TunnelBeamer™ system, an independent survey using the camera based DIBIT system was undertaken on each advance (Figure 26). Correlation between profiles recorded by TunnelBeamer™ and DIBIT was very good, giving a high degree of confidence that the system was reliable. 5 SPRAYED CONCRETE ACCREDITATION TRAINING The use of sprayed concrete is highly dependent upon the avoidance of human failure. Morgan Est therefore recognise the critical importance of training of all staff and operatives involved in the construction,

Figure 28. A completed tunnel driven by TunnelBeamer™ and excavated by the LaserShell™ method of tunneling.

inspection and testing of sprayed concrete tunnel linings. Accordingly, Morgan Est is the only organisation in the UK to run a sprayed concrete accreditation course suitable for the production of permanent sprayed concrete (Figure 27). The course is held at Morgan Est, Tunneling Divisions Research and Development facility near Rugby, UK. The experienced Nozzlemen and Pump Operatives are given a minimum one-week course involving a combination of theoretical and practical work. Only if all of the operatives sprayed concrete test pieces pass the performance criteria (hand spraying and robot spraying) and only if the operative passes a written examination, will a certificate be awarded. No man is permitted to spray concrete unless they are in possession of a current certificate of competence.

6 CONCLUSIONS The key conclusions are: – The developed sprayed concrete mix delivers a high performance rating in terms of medium to long term strength and provides excellent joint integrity. – Claims of admixture manufacturers that concrete can be retarded for long periods without detriment to the performance of the sprayed concrete have been proved to be wrong. Retardation of over 3 hours has been shown to reduce the early age performance to dangerously low levels (in respect of block retention and fall of newly sprayed concrete). – Early age strength gain from the sprayed concrete with low alkali accelerator is comparatively low (approximately J2), even when retarded to give only 2 hours life.

239 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-29.qxd

2/18/04

12:36 PM

Page 240

– High Carbon steel fibres required to prevent brittle failure of the high strength concrete. – High integrity joints and layers can be formed which show high levels of structural integrity and low permeability. Tensile splitting tests and beam tests on samples with and without joints showed very little difference. – Results from durability testing to date show no signs for concern. The concrete is classified as

“highly durable” and can meet all criteria set in respect of concrete that is “comparable to cast in place” can be met. – Rapid advance rates, coupled with systematic face support and early ring closure delivers excellent settlement control capabilities. – The LaserShell™ method of tunneling, utilising TunnelBeamer™ delivers unparalleled levels of safety, quality and efficiency.

240 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-30.qxd

2/18/04

12:37 PM

Page 241

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Ground support design and analysis: Exchange Place Station Improvements M.R. Funkhouser & M.F. McNeilly Golder Associates Inc., Lansing, Michigan & Newark, New Jersey, USA

ABSTRACT: The Exchange Place Station Improvements Project was one of three parts to the Port Authority of NY & NJ (PANYNJ) Downtown Restoration Program (DRP), which was implemented following the destruction of the World Trade Center (WTC) twin towers and the loss of commuter rail service to the Port Authority TransHudson (PATH) WTC Station. The project involved design and construction of six (6) new tunnel crossovers between five (5) existing 90-year-old concrete lined tunnels, and extending the existing station platforms approximately 46 m (150 ft) west. Design of this project involved excavating cavern spans upwards of 18 m (60 ft) in the underlying Manhattan Schist bedrock formation with rock cover as low as 7.5 m (25 ft) and multi-storey buildings directly above planned excavation limits. Design of ground support alternatives were completed using both review of historical precedents and UNWEDGE software analyses to evaluate load carrying capacities of various composite support systems. The design called for staged excavations with support installed at each stage. Ground support consisted of pre-stressed, resin grouted rock bolts, pre-fabricated steel lattice girders and steel fiber reinforced shotcrete (SFRS) liner systems.

1 INTRODUCTION

2 PRE-EXISTING CONDITIONS

Exchange Place Station (EPS), on the PATH commuter rail system, is located adjacent to the Hudson River in Jersey City, New Jersey. The project involved constructing six (6) new tunnel crossovers between five (5) existing tunnels, and extending both of the stations existing platforms to the west. The purposes for these improvements were to reconfigure EPS to allow operation as a “terminal” station, provide PATH with greater operational flexibility, and re-open EPS by July 2003. This work was undertaken as part of PANYNJ’s overall DRP, and was performed under an extremely tight schedule. Exceptional communication and coordination among the Owner, the Contractor and the numerous design consultants was key to the successful completion of the project. Exploratory drilling, core logging, and laboratory testing was still in progress when preliminary ground support designs were being drafted and construction procurement processes were being finalized. With needs to develop initial designs before site specific data was available, it was necessary to rely heavily on the experience of the project design team and use of precedent evaluations to assess the suitability of the project’s preliminary design, with final design analyses advanced concurrently with construction.

The west end of the Exchange Place Station and the locations of the project’s new tunnel crossovers and platform extensions (shown as dark shaded areas) are shown in Figure 1. The station was originally constructed between the late 1890s and 1908, and became operational around 1910. Very little about the station has been modified since its original construction except for the addition of a new head house and escalator banks, which were added at the station’s east end during 1986. Much of what was known about the station and its connecting tunnel configuration was based on available drawings termed “the 1908 Drawings”, which did not appear to be either construction plans or as-built drawings. As originally configured, EPS could operate only as a “through” station, and could not accommodate “terminal” station operations. This is the reason EPS was forced to close following the tragic events of September 11, 2001. Immediately above the station area, there are four (4) buildings ranging in height from 5 to 30 stories immediately overlying planned excavation limits. Foundation information for most, but not all, of these buildings was available at the time when the project’s new crossover locations were being finalized. The pre-existing tunnels and station are approximately

241 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-30.qxd

2/18/04

12:37 PM

Page 242

Figure 1. New crossovers and platform extensions.

18 m (60 ft) below ground surface and approximately 9.0 m (30 ft) below top of rock on average. More specific details of subsurface conditions are described in the following sections.

3 SUBSURFACE INVESTIGATION Subsurface investigations were planned based on a general knowledge of the bedrock conditions known to exist in the area, the existence of buildings, utilities that precluded drilling exploratory holes, and on the need to get as much geotechnical information as quickly as possible. Ultimately, the exploratory program consisted of the following:

• • • •

29 vertical borings drilled from street level; 2 inclined borings drilled from street level; 3 vertical borings drilled from tunnel level; and 4 inclined borings drilled from tunnel level.

Drilling was performed with truck-mounted drilling equipment from the ground surface and with skidmounted drills from inside the tunnels. Holes drilled from the ground surface were wash-bored and cased down to the top of rock, and then cored with NX sized double-tube core barrels. All collected rock core was preliminarily logged at the drill rig at the time of coring, and was then boxed and transported to PANYNJ’s materials laboratory, where it was logged in greater detail and photographed. See Table 1 for a summary of core recoveries and Rock Quality Designation (RQD), (Deere, 1963), values for each core boring. See Figure 2 for plot of RQD versus elevation for each collect core run. As can be seen in Table 1, average core recovery for the project was approximately 96 percent, and RQD was also reasonably high and averaged 84 percent.

Table 1. Summary of core recovery and RQD. Core recovery (%) Boring

No. Runs

Min

Max

Avg

Min

Max

Avg

450 451 452 454B 455 456 457 458B 459 460A 461 462 465B 467 468C 470A 471A 472 473 474C 475 476 477 478 479B 480D 481 482 483 484 485 486 487 488 489

6 14 12 8 12 1 16 15 1 12 14 17 1 15 14 4 15 11 11 11 13 1 12 14 6 12 7 12 5 10 7 9 8 7 8

86 79 98 96 90 100 90 93 100 92 78 67 96 80 95 95 93 97 85 90 55 88 80 60 90 93 88 50 66 75 85 70 65 68 78

100 100 100 100 100 100 100 100 100 100 100 100 96 100 100 100 93 100 100 100 100 88 100 100 100 100 100 100 100 93 100 100 95 100 100

95 97 100 99 99 100 98 98 100 98 96 95 96 98 99 98 99 99 98 97 93 88 97 93 98 98 96 95 84 86 91 87 87 90 92

42 0 98 24 38 66 34 80 82 77 78 33 84 46 92 65 85 85 0 50 40 75 0 57 77 62 58 0 21 42 23 63 60 48 56

96 100 100 100 100 66 100 100 82 100 100 100 84 100 100 96 85 100 100 100 100 75 98 100 96 100 100 96 93 88 96 94 84 100 97

79 78 100 82 78 66 76 92 82 91 93 85 84 87 98 85 94 95 70 84 83 75 81 85 88 91 81 71 59 69 70 75 74 83 81

242 Copyright © 2004 Taylor & Francis Group plc, London, UK

RQD (%)

09096-30.qxd

2/18/04

12:37 PM

Page 243

Table 2. Summary of rock core UCS testing.

Boring

Figure 2.

455 455 455 455 455 458B 458B 458B 458B 458B 460A 460A 460A 460A 468C 468C 468C 468C 468C 474C 474C 474C 474C 474C Min. Max. Avg.

Distribution of RQD with depth.

Virtually all of the discontinuities encountered in the drilling program were foliation partings. Depending on plan locations, boring were advanced to approximately 1.5 m (5 ft) above existing tunnel crown, or advanced to below tunnel invert elevations, if they were located in rock pillar areas. Select borings were oriented by use of carbide tip scribes fixed to the inside of the core barrels. Orientation of these scribes were measured before the core barrel were lowered down the hole, and the resulting lines scratched down the sides of the collected cores were used to orient the core after it is out of the hole. The presence of buildings directly above planned tunnel excavation limits precluded drilling exploratory holes in these areas from street level. In addition, only limited drilling from within the existing tunnels was allowed because construction was underway and the presence of an exploratory drill rig in the tunnels would hinder the Contractors progress. Also, as a part of the exploratory program, cores were obtained from the existing concrete tunnel linings. A total of 102 cores were obtained to evaluate the thickness and character of the existing concrete lining, and to detect and measure voids between the lining and rock. In general, concrete liner thicknesses were found to be thicker than shown on the “1908 Drawings”. Tunnel sidewalls were found to be “tight” against the rock, and voids upwards of 230 mm (9 in) were

10.21 13.41 16.46 19.51 22.25 11.58 14.63 17.37 20.42 21.95 11.58 15.85 18.90 22.25 10.06 13.11 16.15 18.90 23.77 10.67 13.72 16.76 19.81 23.47

Qu (MPa)

Es (MPa)

Es/Qu

18.6 27.8 22.6 30.9 29.9 26.0 20.6 33.7 20.7 74.6 25.4 51.5 38.7 29.5 38.0 49.5 45.2 26.3 42.2 10.2 9.7 12.2 13.6 39.7 9.7 74.6 30.7

966 1,156 894 1,016 6,012 3,172 3,516 4,544 3,668 7,028 5,013 5,511 4,541 4,572 789 1,049 1,477 3,804 663 2,461 158 1,590 2,508 4,566 158 7,028 2,945

52 42 40 33 201 122 171 135 177 94 197 107 117 155 21 21 33 145 16 241 16 130 185 115 16 241 107

observed at the tunnel crowns. In addition, no steel reinforcing was encountered in the concrete lining. The project elevation datum is elevation 300.0 ft (90.0 m) equals mean sea level at Sandy Hook, New Jersey.

4 LABORATORY TESTING A laboratory testing program was developed to measure rock and concrete strength characteristics. This testing program included 24 unconfined compression strength (UCS) tests and 16 direct shear tests on selected rock core samples, and 16 UCS tests on selected concrete liner core samples. Summaries of UCS test results and direct shear test results from the rock core testing are included in Tables 2 and 3, respectively, and test results for the concrete liner tests are included in Table 4. Compressive strengths for the rock samples ranged from 9.7 to 74.6 MPa (1.4 to 10.8 ksi) and averaged 30.7 MPa (4.5 ksi). However, UCS test results appeared to be highly influenced by foliation, even though the foliation dip angles were typically less than 20 degrees. In addition, strength test results were lower than

243 Copyright © 2004 Taylor & Francis Group plc, London, UK

Sample depth (m)

09096-30.qxd

2/18/04

12:37 PM

Page 244

Table 3. Summary of rock core direct shear testing.

Sample No.

Test orientation

DS1-R7 Parallel DS2-R6 Parallel DS3-R10 Parallel DS4-R7 Parallel DS5-R8 Parallel DS6-R8 Parallel DS6-R9 Parallel DS7-R3 Parallel DS1-R7 Perpendicular DS2-R6 Perpendicular DS3-R10 Perpendicular DS4-R7 Perpendicular DS5-R8 Perpendicular DS6-R8 Perpendicular DS6-R9 Perpendicular DS7-R3 Perpendicular Avg. Parallel Avg. Perpendicular

Cohesion (kPa)

Residual  (deg.)

156.5 98.6 46.2 126.2 48.3 39.3 57.2 22.8 87.6 82.7 77.9 117.2 6.9 2.1 48.3 20.7 74.4 55.4

14.9 25.0 36.3 25.0 31.0 23.4 21.8 26.6 29.5 40.9 29.5 16.7 21.8 13.1 14.9 18.4 25.5 23.1

5 GEOLOGIC CHARACTERIZATION The project is excavated in the Manhattan Schist formation, which can be described as gray, schistose gneiss and schist with occasional pegmatite intrusions. Banding and foliation is locally pronounced, and the dominant structural discontinuity found at the site is foliation. Foliation dip angles range from near horizontal (less than 5 degrees) to approximately 40 degrees, and dip directions varied from northeast to northwest. However, recent geologic deformations have imparted high angle jointing that is overprinted onto the rock mass foliation. Faults or major shear zones were not encountered, and foliation partings were found to be fresh to slightly weathered. At the project site, top of rock slopes towards the Hudson River, and top of rock elevations vary from 90.2 to 83.5 m (296 to 274 ft) from west to east across the site. Available information indicates that the top of rock continues to drop off steeply to the east, as you enter further into the river channel. Overburden thicknesses range from about 4.9 m (16 ft) in the western portion of the site to about 9.1 m (30 ft) in the eastern portion of the site, and consist of man-made fill overlying organic silt, sand, and gravel deposits, which are part of the Hudson River estuary. In general, overburden characteristics were of relative minor concern for the project’s tunnel design analyses, and are discussed no further in this paper. A groundwater table was observed in the overburden at 2.5 to 3.5 m (8 to 12 ft) below ground surface. However, water pressures within the rock mass at tunnel level were not found to exist, based on observed “dry” rock surface conditions upon demolition and removal of the existing concrete tunnel linings.

Table 4. Summary of concrete liner core UCS testing. Track

Station

Qu (MPa)

Es (MPa)

Es/Qu

E E F G G G G/E G/L H H H H H L L L Min. Max. Avg.

1228+00 1229+25 1089+35 1228+75 1229+10 1231+85 n/a n/a 1084+25 1085+45 1088+15 1088+45 1088+90 1229+30 1229+90 1231+70

56.1 36.4 35.7 37.8 48.9 26.0 16.2 13.9 27.4 10.1 41.0 42.2 47.3 24.4 6.7 23.1 6.7 56.1 30.8

4,831 3,325 3,635 4,288 4,211 3,150 1,453 1,388 2,787 1,554 3,831 3,940 3,888 2,104 853 2,410 853 4,831 2,978

86 91 102 113 86 121 90 100 102 154 93 93 82 86 128 104 82 154 102

foliation parting. No direct shear tests were run for joints across foliation, because few were observed in the collected cores and adequate samples could not be obtained for testing. Compressive strength testing of the existing concrete liner cores indicated a range of strengths from 7 to 56 MPa (1 to 8 ksi) with an average strength of 31 MPa (4.5 ksi).

expected based on comparisons with available, published test results for Manhattan Schist and similar schist rock formations (Deere and Miller, 1966; Baskerville, C.A. 1992). Direct shear test results indicated that foliation shear strengths have an average of 65 kPa (9.4 psi) cohesion and an average friction angle of approximately 24 degrees. Furthermore, test results did not appear to depend on the orientation of shearing across the

6 GEOTECHNICAL CHARACTERIZATION Design parameters used in the stability analyses were derived from testing core samples, or in the absence of test data, based on available published information. In addition, analyses were performed using both mean and low range values to assess variable design conditions and concerns.

244 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-30.qxd

2/18/04

12:37 PM

Page 245

Table 5. Summary of geotechnical design parameters. Design values Description

Mean

Low

Rock mass classification RQD RMR (Bieniawski, 1976) GSI (Bieniawski, 1976) Q (Barton et al., 1974; Bieniawski, 1989) Classification from RMR Description

90% 64 74

40% 33 43

15 II Good

2.7 IV Poor

Foliation joints Joint spacing (m) Friction angle (deg.) Cohesion (MPa)

0.6 23 0

0.2 20 0

Steep joints across foliation Joint spacing (m) Friction angle (deg.) Cohesion (MPa)

3 50 0

3 50 0

2,723 34.5 2,068 0.3

2,403 13.8 1,379 0.3

3.95

1.31

Rock mass characteristics Unit weight (kg/m3) UCS (MPa) Modulus of elasticity (MPa) Poisson ratio “m” Strength parameter (Hoek et al., 1998) “s” Strength parameter (Hoek et al., 1998)

7 ROCK SUPPORT ANALYSIS AND DESIGN

0.0056

0.0018

Overburden characteristics Unit weight (kg/m3) Friction angle (deg.) Cohesion (MPa) Poisson ratio Modulus of elasticity (MPa)

1,922 30 0 0.3 149.6

1,602 30 0 0.3 149.6

Existing concrete lining Characteristics Unit weight (kg/m3) UCS (MPa) Modulus of elasticity (MPa)

2,403 31.0 2,758

2,243 20.7 2,068

91.44

91.44

0

6

Water conditions Water elev. in overburden (m) Height of water Above tunnel inverts (m) Rock bolt properties Bar No. Steel grade (MPa) Pre-tension load (kN) Lengths (m) Spans  9.1 m 9.1 m  Spans  12.2 m Spans  12.2 m Spacing (m) Spans  9.1 m Spans  9.1 m In-situ stress ratio

See Table 5 for a summary of design parameters ultimately distilled from available data, published literature and past experience. Table 5 also includes design parameters for the rock bolts ultimately used as part of the project’s initial and final support systems. Ultimate punching shear strengths of 1.51 MPa (219 psi) were used in the stability analyses for plain SFRS materials, and were derived from correlations relating compressive strength and shear strength for plain and fiber reinforced concrete and shotcrete (ACI 1984 and 1988, Fernandez et al., 1979, and Mahar et al., 1975). Composite unit shear strengths for SFRS with embedded steel lattice girders were computed in a similar manner with exception that lattice girder steel cross sectional areas were added to the unit cross sectional area of the lining. For SFRS with embedded lattice girders, ultimate punching shear strengths of 2.70 MPa (391 psi) were used in the stability analyses.

9 517.1 133.4 2.4 3.7 4.6 1.2 1.5 1.5

0.5/2.0

The general approach for stability analyses consisted of evaluating design conditions for a select set of design cross sections that represent expected ranges of conditions. Stability analyses were performed using the software program UNWEDGE (Rocscience, 1997/2002). To analyze stability and develop ground support for discrete wedges using UNWEDGE, the following input parameters were used: (a) dip, dip direction, spacing and strength for three (3) discontinuities; (b) excavation geometry; (c) SFRS unit punching shear strength and thickness; and (d) rock bolt strength, length, spacing and orientation. In addition to these UNWEDGE analyses, evaluation of ground support systems was also performed by use of precedent, and further by use of rock mass rating systems Q and RMR. For each modeled span, multiple sets of discontinuity orientations were used to model potential failure modes, and it was found that sizes (weights) of potential wedge failures were most sensitive to the existence and orientation of high angle joints within the modeled spans. Ultimately, foliation dip angles were set at 15 to 20 degrees and high angle joint orientations were rotated to generate potentially unstable wedges. Hence, UNWEDGE was used to analyze worst case credible wedges generated during this phase of the analysis. For several reasons, not the least of which was the mandated construction completion date, it was decided to use SFRS tunnel lining systems to provide both the initial, short-term ground support and the final tunnel lining. Short-term and long-term design conditions were modeled with various shotcrete strengths (strength gain with time to model curing) and with two different

245 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-30.qxd

2/18/04

12:37 PM

Page 246

water conditions: (1) fully drained (zero head); and (2) 3 m (10 ft) of water head above tunnel invert elevations. These variations in loading conditions and support conditions were intended to model to expected extremes. The proposed new tunnel crossovers consist of excavated spans ranging from approximately 6 to 18 m (20 to 60 ft), with median spans in the range of 12 to 15 m (40 to 50 ft). During the project’s early design phase, much discussion revolved around the effective spans for each new crossover and how much three-dimensional (3-D) effect could be relied upon. Ultimately, it was decided that the rib pillar terminal ends would not be stiff enough or strong enough to provide effective support, because they were too narrow and any 3-D effect would be diminished until the reach under consideration was extended several meters further into the rock mass. It was also concluded that reliance on 3-D arching effects would not be prudent due to the relatively thin rock cover and resultant low confinement, and any 30-D benefits were considered a bonus and not relied upon in design. One method used to provide preliminary assessment of rock bolt lengths for underground support is by use of the “Q” System (Bieniawski, 1989). As described in this reference, preliminary rock bolt lengths can be based on the following formula: (1) where L  rock bolt length (m); B  excavated span Length (m); and ESR  effective span ratio. Using Equation 1 and a range of ESR values from 1.3 to 1.6, preliminary rock bolt lengths would be in the approximate range of 3.0 to 3.7 m (10 to 12 ft) for an excavated span of 18 m (60 ft). Another precedent type evaluation was made using the reference (Cording et al., 1971), which provides rock bolt length and equivalent support pressure data for numerous case histories. Review of several key figures from this reference indicates the following: (a) for excavated spans on the order of 15 to 18 m (50 to 60 ft) rock bolts lengths of roughly 4.6 m (15 ft) have been used on other projects (see Figure 3); and (b) for excavated spans on the order of 15 to 18 m (50 to 60 ft) roof support pressures on the order of 70 to 140 kPa (10 to 20 psi) have been used in the past. From the perspective of equivalent rock loads, 70 to 140 kPa (10 to 20 psi) would be equivalent to 2.6 to 5.2 m (8.5 to 17 ft) of rock having a unit weight of 2,723 kg/m3 (170 pcf). However, this precedent evaluation undertaken with particular caution because, unlike the “Q” System and experience reported by (Cording et al., 1971), shallow excavation depths and thin rock cover conditions were known to exist at the EPS project site.

Figure 3. Rock bolt length vs. excavated span (cording et al., 1971), reprinted with permission of ASCE.

The computer software program UNWEDGE (Rocscience, 1997/2002) provided 3-D visualizations of underground excavations with potentially unstable wedges formed by intersecting discontinuities. The software also allowed for the installation of ground support in the form of rock bolts and shotcrete linings of specified strength, length and/or thickness. UNWEDGE considers rock wedges as infinitely stiff, homogeneous masses acted upon by gravity, water pressure, friction, and applied internal supports. Calculations were performed considering planned new tunnel crossover excavations and assumed rock mass discontinuities to evaluate rock bolt length and spacing, shotcrete thickness and strength, and lattice girder strength and spacing to achieve the desired factorsof-safety for the various design conditions. Sensitivity analyses were performed using UNWEDGE to evaluate the relative importance of the different input parameters including planned support elements (rock bolts, shotcrete, and lattice girders), and rock mass strength and loading parameters (discontinuity strength and orientation, rock unit weight, and water pressure). In these UNWEDGE analyses, rock bolts were assumed to be 28.6 mm (1.125 in) diameter, 517 MPa (75 ksi) yield strength steel (US No. 9, Grade 75) “allthread” bars that were readily available in the NYC metropolitan region. These rock bolts have yield strengths of approximately 334 kN (75 kips), and in these stability analyses, 67% of yield strength or 223 kN (50 kips), was used as the maximum design capacity. Ultimately, using the different evaluation techniques, three (3) different design categories were developed for the project, and these categories were located based on the spans within specific areas of known rock mass

246 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-30.qxd

2/18/04

12:37 PM

Page 247

Figure 4. Typical ground support detail.

quality, and are summarized as follows:

•

•

•

For excavated spans greater than 12 m (40 ft): use 4.6 m (15 ft) long rock bolts at 1.5 m (5 ft) spacing each way in the crown, 2.4 m (8 ft) long rock bolts at 1.2 m (4 ft) spacing each way in the sidewalls, lattice girders at 1.5 m (5 ft) spacing, and 280 mm (11 in) thick SFRS lining. For excavated spans between 9 and 12 m (30 and 40 ft): use 3.7 m (12 ft) long rock bolts at 1.5 m (5 ft) spacing each way in the crown, 2.4 m (8 ft) long rock bolts at 1.2 m (4 ft) spacing in the sidewalls, lattice girders at 1.5 m (5 ft) spacing, and 280 mm (11 in) thick SFR shotcrete. For excavated spans less than 9 m (30 ft): use 2.4 m (8 ft) long rock bolts at 1.2 m (4 ft) spacing each way in the crown and sidewalls, and 150 mm (6 in) thick SFRS lining.

See Figure 4 for a composite section showing the recommended ground support for a typical new tunnel crossover. In the design, pre-existing concrete linings to remain were maintained to the greatest extent possible by installing pre-support rock bolts through the existing concrete tunnel linings, because it was important to minimize excavated spans during construction. Because it was known that the pre-existing lining had voids behind it at some locations, it was necessary to contact grout behind these existing linings before installing and tensioning the specified pre-support rock bolts. Contract Documents were developed and issued to the Contractor specifying prescriptive excavation and support sequences, which were necessary given the project’s critical nature, existence of buildings overlying planned excavation limits and thin rock cover conditions. These construction sequences reduced the sizes of the unsupported spans to the greatest extent practical.

8 CONSTRUCTION OBSERVATIONS During construction it was possible to visually observe the actual rock mass conditions and the size and extent

Figure 5. Completed tracks F to H to L crossovers.

of voids behind the pre-existing concrete lining, and in general, conditions encountered were as expected. The dominant rock mass feature was foliation, but there were also high angle joints encountered in the new crossover excavations, which were anticipated but not encountered in the exploratory boring program. Rock mass joints were tight and unweathered, and isolated water seeps were encountered in a few of the headings, but the excavations were typically dry at times of excavation. Drill-and-blast excavation techniques were initially chosen by the Contractor, but were found to be difficult to implement for several different reasons, and were abandoned in favor of mechanical excavation techniques using road-header type equipment. Mechanical excavations were found to be favorable from both the standpoint of perimeter control and excavation rate of progress. It should be noted that use of road-header equipment to excavate the local Manhattan Schist bedrock is not common practice, and this was just one of the unexpected changes that happened during the work. Installation of rock bolts was completed using small rock drills mounted on skid-steer type equipment and handheld jackleg drills. Installation and tensioning of the resin grouted rock bolts proceeded as anticipated. Application of shotcrete materials were performed only during dedicated night shifts, due to logistics. SFRS materials were batched at remote concrete plants, delivered to the project site by ready-mix transit trucks, and delivered to the station level via slick lines drilled and installed from street level. SFRS was applied with handheld nozzles, and access to upper sidewall and crown areas was completed using small man lift equipment. Steel lattice girders had to be individually fabricated because of each new crossover’s changing crosssection. Each lattice girder was individually identified and no two lattice girders had the same radius, so

247 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-30.qxd

2/18/04

12:37 PM

Page 248

fabrication and timely delivery was a significant logistical challenge. Ultimately, all of the lattice girders were fabricated, delivered, and installed as the excavations were opened up in the predetermined order. See Figure 5 for a representative photo of the completed ground support system. Geologic rock mapping was also performed as excavations progressed to verify design assumptions, and confirm that the specified rock support was adequate, and it was found that the design had an adequate factor-of-safety when considering the actual conditions encountered. 9 CONCLUSIONS The Exchange Place Station Improvements project was an extremely complex underground excavation, support, and infrastructure refurbishment undertaking. It was conceived, planned, and executed under an extremely short schedule, and implementation of innovative techniques, such as the use of road-headers to excavate the Manhattan Schist, and use of a SFRS tunnel lining systems, helped to get the project completed on time. A high level of coordination and communication by PANYNJ and its various design consultants and contractors was also critical to making this project a success. ACKNOWLEDGEMENTS The authors would like to thank PANYNJ for providing the artwork used in Figure 1, reviewing this paper and allowing publication for this conference. The authors would also like to thank their colleagues who worked on the Exchange Place Improvements Project and reviewed and provided comments to this paper. We would also like to thank George Yoggi for providing the photo shown in Figure 4. REFERENCES

ACI Committee 544. 1988. Design Considerations for Steel Fiber Reinforced Concrete. ACI Structural Journal/ September–October 1988: 563–579. ASTM. 2001. Designation D 2938-95 Standard Test Method for Unconfined Compressive Strength of Intact Rock Core Specimens. Annual Book of ASTM Standards 2001, Section four Construction, V04.08, (I): 312–314. ASTM. 2001. Designation D 5607-95 Standard Test Method for Performing Laboratory Direct Shear Strength Tests of Rock Specimens Under Constant Normal Force. Annual Book of ASTM Standards 2001, Section four Construction, V0408, (I): 1353–1364. Barton, N.R., Lien, R. & Lunde, J. 1974. Engineering Classification of Rock Masses for the Design of Tunnel Support, Rock Mech., Vol. 6, No. 4, pp. 189–239. Baskerville, C.A. 1992. Bedrock and Engineering Geologic Maps of Bronx County and Parts of New York and Queens Counties, New York, U.S. Geologic Survey, Miscellaneous Investigations Series, Map I-2003, 2 sheets, Scale 1: 24,000. Bieniawski, Z.T. 1976. The Geomechanics Classification in Rock Engineering Design, Proc. 4th Int. Congress on Rock Mech., ISRM Montreax, Vol. 2, pp. 41–48. Bieniawski, Z.T. 1989. Engineering Rock Mass Classification. New York: John Wiley & Sons. Cording, E.J., Hendron, A.J. & Deere, D.U. 1971. Rock Engineering for Underground Cavers. Symposium on Underground Rock Chambers Herndon, Virginia: ASCE. Deere, D.U. 1963. Technical Description of Rock Cores for Engineering Purposes, Rock Mech. and Eng. Geol., Vol. 1. Deere, D.U. & Miller, R.P. 1966. Engineering Classification and Index Properties for Intact Rock. Urbana, Illinois: University of Illinois. Fernandez, G.D., Cording, E.J., Mahar, J.W. & Van Sint Jan, M.L. 1979. Thin Shotcrete Linings in Loosening Rock. Rapid Excavation and Tunneling Conference. V1: 790–813. Hoek, E., Kaiser, P.K. & Bawden, 1998. Support of Underground Excavations in Hard Rock, A.A. Balkema, Rotterdam. Mahar, J.W., Parker, H.W. & Wuellner, W.W. 1975. Shotcrete Practice in Underground Construction Report No. FRAOR&D 75–90. Washington, D.C. Federal Railroad Administration. Rocscience. 1997/2002. UNWEDGE Users Manual, Rocscience, Inc., Toronto, Ontario.

ACI Committee 506. 1984. State-of-the-Art Report on Fiber Reinforced Shotcrete ACI 506.1R-84. Detroit: American Concrete Institute.

248 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-31.qxd

2/18/04

12:38 PM

Page 249

Session 2, Track 4 Ground modification for underground construction

Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-31.qxd

2/18/04

12:38 PM

Page 251

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Cantilever frozen ground structure to support 18 m deep excavation Dong K. Chang, Peter W. Deming & Hugh S. Lacy Mueser Rutledge Consulting Engineers, New York, NY

Peter A. van Dijk Interbeton, Inc., Boston, MA

ABSTRACT: Artificial ground freezing was used to create a massive block of frozen ground to support an open excavation for a cast-in-place tunnel segment connecting a jacked tunnel with an immersed tube tunnel. This work was done for the Central Artery/Tunnel Project in Boston, Massachusetts. Freezing was performed adjacent to active railroad tracks. The massive block of frozen ground was 9 m wide, 26 m long, and 45 m deep. It worked as a cantilever retaining structure to support an 18 m deep excavation in soft ground (Boston Blue Clay). This frozen ground structure performed effectively for more than 18 months, with only minor deformation. The railroad adjacent to the frozen block operated throughout construction without service interruption. This paper describes the frozen ground structure design and construction, and documents ground performance using field data obtained before, during and after construction.

1 INTRODUCTION Part of the Massachusetts Turnpike Authority’s Central Artery/Tunnel (CA/T) project required three tunnels below active railroad tracks. These tunnels (Ramp D, I-90 West Bound, and I-90 East Bound) were constructed by tunnel jacking through artificially frozen ground that provided a stable material and groundwater control. These jacked tunnels were to be connected to immersed tube tunnels, installed across the Fort Point Channels, by cut and cover tunnel segments. A lean concrete gravity wall and a braced soldier pile and tremie concrete wall were successfully utilized to construct Ramp D cut and cover construction. For I-90 WB cut and cover excavation, localized freezing and soldier pile and lagging earth support wall at the northwest corner were successfully used while the remaining excavations were performed in the soil–cement stabilized ground. Developing an excavation support system for I-90 EB cut and cover tunnels was a more challenging task. Numerous obstructions consisting of abandoned timber pile supported masonry piers, just east of the I-90 EB jacked tunnel, made it impossible to install the planned soil–cement ground stabilization and T-shaped slurry wall excavation support system. After evaluating several options, SIWP (Slattery, Interbeton, J.F. White, Perini, the Contractor) proposed and B/PB (Bechtel/ Parsons Brinckerhoff, the Authority’s engineer)

accepted that the most effective design approach to establish the excavation support system for the 18 m deep cut and cover excavation was to freeze a massive block of ground, approximately 9 m wide, 26 m long and 45 m deep. Mueser Rutledge Consulting Engineers, who was the SIWP’s freezing consultant for the jacked tunnels, was hired by SIWP to develop the ground freezing design. Figure 1 shows the site location plan. Since the freezing operation was on-going for the I-90 WB and I-90 EB jacked tunnels, installing new freeze pipes and providing additional frozen ground mass within the same vicinity of the tunnel freezing areas were a significant advantage and a time saving. The ground freezing subcontractor (FreezeWall, Inc.) indicated that the proposed additional freezing would not require additional refrigeration units. The currently operating refrigeration units had sufficient capacity for this additional freezing work. However, there were several design concerns for implementing the proposed frozen ground design. Safely providing a stable and durable 18 m deep vertical frozen ground face was a technical challenge because frozen ground has rarely been used to support a deep excavation, especially when the excavated frozen ground would be exposed for at least 6 months, including the hot summer season. Another design challenge was to evaluate potential heave and the subsequent thaw settlement resulting from the ground freezing, especially at the adjacent active railroad tracks.

251 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-31.qxd

2/18/04

12:38 PM

Page 252

Figure 1. Site location plan.

2 SUBSURFACE CONDITIONS The project site, bounded by active commuter railroad tracks to the north and west and by the Fort Point Channel to the south and east, was formerly a tidal estuary. Subsurface conditions (in descending order from the ground surface) consist of approximately 10 m of granular fill, 4 m of compressible organic clay/silt, 26 m of marine clay (locally known as Boston Blue clay), 8 m of glacial deposits (till), and argillite bedrock. Groundwater levels generally exist 4.5 to 6 m below the ground surface and are influenced by the tidal fluctuations. However, a confined aquifer exists in the till. The fill contains numerous obstructions, reflecting a century of waterfront construction that had been abandoned and demolished and covered with fill. These obstructions include granite and concrete bridge pier foundations, timber piles, bricks, rubble, and buried abandoned railroad track structures. 3 DESIGN OF THE FROZEN GROUND STRUCTURE The initial frozen ground design approach for the I-90 EB cut and cover tunnel excavation was to install a retaining earth support system consisting of a shallow frozen ground mass supported on deep H piles.

Freezing the shallow ground mass to the tunnel subgrade would provide stable earth support for excavation and the deep H piles would provide vertical and horizontal support to preclude both local and deep stability failures. However, it was determined that driving deep H piles in this area would not be feasible because of numerous obstructions existing in the ground. Thus, use of deep freezing was substituted in lieu of the deep H piles. 3.1

The ground freezing design was subdivided into two parts. One is shallow ground freezing to a depth of about 18 m and the other part is the lower freezing to a depth of 45 m. Figure 2 shows the ground freezing area designations and the locations of the freeze pipe installation. Area B covered the area south of the I-90 WB jacked tunnel and northeast of the I-90 EB jacked tunnel end. This was the primary ground freezing area to support the I-90 EB cut and cover excavation. Both shallow and deep freeze pipes were installed. Areas A and C covered the footprint of the I-90 WB and I-90 EB tunnels and were original freezing areas for the jacked tunnels. Freezing in this area was re-established after the tunnel jacking was completed to provide stable ground above the tunnels facing the cut and cover

252 Copyright © 2004 Taylor & Francis Group plc, London, UK

Ground freezing design

09096-31.qxd

2/18/04

12:38 PM

Page 253

Figure 2. Ground freezing area designations and the locations of the freeze pipe.

253 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-31.qxd

2/18/04

12:38 PM

Page 254

excavation. Deep freeze pipes were added in Area C because the cut and cover excavation would be performed prior to the arrival of the I-90 EB jacked tunnel. The shallow frozen ground in Area B was connected to the Areas A and C frozen ground and provided additional frozen ground continuity to support the excavation. 3.1.1 Shallow freezing The shallow freezing was established with 18 m deep freeze pipes. The pipes were typically placed on a 2 m by 2.3 m spacing. But, at the perimeter of the excavation they were placed on 1.5 m spacing and extended 1.5 m deeper to provide an early frozen groundwater barrier. They also brought the face to the colder temperature to provide sufficient coverage of the exposed frozen ground face for temperature maintenance during the summer months. A total of 65 shallow freeze pipes were installed in Area B. 3.1.2 Deep freezing Deep freezing was established with 43 m deep freeze pipes, penetrating through the marine clay (Boston Blue Clay) and bearing into the till. They were installed in rows perpendicular to the excavation face to provide “barrette” shapes. The deep frozen barrettes were intended to support the shallow frozen ground mass and provide lateral stability. The deep pipes were spaced 2.2 m within a barrette, and each barrette was spaced 4.5 m along the excavation support. A total of 24 deep pipes were installed in the primary excavation support area (Area B). In the eastern end of the EB jacked tunnel area (Area C), a total of 18 deep barrette freeze pipes were added to resist ground movement into the excavation. 3.2

Design analyses

Step by step design analyses were performed to evaluate the proposed ground freezing system. The sections below describe these analyses. 3.2.1 Thermal analysis A thermal analysis was performed to evaluate the ground freezing system. The 2-D TEMP/W FEM computer program from Geoslope, Inc. was used to determine the rate of freezing, the extent of the frozen ground, and the frozen ground temperature profiles. The shallow and deep freezing were modeled independently using the freeze pipes as constant temperature sources. An average brine (freezing) circulation temperature of 25°C was used for both shallow and deep freezing analyses. The freezing influence from both the I-90 WB and I-90 EB jacked tunnel freezing that started much earlier than the present freezing area were included in the analysis as additional cold boundary conditions.

The results of the thermal analysis indicated that the shallow freezing with freeze pipes spaced on a 2 m by 2.3 m should provide sufficient shallow frozen ground mass after 90 days of continuous freezing. Additional freezing would make the frozen ground colder and stronger. The results of the deep freezing analysis indicated that at 90 days of continuous freezing, the ground would be frozen between the deep freeze pipes in each barrette and the area of frozen ground would cover about 45 percent of the area below the shallow frozen ground. At 120 days, the barrettes would merge to become a deep frozen mass and the frozen ground area would cover about 65 percent of the area below the shallow frozen ground. Figure 3 shows a typical output (shallow frozen ground temperature profile after 90 days of freezing) from the TEMP/W FEM analysis. 3.2.2 Strength of frozen ground It was very important to estimate the frozen ground strength to perform realistic stability analyses at various construction stages. In general, frozen ground strength increases as the frozen ground temperature decreases, but the strength at a constant temperature decreases with time because the frozen ground creeps under a constant loading. Table 1 shows a summary of the frozen ground strength properties estimated for the stability analyses. These strengths were based on actual frozen soil laboratory tests performed on undisturbed samples obtained from the jacked tunnel box areas (Deming 2000) and available frozen soil test results obtained from the nearby Russia Wharf frozen ground tunnel project (Lacy 2000). The frozen soil laboratory tests included short term unconfined compressive strength tests and creep strength tests. The creep tests were performed under a constant stress level while measuring strains with time. These test results were then used to develop time-dependent constitutive equations, which define the relationship between strain and applied stress at various creep times (Andersland 1994). 3.2.3 Stability analyses The frozen ground stability analyses were performed to make sure that the frozen ground provides the following stability requirements: (a) the available frozen ground strength should be greater than the stresses induced in the frozen ground; (b) the deep rotational slope stability failure through the frozen ground barrettes after the full excavation should have sufficient safety factors; (c) the local bearing capacity failure through the shallow frozen ground should have sufficient safety factors; (d) the sliding failure at the top of till stratum should have sufficient safety factors. The results of the various stability analyses as described above indicated that after 90 days of continuous freezing, the frozen ground (both the shallow and deep) would provide sufficient strengths to support

254 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-31.qxd

2/18/04

12:38 PM

Page 255

Figure 3. Temperature contours after 90 additional days of freezing (TEMP/W FEM thermal analysis). Table 1. Summary of frozen ground strengths used for stability analyses. Assumed frozen ground conditions Average temperature (°C) 7.5 10

Frozen ground strength

Creep age (days)

Allowable strain level (%)

150

2

150

5

the first 10.5 m of excavation. At 120 days of freezing, the excavation could safely advance an additional 7.5 m of excavation to the final subgrade level. Figure 4 shows a typical cross section with the proposed excavation stages and various stability failure modes that were analyzed. In order to evaluate the most critical stability condition, the stability analyses assumed that the excavation would remain open for 150 days for the cut and cover tunnel construction and the brine circulation temperature would be switched to maintenance temperature after the initial 120 days of freezing to hold an average frozen ground temperature of 7.5°C. The results of stability analyses indicated that the safety factors were about 2.0. 3.2.4 Heave and thaw settlement Heave prediction was an important design issue because the nearby railroad operation would be

Compressive (kPa)

Shear (kPa)

Fill Organic clay Marine clay Fill Organic clay Marine clay

1440 770 510 1630 960 620

720 380 260 810 480 310

Ground Surface (Elev. +35)

FILL

El.+25 El.+21

I-90 West Bound Jacked Tunnel (in-Placed)

ORGANIC SILT MARINE CLAY El.-4.5

TILL

SHALLOW FROZEN GROUND

I-90 East Bound Cut-and-Cover Tunnel Excavation El.+24.5

Local Bearing Capacity Failure DEEP “BARRETTE” FROZEN GROUND

Stage 1 Excavation Stage 2 Excavation

El. +17

Deep Rotational

MARINE CLAY

Deep cement Mix Stabilized Ground

Sliding Failure TILL

Figure 4. Typical section across deep frozen soil mass between jacked tunnels (facing east).

influenced by the heave resulting from the ground freezing. It was assumed that the fill and till strata, which exhibit higher permeability, would not produce volume expansion because the excess pore

255 Copyright © 2004 Taylor & Francis Group plc, London, UK

Frozen soil type

09096-31.qxd

2/18/04

12:38 PM

Page 256

water would drain away quicker than the advancement of the freezing front. The organic clay and marine clay having lower permeability were expected to produce heave. It was determined that the shallow freezing produces about 130 mm of heave directly above the freezing area and would not produce heave in the adjacent railroad area because the ground beneath the railroad had been frozen to the same depth for a long time for the jacked tunnels and would resist (hard boundary) heave in these areas. However, the deep barrette freezing would have the potential to produce heave by lifting up the shallow frozen ground. It was determined that the deep barrette freezing would produce about 250 mm of heave directly above the freezing area and would produce up to 75 mm of heave in the railroad area. However, it was determined by the railroad that the slow rate of heave and the relatively small magnitude of heave would not influence the railroad operation because the railroad would have enough time to make track adjustments. The thaw settlement would start after the completion of the I-90 EB cut and cover tunnel construction and the freezing circulation was turned off. The process of thawing would be a slow process and take more than a year. It was determined that thawing of the shallow frozen ground would produce ground surface settlement of up to 200 mm at the center of the frozen ground. The deep barrette frozen ground would produce thaw settlements of up to 130 mm within the frozen ground area and produce about 50 mm of thaw settlement at a 20 m radius from the edge of the frozen ground. 4 INSTALLATION OF FREEZE PIPES AND INSTRUMENTATIONS The freeze pipes (both shallow and deep) were 114 mm OD steel closed-end pipes. They were installed using an ultrasonic drill rig and installation did not encounter any significant problems. Two types of instrumentation devices, temperature probe pipes and inclinometers were installed to monitor the performance of the frozen ground during the freezing and maintenance periods. Three shallow temperature probe pipes (18 m) were installed in the shallow frozen ground and three deep temperature probe pipes (43 m) were installed in the deep freezing area. Temperature readings were used to evaluate the freezing progress of the shallow and deep barrette. Two deep inclinometers were intended to be outfitted with probe extensometer “Sondex” magnets to evaluate elevation changes due to heave and thaw. However, the installation of the Sondex system was not successful. The locations of these instruments are shown on Figure 2.

5 MONITORING, CONSTRUCTION, AND PERFORMANCE After the freeze pipes were installed and connected to the brine circulation system, the freeze plant was switched on. Frozen ground temperatures were continuously monitored in the deep and shallow temperature probe pipes. The data was initially used to calibrate the thermal FEM computer models and this refined the prediction of the freezing progress. During the freezing and excavation, the use of the inclinometers was unsuccessful. The inclinometers did not survive the freezing environment because water infiltrated into the inclinometer casings froze within the casings. Several attempts were made to melt the ice in the casings, but they were not successful and the use of the inclinometers was abandoned. Because of this, optical survey points were installed on the exposed frozen ground faces, in an approximately 3 m by 3 m grid. The optical survey measured three movements (in and out, up and down, and left and right) at each point. The optical survey was performed twice weekly. After 90 days of continuous freezing, it was determined from the temperature data that the frozen ground had sufficiently low temperatures and it would be safe to excavate the first 10.5 m. The excavation of the frozen ground was sloped back at approximately 20 V: 1 H to prevent overhanging. The exposed face was insulated by spraying on polyurethane foam within 12 hours of excavation. The polyurethane was directly attached to the face with chicken wire mesh, which was pinned to the frozen ground by nails. The nails were installed in an approximately 2 m by 2 m grid. A 10 cm thickness of polyurethane foam was sprayed on and painted a reflective white color to minimize heat absorption from the direct sunlight. After 120 days of freezing, a similar assessment as above was made based on the temperature data and the movements of the exposed frozen ground wall face. It was determined that the frozen ground would provide sufficient excavation support to advance the excavation to the final subgrade. The Contractor performed the additional 7.5 m of excavation to the final subgrade and insulated the exposed face with the polyurethane foam. After the excavation was completed, the brine circulation temperature was cut down to a maintenance temperature. This was intended to save freezing energy, especially because the cut and cover tunnel construction was significantly delayed due to delays in other areas of the immersed tube project. However, the brine temperature was switched back to a lower temperature when the temperature data and movement data showed that the frozen ground was showing a reduction in strength due to warming up or accelerated creep rates. The optical survey data indicated that the frozen ground wall has been stable throughout the

256 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-31.qxd

2/18/04

12:38 PM

Page 257

support an 18 m deep excavation without any bracing system was a success. It provided its primary design objectives without encountering any significant problems and demonstrated that this technology can be used in other difficult construction sites. Installation of freeze pipes was much quicker than the original design (T-shaped slurry elements) because freeze pipes were drilled through the obstructions. The frozen ground structure performed effectively. Deformation of the exposed frozen ground walls was minor and the walls were stable for more than 18 months. Whenever the deformation rate of the wall increased, the freezing system was quickly switched back to the lower brine temperature to gain additional frozen ground strength. This flexible strength control system was one of the advantages of using ground freezing. There were no stability problems for the frozen ground mass. Conservative design assumptions used for the stability analyses demonstrated the effectiveness of the frozen ground mass. The railroad adjacent to frozen block operated throughout construction without service interruption. Both the shallow and deep freezing did not generate significant heave and lateral deformation. Thaw consolidation settlement from the organic clay and marine clay was not significant and did not impact the adjacent railroad tracks.

Figure 5. Foam insulation on frozen walls – areas B and C.

ACKNOWLEDGEMENTS

Figure 6. 18 m high Area B wall – west bound at right.

construction period. During the 18 months of the construction period, the maximum cumulative deformations of the wall were less than 15 mm. It was difficult to establish whether there was any contribution of heave at the railroad tracks from the deep ground freezing because of other freezing and thawing activities that were ongoing at the same time. However, the insignificant vertical movements at the face of the frozen ground wall suggest that the deep freezing had no significant impact on the adjacent railroad track areas. Figures 5 and 6 show an overall view of the frozen ground support walls and the completed base slab of the cut and cover tunnels. 6 SUMMARY Utilizing artificial ground freezing technology to create a massive shallow and deep frozen ground mass to

The authors would like to thank the Massachusetts Turnpike Authority and the Federal Highway Administration for their support in publishing this paper, which summarizes our valuable experience and knowledge gained through this project. The author also would like to acknowledge the contributions of their colleagues at the CA/T project, B/PB, SIWP, and FreezeWall, Inc. who installed and operated the freezing system.

REFERENCES Andersland, O.B, & Ladanyi, B. 1994. An Introduction to Frozen Ground Engineering. Chapman & Hall. Deming, O.W., Lacy, H.S., & Chang, D.K. 2000. Ground Freezing for Tunnel Face Stabilization. In L. Ozdemir (ed.), Proceedings of the North American Tunneling 2000 Conference in Boston, Massachusetts, USA: 383–392. Rotterdam: Balkema. Lacy, H.S., Arland, F.J., & Chang, D.K. 2000. Supporting Historic Buildings While Tunneling Below. In L. Ozdemir (ed.), Proceedings of the North American Tunneling 2000 Conference in Boston, Massachusetts, USA: 383–392. Rotterdam: Balkema.

257 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-32.qxd

2/18/04

12:39 PM

Page 259

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

New chemical grouting materials and delivery equipment technologies Gary N. Greenfield Tunneling Director (Retired) Minova USA

Anthony C. Plaisted M.R.S.C. (UK), M.A.H.P. (USA) Technical Manager (Retired) Minova USA

ABSTRACT: The authors present a critical survey of grouting materials technologies that combine “new or new to America” products and their delivery systems for ground water management, backfill grouting and ground support during mining and recent advances in non-explosive demolition in confined spaces. Emphasis is placed on the significance of the necessary and related delivery systems. This paper reviews the changes that have transpired over time in the industry, which have increased the delays to “timely introduction” of novel materials and their appropriate and recommended delivery or application methods for both mining and tunneling projects in North America. Examples sited include several underground mining and tunneling projects that have encountered severe ground and water infiltration problems. The unintended consequences have included, loss production time or the marginal loss of a mine section. Tunnel advance has been halted due to excess water infiltration, affecting standard backfill grout placement with the potential for contaminating nearby fresh water resources. Alternatives that have been considered to alleviate these problems include modification materials for cement and chemical grouts, allied to grout delivery and placement systems, engineered for the job. Increased limits placed on the permitting and use of explosives for demolition both rock and concrete have resulted in evaluating chemical rock splitting mortars as an alternative procedure. Modification of these materials has provided for wet ground and low temperature applications, making it user friendly. To anticipate that any new technologies can be considered and applied within the context of current Contract Documents, the following fundamentals will be offered. Initial presentation of (a) documented field trial data or related site specific application histories, so that (b) materials application methods and equipment are clearly understood, leading to (c) a timely, initial cost/benefit study and (d) followed by evaluation trials properly supervised and witnessed.

1 INTRODUCTION Products for rock support and ground water management that were introduced in the early 1970s as relatively “new or new to America” with application to the tunneling and underground mining industries saw their acceptance by these industries in varying degrees. Over the past thirty years, beginning with polyester resins for rock bolting through trials and targeted project application, they have helped in contributing to the awareness of a new “language”. Words such as rheology (“quality or state of being, to be deformed or to flow”) and thixotropy (“the property of various gels becoming fluid when disturbed as by shaking”) as well as phrases such as grout containment have appeared in product literature and the technical press. The purpose of this terminology is to properly

describe the function of selected modifiers or polymers for example, that alter the properties of “traditional” cement and chemical grouts. The North American underground mining and tunneling industries face many challenges, not the least of which are environmental regulations placed on current products as well as those being developed to address more effective control of “difficult” ground and manage high rates of water ingress. As this paper is being written, a current tunneling related article mentioned, “that the means of constructing this valuable infrastructure are ever changing”. We suggest that in meeting those “means” with new technologies, we cannot always apply the same old methods and materials in our battle with the ground. We have been reminded more than once over the last several years that there is no good ground anymore! Major metropolitan markets

259 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-32.qxd

2/18/04

12:39 PM

Page 260

in the United States do not have the luxury of “offering up” ideal ground conditions under their metropolis for expanding needed transportation, water and waste water facilities. This paper seeks to bring back into focus the need for planning a return to a more thorough investigative “process” for evaluating and applying “new technologies” that are proven or being developed. 2 PRODUCT INTRODUCTION AND COMMENTARY 2.1

Rock reinforcement

2.1.1 Polyester resins Early in 1970, the word “polymers” took on a new meaning for the North American mining and tunneling industry with the introduction of polyester resins using a peroxide catalyst. Produced in paste form, placed in a sealed, thin poly film, the package resembled an oversized “hot-dog” and was offered in several diameters. These resin cartridges were initially developed and placed in British and French coal, metal and non-metal mines as a fully grouted, un-tensioned rock bolt. Initially bolts that complimented the resin, were threaded, deformed steel (rebar type) rods. Application methods, installation equipment and resin product development continues to evolve, currently offering tensioned, un-tensioned and “resin assisted” mechanical anchorages to complement both steel, fiberglass and wood “bolts”. Commentary Polyester resin bonded rock bolt trials made their initial evaluation in the coal fields of Western Pennsylvania, Virginia and West Virginia under the watchful eye of both the operating companies and the U.S. Bureau of Mines (USBM). Underground metal mines, notably White Pine Copper undertook extensive application trials of resin bonded rock dowels in difficult ground conditions. With additional published technical data from Europe and the results of on-going trials in the United States, the tunneling community began to apply this rock support system to projects, such as the first Straight Creek (Eisenhower) Tunnel in Colorado and the Washington Metropolitan Area Transit Authority (WMATA), at Washington, DC’s DuPont Circle Station. Testing with close proximity blasting to installed bolts was carried out for Virginia Electric Power (VEPCO) at their North Anna nuclear station. This resin system was approved for rock foundation reinforcement on two additional reactor bases. Long term trials, evaluation and a report in 1973 by the U.S. Army Corps of Engineers was also published, following completion of a twelve month trial. The Corps then issued an approval document for use of polyester resin bonded rock bolts in all of their districts in the United States.

The rock abutments in Hannibal Shale on the Clarence Cannon Dam (USACOE) in Missouri saw the drilling sub contractor design and build drilling and resin bolt insertion rigs to maneuver on the tiered rock benches. A detailed Corps document was published on this work, in 1974. These independent documents afforded civil and geotechnical engineers a technical reference on which to base their decision for resin bonded rock bolt placement in subsequent project documents as a permanent rock support system. It may also be noted that today, two US based companies supply the American underground coal market with approximately 90% of the industry’s total roof (rock) support requirements with fully bonded and resin assisted mechanical anchorages. The early performance success of this support method enabled development in the 1970s of small diameter carbide bits and related drill steel, providing one inch (25 mm) holes to be drilled. Mine roof bolt manufacturers soon began producing headed rebar for placement in the standard drive head on mine roof bolters. The principal benefits to the mining companies? Improved roof control safety and reduction in maintenance costs over time, when compared to mechanical anchorage systems. It was the design and implementation of polyester resin bonded bolts to pre-bolt and support overcast site locations (for ventilation) in coal mines that provided experience for applying pre-support methods at a Corps of Engineers (USACOE) tunnel site. Work on the tunnel began as the first heading was mined, initiating a “two pass” drill and blast excavation method from the portal. The technique involved drilling the crown for a twenty (20) foot, coupled rock bolt. The upper most ten (10) feet of the coupled rock bolt was fully resin grouted. The lower portion of the bolt was left un-grouted and tensioned from the crown of the first heading. Pulling rounds to begin the second ten feet or final excavation pass in the horizontally bedded rock exposed the bolt at the bolt’s coupling “horizon” for subsequent placement of a plate, washer and nut. In effect, the rounds were shot against a pre-supported back, minimizing rock over-break and providing tunnel mining safety. 2.1.2 Polymer modified cements Applications in “weak”, water saturated rock and underwater concrete structures offered an opportunity for the development, trial and application of a prepackaged, neat cement based, polymer modified grout to be pumped into bore holes prior to equipment (mechanical) or hand insertion of several types of deformed steel rock bolts. Of significance, the thixotropic properties enabled the cement grout to inhibit dilution from water in a “marine” environment. Equally important, grout was not lost to adjacent fractures in the rock structure, due to grout “stop and stay” thixotropic characteristics (resembling stiff, gritty mayonnaise) when grouting pressure ceased.

260 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-32.qxd

2/18/04

12:39 PM

Page 261

Commentary A request for a calculated quantity of pre-placed non-shrink, non-bleed, controlled expansion grout in thirty (30) “plus” feet of drilled hole with manual insertion of a rock dowel into a “foundation” of fractured rock again proved the effectiveness of a formulated thixotropic cement grout. Grout design enabled conventional grouting equipment to pump grout through a two (2) inch line over 200 feet in length to the collar of the holes. The grout was pumped to the mouth of the grout line, the line then pushed to the back of the individual holes. The rising column of grout gave the contractor a visual indication as to the location of the grout column as the hose was ejected by the rising column of grout. Any ground water that migrated into the drilled holes had no diluting effect on the grout. A pull test to 60% of yield of the Grade 60 rock bolt was carried out in 20 to 24 hours when the grout reached 5000 psi (34.5 MPa). Applications on foundation and rock “high walls” in the Colorado Mountains near Central City provided a proven alternative, offering rapid grout pre-placement and insertion of several hundred rock bolts of up to twenty (20) feet in length. On a project in Kentucky short starter dowels for “keying” shotcrete on four (4) adjacent tunnel portals and rock abutments were quickly placed after the drilled holes were “charged” with the thixotropic grout from a grout pump. Several hard rock mines in Canada found this grout material suitable for pregrouting cable bolts. A mined section could be pre-bolted ahead of the advancing face in the stope. The grout was pumped by hand upwards into the drilled hole. The cable bolt of a predetermined length was then fed by hand into the hole, completing the installation. 2.2

Chemical grouting

2.2.1 Sodium silicates, hardeners and fillers Sodium silicates have a time and job proven history for providing ground stabilization and minimizing ground water infiltration in major tunneling contracts. Combined with a compatible hardener and where severe ground water infiltration is encountered, with the addition of reinforcing fillers, effective reduction of water ingress on tunnel and mining sites has been achieved. Such fillers may include pulverized limestone. A sprayable system was developed for coating and reinforcing specific rock types (St. Peters Sandstone and Austin Chalk) to minimize ground sloughing in utility and transit tunnels. Commentary Critical to performance of sodium silicate based grouts is the selection of the hardener, which should be used at the right stoichiometric ratio to the silicate to insure longevity of the set grout. In an Austin Chalk

formation, a sprayed silicate (40% silicate concentration) coating was applied with a hand “back pack” sprayer during the “off ” mining cycle of the tunnel boring machine (TBM). This procedure allowed the chalk to be protected until shotcrete was applied clear of the trailing gear. Initial trials were first conducted with the contractor, which demonstrated excellent bonding at the shotcrete-“treated” chalk interface. Sodium silicate and a hardener system, comprised of specifically selected organic esters was chosen to shut off water while sinking two shafts to a depth of over 3300 feet in New Mexico. The shafts penetrated seven discrete aquifers. The grout was tested at both an independent laboratory and by the owner prior to approval and was found to exceed Environmental Protection Agency (EPA) standards. Approximately 1,500,000 gallons of grout was placed at the site.

2.2

2.3.1 Non-bleed grouts The parameters of a non-bleed and non-settling grout were seen as essential for improving smaller annuli grout placement for tendon grouting of post-tensioned structures, notably nuclear reactor vessels and cable stayed bridges. North Sea sited oil well drilling rigs were one of the first structures to utilize this type of grout. Application has also involved the grouting of steel tendons in fractured rock to stabilize an underground hydroelectric turbine chamber in Latin America. Critical to product quality and performance in eliminating water bleed was the use of high sheer mixing equipment during the introduction and dispersion of a powdered polymer into the grout. Commentary Testing modified grout anti-bleed performance required introducing a new method. A Gelman filter was chosen to test this thixotropic mixture. Grouts, both control and polymer modified, were placed under pressure (80 psi) to quantify bleed rates. As an example, under then current Post Tensioned Institute (PTI) guidelines, a maximum tendon grouting distance of 125 feet was mandated to insure minimum water loss from a cement grout during grout placement between the protective sheath and tendon. The distance was more than doubled, without bleed occurring with the application of this “new technology” grout on three cable-stayed bridges in the United States. On an existing bridge, located in Mississippi, a protective grease was initially placed within the cable stayed sheaths. During a routine inspection, it was determined that the grease contained sufficient moisture to have initiated corrosion on the cables. The thixotropic, non-bleed cementitious grout was selected to be pumped into the annulus, displacing the grease.

261 Copyright © 2004 Taylor & Francis Group plc, London, UK

Cementitious grouting

09096-32.qxd

2/18/04

12:39 PM

Page 262

This material technology was “transferred to tunneling” with the owners requirement on the Channel Tunnel (UK Drives) for a different approach to backfill grouting behind the segmental lining. An expanded explanation of this development is offered within the “Backfill Grouting” segment of this paper. 2.3.2 Ultrafine cements Micro fine cements received a great deal of attention when first introduced to the civil engineering community. A number of papers presented at the ASCE conference on Grouting and Soil Improvement, New Orleans, 1992, reported on introduction, performance evaluation and continuing investigation of both physical and chemical properties of microfine cements. Included, manufacturing processes which were then being evaluated. One concern that was expressed, the necessity of adding a dispersant to the cementitious powder during grout preparation, to insure that the fine particles were wetted out to counter agglomeration, which would otherwise restrict penetration of the mix during injection into soils. Commentary In subsequent investigations into resolving the procedure for mixing two separate components in preparation for grouting, one US based manufacturer introduced an ultrafine cement grout integrating a dispersant into the “base” product during the manufacturing process. Initial site evaluation and application of the combined ultrafine/dispersant blend was carried out within the concrete lined Air Intake Shaft for the Department of Energy (DOE) at the Waste Isolation Pilot Plant (WIPP) site in New Mexico. Reported testing of the various grouts submitted confirmed that it was the only ultrafine grout of those tested that “met established set characteristics and ease of preparation” necessary in the confined space within the shaft. This grout was also chosen for grouting on the Inter Island Tunnel in Boston, through forward probes on TBM driven tunnels in California near Los Angeles and in underground mining applications in Canada. The availability of sub micron or nano range particles is on the horizon. This new technology will open up a new era in cement grouting which will allow us to tackle severe ground conditions with nano fine particles. Evaluation studies we feel will give us the confidence to exploit this breakthrough at the earliest opportunity. 2.4

Backfill grouting

2.4.1

Control modules and polymer modified grout The development of a grout modifier for improving the properties of a cement based grout, specifically designed for backfill grouting in tunneling is the result of technology transfer, based on work previously

carried out to seek elimination of grout bleed in tendon grouting. Construction of the Channel Tunnel between England and France presented the initial opportunity to provide a polymer modified cement to control ground water dilution of ordinary cement grouts during mining. The grout modifier used on the United Kingdom (UK) seaward and landward drives was derived from original work, previously carried out on cement grout modification for tendon grouting by M. Schupack and A. Plaisted. The result was a cement based grout that could be mixed and held (retarded) for up to six hours in grout cars and lines, before hydration began. The grout was thixotropic, did not bleed and resisted ground water washout during placement behind the lining. Controlled set times were carried out with sodium silicate, introduced at the point of injection on the segments. The existing onboard grout pump on each of six TBMs was connected to a purpose built hydraulic powered control module and catalyst pump. The grout line received a grout “gun” fitted with an in-line mixing element and the gun designed for connection to the individual grout ports in the segmental concrete lining. The owner established a maximum grouting pressure of 75 psi (5 Bar) that was “dialed in” and maintained by the control module. The module also provided for variable control of sodium silicate volume to provide management of selected grout setting times. Commentary The “Chunnel” project provides an example of the collaborative effort that resulted in providing an effective backfill grout program that involved an equipment manufacturer specializing in hydraulic controls and pumps, a mining and tunneling specialty chemicals firm and the owners engineering staff. Tests were witnessed, involving all the parties and carried out in as close to actual backfill grout placement conditions as was practical, prior to formal approval. 2.5

262 Copyright © 2004 Taylor & Francis Group plc, London, UK

Chemical rock splitting

2.5.1 Rock splitting mortar Recent developments in non-explosive demolition agents have permitted controlled demolition of rock and concrete structures, enabling the expansive cementitious mortars now to be placed in bore hole diameters from 1–1/8 inches (28 mm) to 3 inches (76 mm) in diameter. The material can be loaded into deep holes and pumped for underwater (marine) applications. Recent applications where explosives are prohibited involve placing demolition mortar in rock adjacent to utilities and surface structures. Currently these applications include creating access for tunnel construction and clearing rock in preparation for new surface construction. As this paper is being written, product development work continues and

09096-32.qxd

2/18/04

12:39 PM

Page 263

innovative drilling tool accessories are being fabricated for evaluation in the field to improve demolition productivity. Commentary Current construction industry interest in the use of nonexplosive demolition mortars continues to increase. In recent months, as engineering firms and municipalities have identified the growing need to limit the disturbance to adjacent businesses during new construction for surface and sub surface facilities, non-explosive demolition mortars are becoming an alternative method “of interest”.

3 CONCLUSION The need today for a more thorough investigative process in evaluating and accepting new technologies was mentioned in the Introduction segment of this paper. The authors identified materials and referenced several projects that benefited over the past thirty years from these “new or new to America” products. Today, products formulated and designed to focus on ground and ground water control for tunneling and foundation engineering must meet stringent environmental regulations, if they are to be considered. Can we reasonably anticipate that any new materials technologies and their related delivery or application methods that are brought forward have the opportunity to be evaluated? It is hoped that as projects are funded for design and construction, these technologies will be considered as solid “candidates” for inclusion in the contract documents. Listing these additional product options (prescriptive approach) in contract documents have the potential for diminishing any unintended consequences that may occur during our “battle with the ground” on a specific project. What about evaluating product performance involving ground support and ground water management? We urge owners, their retained and in house engineering staff and consultants to consider the following steps be taken during planning and the preparation of contract documents. Request potential suppliers who have come forward to provide documented field trial data or related, site specific application histories. Second, that the supplied documentation provide application methods, clearly identified, including required application equipment. An initial cost/benefit analysis should be carried out after evaluation trials that are properly supervised,

witnessed and performed to closely replicate anticipated site conditions. An Engineering News-Record (ENR) Editorial (October 20, 2003) contained a concluding paragraph that provides a final, fitting comment. “The potential is there for healthy innovation in (tunneling) construction with positive contributions from the technical, experiential and financial sides. Now we just have to use it”. REFERENCES Albritton, J.A. 1974. Rock Bolt Field Tests, Clarence Cannon Project. U.S. Army Corps of Engineers District, St. Louis, MO. 1974. American Society of Civil Engineers (ASCE). Developments in Geotechnical Engineering. 1988. Avery, T.S. Optimizing the Performance of Polyester Resin Grouted Rock Bolts. Hershey, PA. 21p. Annett, M.F. and Stewart, J. 1989. Development of Grouting Methods for Channel Tunnel United Kingdom Segmental Lining. British Tunnelling Society, London. 1989. pp.173–178. Annett, M.F. 1994. Grouts in Tunnelling. SCI Conference, Grouts and Grouting. London, England, 1994. Avery, T.S. and Daemen, J.J.K. 1994. The (In?) Significance of Creep in a Prestressed Polyester Resin Grouted Rock Anchor. Rock Mechanics (NARM) Nelson, P.P. & Laubach, S.E. (eds.) Balkema, A.A., pp. 953–960. Brierly, G. 2001. Tunneling: a Battle Against The Ground. TBM: Tunnel Business Magazine. February, 2001. pp. 32–33. Burke, J. 2001. Building a Third Stage for Carnegie Hall. World Tunneling (WT On Site) 2001. pp. 183–185. Greenfield, G.N. and Plaisted, A.C.1994. New Perspectives in Grouting Materials for the 90s & Beyond. Trenchless Technology, 1994. pp. 35–37. Haywood, H.M. 2000. Contractor Outreach and PreQualification Program. North American Tunneling 2000. Ozdemir, L. (ed.) Balkema, A.A., pp. 125–128. Kendorski, F.S. 2000. Rock Reinforcement Longevity. 19th Conference on Ground Control in Mining, Morgantown, West Virginia, 2000, 6p. Moss, T.A., Phillips, S.H.E. and Smith, D.F. 1993. Saline Tolerant Grout Use at a Nuclear Waste Facility. Conference on High Level Nuclear Waste Storage Facilities, Las Vegas, Nevada. 1993. Nelson, C.R. 1977. Spray Grouting for Tunnel Support and Lining. Underground Space. Vol. 1. Pergamon Press 1977. pp. 241–246. Reed, J.J. and Ortlepp, W.D. 1969. Grouted Bolts for Faster Rock Stabilization. The Mines Magazine, March 1970. (Reprint). Reilly, J. 1997. Owner Responsibilities in the Selection of TBMs. International Tunnelling Association (ITA). Vienna, April 1997.

263 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-33.qxd

2/18/04

12:39 PM

Page 265

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Jet grout bottom seal for cut and cover tunnel Thomas M. Hurley Layne GeoConstruction, Bridgewater, Massachusetts, USA

ABSTRACT: The Massachusetts Bay Transportation Authority, Silver Line transitway is a new transit service from Boston’s South Station to the fast developing South Boston Waterfront. The Silver Line will utilize trolley busses within a tunnel section slightly larger than the bus itself. The tunnel alignment is particularly difficult. The first obstacle to the tunnel alignment is the two seven-story buildings, the second the Fort Point Channel crossing and the third the East Cofferdam. Three distinct tunnel methods are employed for this project. NATM with micro-pile and frozen ground underpinning is utilized at the Russia and Graphic Arts buildings, immersed tube tunnels are employed at the Fort Point Channel Crossing, and cut and cover tunneling with a jet grout base seal are employed at the East Cofferdam. The east side of the project utilized a sheet pile cofferdam to construct a cut and cover tunnel. The cofferdam was approximately 140 long by 50 wide with an excavation depth of 60. Layne GeoConstruction designed and constructed a system of secant and tangent jet grouted columns to stabilize and seal the bottom of a sheet pile cofferdam, for the construction of a cut and cover tunnel. The jet grouting was used in lieu of the contract design that specified deep soil mixing. The jet grouting acted, additionally, as a sub-surface strut, providing toe support for the sheet pile system. Moreover, jet grout columns improved weak Boston Blue Clay soils and provided support for the setting of immersed tube tunnel segments outside the cofferdam.

1 PROJECT OVERVIEW The Massachusetts Bay Transportation Authority (MBTA), Silver Line is a new transit service from Boston’s South Station to the fast developing South Boston waterfront, the epicenter of commercial development in the city of Boston. Former commuter parking areas and empty lots have been transformed into a new Federal Courthouse at Fan Pier, and the Seaport Hotel at the World Trade Center Boston. The City has made strides to improve access to the area by constructing a new Northern Avenue bridge, and rehabilitating the deteriorating Congress Street and Summer Street bridges. The Silver Line will utilize the most advanced Bus Rapid Transit (BRT) system, within a tunnel section only slightly larger than the vehicle itself. The alignment of the new tunnel (Figure 1), begins at South Station, integral with the Central Artery/Tunnel (CA/T) contract C11A1, continues north through CA/T contract C17A1, until reaching the beginning of the MBTA Silver Line contract E02CN15, at the intersection of Congress Street and Atlantic Avenue. It is at this juncture that the tunnel passes beneath two seven-story, historic buildings then crosses the Fort Point Channel

Figure 1. Silver line transitway tunnel alignment.

265 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-33.qxd

2/18/04

12:39 PM

Page 266

en route to its termination at the World Trade Center Boston. The tunnel alignment through MBTA Contract E02CN15 is particularly challenging, presenting several obstacles. The first obstacle is the seven-story buildings, the second the Fort Point Channel crossing, and the third the East cofferdam. Three distinct tunnel methods are employed for this project. NATM with micro-pile and frozen ground underpinning is utilized at the Russia and Graphic Arts building, Immersed Tube Tunnels are employed at the Fort Point Channel Crossing, and Cut and Cover Tunneling with a jet grout base seal are employed at the East Cofferdam.

A4 size paper

Letter size paper

Setting

cm

inches

cm

inches

Top Bottom Left Right All other Column width* Column spacing*

1.2 1.3 1.15 1.15 0.0 9.0 0.7

0.47 0.51 0.45 0.45 0.0 3.54 0.28

0.32 0.42 1.45 1.45 0.0 9.0 0.7

0.13 0.17 0.57 0.57 0.0 3.54 0.28

2 EAST COFFERDAM CUT-AND-COVER TUNNEL The east side of the project utilized a cofferdam to construct the cut and cover tunnel. The cofferdam was approximately 140 long by 50 wide with an excavation depth of 60. Steel sheet piling was driven to depths of 80 and the excavation was internally braced with steel wales and struts, and soil stabilization at the toe of the sheet piling, see Figure 2. 3 SUBSURFACE PROFILE The subsurface conditions (Figure 3) at the East Cofferdam are primarily 10 to 20 of urban fill, underlain by

Figure 3. East Cofferdam subsurface conditions.

Table 1. Cost comparison.

Figure 2. Sheet piling soil stabilization.

Technique

Cost

Soil mixed cross walls at 8 on center and perimeter walls extending down through the clay Jet grout cross walls at 8 on center and perimeter walls extending though the clay Slurry wall (unreinforced) cross walls at 20 on center with reinforced perimeter slurry walls extending through clay

$2,000,000

266 Copyright © 2004 Taylor & Francis Group plc, London, UK

$3,400,000 $3,400,000

09096-33.qxd

2/18/04

12:39 PM

Page 267

25 of organics then 50 to 80 of Boston Blue Clay to reach glacial till. The organic and clay materials were soft to very soft as indicated by average N values between 0 and 4 blows per feet, as shown in Figure 3. Soil stabilization was required, as a result of the Engineer’s base heave stability analysis of the soft clay, to ensure a stable bottom and safe toe support for the excavation support system.

using soil mixing, jet grouting or slurry walls. The Engineer’s preliminary cost comparisons (Table 1) indicated a soil mixed cross wall as the most economical approach. The results of the cost analysis are as follows: The soil mix design recommended by the Engineer is shown in Figures 4 and 5. The design specified 3 wide cross and perimeter walls, using 3 diameter triple augers to ensure continuous walls as they were to

4 SOIL STABILIZATION ALTERNATIVES The Engineer evaluated three methods to alleviate base heave failure. 1. Reinforced concrete slurry walls installed vertically from the bottom bracing level to the till or rock. 2. Stabilization of the clay with individual elements using soil mixing, jet grouting, or stone columns. 3. Extension of the excavation support system down through the clay using cast-in-place cross walls constructed with soil mixing, jet grouting or reinforced slurry walls. Reinforced concrete slurry walls spanning through the clay (approx. 50) along the perimeter of the excavation would result in expected lateral movements of up to 6 and was deemed unacceptable. Soil reinforcement using individual elements of soil mixing, jet grouting, or stone columns was discounted due to the complexity and uncertainty of the interaction between the individual elements and the clay. In addition, tension forces would most likely develop in the elements from the basal heave forces. The Engineer recommended extending the excavation support system downward through the clay with interconnected cross walls acting as subsurface struts,

Figure 4b. Soil stabilization detail.

Figure 4a. Contract specified soil stabilization plan.

267 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-33.qxd

2/18/04

12:39 PM

Page 268

Figure 5. Contract specified soil stabilization profile.

function as structural elements. The minimum unconfined compressive strength was 300 psi at any individual location with a 450 psi average. 5 SOIL STABILIZATION ALTERNATIVE EMPLOYED – JET GROUTING Although the Engineer considered three alternatives for soil stabilization on the project, the cost analyses reflected the conservative nature of the Engineer and specialty contractors when queried for budget pricing prior to bidding. In actuality, the jet grout cross and perimeter wall alternative was the most cost effective at bid time, nearly 50% less costly than the Engineer’s estimate for soil mixed cross and perimeter walls. Therefore the jet grouting option was proposed. The jet grout alternate was submitted using the same geometric pattern as the soil mix design above, except individual 5 diameter jet grout columns would be constructed using the double fluid method, in lieu of the 3 diameter triple auger panels. Figure 6 shows a detail of the double fluid method of jet grouting. The double fluid method is a mix in place jet grouting method where neat cement grout is injected at high pressure (up to 8000 psi) and the mixing process is assisted by a cone of air injected though a co-axial nozzle. The engineer expressed concerns regarding the alignment of the individual jet grout columns for the cross walls due to the potential deviations in the jet grout drill sting at the required depths of 105. The potential deviations were a concern due to the structural

nature of the cross wall acting as a subsurface struts. Although Layne GeoConstruction was confident that a deviation from vertical for the drill string of less than 1/100 could be achieved due to the soft soils underlying the fill, there remained questions about obstructions contained within the fill layer which could adversely affect vertical alignment. After further consideration, a complete bottom seal was proposed to satisfy the vertical alignment concerns. The full bottom seal was economically feasible as jet grouting allows for the creation of discrete columns at any starting and stopping elevations, unlike soil mixing where equipment limitations require treatment from the ground surface to the design tip elevation. 6 JET GROUT TEST SECTION Figure 6 illustrates the geometric layout of the jet grout test section consisting of six 5 diameter columns, spaced at 5 center-to-center. Coring of test columns would be performed to verify geometric and mechanical properties of the improved soil. Test jet grout columns were installed from a depth of 98 to 60 below working grade. Prior to installing the columns all pilot holes for jet grout column and core hole locations were provided a steel casing for the purpose of measuring the deviation from vertical. The pilot hole deviations were measured using a gyroscopic survey which produced rectangular coordinate deviations. The data was used in conjunction with coring data to evaluate the jet grout test section.

268 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-33.qxd

2/18/04

12:39 PM

Page 269

Figure 6. Double-fluid jet grouting detail. Figure 7. Jet grout test section – geometric layout.

Five core borings were taken at the locations in shown in Figure 6 using a size P triple barrel wire line coring system and a diamond coated face discharge bit manufactured by Christensen Products.

• •

• •

Core 1 was sampled at the interstitial space and only recovered traces of grouted soil. Cores 2 and 3 were sampled at the tangential point of two columns and recovery was mostly loose fragments of jet grout that exhibited signs of excessive grinding – however drilling resistance was significantly greater than samples recovered in clay above the treatment zone. Core 4 was sampled at the center of the column and produced recovery of full cylindrical cores. Core 5 was located between columns, but 2.1 from the center, slightly less than tangent. Recovery was jet grout and only small amounts of clay, however, the lower 20 of core had no recovery.

The core borings at the column interface (tangent) exhibited poor recovery due to the non-homogeneity of the jet grouted soil at the outer limits of their intended column diameter. This is particularly evident with the double and triple fluid systems of jet grouting where the assistance of the air turbulence has much less concentration farther from the point of origin (nozzle). It is also important to note that coring systems are designed for hard formations and it is difficult to obtain good recovery of jet grouted clays and silts due to their soft nature. In this case it appears that seams of good jet grouted soil, layered between softer clays, were present and caused the harder jet grouted soil to delaminate from the softer material and shift within the core barrel during the coring process. This produces a grinding effect at the bit, and resulted in poor recovery. Hence the drill resistance indicated the presence of jet grouted soil, although the core recovery did not confirm this fact. A split spoon sample SPT 1 was performed as an alternate method to verify the presence of jet grouted

soil at the tangent interface of two 5 diameter columns. The samples recovered exhibited 8 of jet grout with significant clay with average N-values of 21 blows/ft and 16 of jet grout with little clay, and average N-values of 26 blows/ft. The N-value of the clay prior to jet grout soil stabilization was 0–6 blows/ft. The results of the core borings and SPT borings indicated significant jet grout was present at the tangent between two 5 diameter columns, although the homogeneity of the columns at their extreme limit of 5 diameter was variable. The results were deemed acceptable by the Engineer and 5 diameter columns spaced at 5 tangents were recommended for base heave resistance. Due to the variable nature of the jet grout at the extreme limit of 5 diameter the Engineer recommended a 4 center-to-center spacing along the sheet pile walls to ensure adequate toe support. 7 JET GROUTING PRODUCTION The final geometric layout for production work is detailed in Figure 7. The top and bottom elevations of each column were calculated and tabulated based upon the profile grade line of the cut and cover tunnel, in addition to the approximate elevation for the top of glacial till. Detailed shop drawings were approved and work commenced. Two drill rigs were procured specifically for this work with emphasis on the deep treatment zone up to 105 below ground surface. Figure 8 illustrates a drill rig working on production columns. Each drill rig is fully automated to control the retraction rate of the drill rods, ensuring consistent quality. The grout plant consisted of a bulk cement storage silo, a colloidal cement grout mixer, and a high pressure grout pump. To ensure consistent quality, automated features were employed at the grout plant. The grout

269 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-33.qxd

2/18/04

12:39 PM

Page 270

Figure 9. Jet grouting drill rig. Figure 8. Jet grout production work – geometric layout.

mixer accurately measures the mix proportions of cement and water to achieve the desired specific gravity. This feature was pre-programmed to the required mix proportions and batching was performed using the automated features of the mixing plant. The grout plant technician periodically verified the grout specific gravity, by mud balance, as a check to our automated plant function. Jet grout injection pressure was continuously monitored at the internal pressure gauge of the grout pump, and grout volume was measured using the automatic stroke counter at the grout pump control panel. The information was recorded and submitted for inclusion into the project record. In total over 300 columns were installed to treat the Boston Blue Clay between depths of 60 and 105. The work was accomplished in 2–3 months. Core sampling of production jet grout columns was performed at 10 locations near the center of the columns. Unconfined compressive strength testing was performed by an independent laboratory and yielded average strength of 840 psi. Excavation for the cut and cover tunnel proceeded upon completion of the jet grouting work, yielding a dry and stable bottom surface (Figure 9). The cut and cover tunnel was constructed without any dewatering and the structure was placed on top of the prepared jet grout bottom seal. 8 CONCLUSIONS Jet grouting proved to be the most economical alternative for soil stabilization on this project. Cost estimates for the soil stabilization performed by the Engineer prior to the bid proved to be very conservative. Keen observations by the Engineer and Contractor during the test section uncovered the limitations of core sampling jet grouted soils. SPT borings confirmed the presence of jet grouted soil, while the coring operation observed drilling resistance but variable recovery.

Figure 10. Base of excavation, 60 below sea level.

The SPT method proved to be the more reliable method of confirmation of the presence of jet grout when treating soft soils. The jet grouting was successfully completed in Boston Blue Clay. The jet grouting was performed to depths of over 105, with specialized hydraulic drill rigs and grout mixing and pumping equipment with automated features for optimum quality control. ACKNOWLEDGEMENTS The author would like to acknowledge the support and technical contributions of David Shields of GEI Consultants, Inc. of Winchester, Massachusetts, as well as Pier Luigi Iovino, Division President of Layne GeoConstruction, in the preparation of this paper. REFERENCE GEI Consultants, Inc. July 1998. Final Geotechnical Data Report, MBTA Contract E02CN15, South Boston Piers Russia Wharf and Fort Point Channel Tunnel. Prepared for Frederic R. Harris, Inc. Boston, MA.

270 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-34.qxd

2/18/04

12:40 PM

Page 271

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

North airfield drainage improvement at Chicago O’Hare International Airport: soil stabilization using jet grouting D.A. Lewis & M.G. Taube Nicholson Construction Company, Cuddy, Pennsylvania, USA

ABSTRACT: Drainage improvements at O’Hare International Airport, Chicago, IL included the installation of a drainage and storm water system to control the overflow from nearby Willow Higgins Creek. This involved construction of a weir structure at the creek and the channeling of water from the weir to a newly constructed reservoir via three, 3.7-m (12-ft) diameter, underground storm sewer lines. The sewer lines passed beneath an existing, 2.3-m (90-inch) diameter high-pressure water main, which was to remain in service throughout sewer line installation. The soil composition around the water main ranged from medium stiff clays to silty sands and sandy silts. Triple-fluid jet grouting was used to stabilize the variable soil profile beneath the water main in preparation for tunneling and installation of the sewer lines. The varying soil strata presented a challenge to the project team to establish a constant set of jet grouting parameters throughout the stabilized zone while keeping within the specified unconfined compressive strength range for the stabilized zone of 690 to 1380 kPa (100 to 200 psi) at 28 days.

1 INTRODUCTION Willow Higgins Creek runs within the boundaries of Chicago-O’Hare International Airport. During periods of heavy rain, the creek regularly overflows and floods airport property and nearby businesses and homes. To control storm water overflow, the City of Chicago planned to install an intake weir at the creek to channel the overflow to a newly-constructed reservoir via three, 3.7-m (12-ft) diameter storm sewer lines. These lines passed beneath an existing, 2.3-m (90-inch) diameter high-pressure water main. This water main provides service for Chicago’s western suburbs. The invert elevation of the water main was 193.0 m (633.25 ft), approximately 4.5 m (14.75 ft) below existing grade. Soils in the area of the water main crossing generally consisted of medium-stiff clays from existing grade to elevation 191.6 (628.5 ft). The clay stratum was underlain by silty sand to elevation 190.0 (623.5 ft), beneath which sandy silt extended to the storm sewer invert at elevation 186.8 (613.0 ft). Silty clay was encountered between elevations 186.8 m and 185.5 m (613.0 ft and 608.5 ft), overlaying a silt stratum. Given the flowable nature of the silty sand and sandy silt through which the storm sewers were to be advanced, measures were required to eliminate ground loss or heave during tunneling operations that would threaten the integrity of the water main.

Options considered for protecting the 2.3-m (90-inch) diameter water main included temporarily diverting the flow from the water main, or installation of structural framing to support the water main during installation of the 3.7-m (12-ft) diameter storm sewer lines. These two options were determined to be more expensive and more risky than a soil treatment option. Triple-fluid jet grouting was specified to achieve the required improvement. 2 JET GROUTING TECHNOLOGY The fundamental principle of the jet grouting technique is a high-speed erosional jet acting under a nozzle pressure of up to 50 MPa (7250 psi). The soil structure is destroyed and the soil is eroded, and typically mixed in situ with cement-based grouts to form various geometries, depending on the application. The design of the jet grouting work is often the prerogative of the specialty geotechnical contractor and is based on empirical considerations that take into account the specific subsurface conditions, equipment characteristics, and the specification requirements. Given that the jet grouting technique is virtually independent from the soil texture and structure, it can be applied to a variety of conditions. In the majority of applications, a neat cement-water mix is used, with

271 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-34.qxd

2/18/04

12:40 PM

Page 272

Figure 3. Limits of treatment: Elevation view perpendicular to crossing.

parameters was selected in order to provide treated soil that was stiff enough to control ground movement, but not too hard to make excavation difficult. The geotechnical contractor was faced with two challenges:

Figure 1. Limits of treatment: Plan view.

• •

To establish a constant set of jet grouting parameters throughout the varying soil strata while keeping within the unconfined compressive strength range specified for the stabilized zone. To avoid detrimental impact to the high-pressure water line as a result of jetting operations in close proximity.

4 TEST PROGRAM Figure 2. Limits of treatment: Elevation view in direction of crossing.

initial rheological properties typically characterized by low viscosity and low rigidity. Grout additives can be used in particular applications. Jet grouting has been used in the U.S. since the 1980’s and is now a well-accepted ground treatment technique that can provide innovative solutions to very difficult engineering problems such as those at O’Hare Airport. 3 PROJECT OBJECTIVE The intent of the jet grouting program was to treat the potentially unstable soils in order to eliminate ground loss or heave, and thus minimize movement of the water main, during tunneling operations for the sewer lines. The defined treatment area was 19.5 m by 20.3 m (64 ft by 66.5 ft) in plan and extended from elevation 194.5 to elevation 183.9 (638 to 603.5 ft), approximately three meters (10 ft) below existing grade (see Figures 1–3), providing a stabilized soil mass with an unconfined compressive strength of 690 to 1380 kPa (100 to 200 psi) at 28 days. The range of strength

Test columns were installed between elevations 188.1 m (617 ft) and 192.0 m (630 ft) to encompass the full range of soil conditions encountered during the production work. The geotechnical contractor designed the test program to achieve a target unconfined compressive strength of 690 kPa (100 psi), meeting both seven-day and 28-day strength requirements. During test column construction, cement/water ratios, replacement ratios, and lift and rotational speeds were varied while fluid pressure and flow remained constant. For the initial tests columns, the geotechnical contractor selected experience-based parameters of cement/water ratios of 0.5 and 0.6, and replacement ratios of 40, 45 and 50 percent. The replacement ratio is defined as the theoretical volume of soil replaced by grout. Bentonite was added to the mix for both test and production columns to control grout bleed to reduce unconfined compressive strength. Following the initial series of tests performed with cement/water ratios of 0.5 and 0.6, the owner requested that additional test columns be constructed using higher cement/water ratios in order to assess how workable stronger jet grout columns would be. To that end, test columns using 0.8 and 1.0 cement/water ratios were constructed. The strengths of these columns ranged from 627 kPa to 4447 (91 to 645 psi), and it

272 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-34.qxd

2/18/04

12:40 PM

Page 273

Table 1. Test program strength results. Sample retrieval elevation m (feet) 192.9 (633) 192.0 (630) 190.8 (626) 189.0 (620)

Soil description

Range of strength at 5 days: Cement/water ratios 0.5 & 0.6 kPa (psi)

Range of strength at 5 days: Cement/water ratios 0.8 & 1.0 kPa (psi)

Stiff brown clay, dry Gray sandy silt, wet Gray sandy silt, wet Gray silty sand, wet

152–958 (22–139) 283–1,200 (41–174) 469–848 (68–123) 889 (129)*

2,654–3,771 (85–547) 1,882–4,447 (273–645) 627–3,020 (91–438) N/A

*One test only.

was agreed by the owner and engineer that the treated soil would be too difficult to excavate at these higher strengths. 4.1

Verification

The verification program included:

• • • • •

Continuous monitoring of the main grouting parameters over the full length of the column, including fluid pressure, fluid flow, rate of rotation and rate and withdrawal of the monitor. Excavation and visual examination of test columns for diameter measurement and continuity of construction. Wet grab samples obtained at four elevations and strata for future testing. Spoil sampling. Batch plant sampling.

Core samples were also retrieved and independently tested in the laboratory for unconfined compressive strength. In order to shorten the duration of the test program, the curing time for the jet grout samples was reduced from seven to five days. Table 1 presents the range of unconfined compressive strengths obtained at five days during the test column program. These strengths represent approximately 60 percent of the 28-day strengths based on the interpretation of strength versus maturity relationships for concrete (MacGregor, 1988). 4.2

Production parameters

From the results of the test program, optimum jetting parameters of 0.6 cement/water ratio and 50 percent replacement ratio were determined to achieve 1.68-m (5.5 ft) diameter columns for the production work. 5 PRODUCTION WORK Prior to jet grouting, three monitoring points were established on the crown of the water main within the zone where the pipe passed through the area to be stabilized. These points were installed by hand-excavating

Figure 4. Jet grouting in progress.

to the top of the pipe, placing sleeves through which survey rods could be inserted, and backfilling around the sleeves. Pipe elevation was continuously monitored during jet grouting operations. For the majority of the work, jet grouting locations were designed to be on a 1.52 m (5 ft), center-to-center triangular pattern to construct overlapping, 1.7-m (5.5-ft) diameter columns. Adjacent to the water main, columns were to be constructed on a 15 degree batter along a straight line offset 305 mm (12 inches) from the outer extent of the pipe. Battering of the jet grout columns was necessary to treat the soils beneath the 2.3-m (90-inch) diameter water main. Grouting work was initiated along the water main alignment to create a system of primary columns to support the pipe as the intermediate columns were installed. Drilling and grouting operations were accomplished using a 90-mm (3.5-inch) diameter, triple-walled drill rod mounted on a hydraulic track rig (Figure 4). Cement grout was colloidally mixed at an on-site batching plant and stored in agitator tanks before being fed to the jet grout monitor via highpressure pumps. Given the nature of the jet grouting process, ground disturbance is inevitable. The degree of disturbance is a function of the jet grout contractor’s experience and knowledge. Sequencing of column installation,

273 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-34.qxd

2/18/04

12:40 PM

Page 274

adequate spoil return through the drill holes and venting of the pressures introduced into the soil will typically control, but not necessarily eliminate, heave potential. The specifications required that the contractor exercise extreme caution during grouting operations to avoid causing movement that might result in structural damage. No limit of magnitude of allowable movement was specified. To minimize the potential of pipe movement, the contractor implemented the following controls:

• • •

Sequencing of column installation. Close monitoring of spoil return. Continuous monitoring of the top elevation of the water main.

Despite these measures, during the first phase of column installation 12.2 mm (0.48 inches) of movement was recorded at the center of the section being monitored and 9.14 mm (0.36 inches) of movement was recorded at the east and west monitoring points. Work was halted under the water main, while column installation continued at other locations. When work resumed under the water main on a revised installation schedule, heave of the water main once again occurred during the construction of two columns. The observed movements were judged by the geotechnical contractor to be within limits commonly experienced in the industry for similar applications. The movement was also determined by an independent expert to be within the allowable deflection for the water main as designed. The owner, however, mandated zero movement of the pipeline after the initial movement had occurred. The approach to the installation of columns was modified as follows:

• • • •

The method of top of pipe surveying was changed from an automated laser to an optical leveling system. Drill hole diameter was increased from 152 mm (6 inches) to 200 mm (7.88 inches) to increase the annulus size for increased spoil return. Tighter requirements limiting the proximity of jet grout columns to be installed within the area immediately beneath and adjacent to the water main within the same shift were imposed. Vent holes were installed within treated areas to relieve the build up of pressures beneath the water main.

Completion of the support system under the water main and the remaining non-support columns was successfully accomplished in accordance with project specifications. 6 MONITORING AND TESTING During production jet grouting, on-board instrumentation provided continuous, real-time monitoring of

Figure 5. Horizontal borehole.

Figure 6. Installed sewer lines.

fluid pressure, fluid flow, rate of rotation, and rate of withdrawal of the monitor. Wet grab, spoil and batch mix samples were retrieved and evaluated for unit weight, flow and bleed. Following completion of the jet grouting and installation of the jacking and receiving pit, the tunneling contractor bored three, 406-mm (16-inch) diameter horizontal holes along the center of the alignments of the three, 3.7-m (12-ft) diameter storm sewer lines. These pilot holes were advanced to allow the tunneling contractor to assess the performance of the treated soil. Video inspection of the east and west alignments of the horizontal bores demonstrated that the bores remained open within the stabilized mass after the casings were removed (Figure 5). Extraction of the casing on the west alignment beyond the stabilized zone resulted in soil collapse into the borehole. Once exposed in preparation for sewer line tunneling operations, the stabilized vertical face on the north side of the treated area remained intact and stable. Tunneling was accomplished without inflow of soils into the excavation (Figure 6) and with zero movement of the water main.

274 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-34.qxd

2/18/04

12:40 PM

Page 275

Many details of the sampling and treatment requirements were not adequately addressed in the project specifications. For example, for the pilot holes that were advanced the full length of the three sewer lines, the specifications required that the grout installation be verified by video inspection and the areas of inadequate or ineffective grout be re-grouted and re-inspected at no additional cost to the owner. The methods by which the treated soil was to be sampled were not specified. During advancement of the augered pilot hole, samples of treated soil that had been bored through, churned up, wetted and transported via two sets of augers were collected. Upon collection and visual inspection of the highly disturbed samples, it appeared to some that the soil treatment was not adequate, even though the pilot boreholes remained open and the video inspection showed continuous treatment. It was soon realized, however, that the sampling techniques utilized yielded very low quality and highly unrepresentative samples. The excellent treatment obtained was further demonstrated during the subsequent tunneling operations.

often one of educator in order to furnish owners and their engineers unfamiliar with the technology with a level of confidence in the technique itself and with the performance of the product. The geotechnical contractor should be involved not only with the planning of the jet grouting treatment, but also with the establishment of reasonable sampling techniques and verification requirements, which should come under the specialist’s scope. ACKNOWLEDGEMENTS The authors would like to thank the following for their cooperation on this unique and challenging project:

• • • •

Nicholson Construction Company, Cuddy, PA: Jet Grouting Contractor. Plote Construction, Inc., Elgin, IL: General Contractor. Airport Owner’s Representatives, Chicago, IL: Owner’s Representative. Brunzell Associates, Ltd., Skokie, IL: Pipeline consultant for Nicholson.

7 CONCLUSIONS Over its approximately 20-year U.S. history, jet grouting has gained steady acceptance, particularly as an underpinning and excavation support technique, with many successful applications reported in the technical literature. However, soil stabilization by jet grouting to facilitate tunneling through soft, mixed face or flowable conditions has seen much fewer applications. The role of the specialty geotechnical contractor therefore is also

REFERENCES MacGregor, J.G. 1988. Reinforced concrete mechanics and design. New Jersey: Prentice Hall. Pellegrino, G. 1999. Soil improvement technologies for tunneling: selected case histories. Proceedings of stateof-the-art technology in earth and rock tunneling; ASCE metropolitan section spring geotechnical seminar, New York, 26–27 May 1999.

275 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-35.qxd

2/18/04

12:40 PM

Page 277

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Ground freezing and spray concrete lining in the reconstruction of a collapsed tunnel S.J. Munks & P. Chamley Ove Arup and Partners, London, United Kingdom

C. Eddie Morgan Tunneling, Rugby, United Kingdom

ABSTRACT: During the construction of 14 ft diameter flow transfer tunnel for Yorkshire Water in Kingston upon Hull, UK, a 350 ft length of completed tunnel collapsed within a few days of construction. The tunnel was constructed using a full face Earth Pressure Balance (TBM) and a segmental concrete lining in alluvial and glacial soils with a high water pressure. Artificial Ground Freeze with Liquid Nitrogen as the freeze medium was used to stabilize the ground to permit the reconstruction of the tunnel along its original alignment. Excavation and reconstruction was carried out within the frozen ground using sequential excavation techniques and sprayed concrete primary and secondary lining. This paper describes the design of the Artificial Ground Freeze, the reconstruction of the tunnel and a description of the lessons attained from constructing a major civil engineering structure in frozen ground. Details on the extensive monitoring system and safety measures implemented are also given.

1 BACKGROUND 1.1

Introduction

The flow transfer tunnel was being constructed for Yorkshire Water to direct sewage flows to a new wastewater treatment works being constructed to allow compliance with the European Union’s Urban Wastewater Treatment Directive (UWWTD), which was commissioned and implemented through UK legislation to clean coastal waters. Miller Civil Engineering Ltd (MCEL), now Morgan EST Tunneling, was commissioned to design and construct the Flow Transfer Works, and Ove Arup and Partners was appointed to act as Project Managers. The total scheme budget for the transfer tunnel was £67 million. The conditions of contract for the construction of the flow transfer tunnel were based on the Institution of Chemical Engineers (IchemE) Model form for Process Plant, ‘Greenbook’. This provided a Target Cost reimbursable contract incorporating ‘painshare’ and ‘gainshare’ incentives. The 5.2 mile transfer tunnel was located along the north bank of the River Humber and was designed to provide gravity flow from connections to the existing sewerage system to the new treatment works. The 12 ft internal diameter tunnel was constructed using

an Earth Pressure Balanced tunnel boring machine (EPBM). The tunnel was lined with a pre-cast segmental lining comprising a 6 piece reinforced, tapered ring of 1 ft thickness, fitted with EPBM gaskets. The ground conditions surrounding the tunnel comprised alluvial and glacial deposits. The alluvial deposits consisted of clay, silt, sand, gravel and peat, which lay conformably on the glacial deposits comprising clay, fine to medium sand and gravel. The Upper Chalk was at depth beneath the tunnel and rested unconformably with the glacial deposits. Two aquifers were present along the route. The upper aquifer was approximately 2 m below ground level and was hydrostatic. The second aquifer was tidal and was beneath the laminated clays, situated in the lower glacial deposit and chalk. 1.2

277 Copyright © 2004 Taylor & Francis Group plc, London, UK

The collapse

At 00.30 on 16th November 1999, a locomotive driver reported signs of water inflow at a segment joint which was carrying fine sand into the tunnel, at a point some 650 ft behind the TBM. Despite efforts by the tunnel gang to stem the flow of material, water and sand inflow levels rapidly increased and the tunnel became destabilized such that at 03:00 the same day,

09096-35.qxd

2/18/04

12:40 PM

Page 278

Figure 1. Location of collapse.

Figure 2. Extent of crater.

the tunnel was evacuated and subsequently collapsed. The collapse was focused some 20 ft to the east of a maintenance shaft, known as T3, which was located within a car park close to Hull town centre. Immediately following the collapse the A63 in the proximity was closed and the nearby properties evacuated; both as a precautionary measure. The location of the collapse is illustrated on Figure 1. At the location of the epicentre, the tunnel and ground surface sank by some 8 ft within a ‘crater’ spome 200 ft in diameter and it was calculated that approximately 6560 ft3 of ground entered into the tunnel. The extent of the crater is illustrated on Figure 2, an aerial photograph of the site. In total, approximately 350 ft of tunnel was affected by the collapse and required reconstruction. 1.3

Summary of ground and ground water conditions at the collapsed section

Following the collapse, a detail ground investigation was conducted to provide information for the remedial works and to determine the length of the collapsed tunnel section. The ground conditions surrounding the tunnel alignment following the collapse are summarized on Figure 3. 1.4

Cause of failure

An intensive investigation into the cause of the failure was launched and following extensive physical and numerical modeling, it was concluded that longitudinal differential movement had initiated the collapse. A significant factor which influenced the speed and magnitude of the collapse, was the presence of a thin layer of fine, single size Aeolian (Wind blown) sand

Figure 3. Summary of ground conditions following collapse.

resting on top of the Alluvial Glacial Deposits. This material proved to be highly mobile in the presence of high water pressure and was able to exploit a very minor leak and eventually cause total failure of the tunnel. 2 ARTIFICIAL GROUND FREEZE 2.1

Several options, including Tunnel Diversion, Cofferdam Construction, Jet Grouting and Artificial

278 Copyright © 2004 Taylor & Francis Group plc, London, UK

Selection of remedial works

09096-35.qxd

2/18/04

12:40 PM

Page 279

Ground Freeze (AGF) were reviewed for the reconstruction of the tunnel. With the exception of tunnel diversion, all options considered involved stabilizing the ground at the collapsed section and reconstructing the tunnel along its original alignment. Supporting the ground by AGF and forming the tunnel lining with Sprayed Concrete Lining (SCL) was adopted as this was deemed to provide the optimum solution when considering the local ground conditions, safety, program, buildability and cost. Liquid Nitrogen (LIN) was chosen as the freeze medium as it exists at a lower temperature than other freeze mediums, i.e. brine, and had sufficient cold energy to freeze the saline and moving ground water at the location of the remedial works. In addition, horizontal ground freezing using LIN was the only method that would enable the undamaged tunnel to be captured safely. 2.2

Figure 4. Designed horizontal freeze.

A cross section of the designed horizontal freeze structure is given in Figure 4.

Design of freeze system

Three phases existed for the life of the ice structure at Hull, which is common to many Civil Engineering Projects that implement AGF. The first phase was termed the Primary or Active Freeze Period and comprised the development of the freeze until it reached its design thickness. The second phase was termed the Secondary or Passive Freeze Period and comprised maintaining the ice structure at design thickness during the excavation. The final phase, termed the Thaw Period, occurred after the tunnel lining had reached its design strength and the freeze system became redundant. The freeze system used at Hull was an open system, with LIN, which exists at approximately 196°C, being pumped into a series of freeze tubes and released at an exhaust. As the LIN passed through the freeze holes, the LIN warmed and boiled as heat was removed from the ground and the resulting gas was released from the system at the exhaust. The temperature of the gas at the exhaust formed the control for the system and was termed the Set Point. Extensive testing of frozen soils taken from the post collapse recovery zone was undertaken to establish suitable parameters for the design of the frozen ground support structures. The design of the ice structure took into account the highly disturbed ground conditions that existed surrounding the collapse. For this project, Finite Element analysis (FE) was used to design the ice structure, using data from the extensive testing. The FE demonstrated that each freeze tube had to provide a 2.5 ft freeze radius giving a 5 ft diameter cylinder of ice around each tube. On this basis, the freeze tubes were installed at a maximum of 2.5 ft spacing.

2.3

2.3.2 Pressure Relief Hole A pressure relief hole was used to ensure that the excavation area was enclosed from the surrounding ground and ground water before excavation commenced. The Pressure Relief Hole comprised a slotted

279 Copyright © 2004 Taylor & Francis Group plc, London, UK

Development and maintenance of ice structure

2.3.1 Control of Set Points During the Primary Freeze, the Set Point was placed at a low temperature and once the ice structure reached its design thickness, the Set Point was increased. Various temperatures were used for the Primary Freezes at Hull. Initially, low temperatures in the region of 140°C were used as these achieved the designed ice thickness in the shortest time. Although using such low temperatures allowed excavation to commence quickly, within 14 days of switching on the freeze system, the disadvantage was that it allowed high accumulation of cold energy within the excavation zone; this resulted with high compressive strengths of the soil matrix. For the subsequent excavation stages, the Set Point was raised to the region of 120°C during the Primary Freeze; although this increased the lead-time for excavation to approximately 30 days, the accumulation of cold energy in the excavation area reduced which prevented such high gains of the compressive strength of the soil matrix. This reduced the effort required to excavate the frozen ground. Once the Primary Freeze was accomplished, the exhaust temperature was raised to the region of 90°C and then raised in 20°C increments. The Secondary Freeze exhaust temperature was never raised above 40°C.

09096-35.qxd

2/18/04

12:40 PM

Page 280

pipe inserted the full length of the excavation and fitted with a ball valve and pressure gauge. Operation of the Pressure Relief Hole was simple. Prior to commencing excavation, the Pressure Relief Hole was opened and if no water was emitted, the excavation area had to be sealed from the surrounding groundwater. The Pressure Relief Hole also provided an indication on how the freeze was developing, because as the ice structure developed, the pressure of the water trapped within the excavation area increased; the increase of pressure be monitored on the pressure gauge. 2.3.3 Monitoring of ice thickness To monitor the thickness of the ice structure during the Primary and the Secondary Freezes, monitoring holes, which consisted of cased holes with thermocouple arrays, were positioned along side the freeze tubes. Typically, one thermocouple for every 16 ft3 to 66 ft3 of frozen soil was used. Each thermocouple had a Threshold temperature, which when reached indicated that the ice wall had reached its design thickness. During excavation, it became apparent that the ground being excavated was much colder than predicted and therefore the ice structure was thicker than designed. A graphical means of projecting the ice thickness was devised which comprised formulating graphs by plotting the temperature readings in the thermocouples against the distance from the nearest freeze tube. Assuming a conservative linear approach, best-fit lines were plotted on the graphs, extending from the source temperature in the freeze tube, through the known temperature at the thermocouple, to zero. The predicted ice thickness was the resultant at zero. To maintain a conservative approach, the source temperature was taken as LIN’s injected temperature 196°C. This assumed the greatest temperature gradient and therefore projected a conservative ice thickness. An example of a graph used to predict the ice thickness, is given on Figure 5. This graph demonstrates an ice thickness in the region of 1 m. Using a combination of the graphical and the original method for calculating the ice thickness achieved a greater control over the freeze and reduced the compressive strength of the soil matrix and the consumption of LIN.

Figure 5. Predicted ice thickness.

ground within the excavation was excessively cold, a temperature of 105°C having been recorded, it was determined that the ground had begun to shrink and formed cracks. The cracks were not treated as they did not provide a means of water ingress as the ground was sufficiently cold that water entering the system would freeze before it entered the excavation area. The cracks were visually monitored for movement, as block failure had been identified as a low risk. As previously mentioned, a Set Point of 140°C had been used for the initial construction stage to allow the freeze to develop in minimum time; this caused a significant accumulation of cold energy in the excavation area that was not able to dissipate. The cracks were not encountered on the remaining construction stages when a warmer Set Point was used during the Primary Freeze. 3 RECONSTRUCTION OF COLLAPSED SECTION 3.1

2.3.4 Cracks in first excavation phase During excavation of the first construction phase, cracks were observed in the tunnel face, radiating both further into the excavation and outwards towards the freeze pipes. The reasons for the cracks are given below. As ground freezes, it initially expands and once a certain temperature has been attained, which is dependent on the ground conditions, it shrinks. As the

280 Copyright © 2004 Taylor & Francis Group plc, London, UK

Introduction

The reconstruction of the collapsed section of tunnel was conducted in five stages, with two to the west, and three to the east of Shaft T3. The five construction stages were referred to as West 1 (W1), West 2 (W2), East 1 (E1), East 2 (E2), and East 3 (E3). The length of the construction stages was governed by drilling constraints and ranged from approximately 65 ft to 80 ft in length. The tunnel axis was at a depth of 65 ft below ground level.

09096-35.qxd

2/18/04

12:41 PM

Page 281

Figure 6. Construction sequence.

The construction sequence to the east of the shaft is illustrated on Figure 6. The construction sequence to the west mimicked that to the east, but only had two construction stages. Each construction stage was supported and closed to the surrounding ground and ground water horizontally with a circular ice wall and vertically with a frozen bulkhead. The bulkheads were constructed by drilling typically 23 vertical holes from the surface to below the level of the tunnel invert and freezing the surrounding ground. The horizontal freezes were constructed by drilling typically 33 horizontal or slightly inclined holes to the vertical bulkheads and freezing the surrounding ground. Prior to drilling the horizontal freeze tubes, the freeze bulkheads were replaced with a 16 inch reinforced concrete structural bulkhead with a characteristic strength of 25 N/mm2. To allow the excavation of the first set of horizontal freeze tubes, bulkheads were formed at the tunnel eyes within the shaft. The horizontal drill holes for all construction stages, except W2 and E3, splayed away from the tunnel centre line to allow an increase in the excavated diameter from 18 ft to 25 ft. The increase in diameter was to allow for the construction of a drill chamber within the tunnel to facilitate the drilling of the next advance of horizontal freeze pipes. As the final construction stages (W2 and E3) did not require a drill chamber, the freeze tubes remained at a constant diameter around the tunnel. To prevent the LIN flowing too quickly along the inclined freeze holes, they were fitted with weirs, typically at 20 ft intervals where the gradient was steep and 10 ft intervals where the gradient was gentle. The freeze holes were drilled using directional drilling techniques. Once drilled, the holes were surveyed and as an built plot, to 0.01 inch, accuracy was compiled. If necessary, additional freeze holes were drilled. In general, between 1–3 freeze holes had to be redrilled for each construction stage. Once drilled, the 4 inch cased holes were fitted with 2 inch copper tubing and were plumbed into the nitrogen supply. Copper tubing was used as it is ductile and would not fracture in the cold conditions. Where cold energy was not

Figure 7.

Figure 8. Drill chamber.

required to be dissipated, the freeze pipes were insulated, normally with rockwool or foam. An example of a competent ‘as built’ survey is given on Figure 7 and a photo illustration of a drill chamber is given on Figure 8. 3.2

Excavation plant

The choice of plant was limited as it was governed by the size of Shaft T3. Shaft T3 had been constructed as a maintenance shaft and had a 20 ft internal diameter. A Schaeff Roadheader, a 17 tonne tracked machine with the capability of using either a rotary cutter or a

281 Copyright © 2004 Taylor & Francis Group plc, London, UK

‘As Built’ horizontal freeze.

09096-35.qxd

2/18/04

12:41 PM

Page 282

hot water, in addition a geocomposite insulating layer was attached to the side of the tunnel prior to spraying the shotcrete. These measures ensured that the concrete gained sufficient strength before the cold halted the cement hardening reaction. The primary lining had a design strength of 25 N/mm2 at 28 days. Prior to spraying the secondary lining, the tunnel was heated and a waterproof membrane was applied to the primary lining. The 16 inch shotcrete secondary lining comprised a steel reinforced shotcrete layer and was applied following the completion of the construction stage. The secondary lining had a design strength of 45 N/mm2 and was formed from the same mix as the primary lining without the steel fibres. The steel fibres were omitted as structural reinforcement was included in the secondary lining. The secondary lining was tested at 3, 7, and 28 days for strength, with 5 cores for every 300 ft3 sprayed, providing results were constant. If the results were not constant, 5 cores were tested every 150 ft3, in accordance with the Method Statement.

Figure 9. Schaeff.

pneumatic breaker was used for excavation. The Schaeff was lowered into the shaft in pieces and assembled in the pit bottom. In areas where the Schaeff could not be used, a smaller tracked machine with a breaker was utilised. Hand mining was used to trim the edge of the excavation to obtain the profile yet was intentionally limited due to Hand Arm Vibration Syndrome, a condition which is aggravated by the cold conditions. A plate illustrating the Schaeff being used during the excavation stage is given on Figure 9. 3.3

Excavation and tunnel lining

The two excavation stages adjacent to the shaft (W1 and E1) could not be conducted concurrently due to a lack of space. Once the tunneling operations had progressed away from the shaft, excavation proceeded concurrently to the east and west. Prior to the excavation of any of the construction stages and at 6 ft intervals along the tunnel drive, 4 probe holes were drilled to identify if there were any isolated spots of saturated ground. The tunnel was advanced in 3 ft sections, split into crown, bench and invert, with steel lattice girders used to maintain the tunnel profile. Following the completion of each 3 ft advance of the crown, bench or invert, a 12 inch shotcrete primary lining was applied to the tunnel. The shotcrete mix had been designed following extensive trials in cold conditions and comprised an accelerated mix with steel fibres. The steel fibres were used to act as reinforcement to aid in crack prevention. To aid in overcoming the cold conditions, the shotcrete was batched using heated aggregates and

3.4

Challenges faced during the excavation of the tunnel comprised the removal of a locomotive and the removal of the plates from the collapsed tunnel. The locomotive abandoned in the vicinity of Shaft T3 during the collapse. The locomotive had to be cut and removed in sections as excavation progressed as the ground surrounding it was frozen. All excavation that took place around the locomotive was manual to ensure that there was no leakage of hydraulic oils, diesel or battery acid. The segments from the collapsed tunnel were removed as excavation progressed. In some instances, such as in the epicentre of the collapse, the collapsed tunnel segments extended beyond the excavation zone and had to be carefully removed to ensure that the freeze wall was not ruptured. Occasionally, freeze holes pierced through the collapsed tunnel segments. In such instances, the segment had to be removed by cutting along the excavation profile with a diamond saw to avoid dislodging the freeze pipe.

4 MONITORING 4.1

Ground surface

The monitoring regime at the ground surface was designed to confirm stability of the works from the excavation and to monitor the ground movements post collapse. The monitoring was conducted on a 24 hours basis during construction and gradually reduced to monthly readings.

282 Copyright © 2004 Taylor & Francis Group plc, London, UK

Anomalies in excavation

09096-35.qxd

2/18/04

12:41 PM

Page 283

Within 24 hours of a freeze system being switched on, heave was noted at the ground surface in the order of 0.07 inch. Total heave above frozen sections was of the order of 0.5 to 1.5 inches. 4.2

In-tunnel

Monitoring of the in-tunnel deformation was performed using precise three-dimensional surveys of arrays of Bioflex targets spaced along the tunnel at intervals of between 16 ft and 52 ft. Accuracy of the readings, which was affected by the quality of the targets and the atmospheric conditions in the tunnel, in particular the temperature, was 0.02 inch. The results of the deformation monitoring were processed using the Dedalos tunnel deformation program. This allowed the presentation of the vertical, transverse and longitudinal movements at each array to be plotted against time. The targets were positioned with one in the crown, two at shoulder and two at the knee. Following the installation of the secondary lining, the targets were repositioned and subsequently monitored. The results from the monitoring were reviewed daily against the trigger, action, and evacuation levels, set at 0.3%, 0.7% and 1.3% strain respectively. These levels were devised from the initial design of the ice structure and related to the creep and the loss of compressive strength of the ground over time. Only in one instant did the convergence readings exceed 0.3%. Monitoring of heave at tunnel level was conducted by monitoring the segmental tunnel adjacent to the collapsed section. A heave in the region of 1.2 inches was recorded, which correlated with the predicted values. 5 SAFETY A comprehensive safety system was installed, which included the normal rigorous safety procedures for working in confined spaces, in conjunction with incorporating systems for working in close proximity to LIN. A summary of the safety systems installed specifically for working in close proximity to LIN in a confined space, are given below:

•

•

Emergency Stop Buttons – these were installed both in the pit bottom and at ground surface. Once

• •

• •

6 CONCLUSIONS A part of the success of the remedial works at Hull is attributed to the collaborative effort under taken by all those involved on the project, which comprised Yorkshire Water, Ove Arup and Partners, Miller Civil Engineer and their sister company BeMo. The project used challenging civil engineering techniques and it was essential that all parties involved worked together as a team to resolve the problem. ACKNOWLEDGEMENTS The authors would like to thank Yorkshire Water for providing approval to submit this paper and www.petersmith.com for usage of photographs.

283 Copyright © 2004 Taylor & Francis Group plc, London, UK

pressed, the LIN would be cut off from the system, however the LIN would not stop being emitted until the LIN already in the system was released. Ventilation – forced ventilation was implemented at all times. Furthermore, the vent system was linked to the Nitrogen Sensor so that in the event of a leak, the vent system would switch to double speed to force all the LIN fumes from the excavation. Protection – the freeze pipes were protected by the installation of steel stools fitted with ambi-decking in the shaft to prevent damage from the materials being hoisted. Personnel and PPE – An intensive site safety induction was put in place informing personnel of the risks of working in close proximity to LIN. The self-rescue kits for working at Hull were specially designed to work in the temperatures that would be experienced from a LIN leak. Numerous ‘Tool Box Talks’ were given to the operatives to inform them of safe working practices when within close proximity to LIN. All personnel that worked in the tunnel were provided with specialist PPE. The nitrogen pipes were located within covered concrete trenches, with removable steel covering. Others – the pit bottom was decked out to allow access to the freeze pipes at all times. Also, as a preventative measure, the shotcrete equipment was constantly on standby.

09096-36.qxd

2/18/04

1:45 PM

Page 285

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Ground freezing for urban applications P.C. Schmall, D. Maishman, J.M. McCann & D.K. Mueller Moretrench American Corporation, Rockaway, New Jersey, USA

ABSTRACT: Ground freezing occupies a small, but nevertheless important, niche in underground construction where complete groundwater cut-off and in-built ground support are required, but difficult or disturbed ground is present and/or deep excavation is planned. Moreover, since ground freezing is a non-invasive technique that does not disturb the ground, it is particularly well suited to the challenges encountered in urban environments. In conditions that preclude ground displacement techniques, or where the vibration or settlement potential associated with conventional shoring techniques is a concern, ground freezing offers a viable, safe alternative. This paper presents a brief introduction to ground freezing technology, together with case studies of ground freezing performed for a range of applications, including the most extensive urban ground-freezing project to date, on Boston’s “Big Dig.”

1 INTRODUCTION Many urban construction sites are characterized by restricted access, existing structures adjacent to the proposed excavation, active utilities proximate to or within the work zone, and a groundwater table above the bottom of the proposed excavation. Ground support options become limited if vibration or potential settlement associated with conventional shoring techniques are unacceptable, urban subsurface obstructions limit the use of displacement techniques, and/or access restrictions do not permit large construction equipment. The presence of difficult or disturbed ground may also preclude the use of other techniques. When the project calls for a watertight excavation, the ground support options become even more limited, particularly at greater depths. However, the use of ground freezing overcomes these challenges. Ground freezing is accomplished through smalldiameter, closed-end pipes placed in pre-dilled holes. The ground remains undisturbed, eliminating the difficulties associated with displacement techniques. Ground freezing can be accomplished in the full range of soils, from clays to cobbles and boulders, and in pervious or fissured rock, and frozen walls can be formed around underground structures or obstructions. Compared to other groundwater cut-off or excavation support methods, a frozen wall is easily connected to the underlying bedrock and will also adhere to adjoining subsurface installations, if necessary, to provide a composite cut-off structure. By providing a

complete groundwater cut-off, freezing does not impact the surrounding groundwater regime, which is often contaminated in urban areas. 1.1

Simply put, the principle behind ground freezing is the use of refrigeration to convert in situ pore water into ice through the circulation of chilled calcium chloride brine, or, in some circumstances, the evaporation of liquid nitrogen. However, the process is actually quite complex and requires a specialist contractor skilled in refrigeration, thermal analysis, groundwater flow and geotechnical engineering. To create a frozen earth cofferdam, or frozen soil mass, the closed-end freeze pipes are inserted into drilled holes in a pattern consistent with the shape of the area to be stabilized and the required thickness of the wall or mass. A frozen shaft may require freeze pipes hundreds of meters deep. As the brine moves through the pipes, heat is extracted from the soil causing the ground to freeze. The brine is returned to the refrigeration plant through an insulated header and, after re-cooling, is re-circulated within the closed system. The ice acts as a bonding agent, fusing together particles of soil or rock to increase the strength of the mass and render it impervious. Ground freezing is primarily used for shaft and tunnel construction or for the construction or excavation of other underground structures. Once the structure is completed, refrigeration is discontinued and in most cases the ground returns to its normal state.

285 Copyright © 2004 Taylor & Francis Group plc, London, UK

Ground freezing technology

09096-36.qxd

2/18/04

1:45 PM

Page 286

2 MASS FREEZING Although typical applications of ground freezing involve the creation of a peripheral frozen structure for excavation support and groundwater control, there are circumstances under which massive volumes of soil need to be stabilized to facilitate excavation within the frozen, stabilized ground. Such circumstances include projects where a broad spectrum of subsurface conditions exist, water control is paramount, and minimal intrusion or disturbance to the subsurface stratigraphy is required. Ground freezing meets all of these requirements.

2.1.1 Ground freezing design The ground-freezing program was designed by the ground-freezing subcontractor to provide multiple geotechnical functions including groundwater cutoff, encapsulation of the fill debris within a matrix of frozen ground, and improvement in the strength of the organics and marine clay along the tunnel alignment. The freeze system was composed of:

• • •

2.1

Case Study: Central Artery/Tunnel, Contract 9A4, Boston, MA

Jacking of three, massive tunnels just beneath the seven, active Amtrak lines serving Boston’s South Station and the financial district, and through what has been described as “… the most difficult soil conditions imaginable,” is acknowledged as the most demanding component of the overall Central Artery/ Tunnel Project (Rogers and Taylor 2003). The jacked tunnel method was selected to allow full rail service to the station during tunnel construction. The soil profile through which tunneling would take place included 6.1 m of historic fill containing building debris, granite seawalls, piles, wharf structures, and abandoned brick structures (Fig. 1). This was underlain by soft, organic material and a marine deposit of Boston Blue Clay. Groundwater was encountered approximately 3 m below ground surface. Preliminary stabilization concepts included dewatering, grouting the fill stratum and organics, and soil nailing the marine clay. With just 2 m of cover between the box and the rail tracks, the General Contractor was concerned with the potential for unacceptable heave generated by grouting, and also settlement, with removal of the numerous obstructions within the fill, and elected to use ground freezing to stabilize the excavation face.

Figure 1. Excavation of frozen, open-work rubble fill in Boston.

• •

Accurate spacing and freeze pipe location was critical to the success of the stabilization program. Finite element modeling was performed to determine the transient heat flow from the ground to the freeze pipes in order to evaluate freeze pipe disposition, freeze formation period, and freeze plant capacity. With the majority of the excavation work undertaken in the Boston Blue Clay, which also required the greatest refrigeration effort, all design work and detailed thermal analyses were focused on this stratum (Donohoe et al. 2001). 2.1.2 System installation and operation Freeze pipe installation was accomplished from the rail tracks using sonic drilling techniques. Highly maneuverable, high-rail mounted equipment was utilized to precisely locate the pipes among the complex track switchgear, and the sonic drill head could readily penetrate many of the obstructions. This was invaluable in allowing the contractor to maintain the required pipe configuration (Donohoe et al. 2001). Chilled brine was circulated through the freeze pipe system for three to four months prior to tunneljacking. A computerized instrumentation system, also designed by the ground-freezing contractor, was installed and remotely monitored to ensure frozen ground conditions prior to the start of tunneling. The ground-freezing program provided a stable and predictable, completely unsupported, 24.4-m wide by 12.2-m high vertical face that allowed the use of an open-faced tunnel shield (Fig. 2). The tunnel was jacked in place without incident, or without disruption to rail service, and the use of ground freezing was estimated to have resulted in a $4 M saving (Angelo 1999). Ground freezing was utilized for all three tunnel jacks.

286 Copyright © 2004 Taylor & Francis Group plc, London, UK

A massive, centrally-located refrigeration plant, A brine circulation system capable of pumping 15,000 liters per minute to any/all of the three jacked box areas, An insulated brine supply and return manifold system installed completely within the rail track structure, Heat pipes to control the lateral growth of frozen ground beyond the box perimeter, and In excess of 2000, 115-mm O.D. steel freeze pipes installed from 14 m to 18 m below ground surface

09096-36.qxd

2/18/04

1:46 PM

Page 287

Figure 2. Excavation of frozen, free-standing face in Boston.

3 DEEP SHAFT EXCAVATION Ground freezing has been used to allow shafts to be sunk in water-bearing ground to considerable depths. The technique is ideally suited for mixed ground conditions varying from highly permeable sands and gravels to clay and rock, and difficult ground conditions commonly encountered at the soil/rock interface where the geology is generally the most challenging and where displacement or improvement of the ground by other techniques is impractical. With ground freezing, the difficult ground does not need to be replaced and is simply incorporated into the frozen structure. The frozen ground conforms perfectly to the contours of the rock surface, and provides a continuous, high-strength, impermeable material through the soil/rock interface. On any deep shaft project in water-bearing ground, ground freezing is the most practical solution at depths greater than 30 m, and becomes more economic with more difficult ground and with increased depth. 3.1

Case Study: New York City Water Tunnel No. 3, Shaft 22B, Brooklyn, New York

Access Shaft 22B for the New York City Water Tunnel No. 3 is located in Lower Brooklyn, and is bounded on two sides by multi-story apartment buildings and on the other two sides by busy thoroughfares. Subsurface conditions through which the shaft was to be excavated consisted of approximately 9 m of fill, with groundwater encountered 3 m below the surface. The fill was underlain by a 1.5-m thick organic peat layer, beneath which granular sands and gravels extended to a depth of approximately 44 m below grade. The remains of the Gardiner’s Clay formation was

encountered from 44 m to 46 m, effectively creating an upper and lower aquifer within the shaft. Granular soils extended from below the Gardiner’s Clay to decomposed rock at 50 m, with sound bedrock encountered at a depth of 67 m. The shaft was designed with an excavated diameter of 13 m from ground surface into competent bedrock and a decreased diameter for the remainder of the shaft, terminating at approximately 213 m beneath the surface. Ground freezing was used to provide a complete groundwater cut-off and structural support for shaft excavation in the overburden. Freeze pipes were installed using rotary sonic drilling methods. The 2.5-m thick frozen wall was formed using 45, vertical freeze pipes on a 16-m circular pattern, extending 3 m into bedrock. Since the completed frozen wall would extend within 6 m of the adjacent apartment buildings, foundations were carefully monitored for movement during ground freezing operations. Freeze work was completed within eight weeks without damage to the adjacent structures, allowing the General Contractor to begin excavation (Fig. 3). 3.2

Shaft 23B, the break-out shaft for the New York City Water Tunnel No. 3, Stage 2, is located in a heavily trafficked area of South Brooklyn near the mouth of the Brooklyn-Battery Tunnel and at the junction of the Gowanus and Brooklyn-Queens Expressways. At the break-out location, the water tunnel lay more than 105 m below the top of biotite gneiss bedrock, which was overlain by 40 m of soft overburden soils. These consisted of 6 m of surficial fill overlaying

287 Copyright © 2004 Taylor & Francis Group plc, London, UK

Case Study: New York City Water Tunnel No. 3, Shaft 23B, Brooklyn, New York

09096-36.qxd

2/18/04

1:46 PM

Page 288

Figure 3. Shaft 22B, New York City Water Tunnel No. 3.

1.5 m of peat bog (the original shore line), beneath which glacial sands extended to top of rock. Near the soil/rock interface, the sands were interspersed with boulder beds. Groundwater was encountered at approximately 4 m below ground level. Ground freezing was selected to provide watertight support for excavation of the 12-m diameter shaft into bedrock. Forty five, vertical freeze pipes, were installed in an 18-m diameter, circular pattern around the shaft location, and seated 1.5 m into bedrock. During drilling for freeze pipe installation, previously unknown, old building foundations were encountered at a depth of 2.5 m in four of the drill holes, but were successfully penetrated. With the frozen cofferdam in place, the outside edge of the frozen ground extended within 3 m of a multistory apartment building. The use of ground freezing eliminated the potential for foundation subsidence and superficial structural damage, and the technique was later used for the construction of a further four shafts along the route of Water Tunnel No. 3. 3.3

Northeast Ohio Regional Sewer District, Mill Creek Tunnel, Phase 2

During periods of heavy rainfall, the capacity of the Southerly Sewage Treatment Plant in Cuyahoga

Heights, Ohio, is insufficient to handle the additional volume. The Northeast Ohio Regional Sewer District therefore elected to construct a 6-m diameter, 13-km long tunnel through bedrock at 60 m beneath the surface to divert and hold rainwater and residential/industrial wastes until periods of low demand at the plant. The cost of constructing such tunnels to act as reservoirs is more economical than increasing sewage plant capacity. Two of the four access shafts required for tunnel construction lay in an area that was a river valley some 10,000 years ago. Groundwater was present at approximately 18.3 m below the surface. Soils below the groundwater table were primarily glacial, highly permeable, residual, silty sands. A coarser sand and gravel layer was encountered immediately above top of rock. Dewatering to the top of the bedrock was rejected because of the potential for ground loss within the saturated, unstable soils at the soil/rock interface. The Construction Manager for the shaft excavation therefore specified ground freezing to avoid the potential for soil loss in this difficult stratum while at the same time providing groundwater control and support of excavation for the full length of the shafts. Vertical freeze pipes were installed around the proposed, 9.75-m diameter shaft locations and seated 3 m

288 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-36.qxd

2/18/04

1:46 PM

Page 289

Figure 4. Frozen shaft, Mill Creek tunnel. Figure 5. Frozen ground excavation support for tank removal at Huntington Hospital.

into rock. Calcium chloride brine circulating through the freeze system created a 3-m thick, watertight cofferdam through which excavation was able to proceed without incident (Fig. 4). 4 LIQUID NITROGEN FREEZING On smaller projects where the ground is maintained in a frozen state for a relatively short period of time, or in emergency situations, liquid nitrogen is often used as the freezing agent. Although a day of freeze formation with liquid nitrogen is more expensive than with brine, colder temperatures are achievable, allowing the liquid nitrogen to freeze the soils much more rapidly, typically in days rather than weeks, thus offsetting the unit cost. Nitrogen is typically cost effective when the freeze must be formed quickly and remain in place for a limited period, and little or no maintenance of the freeze is required.

ground freezing to a depth of 6 m to provide the structural support needed during excavation and tank removal. Liquid Nitrogen was selected as the freezing agent rather than calcium chloride brine because the excavation would only be open for several days. The freeze wall would be constructed with closely spaced freeze pipes in the form of an arch surrounding approximately two thirds of the excavation zone, leaving one portion open. This would be sloped back during excavation for ease of tank removal. To increase the strength of the freeze wall, water was added to the sands by sprinkler hoses during the initial freezing period. The system was installed with a minimum amount of disturbance. The freeze was formed in a few days, and the tanks were successfully removed in one day (Fig. 5). 4.2

4.1

Huntington hospital, Huntington, New York

Removal of two, abandoned 76000-L oil tanks at Huntington Hospital presented a number of difficulties. The tanks were buried 4.6 m below the surface, immediately in front of one of the hospital’s building walls and close to operating room facilities. Furthermore, a 9500-L diesel underground storage tank was in the immediate vicinity of the tanks to be removed. Prevailing soil conditions consisted of moist sands, with the natural water table well below the bottom of the proposed excavation. A conventional earth support system for the proposed excavation generated significant concerns with the hospital administration. Given the proximity of the highly sensitive operating rooms, vibrating or driving steel sheet piling or H-beams was not permitted and, if conventional drilled in soldier beams and lagging were to be used, there was a potential for ground loss. The earth retention contractor therefore proposed

At the Newtown Creek WWTP, a new 1.5-m diameter transmission line was to be installed using microtunneling techniques. The access shaft to launch the microtunneling operation was located in an extremely congested area of the plant, adjacent to active tanks, and was supported by tight steel sheeting on all four sides. Subsurface soils were a mix of native silty sands and organic silts, with groundwater present approximately 3.0 m below ground surface. The invert of the 1.5-m diameter “eye” through which the microtunneling machine would be launched was at 6.0 m below groundwater level. During excavation of the shaft, a split was uncovered in the sheeting below the eye to be cut. Due to the groundwater pressure outside of the sheets, approximately 15 to 20 m3 of material blew into the excavation, displacing the ground behind the eye and replacing it with loose, sloughed-in material. Although slurry

289 Copyright © 2004 Taylor & Francis Group plc, London, UK

Newtown Creek Wastewater Treatment Plant, Brooklyn, New York

09096-36.qxd

2/18/04

1:46 PM

Page 290

grouting was performed to fill the voids in the now highly unstable, sloughed-in soil, the ground on the backside of the sheets was still an indefinable, loose, water bearing mixture of the various sloughed-in soils. Ground freezing was selected as the most cost-effective and practical approach to stabilizing the soils, because the freezing would provide assured results more or less independent of the soil type and consistency. Since cutting the eye and installing a sacrificial concrete liner in preparation for future tunneling was a one day operation, liquid nitrogen was selected as the freezing agent. The freeze pipes were installed vertically from ground surface to 1.2 m below the tunnel invert on a irregular pattern due to a myriad of subsurface obstructions. Freezing was accomplished in three days and the eye was cut and the liner installed against a vertical free-standing frozen face in one day without incident. 5 CONCLUSIONS Ground freezing is a niche technology, fulfilling a geotechnical need that is difficult, and in some instances impossible, to fulfill through other groundwater control

and excavation support methods. For deep excavation, difficult or disturbed ground conditions, or for the stabilization of complex subsurface profiles, ground freezing offers the assurance critical to successful and timely urban construction. ACKNOWLEDGEMENTS The authors extend their appreciation to the management of Moretrench American Corporation, and its ground freezing division, freeze WALL, for their input and cooperation in the preparation of this paper. REFERENCES Angelo, W.J. 1999. Crucial Link Nears Completion With The Aid Of Soil-Freezing. Engineering News Record, December 13, 1999 Donohoe et al. 2001. Ground Freezing for Boston Central Artery Contract Section C09A4, Jacking of Tunnel Boxes. Proc. Rapid Excavation and Tunneling Conference, San Diego, California: 337–344 Rogers, C.R. & Taylor, S. 2003. The Big Dig’s Big Dig. Civil Engineering, September, 2003: 40–49

290 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-37.qxd

2/18/04

1:47 PM

Page 291

Session 3 Design/build contracting practices

Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-37.qxd

2/18/04

1:47 PM

Page 293

Session 3, Track 1 Predicting and controlling cost and schedule

Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-37.qxd

2/18/04

1:47 PM

Page 295

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

An economic approach to risk management for tunnels Bashar Altabba HNTB Corporation, Boston, MA, USA

Herbert Einstein Massachusetts Institute of Technology, Cambridge, MA, USA

Hugh Caspe HNTB Corporation, Boston, MA, USA

ABSTRACT: Comprehensive and proactive management of risk is vital to the success of tunneling projects. A well planned and executed approach to managing risk is essential for making informed decisions, evaluating options and developing contingency plans to deal with financial and contractual risks and to improve safety on tunneling projects. Prudent and responsible management requires a broad view of all risks, their effects, the probable reaction of other stakeholders and the possible down-the-line consequences. Potential lawsuits, lack of insurance, equipment failure, errors and omissions, results of third party involvement and changed conditions are all risks that need to be considered when creating risk management plans and strategies. This paper presents an economic approach to risk management using qualitative and simplified quantitative measurements to foster confidence that a less risk prone project can be constructed. This approach will help reduce, to manageable levels, the uncertainty and its associated costs, to all parties, that are common to tunneling projects.

1 INTRODUCTION Tunneling projects are affected by unique uncertainties. Numerous judgments and decisions that need to be made in the development of tunneling projects are made without complete information, and therefore give rise to some degree of uncertainty in the outcome. Risk management techniques can be employed to identify project risks and develop strategies to effectively address them. This paper proposes an approach based on a simplified economic model for evaluating tunneling uncertainties and successfully adjusting the tools available to effectively control those uncertainties. The paper will focus primarily on defining the overall process, rather than the detailed application of each tool, to establish a rationale for a comprehensive project risk management approach. The purpose of this process is to carry out a rough preliminary appraisal of the anticipated exposures and decide how to best address them. The proposed approach starts with the identification of major potential sources of exposure and the various scenarios anticipated for their occurrence. The possible consequences of each event are then described in terms of a rough gross monetary range or

non-monetary impact. By assigning probabilities to each occurrence, one can then evaluate the project risks prior to any response. Once consequences of exposure have been identified a decision has to be made on how to best address each element of exposure. If all risks were to be mitigated during design, it would prohibitively drive-up the cost of construction. Various alternate risk management tools are available to maintain a reasonable control over the expected cost of construction while reducing the degree of exposure. These include, Risk Mitigation, Risk Allocation and Risk Absorption. These alternates all have costs associated with them that can be estimated and included in the total cost of the project. Once a strategy is identified, the cost of its implementation is estimated together with the cost should that strategy not be fully successful. This approach is not complex, and is a simple rational process to dealing with uncertainty and insuring that proper account is taken of foreseeable risks. The aim is to allow proactive management in advance, instead of allowing risks to mature, requiring crisis response. Effective project risk management should include the following elements in a thorough and systematic approach.

295 Copyright © 2004 Taylor & Francis Group plc, London, UK

2/18/04

1:47 PM

Page 296

Risk Identification

I. Risk Assessment

09096-37.qxd

Probability of Occurrence P

Consequence of Occurrence C

III. Contingency

II. Risk Response

Risk Evaluation R=PxC

Risk Mitigation

Design Mitigation

Risk Allocation

Performance Mitigation

Residual Risk Contingency

Risk Sharing

Risk Transfer

Residual Risk Contingency

Risk Absorption

Unforeseen Risks

Contingency Planning

Experience

Residual Risk Contingency

Unforeseen Risks Contingency

Figure 1. Risk management approach.

• • •

Risk Assessment. Determine and evaluate the major risk elements involved. Risk Response. Decide on an appropriate measure to deal with the risk. Contingency Planning. Make appropriate allowance for retained risks. See Figure 1.

2 RISK ASSESSMENT 2.1

Definitions

Risk combines the uncertainty and the consequences of an event. This can be expressed using the expected value as: (1) where R  Risk, P[E]  Probability of the event, and CE  Consequence of the event. Events can vary widely from equipment failures, to slowdowns because of different geologic conditions, to traffic delays. In principle, events could also have positive consequences but when used in conjunction with risk, one considers only negative consequences. In conjunction with construction, the consequences will be usually expressed in terms of cost and time, possibly also in terms of accidents and environmental effects among others. The probability term P[E] expresses the uncertainty with which an event can occur and ranges from 0 to 100%. Such probabilities can be estimated

subjectively or they can be drawn from objective statistical (relative frequency) records. An example of the former is the estimation of the probability that a major piece of equipment will fail. This can also be obtained from experience. An example of the latter can be the occurrence of traffic delays where one might have records on the number of occurrences per duration of construction. The definition of risk as given above is only one of many possibilities. Insurance companies for example, use the term risk to express the component CE in the above expression. In most civil engineering applications the definition as given above is used. The following additional points should be made: 1. It is possible that even if an event occurs its consequences may not occur. For instance, a settlement under a building may occur without the expected building damage. This means that there is a additional “conditional” probability P[CE|E] that needs to be applied to adjust the probability of occurrence. Expression (1) then becomes: (2) 2. It was assumed that risk has to be expressed in monetary terms (cost). This is usually done by transforming other consequences (time delays, injuries) into costs. This transformation is, however, not necessary as is done in multiattribute utility analysis, were consequences can be expressed by utilities (non-monetary values) and these in turn can be

296 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-37.qxd

2/18/04

1:47 PM

Page 297

related to different attributes (non-monetary objectives like schedule or safety). 3. The expected risk value model on which this analysis is based assumes that an event will occur a number of times. Should an event not be repetitive the results of this approach may not be applicable. 2.2

Scenario analysis

This is the central aspect of risk analysis and is also where professional experience comes to bear. Scenario analysis consists of a detailed description of the project and, most importantly, the process through which it is implemented. This starts with planning and goes via design to construction and could include operation if so desired. Additionally, aspects where the project implementation process can vary are also identified. Usually these are aspects where the process can go as planned (normal) or where it deviates from the normal. Depending on the phase for which the scenario analysis is conducted, there may be a varying number of different scenarios to consider. These can range from political demands, to changes in the regulatory requirements, legal actions, alignment changes, including many construction changes. These scenarios are the risk events that need to be considered. There are many ways in which the scenario analysis can be conducted. One approach is to have a group of experienced professionals involved in the project brainstorming and documenting the different scenarios with emphasis on identifying the events. A facilitator, one of the experts or an outsider with encompassing understanding, can then go through the scenarios developed, check them for consistency and document a final set of scenarios and associated events. A particularly good example of this process is the Cost Estimation Validation Workshops conducted by the Washington State DOT for all its major projects. Scenario analysis is not only essential as a basis for risk assessment but also encourage a thorough thinking about the specific project early in its planning and sets the stage for further analysis. Given the process, it is quite clear that the individuals involved in scenario analysis have to be those who have major responsibilities on or a stake in the project and are experienced in the specific issues involved in tunneling. 2.3

Consequence estimation

The consequences of the events can be described in terms of, cost, time or other quantitative or qualitative descriptions and specific values associated with event occurrence. The values can be described in different ways. Cost consequences are usually associated with monetary values. If time can then be transformed into cost, it is possible to express risk in purely monetary

terms. Similarly, accidents and environmental failures may also be transformed into costs. Another approach is to create classes of consequences, which can be verbally defined and/or each assigned with a numerical scale. The worst (most severe) class would then be associated with an event that has consequences severe enough to present a likelihood of preventing the project construction. Examples include catastrophic water inflow into a tunnel, major settlements or unacceptable environmental consequence. In any of these methods, it is possible to associate different weights with the consequence. 2.4

Probability in terms of fractions between 0 and 1 or percentages between 1 and 100 can be assigned to each event subjectively. More formal ways of achieving this has originated in economic decision theory. An example is the comparison of the estimated probability with the flip of a coin. This approach has been successfully applied in a number of recent major tunnel projects. Yet, it is important to apply so-called consistency checks. For instance if a probability of 30% is assigned to one event and 60% to another one, it is necessary to make sure that event 2 is really twice as likely to occur as event 1. Another approach of estimating probabilities is through the relative frequency method in which records of past construction projects may shed light on the frequency of certain events. An example being breakdowns of a typical piece of construction equipment such as a TBM machine. 2.5

Risk

Risk can be calculated with the process just described and can be performed using a spreadsheet. As mentioned earlier, of major importance is that risk assessment, scenario analysis, consequence estimation and probability estimation have to be done by experienced personnel.

3 RISK RESPONSE APPROACH Risk response development is perhaps the most delicate part of the risk management process, and it is here where many projects could benefit from a coherent and comprehensive approach. In general for each risk identified, see Table 1, there are a number of options or tools available to deal with that risk. These include mitigating the risk by avoiding or reducing it through design or construction measures, allocating the risk by sharing it or transferring it to another party, or absorbing the risk in a careful contingency

297 Copyright © 2004 Taylor & Francis Group plc, London, UK

Probability estimation

09096-37.qxd

2/18/04

1:47 PM

Page 298

Table 1. Risk identification. General risk types

100%

Potential risk categories High Risk

f. Environmental

Availability and pricing of materials, skilled labor and specialized equipment Soil/water condition Soil/water contamination

b. Schedule delays

c. Cost overruns

III. Technical a. Design

b. Construction performance

IV. Legal a. Contractual

b. Third party

40%

20% Low Risk

g. Utilities II. Organizational a. Management

60%

I. Mitigation

Probability

80% Unknown, unclear, delayed, unassigned

III. Absorption

I. External a. Regulations and permits b. Natural hazards c. Terrorism d. Bankruptcy e. Market condition

II. Allocation Incompetence Lack of organizational structure Inadequate planning Unrealistic schedule Lack of coordination Labor shortage or productivity Material shortage Unforeseen site conditions Owner scope change Lack of site access Schedule delays Inappropriate procurement strategy Contractor claims Under estimating Inadequate data/ground characterization Lack of experience Design inadequacies Last minute design changes Constructability Complexity Inadequate alternates analysis Construction safety Construction quality Rate of production Reliability Differing site conditions New complex technology Third party impact Misinterpretations Misunderstandings Inappropriate contractual strategy Measurement of pay items Variations in quantity Mobilization cost payment method Environmental Personal property

0

4 6 Moderate Consequence

8 High

0% 10

Figure 2. Risk evaluation and response matrix.

plan. The choice will depend on the particular project, the risks involved and the specific circumstances. Each response option needs to be evaluated, assessing its likely effect upon the risk, the feasibility and cost of implementing that option, and the impact of each option on the overall cost of the project. The most cost effective response option for each risk is typically selected, with the realization that the overall risk potential is rarely completely eliminated. Significant judgment should be exercised in determining which particular risk response measures is adopted. The effect of adopting such measures will generally be to reduce the risk but at a given cost to the project. When carrying out the planning and subsequent selection of the primary responses, it is useful to identify the parameters influencing the selection of the appropriate response. In order to facilitate a systematic approach to risk response selection a risk evaluation matrix is prepared representing the expected value of risk, see Figure (2). Equal risk value curves are then plotted and the matrix divided into two zones, high risk events occupying the top right corner and low risk events occupying the rest of the matrix. The clients risk tolerance would then need to be understood to select the appropriate risk value curve above which risks are considered to be high. Some owners may not want to accept any more risk than absolutely necessary. Owners who have a low threshold for risk often choose to specify more costly, conservative designs that can be constructed within lower bounds of risk. The overall shape or the risk assessment and response matrix is the same for all projects. However the values and scales of the axis will change depending on the specifics of the project.

298 Copyright © 2004 Taylor & Francis Group plc, London, UK

2 Low

09096-37.qxd

2/18/04

1:47 PM

Page 299

The following key issues should be kept in mind: 1. The probability estimation, expressed in verbal or numeric terms, carries with it the potential for misjudgment and the risk assessment may suffer from a biased or mistaken probability judgments. As such it may be important to pay more attention to the event impact than to their expected probability. 2. Knowing the client and his or her risk preference is of major importance. The equal value assumption of high probability and low consequence with low probability and high consequence may not be accurate. 3.1

Risk mitigation

Risk mitigations are approaches that recognize the risk and present solutions that either reduce an event probability of occurrence or reduce the probability that the consequence occurs, or both. Such reductions, i.e., mitigations, would lead to a modified residual risk (R) after implementation and would typically involve some cost to the project. One can thus evaluate if the mitigating measures are cost effective. This would occur if the reduction in risk is greater than the cost of mitigation. (3) were R  Risk, R  Residual risk after mitigation, R  Net reduction in risk and CM  Cost of mitigation. As such, for high risks, occupying the top right end of the Risk Assessment and Response Matrix, where the cost of mitigation is lower than the expected value of risk reduction, the rational economic response is to Mitigate. There are two basic mitigation strategies, design mitigation and performance mitigation. 3.1.1 Design mitigation Design mitigation is the basic tool utilized for mitigating risks that can be economically addressed by design. Options available for design mitigation are broad and may include among other approaches:

• • • • • • • • • • • •

Provide adequate time and resources for ground investigation Conduct investigation on the nature and condition of the surrounding structures and utilities in advance Hire experienced engineers Improve communications among all parties Utilize familiar proven technology Encourage peer reviews and use Technical Advisory Board Increase the level of conservatism of the design Use performance specifications were appropriate Specify a construction monitoring program Use the observational method Have regular design reviews Utilize value engineering sessions.

In applying these and other design mitigation measures care must be taken to ensure that the design does not become unduly conservative. Risk avoidance, by taking actions so that the risk event no longer impacts the project objectives, is tantamount to complete elimination of the probability of occurrence or the consequence of an event. This approach is often used when it can be achieved at no or little cost. An example of this approach is changing the alignment to avoid tunneling under a structure to eliminate third part impact. 3.1.2 Performance mitigation The contractor is ultimately responsible for the selection of the means of construction, the equipment and methods for prosecution of the work. Some of the options available that may help enhance performance include:

• • • • • • • • • • •

Performance mitigation is a delicate balance and care must be exercised not to become overly prescriptive in applying performance mitigation measures. 3.2

Risk allocation

For low risk events where the reduction in the expected risk value does not justify the cost of mitigation, the rational response depends on the severity of the expected consequence. High severity level events that the Owner may not want to absorb could be allocated to another party in a better position to bear them. In order to do this one must develop an understanding of the Owner’s risk tolerance. Risk tolerance is the point beyond which the event severity becomes intolerable to the Owner and need to be allocated in some manner.

299 Copyright © 2004 Taylor & Francis Group plc, London, UK

Use rigorous technical and financial prequalification procedures Disqualify bids with un-priced conditions, qualifications and disclaimers Utilize a pre-bid meeting to confirm understanding of what is required and offered Provide contractor value engineering incentives Have a sufficiently long bid period Evaluate contractor’s bid means and methods early in the project to ensure that they meet the project requirements Disclose all available subsurface information to bidders Establish and operate an owner controlled monitoring system with trigger limits and procedures throughout construction for critical elements Execute early site work contracts Pre-select excavation disposal area and construction sites Pre-purchase long lead items where appropriate.

09096-37.qxd

2/18/04

1:47 PM

Page 300

Risk allocation implies shifting, totally or partially, the impact of the risk event to a third party. With this approach risks can either be shared or completely transferred so as to spread the consequences through the use of contract language or by insurance. Its main use is typically limited to financial risk exposure. In allocating risk among the principal parties, experience has shown that the most cost effective strategy is to allocate risk to the party that is closest to its source and hence has control over its occurrence. One has to also ask: Which party would be most able to carry the risk if it cannot be controlled? Which party is best positioned to manage the risk effectively? The client may, however, deem it in his interest to retain some control over the management of the risk after it occurs, as in the case of hazardous materials. If the distribution of risk is not clearly understood, or is patently unfair, then disputes become almost inevitable. It is further important to keep in mind that risk allocation does not remove the risk, but simply gives another party the responsibility of managing it. It is therefore essential that recipients of allocated risks are able to manage the risks allocated to them, otherwise the project will remain exposed. 3.2.1 Risk sharing Tunneling differs from general construction contracting. Unexpected events, including those resulting from unforeseen site conditions, have a greater impact on the progression of work than in other types of construction. Differing contract provisions employed in tunneling projects have a material effect on the distribution of risk among the parties involved. The ideal tunnel contract is one that clearly defines the responsibilities, duties and obligations of each party. Yet standard contract documents typically imply a specific risk allocation, which may not be appropriate for tunneling projects. Good contracting practice will distribute equitably the risk of construction among the parties, to reduce the overall cost of construction and reduce the uncertainties by apportioning the risks. Various types of contracts can be utilized in which the degree of associated risk is shared differently between the parties. The choice of type of contract appropriate for a specific tunnel project is an integral part of the project risk management process. The specific approach for each identified risk and the determination of how they should be shared between the parties should be defined with the insertion of clear language in the contract documents that allocate the responsibility. The following contract recommendations are generally accepted approaches for sharing risk:

• •

Changed Conditions Clause Use of Geotechnical Baseline Report as a benchmark for bidding

• • • • • • •

3.2.2 Risk transfer Risk transfer involves shifting all of the risk to another party. However, risk transfer on a project should not remove ones incentive to managing the risk. The rewards should remain associated with successfully managing the risk. There are many risk transfer tools available including among others:

• • • • • • • • • • •

General Commercial Liability Insurance, which covers third-party bodily harm, property damage or faulty workmanship Professional Liability Insurance, which covers damage and injury resulting from negligence Workers’ Compensation Insurance, which covers injury to workers Builder’s Risk Insurance, which covers damage to work during construction Construction Equipment Insurance, which covers damage to contractor’s equipment Environmental Liability Insurance, which covers injury, damage and clean up cost from release of pollutants Bonding. Bid and performance bonds should be balanced between the rights and the obligations of the parties. The size of the performance bond could be reduced by prequalification if appropriate Guarantees and Warrantees Subcontracting. Highly specialized components are often better subcontracted to a specialty firm Contract Type. Lump sum type contracts can be used to transfer the potential variation in cost of construction risk to the bidder Contract Language. Certain risks can be explicitly transferred in the contract i.e., Construction Safety.

3.3

Risk absorption

Risk absorption entails the acceptance of the risk consequences and planning to deal with them should they occur. This is the rational economic response

300 Copyright © 2004 Taylor & Francis Group plc, London, UK

Means and Methods. The contractor is ultimately responsible for the selection of the construction method, equipment and prosecution of the work Employ performance penalties as well as incentives, such as Liquidated Damages Define types of Third Party Impact and responsibilities Include variations in price clause and suitable mobilization bid items Utilize Escrow Bid Documents and include specific guidelines about what details or backup data the owner wants to be able to track Incorporate a Dispute Review Board for technical issues and not for interpretation of contractual matters Establish an equitable change order process.

09096-37.qxd

2/18/04

1:47 PM

Page 301

when the expected reduction in the value of risk is lower than the cost of mitigation and the consequence severity is relatively low. This option requires the careful estimation of an adequate budget and schedule flexibility for those risk elements the Owner elects to absorb and manage through funds allocated to a contingency, a form of self-insurance. 3.4

Unforeseen risks

Problems do occur on projects which are unexpected simply because they were not anticipated or foreseen during planning and design. Examples may include political or third party involvement, which may delay a project implementation. It is therefore prudent to carry some contingency reserves, both in budget and schedule, for situations that cannot be accounted for or cannot be quantified, to cover unforeseen events. The contingency may be based on experience or historic data and should be modified by special project circumstances. This contingency also covers market fluctuations that are difficult to determine early in the process i.e., market condition, labor supply, etc. 3.5

Contingency

Contingency is the combined cost of all residual risks and unforeseen risks. Residual risks are all the risks remaining after response implementation. This includes risk remaining after mitigation, sharing and risks absorbed. As such contingency allowances are typically included for any remaining risks, which are not specifically covered in the scenarios analyzed, as well as an allowance for unforeseen and unmeasured risks. In general the higher the level of confidence in one’s risk analysis the smaller the contingency needs to be. Two types of contingencies are therefore appropriate, one to cover the expected cost of residual risk and another based on past experience to cover unforeseen risks.

a rational manner. This involves systematically identifying the risks and evaluating their probability and impact, a method to evaluate risk through conceptualizing it as a potential monetary loss, and a procedure to respond to these risks depending on the expected risk value and the severity of the consequence. Where appropriate, response measures are identified to mitigate, allocate or absorb the risk. Residual risks are then handled through a contingency plan. Lastly, an important part of risk management is the continued development of a risk aware culture that accepts known costs now in order to avoid the possibility of unknown and significant costs in the future, while creating confidence that risks are being managed effectively.

REFERENCES ANSI/PMI 99-001-2000. A Guide to the Project Management Body of Knowledge (PMBOK) 2000. Atlanta: Project Management Institute. Abdel Salam, M.E. 1995. Contractual Sharing of Risks in Underground Construction: ITA Views. Tunneling and Underground Technology Vol 10 No 4: pp 433–437. Chernoff, H. & Moses, L.E. 1987. Elementary Decision Theory. New York: Dover Publications. Duncan Luce, R. & Raiffa, H. 1989. Games and Decisions: Introduction and Critical Survey. New York: Dover Publishers. Hatem, D. 1998. Subsurface Conditions Risk Management for Design and construction Management Professionals. New York: John Wiley & Sons. Hilson, D. 1999. Developing Effective Risk Responses. Proceedings of the 30th Annual Project Management Institute Seminars & Symposium. Philadelphia. Institute of Civil Engineers/Institute of Actuaries. 1999. Risk Analyses and Management for Projects (RAMP.) London: Thomas Telford. Parkin, J. 2000. Engineering Judgment and Risk. London: Thomas Telford Publishing. Piney, C. 2002. Risk Response Planning: Selecting the Right Strategy, Fifth Europeans Project Management Conference. Cannes: PMI Europe.

4 CONCLUSION Tunneling is considered to be a risky business. A simple economic approach to risk management is presented to assist in dealing with those uncertainties in

301 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-38.qxd

2/18/04

1:48 PM

Page 303

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Top down construction of Ramp L, Value-Engineering Change Proposal for the Massachusetts Turnpike Authority, Contract CO9A4 W.D. Driscoll & G.A. Almeraris Slattery Skanska, Inc. New York, New York

ABSTRACT: This paper describes a Value Engineering Change Proposal (VECP) to change one of Boston’s Central Artery tunnels from a drilled shaft supported and soil mix restrained ‘Cut and Cover’ tunnel into a ‘TopDown Construction’ slurrywall tunnel. This VECP was conceived and submitted by the Joint Venture of Slattery Skanska, Interbeton, JF White and Perini (SIWP). The schedule and cost benefits of the VECP were recognized with an award in 1999 from AASHTO for ‘Outstanding Value Engineered achievements in Teamwork, Cost Savings, Streamlined Construction Sequencing, and Reduction of Schedule’. Discussed in this paper are many of the technical challenges that had to be overcome both in the design process and during construction. It will describe how the revised construction sequence allowed at-grade and elevated structures to be constructed as soon as the ‘top’ was complete, shaving months off the project construction schedule. This construction method especially limited the risk related to unknown subsurface conditions, which could have caused high cost overruns and project delays.

1 INTRODUCTION The Ramp L Tunnel Structure was a critical section of the Central Artery/Tunnel (CA/T) Project’s CO9A4 Contract. With the overhaul of the City of Boston’s traffic system came an extension of the Massachusetts Turnpike to Logan Airport. For the first time, Interstate traffic traveling from either south or west of Boston could go directly to Logan Airport through a new system of tunnels. One of these tunnels, Ramp L, is the focus of this paper and was constructed in perhaps one of the most challenging locations of the CA/T Project. The Ramp L Tunnel was designed to carry Interstate I-93 traffic from south of Boston into the Ted Williams Tunnel and on to Logan Airport. The most complex portion of the CO9A4 Contract was the extension of the Massachusetts Turnpike (Interstate I-90), which had previously ended at the junction with Interstate I-93 at the South Station railroad tracks. The Contract called for the installation of tunnels under the live tracks at South Station by jacking pre-cast tunnel structures from thrust pits on the west side of the tracks to the east side of the tracks. Ramp L is an off-ramp from Interstate I-93 Northbound that carries traffic to a junction with the new Massachusetts Turnpike extension that merges with traffic heading east to the airport. The extension of the Massachusetts Turnpike to

the airport is the single largest aspect of the CA/T Project, involving billions of dollars in contracts. The opening of this extension was dependent first on the completion of the jacked tunnels and the subsequent cut and cover tunnels and second on the completion of Ramp L from the south. When the Owner accepted the contractor’s Value-Engineering Proposal for the construction method of Ramp L, much of the responsibility for the on-schedule completion of the structure switched to the contractor.

2 2.1

Contract solution for Ramp L

The CA/T Construction Contract CO9A4 called for the cut and cover construction of a new 706-ft long tunnel. The cross-sectional views of the proposed tunnel indicated no reason for concern over the risk associated with the actual construction of this tunnel. However, what was not detailed in the drawings of the tunnel was exactly how to reach the point where the actual tunnel construction could commence. Many months of work were required before the first concrete could be placed for the tunnel invert. Moreover, the work required to get to this point was laden with risk

303 Copyright © 2004 Taylor & Francis Group plc, London, UK

CONTRACT CO9A4, RAMP L

09096-38.qxd

2/18/04

1:48 PM

Page 304

and uncertainty and the potential for huge delays was a reality that could not be overlooked. The construction sequencing as suggested in the Contract Documents required that the tunnel be constructed in three segments, starting from the south and working northward. It is important to note that none of the activities described below could have been accomplished in a single mobilization stage. Some could be completed in two mobilization stages but most required three, one for each segment. The Contract sequencing called for the completion of the construction of the first segment of the tunnel structure before the first step of the lengthy sequence could be commenced in the adjacent segment. 2.2

•

•

•

Construction sequence

Containment sheeting installation: 500 linear feet of sheeting was required to be installed at 180 ft distance from the existing shoreline and existing granite seawall. This work was performed from a barge and was required in both scenarios [Contract method or VECP solution] to provide containment for the backfill and a turbidity curtain for the protection of the Fort Point Channel from pressure injected soil improvement cement.

•

Figure 1. Fort Point Channel Pre-construction.

304 Copyright © 2004 Taylor & Francis Group plc, London, UK

Existing granite seawall removal: A granite seawall that lined the Fort Point Channel had to be removed in its entirety. This wall, 230 ft long and with a 15 ft by 8 ft cross-section, was supported by over three hundred timber piles and consisted of hundreds of granite blocks that were to be salvaged to construct a relocated granite wall at a later date. Land reclamation in the Fort Point Channel: The historic Fort Point Channel (Fig.1) was to be filled with over 150,000 tons of gravel borrow within the confines of the containment sheeting. The gravel was to be placed on the 20-feet thick layer of soft marine sediments [from Elevation 73 to 93] to Elevation 110 over an area of almost 2 acres. After a 25 ft wide strip of backfill was completed, equipment was to be moved on top of the new fill to improve the soil below to a homogeneous 70 psi. A cross-section of the area shows vastly different existing soil conditions between Elevation 68 and 110, however the Contract required reaching a homogeneous improved strength of the entire cross-section. Soil improvement and obstruction removal: Ground stability for the soil mixing equipment was a concern. The Contract indicated a strength for the

09096-38.qxd

•

• •

•

2/18/04

1:48 PM

Page 305

existing organics of approximately 100 psf, and mentioned that the soil was prone to large displacements under any applied loads. The applied surcharge load of the new backfill, from the surface to the top of the soft organic layer, exceeded 2,000 psf, plus the added load from equipment surcharges presented a global stability and constructability issue. The risks of large displacements and instability were major factors in the consideration to submit the Value-Engineering Change Proposal for this work. Additional risk was attributed to the well-documented granite block obstructions that had collected on the bottom of the Fort Point Channel over the past 150 years. The soil improvement had to be executed with equipment situated at grade on Elevation 110. Obstructions littered the subsurface from 25 foot to 35 foot below grade. Removal of the obstructions during soil mixing operations would have been difficult and disruptive. Drilled shaft installation: 86 drilled shafts of 5 ft diameter were planned to be installed as a foundation for the Ramp L tunnel structure, each with an average length of over 110 ft and a minimum embedment of 12 ft into bedrock. Support of excavation: 70,000 sf of sheeting was required on both sides of the proposed cut and cover tunnel. Excavation and installation of the bracing system: Excavation through the proposed 70 psi strength improved soil should not have been an area of concern or a risk item for delays. However, experience on adjacent contracts had proven that the minimum compressive strength of 70 psi was difficult to achieve. To bring the existing subsurface soil, comprised of up to a 20 ft thick layer of soft organic material, and a 17 ft thick layer of granular fill to a homogeneous 37 ft thick layer of 70 psi minimum material had caused many problems. The required soil cement ratios were based on the weaker organic stratum that dictated higher volumes of cement. When these volumes of cement were injected into the gravel layer above, what resulted was a soil mass with a compressive strength far greater than the minimum requirement. Problems ensued during the subsequent excavation, which could no longer be performed by conventional methods. Tunnel construction: Only after the completion of the excavation between the braced sheetpile walls could construction activities for the tunnel begin. Cast-in-place wall construction would follow the invert slabs after and during removal of the cross-lot bracing. Roof construction could commence after shoring towers were installed. It should be noted that many other critical areas of the Contract hinged on completion of the roof slab as the at-grade roadways and the viaduct columns were to be constructed

in this area after the Ramp L tunnel was covered by backfill. 3 CONTRACTOR PROPOSED VECP 3.1

The Contractor decided to propose a change to the Contract construction method, a change that would streamline construction activities and also greatly reduce the risks. As with all subsurface work the risk was clearly recognizable but not quantifiable. Before a redesign could commence, the design parameters of the tunnel must be understood. Within the CA/T Project, the basic concept behind a Value Engineering Change Proposal (VECP) is that it must provide an equal or better end product, with no schedule delays, and monetary savings. In this case, the VECP provided an equal product at a considerable schedule and monetary savings. 3.2

VECP scope of change

The Joint Venture of Slattery, Interbeton, JF White, and Perini (SIWP) proposed an idea to their design consultants, Weidlinger Associates, that was so radically different from the Contract design that it required an almost complete redesign of the structure and its components. [It was awarded AASHTO’s Certificate for the most cost-effective VECP in 1999]. The goal of the VECP was to focus on the work activities described above and reduce the risks with an alternate method of construction. The magnitude of this undertaking is represented in Table 1. The figures represent

Table 1. Breakdown of VECP savings.

Contract solution – deleted Soil mix Drilled shafts Jet grout Support of excavation Cast-in-place walls Miscellaneous VECP scope – added Slurry walls Modified soil mix Glory holes (3) Miscellaneous Engineering

305 Copyright © 2004 Taylor & Francis Group plc, London, UK

Alternate construction method

Qty

US$

150,000 cy 10,000 lf 11,000 cy 70,000 sf 50,000 sf Total As-Bid

9,000,000 5,500,000 1,200,000 1,000,000 3,200,000 487,742 20,387,742

100,000 sf 63,000 cy 1 LS 1 LS 1 LS Total alternate

12,200,000 3,250,000 100,000 348,373 1,185,000 17,083,373

Total savings

3,304,369

09096-38.qxd

2/18/04

1:48 PM

Page 306

Figure 2. Contractor VECP – Ramp L Tunnel.

the gross savings realized and a partial scope of deleted and added work. 3.3

Soil mix, initial method

In the preliminary VECP described above, a reduced quantity of 63,000 cubic yards out of the original 150,000 cubic yards of soil improvement during land reclamation was still necessary. This reduction in treated soil volume was made possible with the new configuration of the tunnel, which created a more rigid structure. In this proposal, only the 20 ft thick organic layer was to receive treatment and it was called ‘modified’ soil mix because it changed the Contract specified soil mix from 70 psi to a soil mix with a reduced required strength of 30 psi. While this proposal represented substantial savings and greatly reduced the risks, SIWP still did not feel comfortable committing to an aggressive accelerated schedule for completion of the tunnel. Any successful soil cement improvement of the organic layer would require deep (unknown) obstructions to be removed during the soil improvement operation itself. Another remaining risk factor was related to the effort that would be required to actually improve the organic

strata from 1 psi to the proposed 30 psi. The assumption that this could be accomplished was based on a belief that the Contract method could be achieved, as the proposed modified soil improvement was to use all of the techniques and equipment specified in the Contract. However, the fact that adjacent contractors had experienced tremendous difficulty in achieving their Contract specified compressive strengths, was something that caused concern to SIWP. 3.4

306 Copyright © 2004 Taylor & Francis Group plc, London, UK

Soil mix, final method

Further development of the VECP addressed the remainder of the ‘modified’ improved soil. Weidlinger Associates re-analyzed the structure without a soil cement buttress surrounding the tunnel but because of the weak nature of the organic soils in the area, the structure would not remain stable during a seismic event. In the Contract design the tunnel had been modeled as a concrete box sitting on top of 120-ft long stilts. The drilled shafts provided no lateral resistance in the event of an earthquake. It was determined that even in the new more rigid configuration of the tunnel that some method of soil improvement would be needed to brace the structure.

09096-38.qxd

2/18/04

1:48 PM

Page 307

SIWP and Weidlinger investigated methods of improving the soil without soil cement. It was decided to analyze the behavior of the structure with 75% of the weak organic layer removed during land reclamation and replaced with gravel borrow. With the improved stiffness of the alternative structure, and a more resistant stronger soil outside the structure the results of the analysis were favorable and the scope of the proposed change was finalized. 3.5

Drilled shafts

The drilled shafts from the Contract design served as the vertical support for the tunnel. The tunnel itself served as the vertical support of the at-grade roadways and viaducts above, as viaduct support columns were designed to rest on top of the roof of the structure. A special slurry panel was proposed that would be constructed in three ‘bites’. The panel was 25 ft long along the baseline of the tunnel and a combination of 60 ft and 120 ft deep. These slurry panels would form the permanent tunnel walls, and replace the drilled shafts for vertical support. The middle ‘leg’ of the slurry panel was designed to replace the drilled shafts and would act as the foundation for the tunnel. 3.6

Sheeting and bracing

The need for temporary steel sheeting during cut and cover tunnel excavation was eliminated as part of the VECP. During the construction phase, excavation could commence immediately upon completion of the slurrywalls, which were to be brought up to just below existing grade and just above the mean high water table. The walls would be utilized for support of excavation in the temporary construction phase by adding rebar to the walls that extended up above the future tunnel roof to just below grade. This would allow the walls to act in cantilever restraining the 25 ft of soil and construction loading applied to the walls. Once excavation was complete to just below the proposed roof slab soffit, a 6 inches thick working slab would be placed and the roof slab installed. The roof slab would now act as a temporary brace between the two walls and allow excavation under the roof to commence. While construction continued under the roof slab, work could commence on the critical activities above the roof level. This aspect of the change is what earned the VECP an award for streamlined construction activities. 4 BENEFITS Other than the obvious benefits in cost savings, further benefits of implementing this VECP are recognized below.

4.1

The baseline schedule required that certain sections of the Ramp L tunnel structure had to be completed before various other components of work could commence. The critical successors to the Ramp L construction could only start upon the completion of sections of the roof that served as the foundation for viaduct columns in 3 separate locations. In another location a 72 in diameter combined sewer pipe had to be cast into the roof slab while the abutment of an access ramp was to be constructed on top of about 100 lf of another section. An active railroad also crossed over the structure and the tracks had to be relocated on top of another section of the Ramp L roof. With the implementation of the top-down method of construction, all of the above work could commence well ahead of the baseline scheduled early start dates. The lengthy processes of removing obstructions during soil mixing operations, improving 150,000 cy of weak soil, installing 12,500 lf of 5 ft diameter drilled shafts, 70,000 sf of steel sheeting, 200 pieces of crosslot steel bracing, 32,000 sf of 6 ft thick invert slabs, and 26,000 sf of 4 ft thick concrete walls were all removed from the critical path leading to the construction of the roof of the structure. Tremendous savings in time were realized, deleting approximately 250 calendar days from the critical path. The activities replacing all of the above listed work were ‘Pre-excavation for slurry walls’ and ‘installation of slurrywalls’. 4.2

Risk reduction

When dealing with subsurface conditions there is always a chance that you will encounter unknown obstructions. It is difficult, if not impossible, to predict the impact of subsurface obstructions on an operation or a schedule. The severity of the impact depends largely on the depth and size of obstruction, the type of obstruction, and most importantly in the case of Ramp L, the type of equipment that encounters the obstruction. Contract CO9A4 had the potential to have the worst possible degree of each impact. If encountered, the granite block obstructions in the Fort Point Channel would have been between 25 feet and 35 feet below grade, outside the reach of conventional excavators. The equipment encountering the obstructions would have been soil mix drilling rigs, the auger of which would not have been able to advance through these obstructions and would have been forced to abandon the obstructed location and setup to drill in another location. An excavator would have to be mobilized to excavate down to the obstruction, remove the obstruction and then backfill the location with suitable material. Assuming the obstructions were within the reach of the available excavators, the disruption to the soil mixing operation would have been enormous.

307 Copyright © 2004 Taylor & Francis Group plc, London, UK

Schedule

09096-38.qxd

2/18/04

1:48 PM

Page 308

Figure 3. Ramp L Looking south at northern interface with cut & cover tunnel.

Other than adding shifts to the operation, there would have been few options to mitigate the delays from the ‘obstructed’ soil mix operation. Having a larger than conventional excavator on appeared inevitable to meet the schedule of the base contract. On the other hand, liquidated damages for failure to meet contract milestones in the area were $5,800 per calendar day. 4.3

Adjacent contracts

On the northern interface of the proposed VECP, the continuation of the Ramp L tunnel structure was to be constructed by another contractor as a conventional cut and cover tunnel (Fig. 3). That contract required land reclamation and soil improvement of the northernmost 55 ft of the CO9A4 Ramp L tunnel. However, after months of obstruction removal and unsuccessful soil mixing at this interface location, the Owner transferred the scope of this work to SIWP to implement their VECP solution in this area. SIWP successfully removed the deep granite obstructions along the alignment of the future slurrywalls with a crawler crane and clamshell bucket under a full liquid head. This allowed the Owner to delete the soil mix over the entire footprint of the tunnel in this area and more importantly, solve the problem that was created by the obstructions that could not be removed, even by an excavator with a 45-ft long dipper.

Just as the northern end of CO9A4 Ramp L required soil mix by another contractor, so also did the southern portion. The first 110 feet of the structure was to have the soil improved to a minimum compressive strength of 70 psi. However, due to difficulties achieving a homogeneous improved subsurface, the contractor was forced to add cement to the grout mixture that ultimately yielded a material with compressive strength upwards of 1,000 psi. Removal of this ‘concrete–like’ material required the rental of a roadheader, which was mounted on the boom of a lowheadroom excavator (Gradall 5200). The tedious mining operation with the roadheader allowed the excavation of the area to proceed more rapidly, and with less damage to the concrete roof and walls, than with a hydraulic hammer. Even though this unforeseen condition considerably lengthened the duration of excavation of the tunnel, and subsequent invert slab placements, the critical work above the tunnel roof proceeded with no impact from the mining operation. Another area that had a large impact on the completion of inverts was just below glory hole 2, located at the approximate midpoint of the tunnel. At this location, a 70 ft long by 18 ft wide by 30 ft deep concrete pier from the recently demolished Broadway Bridge remained in place. Since CO9A4’s base Contract design called for soil improvement over the entire footprint of the reclaimed land, the pier was to be removed by another contractor prior to SIWP’s mobilization to the

308 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-38.qxd

2/18/04

1:48 PM

Page 309

Figure 4. Ramp L Tunnel Profile.

area. Delays experienced on the other contract meant that SIWP’s commencement of work in this area would have been delayed up to 12 weeks. With the implementation of the VECP, there was no longer an immediate need to remove the pier prior to CO9A4 work in this area. SIWP agreed to install the Ramp L slurrywalls and demolish the pier during tunnel excavation. While the low-headroom demolition of this concrete pier had a tremendous impact on the construction activities below the roof, critical work above the roof proceeded with little impact. 5 CONSTRUCTION After installation of the containment sheeting, the reclamation of the Fort Point Channel commenced. Indeed, the anticipated problems with the soft organic strata adversely affected the operation with large displacements of mud in front of the backfilling operation. 30,000 tons of mud was removed and replaced with gravel borrow, as planned, and subsequent activities commenced. Upon completion of pre-excavation for the slurry walls, three digging rigs were mobilized and slurry wall installation began. The walls were 4 ft wide, averaged 25 ft in length, and were 60 ft deep, with a pant leg panel 10 ft long which was dug to a varying depth of 120 ft and had a 6 ft rock socket. Following slurry wall installation, work crews excavated to the soffit of the roof from within the extended slurry wall panels. These panels were placed to above flood tide elevation to protect the work from flooding and to provide earth support.

Once the excavation for the roof was completed, previously embedded female dowel bar inserts were located in the slurry walls, cleaned, and in many cases re-tapped. Following this a tight tolerance mud slab was installed and because it was replacing the soffit forms of the base contract, it had to be removed during tunneling operations to expose the finished surface of the bottom of the roof. A wax film was applied to the mud slab, which prevented the roof slab from bonding to the top surface of the mud slab. Threaded reinforcing dowels were threaded into the dowel bar inserts in the slurrywall keyways. After placement of the roof slab, surface utilities were installed and backfill placed. The surface area above the tunnel was ready for the next stage of construction that included at-grade ramps, elevated viaduct structures and railroad work, all of which were on the project critical path. During roof concreting, two glory holes were accurately positioned to expedite both excavation and concrete operations. These two glory holes and the open south end allowed the advance of up to 5 headings simultaneously (Fig 4). The top heading excavation proceeded with a low-headroom excavator (Gradall 2200) feeding a low-pressure traxcavator (Caterpillar 953). The 953 trammed muck to the glory hole where a ‘fly-by’ track mounted excavator (Caterpillar 350) equipped with an extended boom removed it from the hole and loaded it directly into trailers. This excavator was a roving machine, which serviced the three access points to the tunnel. Upon completion of the top heading, 36 inch diameter pipe struts were installed at the slurrywall joints and pre-loaded to 300 kips. The final bench

309 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-38.qxd

2/18/04

1:48 PM

Page 310

Figure 5. Exposing invert keyways and dowel bars.

excavation then commenced and was followed by trenching for drainage piping and installation of drainage structure. After stoning the tunnel bottom, a mud slab was placed and the dowel bar inserts for the invert exposed. Figure 5 shows a low-headroom excavator stripping the keyway forms from the face of the slurrywall under a strut. In the background of the photograph the adjacent contractor has already placed the invert slab of the adjoining Ramp L tunnel. Once the mud slab achieved an appropriate strength, temporary struts were removed leaving an open tunnel face for ease of invert reinforcing steel, and subsequent concrete placement operations. The tunnel structure was now complete and all that remained was barrier placement, installation of a precast architectural wall and miscellaneous utilities. The precast wall placed on either side of the tunnel atop the barrier produced a utility void for both ventilation and utilities. The roof was cleaned of the wax coating, painted and lighting installed in the embedded unistrut channels. The final work required the connection of two emergency egress structures, which were excavated by conventional support of excavation and constructed as a detached operation.

6 CONCLUSION It is difficult to envision, in retrospect, how the Original Contract solution for the Ramp L Tunnel could have been constructed without major delays to this, adjacent, and follow-on Contracts. The much-feared deep granite obstructions were actually found and removed during the slurrywall pretrenching operation. However, some obstructions proved unreachable by the crawlermounted excavator (Komatsu PC1000) utilized in this effort and were removed during slurrywall operations with the chisel and clamshell. It is important to note as a result of this VECP merely 10% of the original 2 acres of soil improvement now had to be cleared of these obstructions. The concern now was only for the slurrywall alignment and not any longer the entire area of soil improvement. Apart from the obvious advantage of a reduced risk area, somewhat more critical would have been to determine a course of action when deep obstructions would have been encountered during soil mixing operations. The original contract called for 100% coverage by soil cement and the slurrywall pre-trenching operation proved that some obstructions were difficult, if not impossible, to remove by conventional methods.

310 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-38.qxd

2/18/04

1:48 PM

Page 311

It should also be noted that the impact on an urban environment is greatly reduced, firstly by eliminating many of the construction activities requiring large equipment, and secondly by starting construction of the roof before the invert, much of the construction would be performed underground and out of the public eye and ear. Some important items that were discovered during construction of Ramp L should be considered in planning and executing similar work. The female dowel bars, discussed earlier, that provide a connection between the roof and invert and the walls, are fixed to and placed with the reinforcing steel during slurrywall cage fabrication. Clearance tolerances of the roof and invert transverse steel must conform to ACI standards and generally requires that 2 to 3 of cover is provided over the steel. When the 120 ft long rebar cage is lowered into the slurry trench it must be located with

these vertical tolerances in mind to assure proper alignment of all future rebar connections. Another more common issue with slurrywalls is water tightness. While there does not yet appear to be the technology to install a watertight joint between two slurrywalls, one solution to this problem involves the injection of a hydrophilic resin which, when coming into contact with water, expands quickly to cure into a flexible closed-cell foam and seals the joint. From a practical standpoint, the top down construction technique was clearly the more viable option and can be considered as a preferable alternative when surface restoration is critical for succeeding activities. Overall, the project has proven to be a great technical success. The Ramp L tunnel has been open to traffic since January 2003. At grade roadways and elevated structures could be and have been open to traffic since well before that date.

311 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-39.qxd

2/18/04

1:49 PM

Page 313

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Contemporary methods of budget preparation B. Martin & S. Sadek Parsons Brinckerhoff Quade and Douglas, Inc., New York, NY

ABSTRACT: This paper explains methods of applying well known probability techniques for estimating risks, to better assess the project contingency, and prepare better estimates for projects and multi-project programs at the conceptual level, preliminary and final engineering. One of the biggest challenges that faces engineering design firms is how to determine precise cost estimates for new program budgets at the pre-conceptual stage. A Project cost estimate is usually developed in a very early stage of the project when the detailed design is still to be developed and the complete scope is still undefined. In many cases this early estimate becomes “cast-in-stone”, meaning that the project budget is locked with this “speculative estimate”. The traditional approach to overcome this problem is to add a certain percentage to the project cost estimate, usually called contingency. This project contingency percent usually decreases as the design proceeds from the conceptual phase towards the final design. The contingency is generally based on experience and the uncertainty level of different project items. Traditionally, estimators attempt to inflate the contingency allowance to avoid surprises during the design development. If the contingency is overstated, clients with multi-project programs can lose an opportunity of having more projects programmed. If the contingency is too low, the project will be at risk of being shelved or plagued with much unfortunate publicity on cost escalation.

1 INTRODUCTION Engineers have been very successful in predicting micro-cracks in their structures yet are criticized for their inability to predict costs associated with these structures. The literature shows that many projects that initially appeared to be quite economical, feasible and within the available budget, ended up with large budget overruns. A classical example is the Channel tunnel where the initial budget was overspent by billions of dollars as the final cost almost doubled the initial estimate. Another example is the supersonic “Concord”, a joint production between British Airways and Air France. The Concord project cost escalated as much as 700% of its original estimate (Morris & Hough 1987). In the United States, the Department of Energy (DOE) records show that nearly 50% of environmental projects overrun their cost (Deikmann & Featherman 1998). Table 1 shows a list of some other famous construction projects that overspent the initial cost estimate such as Great Belt Tunnel, Øresund Access Links (Denmark), Central Artery Tunnel, Washington Metro (USA), Humber Bridge, Tyne and Wear Metro (UK). A study on cost overrun was carried out by US Department of Transportation and covered 10 US rail

transit projects with a total construction value of $15.5 billion (1988 Dollars). This study showed an average cost overrun by 61% (Flyvbjerg et al. 2003). These projects and many more megaprojects underestimated the initial cost estimate. Although estimators tend to apply contingency factors to inflate the initial estimate, these contingency factors were not sufficient to cover the final escalated cost. In the other hand, a study conducted on 287 construction projects funded by the Hong Kong government showed that contingency was over estimated Table 1. Examples of construction cost overrun. Project*

Approximate cost overrun (%)

Central Artery Tunnel Humber Bridge Great Belt Rail Tunnel Washington Metro Channel Tunnel Øresund Access Links Tyne Wear Metro

196 175 110 85 80 70 55

* These examples and more can be found in Flyvbjerg et al. (2003).

313 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-39.qxd

2/18/04

1:49 PM

Page 314

with an average factor of 3 in these projects (Mak & Picken 2000). The same study showed that when the Hong Kong Government implemented the ERA concept (Estimating using Risk Analysis) the contingency was still over estimated but only by a factor of 2 in a total of 45 construction projects. Although this shows a considerable progress in the engineers’ estimates, it is still considerably higher than they should be. That might be due to unfamiliarity with new implementation and the level of confidence that needed to be developed with the technique. Contingency can be better estimated based on quantitative risk assessment. Different project elements can be assessed separately based on the risk associated with them. This method can estimate the proper contingency for each specific project and should be added to the engineer’s tool kit in the early design phases. Moreover, successful implementation of this technique can contribute to the decision making process and properly address the public concerns regarding cost benefit ratio of the project. This technique will be invaluable if adopted for large projects with high uncertainties such as tunnels construction. This paper proposes a procedure that reduces the risk of having inaccurate Engineer’s Estimate by quantitatively assessing all risk items and analyzing the output for better contingency estimation especially in projects at the conceptual phase. 2 QUANTITIVE TECHNIQUES USED FOR CONTINGENCY ESTIMATION In most cases, the project cost is presented as a deterministic figure comprising a base estimate and the addition of a single contingency amount. Usually the contingency amount is an added percentage. This amount is often presented as a single lump sum with no attempt made to identify, describe and evaluate various categories and possible areas of uncertainty and risk. 2.1

What if scenario

In some cases a more detailed analysis can be generated by assuming a “what if scenario” (Vose 2000) which assumes that every event to have three values: maximum, minimum and most likely. A probability tree is constructed for all the possible events as shown in Figure 1. The project estimated cost is calculated based on this probability tree, which might be a reasonable operation for only a small number of events. A drawback for this technique is that each event is assigned by only three values, without any probability for inbetween outcomes. Additionally, this method assumes equal probability for all events, i.e. the probability of having the maximum cost is equal to the probability of

Max. .

B Max.

Min.

Max.

A

Most Likely

B

Most Likely Min.

Min.

Max.

B

Most Likely Min.

Figure 1. Demonstration for a “What if scenario” for cost estimate of two Events A & B.

having the average and minimum. Thus, this method may predict an unreasonable cost for the project. A more reasonable technique is the use of probability function to simulate each event. The total cost of the project is calculated as probability function based on the probability of occurrence of all events and this can be done using any statistical simulation technique such as Monte Carlo. Obviously, the final output probability function will give a range of estimates, i.e. not a single point estimate, and will leave the engineer to decide which exact value should be used. The common method used for such analysis will be discussed in the following section. 2.2

Statistical simulation techniques

If we consider the project cost is a simple summation of many different items and all these items are variables that don’t have a strictly defined cost value. The question will be how can we add them? Simulation techniques can carry out this addition process in two steps. First step, all items are represented by probability functions that simulate the probable cost for each of them. Second step, the total cost is the total sum of all of the probability function. This can be done by using any of the following techniques. 2.2.1 Exact algebraic simulation If all the risk elements can be simulated by mathematical functions, then, well-established algebraic methods can be used for determining the probability distribution function for the total cost. However, this is only true for simple probability functions (Vose 2000). In fact it is almost impossible to analyze a multifactored problem using this technique due to the

314 Copyright © 2004 Taylor & Francis Group plc, London, UK

Most Likely

09096-39.qxd

2/18/04

1:49 PM

Page 315

mathematical complexity (assuming tens of activities each with a different probability function). The exact solution might not even exist. Instead, random sampling is used, which can produce an accurate and simple analysis tool. 2.2.2 Monte Carlo simulation The Monte Carlo method provides approximate solutions to a variety of mathematical problems by performing statistical sampling experiments on a computer. The method is named after the city in the Monaco principality, because of the roulette, a simple random number generator (Pllana 2000). The development of Monte Carlo methods dated from the end of the nineteenth century. However, the real use of Monte Carlo methods as a research and practical tool stems from work on the atomic bomb during the Second World War. This work involved a direct simulation of the probabilistic problems concerned with random neutron diffusion in fissile material. The main idea of the Monte Carlo simulation is generating random output values based on the probability of its occurrence. In other words, the total cost can be calculated by randomly picking possible cost items and adding them up, so if we run this scheme once we simulate one possible total cost. Repeating this process hundreds and even thousands of times will precisely simulate all the possible scenarios of such random event. Now with the use of computers, Monte Carlo simulation method can generate excellent results, as the procedure of generating output can be repeated as many times as required. Using this simple idea, all scenarios are simulated quantitatively and therefore, complex problems that involve multifactors can be modeled adequately. 2.2.3 Latin Hypercube Sampling Latin Hypercube Sampling (LHS) is a more recently introduced technique for probability simulation (Wyss, G. & K. Jorgensen, 1998). LHS software was originally developed by Sandia National Laboratories, (operated by the Department of Energy’s National Nuclear Security Administration). Most risk analysis softwares nowadays use or at least have the option to use LHS. The idea of LHS is simply splitting the probability distribution into “n” intervals each with equal probability. The random sampling procedure is performed in a way that if one of the intervals is sampled in the first iteration, it is marked as having already been used, and therefore, will not be selected again. Thus, the probability function is modeled more uniformly. Additionally, the number of iterations required to precisely model a probability function by LHS is much less than that required by Monte Carlo. In general LHS is preferable to Monte Carlo. The drawback of using LHS, is that iteration time increases tremendously. It must be noted that Monte

Carlo precision can also be increased by performing more iterations, i.e. more processing time. 2.3

Simulation of a project consists of two parts, simulation of the construction cost and simulation of the construction schedule. The total project estimate should incorporate these two items. In many cases ignoring the schedule simulation might result in underestimate of the construction cost especially in projects with high risk and long duration such as tunnel construction and other heavy construction in urban areas. Additionally, schedule in many cases might be the highest risk against performing the project within budget especially in projects that have a public interest and political weight. 2.3.1 Construction cost simulation Construction cost simulation is a fairly straightforward operation only when required expertise and resources are available. It starts with itemizing all the cost elements. Every cost element is then described with a probability function that can take any shape. The probability shape can be derived from expert opinion and old data of previous projects. When using old data, the original estimate and the final estimate should be recorded. More data will increase the accuracy of final estimate. The expert opinion is crucial in working with this data as experience should be the ruling factor since each project is unique and has its own influencing factors. The selected probability function should include the most likely cost (with no contingency added) as the peak of the probability function. Then one of the simulation techniques (Monte Carlo or LHS) is used to sum all the cost items and come up the probability function of the total cost estimate. Thus the estimator can select the most probable cost assigned degree of confidence (for example, the cost that has 90% or 95% probability). 2.3.2 Construction schedule simulation Construction schedule simulation is based on building a network of activities (FTA 1996), where each activity duration (or group of activities) is presented by a probability function. Then, the total duration of the project is calculated. This technique is commonly known as PERT (Program Evaluation and Review Techniques). The total project duration is estimated finally as a probability function. This function can be obtained by applying either Monte Carlo or LHS on all the activities probability functions. This way, the engineer will have a better picture of all the possible scenarios. In large construction projects such as tunneling, schedule simulation is very important for simulating major uncertainties such as TBM advance rates and

315 Copyright © 2004 Taylor & Francis Group plc, London, UK

Simulation procedure

09096-39.qxd

2/18/04

1:49 PM

Page 316

labor productivity. The actual project cost is very sensitive to such elements and a careful cost estimate should include special consideration for incorporating this fact. Selection of a probability curves based on historical data Choosing the probability function is tricky and one should be very rational and systematic. The first step is usually selecting a probability function shape. A lot of research is performed in selecting the probability function shape (refer to Back et al. 2000, Maio et al. 2000 and Fente et al. 2000). The most commonly used distribution is the triangular distribution, mainly because of simplicity. Next step is defining the parameters maximum, minimum and most likely cost; here comes the difficult part where old projects come into the picture. Old projects data can be gathered then adjusted to take into consideration the difference in situations, project nature and project location. This adjustment is based on expertise and engineer’s judgment. In other words the probability function is the best guess based on the available information. Additional factors such as public concerns, political impact and allocating funding at some stages might be implemented in the analysis. This can be done by adding for example extra cost for delays as a separate item.

Probability Cost at Final Design Stages

2.3.3

Cost at Initial Design Stages

Cost Figure 2. Example of probability function change during design development.

Design Development 100% 90% 60% 30% Preliminary Conceptual

3 AVAILABLE COMMERCIAL SOFTWARE Many commercial software packages can perform this operation for example: @risk by Risk-Modelling, Crystal Ball by Decisioneering, and Project Risk Analysis by Katmar. Moreover, some other software packages are dedicated to simulate scheduling procedure such as P3 and Monte Carlo, by Primavera Systems Inc. Although these software packages provide an excellent tool for analysis, simple spread sheets (for example MS Excel) can also be used for simple projects (with a small number of activities).

4 CONTINGENCY ESTIMATION DURING PROJECT DEVELOPMENT The described procedure can be used at different project development levels. The more the project is developed toward its final shape the more uncertainty is diminished. For example, during preliminary engineering phases, probability functions for most elements can be simulated as a wide cost variable, but at the final design stage the range of cost decreases as shown in Figure 2. Some researchers attempted to predict the

5%

20% 30% Contingency

40%

50%

Figure 3. Example change in contingency estimate during project design development.

accuracy of the estimate based on the estimate quality and level of confidence that the estimator has (Oberlender & Trost 2001 and Trost & Oberlender 2003). This way, some lessons can be learned of how the input probability function can be defined or outlined based on what level of detail we have. Since large project are designed in stages (for example conceptual, preliminary, 30%, 60%, 90% and 100%), it is also prudent to reassess the contingency used at each stage. For example, many uncertainties can be cleared as the design progresses. Thus the proposed quantitative probabilistic approach outlined in this paper should be implemented at each design level coupled with the bottom up cost estimate associated with each design stage. The new design development should refine the cost estimate. Figures 3 outlines the changes in contingency estimate as the design progresses to final stages.

316 Copyright © 2004 Taylor & Francis Group plc, London, UK

10%

09096-39.qxd

2/18/04

1:49 PM

Page 317

schedule activities. It must be noted that a major element in this study was the dependence on previous project data. In reality this is the only way of gaining credibility for this method.

Define Scope Element

Estimated cost per Element

Similar Previous Project

Estimated Durations

4.2

The output probability function will show the total cost as a range of estimates, i.e. not a single point estimate, and will leave the engineer to decide which exact value should be used. For example the value that gives 90% or 95% confidence can be selected. The cost probability shapes can also be shared with the client, and thus the client’s opinion can be also integrated to suit the overall budget allocation.

Careful Selection and Comparison of all Element Similar to Old Project – Initial verses final should be looked at

Probable Cost per Element (Max. Min. & probability shape)

Analysis output

Probable Duration per Element (Max. Min. & probability shape)

5 CONCLUSION Analyze Total Project Cost using Monte Carlo or LHS Simulation

Cost Estimation is a very challenging task that faces every construction project. Our duty as designers is to present our best knowledge, tools and expertise to inform our client with a precise cost estimate. This goal can be achieved by quantitatively simulating all construction cost items and schedule to come up with precise contingencies that cover all possible scenarios. Large projects, specifically tunneling projects, will have a more robust estimate if schedule elements are analyzed in the contingency estimate as well. A database of previous experience with mega-projects can provide an essential resource for carrying out the procedure outlined in this paper with real success.

Probable Total Cost

Yes

Design Development

No Final Cost Estimate

REFERENCES Figure 4. Layout of cost estimate development during design development using historical data and probabilistic approach.

The cost estimate elements should be reassessed at each design stage. The assessment should include scope of work, schedule, and market condition, (labor rates, productivity, materials, etc.). The assessment should be compared with the original scope old records should be used for reassessment. Figure 4 outlines the procedure of applying the probabilistic approach for cost estimate during each design stage of the project. 4.1

Example

The Federal Transit Administration (FTA) had issued a document “Probabilistic Risk Analysis for Turnkey Construction: Case Study” which contains a detailed cost analysis based on simulating both, cost items and

Back, E., Boles, W. and G. Fry (2000) “Defining Triangular Probability Distribution from Historical Cost Data” Journal of Construction Engineering and Management, ASCE, vol. 126, no.1, pp. 29–37. Deikmann, J. and D. Featherman (1998) “Assessing Cost Uncertainty: Lessons from Environmental Restoration Projects” Journal of Construction Engineering and Management, ASCE, vol.124, no.6, pp. 445–251. Fente, J., Schexnayder, C. and K. Knutson (2000) “Defining a Probability Distribution Function for Construction Simulation” Journal of Construction Engineering and Management, ASCE, vol.126, no. 3, pp. 234–241. Flyvbjerg, B., Bruzekius, N. and W. Rothengatter (2003) Megaprojects and Risk, An Anatomy of Ambition, Cambridge University Press, Cambridge, UK. FTA (1996) Probabilistic Risk Analysis for Turnkey Construction: Case Study, Federal Transit Adminstration, Office of Planning, Report: FTA-MD-26-7001-96-2. Maio, C., Schexnayder, C., Knutson, K. and S. Weber (2000) “Probability Distribution Functions for Construction Simulation” Journal of Construction Engineering and Management, ASCE, vol.126, no.4, pp. 285–292.

317 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-39.qxd

2/18/04

1:49 PM

Page 318

Mak, S. and D. Picken (2000) “Using Risk Analysis to Determine Construction Project Contingencies” Journal of Construction Engineering and Management, ASCE, vol.126, no. 2, pp. 130–136. Morris, P. and G. Hough (1987) The Anatomy of Major Project – A Study of the Reality of Project Management, John Wiley & Sons, New York. Oberlender, G. and S. Trost (2001) “Predicting Accuracy of Early Cost Estimates Based on Estimate Quality”, Journal of Construction Engineering and Management, ASCE, vol. 127, no. 3, pp. 173–182. Pllana, S. (2000) “History of Monte Carlo Method” http://www.geocities.com/ CollegePark / Quad/2435/ history.html.

Trost, S. and G. Oberlender (2003) “Predicting Accuracy of Early Cost Estimates Using Factor Analysis and Multivariate Regression”, Journal of Construction Engineering and Management, ASCE, vol. 129, no. 2, pp. 198–284. Vose, David (2000) Risk Analysis A Quantitative Guide, 2nd Edition, John Wiley & Sons, LTD, New York. Wyss, G. and K. Jorgensen (1998) A User’s Guide to LHS: Sandia’s Latin Hypercube Sampling Software, Sandia National Laboratories, Report no. SAND98-0210.

318 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-40.qxd

2/18/04

1:50 PM

Page 319

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Geotechnical mapping methods utilized in the Chattahoochee Tunnel Project, Cobb County, Georgia, USA J. Reineke, J. Raymer & M. Feeney Jordan, Jones & Goulding, Inc., Norcross, Georgia

K. Kilby Engineering Design Technologies, Inc., Marietta, Georgia

ABSTRACT: The mapping of the entire 15,125 meter length of the Chattahoochee Tunnel provided an opportunity to consult with the contractor directly during excavation and installation of support as well as compare actual tunneling conditions versus those predicted in the geotechnical base line. Mapping was conducted using the RMR and Q classifications modified for the low grade metamorphic rocks found along the tunnel alignment. The mapped quantities, predicted quantities, and the actual installed support all fell to within a few percentage points of those expected.

1 INTRODUCTION Conducting an accurate geotechnical baseline study predicting ground conditions for a proposed tunnel is a key element for both the tunnel’s alignment and design engineering. The construction of the Chattahoochee tunnel in Cobb County Ga., which combined both a detailed preliminary geotechnical investigation as well as detailed geotechnical mapping of the excavated tunnel during construction, allowed for a direct comparison of the predicted ground conditions versus those actually observed during excavation. The same combination of engineering, statistical, and geological methods used to identify ground types, predict ground behavior, and estimate tunnel support quantities were used to map the excavated tunnel for the necessity of the installed support as well as evaluate the actual ground conditions versus those predicted. 2 PROJECT DESCRIPTION The Chattahoochee Tunnel is a deep hard rock sewer tunnel under construction in eastern Cobb County, (metro Atlanta) Georgia. The tunnel is 15,125 meters in length with an excavated diameter of 5.58 meters and ranges in depth from approximately 30 to 130 meters below ground surface. Excavation of 14,580 meters was accomplished utilizing two hard rock Robbins Tunnel Boring Machines (TBM). The remaining 545 meters

as well as several shafts, chambers, and connecter tunnels were excavated using traditional drill and shoot methods. The geotechnical investigation began in October 1998, with design work completed in 1999, and construction starting in June, 2000. TBM excavation began in August 2001 and was completed in December 2002. Installation of the cast in place concrete lining is currently in progress as of November 2003. The project is scheduled for completion in the summer of 2004. 3 GEOLOGIC SETTING Located in the Piedmont region of the Southern Appalachians, the tunnel alignment consists of medium-grade (greenschist) metamorphic rocks with a few small bodies of granite (Raymer, Reineke 2001). The most common rock types are biotitequartz-feldspar gneiss and mica schist. The rock is strongly foliated with a fairly uniform dip of about 35 degrees to the southeast. The fractures, or “joints”, in the bedrock exhibit three particular characteristics that are key to this analysis. First, the joints tend to be conduits for groundwater, which leads to weathering. Weathering is especially severe in the feldspar-rich gneisses where the tunnel is shallow. Weathering has been observed to produce soil seams up to about one meter in thickness along the larger joint planes. Second, many joints occur in closely spaced en-echelon swarms of limited

319 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-40.qxd

2/18/04

1:50 PM

Page 320

width and length. The joints in the middle of the swarms are longer and more mineralized, whereas the joints along the edges are short and tight. Third, the rock contains substantial in-situ stress and is undergoing active exfoliation related to geologically recent regional uplift and lithostatic unloading. This active exfoliation produces incipient joints. Incipient joints are partially formed joints in the process of propagating. Along its length, a single incipient joint may be tight, open, or merely a plane of greatly weakened rock. Incipient joints are especially common along micaceous foliation.

install was made by the contractor based on actual observed conditions at the heading. Tunnel mapping was performed immediately behind the TBM by the construction engineer to verify that the quantities and type of installed support were appropriate for the ground conditions. RMR (Bieniawski 1989) and Q (Barton 2000) were used together to classify the mapped ground as Type A, Type B, or Type C. This paper compares the tunnel mapping results to the statistical analysis of expected ground types to illustrate the accuracy of the statistical method and where it could be improved to be more representative of conditions inside the tunnel.

4 GEOTECHNICAL INVESTIGATION 6 GROUND TYPES AND SUPPORT The geotechnical investigation was integrated directly into the design process. The work included 50 HQ diamond core borings, laboratory testing, and field mapping. Cores were taken to a depth of 10 meters below tunnel invert resulting in approximately 3,500 meters of recovered core. Because the work was performed in an urban environment, most of the borings could not be drilled directly on the alignment and many were more than 100 meters away. A concise data report and baseline report were prepared to tell the contractor what conditions to expect. From the core borings the tunnel was divided into reaches based on geologic formation boundaries. For each rock type within these boundaries a series of tests were performed to determine strength and borability rates. These tests included point load testing, Brazilian and UCS tests for intact rock strength, Chechar tests for abrasivity, and packer tests for permeability. 5 GENERAL APPROACH Three ground types were established for the tunnel during the geotechnical investigation. With Type A ground as the best ground and Type C as the worst. The expected behavior for each type was described in the geotechnical baseline report. Expected quantities of each ground type were estimated and a baseline established using a combination of geological correlation and statistical analysis. Geological correlation consisted of recognizing particular features in the core or ground surface and projecting those features into the tunnel at specific locations. Statistical analysis consisted of compiling statistical distributions of core measurements and applying those distributions to the tunnel as a whole. Specific support systems for each ground type were designed based on an engineering analysis of the geotechnical conditions. Each ground type was bid and paid at a separate unit rate based on the support actually installed. The decision of which support to

Type A ground is the best. It consists of hard rock ranging from massive to lightly fractured that can generally stand unsupported or that requires minimal support. For the Chattahoochee Tunnel a light pattern of 4 2-meter bolts were placed on 1.8 meter centers in the crown and upper sidewalls. The continuous use of light-pattern bolts was primarily for the safety of the work force by reducing the risk of isolated rock wedges becoming detached from the tunnel crown. TBM advance is typically limited by the strength of the rock in Type A ground. Type B ground is intermediate. It is described as somewhat blocky with large wedges and thin soil seams, including fault gouge. Rock strength may be reduced slightly by weathering and requires heavypattern rock bolts for support. The design called for 6 2.4-meter rock bolts to be placed in the crown and sidewalls on 1.2 meter centers. Type B support is designed to support large wedges of rock in the crown and upper sidewalls of the tunnel. In practice, 2-meter bolts were placed on 1.2 meter centers with 2-meter spot bolts placed according to the jointing patterns in the rock. The heavy-pattern bolts and spot bolts were typically supplemented with mine straps, welded wire mesh, and rolled channel as needed to support particular conditions. In some Type B areas, TBM advance is limited by the time needed to negotiate the ground and install support, as opposed to the strength of the rock. Type C ground is the most difficult. It is generally blocky, seamy and prone to raveling. Rock bolts are generally ineffective. Rock strength is typically reduced substantially by weathering and thick seams of soil may be present. Designed support for Type C conditions consists of full-circle steel ribs with welded wire mesh and lagging where needed. The steel ribs were spaced on 1.2 meter centers and were capable of supporting a zone of heavily fractured rock extending up to one tunnel diameter above the excavated tunnel. TBM advance in Type C ground is typically limited

320 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-40.qxd

2/18/04

1:50 PM

Page 321

by the time needed to install the support or to negotiate problem conditions, including clogged cutters or poor gripping.

Table 1. Definition of weathering indices.

Index 0

very good

1

good

2

fair

3

poor

4

very poor

5

very poor

7 TUNNEL MAPPING Geotechnical maps were created during the excavation for the purpose of verifying the baseline conditions. Mapping was conducted daily with intervals typically covering between 30 to 100 meters. The map areas were sketched in plan view at 3 meters  2.5 cm (10 ft  1 in) scale representing the tunnel crown from springline to springline. Each map was then broken into a continuous sequence of “structure zones”. Each individual zone was then described in terms of RMR (Bieniawski 1989) and Q (Barton 2000) which had been specifically modified to describe ground conditions in the low grade metamorphic rocks in the Atlanta area. The continuous sequence of RMR and Q scores from each zone could then be tallied to calculate the actual amount of each ground type encountered during excavation. Structure zones are defined as areas of similar general features that can be reasonably mapped at the 3 meters  2.5 cm scale. Rock type, weathering pattern, jointing pattern, and unfavorable joint intersections producing blocky ground were all considered in defining and describing individual structure zones. Once the structure zones were defined each was individually described in terms of the categories needed to calculate RMR and Q for that zone.

8 RMR CATEGORIES There are six RMR categories: rock strength, rock quality designation, joint spacing, joint condition, groundwater, and joint orientation. The procedures for qualifying each of these categories are described in the following paragraphs. Weathering index (WI) is a qualitative rating from 0 to 5 based on the appearance of the rock and jointing in terms of the definitions on Table 1. While WI does not correlate directly with an individual RMR or Q score it does heavily influence the rock strength, joint condition, and joint alteration values. WI was measured for each zone as a whole with indices 0 through 2 concerned only with the joints themselves and weathering indices 3, 4, and 5 related to the rock matrix as well as the joints. In zones that contained more than one WI value the higher index value was recorded. Rock strength Seven major and several minor low grade metamorphic lithologies were observed during the excavation of the Chattahoochee Tunnel.

Definition No visible degradation or staining, even along joints. Joints generally tight, tightly healed with mineral cements, or lined with euhedral crystals. No visible degradation or discoloration of rock matrix. Visible degradation or staining limited to surface of major joints. Visible degradation or discoloration extends into rock matrix for a short distance along joints but does not fully permeate the rock. Rock appears to be full strength except along joints. Visible degradation or discoloration extends through rock matrix. Feldspar and mafic minerals partially but not fully degraded. Matrix strength of rock is considerably reduced, but the rock is not breakable by finger pressure alone. Can generally be recovered using standard rock coring techniques but typically with low RQD values. Generally causes auger refusal for small drilling rigs. Matrix strength of rock is greatly reduced, and rock should mostly be breakable with strong finger pressure. Weathering permeates entire rock. Generally cannot be recovered using rock coring techniques. Rock completely degraded into soil but with the original texture and structure of the rock clearly visible. (In the Atlanta area, this is known as saprolite.) Cannot be recovered using standard rock coring techniques.

These ranged from mylonite, migmatite, gneiss, schist, amphibolite, granite, and interlayered gneiss and schist. While direct measurements of rock strength for each lithology were not made during excavation, extensive point load testing was conducted on the major lithologies during the preliminary investigation. Rock strengths of 4 MPa (580 psi) were recorded during point load testing of all unweathered samples with the exception of one 2–4 MPa (290–580 psi) graphite bearing schist. The rock was assumed to be at full strength unless modified by weathering indices 3–5. The RMR score was reduced depending on the severity of weathering (see Table13 RMR & Q Scoresheet). The rock quality designation (RQD) is measured horizontally as a percentage rock 11 cm (4 in.) measured horizontally between open joints or areas of WI  5 material. The percent RQD is assigned points for use in calculating RMR for each zone as illustrated in Table 2.

321 Copyright © 2004 Taylor & Francis Group plc, London, UK

RMR rating

09096-40.qxd

2/18/04

1:50 PM

Page 322

Table 4. Groundwater.

Table 2. RQD. RQD range (%)

Points

Groups

Points

Descriptions

90–100 75–90 50–75 25–50 25

20 17 13 8 3

Dry Damp Seeping Flowing

15 10 5 0

Joints and rock face dry Area visibly damp, no drips Water visibly dripping or moving Steady flow with constant stream

Table 5. Joint orientation. Table 3. Joint spacing (Js). Spacing code A B C D E F Laminations

Sub codes

B1 B2 C1 C2 D1 D2

Spacing (ft)

Score

6 3.5 to 6 2 to 3.5 1.1 to 2 0.6 to 1.1 0.3 to 0.6 0.2 to 0.3 0.06 to 0.2 0.02 to 0.06 0.006

20 15 15 10 10 8 8 5 5 5

Groups

Points

Explanation

Very favorable Favorable

0 2

Fair

5

Unfavorable

10

Very unfavorable

12

No possibility of wedges Rare wedges on less common joint sets Wedges on less common joint sets Wedges common on major joint sets Large wedges certain

9 RMR CALCULATION

Joint spacing (Js) was measured as an average spacing between the joints of an individual joint set. When two or more joint sets are found within one structure zone the set with the closest spacing is generally used for the RMR calculation. On a few occasions a more widely spaced set may be scored due to its overall structural influence of the interval. The average joint spacing for the chosen set is then given a score based on the values listed in Table 3. Joint condition is a qualitative value assigned based upon the degree of weathering found in the structure zone (Table 13 RMR & Q Scoresheet). For RMR calculation in the Chattahoochee Tunnel a combination of weathering index (see above) and joint alteration values (see Q Categories below) were assessed to assign a RMR value between 0 and 30 for the RMR calculation. Groundwater inflows into individual zones were estimated qualitatively using the descriptions in Table 4 and a contribution to RMR point value assigned accordingly. Joint orientation concerns the interaction and orientation of joints or joint sets in relation to the excavated tunnel. Most wedges and blocks tend to occur within heavily jointed areas with low RQD percentage; however the unfavorable intersection of two isolated joints within otherwise massive rock can result in the formation of a potentially hazardous wedge. These interactions can also be favorable, effectively “locking” potential wedges into place with in the tunnel wall. Joint Orientation scores are subtracted from the RMR score with the point values and descriptions listed below in Table 5.

The scores from each of the six RMR categories are recorded on the RMR & Q score sheet and added together to obtain the RMR total for that particular structure zone (Table 13 RMR & Q Scoresheet). 10 Q CATEGORIES There are five categories which combine to form the Q score. Quantification procedures for each of these categories are described in the following paragraphs. Joint set number (Jn) is a measure of the number of joint sets found within the structure zone. A joint set is defined for mapping purposes to be two or more individual joints that are parallel in both strike and dip orientation. For the purposes of zone description, joints with no parallel counterpart within the zone boundary were considered “random”, even if two or more in the same orientation can be found in adjacent zones. The Jn was assigned a code (see Table 6) during mapping that relates to a specific value in the Q calculation. Jn is also used as a reference value during the determination of the joint orientation score for RMR. Joint roughness (Jr) describes the average surface texture of each joint set (Table 7). This measure helps to quantify the likelihood slippage occurring on that particular joint set. Jr is recorded for all joint sets within the structure zone. In intervals with two or more joint sets, the Jr value for the joint set determined to have the greatest structural influence is chosen. Joint alteration (Ja) describes the average degree of alteration of the each joint set found within the zone in question (Table 8). Weathering Index and Ja

322 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-40.qxd

2/18/04

1:50 PM

Page 323

Table 6. Joint set number (Jn).

Table 9. Stress reduction factor (SRF).

Code

Score

Short description

Code

Score

Short description

A1 A2 B C D E F G H J

0.5 1 2 3 4 6 9 12 15 20

Massive Few random joints One set One set plus random Two sets Two sets plus random Three sets Three sets plus random Four or more sets Crushed rock or soil

A B C D E F G H J

10 5 2.5 7.5 5 2.5 5 2.5 1

Multiple weakness zones (clay, WI  4) Single weakness zones 150 ft deep Single weakness zones 150 ft deep Multiple clay-free shear zones Single clay-free shear zone 150 ft deep Single clay-free shear zone 150 ft deep Loose, open joints Near surface, open joints common Normal hard rock tunnel conditions

Table 10. Groundwater inflow.

x2 for portal; x3 for intersection.

Table 7. Joint roughness (Jr). Code

Score

Short description

A B C D E F G H J

4 3 2 1.5 1.5 1 0.5 1 1

Discontinuous joints Rough undulating Smooth undulating Slickensided undulating Rough planar Smooth planar Slickensided planar Thick clay zones Sandy or gravelly crush zones

Add 1 to score if spacing is 10 ft.

Code

Score

Short description

A B C D

1 0.66 0.5 0.33

Dry, damp, or seeping Flowing under low pressure Gushing through open joints Gushing with outwash of joint filling

Table 11. Ground type. Ground type

RMR value

Q Equivilant1

Q Equvilant2

A B C

61 to 100 41 to 61 0 to 41

6.5 to 500 0.72 to 6.5 0.001 to 0.72

5.5 to 1333 0.25 to 5.5 0.001 to 0.25

1 2

RMRQ  9 ln Q  44 correlation based on Bieniawski, 1989. RMRQ  15 log Q  50 correlation based on Barton, 2000.

Table 8. Joint alteration (Ja). Code

Score

A

0.75

B C D E

1 2 3 4

F G H J K

4 6 8 12 6

L

8

M

12

N

5

P Q R

10 13 18

Table 12. Comparison of percentage ground types based on tunnel mapping, core analysis, and installed support.

Short description Tight, fresh foliation breaks (no mica), qtz filling. Non-slick surface stains Slightly degraded rock, sandy coatings Lean clay wall coatings Soft or low friction coatings – mica, gypsum, etc. Weathered rock 10 cm thick Hard lean clay fillings 5 mm thick Softer lean clay fillings, 5 mm Fat clay fillings 5 mm Bands of weathered rock with hard lean clay 5 mm Bands of weathered rock with softer lean clay 5 mm Bands of weathered rock with fat clay 5 mm Bands of weathered rock, no clay (WI 4) 10 mm Thick bands of hard lean clay Thick bands of softer lean clay Thick bands of fat clay

Basis

Type A

Type B

Type C

Mapped RMR Mapped Q (Bieniawski) Mapped Q (Barton) Core analysis prediction As-Installed

95.5% 95.1% 95.6% 96.0% 94.2%

3.2% 4.1% 4.2% 3.8% 4.0%

1.3% 0.8% 0.2% 0.2% 1.8%

are very closely related, however; WI takes into consideration the rock matrix containing the joint where as Ja deals exclusively with joints and joint fillings. The Ja was categorized and assigned a value for use in the Q calculation. Stress reduction factor (SRF ) is a measure of the number and severity of planes or areas of weakness found with in the described zone (Table 9). SRF is part of the Q calculation. The rock quality designation (RQD) is measured as a percentage rock within the zone 11 cm (4 in.) measured between open joints or areas of WI  5

323 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-40.qxd

2/18/04

Table 13.

1:50 PM

Page 324

RMR & Q score sheet.

material (the same RQD percent is used for both RMR and Q). The percent RQD is the starting point of the Q calculation. Groundwater inflows (Jw) The Q calculation qualitatively scores flow rates using the descriptions found in Table 10.

11 Q CALCULATION The total Q value is calculated by entering the scores for the five measured categories in to Equation 1 below.

324 Copyright © 2004 Taylor & Francis Group plc, London, UK

(1)

09096-40.qxd

2/18/04

1:50 PM

Page 325

12 DETERMINATION OF GROUND TYPE The three defined ground types, A, B, and C, were defined by RMR values during the geotechnical investigation. Unlike the RMR system, not all of the Q parameters are amenable to core analysis without a large degree of interpretation on the part of the person logging the core. Also several of the Q parameters are also very sensitive to scale, which makes them difficult to apply to core. The Q system was, however, very powerful and clear when used at full scale in the tunnel. This required the Q score to be converted into an equivalent RMR value. The Q value is correlated to the RMR score using both Bieniawski’s equation, RMR  9lnQ  44, and Barton’s equation RMR  15log Q  50 for direct comparison. Both of these values and the direct RMR value are used to place each structure zone into one of the three ground type categories according to Table 11.

Secondly, the percentage of each ground type that was predicted in the geotechnical baseline report could be compared to the actual excavated ground conditions. The RMR based on core analysis conducted during the geotechnical investigation corresponded reasonably well with the tunnel mapping results. The mapped quantities of Type A, Type B, and Type C ground were within two percentage points of the estimated quantities as shown in Table 12. The geotechnical mapping program provided valuable insight into the methods used during the geotechnical investigation. The results of the mapping program are used to further refine preliminary investigation data for future projects. In addition the mapping program was able to provide accurate detailed geotechnical feed back to the contractor, construction manager, and owner in real time allowing for quicker resolution of issues when ground conditions are less than ideal.

13 DISCUSSION

REFERENCES

The purpose for the detailed mapping of the Chattahoochee Tunnel was two fold. First, mapping allowed for the direct comparison of the installed support versus the mapped ground type. Table 12 indicates the quantity of installed Type A, Type B, and Type C support based on the construction managers records. Table 12 also indicates that the results from mapping are within two percentage points of the quantities actually installed, regardless of whether or not the RMR or Q is used. The best correlation of mapped ground type with installed support is obtained from the directly mapped RMR, as opposed to Q converted to RMR.

Barton, N. TBM Tunneling in Jointed and Faulted Rock, Rotterdam: A.A. Balkema, 2000. Bieniawski, Z.T. Engineering Rock Mass Classifications, John Wiley and Sons, 1989. Deere, D.U. and D.W. Deere. “Rock Quality Designation (RQD) After Twenty Years,” U.S. Army Corps of Engineers, Contract Report No. 6L-89-1, 1989. Raymer, J. and Klecan, W. Predicting Ground Conditions and Support Requirements for Atlanta-Area Tunnels Using the RMR System, RETC Proceedings 2003, pp. 849–863. Raymer, J. and Reineke, J. Structural Styles Along the Chattahoochee Tunnel, Cobb County, Georgia, Georgia Geological Society Guidebook, Vol. 21, pp. 43–50, 2001.

325 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-41.qxd

2/18/04

1:50 PM

Page 327

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Value engineered design facilitates Grand and Bates Relief Sewer Tunnel Construction, St. Louis, MO James R. Wheeler Design/Build Geotechnical, LLC, Stow, Massachusetts, USA

Nancy E. Thomson Brierley Associates, LLC, Denver, Colorado, USA

ABSTRACT: The Grand and Bates Stormwater Relief Tunnel consists of approximately 6550 feet of tunnel, associated shafts and an outfall structure constructed to provide overflow relief in the Grand Glaise Sewer District in the City of St. Louis, Missouri. Upon receiving a single bid, nearly $3M in excess of the Engineer’s Estimate for the work, the owner and tunnel contractor completed a comprehensive review of the project in an effort to reduce the overall budget while maintaining the important engineering characteristics of the project. To reduce project costs and enhance the overall design and construction of the tunnel, designers considered areas of potential challenges during construction and avenues to apply value engineering approaches in an effort to save the project. Working cooperatively, the owner, in concert with the design and construction team, evaluated all important aspects of the project including the proposed alignment, critical construction access and drop shafts, the outfall structure, and key aspects of the proposed finished structures, in conjunction with the owner’s design requirements and the anticipated site and subsurface conditions. As a result of this value engineered design approach, the project was enhanced to the satisfaction of the owner while permitting the contractor to deliver the project at the owner’s budgeted cost.

1 PROJECT DESCRIPTION The Grand and Bates Relief Phase II Stormwater Tunnel was designed for the owner, Metropolitan St. Louis Sewer District (MSD) by TSI Engineering, Inc., St. Louis, MO (TSI) and CH2MHill, St. Louis, MO. Late in the design phase, Brierley Associates, LLC, Boston, MA (BA) was retained as a specialty tunnel consultant to TSI to complete technical provisions of specifications and prepare the Geotechnical Baseline Report for the work. The project was designed to minimize area flooding that had reached depths of 8 feet (2.4 m) in recent years. As initially designed, the project included the construction of approximately 6060 ft. (1847 m) of rock tunnel 12 feet (3.7 m) in diameter to be advanced by tunnel boring machine (TBM). Approximately 4850 ft. (1478 m) of the alignment was to be constructed from a primary access shaft, T11D, excavated to a depth of 108 ft. (32.9 m) at the west end of the alignment. Falling at a 0.1% grade, the tunnel was to be advanced along Bates Street and curve to meet construction access shaft T2D into which a low point pump station

was to be installed. Upon reaching shaft T2D, plans called for the extraction and remobilization of the TBM to into a new starter tunnel to be constructed at a depth of 60 ft. (18.3 m) to complete the remaining 1750 ft. (533 m) bore. Shaft T1D and an outfall structure were to be constructed immediately adjacent to a Union Pacific Rail line at the Mississippi river. The permanent tunnel liner was to consist of a 108-in. (2.7 m) reinforced concrete carrier pipe in cellular grout. Also included in the contract were eight offset drop shafts to be drilled from the surface, and associated connecting adits. A Plan view of the project is included as Figure 1. 2 GROUND CONDITIONS As part of their design work, TSI completed a detailed subsurface exploration program for the project. Geotechnology Inc., St. Louis, MO was retained to drill a total of 35 test borings and 5 probe holes. Rock core samples were obtained from 21 of the borings. In-situ testing consisted of installing 12 piezometers

327 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-41.qxd

2/18/04

1:50 PM

Page 328

Figure 1. Plan view (Hayward Baker Inc.).

and performing well recharge testing in six of the piezometers, performing 19 double packer water pressure tests in 10 of the borings, performing laboratory tests on select soil and rock samples, conducting a geophysical survey along portions of the alignment, and screening soil samples for indications of environmental contamination. Based on the subsurface information obtained, ground conditions along the alignment consisted of three distinct types of ground: overburden soil, highly weathered limestone bedrock, and competent limestone bedrock. The overburden soils consisted primarily of loessial and alluvial deposits with some fill ranging in depth from approximately 25 to 75 ft. (8 to 23 m). Borings completed in the vicinity of the proposed outfall structure encountered rubble fill above natural soils and bedrock. A generalized profile of the tunnel is included as Figure 2. Beneath the weathered bedrock, the native limestone was encountered and was observed to be generally competent but with zones of lower quality material. These lower quality zones were often associated with seams of shale that were softer and/or more susceptible to disintegration under the action of drilling. Other zones indicated rock associated with sinkhole features and/or geologic faults. In general, the limestone bedrock (St. Louis Limestone) was noted to be white to light grey, fine to coarse grained, thin to massively bedded, hard, slightly argillaceous,

and contained chert lenses and nodules. Soft greenishgrey shale seams were common in the upper surface of the rock, and in some areas the shale had weathered to shaley clay. The carbonate rock of this formation is commonly jointed and fractured, and is susceptible to solutioning and the development of irregular and pinnacled rock surface and Karst features. The groundwater levels observed in the piezometers installed along the alignment indicated that groundwater occurred in the overburden soils. Based on the 19 water pressure tests conducted in the bedrock, the hydraulic conductivity of the limestone was relatively low, with the exception of two tests. One test conducted below the highly weathered bedrock zone suggested a hydraulic conductivity of 1.5 103 cm/sec. The other test suggesting higher permeability was completed within the highly weathered zone near the top of rock and indicated a hydraulic conductivity of 2.5 103 cm/sec. These hydro geologic parameters suggested a potential ground water inflow of approximately 100 gpm (7.6 l/sec). 3 VALUE ENGINEERING ALTERNATIVE After bids were submitted for the work, a Value Engineering (VE) phase of the project began as described in the following sections.

328 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-41.qxd

2/18/04

1:50 PM

Page 329

Figure 2. Generalized tunnel profile (Hayward Baker Inc.).

3.1

Bid phase

In February, 2003 bids for the project were submitted with a single bidder, Affholder, Inc. submitting a bid in the amount of $32.9M, which was more than $2.9M in excess of the estimated cost of the project. In an effort to salvage the project, Affholder, Inc. representatives initiated a series of meetings with MSD and their design consultants to develop a Value Engineered alternative to the proposed design that could streamline and enhance the project, share potential risks related to changing subsurface ground conditions, satisfy the owner’s critical, project specific design criteria, and complete the project at the budgeted cost for the work. In evaluating requirements for the project, it was immediately evident that by lowering the eastern portion of the tunnel by nearly 50 ft. (15.2 m) between Shaft T2D and T1D, a single, tunnel bore could be completed for the length of the project, thereby eliminating a costly TBM remobilization at shaft T2D. And, by relocating shaft T2D to the south side of Bates Street and connecting it to the main tunnel with a short adit, the bore could be completed in a straight line, eliminating the difficult and costly horizontal curves required in the original design. (Refer to Figures 1 and 2). Further, the contractor indicated that a larger diameter tunnel could be advanced for

the project using an available TBM that was just completing a project in Chicago, IL. By increasing the tunnel diameter from 12 to 14.5 ft. (3.7 to 4.4 m), a 132 in. (3.4 m) precast concrete tunnel lining could be installed in the tunnel, effectively increasing the overall storage capacity of the tunnel by over 40%. With these initial alternatives in mind, the owner agreed to permit the contractor to develop a VE proposal that would incorporate these initial enhancements and to propose other design revisions that would permit the project to be constructed for the pre bid budget cost. 3.2

To complete the VE design work for the project, the tunnel contractor retained the services of selected members of the project design team including Brierley Associates to complete all tunneling related aspects of the design. Geotechnology, Inc. was retained to complete supplemental test borings along the proposed deepened tunnel alignment and to provide subsequent field testing and instrumentation services. Revisions to the civil design aspects of the project were developed and coordinated by DMA Associates, Inc., St. Louis, MO, a firm that had extensive working experience with the owner in the past. If a successful

329 Copyright © 2004 Taylor & Francis Group plc, London, UK

Value Engineering design phase

09096-41.qxd

2/18/04

1:50 PM

Page 330

Table 1. Summary of Value Engineering design enhancements. No. Design enhancement

Enhancement details

Impacts to proposed design

Benefits to the project

1

Increase tunnel diameter

Increase excavated tunnel diameter from 12 to 14.5 ft. (3.7 to 4.4 m)

Permit installation of 132 in. (3.4 m) precast concrete liner vs. 108 in. (2.7 m) liner

2

Revise vertical tunnel alignment

Lower eastern end of tunnel by 48 ft. (14.6 m)

3

Revise horizontal tunnel alignment

Delete curves in alignment at intermediate Shaft T2D

4

Relocate and reduce size of shaft T2D

Relocate shaft to south of Bates Street and resize shaft to accommodate revised pump structure

5

Incorporate drop shaft into access shaft T11D

6

Use jet grout ground prestabilization as primary support for access shafts

7

Permit limited drill and blast construction

8

Simplify ground support at highway and railroad crossings

9

Resize drop shafts to accommodate precast concrete riser pipe

10

Relocate/redesign outfall structure and shaft T1D

Increase shaft T11D diameter to 36 ft. (11 m) and incorporate drop shaft T10D in the completed access shaft Construct interconnected jet grout columns to create a 4 ft. (1.2 m) thick self supporting compression ring at Shafts T2D and T11D Controlled drilling and blasting construction permitted in bedrock at shafts and for starter tunnel subject to approval of local fire official Increased bedrock cover facilitated ROW permit application process and permitted the use of a single, simplified crown support system at all critical project transportation crossings Drop shafts sized and construction procedures revised to permit use of 84 in. (2.1 m) precast concrete riser pipes Relocate the outfall structure closer to the Mississippi river, beyond limits of construction related impacts to Union Pacific Railroad and streamline the design of the structure

Deepen excavation of shafts T2D and T1D, permit construction of entire alignment in a single bore eliminating a TBM remobilization, provide additional bedrock cover over highway and railroad crossings Simplify and speed up excavation of the resulting straight line tunnel bore Eliminated curves in horizontal alignment, reduced shaft construction cost, limited construction traffic control requirements Eliminate separate excavation of shaft T10D

Increased storage capacity by 40%, contractor permitted to use available TBM to complete the work Eliminated a TBM remobilization, reduced construction schedule and cost

330 Copyright © 2004 Taylor & Francis Group plc, London, UK

Reduced construction schedule and cost Reduced construction schedule and cost

Reduced construction schedule and cost

Provide positive control of groundwater flow and ground movements, eliminate cost of soldier pile/lagging shaft liner, accelerate shaft excavation Required development of controlled blasting specification, permit approval, and blast monitoring during construction The project specified pattern rock bolting required along the remainder of the alignment was supplemented with 8 11.5 in. channel crown sets

Provided positive control of ground movements and ground water inflow, accelerated shaft excavation process Reduced time and cost for critical phases of bedrock excavation

None

Reduced cost and simplified shaft lining construction

Eliminate tied-back soldier pile/lagging excavation support system, use precast concrete culvert pipes for outfall construction

Reduced cost and simplified construction process

Simplified and accelerated ROW permit application and approval, subsequent simplified construction and reduced cost

09096-41.qxd

2/18/04

1:50 PM

Page 331

Figure 3. Tunnel sections – typical section (above) and section at railroad/highway crossings (below).

VE proposal could be negotiated, the cost for the VE consultants work was to be covered as part of the awarded contract. Working in a spirit of cooperation, meetings were held between representatives of MSD and members of the contractor’s VE consultants to brainstorm potential cost saving alternatives and evaluate each proposed enhancement in accordance with the owner’s overall requirements for the project. Both the technical experience of the consultants and the practical construction experience offered by the tunneling contractor were critical in developing a series of effective revisions to the project that resulted in significant savings. Critical aspects of the design that were revised and accepted by the owner are summarized in Table 1 along with comments summarizing the benefits that each revision provided to the contract. As a result of the Value Engineered alternatives analysis, the contract was awarded to Affholder, Inc. for $31.2M. Discussion regarding the major aspects of the redesign work follows. To provide critical information required to evaluate and design a deepened length of the tunnel between Shaft T2D and the Outfall Structure and at the relocated and deepened access shafts, a total of four supplemental borings were completed to define or verify

the in-situ bedrock characteristics below that of the initial geotechnical investigation. Results for the 592 ft. (180.4 m) of additional rock core obtained from the supplemental borings indicated that the underlying bedrock characteristics were consistent with engineering design characteristics and parameters assumed in the initial design. RQD values in the tunnel zone were typically high, at or near 100%. Water pressure tests indicated very low hydraulic conductivity in the bedrock, ranging from no measurable flow to 8.1 103 cm/sec. Based on these confirmed bedrock characteristics, the temporary tunnel ground support included in the original design, consisting primarily of pattern rock bolting at 4 ft. (1.2 m) centers along the crown of the alignment using 5.5 ft. (1.7 m) #8 resin grouted bolts and welded wire fabric was determined to be appropriate. However, based on the significantly increased bedrock cover between the crown of the relocated tunnel and the critical highway and railroad crossings, additional ground support at these locations was no longer required. Still, to provide a degree of additional positive ground support at these critical locations, C8 11.5 channel sets were incorporated into the crown rock bolting pattern at these transportation crossings. Typical tunnel sections for the project are shown in Figure 3. As noted in Table 1, a number of other significant enhancements to the project involved the redesign and construction of the west access shaft T11D and the intermediate access shaft T2D. The simple relocation of shaft T2D to the south of Bates Street permitted a straightened tunnel alignment and eliminated significant traffic control expenses during construction. Increasing the diameter of shaft T11D by 4 ft. (1.2 m) to incorporate the construction of one of the drop shafts within the completed access shaft was another cost saving measure. Other significant shaft related design enhancements involved the use of jet grout stabilized ground. The initial project design called for the use of jet grout prestabilization of the overburden soils prior to the construction of shafts T11D and T2D to limit ground movements and control anticipated groundwater flow at the irregular soil/bedrock interface. In addition, specifications called for pressure grouting of the upper 20 ft. (6.1 m) of bedrock to seal off potential joints in the bedrock to further limit groundwater flow. Specifications also called for the installation of a contractor-designed temporary soldier pile and lagging shaft liner system within the limits of the jet grout stabilized ground. However, evaluations completed during the VE design indicated that by increasing the thickness of the jet grouted soil mass, typically referred to as soilcrete, to 4 ft. (1.2 m) and ensuring a minimum soilcrete compressive strength of 300 psi (2.1 MPa),

331 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-41.qxd

2/18/04

1:50 PM

Page 332

Figure 5. Completed outfall structure and temporary earth support system at shaft T1D.

Figure 4. Section through shaft T11D.

each shaft could be constructed as a self-supporting compression ring with a minimum safety factor of 2.8. In addition, the soldier pile/lagging liner would be replaced by a 3 to 4 in. (0.1 m) nominal thickness of shotcrete placed over welded wire fabric to provide an internal shaft lining and prevent potential freeze thaw degradation and/or spalling of exposed soilcrete. To complete the pressure grouting of the upper bedrock required in the original design, holes would be cored through the soilcrete and into the underlying rock. Soilcrete core samples obtained during the coring would be tested to confirm unconfined compressive strength. Rock bolts consisting of 30 ft. (9.1 m) lengths of #8 rebar would be grouted into each core hole to provide a shear key to anchor the soilcrete shaft to the bedrock and to serve as vertical spiles that would limit bedrock overbreak as the shaft excavation advanced into the bedrock. The resulting design, shown conceptually in Figure 4, provided self-supporting shafts that required no internal bracing during excavation resulting in reduced shaft construction costs and a simplified and accelerated shaft excavation process. To monitor compressive stresses in the soilcrete shafts during construction, provisions were made for the contractor to install vibrating wire strain gages in the soilcrete around the perimeter of each shaft. Gages would be installed in groups of four at three levels in each shaft and monitored throughout the project to ensure that anticipated soilcrete compressive stresses were not exceeded during construction. In the event that data indicated measured stresses were exceeding anticipated levels, the contractor would respond by applying an additional thickness of

shotcrete to the shaft lining to increase its thickness and composite strength as required. Another aspect of the project that was significantly redesigned and enhanced involved the outfall structure and shaft T1D. As initially designed, the outfall structure was to be constructed immediately adjacent to an active Union Pacific Rail line and required an extensive tied back soldier pile/lagging temporary excavation support system to counteract the significant railroad design loading and limit ground movements and track settlement. Based on the depth to bedrock that was noted in the supplementary boring program, relocation of the shaft to the east of its initially proposed location would permit the excavation of the overburden soils to be completed beyond the limits of influence that would be anticipated to contribute to construction related ground movements. As such, a smaller, internally braced soldier pile and lagging system was designed and constructed by the contractor as shown in Figure 5. This provided temporary support for the excavation to bedrock and permitted the use of lower cost precast concrete box culvert sections for the outfall conduit. 4 CONSTRUCTED PERFORMANCE The contractor was given Notice to Proceed on April 15, 2002; mobilization began and continued through May 10, 2002. Construction and jet grouting of the shafts, undertaken by specialty geotechnical contractor Hayward Baker Inc., commenced immediately. The performance of the self-supporting, jet grouted

332 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-41.qxd

2/18/04

1:50 PM

Page 333

Figure 6. Rock bolted channel crown sets provide positive support at railroad crossing area.

shaft walls at shafts T11D and T2D were closely monitored throughout construction by taking continuous readings from the vibrating wire strain gages embedded in the jet grout columns. The data collected from the instruments was plotted regularly to ensure that the soilcrete compression ring was not experiencing stresses greater than anticipated. The data revealed that the soilcrete shaft walls performed as expected during construction, providing the necessary factor of safety for the excavation at a reduced construction cost. As a result of successful jet grouting operations and subsequent stabilization of the alluvial ground mass in two of the access shafts, the contractor was able to excavate the soil portion of these two shafts at a rate of approximately 8 vertical feet (2.4 m) per shift with no running ground or water inflow problems. Unfortunately, the soil/bedrock interface at the smaller diameter drop shafts that were not jet grouted or prestabilized prior to excavation experienced difficulty associated with significant water inflow rates of up to 100 gpm (7.6 l/sec). The contractor chose to operate one 9-hour shift per day, five days per week during shaft sinking and tunneling operations. During pipe installation, operations utilized two, 9-hour shifts per day. Tunnel excavation averaged approximately 70 feet (21.3 m) per shift, with the highest production rate reaching 105 feet (32 m) per shift. These production rates were possible, in part, because the eastern portion of the alignment was lowered and the entire alignment could be excavated entirely in the St. Louis Limestone, which provided a relatively dry, stable medium for tunneling activities. Any potential problems associated with

Figure 7. Positive ground control at self-supported soilcrete access shaft T11D.

the tunnel crossing beneath Interstate I-55 and in the vicinity of the two Union Pacific Railroad crossings were mitigated by modifying the initial ground support pattern, as shown in Figure 6. The project was completed on June 4, 2003 and was on schedule and under budget. During construction, the owner elected to provide the contractor with additional work that included the installation of the helicoidal drop ramps inside the completed drop shafts. When added to the initial contract price of $31.2M, this change order brought the final construction price to $32.9M. 5 SUMMARY AND CONCLUSIONS The Grand and Bates Sewer Tunnel Project is unique not for its construction, but rather for the problems that were solved by the design and construction team prior to construction. With only one bid submitted for the project, and that bid exceeding the owner’s budget, the project was clearly in jeopardy. However, the past experience of the contractor and the members of the design team, and the willingness of the owner to work proactively towards developing a solution to the problem initiated the period of open discussion, brainstorming, redesign, and negotiations that made the

333 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-41.qxd

2/18/04

1:50 PM

Page 334

completion of the project possible. As a result of the Value Engineered design process that followed the bid, the owner was provided with a project that satisfied project design requirements and exceeded project expectations by providing an additional 40% storage capacity at the budgeted project cost. In exchange, the contractor received the benefits of streamlining key aspects of the construction in accordance with past experience and eliminating or limiting costly and unnecessary aspects of the construction. The use of several innovative underground design and construction techniques developed and implemented by qualified personnel, combined with common sense understanding of the project, and the willingness of all parties to work interactively and proactively throughout the work were clearly the keys to the successful completion of the project. ACKNOWLEDGEMENTS The authors would like to express their appreciation to Ms. Pamela Huntoon, MSD for her understanding, cooperation and proactive support during the VE design and construction process. The authors would

also like to thank Mr. Mark Rybak, Affholder, Inc., Project Manager, for his understanding and cooperation throughout the project and for his input and assistance in preparing portions of the text summarizing construction. In addition, thanks are extended to Hayward Baker Inc. for the graphics used to prepare Figures 1 and 2 and to Mr. Shaun Connors, Brierley Associates, LLC for his assistance in preparing the manuscript and remaining graphics.

REFERENCES Burke, J. “Common Sense Cooperation Saves St. Louis Tunnel Project”, World Tunnelling, 2003, pp 485–488. Camper, K., Thomson, N. and Wheeler, J. “Value Engineered Shafts in St. Louis”, Tunnels & Tunnelling International, May 2003, pp 30–33. Pierce, T. “Safe Passage: Jet Grouted Columns on Grand & Bates Tunnel”, Tunnel Business Magazine, December 2002, pp 20–22. TSI Engineering, Inc., “Geotechnical Baseline Report, Grand and Bates Storm Sewer Relief Tunnel, St. Louis, Missouri”, November 19, 2001.

334 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-42.qxd

2/18/04

1:51 PM

Page 335

Session 3, Track 3 Investigation, inspection and rehabilitation

Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-42.qxd

2/18/04

1:51 PM

Page 337

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Monitoring excavations using 3D Laser Scanning and Digital Close-Range Photogrammetry T. Trupp, L. Liu & Y. Hashash University of Illinois at Urbana-Champaign, Urbana, Illinois, USA

ABSTRACT: Monitoring progress of excavations still demands the commitment of significant resources and number of personnel. Procedures used nowadays are time consuming and provide needed data in an impractical time frame. These procedures do not necessarily provide accurate information on construction progress and excavation geometry. This paper introduces the use emerging technologies, (1) 3D Laser Scanning (3DLS) and (2) Digital Close Range Photogrammetry (DCRP) to significantly enhance collection of as-built data to monitor construction progress. The paper describes the principles of the new technologies as well as their capabilities to determine excavation geometry and excavated volume. The paper discusses the usage of collected field data to enhance project management and control.

1 INTRODUCTION The increase in urbanization, population density and traffic congestion in major metropolitan centers around the world has led to greater demand for underground space. Construction of such underground space poses a tremendous challenge to engineers to provide a safe work environment in a timely and cost effective manner. Open cuts and tunnels have long been used to create underground space, but with the increasing usage of underground space, new constraints on the design and construction of such projects are added. One of the major concerns when constructing underground space is the impact of ground movements related to construction activities. These ground movements can be critical to adjacent sensitive buildings and utilities. Hence, it is very important to predict and control the magnitude and distribution of ground movements, which result from building underground space. Nowadays, sophisticated computer aided tools assist designers to predict such ground movements. During construction extensive instrumentation programs are implemented to monitor ground and building movements. However as built data quantifying construction progress is hardly available. In many cases, contracting practices dictate that the contractor is responsible for temporary support systems, therefore designers usually do not have sufficient control of important details in the construction process. In order to obtain accurate construction progress information,

engineers have to rely or manual field reports, or employ consult surveyors to quantify excavation geometry. This traditional approach to collect field information is labor intensive and leads to an increase in cost. In addition, collected data is not provided in a practical time frame, due to the laborious data collection process. A semi-automated system, capable of gathering accurate as-built data in a practical time frame would be highly beneficial for engineering evaluation and control of construction. Engineers, owners, construction managers and contractors would also gain advantage with such a system. Accurate as built data provides useful information to construction management. For example, calculating the exact volume of excavated material could solve major pay item issues. In general, the collection of accurate as-built field data provides many advantages to the lifecycle of a construction project. This paper introduces two novel technologies; digital close range photogrammetry and 3D laser scanning, which enable capturing of the excavation and construction processes accurately and in a useful time frame. 2 INTEGRATED APPROACH FOR CONTROL OF MOVEMENTS DURING EXCAVATION The use of advanced technologies for monitoring excavation progress is part of a larger ongoing research project to develop integrated tools for control of

337 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-42.qxd

2/18/04

1:51 PM

Page 338

deformations due to excavations. The project is a joint effort between the University of Illinois at UrbanaChampaign and Northwestern University. It aims to develop tools to integrate field data collection with numerical simulation to predict and update excavation performance. A component of the research project focuses on developing numerical methods to predict wall deformations and surface settlements, which occur during deep excavations in urban areas. Accurate collected excavation data would enhance prediction models tremendously. Hashash (2002) presents the usage of field data to be incorporated into numerical simulations to extract soil behavior (Figure 1). Using field data to update design data will improve construction processes. Designers and construction managers can respond to changes immediately and can adjust planning at early stages of a project accordingly. Hashash (2002) proposes a new method for simulation of geotechnical problems through the use of direct field measurements with an application to deep braced excavations. In this approach the soil constitutive model “learns” the soil behavior using measured deformations around a deep excavation. The soil constitutive model simulation capability is enhanced using new or additional data. For a typical braced excavation problem the material constitutive model is initially trained to reproduce known behavior of the in situ soil obtained from typical laboratory tests. Field measurements of excavation response such as lateral wall deformations and vertical surface movements from the excavation stages are then used in an autoprogressive training algorithm to extract additional information about the soil response. The newly developed model can be used to make a forward estimate of further excavation stages or for excavations in similar ground conditions. The proposed method provides the user Engineering Design Objectives: -economic underground space, -minimum impact on existing facilities

Update simulation model of future ground deformations Now: ad hoc, non-systematic Proposed: Intelligent update

with unprecedented flexibility to integrate and learn from field observations. A critical yet often ignored aspect of field observations is the availability of an accurate record of construction staging. This information is normally available at the construction field office. It is perishable as most of that information is lost within the project files once construction progresses to later excavation stages. In order to measure and document the status of an excavation, a surveying crew has to be deployed to the actual site and measure characteristic points in order to model the entire excavation site. This method requires trained surveying personnel and is time consuming. The new technologies, introduced in the following sections, are used to collect accurate field construction progress data. 3 DIGITAL CLOSE RANGE PHOTOGRAMMETRY Digital close-range photogrammetry has various potential applications in construction. They include accurate as-built dimensional data for remodeling, quality control of building dimensions, and monitoring distortion and displacement of structures. Research conducted at the University of California at Berkeley, as well as at the Technical University of Berlin, already demonstrated the potential of automated photogrammetry systems for recording and document historical buildings (Debevec et al., 1996 and Wiedemann and Rodehorst, 1997). Further, digital close range photogrammetry has been used in structural tests in order to record and measure cracks in concrete during laboratory tests (Whiteman et al., 2002). Digital close-range photogrammetry (DCRP) as its term implies, uses digital images to construct threedimensional models. DCRP measurement technology used to obtain spatial information about any object in a

Simulation model of anticipated ground deformations

Adjustment of engineering design & construction activities to control deformations

Data Storage and Display Now: simple graphical display, lacks const. Info. Envision: Virtual Reality

Figure 1. Deformation cycle of a deep excavation.

338 Copyright © 2004 Taylor & Francis Group plc, London, UK

Engineering design

Construction activities Now: limited records Envision: detailed rec.

Field data collection Now: manual, non-real time Envision: sensors, wireless tech.

09096-42.qxd

2/18/04

1:51 PM

Page 339

two dimensional picture. As a form of photogrammetry, this technology derives measurements from electronic images of the object, rather than measuring the object directly. In the close-range process, the camera is positioned at a finite distance ranging from 1 to 150 meters. From multiple positions, the user is able to acquire imagery at many convergent angles both inside and outside of the object. Due to these capabilities, digital close-range photogrammetry is suitable for a large number of applications, ranging from simple manual control point measurements to automated processing. Although digital close-range photogrammetry can be a stand-alone operation, it is equally effective when integrated with other measurement and processing methods. Photogrammetry has been used in various applications from aerial survey to military intelligence. One of the most popular photogrammetry applications can be found in the fields of archeological excavations or facades of historical buildings in need of renovation. In particular, some photogrammetry systems in architecture generate three-dimensional models of buildings including texture mapping. Close range photogrammetry is useful for monitoring excavation projects for many reasons. One of them is the ease of use. Images can be obtained with, what photogrammetrists call “amateur photography”. That means any nonprofessional could take photographs, which can be processed through photogrammetry software in order to gain valuable measurements. Those pictures might not be ideally exposed, well focused, nor generally in an optimal condition for photogrammetric analysis. But it will be possible to extract measurements. However, applying DCRP to excavation projects require some preplanning that means controlling various factors, in order to increase accuracy. The most important factors are exposure and resolution. Points of interest have to be plainly visible and background

and object should be in a clear contrast to each other. In the process of deploying photogrammetric equipment, it is also necessary to pay attention to some physical constraints. Positioning the cameras has to allow that object points are visible from at least two images. As an excavation project progresses, the determination of camera locations can become more and more difficult. The geometric consideration here is the angle that is enclosed by at least two cameras. Highest overall accuracy is guaranteed when the camera axes intersect at a 90° angle. Thus, the closer the intersection of the camera axes approach to 90°, the more accurate. The process of extracting accurate measurements off a two-dimensional picture is nowadays software-aided. Before the cameras are deployed to a construction site, they must be calibrated to acquire camera parameters such as lens distortion, principle point and resolution. Pictures taken from an excavation site have to be uploaded into a software program. The user has to determine (pick) the reference points (targets), which ought to be measured, in fact in every single picture. After marking all the points, they have to be crossreferenced in order to determine their exact location in space. Special targets with unique shape and colors can be used. Using targets in the field will not only increase accuracy, but will also ease the process of marking and referencing the points. Based on shape and colors, the software is capable of detecting, marking and referencing the points automatically. However, after determining the points to be measured in the image space, they are processed through an algorithm, which determines the exact position in a three-dimensional object space. A 3D viewer then allows to start modeling the measured objects. Adding lines, edges and surfaces, a 3D model of the object or scene can be generated. Figure 2 shows a 3D model of a scaled (1:50) excavation model, generated using Photomodeler software

Figure 2. 3D model after processing through PM 5.

339 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-42.qxd

2/18/04

1:51 PM

Page 340

program. The model is basically a box filled with earth to simulate an excavation site. A total of 40 targets were placed to determine the edges as well as to define the contour of 10 m by 10 m grids. Digital images were taken from four different locations, capturing the entire model. After marking and referencing the four points, which describe the edges of the model, the targets of the grid could be marked and referenced automatically. Subsequent to processing the entire set of points, 3D models can be generated. Knowing the 3-D location of the points, a model could be constructed adding lines and surfaces. The latest development of photogrammetry even allows texture mapping. Further it contains the capability to calculate volume. Having created surfaces successfully, a plane can be placed horizontally on top of the model. The program then integrates the space enclosed by the plane and the created surface. Hence, accurate volume measurements can be generated. The model demonstrates that accurate three-dimensional data can be extracted from twodimensional images. However, using digital photogrammetry requires the consideration of some vital factors as described earlier. In case more points have to be measured, more targets have to be placed on site. Presently, the research team are exploring the deployment of webcams with photogrammetry on a construction site, located on the campus at Northwestern University. Past experience through a Webcam, installed by Finno (2003) on a construction site in downtown Chicago, to monitor construction progress, showed the capability of collecting images remotely through the internet. Webcams deliver digital images in set time intervals to any place around the globe. Using DCRP, these photos can then be processed to generate volume or profiles of an excavation site. Combining webcams with digital close range photogrammetry (WDCRP) permits the acquisition of spatial data related to excavation geometry, remotely and in a timely manner. Various issues associated with such a system are currently under investigation to open new possibilities for monitoring excavation and various potential applications in construction. 4 3D LASER SCANNING Recently, developments in 3D laser scanning made several laser-scanning systems available commercially. Researchers around the globe are working on implementing three-dimensional laser-scanning technologies for the construction industry (Wunderlich, 2002 and Kern, 2002). Successful volume calculation using 3D laser scanning has been conducted in a lab experiment at the National Institute of Standards and Technology (Stone et al., 2000). Three-dimensional laser-scanning systems provide new opportunities for

collecting accurate construction field data in the Architecture/Engineering/Construction industry. The scanned data produce accurate as-built information for project control and analysis. This accurate as-built excavation data opens new possibilities for analysis and feedback on ground motion and advanced geotechnical analysis. Many more researchers (Hwang et al., 2003) are exploring new applications for real-time 3D field data collection. Three-dimensional laser scanning is a new technology in the field of surveying. In a relatively short amount of time. Lasers (Light Amplification by Stimulating Emission of Radiation) have been used for years. The probably most known application nowadays is the barcode laser. Grocery stores, while checking out products, commonly use the technology of laser scanners to collect data. Recent developments in the field of three-dimensional laser scanning utilized new equipment, which enhances field data collection tremendously. Since the turn of the millennium, 3D laser scanners open new possibilities in capturing valuable 3D data from various objects. The basic technology of laser scanning (measuring with light) is called LIDAR (Light Detection and Ranging). Similar to RADAR (Radio Detection and Ranging), it uses light to measure range and distance precisely. Basically, a laser-scanning device consists of an emitting diode that produces a light source at a very specific frequency. A mirror, located in the casing, directs the laser beam horizontally and vertically towards the target, covering an area of 40° by 40°. Applying the principles of pulse time of flight method, distance can be determined by the transit time of the reflected laser. The result can be seen on the computer screen immediately. A so-called point cloud is generated during the process of scanning an object. These point clouds can then be processed into accurate 3D models, using specific CAD and rendering software applications. The accuracy of three-dimensional laser scanners primarily depends on three factors; range, point distance and diameter of the laser beam. The latest developments of three-dimensional laser scanners cover a range up to 500 feet, with an accuracy of 6 mm (the diameter of the laser beam). A distance of 1.2 mm between each laser beam allows a high resolution, while the relatively small diameter of the laser beam enables to capture small objects. Conducting scans appears to be relatively simple. The scanner is mounted on a tripod, which can be placed at any desired location. No leveling of the tripod or scanning device is required for scanning multiple sections. Once the scanner is in place and connected to a laptop computer, it scans at a rate of 1 column a second. Since one column contains a maximum of 1,000 points, the scanner is able to capture

340 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-42.qxd

2/18/04

1:51 PM

Page 341

Figure 3. Single scan with Cyrax 2500.

60,000 points a minute. A matrix of 1,000 columns and 1,000 rows provides the highest resolution. It is to be understood that a single scan does not allow generating a three-dimensional model of the complete object. The field of view is restricted to 40° vertically by 40° horizontally. In order to capture the entire object, several scans need to be done and registered (“stitched”) together, as post processing the collected data. While scanning the scene, each scan has to overlap with the adjacent one. The registration of the different scans requires an overlap of approximately 8°. During the registration process of each of the different scans, the algorithm needs three points to position the scans properly. These points ought to be identical points of two point clouds, placed in the overlapping zone. Having marked the starting points, the software references all scanned points relatively to each other. The result, a complete point cloud model, appears in a local coordinate system. Figure 3 shows the result of a single scan at a construction site. 5 RESULTS AND LESSONS LEARNED A field trial of 3D laser scanning was conducted at the Lurie Research Center excavation project in downtown Chicago. A Cyrax 2500 laser scanner from Leicageosystems Inc. was used to carry out the field test, whereby the resolution of each scan was set to its maximum (1,000 by 1,000 points). A total of 16 scans from thee differenct locations were conducted during the excavation process. The entire project site, measured roughly 255 ft 225 ft 25 ft (length width height) took approximately four hours to

scan. During the process of scanning, captured points became visible on a notebook, allowing to adjust scans accordingly. After scanning the site, the collected data was processed through a software program, called Cyclone from Cyra Inc. The software program allows users to register (“stitch”) each scan together to generate a single, to scale 3D point cloud model of the site as shown in Figure 4. The field test focused on measuring the volume of excavated material and generating profiles at any given point. However, the point cloud generated also contained surrounding buildings as well as earthmoving equipment. Various functions of the software program allow to “clean” redundant data. Hence, surrounded buildings and obstacles located inside the excavation were deleted, which resulted in a point cloud describing precisely the surface of the terrain. At this point the collected field data is already useful in many ways. Point clouds can be exported to various CAD applications and rendering applications. These points provide accurate location in a threedimensional space, for modelling and rendering. However, focusing on measuring volume of excavated material accurately, a mesh had to be generated. Cyclone supports a so-called TIN mesh (Triangulated Irregular Network). A TIN mesh consists of contiguous triangles in which no two vertices sharing the same X and Y coordinates, and no two triangles overlap along the vertical axis. That means, data holes created while cleaning the point clouds are now filled and the mesh represents the surface topography from the scanned and processed point clouds. This mesh provides many advantages in terms of calculating volume. A horizontal plane (defined grid) can be set at a

341 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-42.qxd

2/18/04

1:51 PM

Page 342

Figure 4. 16 Scans registered.

The field test proved the feasibility of using 3D laser scanning to capture as-built construction staging in a deep excavation environment in a timely manner. The fact that the research team did not interfere with ongoing construction processes shows the potential 3D laser scanning, which opens new ways in collecting excavation field data for enhancing engineering analysis, project management, and control. 6 CONCLUSION AND FUTURE DIRECTIONS

Figure 5. Excavation profile.

determined height to compute the total excavated volume. The space enclosed by the surface of the terrain and the plane describes the volume of the excavated earth. The exact volume can be calculated by integrating the distance between the plane and the terrain surface instantly. Furthermore, cross-section can be developed anywhere to show the elevation profile as illustrated in Figure 5.

Digital close range photogrammetry and 3D laser scanning open new possibilities for excavation monitoring. The technologies enable the recording of as-built construction data precisely and in a practical time frame. These technologies are expected to make a tremendous impact on the construction of underground space. Although pricy, the technologies have great potential in the future as prices decrease, while functionalities continue to increase. In the future, laser scanning as well as photogrammetry systems could be installed on a construction site temporarily, delivering data to any desirable location around the globe. Integration of collected field data into engineering

342 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-42.qxd

2/18/04

1:51 PM

Page 343

analysis and project management and control processes would highly benefit contractors as well as designers/engineers. Collected as-built 3D terrain models can be automatically imported into numerical simulation applications to improve the fidelity of model simulation. Accurate 3D as-built data can also update as-planned models. Schedules, for example, can be adjusted consequently. Construction personnel will be able to react to changes and differences quickly and efficiently to meet project requirements. There are, however, challenges in collecting accurate as-built data. These challenges, such as target selections, accuracies, and missing data behind objects, have to be field proven. The research team plans to conduct more field trials deploying photogrammetry and laser scanning technologies. An upcoming construction site on the Northwestern University campus will provide optimal field-testing conditions to conduct further field trials. ACKNOWLEDGEMENTS This material is based upon work supported by the National Science Foundation under Grant No. CMS 02-19123 under program director Dr. R. Fragaszy. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation. The authors would also like to acknowledge our research collaborator Dr. Richard Finno and his assistance in providing access to the excavation site for conducting the field trial.

REFERENCES Cyra website, www.cyra.com Debevec, P.E., Taylor, C.J. and Malik, J., 1996. Modeling and Rendering Architecture from Photographs: A Hybrid Geometry- and Image-Based Approach, SIGGRAPH 96. Computer Graphics Proceedings, Annual Conference Series. Addision Wesley, New Orleans, Louisiana. Finno, R., 2003. Personal Communication, Professor, Northwestern University, Evanston, Illinois. Hashash, Y.M.A., 2002. Systematic update of a deep excavation model using field performance data. Computers and Geotechnics, Submitted for review. Hwang, S., Trupp, T. and Liu, L., 2003. Needs and Trends of ITbased Construction Field Data Collection, IT Symposium, ASCE Annual Conference & Exposition, Nashville, TN. Kern, F., 2002. Precise Determination of Volume with Terestical 3D-Laserscanner. Geodesy for Geotechnical and Structural Engineering II: 531–534. Photomodeler website, www.photomodeler.com Stone, W.C., Cheok, G. and Lipman, R., 2000. Automated Earthmoving Status Determination, ASCE Conference on Robotics for Challenging Environments, Albuquerque, NM. Whiteman, T., Lichti, D.D. and Chandler, I., 2002. Measurement of Deflections in Concrete Beams by Close Range Digital Photogrammetry, Symposium on Geospatial Theory, Processing and Applications, Ottawa. Wiedemann, A. and Rodehorst, v., 1997. Towards Automation in Architectural Photogrammetry using Digital Image Processing, CIPA Int. Symposium ‘97, Photogrammetry in Architecture, Archaeology and Urban Conservation, Goeteborg, Sweden, pp. 209–214. Wunderlich, T., 2002. Vektorielle Abtastung mir Laser Scannern – Das Potential raeumlicher Punktwolken. Beton- und Stahlbeton 97, Heft 11: 557–563.

343 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-43.qxd

2/18/04

1:52 PM

Page 345

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Durability and corrosion protection of support systems in soil and rock tunnels M.R. Jafari, V. Nasri & M. Wone STV Incorporated, New York, USA

ABSTRACT: Reinforced concrete and anchors are the two major elements being used widely in underground construction activities for both hard and soft grounds. The former is used either as pre-cast liners for bored tunnels or as the cast in place concrete (or shotcrete) for tunnel support. The later is used for ground reinforcement in various forms to help the rock mass support itself and mobilize the inherent strength of the rock. Various design aspects of reinforced concrete and anchors have been discussed by numerous authors, however, very few attempted to discuss the durability and corrosion aspects of these elements. This paper will discuss the fundamentals of corrosion process and the prevention techniques for both newly build and old structures. As a case study, a summary of corrosion study for a rock tunnel project in New York is presented.

2 MECHANISM OF CORROSION OF STEEL IN CONCRETE

1 INTRODUCTION The corrosion of steel reinforcement in concrete construction is a worldwide problem, which can, on the small scale, cause disfigurement and, on the large scale, lead to structural catastrophe. The issue of reinforcement corrosion in reinforced concrete structures gained national attention in US shortly after the advent of the interstate highway system due to severe deterioration of structures less than 20 years of age and before the major highway development period between 1950’s and 1960’s was complete. The corrosion of existing concrete structures is estimated to be approximately $8.3 billion per year for highway bridges alone, and there are many other types of concrete structures that are also exposed to corrosion as shown in Figure 1. In recent years, more emphasis has been given to creating more durable and cost effective concrete structures. This is due to the increase cost of repair and maintenance, which may be similar to the cost of building a new structure in many cases. Corrosion of metals is a spontaneous processof metals returning to their natural state by oxidationreduction (producing/consuming electron) reaction. Corrosion is caused by a flow of electricity from one metal to another metal or recipient of some kind, or from one part of the surface of one piece of metal to another part of the same metal where conditions permit the flow of electricity.

Corrosion of steel in concrete can occur if the concrete is not of adequate quality, the structure was not properly designed for the service environment, or the environment was not as anticipated or changes during the service life of the concrete. The influences of corrosion of steel reinforced concrete can be summarized as: 1) Corrosion influence on steel reinforcement The corrosion results in the loss in steel section, and consequently a significant decrease in the mechanical resistance of structural members. 2) Corrosion influence on concrete Corrosion of steel reinforcement causes rust stains, cracking, delamination, and spalling. 3) Corrosion influence on steel-concrete bond Concrete and steel have similar coefficient of thermal expansion, thus allowing the two materials to expand and contact together during temperature changes. In good quality steel reinforced concrete, the sufficient bond between concrete and steel guarantees the transfer of stress through shear. However, the rust formed due to the corrosion causes the stress unable to be transferred efficiently. For steel reinforced concrete these influences usually appear together. The combined influences not only

345 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-43.qxd

2/18/04

1:52 PM

Page 346

film to break down. This can be result of one or the combination of following causes. 2.1

Carbonation-induced corrosion

Carbonation of concrete is a process by which carbon dioxide from the air slowly penetrates the concrete, reacts with the soluble alkaline calcium hydroxide and other cement hydrates in concrete to form insoluble carbonates, such as carbonate: (1) Carbonated concrete losses its passivating alkalinity around the rebar. The concrete around the rebar may become very hard and then eventually decompose and crumble. The pH value of concrete will drop over time and once is below 9.5, the passivating oxide film will start to break down (St. John et al 1998). Carbonation is affected by humidity, atmospheric pollutants, and the permeability of the concrete. 2.2 Figure 1. Existing tunnel in Manhattan.

result in the façade problem, but also the progressive weakening of load bearing capacity over a period of time, or even sudden failures. Two kinds of corrosion cells along or between steel rebars with the concrete pore fluids as electrolyte can be developed, ranging from microns to meters in size scale, depending on the particular circumstances. These are: (a) the micro-cell, in which the anode is immediately adjacent to the cathode, and results in localized pitting type corrosion; and (b) the macrocell, in which anode and cathode is separated by some distance. Macro-cell corrosion can be very aggressive and severe and is responsible for much of the severe structural damage experienced on bridges and other structures. Both pitting and general type corrosions might be seen as a result of macro-cell corrosion (Holm 1987). Concrete normally provides reinforcing steel with excellent corrosion protection. The high alkaline environment in concrete results in the formation of a tightly adhering film, which passivates the steel and protects it from corrosion. Concrete with low permeability could minimize the penetration of corrosioninducing substances and also increases the electrical resistivity of concrete, which impedes the flow of electro-chemical corrosion currents. For corrosion to be initiated on rebars, it is necessary for the protective

Penetration of chloride ions into concrete may cause some of the chloride become bound into the cement paste in the form of calcium chloroaluminate hydrate (3CaO.Al2O3.CaCl2.10H2O) dependent on the amount of aluminate phase present in the cement, while the remainder is present as free chloride ions in the pore solution (St. John et al 1998). These free chloride ions may result in a shift of electrical potential, and non-uniform penetration of chloride ions leads to the formation of corrosion cells. The concentration of chloride ions can cause a significant decrease in resistance of concrete which acts as electrolyte in corrosion process. More importantly, chloride ions can breakdown the passive film on steel surface. The exact mechanism of the breakdown of the protective passive film is still unclear, however, three modern theories such as oxide film theory, adsorption theory and transitory complex theory have been reported to explain chloride-induced corrosion (Kitowski and Wheat 1997, Fraczek 1987 and ACI 222R-96). Having identified in general terms the mechanism leading up to corrosion of steel, the influencing factors can be identified as follows: 1) The concentration of chlorides in the solution to which the concrete is exposed. In conditions in which the concrete is permanently submerged and concrete is saturated, this will determine the concentration gradient through the pore solution. 2) The level of chloride at the surface of the concrete, which defines the concentration gradient when the

346 Copyright © 2004 Taylor & Francis Group plc, London, UK

Chloride induced corrosion

09096-43.qxd

2/18/04

1:52 PM

Page 347

concrete is not permanently submerged. This is determined by: a) The salt exposure, defined by general environment (macro-climate), specific location and concentrations of the solutions to which the concrete is exposed (meso-climate) b) The content and type of cement with which the chlorides react and bind c) Effects of chloride condensation 3) The resistance of the concrete in the cover zone to chloride penetration as affected by: a) The depth of cover b) The concrete mix design and, in particular, the grade of concrete defined by water/cement ratio or compressive strength c) The cement type as it affects the pore structure and the degree of chloride binding d) The use of admixtures, either water reducing or waterproofing e) The degree of curing f) The moisture content of the concrete g) The temperature 4) The ratio of free/total chloride in the concrete and chloride content in the pore water. These are influenced by: a) The fineness of the cement b) The chemical composition of the cement and the use of the mineral additives such as PFA, GGBS or SF. c) The nature and concentration of the exposure solution d) The alkalinity of the cementing system and the presence of sulphates e) The temperature f) Carbonation 5) The threshold level of chloride at which corrosion is initiated. This is normally expressed as a ratio of chloride ions either to the hydroxide ion content of the pore water or to the cement content of the mix or to the concrete weight. The critical ratio may be influenced by: a) The pH of the pore solution b) The temperature c) The availability of oxygen which influences repassivation and initial steel potential d) The condition of the surface of the steel when embedded in the concrete Table 1 presents the recommended permissible limits for chloride ion (Cl) in concrete prior to service exposure, expressed as a percent by weight of cement by the American Concrete Institute (1987). 2.3

Other contaminants

Sulphates present in the soil will penetrate the underground concrete and reduce alkalinity. In the case of

Table 1. Permissible limits for chloride ion content in concrete. Type of concrete

Percentage

Prestressed concrete Reinforced concrete in a moist environment and exposed to chloride Reinforced concrete in a moist environment, but not exposed to chloride For above-ground structure where concrete is dry

0.06 0.1

No limit

chemical industries, concrete members are exposed to toxic compounds of sulphates, phosphates and nitrates. They penetrate into concrete and reduce the alkalinity of the protective cover around the steel and induce corrosion. Certain compounds of sodium and of magnesium are aggressive to concrete and are responsible for the cationic action. This is the replacement of the lime in the cement by another cation, which is effected by the ions of magnesium and ammonia, resulting in a weakening of the concrete cover. 3 CONCRETE SERVICE LIFE PREDICTION The unexpected, premature deteriorations in reinforced concrete structures have generated several theories and models to predict concrete service life, as a function of different sources of aggressive agents and of different rate-determining parameters. Service life prediction is a complex matter in which both technical topics and economical consequences are involved. This concept has been expressed in different ways, however, generally being the period in which a structure fulfills its structural requirements. The simplified model of the corrosion process Tuutti (1982), shown in Figure 2(a) assumes that once corrosion is initiated then the rate of corrosion will be constant. However, recent research (Bamforth 1994) has indicated that a non-linear relationship exists between the corrosion rate and the chloride content at the rebar depth, suggesting that the rate of corrosion will progressively increase as the level chloride builds up at the surface of the steel as shown in Figure 2(b). The residual lifetime of the structure depends on the rate of deterioration. An unacceptable degree of corrosion, not quantified by Tuutti or Bamforth, is reached when a repair should be undertaken. The corrosion initiation time at which the steel becomes active primarily measures the life period of the structure. The Propagation period, at which the damage becomes visible, was found to vary from 1 to 10 years for cover thickness varying from 25 to 50 mm.

347 Copyright © 2004 Taylor & Francis Group plc, London, UK

0.15

09096-43.qxd

2/18/04

1:52 PM

Page 348

The quantification of this deterioration period becomes of crucial importance in the assessment of damaged structures. Different laws of diffusion of chlorides and carbon dioxide have been proposed in order to calculate the time of corrosion initiation as a function of different parameters, however, the propagation period has received less attention, perhaps due to the scarce data offered by the literature on deterioration rates. In addition, the determination of an unacceptable level

of corrosion has been described more qualitatively than quantitatively. For this propagation period model, three different steps need to be set out in order to calculate residual service life of corroding structures: (1) a more accurate definition of capacity of the structure; (2) the selection of the deterioration determining parameters, that is, the parameters that need to be measured to be able to quantify the damage; and finally, (3) the transformation of the experimental data of steel corrosion rates into a form applicable to the determining parameter. 4 LEVEL OF DETERIORATION The Comité Eurointernational du Beton (CEB) (1983) has introduced a method in its Bulletin No. 162 for classifying the damage level of reinforced concrete building and the urgency of repairing or strengthening a structure after damage. Table 2 is the reproduction of the levels of deterioration (A, B, C, D, and E) as classified in the bulletin. Combining these levels with the calculation of “capacity ratio”,   R/S (R is the load-bearing capacity and S is the action effect this system or element would be required to resist according to National Codes), the residual stiffness may be approximately estimated. Thus, capacity ratio values lower than about 0.5 as shown in Table 3 (same source) would require immediate repair action. Higher values of  would allow up to 1 or 2 years before any intervention and values Table 3. Pseudo-quantitative estimation of capacity ratio for building-elements after chemical attack.

Figure 2. Tuutti and Bamforth model of service life.

Construction

R.C. Element damage level

New Old

A 0.95 0.85

B 0.8 0.7

C 0.6 0.5

D 0.35 0.25

Table 2. Damage levels of reinforced concrete. Damage levels Visual indications

A

B

C

D

E

Color changes Cracking

Rust stains some longitudinal

As in A Extensive

As in A As in C

As in A As in C

Spalling

–

As in A Several longitudinal, some on stirrups Some

Extensive

As in D

Loss in steel section

–

5%

10%

In some areas steel no more in contact with concrete 25%

Deflections

–

–

–

Possible

348 Copyright © 2004 Taylor & Francis Group plc, London, UK

Some stirrups broken, main bars buckled Apparent

09096-43.qxd

2/18/04

1:52 PM

Page 349

close to one would tolerate longer periods of time (10 to 20 years) before repairing. In Table 2 levels C and D would be those requiring a rapid intervention otherwise the structure would have run out its residual service life, whereas level B and A indicate a longer residual service life. 5 DETERMINING PARAMETER OF LOAD-CARRYING CAPACITY LOSS In Table 2 five parameters (color changes, cracking, spalling, loss of steel section, and deflections) define the level of deterioration. Among these, only crackingspalling and steel section loss will be considered for the discussion. Cracks running parallel to the rebar are the common external sign of steel corrosion. However, during the reinforced concrete corrosion process, the oxide generated may either crack the cover or may diffuse through the pore network producing brown spots in the concrete surface especially in very wet concrete. For this reason many authors prefer to work with reduction of the bar diameter or bar section (attack penetration). Attempts have been made to calculate the stress needed for causing spalling the concrete cover by generation of the oxides and, therefore, to design the bar diameter/cover ratio in order to avoid cracking if corrosion develops. However, since cracking can be suppressed in very wet concrete, although still corrosion is high. This may suggests that the cracking of the cover might not be a reliable indication of the level of unacceptable deterioration. Therefore, the reduction of the bar diameter or bar section (attack penetration) was chosen as the rate determining parameter because in both cases (cracking or diffusion of the oxides through the pores), this reduction occurs as a consequence of the metal loss. It should be noted that the reduction of the bar diameter needs to be either generalized or to take place in the critical zones of the structure in order to assume it affects the load carrying capacity. Based on previously suggested levels of deterioration, reduction in bar section between 10 to 25% in the critical zones of the structure indicates the depletion of its residual service life, whereas reduction of up to 5% (even with cracking and spalling) will indicate an early stage of deterioration with remaining service life depending on the real corrosion rate of the steel.

produced by calculating the penetration attack in millimeters per year for bars of 10 and 15 mm diameters from the values of the possible corrosion rates. The corrosion intensity values, icorr, were transformed into percentage of reduction in bar diameter or bar section (1 A/cm2 is equivalent to about 11 m/ year). The corrosion intensity value, icorr, can be evaluated using the electrochemical technique known as polarization resistance. Assuming the corrosion rate remains constant, the prediction of the number of years to reach a deterioration level (5, 10, or 25%) is easily attained. For instance, if the corrosion rate is 5 A/cm2 (0.05 mm/year) a 25% reduction in bar section is reached in 12.5 and 25 years after depassivation for a bar of 10 and 20 mm diameters respectively. Hence, the remaining service life would be double for that of 20 mm. This may suggest that in a corroding structure, a few bars of large diameter seem safer than numerous thinner ones. 7 CORROSION CONTROL IN CONCRETE Given the importance of the costs associated with the corrosion of infrastructures, it is extremely important that all possible methods applicable to controlling corrosion in existing concrete structures be developed in order to prevent premature deterioration of these

6 RESIDUAL SERVICE LIFE IN CORRODING CONCRETE STRUCTURES Figure 3 represents the attempt to predict the residual service life of concrete shown in Figure 2 by a simple and practical methodology. This figure has been

Figure 3. Example of residual service life for bars of 10 and 20 mm diameters.

349 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-43.qxd

2/18/04

1:52 PM

Page 350

structures. Equally important is developing methods to avoid this costly corrosion problem in all new concrete structures to be constructed in the future. Accordingly, the control methods can be divided into two major areas:

• •

Corrosion control in new concrete construction Corrosion control for rehabilitation of existing concrete structures

Many techniques and products have been identified to stop or retard active corrosion of reinforcing steel or to prevent corrosion from initiating on new concrete structures. Even though some of these techniques and products are in experimental or development stages, most of them have gained acceptance in concrete repair. According to their working principles, they can be categorized as (1) mechanical, (2) electromechanical methods. 7.1

Mechanical techniques

Based on the corrosion mechanism two important parameters affecting the corrosion of reinforcing steel in concrete are: a) The environmental condition b) Concrete quality The mechanical techniques for retarding and possibly preventing oxygen, chloride and water migration can be applied in two ways to incorporate materials within the concrete acting as physical barriers to the access of oxygen, chloride, and water. These methods are: 1) The concrete admixtures 2) The protective coatings Concrete admixtures are those ingredients in concrete other than Portland cement, water and aggregates that are added to the mixture before or during mixing. Admixtures can be classified by function in (Kosmatka and Panarese 1998): i) ii) iii) iv) v) vi) vii)

Air entraining admixtures Water reducing admixtures Retarding admixtures Accelerating admixtures Superplasticizers Fine mineral admixtures Miscellaneous admixtures, such as workability, bonding, damp proofing, permeability reducing, grouting, gas forming, coloring, and corrosion inhibiting.

Concrete quality is very important for corrosion prevention or control. A good quality concrete shall have low permeability, high bond strength and good durability. Lower water/cement ratio and/or dense concrete have a very low permeability resulting in a good performance in environments of freeze/thaw, wet/dry,

de-icers, sulfate, and alkali reactivity. The addition of admixtures could enhance the concrete quality for corrosion prevention/control without increasing the cost. The disadvantage of using admixtures is that this option is effective for new structures that have not been contaminated sufficiently to initiate active corrosion and can be categorized as corrosion preventive technique rather than corrosion control. Corrosion inhibitors are materials to reduce or block metal loss due to corrosion attack. They can be classified into three general groups according to their working mechanism as: (a) anodic inhibitors, (b) cathodic inhibitors, and (c) adsorption inhibitors. Anodic inhibitors are those chemicals that function by stifling the reaction at the anode by reacting with the ferrous ions from the corroding steel to increase the polarization of the anode and forming an extremely insoluble, tightly adhering thin passive film or salt layers right on the steel surface. This could prevent further contact between corrosive solution and the metal surface. Effectiveness of anodic inhibitors requires that its concentration be maintained above a critical level. If the content of inhibitors is or gradually becomes too low to cover all surface of steel acting as an anode then as a result the dangerous situation of small anode and large cathode combination will be obtained resulting in an increase in the corrosion rate and in extreme case leading to pitting. Cathodic inhibitors affect cathodic reactions. Some of them, such as salts of zinc, magnesium and calcium, can react with the hydroxyl ions to precipitate insoluble compounds on the cathode site and prevent access of oxygen; others such as arsenic, bismuth, antimony and some organic compounds can react with hydroxyl ions from a layer of adsorbed hydrogen on the cathode surface. Cathodic inhibitors are generally less effective than anodic inhibitors. The precipitate formed is not totally insoluble and it does not tightly bond to steel surface. In addition, the mechanism of corrosion prevention is indirect, whereas the anodic inhibitors directly prevent the steel from corroding. However, cathodic inhibitors are safe because the active cathode area is reduced even using very small amount of the inhibitor. Absorption inhibitors function by being adsorbed all over the steel surface and double acting, i.e., simultaneously retard both the anode and cathode process. Adsorption inhibitors can limit the diffusion of oxygen to the steel surface, trap the iron ions on the steel surface, and reduce the rate of dissolution (Pyc 1997). It is found that a combination of two inhibitors may result in a significant effectiveness and also eliminate the risk of pitting for small inhibitor concentrations. The advantage of using corrosion inhibitor is that inhibitors are distributed throughout the concrete and protect all the steel. The disadvantages of using inhibitors are: (1) the concentration amount of anodic

350 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-43.qxd

2/18/04

1:52 PM

Page 351

inhibitors is critical, (2) some inhibitors may have adverse effects on the concrete properties, e.g., chromate and sodium nitrite can decrease the concrete strength, (3) some are toxic, e.g., chromate, and (4) attention must be paid in treatment with inhibitors since unless the concentration around and along the rebars is reasonably uniform, corrosion could be activated rather than inhibited due to formation of dissimilar electrolyte corrosion cells. Protective coatings can be categorized into three groups: organic (epoxy) coating, inorganic coating or metallic clad, and solid metallic coating. The most common and widely used protective coating for the steel protection in concrete is the epoxy coating. Belonging to the poly addition plastic family and being thermosetting plastics, epoxy resins are reported to have good long-term durability in concrete and be resistant to solvents, chemicals, and water. Epoxy resins also have desirable mechanical properties such as high ductility, small shrinkage upon polymerization and good heat resistance. By appropriate quality control during coating application and subsequent handling, a defect free epoxy coating with adequate thickness can act effectively as a barrier that prevents ingression of chloride ions, water, and oxygen. American Concrete Institute (1996) summarized the methods of excluding external sources of chloride ions from concrete as below: a) Waterproof membranes Since external sources of chloride ions are waterborne, a barrier to water will also act as a barrier to any dissolved chloride ions. The most common types of the waterproof membranes are sheet systems and the liquid applied materials. The performed sheets are manufactured under factory conditions but are often difficult to install, usually require adhesive, and are highly vulnerable to the quality of the workmanship at critical locations. Although it is more difficult to control the quality of the liquid applied systems, they are easier to apply and tend to be less expensive. Field performance of waterproof system has been found to depend not only on type of waterproofing material used, but also on the quality of workmanship, weather conditions at the time of installation, design details, and the service environment. The major problem encountered in applying waterproofing membranes is blistering caused by expansion of entrapped gases, solvents, and moisture in the concrete after application of the membrane. b) Polymer impregnation Polymer impregnation consists of filling some of the voids in hardened concrete with a low viscosity (or mixture of monomers) that polymerizes in situ to form a network within the pores. Impregnation results in markedly improved strength and durability (e.g., resistance to freeze-thaw damage and corrosion) in comparison with conventional concrete. Laboratory

studies conducted on polymer-impregnated concrete (PIC) indicated that this material is strong, durable, and almost impermeable. The physical properties of PIC are determined by the extent to which the ideal processing conditions are compromised. Most field applications of PIC have been aimed toward producing only a surface polymer impregnation, usually to a depth of about 25 mm. The disadvantage of PIC is the relatively high cost, as the polymer is more expensive than cement and the production process is more complicated. The principal deficiency identified to date has been the tendency of the concrete to crack during heat treatment. c) Polymer concrete overlays Polymer concrete overlays consist of aggregate in a polymer binder, e.g., epoxy or polyester, and fine aggregates.

• • •

Latex modified concrete (LMC) is a conventional Portland cement concrete in which an admixture of styrene butadiene latex particles suspended in water is used to place a portion of the mixing water. The water of suspension in the emulsion hydrates, the cement and the polymer provide supplementary binding properties to produce a concrete having a low watercement ratio, good durability, good binding characteristics, and high degree of resistance to penetration by chloride ions. The performance characteristics of latex modified concrete overlays have been satisfactory, though extensive cracking and some debonding have been reported, specially in overlays 20 mm thick that were not applied at the time of the original structure construction. 7.2

Electrochemical techniques

Compared to the aforementioned mechanical techniques, the electrochemical corrosion protection techniques are based on the electrochemical principle of corrosion. There are mainly four methods. 7.2.1 Cathodic protection Cathodic protection (CP) prevents corrosion by reversing the electrochemical process that causes corrosion. An electrode is inserted in the concrete near the reinforcement and connected to the positive terminal of a direct current power source. The magnetic terminal is connected to the reinforcement. The positive current supplied to the steel bar overcomes any corrosion that was previously flowing in the reinforcement. In other words cathodic protection can be described as “use corrosion to treat corrosion”. It is believed that the cathodic protection is the only rehabilitation technique that has proven to stop corrosion

351 Copyright © 2004 Taylor & Francis Group plc, London, UK

Portland cement concrete overlays Low slump concrete overlays Latex modified concrete overlays

09096-43.qxd

2/18/04

1:52 PM

Page 352

in salt contaminated bridge decks regardless of chloride content of the concrete. FHWA estimated that about $50 billion in repair costs could be saved over the next 30 years by the use of cathodic protection (Scheffey 1981). Nowadays the cathodic protection can be accomplished by two widely used methods: a) Impressed current cathodic protection system An impressed current cathodic protection (ICCP) system involves impressing a direct current between an inert anode and the steel reinforcement to be protected. Since electrons flow to the steel and thus it becomes cathode, it is protected from becoming the source of electrons, i.e., anode. A typical ICCP system requires the two basic components: external DC power source (rectifier or battery), and current distribution hardware (anode). On the whole two kinds of current distribution anode systems are most commonly used: conductive overlay/coating method and distributed-anode method (or called slotted system). b) Sacrificial anode system The protection is achieved by connecting a more active metal (with a higher natural electromotive force) such as zinc or magnesium through a metallic conductor or wire to the reinforcement to be protected. This produces a galvanic cell in which the active metal works as an anode and provides a flux of electrons to steel, which then becomes the cathode and thus is protected. The anode is progressively destroyed; hence it is called the sacrificial anode. The main advantages of CP system are: 1) The system can effectively stop on going corrosion because the chloride ion is attracted to the anode, thus reduces the chloride concentration and increase the pH value in the vicinity of the reinforcing steel. 2) It is easy to install and the operation is environmental friendly and relatively quiet. 3) It is applicable to any structural geometry. The main disadvantages of CP system are: 1) CP is the nonstructural component of structures and can only increase the dead load. 2) It needs to be periodically monitored and maintained for the life of structures. 3) If the system breaks down, loses its power or most of the sacrificial anode is consumed and the protection ceases. 4) Special contractors and inspectors are required. 7.2.2 Electrochemical chloride extraction Electrochemical chloride extraction (ECE) is essentially a process whereby chloride ions are removed from chloride-contaminated concrete through ion migration. In this system an anode embedded in electrolyte

is applied to the surface of the concrete. The anode and reinforcing steel in the concrete are connected to the two terminals of a direct current (DC) power supply so that the anode is positively charged and the rebar is negatively charged. Chloride ions being negative ions migrate toward the positive electrode, i.e., the anode. Since this is external to the concrete, the chloride ions will leave the concrete and concentrate around the anode, reducing the chloride content around the negatively charged reinforcing steel where the concrete for all practical purposes becomes free of chloride. Simultaneously, the electrolytic production of hydroxyl ions at the reinforcing steel surface results in a high pH environment around the steel. This system of corrosion protection will provide a chloride free and highly alkaline environment. 7.2.3 Cast-in galvanic anode Cast-in galvanic anode is recently developed technique consisting of anodes, which are typically incorporated into patches. The anode provides galvanic protection to the rebars directly surrounding the patch without any reduction of chloride concentration in the surrounding concrete. This technique may not be as effective as CP or ECE techniques, but shows significant improvement over patch repairs (Whitmore and Allies 1999). 7.2.4 Re-alkalization Re-alkalization is an electrochemical treatment to prevent or halt ongoing reinforcement corrosion by increasing the pH value of the carbonated and or chloride contaminated concrete to more than 10.5 (Norcure 1999), which is sufficient to maintain passivity of the reinforcement. Very similar to ECE technique, re-alkalization is performed by applying an electric field between the reinforcement in the concrete and an anode system consisting of an anode mesh. Re-alkalization has the similar advantages and disadvantages as that of ECE technique. But the treatment time is less than one week, which is much less than ECE effective time. Besides, re-alkalization can be performed under all weather. 8 DESIGN ASPECTS OF CORROSION PROTECTION OF METAL TENSIONED SYSTEM Geotechnical applications of metal-tensioned systems include ground anchors, rock bolts and soil nails. Tensioned elements of the system include bar and strand components. The steel grade and level of pressure employed in these systems are relevant to the type of corrosion problems that may occur, and prediction of service life. Soil nail systems use bar elements, but ground anchors and rock bolts may be either bar or

352 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-43.qxd

2/18/04

1:52 PM

Page 353

strand. Bar elements are available in a variety of steel grades ranging from grade 60 to 160. Strand elements are manufactured from Grade 250 and 270 high strength steel. 8.1

Service life of anchors

For corrosion protection there are two possible service lives: temporary (less or equal than 24 months) and permanent (more than 24 months). The anchor used for tunneling purpose should have minimum life span of 100 years. 8.2

Assessment of corrosivity

Corrosiveness of the environment is determined by test and/or field observations. The parameters affecting the corrosiveness of an environment are summarized as:

• •

Type of soil Soil resistivity

Table 4. The classification of soil corrosiveness. Global corrosiveness index 13 9–12 5–8 4

Soil feature

Classification

Highly corrosive Corrosive Average corrosiveness Slightly corrosive

I II III IV

• •

Moisture content pH

As a general rule ground is considered to be aggressive when the level of pH is less than 4.5 and the resistivity is below 2000 ohm-cm. However, based on the abovementioned parameters and the summation of the weighing ascribed to each parameter the global corrosiveness index of the soil can be determined. On the basis of the overall corrosiveness index obtained, soils can be classified as shown in Table 4. Table 5 presents the additional sacrificial thickness required by unprotected steel for different soil corrosion classification in relationship with the service life of the structure. Current guidance documents (PTI 1996) recommend incorporating corrosion protection measures into the design of metal-tensioned systems. Corrosion protection measures include the use of coatings, protective sheaths, passivation with grout, encapsulation and electrical isolation. PTI (1996) recommended Classes I and II for corrosion protection of permanent anchors. The corrosion protection requirements recommended by PTI is presented in Table 6. Most temporary anchors will not require anchorage or pre-stressing steel corrosion protection (PTI 1996). However, for aggressive environments or when the service life exceeds the 24 months, a class II protection shall be required as a minimum and protective cover may not be required for temporary anchors.

Table 5. The required sacrificial thickness. Combined soil corrosion index/classification  4/IV 5 to 8/III 9 to 12/II 13/I

Short-term  18 months (mm) 0 0 2

Medium-term 1.5 to  30 years (mm)

Long-term 30 to  100 years (mm)

2 4 8 Unsuitable for unprotected steel

4 8

Table 6. Corrosion protection requirement (PTI 1996). Protection requirements Class

Anchorage

Unbonded length

Tendon bond length

1) Encapsulated tendon

Trumpet Cover if exposed

Grout-filled encapsulation, or Epoxy

II) Grout protected tendon

Trumpet Cover if exposed

Grease-filled sheath, or Grout-filled sheath, or Epoxy for fully bonded anchors Grease-filled sheath, or Heat shrink sleeve

353 Copyright © 2004 Taylor & Francis Group plc, London, UK

Grout

09096-43.qxd

2/18/04

1:52 PM

Page 354

9 CORROSION STUDY FOR A ROCK TUNNEL IN MANHATTAN For the proper design of rock support elements of the Manhattan tunnels, it is important to evaluate the corrosivity of the surrounding environment. The corrosiveness of the environment was evaluated based on groundwater laboratory testing in three monitoring wells in the Manhattan tunnel area by two Environmental Laboratories (Table 7). Five groundwater parameters representing the corrosion conditions of the underground environments are: the pH value, the dissolved oxygen concentration, the chlorides content, the sulfates content, and the specific conductance or groundwater resistivity. Specific conductance data in µmhos/cm presented in the table was converted to resistivity in ohm-cm. Water resistivity from soil borings varied from 500 ohm-cm to 1,500 ohm-cm. Resistivity data is the most important factor to evaluate water aggressiveness. Lower resistivity enables corrosion reactions to occur more easily. All water samples show resistivity below 2,000 ohm-cm, which indicate a high corrosiveness. In many cases the pH value of the groundwater has a strong influence on corrosion rates. Groundwater which is highly acidic (pH  4) or highly alkaline (pH  9) usually has significant higher corrosion rates than neutral groundwater. The pH value of water samples is relatively consistent and varied from 6.2 to 6.9. These values represent predominantly neutral or slightly acidic water conditions. A significant concentration of chloride and sulfate ions makes water aggressive. Usually a concentration over 500 mg/l is considered high. All the tested samples showed a relatively low content of sulfate and chloride ions in water. One sample showed a moderate content of chloride (550 mg/l). So, based on the amount of sulfates

and chlorides, the water in the Manhattan tunnels can be considered moderately corrosive. A relatively high concentration of dissolved oxygen was found in all water samples. This concentration is about 30% higher than usual in neutral waters. The high concentration of oxygen will promote a corrosion reaction of metal. In dynamic water systems such as seepage through the tunnel walls, present concentration of oxygen can significantly increase the corrosion rate. In addition to environmental corrosion studies, a stray current survey was conducted in the site for the existing tunnels. The result of this survey showed different stray current levels on underground metallic structures with average stray current activity of 140 millivolts. Dynamic transit DC stray currents can cause severe corrosion within a very short period of time. One ampere of stray current discharging into the electrolyte (rock/soil/water) will remove twenty pounds of steel in a period of one year. If, for example, just 5 milliamperes flowed off a 1-inch diameter dowel, in one year 1.6 oz of metal would be removed. From a few milliamperes up to 20 amperes were measured in different locations during stray current survey. The evaluation of the environmental corrosivity of the Manhattan tunnels was made based on the assessment of the global corrosiveness index that is presented in Table 8.

Table 8. Assessment of global corrosiveness index. Resistivity, ohm-cm

Water pH Table 7. Groundwater chemical test results in Manhattan segment. Boring no.

Sulfates, mg/l

Test type

Units

1

2

3

Dissolved oxygen pH Specific conductance Total dissolved solids Alkalinity, total Sulfate Hardness as CaCO3 Chloride

Mg/L

10.88

9.46

9.63

mhos/cm

6.17 2000

6.75 650

6.9 750

Mg/L

1700

420

470

Mg/L Mg/L Mg/L

110 120 750

83 42 200

84 180 330

Mg/L

550

100

63

Chlorides, mg/l Dissolved oxygen, mg/l Stray current activity, mV

354 Copyright © 2004 Taylor & Francis Group plc, London, UK

700 700 to 1,000 1,000 to 1,200 1,200 to 1,500 1,500 to 2,000 2,000 0 to 2 2 to 4 4 to 8.5 8.5 1,000 500 to 1,000 500 1,000 500 to 1,000 500 10 7 to 10 5 to 7 5 300 100 to 300 20 to 100 20 Global corrosiveness index

10 8 5 2 1 0 5 3 0 3 5 3 0 5 3 0 5 3 1 0 8 6 3 0 Sum of above:

09096-43.qxd

2/18/04

1:52 PM

Page 355

10 CONCLUSION This paper presented the fundamentals of corrosion process for various support systems used for underground construction activities accompanied by techniques to prevent or protect these support systems during their life span. An example of corrosion study of a rock tunnel in Manhattan is also presented.

REFERENCES ACI committee 201, “Guide to durable concrete,” ACI Manual of Concrete Practice (1987). ACI committee 222, “Corrosion of Metals in Concrete,”. A report prepared by American Concrete Institute (1996). Comité Eurointernational du Beton (CEB), “Assessment of Concrete Structures and Design Procedures for Upgrading,” Pub. No. 162, 1983, pp. 87–90. Bamforth, P. B. (1994). “Specification of onset of concrete for the protection of reinforcement in chloride contaminated environments”. UK Corrosion and Eurocorr 94, Bournemouth, Vol. III, pp. 249–258. Fraczek, J. (1987). “A Review of Electromechanical Principles as Applied to Corrosion of Steel in a Concrete or Grout Environment”, CORROSION, CONCRETE, AND CHLORIDES-Stell Corrosion in Concrete: Cause and Restrains, pp. 18–21.

Holm, J. (1987). “Comparison of the Corrosion Potential of Calcium Chloride and a Calcium Nitrate Based NonChloride Accelerator - A Macrocell Corrosion Approach”, CORROSION, CONCRETE, AND CHLORIDES-Steel Corrosion in Concrete: Cause and Restrains, pp. 40. Kitowski, C.J. and Wheat, H.G. (1997). “Effect of Chlorides on Reinforcing Steel Exposed to Simulated Concrete Solutions”, Corrosion, Vol. 53, No. 3, pp. 216–226. Kosmatka, S. and Panarese, W.C. (1998). Design and Control of Concrete Mixtures (13th Edition), Portland Cement Association. Pyc, W.A. (1997). Performance Evaluation of EpoxyCoated Reinforcing Steel and Corrosion Inhibitors in a Simulated Concrete Water Solution, Master’s Thesis, Department of Civil Engineering, Virginia Polytechnic Institute and State University, Blacksburg, Virginia. Romanoff, M. (1989). Underground Corrosion, NACE. Scheffey, C.F. (1981). Bridge Deck Deterioration - A 1981 Perspective, FHWA Memorandum, Federal Highway Administration Office of Research. St. John, D.A., Poole, A.B. and Sims, I. (1998). Concrete Petrography - A Handbook of Investigative Techniques, John Wiley & Sons Inc., pp. 284. Tuutti, K. (1982). “Corrosion of steel in concrete”. Report prepared by Swedish Cement and Concrete Associations, Stockholm, pp. 469. Whitmore, D.W. and Allies, J. (1999). “Halting Corrosion Using Electrochemical Methods”. Materials and Construction-Exploring the connection, The Fifth ASCE Materials Engineering Congress, Cinvinnati, Ohio.

355 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-44.qxd

2/18/04

3:01 PM

Page 357

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Investigation of complex geologic conditions for the Second Avenue Subway tunnel alignment in New York City, New York C.P. Snee & M.A. Ponti, Jr GZA GeoEnvironmental of New York, USA

A.N. Shah MTA-New York City Transit, New York, USA

ABSTRACT: The proposed Second Avenue Subway Project in New York City is approximately 8 miles long, up to a depth of 120 feet beneath grade and runs the length of the eastern side of the island of Manhattan from Harlem in the north to the financial district in the south. A comprehensive subsurface investigation has been carried out that included archive searches, rock exposure mapping, test borings, in situ soil and rock testing, instrumentation, monitoring and laboratory testing over the full length of the alignment. The exploration encountered a 1.3 Ba. geological history from Precambrian to Devonian tectonites from the creation of the continental USA, systematic brittle and ductile faulting and shearing from repeated orogenies, proto and post glacial erosion and culminating in Pleistocene and Holocene weathering and deposition. The geological structures created by these events present significant challenges for design and construction of this mega-project and their investigation is a critical part of the design process. This paper presents the geological techniques that were used to investigate soil-rock-groundwater systems in order to advance the understanding of the fundamental structural geology of central Manhattan.

1 INTRODUCTION The Second Avenue Subway Project is a major capital expansion project of the New York City subway that will provide a dedicated line for the east side of Manhattan with a link to the existing subway network. The proposed alignment runs from Park Avenue and 125th Street in the north to the financial district in the south with possible extension to Brooklyn (see Fig. 1). The project is approximately 8 miles long and includes sixteen stations, of which six are to be mined caverns. In addition there are numerous multi-track tunnels, crossovers and connections that will be constructed in caverns. The nominal tunnel diameter is 22 ft and the cavern spans range from 40 ft to 100 ft. All caverns have rock cover less than their span and significant lengths of tunnel will be in mixed faces of soil and rock. A shallow alignment entirely in cut-and-cover excavation was designed in the 1970s and two sections were constructed between Chatham Square and Grand Street and between 99th Street and 120th Street. The new proposal is a deeper option to reduce the urban impact of the work although some of the existing structures will be incorporated. At a very early stage in the design process the significance of the fundamental

Figure 1. Project location (not to scale) showing the alignment and the area of geological interpretation.

357 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-44.qxd

2/18/04

3:01 PM

Page 358

geological structures were recognized as being a key to understanding the rock mass behavior. The exploration program included geological studies from microscopic to regional in addition to conventional geotechnical methods to advance this understanding. The preliminary engineering is being undertaken by a joint venture of DMJM  Harris and Arup with specialist subcontractors. GZA GeoEnvironmental is undertaking the subcontract for the site investigation and geological interpretation in association with Mueser Rutledge Consulting Engineers. 2 GEOLOGICAL INVESTIGATION The geological investigation of the site started with collection and assimilation of existing information. The earlier study for this project included a boring program but the data had limited application because the work preceded current standards for logging rock core samples, there was little attention to geological detail and the depth of exploration was insufficient. Other sources of information were of varying levels of quality and relevance for the same reasons. However, the database included more than 600 historic borings and these were sufficient to develop a preliminary conceptual geological model for the alignment. In view of these issues a new exploration program was designed to obtain data to current standards, to check, correlate and enhance existing boring data, to transform the conceptual geological model to a definitive model, and to compare this with the published geological model. Historic maps showing geomorphology, geology, land use and progressive development from early colonization of Manhattan were used to make initial interpretations of structural geology, particularly the location of major fault trends because these were postulated to be a significant influence on the natural drainage pattern of Manhattan Island before its development. This preparatory work was used to plan the exploration program with clearer focus on geological zones of importance and their relevance to design of the project. The new investigation program included: – 321 exploratory borings to sample soil and rock. – Laboratory testing of rock samples for strength, deformability and abrasivity. – In situ testing for primary and fracture permeability. – Orientation of geological fabric, discontinuities and structure. – Petrographic analysis. – Rock exposure mapping. The program also included in situ and laboratory testing to obtain design parameters for the soil and rock but these are not included in this paper. Additional

ground investigations are proposed at the time of writing including angled borings, seismic refraction, in situ stress testing and other test borings to define the model further. The primary source of geological information was rock core, although proximate excavations for deep basements and tunnels were mapped as part of the program. All core drilling was by NQ double tube wireline equipment to produce nominal 2 diameter core. The rock cores were logged in general accordance with ASTM and ISRM because these are standard practice for underground projects. The methods provide qualitative and quasi-quantitative detail of top of rock, color, strength, grain size, weathering, orientation and spacing of fabric and fractures and a generic geological name. These methods can provide sufficient information for general definition of a geological domain and statistical approaches to geotechnical characteristics for analysis and design. However, in a complex geological setting such as the Manhattan Prong it is necessary to supplement the conventional methods. The process that was developed and operated for the Second Avenue Subway project had three basic components: – Fabric and petrographic logging of rock core in a field laboratory. – Fracture and joint set orientation and classification by core logging, core scribing and borehole geophysics. – Petrographic analysis by thin-section. These methods are discussed in detail in Section 4. 3 GEOLOGICAL SETTING The geological setting of New York City has posed many challenges to the construction industry and particularly to subsurface projects. The rock types encountered ranges from Precambrian to Devonian in age. The Pleistocene glaciation has added further complications with subsequent active erosion in Holocene times. The erosion and deposition has accumulated vast glacial till, modified glacial drift, sand and gravel and glacio-lacustrine silt, clays and marshland. The underlying bedrock geology of New York City is highly complex. However, some of the observations made during the on-going subsurface investigations have revealed clarifications of the current geological knowledge of the area. The Precambrian to CambroOrdovician crystalline rocks of New York City are divided into two major units separated by Cameron’s thrust fault, a regional NE-SW trending structural feature which dips due southeast. The thrust fault extends from Connecticut through the Bronx, Manhattan, through Staten Island and further south into central and southern New Jersey. This regional feature has

358 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-44.qxd

2/18/04

3:01 PM

Page 359

been classified as a suture of the proto-american plate. The New York City Group or Manhattan Formation are found west of this major thrust fault and the rocks to the east of this fault are known as the Hartland Formation (Hutchinson River Group). The Cameron thrust faulting has affected both these units and imparted various structural features such as faults, shears and joint systems. The position of this major suture is under review and so a definitive formation name cannot be given at this time. However, the published geological map (Baskerville, 1994) shows the dominant rock formation in Manhattan to be the Hartland. The project area mainly consists of the Manhattan schist rocks, calcareous rocks of the Inwood Marble and Fordham Gneiss. The Fordham Gneiss has yet to be definitively identified by the exploration program. Manhattan schists are typically crystalline variations of essentially quartz and mica composition with quartz and feldspar rich zones, garnetiferous biotite and muscovite mica schist, quartz-hornblende-mica-garnet schists, and chlorite schists. Numerous pre and post to late thrust kinematic pegmatite intrusions of varying size have been emplaced within these schists typically along and occasionally across the foliation and along other fractures. This activity is highly noticeable in the midtown area of Manhattan Island where these intrusions have locally elevated the metamorphic grade of the schist’s and modified their textures and structures to resemble almost aplitic gneissic to granitic rocks (Fluhr, 1941). The Inwood calcareous rocks have been observed from 106th Street to 128th Street. They vary in composition from calcareous schist to dolomitic marble. The calcareous rocks are intercalated with schist beneath 125th Street between Lexington Avenue and the low-lying areas to the Harlem River in the east. The rocks of the Manhattan area have undergone multiple deformation events (D0 to D5) causing three identifiable foliation from S0 to S3 as follows. Deformation Fabric event S0 S1

D1 ductile D2 ductile

S2

D3 ductile D4 ductile

S3

D5 brittle

Strike

Type location

Unknown EW-ENE

Unknown Central Manhattan WNW-NE North and south Manhattan NNE-NE 103rd to 107th Street EW-WNW 34th to 62nd Street and 96th to 125th Street

Plate No. 1 2

The D4 event is contemporaneous to Cameron’s Thrust and the development of en-echelon faults. 4 STRUCTURAL INTERPRETATION It is very difficult to distinguish between folded, faulted and unfaulted ground using conventional methods of core logging even if the full range of ground types and rock mass conditions are intercepted by the boring. It is possible to make general interpretations of the structural geology with rock outcrops to supplement the borings but these are rare in Manhattan. In view of the significance of the structural geology to the project it was considered that methods must be developed that improve these conventional methods. The basis of the methods devised for the project was to identify telltale signs of particular types and phases of deformation and structure in the cores such as fault planes, and joint sets and where possible orient the features. 4.1

3 4 5 Plate 1. Non-crenulate fabric of quartz-mica schist with quartz and garnet augen (view length 200 mm).

359 Copyright © 2004 Taylor & Francis Group plc, London, UK

Fabric and petrographic logging

The objective of this examination was to classify the rocks by stratigraphy, genesis and deformation event. This can be particularly difficult in ancient environments with multiple superposed deformation events because evidence of the original depositional environment is all but obliterated. However, if the fabric and mineral assemblage of the rocks are examined closely the various forms of deformation such as crenulate and convoluted foliation and annealed breccia can be seen, as shown in the following plates of a selection of NQ rock cores. Plate 1 shows the common sub-parallel foliation that is characteristic of a large proportion of the schistose rocks in Manhattan. This pervasive and consistent fabric is normally associated with locations that lack major faulting and folding and may be interpreted to be close to the original S0 fabric. The orientation of the schistosity does not vary dramatically in these regions and so geological markers such as pegmatite, amphibolite and granofels can be established and used for structural interpretation. However, it is important to establish the limit of the zone, a key to which is dramatic changes in dip of foliation or transition to convoluted foliation as discussed below.

09096-44.qxd

2/18/04

3:01 PM

Page 360

Plate 2 shows the dominant and parasitic foliation of two deformation events. Plate 3 shows even more convoluted schist which is evidence of multiple superimposed events creating almost random foliation directions often associated with regions that had undergone faulting and shearing, intrusion, post-orogenic metamorphism and pegmatite formation. Plate 4 is from an area that has undergone all of the deformation events with particularly strong evidence of Cameron’s Thrust (D4 event) that juxtaposed the Inwood Marble with the Manhattan schist rocks. This created an intercalated schist-marble complex with convoluted foliation. This type of rock can be used to interpret proximity to this major regional thrust in the north of Manhattan. Annealed breccia and healed offset micro-faults as shown in amphibolitic schist in plate 5 is distinct in pegmatite and quartz vein rich regions. These features can be used to predict the proximity of major pegmatite and local faulting. Plate 6 shows a zone of extremely fractured and degraded rock. The rock has been reduced to silty sandy gravel in places and the intact pieces are friable. There is alteration in the form of secondary mica and distortion of the schistosity. These are the characteristics of faulted ground. If there is three or more borings in close proximity that intercept this feature then it is possible to estimate the orientation of the fault. However, it is often the case that there is limited corroborating information from adjacent borings because the boreholes are too widely spaced or the feature is steeply dipping. If the methods of fabric and petrographic logging described above are applied to the core samples in the area the possibility of establishing geological control to define this type of feature is increased. 4.2

Plate 2. Convoluted and slightly crenulate schistosity typical of southern Manhattan (view length 200 mm).

Plate 3. Highly convoluted and highly crenulated schist from central Manhattan (view length 200 mm).

Plate 4. Convoluted foliation in garnetiferous calcic schist from intercalated region of northern Manhattan (view length 200 mm).

Plate 5. Healed micro-offsets of brittle fractures in amphibole schist (amphibolite) with convoluted foliation and quartz veins (view length 200 mm).

Fracture logging

The conventional methods of fracture logging provide basic spacing and dip angle data but it is not possible to make direct interpretation of these data into structural groups or joint sets because the dip direction is unknown. This is a serious weakness of conventional methods when used for structural geology, yet it is critical to categorize and orientate the joint sets and distinguish the discrete features such as faults for structural interpretation. Therefore, the detailed fracture logging of the core (Plate 7a) was enhanced by imaging the borehole wall with an acoustic televiewer (Plate 7b). This technique involves measuring the travel time of an ultrasonic signal projected onto the borehole wall and returned to a receiver in a sonde that is lowered down the borehole at a fixed rate. A caliper measures the diameter of the borehole so the travel time can be converted to a sonic velocity. In general, fractures, voids and decomposed material show lower sonic

Plate 6. Highly fractured, weathered and altered quartz mica garnet schist from central Manhattan (view length 24).

Plate 7. a) Digital photograph of core and b) reconstructed core from acoustic televiewer image of borehole wall.

360 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-44.qxd

2/18/04

3:01 PM

Page 361

velocity than intact crystalline rock. The differences in velocity are converted to variations in color; the lower the velocity the darker the image. Figure 4 shows the outcome of this analysis against the actual core stick from the boring. This shows very good similarity between the two methods. The significance of the acoustic televiewer is that the image is oriented so the dip and dip direction of the features, such as fractures can be determined mathematically from the data. Combining these methods produces the engineering characterization of planes such as joints, faults, fracture zone boundaries, their spacing, dip and dip direction and provides spatial control of the data. 4.3

Coale Kill or Duffore's Mill Stream

Petrographic analysis by thin section

The examination of thin sections by petrological microscope has several geological and engineering objectives such as the proportion of hard minerals, the degree of alterations and decomposition, and the extent of mineral segregation. These are critical concept for understanding the engineering properties of the intact material where anisotropy may have influenced the behavior of the specimen under load. For example, the schistosity may lead to premature failure of samples under compression and the thin section may reveal the characteristics of the fabric in rock types where this occurs. The examination can identify alteration and weathering associated with the faulting and hydrothermal action. For example, this information can be used to distinguish between contemporaneous alteration and Pleistocene or Recent weathering. 4.4

Interpreted fault trend from boring data

Faults

The orientation of faulted and brecciated ground was achieved in part by interpretation of the acoustic imaging data and correlation between borings of the geological tell-tales described above. The acoustic image was particularly important for orienting geological markers. The data were used to the approximate the thickness of the faults and shears and their relative orientation to the proposed tunnel alignment. For example, there is a zone of disturbed rock in the area of 48th Street and 2nd Avenue created by a complex of conjugate faults juxtaposing a major pegmatite with intercalated garnetiferous schist and granofels. This is a very complex structural region that could only be interpreted by interpolation between closely spaced and very carefully logged boreholes with fracture orientation by acoustic imaging of the borehole wall. The interpreted trend of the dominant fault of the conjugate system is placed on the Viele Map of 1874 for the Borough of Manhattan (Fig. 2). The Viele map shows a fluvial meander and the boring data confirm this by braided channel and flood plain deposits. The channel corresponds to one of the

Figure 2. Extract from the Viele map of 1874 between 40th Street to 58th Street.

major structural trends following a NW-SE orientation toward the East River. This diversion coincides with a large pegmatite slab at 2nd Avenue and 47th Street that forms the southern bank of the channel. This is a relatively deep incision for mid-town and the aperture of the fractures and prevalence of oxidation and detritus in the fractures support a hypothesis that the channel may continue to show post-orogenic brittle movement. The same approach as described above has been applied to a section of central Manhattan and the preliminary interpretation of the trend or strike of the major faults identified during the exploration to date is shown in Figure 3. This exercise shows that the most significant fault/shear/fracture zones are oriented NNW-SSE, WNW-ESE and NNE-SSW. The engineering classification of the material within the fault zone can be applied to the specific engineering circumstance in terms of the location within a particular construction element. This is not within the scope of this paper. 5 CONCLUSIONS The 2nd Avenue Subway Project will be one of the largest and complex construction projects in the USA and a critical part of the success for the project will be the definition of the fundamental structural geology. This is being achieved for the project by adopting conventional and specialized geological techniques. The following conclusions are made based on this preliminary phase of interpretation.

361 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-44.qxd

2/18/04

3:01 PM

Page 362

iii. The conventional methods have been enhanced successfully by applying fabric and petrographic logging and combining fracture logging of the core with reconstructed acoustic images from the borehole wall. iv. The use of these combined geological techniques has improved the ability to predict and interpret fundamental geological structure for the tunnel alignment. v. The preliminary interpretation of the data confirms the dominance of NW-SE trending faults with NNE-SSW subordinate faults. vi. The interpretation reveals that significant faults occur throughout the Manhattan Prong but their location can be better estimated using the techniques described in this paper.

6 LIMITATIONS The views and opinions expressed in this paper are those of the authors and not their companies, their employers or their subsidiaries. The paper is based on the preliminary interpretation of the exploration program for the project.

ACKNOWLEDGEMENTS The authors gratefully acknowledge Anil Parikh and Madan Naik of New York City Transit, DMJM  Harris Arup Joint Venture for their permission to publish this paper and GZA GeoEnvironmental Inc. particularly R J Palermo the Chief Geotechnical Engineer and T Kwiatkowski for their support.

REFERENCES

Figure 3. Strike of major faults identified by the structural interpretation of the exploration data for the 2nd Avenue Subway Project.

i. Major infrastructure projects in complex geological environments require a wide range of exploratory and interpretive techniques to establish the fundamental geology of a region. ii. Conventional methods and standards for logging rock core do not include sufficient geological depth for these studies.

Baskerville, C.A., 1982. The foundation geology of New York City: Geological Society of America. Reviews in engineering geology. Vol. 5, pp. 95–117. Baskerville, C.A., 1994. Bedrock and engineering geologic maps of New York County and parts of Kings and Queens counties, New York and parts of Bergen and Hudson counties, New Jersey, Miscellaneous Investigation Series Map, I-2306. Two sheets, scale 1:24,000. Shah, A.N. et al. Geological hazards in the consideration of design and construction activities of the New York City Area, Environmental & Engineering Geosciences, Vol. IV, No. 4, Winter 1998, pp. 524–533. Viele, E., 1874. Topographic Atlas of the City of New York.

362 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-45.qxd

2/18/04

1:53 PM

Page 363

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

An automated structural monitoring system for the Federal Reserve Bank of Boston Thomas L. Weinmann Construction Technology Laboratories, Inc., Skokie, IL, USA

Lewis Edgers Tufts University, Medford, MA, USA

ABSTRACT: Among the burdens of massive underground construction programs in urban areas is guarding public safety and protecting building owners’ property. This paper describes a unique monitoring system installed for the Federal Reserve Bank of Boston (FRB), and explains the decision-making strategy in the case of unplanned consequences of underground excavation nearby. The FRB building, a 21-year old, 33 storey facility, stands immediately adjacent to the 34 m (110-ft) deep, open-cut I-93 construction of Boston’s Central Artery on Atlantic Avenue. Each of the Central Artery Tunnel’s prime construction contracts provides for geotechnical and structural monitoring of its own highway segment as construction proceeds, assessing and assuring that ground movements remain within anticipated range and construction components are not overstressed. The structural monitoring of the Federal Reserve Bank’s foundation has been the responsibility of a team gathered in 1993. As a part of this team, the CTL (Construction Technology Laboratories, Inc.) structural health monitoring specialists designed and installed a 500-sensor system utilizing automated data acquisition/communication focused specifically on protecting the needs of this facility, and fully integrating it with the geotechnical monitoring program. The automated system consists of more than 300 strain gages on building components monitored continuously. All strain gage data is pre-processed by the system’s central computer and resolved into meaningful parameters such as column bending strains, and automatically compared against predetermined specific response thresholds. These were devised to reflect anticipated impacts on the structure at each individual stage of the construction process. Active data communication links connect the Bank’s operations staff, the CA/T project managers, the contractor and system operations staff at the CTL Structural Laboratory in Chicago.

1 INRODUCTION The Federal Reserve Bank (FRB) of Boston is a critical regional bank-processing center for the Northeastern US. As part of the Central Bank System, this facility processes approximately $90 billion in wire transfers daily. The result of any disruption to bank business could adversely impact the entire New England economy. The bank is a 7 day, 24 hour operation and therefore every effort must be made to ensure uninterrupted operation. The facility consists of a 33-storey tower (Fig. 1) and 4- and 5-storey operations wings, which house extensive security and operations functions. The main tower structure includes two steel braced frames that carry the wind and gravity loads to the two 2.4 m (8 ft) thick, 29.6 m (97 ft) by 14.5 m (47.5 ft) tower mats. The tower is connected to a 4-storey operations wing

over a two-level basement that extends beyond the tower structure to the site limits bounded by Atlantic Avenue, Summer Street, Dorchester Avenue and Congress Street. Spread footings and slab-on-grade foundation support the two-level basement near Atlantic Avenue. The upper level of the basement is used for bank records, Property Management offices and loading dock. The lower level (B2) of this basement is mechanical and parking area. This crucial complex sits immediately adjacent to a portion of the 110-ft deep, open cut excavation for Boston’s Central Artery/Third Harbor Tunnel (CA/T) Project. This section of the CA/T project consists of a 630 m (2065 ft) long, three- to five-lane cut-and-cover highway tunnel for the northbound central artery, overlain by a bus transitway tunnel (Fig. 2). The excavation for this tunnel alignment along Atlantic Avenue reaches its greatest depth, approximately 34 m (110 ft)

363 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-45.qxd

2/18/04

1:53 PM

Page 364

Figure 3. 3-D view of Dewey Square intersection.

This paper describes the process for selection, installation and monitoring of instrumentation and the processing and evaluation of such data.

Figure 1. Federal Reserve Bank of Boston.

2 BACKGROUND

Figure 2. Location of Federal Reserve Bank relative to Central Artery Tunnel.

in order to pass beneath the MBTA Red Line (Fig. 3), an active subway carrying more than 300 trains per day beneath Summer Street. As a result of the importance of this facility and the proximity to construction activities, a structural performance monitoring program was developed to provide the owner with real-time data pertaining to building response from adjacent construction activities.

During initial design of the CA/T project, both historic and prominent structures along the project were evaluated for potential disturbances from adjacent tunnel construction. The FRB assembled a team of consultants to work with the project managers, Bechtel/ Parsons-Brinckerhoff (B/PB) and the Massachusetts Turnpike Authority (MTA) in development of measures to minimize the effects of the CA/T excavation to the FRB. This team consists of Dr. Lewis Edgers (geotechnical), Mr. Ken Wiesner of LeMessurier (structural) and Mr. Thomas Weinmann of CTL (instrumentation). This team was formed in 1992–1993 with the initial task of reviewing design/construction/ specification documents for the CA/T project in the immediate vicinity of the FRB. Pre-construction analyses of the effects of excavation induced ground movements on the FRB were performed by the section designers (1,2). Initial analysis determined potential soil displacements under the FRB in close proximity to the excavation of approximately 0.9 cm (0.35 in.) and 2 cm (0.80 in.) vertically and horizontally, respectively (Fig. 4). These soil movements are the result of potential slurry wall displacements of up to 2.4 cm (0.95 in.) (Fig. 5). Because they are highly indeterminate, some large modern buildings are sensitive to even small differential foundation movements. On the other hand, the

364 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-45.qxd

2/18/04

1:53 PM

Page 365

basement slabs and additional cracking in the foundation walls at abutting corners. 3. There existed marginal strength in the connection between the steel column and two-way flat concrete slab at the first basement level for resisting the bending type loads imposed by the laterally displaced foundation and column line. 4. The high-rise structure may experience some cracking of the first basement and ground level slabs due to displacement of the north tower mat. 5. The auditorium structure may experience localized cracking of slab, wall, and architectural elements. These analyses showed that the FRB structures are more sensitive to horizontal than vertical ground movements. The studies served as a basis for a plan of action that included pre-construction structural strengthening and performance monitoring. The structural strengthening included increasing the moment carrying capacity of the column/floor slab connection at selected locations on the B1 Level. This was implemented to specifically address item 3 above. One-inch thick bearing plates were welded to the base of the columns on top of the B1 level floor slab and column capitol/drop panel thickness was built up at the underside of the B1 level floor slab. An instrumentation program was initiated to monitor the performance of the FRB structure resulting from adjacent deep excavation. This program was developed jointly by the FRB Consulting Team, B/PB Project Management Team and included both geotechnical and structural instrumentation.

Figure 4. Predicted soil movements.

3 INSTRUMENTATION

Figure 5. Predicted slurry wall displacement.

benefits of structural creep and relaxation on building response are unknown. There are virtually no data available on the response of large structures like the FRB to excavation induced ground movements or to seasonal effects. The site-specific FRB analyses (1,2) concluded that: 1. The foundation movements will not cause collapse or overall stability problems. 2. The two-storey below grade basement structure may experience some cracking in the first and second

The CA/T designers established an array of project instrumentation in order to monitor the performance of the open cut excavation support system. This instrumentation included vibrating wire strain gages installed on selected struts and soldier piles, vertical inclinometers in the SPTC wall and in soil outside the excavation, ground water observation wells, piezometers and vertical and horizontal deformation monitoring points. Based on the results of initial analyses and the importance of the FRB, additional instrumentation was included to supplement the original plan. These instruments were selected to measure key parameters identified by the pre-construction analyses. Criteria for the selection of instruments included accuracy, repeatability, durability and cost. In all instances the instruments selected were instruments with proven reliability, generally in conformance with project wide specifications and agreed upon by all parties. This instrumentation included vibration monitoring, monitoring of horizontal and vertical displacements, tower tilt and structural monitoring using strain gages.

365 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-45.qxd

2/18/04

1:53 PM

Page 366

The supplemental instrumentation program incorporated redundancy in the types of instruments selected and made use of both manually read and automated instrumentation. The automated instrumentation consisted primarily of strain gages that were read continuously for the duration of the project. The manually read instrumentation was read approximately every 4 to 6 weeks unless the automated system triggered alarms, initiating the response plan that included more frequent reading of manually read instrumentation. 3.1

Manual readings

The instrumentation program used some sensors that were best suited for manual readings. These included vertical inclinometers, convergence gages, tilt meters, mechanical strain gages and a precision survey network. These instruments were generally read on a 4–6 week interval. Three vertical inclinometers were installed in the B2 Level of the FRB. These instruments were installed to an approximate depth of 34 m (110 ft) and measured the horizontal movements of the B2 level floor slab and soil at 0.6 m (2 ft) vertical increments (Fig. 6). Thirty-three pairs of convergence gage anchorage points (CGAPs) were installed on selected columns in the B2 Level of the FRB. These instruments were used to measure relative horizontal movement of select basement columns, exterior basement wall and main tower mats (Fig. 7). Two biaxial tilt meters were installed on the tower mat foundations at the B2 level. These instruments were used to measure tower tilt (Fig. 8). Forty-one columns were instrumented with strain gages to measure bending strains in the columns above and below the B1 level floor slab. At each of these locations (Fig. 9), both mechanical and vibrating wire strain gages were installed. The vibrating wire strain gages were connected to an automatic data acquisition system and will be discussed in subsequent sections. The intent of the mechanical strain gage is to provide redundancy to corresponding vibrating wire strain gages in the event of failure or unreliability of the automated system. After the first year of the monitoring program, manual reading of the mechanical strain gages was discontinued. It was agreed that the readings from the mechanical strain gages did not provide the repeatability and resolution of the data obtained from the automated system. The gages are still in place and could be read and still provide a relative measure of strain in the columns, although not as accurate or with the same magnitude of resolution as the vibrating wire strain gages. A precision survey network was established on the B2 Level of the FRB. This survey network was used

Figure 6. Inclinometer locations.

Figure 7. Convergence gage locations.

Figure 8. Tilt meter locations.

to measure both vertical and horizontal displacements in the B2 Level. Eighty-three vertical displacement measurement points (VDMPs) were installed (Fig. 10) to measure relative and absolute vertical settlements of select columns, wall areas and tower mats.

366 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-45.qxd

2/18/04

1:53 PM

Page 367

Figure 9. Strain gage locations.

Figure 11. Horizontal DMP locations.

Figure 10. Vertical DMP locations.

Figure 12. Survey network schematic.

Thirty horizontal displacement measuring points (HDMPs) were installed (Fig. 11) to measure horizontal movements (spreading) of select columns and wall areas. The HDMPs provide some redundancy for the data obtained with the CGAPs. This precision survey network provided orders of magnitude of greater accuracy than regular construction surveys. Monumentation included stainless steel machined targets permanently affixed to the structure. Station setups were permanent fixtures on the floor (Fig. 12) with all equipment being dedicated to the FRB.

These units ran continuously and triggered data collection/storage based on predetermined response values. Forty-one columns were instrumented with vibrating wire strain gages (VWSGs) to measure bending strains in the columns above and below the B1 level floor slab (Fig. 9). Each column had 8 gages for a total of 328 VWSGs. The vibrating wire strain gages were connected to an automatic data acquisition system and controlled by a central computer. The central computer has the ability to display and archive data as well as control remote access to the system.

3.2

3.3

Automated readings

This program also used sensors that were best suited for automated readings. This allowed for collection and interpretation of data in real time. These sensors included seismographs and vibrating wire strain gages. Four seismographs were installed in the FRB to monitor and protect sensitive Bank equipment from excessive construction vibrations. Two were installed in the B2 Level in close proximity to construction while the other two were installed in operation sensitive areas such as the UPS room and the Computer center.

The data logging system for the VWSGs collects data continuously, processes data to individual column bending strains, compares these values to predetermined response values and utilizes dial out modems to alert personnel via pager networks (Fig. 13). Upon notification of such alarms, personnel have the ability to remotely access the central computer for review of current data and printing of reports. The remote access feature is used by Bank and BP/B personnel in Boston as well as CTL in Skokie, Illinois.

367 Copyright © 2004 Taylor & Francis Group plc, London, UK

Automated system alarms

09096-45.qxd

2/18/04

1:53 PM

Page 368

Contingency plans were developed early in the design phase to enable prompt action in the event that action is required. These plans included additional bracing levels and tieback support to reduce slurry wall movements. The contingency plans required the contractor to have on-hand the necessary materials to implement these measures. The success of the management plan has been the result of regular meetings of the CA/T design and construction team with the FRB Consultants and representatives. The meetings allow for discussion of current instrumentation measurements in relation to construction activities and future construction and measurement schedules.

Figure 13. ADAS location schematic.

The software is unique in the design of eliminating “false triggers” by using software algorithms that require sustained levels of data response to eliminate triggers caused by occasional data spikes, wind gusts or fluctuations of data at borderline response values. 4 DATA MANAGEMENT The instrumentation program was developed jointly by the FRB Consultant Team and the Project Managers – B/PB. Instrumentation was installed well before construction to verify the reliability, repeatability and accuracy of the instruments. Also, this pre-construction time period allowed for obtaining baseline data resulting from seasonal effects only. As part of this development process, instrument response values, response procedures and contingency plans were established. Each and every instrument in the Geotechnical Instrumentation Specification was assigned a threshold and limiting value. Limiting values were determined from the analyses described earlier in this paper. Threshold values, typically one-half to two-thirds of the limiting values, provide some reserve capacity below the limiting value. The response value of the instrument was generally the threshold value established for that instrument. The goal of the program was to act upon the response of the threshold value to prevent reaching the limiting value. Automated data was generally reviewed on a weekly basis. If a threshold value was reached, further action was required. This usually started with review of construction activities in the area of the instrument. Data from corresponding manual instruments were also reviewed and correlated with construction activities and baseline seasonal variations. A Contractor/ Engineer meeting may be required to discuss response values, need for more frequent meetings or modifications to construction procedures. As a last resort, predetermined contingency plans could be implemented. The response time for these activities was 24 hours.

5 INSTRUMENT MEASUREMENTS In general, the results of excavation in the vicinity of the FRB have shown slurry wall movements well below those predicted in the design phase. Not anticipated in the design phase was the need for substantial construction dewatering in the vicinity of Dewey Square. The underpinning of the MBTA Red Line at South Station required the lowering of the piezometric surface to below 6-m (20-ft). The FRB has experienced settlements due to compression of the underlying soils caused by this local ground water lowering (Fig. 14). However, measured horizontal movements have been small due to reduced hydrostatic pressures on the very stiff CA/T bracing system. 5.1

5.2

Tilt meters

The tilt meters inside the FRB on the tower mat foundations show minimal tilt other than that attributed to seasonal effects. The instrument was read with a

368 Copyright © 2004 Taylor & Francis Group plc, London, UK

Inclinometers

The inclinometers inside the FRB have shown minimal horizontal displacements under the building. The inclinometers in the SPTC wall in front of the FRB have shown some horizontal movement, but much less than predicted. The quality of data obtained from these instruments was dependent on the instrument and susceptibility to damage. The inclinometers in the FRB were read using a dedicated instrument stored in the FRB. The inclinometers in the SPTC wall were read using project wide instruments in a construction environment. Over the course of the project, these instruments became worn, damaged and often replaced. The installations in the FRB provided a relatively clean environment with easier access for reading the instrument. The installations in the SPTC wall were in a somewhat hostile environment and much more susceptible to damage.

09096-45.qxd

2/18/04

1:53 PM

Page 369

Figure 14. Piezometric elevation settlement.

dedicated readout box in a relatively clean environment and provided stable long-term data. 5.3

Convergence gage anchorage points

The CGAPs inside the FRB show relatively stable readings over time. The longer the measurements span, the larger the fluctuation in the reading. These instruments are operator sensitive therefore; fluctuations in readings may be attributed to a change in reading operator. 5.4

Vibration monitoring

Vibration monitoring in the FRB provided insight to the levels of vibration seen in specific areas of the facility. All vibrations recorded were well below response values. 5.5

Horizontal displacement measurement points

The HDMPs inside the FRB were part of the precision survey network. These instruments provided stable long-term measurements. Data from these instruments correlated well with the CGAP data. 5.6

Vertical displacement measurement points

The VDMPs inside the FRB were part of the precision survey network. These instruments provide stable long-term measurements and were extremely useful in determination of relative and absolute settlements due to dewatering. Data from these instruments were used to generate settlement contours (Fig. 15). 5.7

Vibrating wire strain gages

The VWSGs in conjunction with the data loggers and central computer provided a means for real-time structural response. Most of the above instruments show the

Figure 15. FRB settlement contour.

symptoms of response to the excavation/dewatering, while the strain gages show the actual structural response in real-time. These instruments provided stable, long-term measurements and were useful in providing real-time notification of instrument response. Early on in the project, an item of interest apparent in the strain gage data plot (Fig. 16) is an event of August 26, 1995. The strain gage data plots at selected locations (E9 for example) showed a distinct increase in tensile strain for all gages on this column. This measured tensile strain increase, or compressive strain reduction, is attributed to the removal of the landscaping berms located on the plaza in front of the FRB Building, and was reflected in the operation early on for the intended use of the instrumentation program for the FRB Building. The actual construction excavation began in early 1996. The instrumentation was installed and the system commissioned in early 1995. It was important to establish seasonal trends in response of the structure. As the seasons change the ground freezes and thaws affecting the structural response of the building. For this reason, data was collected for all instruments to provide seasonal adjustments to alarm response values. Seasonal response for average bending strains in the instrumented columns is shown in Fig. 17. As shown in Fig. 18, a number of noted events are reflected in the strain gage data for Column D13. This column is in close proximity to the slurry wall and at

369 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-45.qxd

2/18/04

1:53 PM

Page 370

Figure 17. Seasonal response for average bending strains.

Figure 16. Column E9 strain gage data with photo of column location relative to landscape berms.

the same elevation as level of slurry wall bracing struts (Fig. 19). Of particular interest is the major axis bending strains shown on the B2 Level plot. A seasonal trend is apparent in the data. This could be the result of cold weather causing contraction of the bracing strut thereby relieving some of the horizontal force applied to the wall/floor slab. Superimposed on this trend are the effects of excavation in this vicinity as it progresses from the elevation of this floor slab downward. In January 2000 the threshold response value for the major axis bending strain for Column B2D13 were exceeded. This alerted personnel to review construction activity and adjacent instrument data. Data from the inclinometer in the slurry wall immediately

Figure 18. Column D13 strain gage data.

Figure 19. Slurry wall bracing struts.

370 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-45.qxd

2/18/04

1:53 PM

Page 371

adjacent to this column indicated movements that corroborate the strain gage data. Data from subsequent DMP surveys also indicated movements that corroborated the strain gage data. This information was reviewed and a determination was made to increase the response value for this column location and increase reading frequency of surrounding instruments. The above example shows the value of real-time data acquisition, automated alarm triggering, supporting instrumentation, understanding of seasonal trends and construction activity and cooperative data management efforts.

gages provided the most cost effective means for performance monitoring. The instruments requiring manual readings became a costly endeavor over a 8-year period. The data from these instruments still had to be manually input for processing and presentation. The initial cost of the automated system will offset the potential costs if these instruments were read and processed manually. The benefit of having real-time data with alarm notification and remote access for instant review not only provided an overall cost savings, but a continuous level of comfort not achieved with instruments read on a monthly or weekly basis.

6 SUMMARY

ACKNOWLEDGEMENTS

This paper describes the instrumentation system that was developed to measure the effects of an adjacent deep excavation on a 33-storey structure. The proper selection of instruments is important to achieving this goal. Of greater importance, in the authors’ view, is the interpretation of data and understanding of activities that impact such data. Accumulation of baseline data provides a means for determining seasonal trends, repeatability and accuracy of an instrument and most importantly, reliability of the instrument. Instruments that are operator sensitive require dedicated or skilled operators. Whenever possible provide “controls” for comparing operator readings or when changing readout instruments. When backup readout devices are planned to be used, take “tie-in” readings with both readout devices in anticipation of using either one. Take proper care of instruments and readout devices. Try to use all the reliable data at your disposal to look at the big picture. From an economics view, the Automatic Data Acquisition System reading the vibrating wire strain

The instrumentation program was developed as a cooperative effort between the abutter consultants and the CA/T project managers, Bechtel/Parsons-Brinckerhoff (B/PB). The authors acknowledge the Federal Reserve Bank of Boston, B/PB and the Massachusetts Turnpike Authority (MTA) for their support in the preparation of this paper. The views expressed in this paper are those of the authors and are not necessarily those of the MTA.

REFERENCES Stevenson, S. Federal Reserve Bank of Boston, Foundation Movements, Task 1, Structural Analysis. Prepared for the Massachusetts Highway Department, CA/T Contract C11A1, November 1993. Stevenson, S. Federal Reserve Bank of Boston, Foundation Movements, Task 2, Structural Analysis of Auditorium Area. Prepared for the Massachusetts Highway Department, CA/T Contract C11A1, August 1994.

371 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-46.qxd

2/18/04

2:02 PM

Page 373

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

A deep horizontal boring – technical and contractual issues J. Glastonbury, K. Ott & J. Freitas Parsons Brinckerhoff Quade and Douglas Inc., New York, USA

B. Russell Parsons Brinckerhoff Energy Storage Services Inc., Houston, USA

M. Wooden Malcolm Pirnie, Inc., Atlanta, USA

W. Meakin & J. Canale New York City Department of Environmental Protection, USA

ABSTRACT: The Rondout West Branch tunnel of the Delaware Aqueduct provides up to 900 million gallons per day of drinking water to New York City. Various studies have confirmed that the 60 year old structure has developed leaks in the vicinity of Roseton in the mid-Hudson Valley, approximately 60 miles north of New York City. An investigative program was developed which included a long directionally drilled borehole to gather hydrological and geotechnical information on areas of suspected leakage. This paper highlights some of the technical and contractual challenges faced in drilling a borehole in close proximity to a deep pressurized structure. As well as discussion on the drilling and investigative methods, this paper outlines the success at maintaining good directional control on the hole while at the same time recovering valuable data on ground conditions. 1 BACKGROUND 1.1

Delaware Aqueduct history

The Rondout-West Branch (RWB) Tunnel of the Delaware Aqueduct is a 45-mile long, pressurized water tunnel that supplies up to 900 million gallons of water per day (MGD) to New York City. It commences high in the Catskill Mountains, bringing water down to the West Branch Reservoir in the midHudson valley. The tunnel was constructed in the late 1930s and early 1940s using drill and blast excavation methods. It was commissioned in 1944 following completion of hydrostatic testing and since commencement of service the tunnel has been dewatered twice for inspection purposes. 1.2

Tunnel construction

Initial ground support varied throughout the length of the tunnel from unsupported spans to the use of steel sets and timber lagging. The tunnel structure was completed with a 13-foot 6-inch diameter concrete lined waterway. Invert elevations along the tunnel vary from approximately 500 feet to more than 2000 feet below ground surface.

During construction of the tunnel in the Roseton area (immediately west of the Hudson River), workers encountered faulted, folded and solutioned limestone and experienced large water inflows into the excavation. A steel inner-liner was placed within the final concrete lining along a section of tunnel about 1,100 feet long where high groundwater inflows and poorer ground conditions were encountered. Grouting was carried out at various locations in this area of the tunnel during construction in attempts to reduce water inflows and improve ground conditions. This was reported to have reduced water inflows from peaks of the order of 1800 gpm down to 240 gpm. General contact grouting (at low pressures) and cutoff grouting (at higher pressures) were also carried out during initial construction along this section of lined tunnel in order to fill voids behind the liner and further reduce any inflows. 1.3

The RWB tunnel in the vicinity of Roseton passes entirely through the Wappinger Limestone. This formation is characterized by tensile normal faults and jointing as well as compressional structures ranging

373 Copyright © 2004 Taylor & Francis Group plc, London, UK

Tunnel geology

09096-46.qxd

2/18/04

2:02 PM

Page 374

from regional scale folding and thrusting, through to smaller-scale folds, minor slip faults and narrow crush-zone bands. These crush zones were reported in construction records to vary from weak gouge-clay bands to bands of stronger rock in which some structure has been obliterated by fragmentation and deformation. The tunnel in the Roseton area is traversed by two major fault systems. The most prominent of these runs northeast-southwest (part of the New Hamburg thrust fault), with the second system running northwestsoutheast. The Wappinger Limestone varies from massive to closely jointed in the Roseton area with local bedding orientation exhibiting a gently east-dipping orientation. 1.4

Observed leakage

Various hydraulic studies have confirmed that surface water expressions in the vicinity of Roseton are linked to leakage from the RWB Tunnel. Up to 34 MGD is estimated to be leaking from the tunnel at peak flows and about 10 MGD is thought to be transmitted through these surface water expressions at Roseton. The realization that water losses of this magnitude were occurring from the tunnel prompted the development of an investigation program. This program is aimed at identifying and characterizing the areas of leakage, with the ultimate objective of developing appropriate tunnel repair solutions.

2 HORIZONTAL BORING PROGRAM 2.1

Objectives

One of the major components of the investigation program was a geotechnical boring, principally designed to investigate the location and characteristics of sources of leakage in the vicinity of the steel lined section of the RWB Tunnel, near Roseton. Roseton Horizontal Boring 1 (RHB-1) was a directional borehole drilled to collect geologic and hydrogeologic data in the suspected area of water leakage. A horizontal boring program was favored over conventional vertical boreholes because it provided a more continuous record in the zone of interest and also minimized permit and site access requirements. The objectives of RHB-1 included: 1. Examination of ground conditions in the vicinity of the tunnel (including any possible changes in ground conditions since tunnel construction). 2. Studies of rock mass groundwater regime in the vicinity of the leak(s), including collection of groundwater pressure data along the borehole to assist in locating the leak and major flow pathways.

The results obtained from the RHB-1 boring would ultimately be used, with information from other elements of the investigation program, to develop tunnel remediation plans. 2.2

The borehole was situated in the Roseton area of the Town of Newburgh, Orange County, NY. The collar of the borehole was positioned on a ridgeline approximately 3000 feet west of the banks of the Hudson River. The site was selected on the basis of its proximity to the steel lined section of the tunnel and also because it afforded a suitable elevation difference (between ground surface and tunnel) over which the drilling angle could be progressively changed. RHB1 was drilled parallel to the tunnel. The hole was located such that horizontal rock coring could commence approximately 200 feet prior to the start of the steel lined section of the aqueduct and be completed approximately 200 feet past the end of the steel lining. RHB-1 commenced at the surface with an inclination of 64 degrees from horizontal and progressively built to a horizontal alignment over a measured borehole distance of approximately 1370 feet (hereafter referred to as the “build-section”). The vertical elevation difference between the borehole collar and end of the build section was approximately 725 feet. Figure 1 shows the vertical alignment of the RHB-1 boring in relation to the tunnel. Following completion of the build section, rock coring was conducted along a horizontal alignment situated 40-feet north and 20 feet above the tunnel centerline, as illustrated in Figure 2. In light of the fact that drilling was occurring in close proximity to a high-pressure structure a 3-foot radius of control was established around the target borehole alignment (for the horizontal portion of hole) prior to commencement of drilling. This radius of control was established as a boundary on allowable deviation of the borehole. Results of steering control are presented in a later section of this paper. In total, RHB-1 was drilled to a final measured borehole depth of 2887.5 feet, terminating approximately 725 feet below collar elevation and approximately 2580-feet horizontal distance from collar. The boring was completed over a four-month period on a 24-hour per day, seven day per week basis. 2.3

Drilling and steering methods

Drilling of borehole RHB-1 was conducted using a platform mounted LS244EC drilling rig, equipped with wireline capability. Drilling through the build section of the hole was carried out using tri-cone roller bits of varying sizes to allow for staged casing

374 Copyright © 2004 Taylor & Francis Group plc, London, UK

Borehole location and alignment

09096-46.qxd

2/18/04

2:02 PM

Page 375

100 Ground surface

Elevation (feet)

0 RHB-1 profile

-100 -200

Hydrological connection with surface springs noted

Steel lined section of tunnel

300 -400

RWB Tunnel

-500

1524+32

1522+32

1520+32

1518+32

1516+32

1514+32

1512+32

1510+32

1508+32

1506+32

1504+32

1502+32

1500+32

Tunnel Station

1498+32

-600

Figure 1. Vertical alignment of RHB-1 borehole in relation to Rondout-West Branch Tunnel.

40’ 0” 3.895” DIAMETER BOREHOLE

ROCK SURFACE

EL -568 STEEL INTERLINER 20’ 0”

13’6” DIAMETER WATERWAY 30’ 4”

STEEL RIB SUPPORT

3’ RADIUS OF CONTROL TARGET BOREHOLE ALIGNMENT

El -588 (VARIES)

Figure 2. Profile of RHB-1 borehole in relation to Rondout-West Branch Tunnel. installation. Once the hole reached a horizontal alignment, coring commenced using HQ triple tube core barrels. Steering (or directional control) of RHB-1 in the build section of the hole was accomplished using MWD (“measure while drilling”) technology. The MWD downhole assembly contained orthogonally mounted accelerometers and magnetometers that calculate tool face inclination and bearing. Regular readings were taken while drilling and were transmitted through the ground as electric signals and received at surface antennae. The MWD system enabled regular monitoring of the azimuth and bearing of the hole as drilling progressed. Adjustments were made to hole orientation

using a bent sub (otherwise referred to as a “mudmotor”) located immediately behind the drilling bit. The orientation and position of this bend in the mud motor allowed for regular minor corrections of hole orientation. A further check on MWD orientations was provided by regular downhole gyroscopic surveys. Blow-out prevention (BOP) equipment was assembled at the collar of the borehole after the hole was progressed to a measured depth of 267 feet and casing had been set and cemented into competent rock. The BOP assembly served to provide a shut-off system for the borehole in the event that high water pressures were encountered during drilling. Various proprietary drilling polymers were used during progression of the hole in order to stabilize the

375 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-46.qxd

2/18/04

2:02 PM

Page 376

hole, lubricate and cool drilling equipment and enable return to the surface of drill cuttings. Various loss containment materials (LCM’s) were also added to the drilling fluid at various stages during progression of RHB-1 for the purpose of limiting the loss of drilling fluids into the ground.

understanding and confidence of the borehole behavior, the interval between gyro surveys was progressively increased to approximately 200 feet by the time the end of the hole was reached.

2.4

The following two geophysical tools were run in the hole:

Hole progression

Drilling of the build section of RHB-1 was progressed using a variety of tri-cone roller drill bit sizes, with details presented in Table 1. No rock coring was conducted in the build section, but rather drill cuttings were sampled at five foot intervals. Casing was installed in four stages through the build section of the hole to a total depth of 1370 feet, as indicated in Table 2. The approximate duration of other key activities during the boring program are indicated in Table 2. Within the horizontal portion of the hole, the typical progression sequence involved: 1. Coring for a prescribed distance (initially adopting a distance of 30 feet and ultimately working up to approximately 200 feet); 2. Conducting a gyroscopic survey of hole orientation; 3. Conducting geophysical logging; 4. Conducting appropriate hydrological testing if appropriate; and 5. Recommencement of coring. In the initial stages of the cored section, gyro surveys were conducted at 30 foot intervals. With increased

Table 1. As-built details of RHB-1. Measured depth

Bit size

0.0–11.4 ft 0–59.5 ft 0–267.3 ft 0–1371.7 ft

Backhoe excavation 19 tricone roller bit 14.75 tricone roller bit 6.75 tricone roller bit (to 905) 6.5 tricone bit (905 to 1371.7) HQ3 triple tube coring

1373.4–2887.5 ft

2.5

1. Acoustic televiewer tool (ATV); and 2. Natural gamma tool. The ATV tool provided a radial image of the borehole wall using an acoustic beam, enabling identification of discontinuities as small as 0.1 mm width within the rock mass. The ATV tool was also equipped with magnetometers and accelerometers enabling constant image adjustment and measurement of borehole deviation (reported as travel time measurement from the tool to the borehole wall). This travel time measurement provided a “de-facto” indication of borehole size, as a replacement for traditional caliper logging tools. The natural gamma tool was similarly run in the hole in stages throughout the course of drilling. This tool provided a means for identification of zones of higher gamma ray production (natural radioactivity) most typically associated with trace element concentrations of uranium, thorium and potassium. The natural gamma log was particularly useful for stratigraphic correlation and identification of clay or weathered zones within the limestone strata. The data gathered from this tool complimented the ATV data and was particularly valuable in the build section of the hole where no core was recovered. 2.6

Table 2. Progression of RHB-1.

Downhole logging

Core recovery

At the commencement of drilling it was anticipated that several alignment correction runs would be required and hence core recovery would be reduced. As a result of good alignment control, core recovery from the horizontal portion of the hole was in excess of 96% which was beyond initial expectations given that coring was not possible during borehole alignment corrections.

Activity

Duration

2.7

Stabilizer casing to 11 feet Conductor casing to 60 feet Surface casing to 267 feet Completion of build section to 1370 feet Installation of build section casing Completion of coring to 2887.5 feet Abandonment

1 day 3 days 5 days 25 days 2 days 63 days 3 days

Abandonment of RHB-1 commenced after completion of coring to a distance some 200 feet beyond the end of the steel-lined interval of the tunnel. Abandonment was completed using a combination of packers and cement grout. Packers were run in the hole and set at specific depths in order to isolate high pressure and/or high water inflow zones. Cement grout was pumped into the

376 Copyright © 2004 Taylor & Francis Group plc, London, UK

Hole abandonment

09096-46.qxd

2/18/04

2:02 PM

Page 377

hole above and below each of these packers. Cementing of the borehole was completed up to ground surface.

4 TECHNICAL CHALLENGES 4.1

3 GROUND CHARACTERIZATION 3.1

Rock characteristics

Wappinger Limestone was encountered over the complete length of the borehole, varying from unweathered to extremely weathered. The position and orientation of bedding and other geological structures was determined from both oriented core and ATV logging and was observed to match closely with original tunnel construction mapping records. Extensively weathered and variable decomposed zones of limestone were encountered at a few locations along the length of hole coincident with the steel lined section of tunnel. These zones correspondended closely with fault zones mapped in tunnel construction records. In areas surrounding these fault zones, the limestone was typically slightly weathered and high strength and often contained vugs or open defects which could act as conduits for groundwater flow. 3.2

Hydrogeological characterization

Two general types of flow paths were identified in the investigation. The first of these is characterized as a conduit(s) composed of discrete open discontinuities within an otherwise competent rock mass. The second type of flow path is characterized by faulted, decomposed limestone with occasional soil bands (composed of clay, silt, sand and gravel). Rock within this type of flow path has a typically high porosity and the rock mass is predisposed to a higher mass permeability. On two occasions hydrological connectivity was observed between discontinuities in the build section of the borehole and various surface water bodies (as indicated in Figure 1). In contrast, direct connectivity between discontinuities in the horizontal portion of the hole and surface water expressions was not observed despite significant drilling fluid losses over this portion of the hole. Occasional zones of high water pressure and variable flow were observed in areas of otherwise competent rock. These discrete open defects within competent rock were observed to have the potential to provide high permeability flow paths. However, water inflows and drilling mud losses were observed to be highest in the faulted, decomposed zones. Hydrological test data suggests that larger flows are concentrated in these heavily fractured and weathered zones. The permeability results and ancillary information indicate that the two different types of flow path within the rock mass will likely require different engineering solutions for stemming their flow.

At the outset of the boring program, it was recognized that one of the primary technical issues would involve dealing with potentially high water pressures in the borehole. At the drilling site, the calculated hydraulic grade line was some 460 feet above ground surface at full tunnel operating flows. Consequently, maximum anticipated pressures at the borehole collar were of the order of 200 psi, assuming direct connectivity with full tunnel pressures. During the course of drilling, the maximum recorded pressures at drill hole collar reached 175 psi. Values recorded along the length of the cored section of hole were typically 30 to 40 psi. These lower values are partly a reflection of reduced flows in the tunnel during the boring program and also increased column weight in the borehole due to the drilling mud, but are also considered to reflect head losses through the rock mass between the tunnel and borehole. The blow-out prevention assembly at the borehole collar provided a good means of control of the water flows associated with these high pressures. Flow from the borehole reached peaks of the order of 100 gallons per minute out the top of the HQ drill rods. 4.2

Hydrological connectivity

One of the primary challenges faced in drilling of this borehole was the issue of hydrological connectivity with surface water bodies which were known to be connected to tunnel leakage points. On multiple occasions, drilling fluids were observed to have migrated to these water bodies. Appropriate regulatory reporting of these episodes was carried out. While these events provided valuable information on the hydrogeological characteristics of the ground in the vicinity of the tunnel, attempts were made to reduce the likelihood of repeating such drilling fluid losses. These measures included use of loss containment materials in drilling muds and modifications to hydrological testing procedures. Standard packer tests on isolated intervals within the hole were replaced with a more passive test method involving measured inflow rates and pressures into specific borehole intervals as opposed to injection of fluid into the rock mass. 4.3

Borehole control

Drilling in close proximity to a pressurized structure posed additional challenges to the execution of this boring program. Accurate measurement and control

377 Copyright © 2004 Taylor & Francis Group plc, London, UK

High pressure conditions

09096-46.qxd

2/18/04

2:02 PM

Page 378

of hole orientation was perhaps the most important challenge faced in drilling this hole, hence the establishment of a tight target alignment. Within the horizontal section of the borehole there were two departures from this 3-foot radius of control limit. These departures occurred over short intervals in a direction away from the RWB tunnel. On these occasions, maximum deviation was of the order of 6 to 13 inches beyond the control envelope. Coring continued through these departure zones since it was possible to predict the path of the bore with a high degree of confidence. Gyroscopic surveys indicated that the alignment was within the 3-foot radius of control at all other locations within the horizontal portion of the borehole. Only one correction of the hole alignment was carried out using a mud motor steering tool. A complete change of drill bit assembly was required to undertake this alignment correction and no rock coring was possible over this 18-foot interval. This was the only alignment correction needed in the horizontal portion of the hole and represented a total of approximately 1.2 per cent of the horizontal cored section of the hole, contributing to the excellent core recovery reported in Section 2.

5 CONTRACTUAL ELEMENTS 5.1

Project team

The boring program was commissioned by the New York City Department of Environmental Protection, Bureau of Environmental Engineering (NYCDEP) as part of an overall investigation of the RondoutWest Branch Tunnel. NYCDEP’s consultant for this investigation, Malcolm Pirnie, Inc., was the boring program manager. Malcolm Pirnie’s responsibilities included contract administration and oversight, regulatory compliance monitoring, and community outreach coordination. Parsons, Brinckerhoff Quade & Douglas (PBQD) served as the boring program technical manager as a subconsultant to Malcolm Pirnie. PBQD played a lead role in developing the technical aspects of the program, and was responsible for continuous oversight of boring activities. PBQD also worked with the drilling contractor to adjust the investigational methods based on conditions encountered at the site. Drilling was conducted by the Lang Exploratory Drilling division of Boart Longyear Company as a subcontractor to Malcolm Pirnie. Boart Longyear was selected for this program based on its experience in gas and minerals exploration projects where significant drilling depths, tight steering control, and control of high pressures are routine requirements. Boart Longyear’s team included Scientific Drilling, Inc.

(steering and borehole surveying), and Colog, Inc. (geophysical logging). Additional parties to the boring program are identified in Figure 3. 5.2

Among the complicating factors that contributed to the complexity of this program was that NYCDEP did not own or otherwise control the property from which the program was implemented. Due to the depth of the RWB Tunnel, NYCDEP maintains no surface easements for this structure, and owns only limited land areas at shaft locations. The site selected for the boring program was located on property owned by Dynegy Northeast Generation, adjacent to one of its power generation facilities. Therefore, a formal property access agreement was required. Fortunately for NYCDEP and its program team, Dynegy granted permission to use its property for implementing the program, and was most cooperative and accommodating throughout the course of the operation. 5.3

Contract approach

The unique nature of this project necessitated a hybrid contracting approach that combined lump sum with time and materials payment terms. In the early stages of development the project team recognized that a great deal of uncertainty existed regarding subsurface conditions in the area to be explored, thus a contract based on traditional unit price pay items (such as cost per foot drilled) could not be accurately developed. The solution to this challenge was to negotiate a contract that assigned a lump sum price to the quantifiable aspects of the program. These included site preparation, mobilization, tooling and off-site preparation, and data reporting. The remaining work elements were divided into a series of time and materials based line items, and unit prices were negotiated for each. These included line items to equitably compensate the drilling team for project delays and unforeseen conditions beyond its control. The lack of a contractual relationship between PBQD and Boart Longyear was a key factor in the success of the boring program. By maintaining an independent perspective, PBQD was able to critically evaluate the driller’s progress and balance the technical and financial interests of the client. In turn, Boart Longyear proved adept at adjusting its approach and developing innovative solutions to address each new challenge. Together with Malcolm Pirnie, the parties were able to collaborate effectively to adapt the boring program to the dynamic conditions encountered in the hole and achieve the program objectives.

378 Copyright © 2004 Taylor & Francis Group plc, London, UK

Site access

09096-46.qxd

2/18/04

2:02 PM

Page 379

CLIENT New York City Department of Environmental Protection

PROPERTY OWNER

PROGRAM MANAGEMENT

Dynegy Northeast Generation Inc. Newburgh, NY

Malcolm Pirnie, Inc. Fairlawn, NJ

REGULATORY AUTHORITIES New York State Department of Environmental Conservation

TECHNICAL MANAGEMENT Parsons Brinckerhoff Quade & Douglas, Inc. New York, NY

DRILLING SUBCONTRACTOR Boart Longyear Company/ Lang Exploratory Drilling Salt Lake City, Utah

LAND SURVEYING Geod Corporation Newfoundland, NJ

LABORATORY TESTING GeoTesting Express, Inc. Boxborough, MA

BOREHOLE GUIDANCE

GEOPHYSICAL LOGGING

SITE PREPARATION

Scientific Drilling, Inc. Casper, WY

Colog, Inc. Golden, CO

Hudson Canyon Construction Millwood, NY

Figure 3. Organizational chart of parties involved in RHB-1 boring program.

6 INVESTIGATION SUMMARY The horizontal boring program at Roseton succeeded in its objective of collecting information on ground conditions around the tunnel and delineation of particular zones of interest in the vicinity of a steel lined section of tunnel suspected of having leaks. The boring program revealed good correlations in geologic conditions with original tunnel geologic mapping data and assisted in identification of key areas of interest with respect to possible leakage conduits. The various technical challenges faced in this boring program required the use of particular technologies and modified test methods to achieve the overall program objectives. In order to achieve the required

scientific outcomes, while balancing various technical and operational constraints surrounding this program, several contractual issues were addressed. The information gathered from this program will facilitate future studies of the tunnel and assessment of appropriate repair measures. ACKNOWLEDGEMENTS The authors of this paper would like to acknowledge the involvement of various individuals and companies (many of which are identified in Figure 3) in this technically challenging investigation. Site personnel from Boart Longyear, Scientific Drilling and Colog

379 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-46.qxd

2/18/04

2:02 PM

Page 380

provided invaluable contributions to the success of this investigation. Several PBQD and Malcolm Pirnie personnel provided on site efforts at various times over the duration of this project including N. Azzam, A. Benslimane, B. Bergeson, J. Choi, M. Chung, J. Dekoskie, J. Freitas, J. Glastonbury, S. Haq, S. Kota, S. Parry, C. Stewart and B. Russell.

The authors also appreciate the assistance and review offered by colleagues from PBQD Inc, Malcolm Pirnie Inc. and NYCDEP in preparation of this paper. Finally, the authors are grateful for the approval to publish details on this project provided by the New York City Department of Environmental Protection.

380 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-47.qxd

2/18/04

2:02 PM

Page 381

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Rehabilitation of the Big Walker Mountain Tunnel in Bristol, Virginia D. Kukreja Virginia Department of Transportation, Richmond, Virginia

P. Moran Dr. G. Sauer Corporation, Herndon, Virginia

ABSTRACT: The Big Walker Mountain Tunnels are a pair of two-lane highway tunnels built in the late 1960s and early 1970s and are operated and maintained by Virginia Department of Transportation (VDOT). Due to their age and condition VDOT contracted TAMS an Earth Tech Company to provide a structural rehabilitation design for the tunnels. The design included concrete repairs, replacing tiles, railings, curbs, cleaning the drainage system and chemical grouting the ceiling slab as well as repairing areas of the tunnel lining that were leaking. The Dr. G. Sauer Corporation (DSC), as a specialty subconsultant to TAMS, provided a tunnel lining and drainage rehabilitation alternative to the typical rehabilitation grouting. The design has the advantage that it will not adversely impact the existing drainage system unlike typical grouting operations. The key to the design is to understand DSC’s four main principals of rehabilitation which are explained herein. This paper will explain the history of the tunnels, the tunnel lining and drainage rehabilitation.

1 INTRODUCTION TO THE TUNNELS 1.1

History

In the 1950s when the Interstate Highway concept was passed into law, it included a “Great Lakes to Florida” route. One of the final segments of this link was the I-77 segment connecting I-77 in West Virginia to Virginia’s I-81. This highway includes 2 pairs of twolane horseshoe-shaped tunnels; one of which is called the Big Walker Mountain Tunnels and is located in rural Bland County in southwestern Virginia. The design for the Big Walker Tunnels and approaches was developed in the 1960s by Singstad & Kehart consulting engineers of New York City who employed a young tunnel engineer that would nearly 40 years later review the tunnel rehabilitation design as a Senior Tunnel Engineer of Federal Highways, Anthony S. Caserta, P.E. 1.2

Tunnel specifications

The Big Walker Mountain Tunnels, each with a length of 128 m (4229 feet), slope upwards towards the north on a 3.5% grade and are separated horizontally by 8.2 m (27 feet) of ground. Inside, each of the tunnels contain two, 4 m (12 foot) wide traffic lanes, a 79 cm (2.6 foot) wide sidewalk which houses utilities and a 47 cm (1.5 foot) wide ledge (see Figure 1). The roadway

has a 5 m (16.5 foot) vertical clearance above which is a 15 cm (6 inch) thick ceiling slab. The area above this slab is split in half with a vertical cast-in-place concrete wall and provides the exhaust and fresh air ventilation ducts. There are two cross passages connecting the tunnels with a clearance of 1.8 m wide by 2.1 m high (6 feet wide by 7 feet high). 1.3

381 Copyright © 2004 Taylor & Francis Group plc, London, UK

Tunnel cost versus traffic volume

The construction for the 37 kilometers (23 miles) of roadway, bridges and tunnels cost nearly $50 million and was divided into numerous contracts. The tunnel contract was awarded to C.J. Langenfelder & Son, Inc. of Baltimore, Maryland in September of 1967 in the amount of $22.6 million (Hanes 1973). After the tunnels and the highway were dedicated on June 23, 1972 and opened, the traffic volume was estimated at 5600 cars per day (VDOT 1975). Now more than twenty years later, the traffic has increased to 28,000 cars per day (VDOT 2002). With so many motorists using the tunnels combined with the unusual facts that motorists are permitted to change lanes while inside the tunnel and that there is no restriction on hazardous material hauling, it is extremely important to keep up with normal maintenance issues. Without proper and regular maintenance combined with rehabilitation as required,

09096-47.qxd

2/18/04

2:02 PM

Page 382

the tunnels become unsafe for the public and the initial investment is diminished.

2 TUNNEL DESIGN AND CONSTRUCTION In order to understand the problems affecting the tunnel today we first had to understand how the tunnels were constructed. The majority of the drill and blast tunneling was done through competent rock consisting of Sevier shale which was yellow and slightly calcareous to sandy overlain with Clinch sandstone which was coarse (Campbell 1896). The design prescribed five construction methods as listed below: 1. Roof bolts with channels. 2. Concrete and grouting of rock seams. 3. Structural steel support without lagging but blocked against rock, concrete and grout.

4. Structural steel support and 1.8 m (6 foot) steel lagging channels, dry stone packing to be grouted beyond the lagging. 5. Structural steel support and 1.8 m (6 foot) steel lagging channels, concrete and grout beyond the lagging. When seams were encountered, they were to be thoroughly cleaned out, grout and weep holes drilled, and the seam caulked then grouted. After the rock was supported, a continuous, reinforced concrete footing was constructed onto which a 51 cm (20 inch) thick reinforced concrete lining was cast in 11.4 m (37.5 foot) sections. At the end of every section is a 10 mm (3/8”) transverse expansion joint which consists of a 23 cm (9”) wide PVC waterstop, 10 mm (3/8”) premolded cork joint filler and a continuous formed open joint drain in which water is collected (see Figure 2). The joint drain is formed with half of a 7.6 cm (3”) asbestos cement pipe at the end of one pour and half

Figure 1. Big Walker Tunnel cross section.

382 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-47.qxd

2/18/04

2:02 PM

Page 383

of a 10 cm (4”) asbestos cement pipe left in place in the adjacent pour. Each open joint drain has a 7.6 cm (3”) cleanout above the sidewalk and ledge so the water from the open joint drain is deposited into a 10 cm (4”) cast iron drain pipe on both sides of the tunnels. The 10 cm (4”) drain pipe carries the water longitudinally about 1.8 m (6 ft) to a 1.2 m 0.9 m (4 ft 3 ft) pocket of gravel where the water is filtered before it flows through a stainless steel screen and into the 15 cm (6”) transverse drain (see Figure 3). The 15 cm (6”) drain flows to the 38 cm (15”) perforated clay pipe that runs longitudinally through the center of one lane of the tunnel (see Figure 4).

3 REHABILITATION 3.1

Inspection

Vincent F. Schimmoller, deputy executive director of the Federal Highway Administration (FHWA) is quoted as saying “We can’t simply design and build tunnels and expect them to take care of themselves. We must apply sound engineering and business principles to maintain our tunnels.” (Botelho 2001) In July 1986, a structural inspection was performed which located and documented numerous cracks in the tunnel lining and expansion joints that were leaking and had caused deterioration to areas of the tunnel lining and finishes. In fact during the winter, the active leaks frequently caused icy conditions near the tunnel portals therefore it was necessary for the tunnel maintenance staff to temporarily install heat tape to active leaks to prevent freezing. As a result of leaks that have existed for a period of time, adjacent wall tiles were delaminated (see Figure 5), the roadway pavement had deteriorated and formed potholes, the ceiling slab had deteriorated in areas and rebar was exposed,

Figure 2. Tunnel expansion joint section.

Figure 3. Drainage section showing gravel pocket.

Figure 5. Delaminated tiles in the Big Walker Tunnel.

Figure 4. Big Walker Tunnel drainage system.

383 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-47.qxd

2/18/04

2:02 PM

Page 384

Figure 7.

Ice buildup in Lehigh Tunnel No.1 vent duct.

Figure 6. Ponding on the ceiling slab.

all of which posed a safety hazard to the public. Above the ceiling slab and out of the public view, a pipe that drains water off of the ceiling slab was obviously clogged and created a pond on the slab (see Figure 6). This water eventually made its way down the sidewall and resulted in delaminated wall tiles and deteriorated concrete. 3.2

“Chasing the water”

The initial tunnel rehabilitation TAMS proposed to VDOT was to perform crack and joint grouting. Past experiences in many tunnels is that grouting may be successful in plugging and drying up existing leaks however typically new leaks appear later when the water has found a new path of least resistance. Grouting is performed on the new leaks and the cycle continues. This is called “chasing the water” and the cycle can be endless. The Big Walker Tunnels were constructed with a perimeter joint drainage system. If grouting was performed, there would be a risk of clogging the joint drains and gravel pockets with grout rendering the joints drainage system ineffective. Without a working drainage system there is a high probability that many more leaks would form. One of the most important issues is cost. Chasing the water necessitates a cost on a regular basis with no guarantee that the leaks will ever stop. The alternative to grouting and preventing the water from entering the tunnel is handling the water and giving the water a managed flow path. 3.3

Development of a rehabilitation alternative

The Dr. G. Sauer Corporation approached VDOT and requested a site visit of the Big Walker Tunnels. It was observed that the tunnel had some active leaks which may have been fed by active springs. DSC doubted whether grout could adequately plug the leaks and

convinced VDOT to allow them the opportunity to propose an alternative rehabilitation method. After reviewing the inspection reports and videos and noting the typical problems encountered with grouting operations, DSC proposed a sound and proven rehabilitation philosophy based on DSC’s four principals of rehabilitation which include managing the water and provided examples of where the system had been implemented successfully. The Felbertauren Tunnel in the Austrian Alps was the first tunnel where DSC’s rehabilitation alternative was successfully applied. Built in the late 1960’s through metamorphic rock with drill and blast construction methods, the Felbertauren Tunnel’s castin-place concrete lining was leaking along nearly 45% of its length shortly after completion of construction (Sauer 1987). In 1985, after grouting attempts failed to seal the leaking tunnel, DSC developed a method of controlling the water specifically tailored to the tunnel’s construction, the type of leaks and temperature influences that was implemented creating a tunnel which remains dry nearly 19 years later. The Felbertauren Tunnel rehabilitation design philosophy was utilized and adapted in 1994 for a leaking tunnel on the Pennsylvania Turnpike. The Lehigh Tunnel No. 1 was constructed in 1957 using drill and blast methods and was finished with a cast-in-place concrete lining. The lining was leaking so severely that in the winter, tunnel maintenance crews had to frequently remove large icicles in the ventilation ducts (see Figure 7). Since the rehabilitation was completed in 1995, the tunnel has been dry (Mergelsberg 1996). The rehabilitation method is successful if the design accounts for the geology, the tunnel construction and most importantly the water. 3.4

In order to properly design and implement the tunnel leak remediation, DSC’s four principals of rehabilitation

384 Copyright © 2004 Taylor & Francis Group plc, London, UK

Understanding the water

09096-47.qxd

2/18/04

2:02 PM

Page 385

Figure 8. Membrane installed in Lehigh Tunnel No.1 vent duct.

have to be addressed. 1. Where is the water coming from? 2. How can the water be drained? 3. How will the drainage system be maintained in the future? 4. How to prevent water from freezing and clogging the drainage system? 3.4.1 Where is the water coming from? Previous inspections of the Big Walker Mountain Tunnels documented that water was leaking through the existing expansion joints and through cracks in the lining. A large concentration of leakage was located within 120 m (400 feet) of the portals due to the geology and hydrogeology at tunnel portals. The leaks located at the ceiling slab joint are typical to most tunnels built through rock in the 50s through the 70s as the tunnel lining was constructed in stages with a joint at the ceiling slab. The water travels through the construction joint and shows up inside the tunnel under the ceiling slab. Also, as documented by video tape, some joint drains were clogged with calcium deposits and some of the grout pockets were clogged with concrete. It was not known how concrete made its way to the gravel pocket but the theory is that it occurred during the construction of the tunnel. The clogged joint drain in the expansion joints, forced the water to search for an alternative flow path which consisted of making its way around the existing 9” PVC waterstop and a 3/8” thick joint sealer (see Figure 2). It is important to intercept the water where it enters the tunnel either by drilling relief or collection holes and/or providing new joint drains. 3.4.2 How can the water be drained? Above the ceiling slab, in the ventilation exhaust and fresh air ducts, it is sufficient to simply cover the leaking tunnel concrete lining 384with PVC membrane. This prevents the water from dripping onto and

slowly damaging the ceiling slab and it also prevents icicles from forming and creating obstructions in the air ducts during the winter. On the ceiling slab where the slab meets the tunnel arch, a perforated drain pipe is wrapped into the PVC membrane and collects the leaking water. A hole is drilled through the ceiling slab and a pipe is grouted in place. Then the pipe is connected to the drainage pipe in order to convey the water below the ceiling slab where it ties into the existing joint drain or into a new vertical drain as described below. For tunnel lining leaks below the ceiling slab in joint locations, an area at the highest point of the leak must be opened and a section of the joint drain must be removed in order to test the joint drain to see if it is clogged. If the drain is not clogged then a funnel shaped mortar bed is formed around the top of the existing joint drain and 5 cm (2”) diameter inclined holes are drilled into the rock. These collection holes will convey ground water to the opened section of the lining. The funnel shaped mortar bed will direct the water to the existing joint drain. In the event that the existing joint drain is clogged, the joint drain must be removed to the top of the sidewalk or ledge and replaced with a 10 cm (4”) diameter pipe (see Figure 9). Due to the difficulty and cost of replacing the joint drain below the sidewalk and ledge, the existing joint drain must be capped and a slot for a horizontal drain pipe must be cut above the sidewalk or ledge to the existing gravel pocket. Although the drain pipe is termed horizontal, it is installed with a minimal slope to direct the water away from the joint and towards the existing gravel pocket. At the lower end of the future horizontal drain pipe, a 7.6 cm (3”) diameter sidewall diversion hole is drilled downward aiming for the existing gravel pocket. Care must be taken to document the location of the sidewall diversion hole as it may be necessary to drill from another direction to meet the diversion hole and drain the water if the gravel pocket is clogged (see Figure 10). The next step is to run water in the diversion hole to see if the water drains either into the gravel pocket or into the surrounding ground or whether the water leaks into the tunnel at any other locations. If the water drains properly then the vertical and horizontal drain pipes are installed. If the water does not drain, then an additional sidewall diversion hole can be drilled and tested. If the sidewall diversion holes fail the water test, then it is necessary to drill a lateral diversion hole. This hole is drilled from the existing lateral drain pipe through the gravel pocket then the sidewall diversion hole is retested with water. The lateral drain pipes are located near each expansion joint however a manhole does not exist at every lateral diversion pipe. So it may be necessary to open the roadway in order to access the lateral diversion pipe. If the water added at the sidewall diversion hole does not communicate with the lateral diversion pipe,

385 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-47.qxd

2/18/04

2:02 PM

Page 386

Figure 9. Proposed sidewall drainage system.

continue drilling additional sidewall diversion holes to meet the lateral diversion hole until the sidewall diversion hole passes the water test. Then install the vertical and horizontal drain pipes and connect with fittings to the sidewall diversion hole. 3.4.3

How will the drainage system be maintained in the future? This is one of the most important questions. If the new drainage system clogs with sediment, etc. it is important that the owner can clean and maintain the system otherwise another drainage system will have to be installed

and the previous one goes to waste. The new drainage system below the ceiling slab can be maintained by flushing water from the top of the vertical drain which is covered with an access panel and through the clean out located at the low end of the horizontal drain pipe which is also covered with an access panel. Above the ceiling slab, the new drainage system can be maintained at the cleanout which is installed at one end of the waterproofed area and is covered with an access panel. Drainage systems should be flushed on a regular basis to prevent calcification and sedimentation.

386 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-47.qxd

2/18/04

2:02 PM

Page 387

Figure 10. Proposed sidewall drainage below the sidewalk.

Figure 11. Vent duct waterproofing and insulation.

3.4.4

How to prevent water from freezing and clogging the drain system? Even a clean drainage system can become clogged if water is permitted to freeze in the pipes. Therefore, all of the access panels that are covering the vertical and horizontal drain pipes as well as the area at the

top of the vertical drain pipe where the collection holes drain, are thoroughly insulated. In the air ducts, the PVC membrane is also protected with rigid insulation and in addition, a heating strip is installed in the drain pipe that is wrapped in the membrane (see Figure 11).

387 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-47.qxd

2/18/04

2:02 PM

Page 388

after the rehabilitation of the Big Walker Mountain Tunnels has been completed, one entrance will be significantly drier and safer. Fortunately, for the client, the rehabilitation design can be easily modified and adapted for the tunnels exiting the County as well.

ACKNOWLEDGEMENTS It was a pleasure working with the open minds at the Virginia Department of Transportation and the Pennsylvania Turnpike Commission to develop a custom rehabilitation system that can be implemented as needed to deal with tunnel leaks. Also, thanks to Mr. Caserta for his structural engineering during the design of the Big Walker Mountain Tunnels and for his assistance during the tunnel rehabilitation more than 30 years later.

Figure 12. Big Walker Mountain Tunnels portal.

4 REHABILITATION CONSTRUCTION Due to the reduction in government funds which is common to all DOTs nationwide, rehabilitation of the Big Walker Mountain Tunnels has not been advertised for construction at the time this paper was written. We look forward to updating this paper in the future with the contractor’s experiences and recommendations.

5 CONCLUSION Old tunnels don’t have to leak nor do they have to be grouted and regrouted over and over to eliminate leaks. In the past grouting was performed on many leaking tunnels with limited success. Some leaks may be eliminated while new leaks appear later. Also it is difficult to predict the cost of a grouting operation therefore the cost for the rehabilitation is an open ended question for which funding becomes very difficult. It is better to work with the water than to fight the water by understanding where the water is coming from and providing a drainage path for the water that is maintainable in the future and that will not freeze. The simplicity of this rehabilitation method is that it only has to be performed one time in order to be successful. It has been said that Bland County is the only county in the US that you enter through a tunnel (Big Walker Mountain Tunnel) and leave through another tunnel (East River Mountain Tunnel). In the future,

REFERENCES Big Walker Tunnel Inspection videos performed by Parsons Brinkerhoff, 1986. Botelho, F. 2001. A Light at the End of the Tunnel, Public Roads, Vol. 65, No. 1, July/August 2001. U.S. Department of Transportation, Federal Highway Administration. Campbell, M. R. 1896. Description of the Pocahontas Sheet. U.S. Geologic Survey. Commonwealth of Virginia Department of Transportation, Average Daily Traffic Volumes with Vehicle Classification Data on Interstate, Arterial and Primary Routes, 1975. Commonwealth of Virginia Department of Transportation, Average Daily Traffic Volumes with Vehicle Classification Data on Interstate, Arterial and Primary Routes, 2002. Hanes, J. C. & Morgan, J. M., Jr. 1973. The History and Heritage of Civil Engineering in Virginia, Part II, The Big Walker Tunnel. Virginia Section American Society of Civil Engineers. Mergelsberg, W., Gall, V. & Sauer, G. 1996. Achieving Dry Cut-and-Cover Stations, North American Tunneling. Proceedings of the International Conference on North American Tunneling ’96 and the 22nd General Assembly of the International Tunneling Association Washington, D.C. 21–24 April 1996. Rotterdam: Balkema. Sauer, G. & Garrett, V. K., Jr. 1987. Achieving a Dry Tunnel, North American Tunneling. Proceedings of the International Conference on North American Tunneling ’87. Rotterdam: Balkema.

388 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-48.qxd

2/18/04

2:03 PM

Page 389

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Corrosion evaluation of the Manhattan rocks and corrosion protection of the rock reinforcement system for subway tunnels M. Ryzhevskiy STV Incorporated

M. Berman Parsons Brinckerhoff

ABSTRACT: Durability of any rock support/reinforcement system for tunnels, from the mechanical point of view, depends on the lifetime of the applied materials for this system. The lifetime of the applied materials is function of the resistance of these materials against different forms of corrosion. The materials most often used for rock support/reinforcement systems are steel (for rock bolts, dowels, ribs and lattice girders) and steel reinforced concrete or shotcrete. Therefore, corrosion risk and installation method normally govern selection of an appropriate rock support/reinforcement system. This paper provides a general overview of the corrosion evaluation of underground transportation structures, and provides recommendations for corrosion protection and service life prediction for the initial (primary) rock support/reinforcement system, such as rock bolts or dowels and steel-reinforced shotcrete, used for the Manhattan tunnels of the East Site Access project in New York City.

1 INTRODUCTION East Side Access (ESA) is a multi-billion-dollar project that will provide the connection of the Long Island Rail Road to Grand Central Terminal (GCT) on the east side of Manhattan. It will increase capacity for the commuter rail lines of the Long Island Rail Road (LIRR) and will provide direct access between suburban Long Island and the east side of Manhattan. It will include a new passenger terminal in the east midtown of Manhattan. The Metropolitan Transportation Authority (MTA) is the lead agency for this project. The East Side Access connection will be achieved by constructing a 4,600 ft tunnels from the LIRR Main Line in Sunnyside, Queens to the existing tunnel under the East River at 63rd Street. LIRR trains will use the lower level of this bi-level structure. A second 5,000-ft tunnel from 63rd Street will carry LIRR trains under Park Avenue and into a new LIRR terminal under the lower level of GCT. Construction began in 2000 and is to be completed in 2012. Construction of the Manhattan tunnels will require a special rock support/reinforcing system, which will predominantly consist of steel and reinforced concrete elements, such as rock bolts/dowels,

steel mesh/fiber reinforced shotcrete and steel lattice girders. The project design criteria require all tunnel structures to have a service life of 120 years. 2 BACKGROUND Rock bolts (tensioned rock reinforced elements) or dowels (intentioned rock reinforced elements) integrated with steel (mesh or fiber) reinforced shotcrete have been widely applied in tunneling as rock support/reinforced systems since the mid-1960s. They are performing well in a variety of environments and applications. The service lifetime of the new subway tunnels is expected to be 120 years. This requires appropriate design of the tunnel structural elements, based on modern corrosion protection methods, and development of a special installation procedure to eliminate or minimize damage of structural elements. The placement of any metallic materials below ground can result in a continuous process of corrosion. Metallic elements such as rock bolts, dowels, lattice girders will corrode because of the migration of ions from the surface, resulting in a reduction in

389 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-48.qxd

2/18/04

2:03 PM

Page 390

the thickness of the steel element over a period of time, and consequently reduction in structural capacity of the element. Ultimately, if the corrosion rate is high enough, the structural element may fail and lead to failure of the entire tunnel structure. The corrosion protection design should therefore be based on corrosion evaluation of the environment surrounding the tunnel. This paper provides recommendations for corrosion protection in the context of the initial (primary) rock support/reinforcement system, such as the rock bolts and dowels and steel-reinforced shotcrete used for the ESA Manhattan tunnels. Two forms of corrosion – pitting and uniform corrosion – most often affect rock support/reinforcement elements in tunnels and other underground structures. Stray current and chemical aggressiveness of underground water are the general cause of these forms of corrosion. Corrosion can be localized, taking the form of pits in the surface, or it can be more generalized, forming large areas of metal wastage. The corrosion can occur in rock in areas where joint-water can get in contact with the rock bolts/dowels through cracks in the cement or resin grout, if the steel bar of rock bolts/dowels has no additional protection, or where the cement or resin grout has not appropriately encapsulated the steel bar, and the steel bar is not provided with the double corrosion protection system. 3 FACTORS INFLUENCING CORROSION In order to adequately design a corrosion protection system for rock support/reinforcing elements to meet the requirements of the existing ground conditions, it is important to identify the environmental conditions subjecting the material to corrosion over time. To evaluate the aggressiveness of the ground and water the following factors should be analyzed: – – – – – – – – – –

Ground and water resistivity Groundwater contents Groundwater elevation (stable or fluctuating) Water permeability Water dissolved oxygen content pH value of the groundwater Chemical composition of the environment Stray-current presence Stress factor of the structural elements Range of temperature.

All of the above-mentioned factors play vital roles in the design of an appropriate corrosion protection system for rock support/reinforcing elements, and all need to be measured and evaluated. For this purpose an extensive ground investigation program has been conducted.

4 CORROSION STUDIES Manhattan tunnels cross predominantly rock formations. The rocks underlying Manhattan belong to the New England Upland, and are locally known as the Manhattan Prong. This rock consists of schist, gneiss and marble. The tectonic history of the rock has left the Manhattan Schist fractured and dislocated. The main rock types recovered from borings along the Manhattan alignment are metamorphic, dominated by schist and gneiss. The essential minerals are muscovite, biotite, quartz and feldspar (plagioclase, microcline and orthoclase). According to the “Practical Rock Resistivity Table” (Figure 1), the resistivity of this rock can range from 1,000 to 100,000 ohm-cm. Rock resistivity of Manhattan schist is determined by the entrapped water and the ions dissolved in it. The generally adopted corrosion severity ratings are 8,000 to 15,000 ohm-cm. Manhattan is slightly raised above sea level, and is bounded by the East River to the east and the Hudson River to the west. The area is heavily urbanized, with the exception of Central Park, and infiltration of rainfall is therefore low. More intense conductive fracturing will occur at the locations of the buried stream channels where the watercourses followed the weaker rock in the shear zones. These fractures will be conduits for the groundwater with much greater hydraulic conductivity than other fractures in the undisturbed rock mass. The groundwater levels measured in observation wells range from 4.5 m below the street level along Park Avenue to less than 1.5 m below the invert of the existing lower level of Grand Central Terminal. The permeability was determined from in-situ packer testing, and varies from 107 m/sec to 104 m/sec. The sources of groundwater recharge in Manhattan are surface infiltration, leaking sewers, drains and water lines, and the adjacent East River and Hudson River. On a regional scale surface water bodies also include the Harlem River and New York Bay. The network of fractures will control the groundwater conditions for the rock mass. The permeability of the discontinuities will be influenced by several factors, including the intimacy of adjacent surfaces, alteration processes that have removed or placed minerals on fracture surfaces, and joint wall material that has been fragmented or crushed by faulting and shearing. The groundwater chemical test results generally indicate a moderately low concentration of chlorides contents, which ranges from 60 to 100 mg/l, and rarely up to 550 mg/l, and a low concentration of sulfates contents, which is 40 mg/l and rarely ranges from 120 to 180 mg/l. pH values classified as neutral and vary from 6.17 to 6.9. The concentration of dissolved oxygen is perhaps one of the most important factors influencing the rate of corrosion for all structural elements. At ordinary

390 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-48.qxd

2/18/04

2:03 PM

Page 391

Alluvium Clay Shale Sandstone Quartzite Schist Gabbros Granite

01E + 00

01E + 02

01E + 04 Resistivity, Ohm.m

01E + 06

01E + 08

Figure 1. Rock resistivity – overview.

temperatures, the absence of dissolved oxygen will greatly slow down the corrosion process. When oxygen has access to a moist metal surface, corrosion is promoted. The test results show the presence of dissolved oxygen in the groundwater also. The content of dissolved oxygen in the groundwater ranges from 9 to 10 mg/l. Other found components of the groundwater, such as CaCO3 and dissolved solids, do not play a significant role in the estimation of environment corrosivity. The specific conductance of groundwater itself ranges from 650 to 2,000 mhos/cm. The specific conductance data in mhos/cm was converted to resistivity in ohm-cm. A resistivity datum is one of the most important factors to evaluate soil aggressiveness. Lower resistivity enables corrosion reactions to occur more easily. Groundwater samples show the resistivity below 2,000 ohm-cm, which usually indicates a high corrosion activity. In addition to the ground and groundwater chemical studies, a stray current survey was conducted in Grand Central Terminal’s upper and lower levels and Metro-North Railroad’s existing tunnels. Stray direct currents (dc) are defined as currents that leak from electric power circuits and flow through the earth and in other adjacent metallic structures. In the area where stray currents leave a metallic structure and flow into the earth, corrosion occurs. In order to test for the presence of stray currents, it is common to measure the voltage (potential) of a structure. This is done by using a high resistance voltmeter with one terminal connected to the tested structure and the second terminal to a copper-copper sulfate

reference electrode. This electrode provides a stable reference point and is widely used in the corrosion/ cathodic protection industry. The voltage pattern of structures that may be subject to stray currents from transit systems is distinctive and will clearly indicate whether that structure is being subjected to transit stray currents. Because the resistance-to-earth of the structure is constant, voltage variations must be due to fluctuating of stray currents (by Ohm’s Law: V  IR). The results of these voltage measurements can indicate the presence of stray currents if the readings fluctuate. The results of this survey showed different stray current levels on underground structural elements, with an average stray current activity of 140 millivolts with a variation from 0 to 1,600 millivolts. Dynamic transit dc stray currents can cause severe corrosion within a very short period of time. One ampere of stray current discharging into the electrolyte (soil or groundwater) will remove twenty pounds of steel in a period of one year. If, for example, just 5 milliamperes flowed off a 1-inch diameter rock bolt/dowel, in one year 1.6 oz. of metal would be removed. From a few milliamperes up to 20 amperes were measured in different locations during the stray current survey. 5 EVALUATION OF THE ENVIRONMENT CORROSIVITY In order to predict the environmental corrosivity the Project Corrosiveness Model was developed, allowing us to obtain the Numerical Corrosiveness Index based

391 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-48.qxd

2/18/04

2:03 PM

Page 392

Table 1. Assessment of numerical corrosiveness index.

Test type

Rate

Ground features

Resistivity, ohm-cm

500 500 to 1,000 1,000 to 2,000 2,000 to 3,000 3,000 0 to 2 2 to 4 4 to 8.5 8.5 1,000 500 to 1,000 500 1,000 500 to 1,000 500 300 100 to 300 10 to 100 10 10 5 to 10 1 to 5 1 Integral index

Very corrosive Corrosive Moderately corrosive Mildly corrosive Progressively less corrosive Very corrosive Corrosive Progressively less corrosive Corrosive Very corrosive Corrosive Progressively less corrosive Very corrosive Corrosive Progressively less corrosive Very corrosive Corrosive Mildly corrosive Progressively less corrosive Very corrosive Corrosive Mildly corrosive Progressively less corrosive

pH

Sulfates, mg/l

Chlorides, mg/l

Stray current Activity, mV

Dissolved oxygen, mg/l

Table 2. Classification of ground corrosivity using the numerical index.

Index A 13 9 to 12 5 to 8 4

Ground features Highly corrosive Corrosive Moderately corrosive Slightly corrosive

Average velocity of corrosion, mm/year

Classification

1

I

0.1–1 0.01–0.1

II III

0.01

IV

on environmental characteristics obtained from the test results (Table 1). Based on the overall Numerical Corrosiveness Index, the ground corrosivity can be defined (Table 2). According to Table 2 presented above and the available geological data from the Manhattan tunnel investigation, the Numerical Corrosiveness Index can be calculated as following:

ASTM G57 or EPA Method 9050A

ASTM G51 or EPA Method 9045 EPA Method 9038A EPA Method 9252 24-Hour voltage Recording

Index 10 8 5 2 0 5 3 0 3 5 3 0 5 3 0 8 6 3 0 5 3 1 0 Sum of above: A

Thus, the site can be considered highly corrosive. The tunnel’s structural elements in this environment require special corrosion control measures to be applied. 6 CORROSION PROTECTION MEASURES Corrosion protection measures for the rock bolts/ dowels are one of the most controversial in tunnel engineering. Different tunnel designers recommend different corrosion protection methods to extend the lifetime of rock bolts/dowels. For the rock reinforcing elements of the East Side Access tunnels, due to its expected life and to minimize future maintenance, the following factors were considered: – The designed tunnels are considered first category value underground structures with a lifetime of over 120 years. – The chemistry of the groundwater is not wholly known, but the water aggressiveness in some areas is higher than average and stray current presence makes the environment highly corrosive. – Failure of any structural element is very critical from both personal safety and economic points of view.

392 Copyright © 2004 Taylor & Francis Group plc, London, UK

Test Standards

09096-48.qxd

2/18/04

2:03 PM

Page 393

Recommendations of the American Post Tensional Institute (PTI) have been considered and have been applied for the development of the Manhattan tunnel reinforcement system. PTI recommends installing the rock bolts/dowels a with double corrosion protection system for the corrosive environment. Corrosion protection can be achieved when one or more reactions of the corrosion cell are prevented: the anodic, the cathodic or the electrolytic. This means that no humidity and/or no oxygen and/or no carbon dioxide and/or no chloride ions may arrive at the steel surface. The corrosion protection of the tunnel supporting elements, as well as protection of the entire supporting system, must meet some fundamental requirements: – No adverse effect on the strength and/or ductility of the structural elements – Compatibility to each other with respect to their physical and chemical characteristics – Resistance to possible mechanical damages during installation and service – No adverse effect to the environment – Durability for the required life of the structure – Practical and easy applicability – Controllability – Economical during construction and maintenance – Particular requirements for special characteristics, e.g., bond to steel and to grout. With respect to these requirements, the seven main methods of ensuring durability of the rock bolts/dowels in different corrosion environments are analyzed. These methods include: 1. Creation of alkaline environment (by grouting) 2. Encapsulation in injected material (cement or resin) 3. Bitumen tape 4. Corrugated plastic sheathing (extruded polyethylene) 5. Provision of sacrificial additional thickness of the steel bars 6. Hot dip galvanizing 7. Epoxy or polyurethane coating. The first two of the above-mentioned methods can be used only during rock bolts/dowels installation and provide an additional (second) layer of corrosion protection. Methods include injection into the bored holes before, during or immediately after the rock bolts/ dowels installation of cement or resin-based grout. If the injection pressure is less than two bars, the grout fills only the inner space of the bored holes and wide cracks around them, and encapsulates the steel bars of the rock bolts/dowels. The injection pressure higher than 2 bars can create a more effective corrosion protection layer by filling most cracks and joints around the boreholes, and preventing any humidity around rock bolts/dowels.

The third and fourth methods are more appropriate for soil application and rarely are used for rock bolts/dowels. By increasing the rock bolts/dowels diameter, it is possible to create a sacrificial area for the steel bar if corrosion occurs. This method cannot be recommended in a highly corrosive environment, because it requires rock bolts/dowels that are too heavy, and drilling larger diameter holes, which are much more expensive. In addition, this method requires confidence in the ability to predict corrosion rates under non-homogeneous field conditions. Similar concerns exist for galvanizing. Zinc coating may slightly increase the life of steel bars, and the galvanic effect of zinc may cathodically protect the steel surface of the bar, but it does not guarantee the service life of tunnel support elements in the long term, especially in the presence of stray currents. Within the threaded length the zinc coating may be of minimal thickness. In addition, chipping of the zinc coating during bar handling and installation can occur. Epoxy or polyurethane coating can provide suitable protection. Epoxy or polyurethane coating theoretically has a long life in any kind of rock formations, often in excess of the service life of the structure. However this is often reduced by local corrosion due to surface damage or poor application. It is possible to eliminate or significantly reduce this by encapsulating coated rock bolts/dowels into cement/resin-based grout. Cement or resin grout is considered acceptable as an impermeable protective encapsulation, provided that the crack width within a grout cover can be demonstrated not to exceed 0.1 mm. The grout must be high quality. Cement grout is unlikely to provide complete isolation of the steel bar from the environment, and may be classified as partial protection that delays the onset of corrosion. Resin grout with excellent dielectric characteristics is the best choice for protection of rock bolts/dowels in the corrosive environment.

7 CONCLUSIONS AND RECOMMENDATIONS To evaluate the corrosiveness of the Manhattan tunnels environment, a Numeric Corrosiveness Index was designed. This table may be successfully used to evaluate the corrosiveness of any underground environment after a corrosion survey is conducted. Based on the corrosion evaluation of the Manhattan tunnels, the following steps were implemented in the tunnel supporting system design: – Corrosion analysis of the water samples identified the Manhattan tunnels environment as corrosive. But in combination with stray current activity, this environment may be considered highly corrosive.

393 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-48.qxd

2/18/04

2:03 PM

Page 394

– The temporary rock bolts or dowels for the Manhattan tunnels (serving for two years or less) should be fully and properly resin-grouted. – The rock bolts or dowels that will be used for more than two years should have an epoxy or polyurethane coating, encapsulated in resin grout. – Epoxy-coated rock bolts/dowels with a factoryapplied epoxy coating in accordance with ASTM 775 are recommended.

REFERENCES Recommendations for Pre-stressed Rock and Soil Anchors 1986, American Post Tensioning Institute Ryzhevskiy M., 1989, The organization of the application of the resin rock bolts in underground construction, Mine Journal, N12, Moscow

394 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-49.qxd

2/18/04

2:04 PM

Page 395

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Rehabilitation of the Amtrak Long Island City ventilation structures S.G. Price Granite Halmar Construction Company, Inc.

ABSTRACT: In December 2002, Granite Halmar was awarded a contract by the National Railroad Passenger Corporation (Amtrak) for the demolition and reconstruction of the ventilation structures in Long Island City, NY. The project is located in the southern portion of Queens NY along the East River. The project is a part of a general Fire and Life Safety upgrade for the tunnel system into New York City’s Pennsylvania Station. The primary purpose is to improve ventilation and emergency egress from the tunnels. The project involves demolition of the nearly 100-year-old ventilation structures down to the track level, reconstruction of the new ventilation shafts, and the construction of a new ventilation building and electrical substation. The major challenge of the project is that Amtrak and the Long Island Railroad will continue to operate more than 400 trains per day through the tunnels during demolition and construction. Scheduled completion is early 2007.

1 INTRODUCTION The National Railroad Passenger Corporation (Amtrak) is currently engaged in a Fire and Life Safety Program to upgrade emergency egress, ventilation, and fire fighting capabilities in its nearly 100-year old infrastructure. Central to this upgrade are the tunnels and ventilation structures approaching New York City’s Pennsylvania Station. These routes, two tunnels under the Hudson River to New Jersey and four tunnels under the East River to Queens, NY, complete the link of Amtrak’s Northeast Corridor, which provides service from Boston to Washington, D.C. In December 2002, Granite Halmar Construction Company, a subsidiary of Granite Construction Company (Watsonville, CA) was awarded a $66.2 million contract for the demolition and reconstruction of the Long Island City Ventilation Structures. Notice to proceed was given in January 2003, and construction will be completed in 4-years. 2 THE FIRE AND LIFE SAFETY PROGRAM Amtrak has undertaken an aggressive program to improve safety and firefighting capabilities in the tunnels and shafts around the Pennsylvania Station. The facilities are a collection of abandoned electrical systems in various stages of disrepair after having undergone several modifications over the past 100-years.

Figure 1. Abandoned electrical systems.

The system has been prone to electrical fires, which has raised concern for firefighting capabilities in the tunnels. Existing fire lines extend less than 100-ft from the ventilation shafts. In the event of an emergency, passengers would need to evacuate more than 60-ft to the surface via a series of narrow and poorly lit spiral stairways. The Fire and Life Safety Program involves several contracts including ventilation upgrades at Penn Station, reconstruction of the ventilation structure in Weehawken, NJ, installation of fire standpipes in the Hudson and East River Tunnels, demolition and reconstruction of the ventilation structures in Long

395 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-49.qxd

2/18/04

2:04 PM

Page 396

Figure 2. Existing spiral stairway.

Figure 3.

Existing north ventilation building.

Island City, Queens, and a future contract to demolish and reconstruct the ventilation structures on 1st Avenue in Manhattan. 3 THE LONG ISLAND CITY VENTILATION STRUCTURES Amtrak and the Long Island Railroad use the Long Island City Ventilation complex jointly. Combined, they operate more than 400 trains per day into and out of Pennsylvania Station. Since there are no alternative train routes, all construction activities must be designed to support continuous operations of these railroads and emergency egress for passengers. Work at the tunnel levels is allowed only during scheduled weekend track outages. All other work must be carefully scheduled and protected from train operations. The work at Long Island City involves demolition of the existing ventilation buildings, substations, and ventilation shafts and construction of new ventilation shafts. After the shafts are reconstructed, a two-story ventilation structure will be built over the entire site to serve all four tunnels. The existing ventilation structures are two story brick buildings that were constructed on top of two rectangular caisson foundations. Each ventilation structure services two tunnels. The caissons were initially used as construction shafts for the East River tunnels into Pennsylvania Station. Each caisson structure is a 70 35-ft with 4-ft thick steel lined walls. The caissons extend through the 40-ft deep overburden and continue through bedrock to a total depth of about 80-ft below the surface. The ventilation shafts, sumps, pump rooms and stairways are concrete structures that were constructed within the caissons. There are also two electrical substations on the site that supply power to the third rail for the LIRR trains.

Figure 4. Existing north substation.

4 TEMPORARY ELECTRICAL SUBSTATIONS The East River tunnels have two existing electrical systems for train operations. The Amtrak trains run off an overhead catenary system that is powered by remote substations at Penn Station and the Sunnyside Yards. Long Island Railroad trains run off a third rail system that is powered locally by two substations on the project site. Two new substations will be housed within the new ventilation structure. However, since the work involves demolition of the existing substations, the contract requires installation of temporary substations in order to maintain continuous service throughout construction. The temporary substations consist of four trailer units manufactured for the LIRR by Impulse. Since there is no room for the temporary substations on the cramped construction site, the substation trailers had to be installed on LIRR property directly across the street from the project.

396 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-49.qxd

2/18/04

2:04 PM

Page 397

be installed offsite on the LIRR property and overhead wires run across Second Street and through the future ventilation building. The temporary C&S systems will be removed after the permanent equipment is installed within the new ventilation building. 6 TEMPORARY TUNNEL SHIELDS

Figure 5. Temporary substations.

Power cables will then need to run across Second Street and through the construction site to each of the four vent shafts. The eighty-four 2000 Kcmil temporary cables are required to run from the temporary substations to splice boxes at track level in each of the four tunnels. The problem is that once installed, these temporary cables cannot be relocated and must remain in service until the permanent substations are operational. Therefore the temporary cables must be routed so that they do not interfere with future demolition and construction of the vent shafts and ventilation building. This will require careful placement in order to allow construction of piles, pile caps, grade beams, slabs, columns, walls and permanent cables trays. To accomplish this, Granite Halmar and electrical subcontractor Mass Electric designed an elevated cable tray to carry eight-four 2000 Kcmil electric cables over Second Street and through the middle of the future ventilation building. Once the temporary substations are installed they must be operated for 60-days before the existing substations can be decommissioned and demolition can begin. In order to assure continuous train service during this testing phase, the system has the capability to quickly switch back to the existing substations in case of a failure of the temporary substation.

5 TEMPORARY COMMUNICATIONS AND SIGNAL WORK The future ventilation building will also house communication and signal (C&S) systems used by Amtrak and LIRR operations. Verizon also uses the facility for communication lines between Queens and Manhattan. Similar to the electrical systems, the C&S systems must remain in continuous operation during construction. Therefore temporary C&S equipment must also

Except for scheduled track outages, continuous train service must be maintained throughout demolition and construction. Amtrak and LIRR, the two largest passenger railroads in the country, currently operated more than 400 trains per day through the East River tunnels at speeds of about 60-mph. On average this means that each of the four tunnels handles one train every four minutes. Trains approaching and leaving the ventilation shafts also create tremendous air blasts that must be controlled before work can begin in the shafts. In order to isolate moving trains from the work and control the air blasts, the contract requires temporary shields to be installed in each of the tunnels. The shields are 30-ft long airtight canopies that fit between the caisson walls at track level. The trains pass through the canopies without disturbing the work in the shafts. The contract required that the shields be designed to withstand 85 psf loading. The shield design in the contract was a steel frame that was capped by an arch made from liner plates. Granite Halmar’s modified design was rolled steel channels and 1/2-in thick steel skin. Installation of the temporary shield involved several operations that took place over a series of weekend track outages. A typical weekend outage starts about 11 PM on Friday night. However, before work can begin, Amtrak forces must install grounding devices and lower the overhead catenary. Therefore work usually cannot begin until about 2 AM on Saturday morning. Granite Halmar then has continuous access to the track until about 2 AM on Monday. Amtrak forces then have until 5 AM to re-hang the catenary and restore service. Primary access to the tunnels for weekend work is by work trains that enter the tunnels at the Sunnyside Yard. The tunnel portals are about 1/2 mile from the Long Island City vent shafts. The first operation was to remove the fresh air flue at the bottom of the shaft. The flue was essentially a 30-ft long steel arch at the bottom of the shaft. As originally constructed, the flues were nozzles to divert fresh air supplied by the low velocity fans in the existing ventilation buildings. Flue demolition was complicated due to the proximity of the overhead catenary and various communication and signal lines that run along the tunnel walls. These utilities had to be continuously protected and relocated during the flue demolition.

397 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-49.qxd

2/18/04

2:04 PM

Page 398

start. Once the electrical systems are cleared from the ventilation shafts, each shaft will be sealed with an 18-ft diameter steel plug. Although the passing trains are isolated from the work by the previously installed shields, the shields alone are not capable of protecting the trains from the potential of falling concrete debris during demolition. The steel plugs will be designed to withstand a 400 psf distributed, or 2000 lb concentrated impact load. During shaft demolition, the DC power, communications and signal cables that were previously installed in the fresh air flues will be exposed. Therefore it will be necessary to support and protect these cables as demolition progresses. A system of pipe struts will be used to support the caisson walls. The struts will also support temporary fire standpipes and discharge lines. The primary concerns during all phases of construction are to maintain continuous train service and emergency egress from the tunnels. During demolition, the existing spiral stairways will be maintained until demolition reaches the tunnel level. Then a temporary stair tower will be installed and the spiral stairs will be demolished. About 2,600 CY of concrete will be demolished and removed from each caisson.

Figure 6. Flue demolition.

8 CONSTRUCTION OF NEW VENTILATION SHAFTS

Figure 7. Completed tunnel shield.

In order to maintain clearance for the train envelope, the shields were designed to the same radius as the tunnel. Therefore, it was necessary to remove about 9-inches of concrete from the tunnel arch at the bottom of the shaft. Concrete was also removed from the tunnel walls in order to make room to relocate fire lines and discharge lines. Concurrent with demolition work, it was also necessary to protect, relocate and restore existing communication, signal and catenary systems in the tunnels. After the shields were installed, they were then grounded and made airtight with a fire resistant sealing material. Fiberglass panels along the crown of the shield prevent arcing from the overhead catenary. 7 SHAFT DEMOLITION After the temporary substation has been successfully operated for 60 days the existing substations and systems will be decommissioned and demolition can then

Four new 16-ft by 18-ft ventilation shafts will replace the 18-ft diameter shafts. The new configuration will also allow for four new 18-ft by 8-ft utility shafts that are isolated from the tunnels. This will allow maintenance and repairs to be done without interruption of train service. The primary egress from each pair of tunnels will be through a pair of main stairways to the center bench wall. A stairway and crossover from the outer bench walls will also connect to the main stairway. The new stairways will be 4-ft wide. 9 CONSTRUCTION OF NEW VENTILATION BUILDING The new ventilation building is a two story concrete structure. Where there were originally two smaller ventilation buildings, a single building will now serve the four tunnels. The building foundation is comprised of 160 micropiles, 28 pile caps, and interconnecting concrete grade beams. The 10.75-in diameter micropiles will be augured to an average depth of 41-ft. The building will have two 20-ft by 20-ft valve rooms. Each room will contain control valves for the fire lines for two tunnels. Excavation for the valve rooms will be about 13-ft below grade and therefore

398 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-49.qxd

2/18/04

2:04 PM

Page 399

will require support of excavation and dewatering. Because they will be below the water table, the valve rooms will have a membrane waterproofing system. The first floor of the ventilation building will house two new 2500 kva substations, communication systems and signal systems. The substations will provide electric power to the LIRR third rail system. The second floor of the ventilation building will house three new 350 horsepower fans. The fans will be reversible and each capable of providing 228,000 cfm of airflow. The contract also includes a SCADA system to coordinate alarms systems and control of fans and dampers. The SCADA system will be integrated into the central system at Penn Station. When a fire is detected in the tunnels, the SCADA systems will have the ability to automatically set fans and dampers at all ventilation shafts so that smoke will be drawn away from evacuating passengers. The New York Power Authority (NYPA) holds a permanent easement through the new ventilation building for a pair of 345 kV power lines. The cables are buried in thermal insulating sand. The contract includes construction of new reinforced concrete trench walls that are supported on micropiles. Granite Halmar will then replace the thermal sand and the trench will be capped with precast concrete planks that are integrated into the building floor slab. This will allow for easy access for NYPA to maintain or repair the 345 kV cables after the ventilation building is completed.

The project environmental report identified a significant quantity of asbestos that was used in insulation, conduits and roofing materials. The existing substations are known to contain PCB contaminated oil that must be abated prior to demolition. There are also significant quantities of lead pain throughout the site. The Contract anticipated the likelihood of encountering contaminated soils. Therefore all excavated materials must be segregated and tested. Contaminated soils will be appropriately disposed of. Dewatering will be required for excavation of the valve rooms and pile caps. Therefore, the potential for groundwater contamination also exists. Groundwater sampling and testing is required as part of the permitting process. Contaminated groundwater will be treated before it is disposed of. Several treatment options are available, depending on the nature of the contamination. 11 PROJECT SCHEDULE Installation of the temporary substation will be completed in April 2004 and the 60-day redundant test will conclude in June 2004. Demolition of the ventilation buildings and shafts will be completed in early 2005. The new ventilation building structure will be completed in 2006. Installation of the mechanical and electrical systems will be done in 2006. 12 CONCLUSION

10 ENVIRONMENTAL ISSUES Environmental hazards are a common problem in infrastructure rehabilitation. Materials that were commonly used in construction 100-years ago are now recognized to be sources of contamination and health risks.

When completed, the Fire and Life Safety Program will significantly improve passenger rail service along the busiest corridor in the nation. The work involves many construction challenges that will require extensive engineering and planning in order to complete these projects without jeopardizing existing structures and yet maintain service continuity.

399 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-50.qxd

2/18/04

2:09 PM

Page 401

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Using seismic tomography and holography ground imaging to improve site investigations E.J. Kase & T.A. Ross NSA Geotechnical Services, Inc., Golden, Colorado USA

ABSTRACT: Unforeseen, variable subsurface ground conditions present the greatest challenge to the heavy construction and civil engineering industries in the design, construction, and maintenance of tunnel projects. A detailed, accurate site investigation will reduce project risk, improve construction performance and safety, prolong the life of the tunnel, and prevent waste in over-design. Presently, site characterization and geotechnical engineering are limited by the inability to adequately describe these subsurface ground conditions. NSA Geotechnical Services has successfully applied seismic tomography and holography ground imaging technologies on tunneling and heavy civil excavations worldwide. Tomography and holography are proven inversion technologies for estimating location and extent of material property variations causing changes in signal waveforms. This paper presents various applications of these technologies, illustrating how seismic imaging can provide accurate information regarding ground conditions associated with tunneling projects. With this information, engineers can complete projects safely, within time and budget constraints.

1 INTRODUCTION One of the key technology issues for the tunneling industry is accurately identifying and characterizing geologic and geotechnical risks on projects of all sizes. Unforeseen, highly variable subsurface ground conditions and obstacles present serious challenges to the heavy construction and civil engineering industries to design, engineer, construct, and maintain the massive infrastructure supporting the world economy. Hundreds of millions of dollars are spent on geotechnical contingencies supporting major construction projects. Efficient, economic, and safe excavation of tunnels depends critically on a detailed understanding of rock conditions to be encountered at the working face. In most countries, current practice usually involves completion of a Geologic Baseline Report (GBR) during the design phase of the project that is included in the bidding documents. Incomplete information in the GBR or the possibility of encountering differing conditions during excavation can only be discovered by examination of the rock mass as it is revealed at the working face. Using only the GBR database to determine ground support needs or to project the geology and structure ahead to anticipate what may be revealed in the next excavation cycle often leads to surprises. These can take the form of unanticipated

joints, faults, shear zones, karst conditions, or other changes in the structure that could lead to potentially hazardous conditions causing work delays and resultant claims and disputes. Detailed and accurate knowledge of ground conditions reduces project risk and expense associated with “change of conditions” claims or litigation, improves construction performance and safety, prolongs structure life, and prevents excessive waste in over-design, to name just a few of the benefits. All forms of major construction projects are impacted, including tunneling and microtunneling (transportation/utility), foundations and structures (buildings, piers, retaining walls, etc.), bridges and highways (abutments, roadcuts, slopes, etc.), dams, and impoundments. The methods used predominately today for site characterization and geotechnical engineering supporting these construction activities are greatly limited by the inability to adequately describe the physical and engineered attributes of the groundmass. For example, core drilling describes only a small portion of the ground volume, with subsequent laboratory investigations testing only those samples surviving acquisition, handling, and preparation. This often results in grossly skewed estimates of rock and soil strength conditions, further requiring arbitrary scaling for design based on empirical experience and intuition.

401 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-50.qxd

2/18/04

2:09 PM

Page 402

New methods of risk identification during site investigation and the active tunnel excavation can have a significant impact on the competitive advantage of contractors on overall project costs for the owner by mitigating these risks. To overcome the inability to accurately characterize the three-dimensional groundmass, NSA Geotechnical Services, Inc. (NSA) provides seismic tomography and holography imaging to characterize ground conditions associated with a civil construction or mining project.

2 SEISMIC CROSS-HOLE TOMOGRAPHY 2.1

Theory

Seismic tomography is based on the principle that acoustic waves have different propagation velocities and attenuation through different ground types. That is, seismic waves travel faster and have less attenuation in strong, competent material and slower velocity with greater attenuation in weaker materials (e.g. voids, broken or weathered rock, soil). Velocity tomography images represent the ground velocity as measured between seismic sources and receivers. Attenuation tomography images represent the relative attenuation rates within the surveyed area. There are numerous factors that may cause variations in velocity and attenuation. Different ground types usually have different material/seismic properties, but variations within the same ground type are also commonly encountered. Variations in stress, fracture extent, water saturation, soil compaction, etc., all may have a significant effect on velocity and attenuation. In areas where geological features such as fracture zones, faults, subsidence zones, or cavities exist, the seismic waves may travel at a lower velocity,

or may travel across an increased distance to pass around the anomaly and suffer increased attenuation. The same type of behavior may be noted in rocks of varying lithology as harder, more competent materials propagate seismic waves at higher velocity and lower attenuation than softer, less competent or less consolidated rocks. 2.2

Bridge foundations – LACSD Joint Outfall Replacement Sewer

NSA conducted a cross-borehole seismic tomography survey along the proposed alignment of the LACSD Joint Outfall “H” Unit 1B Replacement Sewer Tunnel Alignment, Section 1, LACSD, where it crosses under Atlantic Avenue. The primary purpose of this survey was to determine if the pilings for nearby Atlantic Avenue bridge piers cross into the proposed tunnel alignment. The images indicate that the ground under the Atlantic Avenue Bridge through which the tunnel was projected to pass was not obstructed by bridge pier pilings. There were two higher seismic velocity regions within the imaged area under the Atlantic Avenue Bridge, but outside the proposed tunnel alignment that could be indications of concrete pilings. One was on the west side of the tunnel, between 1820 and 1850 and a minimum of 12.5 feet (4.8 meters) from the proposed tunnel centerline (see Figure 1). The other was on the east side of the tunnel between 1780 and 1800 and a minimum of 15 feet (5 meters) from the proposed tunnel centerline. There were also some higher velocity zones below the tunnel invert that probably indicate harder and/or compacted ground. No bridge foundations or other obstructions were encountered during construction of the tunnel.

Velocity (ft/s) 2500 -30

2000 -15

1500

0 15

25 17+10

1000

500 17+50

18+00

18+50

Approximate edge of overpass

Figure 1. Plan view seismic velocity tomograms at 47.5 feet elevation.

402 Copyright © 2004 Taylor & Francis Group plc, London, UK

19+00

09096-50.qxd

2.3

2/18/04

2:09 PM

Page 403

Subsidence – Smithsonian Museum

NSA conducted a cross-borehole seismic tomography survey to investigate subsurface conditions contributing to surface depressions at the Washington D.C. Smithsonian Museum of History & Technology playground. Review of the horizontal tomographic slices revealed that relatively low ground velocity zones parallel the Metro tunnel through the imaged area indicating the presence of weaker, lower density soils or ground. Some of the weaker ground extended to as much as 30 feet (9 meters) below the surface elevation, which was approximately equal to the elevation of the tunnel invert. Some of these weaker ground zones that extend to the tunnel invert correlated with zones that have been documented as depressions or sink holes on the surface; however, there were additional “deep” weak zones that have either not manifested in surface depressions or have not been documented. From the image in Figure 2, it appeared that the weak ground zones and surface depressions might be a consequence of the Metro tunnel construction. This was based on the fact that the two zones on either side of the tunnel parallel the tunnel alignment and do not extend beyond the tunnel invert. 3 SEISMIC REFLECTION TOMOGRAPHY 3.1

Theory

Seismic reflection has been used in the petroleum industry for over 50 years. Due to advances in computer technology in the past few years, its application for shallow engineering studies has dramatically increased. Reflection surveys conducted from the surface are now a common tool used in both civil and mining applications. The basis of the technique lies in the existence of contrasts in acoustic impedance (the product of density and seismic velocity) between different geologic units. Where such contrasts exist, seismic energy incident on the boundary is partially reflected or scattered. These reflections may be

Figure 2. Perspective view of imaging site showing lowvelocity zones, with forward model.

detected with appropriate sensors, and the data analyzed to supply the seismic velocity of the material and the location of the reflecting boundaries. The TRT™ method uses a three-dimensional array of receivers installed around the tunnel at some distance behind the tunnel face. The placement of receivers is predetermined by the tunnel geometry and forms the basis of the TRT™ data acquisition system methodology. Up to 10 pre-amplified accelerometers are placed in a predetermined pattern on the rock surface in the vicinity of the probes. The signal produced by a hammer blow or small explosive charge is collected in a standard 24-channel seismograph. The direct waves, as well as the waves reflected from anomalies ahead of and around the tunnel are detected by the receivers in the array and are used to build a velocity model and to image ground anomalies. Once collected, the data are transferred and processed on a laptop computer. A three-dimensional reflection tomographic image of conditions ahead of the tunnel face is produced. As the tunnel excavation advances, the array is reinstalled, and data acquired. The entire data collection process takes one to two hours and is conducted every 80 meters to 100 meters of tunnel advance to maintain complete coverage of the drivage. The velocity model is defined within a rectangular block selected to include zones of interest. Normally, each rectangular block is oriented parallel to the tunnel axis and to the vertical direction. The velocity model forms a base for the TRT™ system to generate an image of anomalies of weaker and stronger rock ahead of and adjacent to the surveyed tunnel. In a homogeneous media, for each source and receiver of known location, the locus of all possible reflector positions with a given two-way travel time defines an ellipsoid in three-dimensional space. For a sufficient number of sources and receivers forming a three-dimensional array, each boundary reflecting seismic waves can be identified as an area where a majority of ellipsoids for pairs of sources and receivers intersect. Thus, each grid point in the volume examined may theoretically give rise to a reflection or scattering event. A discrete image of reflecting or scattering anomalies is calculated for each point of a three-dimensional grid within a selected block of the rock space that includes all sources and receivers. The image is then smoothed by interpolation. A discrete value for the image at each point in the grid is calculated by stacking all seismic waveforms with each waveform shifted proportionally to the total distance from the source via this grid point to the receiver. The shift is calculated using the velocity model defined for the volume of the survey block. Using this approach, the resulting image is similar to a holographic reconstruction. The polarity of the value is positive for reflections from a weak to strong rock transition, and

403 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-50.qxd

2/18/04

2:09 PM

Page 404

is negative for reflections from a strong to weak rock transition. The spacing of grid points sets the range for the desired resolution of the image. The shortest wavelengths of recorded seismic waves determine the available level of resolution. The dimensions of the block, which control the extent of the image, are inversely proportional to the desired resolution of the image. This limitation stems from constraints that control the time of computations required for the TRT™ to generate an image within a preset block. Typically, the initial velocity model is built by extrapolating the velocity tomogram obtained for direct waves and using other available data (velocity measurements, geological data, known voids). As the tunnel excavation advances, the velocity model should be continuously updated and improved based on comparison between tomographic predictions and the ground truth of geological mapping and observed rock conditions. 3.2

Fujikawa Highway tunnel

The 4520 meters Fujikawa Highway tunnels in Shizuoka Prefecture, Japan, were driven through fractured and sheared andesites with tuff breccia lenses. TRT™ was applied in two sections. The objective of the TRT™ survey for the first section was to

delineate the boundary of the gravel deposit overlying an andesite and tuff breccia formation above the alignment of the TBM-excavated tunnel. The objective of the survey for the second section was to image ahead of the TBM where ground conditions had deteriorated due to frequent faulting and shear zones where the tunnel passes under the Umuse River flowing into the Fuji River and continues in a 470 meters section between Sta. 104140 to Sta. 103670. 3.2.1 Gravel boundary Based on seismic reflection, the original assessment of the gravel bottom was at elevation approximately 220 meters (722 feet) ASL, or 55 meters (180 feet) above the centerline of the TBM tunnel. After final adjustment for attenuation of seismic waves, another horizontal reflective boundary was detected at elevation 200 meters (656 feet) ASL, or 35 meters (115 feet) above the centerline of the TBM tunnel. A nearly identical sequence of reflective anomalies was reconstructed for the TRT™ survey in the TBM tunnel. However, the elevation of the lowest boundary was at 202 meters (663 feet) ASL, or 37 meters (121 feet) above the TBM tunnel centerline. Combined images for the pilot tunnel and for the TBM tunnel shown in Figure 3 allowed three-dimensional assessment of the shape of the gravel bottom boundary. The horizontal

Figure 3. Isometric projection due east (left), and due northeast (right) of vertical and horizontal tomograms and contour reflective anomalies above Fujikawa tunnels.

404 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-50.qxd

2/18/04

2:10 PM

Page 405

alignment of this boundary appeared to correspond well with surface reflection data, but its elevation was higher by approximately 5 meters to 7 meters (16 feet to 23 feet). Also, the boundary appeared to dip slightly from the TBM tunnel toward the pilot tunnel. No other explicit horizontal boundaries were identified in the elevation range between 200 meters (656 feet) ASL and the TBM tunnel. 3.2.2 Imaging ahead of TBM Three separate but continuous sections up to 150 meters each were imaged. The images were then correlated with the machine forces and observations. Figure 4 shows the original image between Sta. 104138 and Sta. 103950 along with forecast comments, geologic information, and observations of encountered conditions. 3.3

Hollywood Bypass tunnel

NSA conducted a seismic tomography survey to identify features in a selected section of the Hollywood Bypass tunnel to determine the need for grouting. Specifically, the objective was to produce two-dimensional and threedimensional tomographic images of the ground conditions within the grout curtain zones to identify and delineate various geologic formations, contact zones (sandstone/shale and basalt), and any anomalies within the survey area and/or volume and to identify permeable zones, channel-like features, and unseen weak zones within the survey area and/or volume. A possible water-saturated joint was detected in front of the TBM between Station 5875 and

5925. Figure 5 illustrates the location of this fault crossing the tunnel alignment. 3.4

Lake Mead tunnel

NSA conducted a seismic tomography survey to identify features in the Lake Mead tunnel to determine the orientation and extent of faults in advance of tunnel excavation. Open joints and shear zones in the amphibolite bedrock of Saddle Island carry significant amounts of water. The groundwater inflow exceeded the levels expected prior to start of the IPS-2 intake tunnel excavation. This water appeared to be connected to and recharged from Lake Mead. The joint systems and shear zones that are exposed to Lake Mead and that cross the IPS-2 intake tunnel alignment were considered the major contributor of groundwater inflow into the tunnel both during construction, and when the tunnel was to be dewatered. According to experience, the water-bearing fissures have to be grouted ahead of the tunnel excavation to limit present and future groundwater inflow. The array of seismic sensors used for the TRT™ survey was attached to the tunnel walls approximately 70 feet (21 meters) behind the tunnel face. A rotarypercussive drill at the face area provided a convenient and effective source of seismic waves. The image in Figure 6 was generated using the TRT™ technique based on recording seismic waves reflected from structural anomalies in the rock mass ahead of and around the IPS-2 tunnel. Typically, for unknown ground conditions, a routine probe drilling up to 80 feet (24 meters) ahead of

Figure 4. Horizontal section of TRT seismic reflection image and correlation between forecasted and detected rock features in Fujikawa tunnel.

405 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-50.qxd

2/18/04

2:10 PM

Page 406

the tunnel face has to be conducted to identify zones that require grouting, and to assess the proper range of grouting beyond the tunnel face. The probe drilling results in downtime in tunnel excavation and substantial cost increase. Also, the drilling does not allow assessing the extent of water-bearing zones. Furthermore, the extent and effectiveness of grouting remains unknown. By employing seismic imaging using NSA’s TRT™ technology, the contractor was able to reduce downtime and costs associated with investigating unknown ground conditions. Further, with the more descriptive, more reliable results from TRT™

+2

5 5 5 Possible water57 +7 5 saturated joint or 57 +2 low p-wave velocity 5 anomaly 56 +7 5 58

+2

56

+7

+2 5 Reflection coefficient

55

+7

59

58

5

-100

0

100

Figure 5. Isometric projection of p-wave reflection image extended 220 feet ahead of the face.

imaging, it was possible to determine the extent of the water-bearing zones, providing information for the grouting operation. 4 CONCLUSIONS Encountering unforeseen geologic conditions while tunneling is the single most important contributor to cost overruns and increased risk in tunneling. Seismic imaging effectively reduces that risk by providing the operator with a timely and complex analysis of the upcoming geologic conditions without requiring expensive and highly trained staff on site. RockVision3D™ has been demonstrated to be an effective and non-intrusive tool for imaging ground anomalies in three-dimensions along and adjacent to tunnels in a number of different geologic settings. RockVision3D™ has been used effectively prior to construction to characterize ground conditions along a tunnel alignment. It has also been used as a tool when problems were encountered during construction or to understand phenomenon such as subsidence after construction. RockVision3D™ can also be used to assess the effectiveness of grouting operations. TRT™ has also been demonstrated to be an effective and non-intrusive tool for imaging ground anomalies in three-dimensions along and adjacent to

Figure 6. Side view of vertical section through the center line of the tunnel showing the shaft boundary and approximate profile of the lake bottom, and a local anomaly 20 feet behind the face that matched a water-carrying fracture zone.

406 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-50.qxd

2/18/04

2:10 PM

Page 407

tunnels in a number of different geologic settings. With a proven and accurate detection range of 50 meters to 120 meters along the alignment, and with a peripheral range of 20 meters to 30 meters, TRT™ provides multiple interrogation possibilities from the digitized grid of data produced by a multitude of convenient sources. TRT™ can be used in tunnels as they are constructed, providing valuable information regarding upcoming ground conditions while minimizing downtime. TRT™ can also be used to assess ground conditions around an existing tunnel. When these tools are used by an experienced site tunnel geologist or engineer, the ability to interpret expected conditions and anticipate ground control measures is greatly enhanced over any other method in use today. This can greatly reduce cost overruns due to change of conditions and reduce risks for tunneling projects, both large and small.

REFERENCES Nur, A., 1987. Seismic rock properties for reservoir descriptions and monitoring. In G. Nolet (ed.) Seismic tomography with applications in global seismology and exploration geophysics, Dodrecht: D. Reidel Publishing Co. Qin, F., Luo, Y., Olsen, K., Cai, W., and Schuster, G. 1992. Finite-difference solution of the eikonal equation along expanding wavefronts. Geophysics, 57(3): 478–487. Shea-Albin, V.R., Hanson, D.R., and Gerlick, R.E. 1991. Elastic wave velocity and attenuation as used to define phases of loading and failure in coal. USBM Report of Investigation 9355, 43 pp. Yu, G., 1992. Elastic properties of coals. Ph.D. Thesis, MacQuarie Univ., Sydney, Australia, 133 pp.

407 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-51.qxd

2/18/04

7:57 PM

Page 409

Session 3, Track 4 Machine mining – soft ground to hard rock to everything in between

Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-51.qxd

2/18/04

7:57 PM

Page 411

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Conditions encountered in the construction of the Braintree-Weymouth Tunnel Project, Boston, Massachusetts D.W. Deere Tetra Tech RMC

J. Kantola Modern Continental Construction Company

T. Davidson Shaw Environmental & Infrastructure

ABSTRACT: The Braintree-Weymouth Tunnel and Shafts Project was constructed between 1999 and 2003. The project is a continuation of the previously constructed Inter-Island Tunnel. The two tunnels convey wastewater to the MWRA (Massachusetts Water Resources Authority) wastewater treatment facility at Deer Island and return thickened sludge back to the Fore River Shipyard in Quincy, Massachusetts through various ductile iron pipes. The Braintree-Weymouth Tunnel was constructed in two segments from a common shaft: a 2,350-foot length in a southerly direction by drill-and-blast techniques (approximately 14 feet high by 18 feet wide) and a 12,350foot length (13-foot diameter) in a northerly direction by Tunnel Boring Machine (TBM). The tunnel is located in the southern margin of the fault bounded structural and sedimentary Boston Basin. All total, eight different and distinct rock types were encountered throughout the tunnel alignment. The majority of the tunnel was driven through the Cambridge Formation, which was comprised primarily of argillites with interbedded sandstones and quartzites; all locally intruded by mafic dikes. This paper discusses the general conditions encountered and their effects on TBM tunneling through this common formation of the Boston area. Boreability of the tunnel was impacted by several factors including zones of very high strength rock, degrees of rock mass fracturing, and occasional TBM steering corrections. The presence of foliation shear zones (aka, kaolinized argillite) affected tunnel progress and support.

1 INTRODUCTION In October 2003, Modern Continental Corporation, of Cambridge Massachusetts, successfully completed the Braintree-Weymouth Tunnel and Shafts (BWTS) Project for the Massachusetts Water Resources Authority (MWRA). The tunnel alignment, shown on Figure 1, is a continuation of the previously constructed Inter-Island Tunnel. The two tunnels meet at Nut Island, where the BWTS project continues in a southerly direction to the Fore River Shipyard in Quincy, Massachusetts. The tunnel system is for conveying wastewater through a 42-inch diameter ductile iron pipe to the MWRA wastewater treatment facility on Deer Island. The tunnels will also return thickened sludge through twin 14-inch diameter ductile iron pipes back to the Fore River Shipyard, where it will be converted into

fertilizer pellets. The BWTS tunnel continuation consists of 2.8 miles of tunnel and three 200-foot plus deep shafts. The three shafts included a main work shaft in North Weymouth and two smaller raise bore shafts at the north (Nut Island) and south (Fore River Shipyard) ends of the project. The smaller shafts were used to bring the ductile iron pipes up from the tunnel grade to the facility grades at the pelletizer and treatment plants, respectively. The three shafts were used to bring various combinations of the ductile iron piping up from the tunnel grade to the facility grades. The tunneling work on the BWTS contract was mined through one main access shaft on a spit of land along the tidally influenced Fore River in North Weymouth, Massachusetts. The shaft became known as the North Weymouth Shaft (NWS) and its location is shown on Figures 1 and 2. From here, two tunnel headings were driven in the northerly and southerly

411 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-51.qxd

2/18/04

7:57 PM

Page 412

Figure 1. Location map.

directions. In the southerly direction, a 14-foot horseshoe tunnel was drilled and blasted 2,350 feet towards the Fore River Shipyard. In order to install a Tunnel Boring Machine (TBM) for the northerly heading, a 16-foot horseshoe starter tunnel was drilled and blasted 260 feet north of the shaft’s center. A TBM was then used to drive a 13-foot diameter tunnel 12,100 feet to Nut Island in order to marry up with the previously completed Inter-Island Tunnel. 2 GEOLOGY This tunnel project offered a unique view along a continuous 2.8 mile length of bedrock located within the southern margin of the Boston Basin which typically has less than 5 to 10 percent outcrop exposure. The Boston Basin is comprised of polydeformed and

faulted meta-sedimentary strata of the Cambridge, Roxbury, and Weymouth Formations. These formations have been turned on end vertically and traversed at right angles to 30 degrees across the bedding along the north and south tunnel transects, respectively. The tunnel was subdivided into distinct subreaches based upon stratigraphy or faulting. The geologic subreaches, as well as engineering properties of Rock Mass Rating (RMR), unconfined compressive strength (UCS), Brazilian tensile strength (BTS), and Cerchar abrasivity (CAI) are summarized on Table 1. A geologic map of the alignment is shown on Figure 2 and a geologic section on Figure 3. 2.1

The southern margin of the TBM tunnel and the entire drill and blast tunnel were constructed in the

412 Copyright © 2004 Taylor & Francis Group plc, London, UK

Weymouth Formation

09096-51.qxd

2/18/04

7:57 PM

Page 413

Figure 2. Tunnel geology.

413 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-51.qxd

2/18/04

7:57 PM

Page 414

Table 1. Geologic Subreaches. RMR

UCS (psi)

BTS (psi)

CAI

Sub-reach

Rock type

Stationing

Avg.

Range

Avg.

Max.

Avg.

Avg.

TBM Mining A

Weymouth Quartzite

9845 – 10407 (10262 TBM) 10407 – 10882 10882 – 13178 13178 – 15486 15486 – 15989

58.8

52–66

21,108

21,108

–

3.5

62.5 53.6 58.3 67

54–67 14–74 47–76 57–71

7,952 21,059 14,079 24,598

8,000 40,000 26,000 32,500

1,440 1,925 1,594 2,672

3.0 2.8 1.9 4.0

15989 – 16025 16025 – 16364 16364 – 22367

58 69.8 57.7

53–62 50–72 32–67

12,741 25,595

14,900 42,000

1,705 2,296

3.3 2.9

9845 – 8671 8671 – 7858

51.6 47.8

48–53 28–58

8525 – 8327 7858 – 7630

50.3 48.4

49–51 34–58

B1 B2 C D1

Weymouth Argillite Cambridge Argillite Cambridge Argillite Roxbury Conglomerate Roxbury Argillite (maroon) Roxbury Volcanic Flow Cambridge Argillite

D2 E Drill and Blast F Weymouth Argillite G Quincy Granite Weymouth Quartzite/Argillite (within granite) H Weymouth Argillite

Figure 3. Braintree Weymouth Tunnel – generalized geologic cross section.

Weymouth Formation that consists of maroon to gray, very thinly bedded to laminated argillite and siltstone (2,050 feet) with localized sandstone and quartzite (750 feet). Bedding and lamination is steep to vertical with tight isloclinal folds observed in siltstone and interbedded argillite members. Beds are locally offset from three inches to eight feet. Joint spacing within the Weymouth Formation generally is closely to moderately spaced with localized shears and fault zones. Quartzite members were blocky, very light gray to buff white, lithic arenites with interbedded gray to maroon argillite. Compressive strengths in the argillite averaged about 8,000 psi, and averaged about 21,000 in the

quartzite. The Weymouth Formation was intruded along bedding by the Quincy Granite in two separate locations along the drill and blast tunnel. The granite was medium to coarse grained, blocky, and locally faulted. 2.2

The Cambridge Formation along the alignment was comprised of argillite (7,900 feet) and localized sequences of argillaceous sandstone (1,950 feet) and quartzite (250 feet) which were all locally intruded by mafic dikes (450 feet). The Cambridge Formation was the predominant rock type in tunnel subreaches

414 Copyright © 2004 Taylor & Francis Group plc, London, UK

Cambridge Formation

09096-51.qxd

2/18/04

7:57 PM

Page 415

B2, C and E, and accounted for approximately 11,100 feet of the 12,100-foot northward bound TBM tunnel bore. The Cambridge Formation was first encountered in the TBM tunnel at station 108+82 at a healed thrust fault which marked the boundary between the older Cambridge Formation (hanging wall) and the younger footwall sequence of the Weymouth Formation. The Cambridge Formation contained steep to vertical overturned lamination and/or bedding (dipping 60 to 90 degrees north northeast). The Cambridge Formation members were comprised of varying percentages of quartz, albite, chlorite and K-phyllosilicate minerals that varied from muscovite to illite throughout the entire tunnel transect. The dominant rock type of the Cambridge Formation was the Cambridge Argillite which was a gray, very thinly laminated and locally sandy rock. Laminations ranged from 1 mm to 5 cm in thickness within the typical dark gray to gray Cambridge Argillite members in subreaches B2 and C. Although the laminations or bedding banding in the argillite were very closely spaced, the actual joint discontinuities parallel to bedding were wider spaced. The typical joint spacing measured parallel to bedding was 2 feet. This feature distinguishes argillite from slate or shale. Locally the Cambridge Argillite contained intervals of very thinly bedded argillaceous sandstone and thinly to thickly bedded quartzite. The argillaceous fine sandstone was blocky, light gray to gray, very thinly bedded to thickly bedded with localized cross bedding. The bedding was overturned and dipped steeply to vertically to the north-northeast. The interbedded quartzite was massive, very light gray to buff white, medium to locally coarse grained, thinly to thickly bedded with localized cross bedding and alteration. Sandstone compositions range from lithic arenite in the southern reaches of the tunnel (subreach B2) to quartz sub-lithic arenite in the northern tunnel margins (subreach E). The Cambridge Formation was also locally intruded by mafic dikes. Argillites in subreach E differed from those in subreaches B2 and C in that they contained distinct bedding and gradational transitions to argillaceous fine sandstones and sub-lithic arenites with localized cross-bedding suggesting a shallower depositional environment than typical Cambridge Argillite (subreaches B2 and C). Localized cross-bedding indicated the beds were overturned and dipping to the north-northeast. In addition, the argillites in subreach E tended to have higher quartz content with lighter gray to light gray coloration with interbedded dark gray argillites. Shearing was common locally parallel to the bedding and appeared to be foliation/bedding plane shear zones (Deere, 1973). These appeared along and near contacts of quartzite and argillite, and as described by Deere, were probably due to slippage between beds of differing properties due to regional folding. They

appeared as white, crushed sand in the quartzites, and kaolinized zones (Kaye, 1967) in the argillite. The shears ranged in thickness from less than one inch to two feet. Compressive strengths in the Cambridge Formation ranged from 10,000 to 19,000 psi within the argillite in subreaches B2 and C with locally higher compressive strengths adjacent to intrusive dikes. The interbedded sequences of sandy argillite and argillaceous fine sandstone had considerably higher compressive strengths and ranged from 25,500 to 40,000 psi. Compressive strengths within interbedded quartzite ranged from 22,000 to 32,000 psi. Extremely blocky ground was encountered between stations 19500 and 19600 within argillaceous sandstone where eight distinct joint sets were identified. Joint spacing ranged from generally close to moderately spaced with localized zones of very closely and widely spaced joints. Five predominant joint sets were visible along the entire TBM tunnel transect with localized zones of up to eight distinct joint sets. 2.3

3 TUNNEL BORING MACHINE The 13-foot diameter Nut Island drive was excavated using a new telescoping double shielded TBM built

415 Copyright © 2004 Taylor & Francis Group plc, London, UK

Roxbury formation

Mining also penetrated units of the Roxbury Formation, including conglomerate and lithic sandstone (500 feet) that graded into maroon silty sandy argillite (50 feet) in fault contact with volcanic flows (350 feet). The conglomerate member of the Roxbury Formation was limited to subreach D1 and graded from a predominantly lithic graywacke with several outsized clasts to a massive, matrix-supported to clast-supported conglomerate with gravel to boulder sized clasts which were locally deformed and elongated along a locally well developed cleavage dipping steeply to the south. Individual clasts, including quartzite and volcanic rocks, were round to locally sub-angular. The conglomerate graded from lithic sandstone to a maroon silty to sandy thinly laminated argillite in cross-fault contact with older volcanic flows. The volcanic flows (subreach D2) with localized flow breccia and amygdaloidal zones ranged from maroon to greenish gray in color. Quartz filled amygdules ranged from 1 to 3 cm in diameter. Compressive strengths in the Roxbury Formation varied in the different rock types. The conglomerate strengths ranged between 22,000 psi and 33,000 psi while the volcanic flows had between 10,500 psi and 18,000 psi compressive strengths. The conglomerate was generally massive with widely spaced joints, while the volcanic flow contained moderately spaced joints.

09096-51.qxd

2/18/04

Table 2.

7:57 PM

Page 416

Specifications of TBM.

mesh. The standard bolt pattern, or the supplemental version, was used almost exclusively throughout the tunnel. Approximately 105-feet of the TBM bore required additional steel ribs and lagging and are discussed below under Cave-In and Fractured Sandstone, respectively.

General Machine Manufacturer: Wirth built in Erkelenz, Germany Model: TB 400 H/TS Total Power Installed: 2,700 horsepower Weight with Backup Gear: 413 tons Maximum Design Thrust: 55,000 pounds per cutter Turning Radius: 1,000 feet Conveyor Capacity: 275 cubic yards per hour Probe Drill: (1) Atlas Copco COP 1238 Ground Support Drills: (2) Atlas Copco COP 1038S Number of Thrust Cylinders: 7 Stroke of Thrust Cylinders: 4.58 feet Total Thrust Force: 1,720 tons Cutters Type: 17-inch Disc, back loading Quantity: 6 center, 21 face, 2 gauge, 1 reamer Kerf Spacing: 3.125 inches Cutterhead Maximum Operating Cutterhead Thrust: 1,720 tons Speed: 0–13 revolutions per minute Rotational Direction: Reversible Torque at 13 revolutions per minute: 650,000 foot-pounds Break out Torque: 2,175,000 foot-pounds

5 PROBLEMS ENCOUNTERED DURING CONSTRUCTION 5.1

by Wirth of Germany. The main machine specifications are presented on Table 2. Up to seven thrust cylinders could be used for advancing the cutter head while mining, though normally only six cylinders were used. The maximum thrust per cutter was designed for 55,000 pounds. The majority of the tunnel was driven with a thrust of about 45,000 pounds per cutter. If the shield became stuck, the machine could also operate in a high pressure mode. The high pressure mode increased the propel pressures from a normal 3,500 psi to 5,000 psi, and powered up the seventh thrust cylinder. Though it was not recommended, the machine had to be operated under the high pressure mode when mining through the Roxbury Conglomerate Formation encountered beneath Hough’s Neck (Figure 2). Under this condition, the thrust per cutter exceeded the maximum design load by up to 25 percent. The TBM was also equipped with one probe drill, two ground support drills, and a ring beam erector. Very limited probe drilling was performed, and the ring erector was never used during the project. However, the ground support drills were used to set a standard pattern of bolts throughout the tunnel drive.

4 GROUND SUPPORT The ground support used in the TBM drive consisted of 5.5-foot SwellexTM rock bolt pattern of three bolts in the crown on five foot centers. The standard pattern was at times supplemented with additional bolts and wire

5.2

Cave-in

Only one zone of highly fractured ground between stations 11367 to 11420 required substantial support. Originally this ground was supported using the standard bolt pattern, locally supplemented with additional bolts and mesh. After 24 hours, a 25-foot high chimney developed above the crown between stations 11401 and 11406. In order to secure this area 270 degrees of the tunnel circumference had to be supported with ribs and lagging. Fourteen days after the chimney in the crown developed, the eastern sidewall of the tunnel collapsed between stations 11367 and 11395. Ribs and lagging were installed along 180 degrees of

416 Copyright © 2004 Taylor & Francis Group plc, London, UK

Shaft

The access shaft was constructed through 80 feet of overburden that was supported in a 34-foot by 22-foot elliptical opening. Throughout the shaft sinking operation, locally fractured and altered rock and salty groundwater inflow had to be dealt with. Ring steel was used to support the rock, and grouting (cement and chemical) was attempted to control the water. In order to complete the shaft, 27 rings of W8 35 steel installed on four foot centers, with shotcrete, were used for support. The turn-under was supported with spiling and long rock dowels. A geologic section of the shaft is shown on Figure 4. Shaft conditions prior to design were investigated by a single vertical borehole (BW-88) down the center of the shaft. As shown on Figure 4, RQDs within the borehole were typically high (80) and indicated primarily good rock. As relatively massive quartzite was expected, the shaft support was designed with radial rock bolts and shotcrete. Upon initiation of the rock excavation, the conditions encountered were considerably poorer than anticipated from the borehole results. This was due to the presence of a hydrothermally altered diabase dike in the northern shaft wall, as well as locally thin vertical interbeds of sheared and altered argillite and quartzite (foliation shears). In order to safely advance the shaft, W8 35 steel sets and shotcrete were required. This example illustrates the limited geologic information resulting from investigating vertical rock structure with a vertical borehole.

09096-51.qxd

2/18/04

7:57 PM

Page 417

Figure 4. Geologic section of shaft.

417 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-51.qxd

2/18/04

7:57 PM

Page 418

the tunnel circumferance to support this collapsed sidewall. The failure was caused by slippage of a very tall wedge bounded by two distinct shears. The material in the wedge consisted of both blocky quartzite and altered argillite. This feature could be considered a foliation shear zone as described by Deere (1973). 5.3

Fractured sandstone

Hard, unusually tough, blocky, and very fractured ground was encountered in sandstones between stations 19492 and 19550. Eight distinct joint sets resulted in fist-sized wedges of very high strength rock that caved and then lodged between the TBM telescoping shield and tunnel wall. Additional support including ribs and lagging were installed along 270 degrees of the tunnel circumference between stations 19530 and 19545 to support the extremely fractured ground conditions. This phenomenon completely stopped the TBM moving as the machine became seized. The contractor attempted to pull the rear shield forward by anchoring the front shield with the two stabilizer assemblies. These stabilizer assemblies are similar to gripper pads and are located on the upper quadrant of the front shield. Due to rock fracturing and fallout, the gap between the tunnel wall and stabilizer assemblies had to be cribbed. Under maximum retraction pressure of the telescoping jacks, the stabilizer pads lost pressure and could not anchor into the fractured rock. The result was the front shield moved towards the rear shield. Next the contractor attempted to thrust off the tail shield with up to four push jacks with no success. The contractor eventually resumed forward progress of the rear shield by overriding the pressure sensors and applying a constant but low pressure to the stabilizer pads and telescoping jacks (the shield retraction jacks). While applying this steady pressure, the annular space between the tunnel wall and telescoping shield was flushed with high pressure air and water. Though considerable effort was expended, the shield was freed and mining resumed. 5.4

Boreability

The overall penetration and TBM advance rates were lower than planned for at bid time by the contractor. Although there are several factors that can be attributed to overall reduced penetration (machine factors such as steering), the primary geotechnical factors were higher strength rock than anticipated and a lesser degree of fracturing of the rock mass. Many high strength interbeds of sandstone, quartzite, and dikes were encountered within the Cambridge Formation. Unconfined compressive strengths exceeding 40,000 psi were locally measured within these units.

The jointing and fracturing of the rock mass were less than expected as evidenced by the minimal support required to drive the tunnel. There are numerous empirical data from TBM tunnels in hard rock that demonstrate the relationship of increased TBM penetration with increased degree of jointing and fracturing of the rock mass. This relationship has been quantified by Dr. Amund Bruland and others at the Norwegian University of Science and Technology (NTNU) in Trondheim (Bruland, 1998). The degree of fracturing of a rock mass can be translated into a single numerical value called the “fracturing factor” ks. To obtain a fracturing factor, a given section of tunnel must be geologically mapped. The spacing between joints determines the fracture class. If joints are grouped into joint sets, each joint set will have its own fracture class. Having obtained the fracture class and angle between the jointing and the tunnel axis, the fracturing factor is determined by using a graph which correlates the angle to the fracturing factor for various fissure and joint classes. If joints are grouped into joint sets, each joint set will have its own fracturing factor which must be combined into an overall fracturing factor. Once the total fracturing factor is known for a given section of tunnel, the entire process is repeated for the next section. Ks-tot is the fracturing factor for a single mapped section with several joint sets. The plot of ks-tot versus tunnel station for the entire TBM tunnel is shown on Figure 5. The plot of ks-tot has been dampened by using a 12 point moving average (12 40 feet  480 feet). Also plotted on the graph is the actual field TBM penetration rate as recorded from shift reports. The plot is based on a 25 point moving average (25 19 feet average per shift  475 feet). The plot shows good correlation between the penetration rate and fracturing factor. For example, when the fracturing factor decreased (joint spacing increased) in subreaches B-2 (sandstone), D-1 (conglomerate), and the last half of E (sandstone), the penetration rates also decreased. The importance of a fracture factor in TBM performance predictions is discussed by Dollinger and Raymer (2002). They recommend that fracture factors be included in pre-bid geotechnical documents for tunnels in strong rock. 6 CONCLUSIONS The project was successfully completed, although challenging geotechnical conditions were encountered. The presence of numerous interbeds of very high strength sandstone and quartzite were somewhat surprising when compared to other tunnels previously

418 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-51.qxd

2/18/04

7:57 PM

Page 419

Figure 5. Fracturing factor vs. TBM penetration.

excavated in the Cambridge Argillite. Foliation shears were common in this interbedded formation. Geologic field mapping coupled with field penetration data indicated a good correlation between fracturing factors and machine performance. It is paramount when planning future tunnels in the Cambridge Formation that representative sampling be performed by orienting borings across the bedding. Measurement and calculation of fracturing factors from borehole data will greatly aid estimates of TBM performance. ACKNOWLEDGEMENT The authors thank the MWRA for permission to publish this paper. This article represents the opinion and conclusions of the authors and not necessarily those of the MWRA.

REFERENCES Bruland, Amund. 1998. Project Report 1D-98 Hard Rock Tunnel Boring: Performance Data and Back Mapping. Doctoral Dissertation. NTNU-Anleggsdrift Trondheim, Norway. pp. 71–74. Deere, Don U. 1973. The Foliation Shear Zone, An Adverse Engineering Geologic Feature of Metamorphic Rocks. Reprinted from the Journal of the Boston Society of Civil Engineers. October. Dollinger, G. and Raymer, J. 2002. Rock Mass Conditions as Baseline Values for TBM Performance Evaluation. North American Tunneling 2002 (Ozdemir ed.). A.A. Balkema, Rotterdam 2002. pp. 3–7. Kaye, Clifford. 1967. Kaolinization of Bedrock of the Boston, Massachusetts Area. United States Geological Survey Professional Paper 575-C. pp. C165–C166.

419 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-52.qxd

2/18/04

7:58 PM

Page 421

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

The Manapouri Tailrace Tunnel No. 2 construction – a very large TBM tunnel in very strong rock D.W. Deere Tetra Tech RMC

S. Keis Meridian Energy, Ltd.

C. Watts Halcrow, Ltd.

ABSTRACT: The Manapouri Tailrace Tunnel No. 2 (2MTT), located in southern New Zealand, was constructed between 1997 and 2002 as part of Meridian Energy, Ltd.’s 700 MW hydroelectric facility. The 10 km long tunnel is one of the World’s largest full face TBM tunnels, at 10 m diameter, driven in strong to very strong rock. Many obstacles were traversed by the contractor. The TBM passed through five major fault zones, 5 to 120 m wide with cumulative groundwater flows up to 1,000 l/s; however, the most difficult ground for the TBM was the hard, blocky ground. This project pushed the envelope in hard rock TBM tunneling due to tunnel size, water inflows, the strength of rock, environmental considerations, and the isolated nature of the site. Construction difficulties and ensuring contractor claims resulted in exhaustive studies of rock properties and TBM behavior which now provide important insight into hard rock boreability.

1 INTRODUCTION

2 GEOLOGY

The 700 MW Manapouri Hydroelectric Power Station is located in the Fiordland region of southwestern New Zealand (Figure 1). The facility has been operating since 1969 and is currently owned by Meridian Energy. A second tailrace tunnel was constructed between 1997 and 2002 in order to allow the power station to operate at full efficiency, since friction losses had reduced the maximum output from 700 MW to 585 MW. The second tunnel was constructed by TBM parallel to the original tailrace tunnel. It had an excavated diameter of 10 m (Figure 2) and a length of 10 km. The contractor for the project was the joint venture FDI (Fletchers, Dillingham, Ilbau). The site lies within a temperate rainforest with an annual rainfall of 6 m. The project also lies within New Zealand’s largest National Park. For additional information on this fascinating project, the reader is referred to “Second Manapouri Tailrace Tunnel Not Just A Walk in the Park” by Martin, Heer, and Macfarlane published in the 2003 RETC Proceedings.

2.1

The first and second tunnel were excavated in Palaeozoic aged migmatitic metamorphic rocks and Cretaceous aged igneous intrusive rocks. The metamorphic rocks contain minerals of the amphibolite facies and exhibit features that indicate several phases of metamorphism and intrusion in their long history. Another dominant geological feature close to the project (45 km northwest) is the tectonic boundary between the Australian and the Pacific Plates, where the Australian Plate is being subducted beneath the Pacific Plate. This tectonic activity has played a part in exhuming these rocks from depths in the order of 10 km to their present near-surface position. 2.2

Rock types intercepted

The TBM part of the tunnel was divided into four reaches during the design phase as shown on the

421 Copyright © 2004 Taylor & Francis Group plc, London, UK

Geologic setting

09096-52.qxd

2/18/04

7:58 PM

Page 422

Figure 1. Location map.

422 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-52.qxd

2/18/04

7:58 PM

Page 423

Figure 2. 10 m diameter smooth bore.

Figure 3. Geologic profile.

geologic profile (Figure 3). This subdivision was based broadly on anticipated geological conditions. Reach 1 (1,770 m) was excavated in mixed metasediments, consisting primarily of non-banded and banded gneiss and interbedded metasedimentary calc-silicate rocks. The majority of Reach 2 (2,530 m) consisted of gabbro/diorite and diorite gneiss, but also contained banded gneiss with some interbedded calc-silicate rocks. Reach 3 (1,800 m) consisted of banded gneiss, amphibolite and amphibolite gneiss with minor amounts of intruded pegmatite and granite. Reach 4 (3,729 m) intercepted banded and non-banded massive

gneiss with interbeds of calc-silicate and intruded pegmatite and granite which were more prevalent in this reach. Description of Gneiss: Generally the gneiss can be described as poorly foliated to well foliated, well banded to non-banded and generally the dominant minerals were hornblende (40%) and biotite. The average UCS for the gneiss was 138 MPa, ranging from 45 to 379 MPa. Gneiss was the dominant rock type in the tunnel (5,175 m or 53%). Within Reach 1 and part of Reach 2, a dark gray, very poorly to well foliated, fine grained gneiss (contractor named this rock metaandesite and metadolerite)

423 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-52.qxd

2/18/04

7:58 PM

Page 424

was present as cross cutting features, inclusions, and bands within the other units. The fine grained gneiss makes up nearly 20% (354 m) of the rock excavated in Reach 1 and 8.8% (223 m) of the rock excavated in Reach 2. Fine grained gneiss bands were typically 0.5 to 2 m, but up to 20 m wide. The foliation was normally subparallel to the contacts of the bands and was typically discordant with the foliation of the adjacent gneiss. Description of Metasediments: The metasediments consist mainly of calc-silicates (730 m or 7.5%) and minor amounts of marble (45 m or 0.5%). The calcsilicate can be described as very poor to moderately foliated, well banded, and multiple lighter colors (cream, white, light green, pink, and buff). Calcsilicates were regularly inter banded with biotite gneiss. Typically, the dominant minerals of a calcsilicate are quartz and Ca-minerals such as epidote and pyroxene. The average UCS for the calc-silicates was 171 MPa and ranged from 41 to 309 MPa. Marble can be described as variably banded, cream in color, and soft. Description of Amphibolite/Amphibolite Gneiss: Amphibolite/amphibolite gneiss contains greater than 40% amphibole (hornblende). Generally these rocks can be described as non to well foliated, non to well banded and darker in color due to the higher

hornblende content. The average UCS was 131 MPa which ranged from 20 to 273 MPa. This rock type was present in the tunnel for 920 m or 9.5%. Description of Pegmatite/Granitics: Pegmatite/ granitics were massive (not foliated) and typically white in color. The average UCS was 174 MPa and ranged from 50 to 225. The pegmatite/granitics were generally present as narrow bands (0.1 to 0.5 m thick) distributed within the other rock types. However, occasionally pegmatite/granitics formed the dominant rock type in the tunnel (950 m or 9.7%). The tunnel exposures typically showed the pegmatite/ granitics cross cutting the other rock types indicating that these dikes were the youngest rocks (Figure 4). Description of Gabbro/Diorite: Gabbro/diorite were massive (non foliated), non-banded and darker in color. The average UCS was 161 MPa which ranged from 89 to 228 MPa. In the field it was not possible to distinguish between these two rock types and the term gabbro/diorite was used. Petrology investigations indicate that diorite was the dominant rock type with only minor amounts of gabbro intercepted. The other rock type that makes up this unit of gabbro/diorite is diorite gneiss. This rock has a similar appearance to the gabbro and diorite, but contains a weak mineral alignment or foliation. Often the gabbro/ diorite and diorite gneiss would occur together with

Figure 4. Rock mass at the face.

424 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-52.qxd

2/18/04

7:58 PM

Page 425

gradational or indistinct boundaries between them. This was the second most common rock type intercepted in the tunnel (1,990 m or 20%). 2.3

Foliation

The extent of foliation development was variable from very poorly to well developed and was almost always parallel to the metamorphic banding. However, from approximately Station 5836 to 6704 most of the metamorphic rocks were well foliated with foliation striking sub-perpendicular to the tunnel alignment and dipping at between 60 and 90° to the west and east. The degree of foliation evident over this section of tunnel was such that it may be described as semi-schist. This well foliated section of tunnel was characterized by some of the most sustained and rapid rates of penetration and daily advance rates recorded for Tunnel 2. The best maximum daily rate of 38.4 m, the best weekly advance of 169.6 m, and the best monthly advance of 604.3 m was achieved. The average ROP for the best month was 1.96 m/hour and the utilization for the month was 42.7%. While most rock characteristics remained essentially the same (rock types, joint and shear occurrence and characteristics, and presence of variable lithologies within the heading) the presence of well-developed foliation and its orientation with respect to the tunnel appears to have had a significant effect on penetration rates. The assistance that foliation was providing to the TBM was evident in the tunnel walls over this section where foliation planes are well defined and act as rock mass discontinuities.

some overbreak in the crown (0.1 to 1.2 m). The fault zone was moist with the occasional seep intercepted. A total of 46 sets were installed. The Disaster Burn Fault Zone (Station 5414 and 5428) strikes approximately normal to the tunnel axis, dips to the east at 60°, and consists of a 4 m wide zone of sheared and crushed material. The sheared and crushed material contained groundwater seeps in the crown and did not stand well, forming overbreak above the head of the TBM, up to 3 m forward of the fingers and 3.5 m above the crown. This ground probably represented the poorest tunneling conditions. The overbreak was stabilized using shotcrete and timber cribbing in combination with steel sets (15 total). The Mica Burn Fault Zone (Station 6704 and 6824) hanging wall strikes nearly normal to the tunnel axis and dips to the west at 70°. The hanging wall consists of a 3 to 5 m wide zone of crushed material composed of silt, sand and gravel sized rock fragments with occasional seams of clay gouge (1 to 10 mm). The internal fault material consisted of pervasively shattered rock with numerous discrete sheared and crushed zones. While this was the largest fault zone encountered, it was characterized by relatively good tunneling conditions with some of the best daily advance rates while standing steel sets (15.8 m/day). Steel sets (121) were stood throughout this zone, but none were observed to be taking any appreciable loads and overbreak was minimal. This behavior must in part be associated with the absence of any significant groundwater through this zone.

2.5 2.4

Description of fault zones

Five major fault zones were intercepted in the tunnel (Figure 3). The Wilmot Fault Zone consists of a 1 to 3 m wide crushed and sheared zone that strikes at 20° off the tunnel alignment and dips to the north east at 50 to 70°. The crushed and sheared materials when exposed in the crown produced up to 3 m of overbreak. This section of the tunnel was supported with 20 steel sets at 1.2 m centers. The Stella Burn Fault Zone (Station 2695 to 2788) consists of zones of blocky and seamy rock cut by numerous sheared zones. This feature produced up to 2 m of overbreak in the crown and was supported with 65 steel sets at 1.2 m centers. The Disaster Branch Fault Zone (Station 4088 to 4135) is a 16 m fault zone that strikes approximately 45° off the tunnel alignment and dips steeply (60°) to the west. The hanging wall consisted of a 0.5 to 1 m thick crushed zone with some occasional gouge material 1 to 70 mm thick. Internal fault material consisted of moist crushed/sheared/shattered rock with soil like properties. This material stood well with

In Tunnel 1 high groundwater flows under pressure were intercepted which delayed progress and caused washouts. Extensive grouting was completed ahead of the tunnel face in Tunnel 1 where water was intercepted, but was believed to have limited success. Within Tunnel 2 similar groundwater conditions to Tunnel 1 were anticipated (Figures 5 and 6). To allow the TBM to successfully tunnel through these conditions, the GBR and Technical Specification required that two drainholes be installed at a minimum of 30 m ahead of the TBM face over the following specified tunnel stations: 3600 to 4700, 6600 to 7500, and 8100 to 8500. The sections of tunnel that required drainholes totaled 2,400 m, and represented areas where high groundwater inflows or washouts were encountered during the excavation of Tunnel 1. 2.5.1 Drainhole drilling A total of 30 drainholes were drilled between 3600 and 4700. These drainholes were 125 mm in diameter and were drilled up to 120 m ahead of the face. No drainholes were drilled in the remaining two

425 Copyright © 2004 Taylor & Francis Group plc, London, UK

Groundwater conditions encountered

09096-52.qxd

2/18/04

7:58 PM

Page 426

Figure 5. Groundwater at the heading.

Figure 6. Measurement of groundwater flows.

426 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-52.qxd

2/18/04

7:58 PM

Page 427

specified drainhole sections as less groundwater was intercepted and controlled with probe holes (50 mm diameter holes drilled using the rock bolt drill). However, it was determined that drainholes would be drilled if the combined flows from 2 probe holes exceeded 10 l/s or if the probes indicated very soft or running ground. The contractor used a DTH drill rig to drill the drainholes that had to be mobilized each time a drainhole was required. This meant that considerable time was required to set up the drill rig and because of the size of the rigs only one hole could be drilled at a time. The maximum flow from a drainhole was measured at 164 l/s (2,580 gpm) and the maximum pressure was measured at greater than 35 bar (508 psi). During the drilling of some drainholes, problems were encountered advancing the drains when high groundwater flows and pressure were intercepted. The water pressure in the drainhole is believed to have exceeded the hammer pressure (25 bar or 362 psi) restricting progress. However, progress through these high groundwater areas was achieved by intercepting the feature with multiple drains. When the advance of the drainholes was stopped by a high groundwater bearing zone, the drilling rig was disestablished and the TBM mined up to 20 m from the end of the drains. Additional drainholes were then drilled into the zone and were generally successful at penetrating the zone. The time delay between the first and second pair of drains is inferred to have allowed the water pressure to drop sufficiently for the second pair of drainholes to progress through the zone. We believe that the drainholes were a successful tool in progressing the TBM through these high groundwater bearing zones. The drainholes were time consuming to drill, but the negative effects are believed to be outweighed by the positives, which are discussed below:

3 TUNNEL BORING MACHINE

– The drains were drilled out ahead of the tunnel and reduced the groundwater pressure in the zone before it was excavated by the TBM head. On average the water bearing zone had two days or more drainage before encountered by the TBM. The lower water pressure within the zone reduced the chance of a washout if fine sheared and crushed materials were intercepted. – A major problem when excavating the high flow areas was the excess water washing fines off the inclined conveyor belt between the TBM and the backup. The drainholes intercepted the high groundwater bearing zones and diverted water away from the head of the TBM reducing the amount of water within the muck and on the belt. The large drainhole flows were piped and discharged down the tunnel away from the working area.

Transfer size Back-up power

The 2MTT was excavated with a 10.05 m diameter Robbins open-beam machine. The basic specifications of the TBM are shown in Table 1. It was fitted with 68, 432-mm diameter disc cutters at a recommended cutter load of 27 tons. Eleven drive motors at 315 kW provided a total cutterhead drive power of 3,465 kW (4,220 hp). The machine was capable of generating a maximum thrust of about 2,710 tons. Table 1. Specifications of Tunnel Boring Machine (TBM) model 323-288. Main bearing Cutters Number of disc cutters Max. recommended individual cutter load Cutterhead Recommended normal operating thrust Max. thrust Cutterhead drive Cutterhead power Cutterhead speed Approximate torque (high speed) Approximate torque (low speed) Thrust cylinder boring stroke Gripper pads Size of pads Maximum pressure on tunnel wall Electrical system Motor circuit Lighting system Secondary voltage

Total power installed (TBM and trailing gear) Machine conveyor Width Capacity (approximate) Belt speed Weight (approximate) TBM Weight of TBM including trailing gear (approximate) Heaviest piece Length of TBM Length of TBM including trailing gear

427 Copyright © 2004 Taylor & Francis Group plc, London, UK

Three roller (3-roller) Series: 17 (432 mm) 68 267 kN (27 tons approximate) 18156 kN (1816 tons) 27101 kN @ 34.5 MPa Electric motors/air clutches 3465 kW (11 315 kW) 5.07 rpm (high) 6344 kNm 9516 kNm 1.83 m 6.44 1.95 m each (2 Nos.) (measured on arc) 2.84 MPa (28.4 tons/m2) 660 VAC, 3 phase, 50 Hz 11000 V. 50 Hz 660 VAC, drive motors 400 VAC, hydraulic pump motors 3 2500 kVA 2500 kVA (including 500 kVA for TBM 500 kW (approximate) 1370 mm 1500 tons 122 m/min. 925 tons 1500 tons 96 tons 25 m 495 m

09096-52.qxd

2/18/04

7:58 PM

Page 428

At the time of manufacture, the TBM was the most powerful machine ever built. For the installation of rock support, the machine was equipped with the following features: – Ring beam erector to allow the erection of steel sets under the “finger shield” immediately behind the cutter head. – Two hydraulic drills COP 1032, one of each mounted on either side of the main beam and behind the ring beam erector for rock bolting and probe hole drilling. – Two hydraulic drills COP 1032 mounted at the front end of the TBM backup (BU) for supplementary rock support. – Shotcreting robot on the BU for the application of dry shotcrete for support and lining covering 270° circumference over the arch. As high water inflows were expected during the excavation of the tunnel, the BU designed by Rowa in Switzerland, featured a 90° open invert which allowed: – Unrestricted flow of water away from the face and TBM. – Invert clean-up prior to the placement of the rail track by means of a rubber tired loader via a vertical conveyor back-up onto the main trailing conveyor. To prevent damage to conveyor system, boulders which had passed through the muck buckets in the head were diverted to the invert by means of a vibrating screed mounted above the trailing conveyor, where they could be picked up by the loader. The 40 m long double story BU-1 section was supported by hydraulic legs off the tunnel haunches allowing it to step over tunnel ribs where necessary. The remaining BU section carrying the trailing conveyor run on a monorail system suspended from the tunnel roof. It became apparent during the excavation that this configured TBM/BU system suited the tunneling conditions actually encountered very well.

spaced at 1.5 m circumferentially and 1.8 m longitudinally. To reduce the number of bolts installed in intact rock mass class, a subclass of Type T1 was developed during construction. This subclass was called Type T1 Intact and contained only six bolts at 1.75 m spacing around the circumference. A total of 6,023 m or 62% of the tunnel was supported with Type T1. The rock bolts used on the project were epoxy coated 32 mm diameter bolts installed with resin capsules. Type T1A: This ground support type consisted of pattern rock bolts in the crown and side walls, and possible shorter 1.2 m long supplementary bolts in the invert. The pattern bolts in the crown and side walls consisted of 12, 3 m long bolts spaced at 1.8 m circumferentially and 1.2 m longitudinally. Shotcrete and welded reinforcing fabric (WRF) were applied to areas where overbreak had occurred as a result of spalling. A total of 832 m or 9% of the tunnel was supported with Type T1A. Type T2: This ground support type consisted of pattern rock bolts in the crown and side walls, and shorter 1.2 m long supplementary bolts in the invert. The pattern bolts in the crown and side walls consisted of 12, 3 m long bolts spaced at 1.5 m circumferentially and 1.2 m longitudinally. Shotcrete and WRF were applied around the tunnel perimeter. A total of 1,835 m or 19% of the tunnel was supported with Type T2. Type T3: This ground support type consisted of 200 mm deep expanded steel sets spaced at 1.2 m centers with WRF for lagging. The final lining consisted of nominal 375 mm thick cast-in-place concrete. A total of 923 m or 10% of the tunnel was supported with Type T3. In ground support terms, the tunneling conditions were better in Tunnel 2 than anticipated with more Type T1 and less Type T2 installed.

5 PROBLEMS ENCOUNTERED 5.1

4 GROUND SUPPORT The rock mass classification system modified from the Terzaghi rock mass classification system was used to predict the ground conditions in Tunnel 2 from the Tunnel 1 geological logs. Without more detail on the discontinuities in Tunnel 1 and other data such as RQD, it was considered that other rock classification systems such as Q or RMR could not be used. The rock mass classification system and ground support types are presented in Table 2. Type T1: The ground support system consisted of a pattern of rock bolts in the upper 120° of the crown. The pattern contained eight, 3 m long rock bolts

Blocky ground conditions were encountered over a total distance of some 2,700 m or 28% of the tunnel. The rock in these sections was stabilized using Type T2 and Type T3 support (for details of the support classification system see Section 4). While only minor or moderate overbreak was encountered in the majority of T2 ground, the installation of rock support was time consuming and made it necessary for the TBM to stop after each stroke to allow rock support to be completed (except shotcrete) prior to advancing the machine. As discussed in Section 3, less Type T2 ground support was used than anticipated, and hence less blocky rock was intercepted than anticipated.

428 Copyright © 2004 Taylor & Francis Group plc, London, UK

Blocky ground

09096-52.qxd

2/18/04

7:58 PM

Page 429

Table 2. Tunnel No. 2 ground support types and rock mass classification. Tunnel No. 2 stabilization types

Rock mass classification

T1 62% of Tunnel

Intact

Summary rock mass geological description – Strong, unweathered rock, uniformly textured with tight joints at a spacing of greater than 1.0 to 1.5 m. – Will also intercept rock containing shearing zones less than 0.6 m in thickness and the rock mass on either side of these features will be more closely jointed (shear zone treated with supplementary stabilization). – Strong, unweathered to slightly altered rock, with texture ranging from massive to well foliated or banded. Joints approximately 0.3 to 1.0 m spacing and may be iron stained. – Will also intercept rock containing sheared zones less than 0.6 m in thickness and the rock mass for a meter either side of these features will be more closely jointed (sheared zone treated with supplementary stabilization).

Moderately blocky, schistose

T1A 9% of Tunnel

Spalling ground

– Generally occurs (but not exclusively) east of the Mica Burn Fault in intact to moderately blocky ground.

T2 19% of Tunnel

Blocky and seamy

– Strong, unweathered, occasionally altered to decomposed rock. Joints approximately 0.150 to 0.6 m spacing and often open containing calcite, iron oxide, epidote (light green mineral) and may be slickensided (“slippery”). – Rock mass also contains sheared zones up to 0.6 m in width. – Fault and sheared zones 0.6 to 3 m in width, consisting of closely fractured rock with crushed and gouged seams. The rock mass for a meter on either side of the feature is typically moderately blocky or blocky and seamy.

Moderately wide faults or shears (0.6 to 3 m) T3 10% of Tunnel

– Fault and sheared zones greater than 3 m in width, consisting of closely fractured rock with crushed and gouged seams. The rock mass for 1 to 10 m or more on either side of the feature is typically moderately blocky or blocky and seamy.

Wide faults and shears (greater than 3 m)

In most T3 ground conditions, moderate to substantial overbreak was experienced. Quite often relaxed blocks and wedges of rock had to be removed from the crown before steel sets could be installed. Such operations were time consuming and at times caused damage to electrical and/or mechanical components of the TBM when pieces of rock crashed down to the tunnel invert from the crown 10 m above. On occasion, wedge shaped blocks were dislodged from the face or the crown in the blocky ground. These hard pieces of rock were often too large to fit into the muck buckets until they were broken up by the rotating cutter head. This caused considerable wear to the head face plate and damage to cutter mountings and bucket lips. In blocky conditions, cutter bearings had to absorb high impact loads when hitting voids where rock wedges had already dropped out from the face. As a consequence, an increased number of hub bearings seized in these conditions and the thrust on the TBM had to be reduced to control vibration and avoid further damage to cutters.

Over the duration of the drive, the TBM had to be stopped for more than two days at a time on 12 occasions to repair damaged cutter mountings and bucket openings. During such stoppages, wear plates around the cutter housing and hard facing were added to the head. On one occasion, the TBM was stationary for a three week period to complete major head repairs. Sharp boulders and pieces of steel originating from the cutter head caused substantial damage to the belt of the tunnel conveyor on at least three occasions, requiring the premature replacement of several lengths of conveyor belting. 5.2

429 Copyright © 2004 Taylor & Francis Group plc, London, UK

Groundwater inflows

The large groundwater flows were generally confined to water bearing shattered zones 100 to 2,000 mm wide and adjacent jointing that contained open features (1–20 mm). These features contained only minor sheared material and washing of fines was not a significant problem. When high groundwater flows were intercepted in the tunnel and by the drainholes this

09096-52.qxd

2/18/04

7:58 PM

Page 430

resulted in high heading flows (water flowing into the tunnel from the face to 30 m down the tunnel). For the majority of the tunnel, the heading flows were estimated to be less than 20 l/s (300 gpm). However, heading flows of greater than 125 l/s (2,000 gpm) were exceeded three times in Tunnel 2 when high groundwater bearing zones were intercepted by the TBM and drainholes. The maximum heading flow intercepted was between Station 4225 to 4285 (60 m). The heading flows over this interval of tunnel ranged from 174 to approximately 550 l/s (2,750 to 8,700 gpm). This heading flow was the combined flow from drainholes and flow from the face, which intercepted two water bearing features. The sustained cumulative portal flows peaked at approximately 1,000 l/s (15,840 gpm). The total cumulative portal flow increased sharply from 378 to 1,000 l/s from Station 4020 to 4390 where several water bearing features were intercepted. This equates to an increase in groundwater flow of 1.67 l/s/m (0.13 gallons/second/foot) of tunnel. After this peak, the flows reduced to 700 l/s (11,100 gpm) at approximately Station 6000 and then remained relatively constant between 700 and 800 l/s (11,100 to 12,600 gpm) until the end of the tunnel. Groundwater was pumped into 400 mm diameter (dirty water) pipes and 600 mm diameter (clean water) pipes installed along the tunnel wall over the entire length of the tunnel. Thirty-eight sumps fitted with submersible pumps located in wet areas along the tunnel were used to deal with groundwater inflows.

In wet areas, considerable control measures consisting of drain pipes and panning had to be installed prior to the application of support and lining shotcrete. 5.3

Figure 7. Penetration rate vs. cutter load.

430 Copyright © 2004 Taylor & Francis Group plc, London, UK

Boreability

The contractor, FDI, filed a large claim for a lower than expected penetration rate in the first 1,770 m (Reach 1) of the tunnel. The estimated penetration rate used at the time of bid was 1.49 m/hour at a cutter load of 27 tons. The contractor actually achieved a penetration rate of 1.25 m/hour (4 mm/revolution) with an average thrust of 22 tons/cutter. In order to evaluate the contractor’s claim, Meridian Energy commissioned the most extensive suite of linear cutting tests ever performed for a tunneling project. Seven large rock blocks (0.5 m3 each) were removed from the tunnel walls and shipped to the Colorado School of Mines, Earth Mechanics Institute for testing. The linear cutting machine (LCM) allows one to set depth of cut and measure the cutter loads. Cutter spacing in the field is replicated by setting the same distance between cutter passes. The cutters used in the test were taken from the job site and testing was also done with both a new and worn cutter to replicate field conditions. Figure 7 shows the relationship of penetration rate vs. cutter load for hard rocks developed by many years of field experience and data by Atlas Copco Robbins Company (Janson, 1995). The curves show a linear relationship beyond a threshold thrust of

09096-52.qxd

2/18/04

7:58 PM

Page 431

17 tons/cutter. Gneiss boreability lies within a range or band of boreability. Underneath the gneiss band, indicating more difficult boring respectively, are granites and mafic rocks such as diorite and gabbro. Above the gneiss band lies phyllites, a schistose rock whose steep linear relationship displays that it cuts quite easily. The Manapouri linear cutting test results are superimposed on Janson’s graph on Figure 8. Samples which were tested at various penetration (or cutter loads) are plotted so that the linear relationship between cutter load vs. penetration could be investigated. The samples were assigned the rock names given by the contractor in his claim. Paragneiss (a gneiss formed by metamorphosis of sedimentary rock) was a classic banded gneiss and is plotted by diamonds on the graph. The cutter relationship was a steeper slope than the standard gneiss range, indicating it is on the easier end of the scale of cutting for gneiss. The rock chips broke out on foliation controlled partings during testing. This rock made up about one-third of Reach 1. Metadolerites (fine grained gneiss) appeared in the tunnel as mafic bands both concordant and discordant to foliation and appeared to be metamorphosed dikes. On Figure 8 the metadolerite data is shown as dots. The rock plotted, as expected, near the historical

database for mafic rocks and was more difficult to bore than a classic gneiss. The metadolerite makes up approximately 14% of the rock excavated in Reach 1. One of the strongest rocks at the site was calcsilicate (up to 306 MPa). The rock had the appearance of a greenish quartzite. The high strength and difficult boreability could be attributed to an epidote (a very hard, pistachio green, mineral) content of about 30%. The calc-silicate baseline UCS strengths were exceeded and the contractor was awarded a DSC for calc-silicate strength. Another primary rock type encountered in Reach 1, granitic gneiss (non-banded gneiss) is also plotted on Figure 8. The inverted triangle represents the as-built, actually achieved, penetration rate of 4 mm/revolution (1.25 m/hour) at an average cutter load of 22 tons. The large X demonstrates the contractor’s pre-bid estimate of 1.49 m/hour (5 mm/revolution) at 27 tons/cutter. Figure 9 shows a linear regression line based on a weighted average of all LCM data. It falls directly on the lower bound line for gneiss. The actual penetration rate achieved (inverted triangle) lies above this line because the LCM data are from intact rock blocks, whereas the actual field data include the beneficial effect of rock mass jointing. The LCM data plot close to or slightly above the contractor’s pre-bid estimate indicating the rock mass bored as expected.

Figure 8. Linear cutting machine results.

431 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-52.qxd

2/18/04

7:58 PM

Page 432

Figure 9. Linear regression of LCM data.

The pre-bid penetration estimates were not achieved simply because the pre-bid assumption of 27 tons/cutter was not utilized during construction of the tunnel. The reasons for not utilizing maximum thrust during driving of the tunnel through massive to normally jointed rock are still not clear. However, thrust reduction was required in blocky rock in order to limit vibration and damage to the TBM.

6 CONCLUSIONS

cutter head and downtime to a TBM. When excavating blocky rock lower cutter loads are required to minimize damage to the TBM. ACKNOWLEDGEMENTS The authors thank Meridian Energy, Ltd. for permission to publish this paper.

REFERENCES

The contractor traversed many obstacles in completing this large diameter tunnel. Although the rock was strong to very strong, LCM tests confirmed that it exhibited normal boring relationships of thrust vs. penetration for a gneiss. In preparation of pre-bid estimates, it is important to evaluate the estimated thrust that will be utilized during construction. Drainhole drilling was an effective method for controlling very high groundwater flows in a TBM tunnel. When planning future large diameter TBM tunnels in very high strength rock, the understanding of the occurrence and consequences of blocky rock conditions is very important. Loosened hard blocks at the face can cause considerable structural damage to the

Janson, H. 1995. Hard Rock Tunnel Boring Machines, Colorado School of Mines Mechanical Tunneling Short Course, October 2–4, 1995. Martin, Tom, Heer, Brian (deceased) and Macfarlane, Don. 2003. Second Manapouri Tailrace Tunnel Not Just “A Walk in the Park.” Rapid Excavation and Tunnel Conference, 2003 Proceedings.

432 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-53.qxd

2/18/04

2:13 PM

Page 433

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

South Austin Regional Waste Water Treatment Plant Interconnect Tunnel Project Saad Cheema Brierley Associates LLC

Kevin Koeller Turner, Collie & Braden, Inc.

Randy Pohren City of Austin

Gregg Sherry Brierley Associates LLC

Ross Webb Affholder, Inc.

ABSTRACT: This paper discusses a case history of a unique tunneling project involving two different construction techniques in soft ground conditions. The challenging part of this project was the connection to existing facilities without disturbing or interrupting the process flow of an operational wastewater treatment plant. The tunnels were constructed with a single shielded TBM and a Roadheader in clay and clay shale. Due to extensive work during design phase and a skilled contractor, the project was completed accident free, under budget, ahead of schedule, and without any claims for different site conditions.

1 INTRODUCTION

2 BACKGROUND

In the mid 1990’s, the City of Austin began a process to expand its existing South Austin Regional Wastewater Treatment Plant (SARWWTP), from a 50 million gallons per day (MGD) to a 75 MGD treatment facility. This expansion project also included improvements to existing facilities required to support the total expanded treatment facility. Turner Collie and Braden served as lead consulting engineer for preliminary engineering, design and construction phase services. The SARWWTP Expansion and Improvement Project included seven separate phased construction projects with a total estimated construction cost of $100 million. The second of these construction projects was the Lift Station Interconnect Tunnel Project. An extremely challenging design and construction tunneling project, it was completed to project specific requirements ahead of schedule and under budget. This paper discusses the history of this project from design through construction.

To meet the needs of the SARWWTP operational and maintenance requirements, minimize the need for clarifications and change orders during construction, and generate more competitive bids during construction, the City of Austin required and implemented a higher level of design standards. Termed “detailed design,” the entire SARWWTP Project, including the Lift Station Interconnect Tunnel Project, was designed to this higher standard. The design team worked to this high standard, anticipating unknowns and defining project specific requirements. The SARWWTP receives flow from the southern half and downtown areas of Austin via two tunnels. The 84-inch Onion Creek Tunnel, serving southern Austin, terminates at Lift Station No. 1, and the 96inch Govalle Tunnel, serving downtown Austin, terminates at Lift Station No. 2. Both of these on site existing lift stations are massive circular structures of comparable stature and design and have a wet-well dry-well

433 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-53.qxd

2/18/04

2:13 PM

Page 434

configuration. Lift Station No. 1 is 90-foot diameter, 86fi feet deep and the 84-inch Onion Creek Tunnel enters the lift station at a depth of 63.15 feet below grade. Lift Station No. 2 is 110-foot diameter, 97.75 feet deep, and the 96-inch Govalle Tunnel enters the lift station at a depth of 83.12 feet below grade. Similarly designed, each lift station consists of a wet well with two separate chambers connected to a common inlet distribution box. Sluice gates in each lift station’s inlet distribution box are used to isolate each wet well. The only access to each lift station’s distribution box is via an access hatch in the top slab of the lift station. Numerous underground utilities serve and surround each lift station, making it impossible to excavate or open cut near the lift stations. As part of the overall 25 MGD plant expansion, the design team confirmed flows from the SARWWTP service areas, but did not define the division of this flow between each of the two tunnels. It was uncertain which service area would grow during the 20-year design life of the facility; options for plant expansion were either to estimate flow in each tunnel or to interconnect the lift stations. The design team’s evaluation determined that there was adequate capacity for the 25 MGD plant expansion if the lift station wet wells were interconnected. The solution was to connect the two lift stations, treat both service areas as one large composite area and allow for the optimization of existing facilities. A main concern was the SAR plant operation, which has an average daily flow into both lift stations of 40 to 45 MGD with peak flows during wet weather events in excess of 150 MGD. Because of the requirement to maintain both lift stations in service during the tunnel construction, the design team developed the concept of making the connection to the existing lift stations from within the new tunnel. Then, the design team decided to construct two access shafts outside

the limits of the lift stations and tunnel between these access shafts and from the access shafts to each existing lift stations. 3 DESIGN CONSIDERATIONS After deciding to connect the two existing lift stations, the design team began work on the tunneling portion of the project. The professional design firms of Fugro South and Brierley Associates assisted Turner Collie and Braden with developing a geotechnical baseline report (GBR). Using the ASCE Publication, Geotechnical Baseline Reports for Underground Construction, as a guide, the GBR presented an interpretive summary of the results of geotechnical site investigations and provided a geotechnical basis for designing the project. To prepare the GBR, nine test boring logs were performed and included in the contract documents. Five of the test borings were drilled specifically for this project and four were included from previous geotechnical investigations at the site. Bore holes were drilled at the two access shaft locations, outside of each lift station, and along the pipeline route. The boring results were described in the GBR together with ranges and averages of water content, liquid limit, plasticity index, and consolidatedundrained triaxial compressive strengths. In addition, this geotechnical information was shown diagrammatically on Subsurface Profile Drawings included in the contract documents. These Subsurface Profile Drawings represented an alignment stratigraphy for each of the tunneling sections on the project and indicated the type of material where the tunneling would likely take place. The GBR summarized and interpreted the results of geotechnical investigations, discussed the project site in terms of the project geologic setting, and anticipated soil, bedrock, and groundwater conditions

Lift Station No. 1

Lift Station No. 2

Access Shaft

Access Shaft Interconnect Tunnel

Figure 1. South Austin Regional Waste Water Treatment Plant.

Figure 2. Project site sketch showing routing of the interconnect tunnel.

434 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-53.qxd

2/18/04

2:13 PM

Page 435

within the construction zone of the proposed tunnels. The GBR also included baseline conditions for groundwater and estimated inflow of water during tunneling operations. The contractor was advised to make provisions to control the water during tunneling operations to prevent unstable soil behavior during tunneling. 3.1

Excavation concerns

In the GBR, tunnel excavation was identified with different types of ground excavation depending upon the segment of tunnel being constructed. For the two tunnel segments between each of the new access shafts and the existing lift station, some type of “hand” excavation was required because there was no access to remove a Tunnel Boring Machine (TBM) at the lift stations. Furthermore, the design identified numerous concerns with the soils in the vicinity of each of the existing lift stations. During the construction of the existing lift stations in the 1980’s, a large volume of existing soil material was disturbed. The existing lift stations were constructed using sloped excavation methods, formed cast in place concrete construction, followed by backfilling the excavated areas around the perimeter of each lift station. In order to make the connections to the existing lift stations, the tunnels had to advance through this material; however, the limits of the backfill and the quality of these backfill operations were unknown. To address this problem, the design team estimated and identified approximate zones of excavation. The GBR advised bidders that the existing backfill material might contain concrete rubble, wood scraps, and other materials, including a solid wall of sheet piles and that granular material and coarse gravel, also might exist from the exterior face of the concrete out a distance of 3 feet for the entire depth of the lift station. These unknowns about the backfill material adjacent to the existing lift stations required that a tunnel excavation methodology be developed and included in the GBR. The design team reviewed several different methods to stabilize this backfill and selected the method of compaction grouting. To provide stable ground conditions and reduce the amount of water inflow into the face of the excavation, the tunnel excavation methodology specified to stop a distance of 40feet from the existing lift station wall and perform compaction grouting. Once the grouting was completed, four probe holes were to be drilled into this grouted zone to check the condition of the stabilized ground and serve as drain holes during tunneling. The other tunnels on the project between each of the access shafts were designed using an Earth Pressure Balance Machine (EPBM). Specifications were developed for both tunneling segments, the clay shale and the terrace deposits, including lean and fat clays and sand layers. The design team anticipated

that, for the terrace deposits, soil conditioning would be required to prevent the sand layers and fat clays from clogging the TBM. The design team also specified face control during construction. The clay shale was found to be relatively stable, but there was a tendency for the clay shale to deteriorate and/or swell over short periods of time. The design team prescribed that, if tunneling was stopped for any reason, completely shoring the face of the excavation would be necessary to prevent this deterioration. The design team’s previous experience with tunneling projects, led the designers to provide an overview of proposed construction, anticipate subsurface conditions, and emphasize information obtained during the subsurface exploration program that would impact tunnel construction. The GBR established baseline conditions for subsurface ground conditions in and around the proposed construction horizon. The GBR also proscribed how the contractor would evaluate excavation requirements, shoring, and tunneling to complete the project, evaluate actual geotechnical conditions, and determine which conditions qualified as a “Changed” or “Differing Site Conditions.” 3.2

Figure 3. Interior of wet well at Lift Station No. 2.

435 Copyright © 2004 Taylor & Francis Group plc, London, UK

Connection to lift stations

The requirement to maintain operation of the lift stations and the inability to stop flows into the existing lift stations led to the development of a detailed lift station penetration procedure, called a sequence of construction. This procedure outlined operational constraints of the plant and implemented a format to communicate the contractor’s plan for executing the work. During the design phase of the project, items of concern included minimizing interruption of plant operations, avoiding damage to operating facilities, wall thickness of the existing lift station, radial penetration, and gases. The contractor was to submit a debris protection plan that avoided interference with the operation of the lift station and prevented any debris from entering into the influent well of the lift

09096-53.qxd

2/18/04

2:13 PM

Page 436

station. During tunnel construction and high flow conditions, the water level in the wet wells and influent well would rise above the elevation of the tunnel penetration. Consequently, a flood protection plate was designed to be placed on the inside the wet well to cover the location where the tunnel penetrated the wet well inlet distribution box. 4 CONSTRUCTION The final design of the tunnel was completed in May 2001, and the project was advertised for bidding shortly thereafter. The Engineer’s cost estimate for the project was $3.97 million. Because of the complexity of the project, the bid form required the bidder to provide evidence of a minimum of three successfully completed similar projects in excess of $2,000,000 completed in the past five years. A mandatory pre-bid meeting was held at the site to advise bidders of specific contract requirements. Bids were received from five contractors and Affholder, Inc. of Saint Louis, MO was the low bidder. It should be noted that the low bid was below the Engineer’s estimate and the difference between the low bid and second low bidder was slightly over $200,000.00 on a $3.725 million project. The notice to proceed was awarded to the construction contractor, Affholder, on December 3rd, 2001. Shortly thereafter, Affholder mobilized equipment and construction materials on site. 4.1

Construction of shafts

The access shafts were excavated using conventional mechanical methods. A 4-foot high collar was constructed at each shaft location before starting the excavation. The shafts were excavated in 4-foot lifts and were supported by liner plates and 8-inch steel ribs at 4-foot centers. A concrete mud slab was constructed to prevent soil deterioration and to provide a firm-working surface for tunneling activities. To minimize groundwater inflow through the terrace deposits, the shafts were grouted behind the liner plates. Any groundwater seeping through the liner plates of each shaft was collected in the sump and pumped to the surface for proper disposal. 4.2

pressure was balanced with pressure applied by the excavated soil on the exposed ground. Muck was removed with a 300-degree muck ring mounted in the center of the cutterhead, transferred through the pressure relief gates to a conveyor, and transported to the rear of the machine where the muck cars were filled. An electric locomotive moved the muck cars to the shaft where a crane hoisted the cars to the surface for final disposal. After constructing a backstop and launch pad, the tunnel boring machine was lowered into the access shaft, set in place, and aligned. The machine then pushed off the backstop until the tail shield was buried. The machine excavated approximately 20-feet of starter tunnel before the pipe jacking system was lowered into the shaft and installed against the backstop. The first segment of pipe was attached and sealed to the tail ring of the machine with a gasket to control water infiltration in the tunnel. The machine advanced in the tunnel with a 5-foot stroke and pipe was then jacked to complete one cycle of operation. The jacking frame had three jacks capable of producing thrust of 250 tons each and a push ring to distribute the jacking force uniformly to the pipe bell being jacked. The jacking system used for pipe jacking

Figure 4. Jacking and boring operations.

Construction of tunnels by TBM

Construction of the tunnels between the access shafts was performed by pipe jacking method, by using Single Shield 5.75-foot diameter Tunnel Boring Machine (TBM) with 63-inch diameter Hobas pipe jacked behind the TBM. The machine, Decker Model 70, was equipped with a muck ring and pressure relief gates. In closed mode, the cutterhead was allowed to fill up with soil to maintain face pressure. The earth

Figure 5. Tunnel boring machine upon completion of tunnel operations.

436 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-53.qxd

2/18/04

2:13 PM

Page 437

Construction of each tunnel from each access shaft to each of the existing lift stations was performed by two-pass method using an Alpine Roadheader. After completing a 8-ft diameter and 6 linear feet of length starter tunnel with a Mini-excavator, the Roadheader was lowered in the access shaft. The muck from the face of the tunnel was transported to the rear of the Roadheader by the apron and conveyor system of the machine. The spoil was placed directly into the two

cubic yard muck box behind the machine. The muck box was transported through the tunnel with a 2-ton battery locomotive riding on 12-gauge rail. The filled muck box was brought to the shaft area and lifted to the surface by a crane. As with the other tunnel segments constructed, this tunnel’s line and grade was controlled by a laser and fans located at the access shaft provided tunnel ventilation. Steel ribs and wooded lagging supported the majority of the tunnel; however, as the excavation neared the lift stations, the tunnel diameter increased from 8-feet to 10-feet and transitioned to steel liner plates. The first two pass tunnel constructed was the segment to Lift Station No. 2. Intact clay shale was encountered along the length of the tunnel. At approximately 40 feet from the lift station, a single probe hole was drilled 30 feet deep that indicated no change in ground conditions. Excavation proceeded 20 feet further, where, approximately 20 feet from the wall of the lift station, 11 additional probe holes were drilled. Results from these probe holes indicated intact clay shale to the wall of the lift station and no groundwater. The project team decided that because no fill material was encountered to inject lowpressure grout to fill and stabilize possible unstable ground conditions. After completing the grouting operations, excavation continued to the lift station wall. It was at this time that it was identified that the lift station was cast directly against the clay shale. Affholder then installed the flood protection plate inside the wet well of lift station No. 2 and began breaking through the 4-foot wall of the existing lift station. The Roadheader was removed and transported to the other access shaft for the next segment of hand mining. The two pass tunnel to Lift Station No. 1 was constructed using the Roadheader alpine mining machine in a similar manner to tunnel to Lift Station No. 2. Intact clay shale was encountered along the length of

Figure 6. Alpine mining machine for construction of twopass tunnels.

Figure 7. Compaction grouting operations near Lift Station No. 1.

had a stroke length of 8 feet with each pipe segment 10 feet long. A 4-foot Hobas pipe segment served as spacer to fill the gap to jack the entire segment in the tunnel. The 10-foot pipe segment was jacked in twocycle operation. During the pipe jacking operation, bentonite was pumped in the annular space created by the over cutting of the TBM in holes provided in the pipe for grouting. This slurry was pumped, as needed, to reduce friction and maintain the opening around the pipe being jacked. The advance rate of the cycle was dependent on the haulage operation during the excavation of the tunnel. In the beginning, a single muck car was operated until the conveyor system on the machine was modified to accommodate two muck cars in a single pass. Fans were installed at the top of the access shafts and an 8inch diameter line continuously ventilated the tunnel. Line and grade of the Tunnel Boring Machine and pipe were controlled with the use of a laser. A target on the TBM gave the operator visual position with respect to the laser. A second laser was mounted on the machine head to indicate the position of the head. These methods resulted in the 768-foot tunnel between access shaft No. 1 and No. 2 being constructed in 19 working days and the 524-foot tunnel between access shaft No. 2 and Junction Box A, being completed in 16 working days. 4.3

Construction of tunnels by roadheader

437 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-53.qxd

2/18/04

2:13 PM

Page 438

the tunnel and at a distance of approximately 40-feet from the wall of the lift station, four probe holes were drilled. Results from these probe holes indicated intact clay shale a distance of 25-feet and then a material transition to a red sandy clay, the apparent backfill around the lift station. The project team decided to excavate to a distance of 10-feet from the transition to the backfill material and perform additional probing. Then, a total of 12 probe holes were drilled to the depth of the lift station, where a hard intact surface was encountered. This probing confirmed the transition from clay shale to backfill at an approximate 45 degree slope. Although some ground water was initially encountered, these probe holes quickly dried up. As outlined in the project requirements, compaction grouting was then performed. Shortly thereafter, probing and excavation followed and mining proceeded through this stabilized backfill material to the wall of the lift station. The two pass tunnel to Lift Station No. 1 was completed in 37 working days. The two pass tunnel to Lift Station No. 2 was completed in 32 working days. 4.4

Connections to the existing lift stations

As noted earlier, upon completion of the hand mining at Lift Station No. 2, Affholder installed the flood protection plate on the inside of the wet well. During hand mining operations to Lift Station No. 1, the contractor had a separate crew penetrating the 4-foot wall of the existing lift station. At Lift Station No. 1, a similar operation occurred. After breaking through these existing walls, new dowels and reinforcement were installed and a single section of Hobas pipe was placed at proper line and grade and cast in place with concrete. The penetration at each of the lift stations required close coordination with plant personnel to avoid any interruptions to plant operations and to insure the safety of Affholder’s underground work crews. Plant

Figure 8. Breaking through the 4-foot thick wall of Lift Station No. 2.

operations notified Affholder when the water level in the lift stations rose to within several feet of the new tunnel. A few times during construction, Affholder’s crews vacated the tunnel during high water levels in the lift station. It is noteworthy that the project was completed without a single accident, a testimony to excellent project coordination and the skill of Affholder’s work crews. 4.5

Upon completion of tunnel connection at lift stations, steel angle segments were installed at grade and the invert of the tunnel segments grouted. Then 20-foot long Hobas pipe was placed on this invert and pulled into place using a hydraulic tugger. Each pipe segment was blocked to protect it from floating during the grouting process. Upon completion of the placement of the tunnel carrier pipe, it was grouted into place. 4.6

Monitoring during construction

As stated earlier, each of the existing lift stations were surrounded by underground utilities. During construction, tunnel monitoring was performed to identify and control settlement of existing facilities resulting from tunnel construction. Throughout the construction project, these devices were monitored and no movement was recorded. Plant operations continued without interruption throughout all construction activities. 5 CONCLUSION As of November 2003, the lift station interconnect tunnel has been in service for nearly one year, and the City of Austin is extremely pleased with its operation. The project’s design and construction teams met and

Figure 9. Removing the flood protection plate in Lift Station No. 1.

438 Copyright © 2004 Taylor & Francis Group plc, London, UK

Pipe placement (two-pass tunnels)

09096-53.qxd

2/18/04

2:13 PM

Page 439

overcame complex challenges and to the credit of those involved, the project was constructed accident free with very minimal changes, ahead of schedule, and under budget. The project’s success resulted from the following major factors. First, an experienced project team produced a quality set of construction documents where they defined expectations, established project specific requirements, and identified unknowns. The contractor then constructed this project in accordance with these documents with the exception of several minor improvements. Second, because the ground conditions during construction were almost exactly

those predicted in the GBR, the project proceeded as designed by the Owner and Engineer and as bid by the contractor, Affholder. Last, teamwork, evident throughout the project, ensured that potentially difficult decisions were solved with efficiency and relative ease. All team members respected the objectives of the project team and other team members. In addition, Affholder’s skilled crews were experienced and knowledgeable of the project requirements. As detailed in this paper, the South Austin Regional Wastewater Treatment Plant Interconnect Tunnel Project is evidence that proper planning, detailed design, and effective teamwork pay off.

439 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-54.qxd

2/18/04

7:59 PM

Page 441

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Tunneling through an operational oilfield and active faults on the ECIS Project, Los Angeles, CA, USA Eric Keller City of Los Angeles, Bureau of Engineering, Los Angeles, CA, USA

Matthew Crow Parsons Brinkerhoff/Brown & Root Services J.V., Los Angeles, CA, USA

ABSTRACT: A 2.5 km (1.8 mile) long tunnel has been driven by a 4.7 m (15.5 ft) diameter EPB-TBM through sands and gravels of the Lakewood formation and very stiff massive clays and silts with minor occurrences of sand of the San Pedro formation. The Unit 1, western section, of the East Central Interceptor Sewer (ECIS) project tunnels beneath major utility lines with shallow cover, crosses the seismically active NewportInglewood Fault Zone beneath Baldwin Hills, and navigates through the Inglewood Oil Fields. The paper describes the hazard assessment and construction considerations of tunneling through faults and outlines risk avoidance planning to avoid abandoned oil wells. The paper then describes the performance of the TBM in the different ground conditions encountered, the forward probing method used to monitor for gas and the search for existing oil wells using a magnetometer both in a probe hole from the tunnel and also from the surface. Procedures are included describing the re-abandonment of an existing oil well discovered elsewhere on the project, and how successful implementation enabled the project to proceed safely.

1 INTRODUCTION 1.1

ECIS project

The NOS-ECIS alignment presented in Figure 1 is approximately 18.5 km long (11.5 miles), and extends from the NORS connection located in the Baldwin Hills area of Culver City, westerly along Exposition Boulevard towards downtown Los Angeles. The eastern end of the alignment terminates just east of the

Los Angeles River, near the intersection of Mission Road and Jesse Street. The alignment is primarily located within the densely developed urban area of central and south central Los Angeles. The tunnel is divided into four construction units with start and end points corresponding to locations of working and retrieval shafts used for tunneling. A general presentation of the design is presented elsewhere by Hanks et al. (1999). Aspects of construction for portions of the project are presented by Crow et al. (2003), and Budd & Goubanov (2003). This paper addresses aspects of tunnel construction on Unit 1. 1.2

Figure 1. Plan of ECIS sewer tunnel alignment.

The tunnel was driven 2.5 km (8200-ft) from the Siphon Structure near the intersection of Jefferson Boulevard and LaCienega Boulevard to the North Outfall Replacement Sewer (NORS) connection structure (Figure 2). The tunnel drive was downhill at a constant gradient of 0.12% from east to west over this reach. The EPBTBM used for mining the tunnel was manufactured in Toronto, Canada by Lovat Inc. and commissioned as “Angie”. The TBM was 4.72 m (15.5-ft) in diameter and weighed approximately 283 tonnes (624,000 lbs).

441 Copyright © 2004 Taylor & Francis Group plc, London, UK

Unit 1

09096-54.qxd

2/18/04

7:59 PM

Page 442

Figure 2. Exposed previous shaft shoring around the existing NORS connection structure.

Figure 4. Launch of Unit 1 TBM at Siphon Outlet Shaft from staging above Unit 2 TBM servicing arrangements.

Figure 5. Longitudinal section of ECIS project.

2 UNDERGROUND CONDITIONS 2.1

Figure 3. Lovat EPB-TBM, “Angie”.

Along the alignment, one maintenance hole was constructed 1680 m (5512 ft) from the NORS connection shaft, which served as a ventilation shaft. Cal-OSHA required the installation of an emergency rescue chamber after a maximum distance of 1524 m (5000-ft) without a ventilation shaft. All excavated soil was removed from the Siphon Outlet Shaft (Figure 4) at the intersection of LaCienega and Jefferson Boulevards.

2.2

Geological setting

The project is located in the northern margin of the L.A. basin (Yerkes et al., 1965), described as a 75 by 20 km lowland coastal plain that slopes south and

442 Copyright © 2004 Taylor & Francis Group plc, London, UK

Topography

The longitudinal section of the tunnel alignment is presented on Figure 5. The tunnel alignment on Unit 1 is beneath the steep hilly terrain of the Baldwin Hills from the intersection of La Cienega Boulevard and Rodeo Road to the NORS connection over a distance of about 1.8 km (5905-ft). The alignment of the eastern section of Unit 1 is beneath ground of low relief. The tunnel invert is approximately 16 m (52.5-ft) below ground at the eastern end (siphon outlet) and approximately 23 m (75.5-ft) below ground at the NORS connection. At the deepest location beneath Baldwin Hills, the tunnel invert is 112 m (367.5-ft) below ground surface.

09096-54.qxd

2/18/04

7:59 PM

Page 443

west. Figure 5 shows the profile and the geological conditions through which the tunnels were driven. Additional information on the geotechnical aspects of the project are presented elsewhere in the proceedings by Seeley (2004). 2.3

Predicted geological conditions

The initial 230 m reach of tunnel was expected to be driven entirely within the silty sands and silts of the Lakewood Formation with occasional cobbles and boulders. A 100 m reach around Station 2210 of mixed face conditions resulting from the transition into the San Pedro Formation from the overlying Lakewood Formation was anticipated. The tunnel was then to be driven entirely through the San Pedro Formation. The Baldwin Hills Fault was expected at Station 1690 and another Fault expected at Station 1810. Changes in lithology were to be expected across these faults accompanied by groundwater under a hydraulic head of up to 10 m above the tunnel invert. The Inglewood Fault was anticipated to be at Station 1510, where it would encounter the eastside of the graben that has down-dropped the San Pedro Formation and the overlying Lakewood Formation. This particular fault zone was expected to be roughly 3 m wide and to contain sheared clay silts and sands that are moist to wet. Change in lithology was expected from one side of the fault to the other. Within the fault, groundwater was likely to be found under a hydraulic head up to 20 m above tunnel invert. The tunnel was then expected to be within the Lakewood Formation west of the Inglewood Fault. The next fault to be encountered was at Station 1460, along the west side of the graben mentioned above. The faults at Stations 0095, 1160 and 1450 were anticipated to contain sheared clayey silts and sands, described as moist to wet. Again, changes in lithology were to be expected at these faults from one side of the fault to the other. Within this section, groundwater could be anticipated under a hydraulic head of up to 30 m. All of the faults were anticipated to be potential traps for methane, hydrogen sulfide and other gases associated with oil fields. The final 100 m reach of the tunnel was anticipated to be a mixed face of the San Pedro Formation overlain by the Lakewood Formation. As the contact between the San Pedro and Lakewood Formations dips to the west, the final 40 m was expected to be completely within the Lakewood Formation, which was previously encountered during construction of the NORS tunnel.

Mining across the area identified as the Baldwin Hills Fault (Station 1810), there did not appear to be any significant change in material. Methane gas registered at 15% and 18% LEL roughly 30-50 m past the proposed fault location. The gas occurrences were read by the handheld instrument at the discharge point of the TBM screw conveyor, typically only associated with the first TBM advance of the day, and quickly dissipated. The identified fault near Station 1690 manifested with an increase of groundwater, causing the annular grout to become excessively wet and blow through the tail seal brushes. Excavated soil between Stations 1618 to 1410 was described primarily as soft clay with fine sands. Near Station 1570, a pocket of methane gas registered a 55% LEL and quickly dissipated, accompanied by an increase in volume of cleaner groundwater. After mining through the Inglewood Fault (Station 1510), material did change to a sandy, silt clay within the zone of Stations 1470 to 1495. The increase in groundwater was enough to require pumping. The TBM was out of the groundwater condition by Station 1460. Near the Station 1160 fault, material was still described primarily as a clay with fine sand, with pieces of sea shells were noted as present in the excavated soil until Station 1115. The soil between Stations 944 through 831 was described as especially hard clay by the Inspector. From Stations 630 through 473, in some of the deepest sections from the surface topography, the very hard clay was described as “squeezing ground”. During probe hole drilling, the steel augers became almost impossible to remove requiring probe holes to again be drilled the following day. After the Station 0095 fault, the excavated soil became soft, silty clay with an increase in sand and later also gravel. One unique feature was the discovery of a preexisting void at Station 1290 during the normal probe hole drilling ahead for gas testing. The void encountered extended approximately 6 m (20-ft) horizontally from the TBM face. The probe hole was advancing through the same tight dry clay that had been the case for a number of previous probes. No water or gas was encountered when the void was found. PVC pipe casing was inserted into the probe hole and used for pumping approximately 4.6 m3 (6 cu yd) of grout to fill the void. The same grout normally used for the segment annulus was used to fill this void. 2.5

2.4

Underground conditions as-encountered

In general, the ground conditions encountered were similar to those predicted but with the following differences.

A 150 m reach of the Unit 1 alignment (Station 1900 to 2050) was beneath soil containing weathered gasoline and sewage. The contamination area was at the intersection of LaCienega Boulevard and Rodeo

443 Copyright © 2004 Taylor & Francis Group plc, London, UK

Contaminated soil

09096-54.qxd

2/18/04

7:59 PM

Page 444

Road where the tunnel passes beneath the existing North Outfall Sewer (NOS) and both existing and former gasoline service stations. The contract provided a unit price bid item for the excavation, handling and disposal of this material if necessary, including specific requirements for safe handling of the material. The tunnel horizon did not encounter any contaminated soil within the possible identified zone of 150 m. However, the unit bid item was utilized for the excavation, handling, and removal of oil contaminated soils within the ECIS-NORS connection shaft worksite area.

3 OIL WELLS 3.1

potential to produce methane gas as a by-product. From previous studies, it was known that prior to 1942 wells were abandoned with drilling mud rather than a deep 150 m cement plug, possibly allowing gas to migrate to the sewer invert and be encountered during construction of the tunnel. The alignment was set to avoid these wells, however it was expected that unrecorded or “wildcat” oil wells might lie on the tunnel alignment. Precautions were therefore taken to investigate for their presence. These unknown wells were not likely to have been abandoned to modern standards, so significant risk to tunneling, but also the risk of cost being incurred in treating the wells to current standards should that be required.

Determination of position of oil wells

The alignment passes through the Inglewood oilfield. This oil field produces from sandy reservoirs hundreds of meters deep, but oil and gas commonly seep towards the surface. Not only was there a possibility of intercepting these deposits, but there was also the possibility of encountering old wells. Prior to the award of the contract, the City produced a plan that identified the boundaries of the oil field, based upon the economic limits of ongoing oil production leases. Numerous oil and gas seeps were also reported to the City (Geotechnical Services and Street Services) over the years, manifesting the migration of oil and gas outside of the oil field along faults, fractures, bedding planes and within permeable sediments. The contract plans incorporated the location of both existing and abandoned oil wells on record with the State Department of Oil & Gas and Geothermal Services (DOGGR). An extract from the plans included as Figure 6 shows the tunnel alignment passing through the operational Inglewood oilfield; the black circles indicating oil wells. Facilities associated with past and present oil production, including sumps, drilling ponds, areas of past spills, storage tanks, exploratory and production wells may also be responsible for release of hydrocarbons into the subsurface soils. Moreover the presence of crude oil in the subsurface due to past leakage or spillage is a potential source of contamination with

3.2

3.2.1 Historical research A combined effort by the City Geotechnical Engineering and Construction Management Staff, the contractor, and a specialist oil field consultant was made to research information from the following sources:

• • • • • •

Figure 6. Tunnel alignment passing through the operational Inglewood Oilfield.

Survey and Operational records of all past and present wells on the oil field lease of Plains Exploration & Production, (PXP). Historic aerial photographs from University of Southern California archives. Los Angeles County area maps. Historic and current oil field maps (DOGGR) Topographic maps for features of possible past oil well working sites. ECIS alignment was surveyed, marked along the ground surface, and walked for visual surface indicators including graded areas and foundations.

3.2.2 Surface magnetometer survey The City commissioned MACTEC Engineering & Consulting, Inc. to perform a geophysical investigation to search for oil wells within a 12 m by 760 m corridor corresponding to the ECIS alignment through the

444 Copyright © 2004 Taylor & Francis Group plc, London, UK

Risk management

As the possibility of encountering an unrecorded oil well during tunneling had been identified as a significant project risk by the City and the Contractor, a joint task force worked with the State Department of Gas and Geothermal Reserves (DOGGR). Based upon a review of the DOGGR maps, it was anticipated that tunnel mining would be no closer than 6 m to any mapped oil well, and that there were likely to be some 18 oil wells within 45 m of the tunneling activities. However, to reduce risk of unmarked oil wells, historical documents were also researched and magnetometer studies were performed to search for possible oil wells, observed as magnetic anomalies from both the alignment surface and from within the TBM.

09096-54.qxd

2/18/04

7:59 PM

Page 445

subject oil field in Culver City. Specifically, the magnetometer would search for undocumented steel well casings. The magnetometer measures the magnetic field intensity using the hand held equipment as shown in Figure 7. This equipment consisted of a Geometrics Model G-858 magnetometer and GPS system (Trimble ProXRS) for horizontal positioning. A second base station magnetometer was installed at a fixed location to monitor the natural time-varying magnetic drift cycle and record any bursts of magnetic noise that might affect the survey data. Prior to the survey a test survey was performed of a known abandoned well location (TVIC-15) in order to establish a characteristic magnetic signature for abandoned wells within the survey area. Total magnetic field of the test magnetic anomaly was approximately 53,000 nanoTeslas (nT). In addition, a utility locating survey was performed to identify potential sources of magnetic interference (i.e. buried metallic utilities) so the associated responses would not be mistaken for indications of abandoned wells. The utility locating survey used a Radiodetection Corp. RD-400 radio-magnetic utility locating system and a Fisher Model TW-6M-Scope. 3.2.3 Results of surface magnetometer This information was processed and presented graphically in relation to the tunnel alignment as shown by the example in Figure 8. Fortunately, the survey indicated no wells in the tunnel alignment. The most significant magnetic anomalies appeared over a surface pad with some abandoned structures, which was considered as a possible previous oil well site. However, the signature of the anomaly was only indicative only of relatively small, shallowly buried

metal objects (Figure 8) dispersed over an 8 m area, and did not match the signature of the known oil well test survey. 3.3

Forward probing for gas, and tunnel magnetometer survey

Contract Specifications required the contractor to probe ahead of the TBM once gas might be encountered beginning at the Baldwin Hills Fault (Station 1810). The Unit 1 tunnel was classified as “Gassy” by the Cal-OSHA Mining & Tunneling Unit. Probing to perform gas testing was performed in accordance with the Cal-OSHA Tunnel Safety Orders, requiring a minimum of at least 6.1 m (20-ft) of tested ground to remain beyond the face of the TBM. Upon reaching Station 750, specifications required magnetometer surveys for the remainder of the drive to Station 000 in order to locate any possible abandoned oil well casings. After probe drilling roughly 58 m (190-ft) with 19 sections of 100 mm (4-in) drilling auger, the drill sections were removed and replaced with a 46 m (150-ft), 75 mm (3-in) PVC casing. The magnetometer instrument, a FVM-400 Vector Fluxgate Magnetometer by MEDA, was inserted in the casing and forwarded to the end of the casing with a fiberglass rod. The instrument was then backed out of the PVC casing at 0.3 m intervals and a data reading taken. Following data collection with the magnetometer, the same basic procedure was repeated but with the insertion of an inclinometer, specifically a Little Dipper by Applied Geomechanics. Data was stored in a Handspring Visor PDA and utilizing Palm OS-compatible software. The inclinometer provided information to determine deviation of the probe hole from the TBM to the end of the casing. Data collection from the magnetometer was downloaded to produce a graph of the magnetic field against the Station, for the x, y, and z components. A near constant slope of the graph downward after leaving the TBM influence area represented a “normal” condition. Actual plots depicting a spike in the slope on any

4.7m

Possible pipeline 250kg 1m bgl

Tunnel Centerline

Small buried metal object 25 kg 0.5m bgl Figure 7. Surface hand held magnetometer with GPS for horizontal positioning.

Figure 8. Magnetic total field contour map above tunnel alignment oil-well abandonment.

445 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-54.qxd

2/18/04

7:59 PM

Page 446

Magnetic Field X Component (nanoteslas) -10000

-20000

-30000

-40000 Magnetic Field Y Component (nanoteslas) 20000

10000

0

-10000 Magnetic Field Z Component (nanoteslas)

30000

20000

Face Station 696.0

Face Station 532.5

40000

Distance from TBM face (m)

Figure 10. Face of TBM after excavation of NORS shaft.

10

20

30

40

10

20

30

40

10000 Distance from TBM face (m)

Figure 9. Magnetic field graphs – xyz component for TBM face stations 0532.5 and 0696.0.

of the three components indicated a magnetic anomaly in a particular Station range. The left side of Figure 9 for the magnetometer probe readings forward of Station 0532.5 indicate apparent x & z anomalies, whereas the probe readings forward of Station 0696 depicted on the right indicate “normal” xyz. The total magnetic field at the station of an anomaly was determined separately by the contractor’s geophysicist, Spectrum Geophysics, after removing the Earth’s magnetic field and the field from the TBM from the data. The largest Total Resultant Magnetic Field was 27,646.0 nT. When anomalies were significantly present, the speed of the TBM was reduced and the mining operation proceeded ahead carefully. In some instances, an actual surface exploration was made along with review of DOGGR records to again check if there was any overlooked possibility of an oil well. Communication with the on site oil-production company, PXP, was also very helpful in the process of reviewing records. Mining of the Unit 1 tunnel was successfully completed on August 8, 2003. Due to site access problems, there was a significant delay in mobilization for NORS shaft excavation. To enable completion of the tunnel drive, the TBM was not driven into the shaft, rather the shaft was later constructed around the TBM. The completed shaft is shown in Figure 10. 3.4

Treatment of existing abandoned oil well

Although no oil wells were encountered during Unit 1 tunneling, an abandoned well was discovered at the other end of the project alignment as excavation began for slurry trench guide-walls at the Mission & Jesse work site (Figure 11). The abandoned oil well

Figure 11. Abandoned oil well at Mission & Jesse Site.

had been cut-off below the surface and as luck would have it, was located within the rectangular shaft site perimeter. Fortunately the oil well was not located in conflict with the actual slurry wall footprint. A more thorough review of DOGGR records was performed for oil fields along the western portion of the project alignment, but the single isolated well on the easternmost portion of the alignment went undetected. Removal of the oil well was critical as shaft design required excavation to 26 m (85-ft) below the ground surface. The shaft location could not be moved due to the confined nature of the site. Staff worked quickly with a DOGGR field engineer to identify the well, which was recorded abandoned in the 1940’s. However, to modify the well casing in any way would require the well to be treated according to present day standards for abandonment. By way of a change order, the contractor provided an oil field welder to remove the well cap and allow DOGGR to determine

446 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-54.qxd

2/18/04

7:59 PM

Page 447

Target of laser guidance system Thrust cylinder

Tool gap Plenum EPB pressure sensor

Annulus backfill grouting

Screw conveyor Bulkhead

Segment erector

Note:

Articulation cylinder Lower part of muck ring 0

8

6

10

12

UNIT 3W Av. = 492 m/month

UNIT 2 Av. = 353 m/month

UNIT 3E Av. = 268 m/month

14 Length [m]

180+00

170+00

160+00

150+00

140+00

130+00

120+00

110+00

90+00

100+00

80+00

70+00

60+00

50+00

40+00

30+00

UNIT 1 Av. = 247 m/month

UNIT 4 Av. = 265 m/month 2 x 10 hr shifts 1x 10 hr shift

Figure 14. ECIS tunnel production.

4 TBM PERFORMANCE 4.1

EPB TBM

The contractor, a joint venture of Kenny Construction Co./J.F. Shea Co. Inc./Traylor Brothers, Inc./FrontierKemper Constructors Inc. (KSTF-K) chose to use a TBM machine manufactured by Lovat. A cross section of the EPB machine is presented in Figure 13; however a comprehensive description of the machine is contained in Crow & Holzhauser (2003) and Budd & Goubanov (2003). 4.2

TBM production

Although the TBM performance for the remainder of the ECIS project was presented in Crow and Holzhauser 2003, a brief summary is provided herein. Tunneling of the 2.5 km long tunnel on this Unit started on October 13, 2002 from the Siphon Outlet Shaft to the East and was completed on August 8, 2003. An overall average progress rate of 247 m/month has been achieved.

447 Copyright © 2004 Taylor & Francis Group plc, London, UK

20+00

Oct-03 Aug-03 Aug-03 Jul-03 Jun-03 May-03 Apr-03 Mar-03 Jan-03 Dec-02 Nov-02 Oct-02 Oct-02 Aug-02 Aug-02 Jul-02 Jun-02 May-02 Apr-02 Mar-02 Jan-02 Dec-01 Dec-01

+

whether the previous plug was secure and that there was no gas leakage. The City enlisted a DOGGR approved consultant, Sampson Oil, to perform the oil well research, coordinate the permit application with DOGGR, and oversee the oil well abandonment process. The contractor hired subcontractor Oil Field Services, Long Beach, a capable contractor who was able to mobilize, drill out the existing well 215 m (700-ft), install a new cement plug, and demobilize within five days. The equipment used is shown on Figure 12. The cement plug of a mix proportion of 100 kg of cement to 45 liters of water was poured in two stages of total volume 20 m3 (725-ft3). Considerable uncertainty surrounded the treatment of this old well, as it might have been filled with discarded oil pipe, pipe retrieving tools, wood, or any other kind of debris available at the time to fill the hole in. Fortunately, the hole was fairly clean except for a difficult wooden plug, and drilling took fewer than three days as opposed to potentially 2–3 weeks. After the slurry trench reinforced concrete guide walls were completed followed by shaft excavation, the existing well casing was cut off in segments as the excavation progressed. In this situation, the well was finally cut-off roughly 1.5 m (5-ft) below the working level of the shaft and sealed with a welded steel plate cap. DOGGR and a representative from LAFD inspected the plugged well casing prior to welding of the final steel plate cover. The effort required to abandon this unmarked oil well properly provided the necessary motivation early in the project to further research the alignment of the Unit 1 tunnel for any other unmarked oil wells which might exist. Potential impact both to project schedule and budget was a real concern. If encountered, the process for abandonment of oil wells along Unit 1 would have required further efforts to obtain surface rights, locate and excavate for the well head, and provide access for the necessary equipment. Fortunately, the oil well at the east-end of the project was the only one encountered.

4

Level of EPB pressuresensors 1m above and below Here rotated into display

Figure 13. Cross-section of Lovat 4.72 m diameter EPB-TBM.

10+00

Figure 12. Treating abandoned oil well at Mission & Jesse site.

2

Belt conveyor

09096-54.qxd

2/18/04

4.3

Settlement

7:59 PM

Page 448

The surface settlement measured along Unit 1 in both granular and cohesive materials was less than 10 mm. Subsurface measurements using multi-point borehole extensometers were undertaken. In general, the lower anchor located at 1.5 m above the crown of the tunnel measured less than 32 mm settlement. 5 CONCLUSIONS

•

• • •

Tunnels can be successfully driven through active oilfields, but the importance of adequate site investigation, risk assessment, good planning and implementation of these plans undertaken should not be overlooked. Tunneling through faults in soft ground with a closed face TBM can be achieved with good site investigation, planning and co-ordination on the part of the owner, designer and contractor. Surface and forward probe magnetometers are useful tools in reducing the risks of tunneling through active oilfields, but they must be used with care and their records interpreted by experienced geophysicists The removal of the environmental nuisance of inadequately abandoned oil wells can be safely and readily achieved with the assistance of experienced oilfield engineers and contractors.

ACKNOWLEDGEMENTS The authors wish to acknowledge the many individuals comprising the project team for the City of Los Angeles, which was led by Baron Miya of the Bureau of Engineering. Thanks also to the team led

by Ted Budd of Kenny Construction who performed the real work of constructing the project, the largest Public Works project awarded by the City of Los Angeles to date. Special thanks to Chief Resident Engineer John Critchfield, who skillfully led the Construction Management Team. Resident Engineer Tom Saczynski, who provided tunneling expertise over both Unit 1 & 2. Jorg Holzhauser who provided valuable TBM performance expertise. Lastly, thanks to the Bureau of Contract Administration efforts to ensure a quality product, especially inspectors Mel Stanley, Jeff Kemper, Carlos Tirres, and Paul Hernandez.

REFERENCES Budd, T., and V. Goubanov. 2003. Case History – East Central Interceptor Sewer, Los Angeles, CA. Proceedings of the Rapid Engineering and Tunneling Conference. Crow. M., B. Miya and T. Budd. Construction of the east Central Interceptor Sewer in Los Angeles using EPBTBMs above the groundwater table. Proceedings of Underground Construction, London, 2003. Crow, M., and J. Holzhauser. 2003. Performance of four EPB-TBMs above and below the groundwater table on the ECIS Project, Los Angeles, CA, USA, 2003. Rapid Excavation and Tunneling Conference. pp. 905–926. Hanks, K., Fong, W. Edgerton, and B. Miya.1999. City of Los Angles Large Diameter Interceptor Sewer Tunnels. Proceedings of the Rapid Engineering and Tunneling Conference. Seeley, T. East Central Interceptor Sewer-EPB mining above and below the water table, 2004 North American Tunneling Conference. Yerkes, R.F., T.H. McCulloch, J.E. Schoelhamer, and J.G. Vedder. 1965. Geology of the Los Angeles Basin, California – An Introduction, U.S. Geological Survey Professional Paper 420-A.

448 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-55.qxd

2/18/04

2:14 PM

Page 449

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Rock tunneling at the Mill Creek project M. Schafer, B. Lukajic & R. Pintabona Montgomery Watson Harza, Cleveland, Ohio, USA

M. Kritzer, T. Shively & R. Switalski Northeast Ohio Regional Sewer District, Cleveland, Ohio, USA

ABSTRACT: The Mill Creek Phase 2 Tunnel (MCT-2) is the largest tunneling project undertaken by the Northeast Ohio Regional Sewer District (NEORSD). The 13,000-ft long tunnel was mined using a 24-ft diameter boring machine in shale ranging from 160 to 260 feet in depth. The tunnel will be utilized to convey and store combined storm and sanitary sewage collected from the member communities in the greater Cleveland area. The project was conceived as the backbone of an integrated solution to convey and store flows, relieving the existing undersized sewers. This paper will discuss the design of the tunnel and describe the construction progress to date. The discussion will include design of tunnel support and criteria for selection of tunnel boring machine (TBM).

1 OVERVIEW OF THE PROJECT The Mill Creek project is located in the Greater Cleveland area and serves 134,000 people in 11 communities. A three-phase tunneling construction approach encompasses fourteen (14) shafts and three (3) tunnels, totaling about 42,000 feet in tunnel length. The f irst phase, a 10-ft diameter conveyance tunnel was predominantly completed in 1999. The second phase project, consisting of a 24-ft excavated diameter tunnel and four large diameter shafts is nearing its completion, while the third phase, also consisting of a 24-ft excavated diameter tunnel is currently under construction with planned completion in 2006. The total contract cost for the three phase development will be about $150,000,000.

Shale bedrock units, as observed from other underground projects in the region, yield very little water. Gas, primarily methane, is commonly encountered in the Cleveland Shale and Chagrin Shale. 3 TUNNELING METHOD A full-face tunnel boring method was adopted to excavate the Mill Creek tunnel. For this purpose, the Contractor chose to deploy an open face (24-ft diameter), Robbins type machine, Figure 1.

2 GEOLOGICAL SETTING The tunnel horizon is situated within the Chagrin Shale rock formation. The Chagrin is approximately 500 feet thick in the Cleveland area and underlies the Cleveland Shale. No known major structural features are located in the project domain. The shale is classified as weak to strong rock and thin to massively bedded. The project area shale is also known to contain zones of thin bedding with siltstone, limestone and sandstone interbeds. In addition, two joint sets are present and strike approximately northeast and northwest. Joint spacing is irregular.

Figure 1. View of tunnel boring machine-cutting head in the tunnel.

449 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-55.qxd

2/18/04

2:14 PM

Page 450

A Geotechnical Baseline Report (GBR) prepared for the project provided the following guidelines for selection of the TBM: First, the TBM must be suitable to negotiate through thinly bedded, closely jointed rock. The GBR stated that overbreak and rock falls behind the tunnel face were potential problems during excavation. Additionally, the formation of wedges bounded by joints, bedding planes, and the tunnel perimeter would occur in the arch above springline. It is for this reason that the TBM was equipped with a protection shield between the cutterhead and the point of primary support installation. A total length of the shield was approximately 50 feet with the back portion consisting of a finger shield. This arrangement permitted the erection of the primary support system in a protected environment. Secondly, the TBM must be capable of excavating efficiently through rock formations of variable strength. The rock unconfined compressive strength ranged from 2,000 to 12,600 psi. Thirdly, the GBR also indicated that lenses of siltstone with unconfined compressive strengths in the range of 10,000 to 18,000 psi would be encountered. The presence of sandstone and limestone lenses of similar strength and thickness were also to be anticipated. Finally, the GBR recommended that the jack bearing pad surface areas be increased to minimize local overstressing of the tunnel walls. 4 TUNNEL SUPPORT DESIGN 4.1

Description of tunnel support

It was determined during the design stage that the preferred lining method would be a two-pass lining system. In this manner, a primary support (first pass) is installed concurrently with tunnel excavation. The final lining (second pass) is installed subsequent to the completion of excavation. This two pass lining system consists of a primary support system of steel ribs and timber lagging, and a final lining of cast-in-place reinforced concrete. Although this support alternative has a slightly higher unit cost than some of the other alternatives evaluated, it was determined that it has the greatest probability of successfully achieving desired performance requirements. Furthermore, this lining system takes advantage of existing experience held by the local labor force. Performance of steel-rib and lagging primary support systems in NEORSD’s previously constructed tunnels proved to be satisfactory. In addition to local contractors having extensive experience in installation of this type of support, the system proved to be very efficient in providing worker’s safety and enhancing tunnel stability.

4.2

The tunnel primary support system was specified to consist of Grade 50, W8 31 expanded circular steel ribs at 5-foot centers and 7-inch thick timber lagging spaced at a maximum of 24 inches along the tunnel perimeter. The specifications required steel ribs to be expanded to a point of intimate contact with the excavated rock surface. Anticipated rock behavior during mining operation was considered a primary criteria in designing the tunnel support. It was recognized that the shale rock units characteristically contain layers or partings of weaker material. It was assumed that these partings will sometimes break along these layers during excavation. The fissile nature of the shale accentuates this phenomenon. Particularly, in soft, thinly bedded shale, overbreak and rock falls behind the face were considered as potential problems during tunneling. At locations where weak partings or thinly bedded zones were anticipated in close proximity of the tunnel crown, support requirements were specified to increase above average requirements, meaning closer rib-lagging spacing. Based on the exploratory borehole data, local tunneling experience and empirical rock mass classification systems, the rock loads for primary rock support were determined to be 0.25B to 0.4B, where B is the tunnel excavated diameter (24 feet). In the areas of lesser rock quality, higher rock loads were recommended (0.5B to 0.7B). Minor modification of the above criteria were made during construction, which resulted in use of lighter ribs and lagging. A view of primary support in the tunnel is shown in Figures 2, 3 and 4. 4.3

Final lining design

Design of final lining for the Mill Creek tunnel was selected in accordance with the requirement for permanent tunnel support, ground water control and hydraulics. Three load cases were considered, with constant rock loads of 0.5B to 0.7B for each load case: 1. Loading imposed on the liner from internal pressure with no external hydrostatic pressures. 2. Internal hydrostatic pressure with external water pressure. 3. External water pressure with no internal pressure. If the internal hydrostatic pressure exceeds the ground water pressure, the liner must be able to carry the difference between these two pressures. Also, the liner carries the unbalanced internal pressure as a composite structure with the surrounding rock. The final lining consisted of cast-in-place reinforced concrete. Where the lining meets the shafts, it was prudent to allow for this section of the liner to be more reinforced, as the liner no longer forms a closed circle.

450 Copyright © 2004 Taylor & Francis Group plc, London, UK

Primary support design

09096-55.qxd

2/18/04

2:14 PM

Page 451

Figure 2. Tunnel configuration and support. Figure 4. A close up of primary support. 15,000 Total: 13,000 feet

Linear feet

10,000

5,000

0

Jan

Mar Feb

Figure 3. View of installed primary support.

5 TUNNEL CONSTRUCTION

Sept Aug

Nov Oct

Dec

tunnel, along individual bedding planes. This was evident in the areas where the tunnel intersected thin beds at an oblique angle or became essentially parallel to the low dip bedding formations. 5.2

Excavation

Excavation of the tunnel began in January 2001 and continued until December 2001. The tunnel was constructed using a full face Robbins boring machine. The Contractor chose to use a conveyer mucking system in conjunction with the TBM. Typically the excavation sequence consisted of tunnel boring (advancing in 4-ft. increments), rib-lagging installation and continuous mucking via conveyor system. Primary support was installed within the finger–shield, located immediately behind the primary TBM shield. Mining production rates ranged from 6 to 10 ft/hour. A graphical presentation of the excavation progress is shown in Figure 5. As per GBR, some overbreak occurred in the crown of the

Concrete lining

The final lining (second pass), consisting of cast in place reinforced concrete, was placed after completion of all tunnel excavation. The Contractor chose to use prefabricated steel forms, with concrete lining process consistently advancing at 96 feet per day. The specifications required that the forms remain in place until a strength of 1,000 psi. was achieved and not less than 8 hours of curing time had expired. In preparation for the lining operation, the Contractor submitted multiple concrete mix designs for review. Approval was granted on two mixes giving the Contractor an option to use during both cold and warm weather conditions. Regular tests were performed in the field to verify the mix met the specification requirements. A view of concrete placement set up is shown in Figures 6 and 7.

451 Copyright © 2004 Taylor & Francis Group plc, London, UK

July June

Figure 5. Excavation progress.

It was decided that the liner should be more heavily reinforced for a distance equal to the diameter of the shaft and over the extent of the access adit junctions as well. Figure 2 shows tunnel-lining design.

5.1

May April

09096-55.qxd

2/18/04

2:14 PM

Page 452

grout the tunnel was a total of 33 days, which is about 60 percent of the original forecast. Approximately 1,135.5 cubic yards of grout was required between the reinforced concrete lining and the rock surface. 5.4

Contract arrangement

A traditional approach of design, bid and construct was followed for this project. During the design stage of the project, a partnering workshop was arranged to include both the Owner (NEORSD) and Consultant (MWH) staff. The main objective of this process was to improve the efficiency of the project team and ensure proper communications at all levels of the project. Figure 6. View of placed liner.

6 CONCLUSIONS The project, once fully in operation, will provide relief to an overstressed combined sewer system. Developing a project of this magnitude required extensive effort in planning by both the Owner (NEORSD) and the Consultant (MWH), as well as an acceptance and support from the Ohio Environmental Protection Agency and the surrounding communities. In terms of both the cost and schedule, it has been demonstrated that a traditional competitive bidding process was a right choice for this project. ACKNOWLEDGEMENTS

Figure 7. Concrete forms.

5.3

Backfill grouting

The contractor’s construction schedule called for approximately 53 days to complete contact grouting. The actual construction time required to drill and

The authors would like to thank Northeast Ohio Regional Sewer District, Charles Vasulka, Director of Engineering for his review of the paper. Special thanks go to Carol Chavis for managing the paper design and production.

452 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-56.qxd

2/18/04

8:01 PM

Page 453

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Construction of the Dougherty Valley Tunnel, San Ramon, California, USA G.S. Nagle URS Corporation, Oakland, California

H. Thom Central Contra Costa Sanitary District, Contra Costa County, California

ABSTRACT: The Dougherty Valley Tunnel and Trunk Sewer project has been constructed by the Central Contra Costa Sanitary District (CCCSD) to provide wastewater services for the new Dougherty Valley residential development, consisting of approximately 10,000 new homes in the southern part of Contra Costa County, California. The tunnel and trunk sewer system conveys wastewater by gravity flow from the new residential development to a pumping station in San Ramon, California. The tunnel was constructed through a northwesttrending ridge between the Dougherty Valley and the San Ramon Valley from a portal site on the west side of Dougherty Road to a shaft located on the east side of Alcosta Boulevard. Cover above the tunnel ranges from about 20 feet at the East Portal and West Shaft to a maximum of about 320 feet near the mid-point of the alignment. The tunnel is 10 feet in diameter and about 4,640 feet long and was constructed through weak sedimentary rock (sandstone, siltstone and claystone). The tunneling was performed with a Technicore Earth Pressure Balance Tunneling Machine (TBM). Steel ribs and timber lagging were installed for initial support. The hole through occurred on April 23, 2002. The tunnel was classified as gassy by Cal OSHA. Three pipes were installed in the tunnel including a 60-inch inside diameter (ID) reinforced concrete sewer pipe (carrier pipe), and two 12-inch ID ductile iron bypass pipes. This paper discusses construction of the Dougherty Valley Tunnel and Trunk Sewer project.

1 INTRODUCTION 1.1

Background

In June of 1997, CCCSD annexed the eastern half of the Dougherty Valley into their service area. The Dougherty Valley is located in southern Contra Costa County, California (see Figure 1). The development includes more than 10,000 new homes. The wastewater facilities to service the new homes in support of the development include the pipeline, which was originally planned as a pumped system up and over the ridge between Alcosta Boulevard and Dougherty Road but was eventually built as a tunnel through the ridge. The developer, Windemere Company paid for the entire cost of the project. The 4,640-foot long tunnel collects wastewater at the southern end of the Dougherty Valley development, conveying the wastewater through the Dougherty hills to the west side of the ridgeline. The tunnel then transitions to a 24-inch trunk sewer along residential streets for 3,300 feet to the existing San Ramon pump station. From the pumping station, the wastewater is

pumped north through parallel 2-mile long 24-inch diameter force mains and then flows by gravity to the treatment plant 20 miles away. Expansion of the pumping station from 4.8 mgd peak wet weather capacity to 16 mgd, and the construction of the parallel force mains were implemented under different contracts. The trunk sewer and tunnel were constructed under one contract. Construction of the tunnel project was delayed for more than 2 years because of litigation, including litigation between the District and the City on the issuance of an encroachment permit. The court ruled in favor of the District and the City appealed. The litigation was ultimately resolved by a negotiated settlement agreement, which included a severe time constraint for construction. The City ultimately issued the encroachment permit in June of 2001; however, major public relations problems existed in the residential area throughout the duration of construction. 1.2

In 1995, CCCSD performed preliminary evaluations related to the proposed pipeline and determined that

453 Copyright © 2004 Taylor & Francis Group plc, London, UK

Preliminary design studies

09096-56.qxd

2/18/04

8:01 PM

Page 454

Figure 1. Dougherty Valley, Contra Costa County, California.

as an alternative to a pumped system, it would be feasible to construct the pipeline in a tunnel and allow gravity flow of the wastewater. The tunnel option was determined to be the favorable option because of operations and maintenance concerns and costs related to the pumped system. Based on the preliminary evaluations, CCCSD elected to design and construct the pipeline in a tunnel. URS Corporation (formerly Woodward-Clyde Consultants) prepared a Preliminary Design Report for the Dougherty Valley Tunnel (Woodward-Clyde consultants, 1996). Roadheader, TBM and drill-and-blast construction methods were all considered feasible construction methods at the time of preliminary design. 1.3

Final design

Final design for the Dougherty Valley Tunnel occurred between 1998 and 2001. The final design team for the tunnel included URS Corporation (prime design consultant) and Montgomery Watson (subconsultant for wastewater and structures). The tunnel and trunk sewer were bid as one construction contract. Final design for the trunk sewer occurred concurrently. The final design team for the trunk sewer included Brown and Caldwell (prime design consultant) and URS (subconsultant for geotechnical and microtunneling). Final design for the tunnel was developed based on using a

Figure 2. Typical tunnel sections.

TBM, digger shield, or roadheader for construction of the tunnel. Drill-and-blast construction was not allowed due to noise constraints. The contractor elected to construct the tunnel with a tunnel boring machine. The design was based on using steel ribs and lagging or precast concrete segments for ground support. Figure 2 shows typical tunnel sections as designed. The contractor elected to construct the circular tunnel section with steel ribs and timber lagging for support. However, the contractor moved the two 12-inch ID ductile iron bypass to the invert where they were cast in a concrete slab.

454 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-56.qxd

2/18/04

8:01 PM

Page 455

2 GEOLOGY OF THE PROJECT SITE 2.1

Geotechnical investigations

Geotechnical investigations for the tunnel consisted of geologic mapping, and drilling and logging 13 exploratory soil and rock core borings. Several of these borings were completed for other tunnel alignments that were considered during preliminary studies for the project and were not on the constructed tunnel alignment. Four of these borings were completed in November and December of 1996, and the remaining borings were completed between September and November 1998. The field exploration work also included borehole permeability tests (packer tests) and installation of open standpipe piezometers. Groundwater levels, as well as gas and oxygen concentrations, were measured in the piezometers following installation. A laboratory testing program was conducted on soil and rock samples recovered from the borings. Tests performed on soil samples included grain size analyses, density evaluations, Atterberg Limit determinations, and strength tests. Tests performed on rock core samples included the same tests that were performed on soil samples, and in addition, free swell tests, slake durability tests, petrographic analyses, and X-ray diffraction evaluations. The primary geologic units that were encountered during construction are described below. An interpretive geologic profile showing the subsurface conditions along the tunnel is presented on Figure 3. 2.2

Topsoil/colluvium

The topsoil/colluvium present at the project site included soft to hard silty and sandy clay. The strength of the clay deposits varied depending on the moisture content and was typically soft when saturated during and after the winter rains. The thickness of the deposits ranged from about one to seven feet. Colluvial soils consisting of topsoil and highly weathered bedrock fragments were present in a thick veneer on the lower hill slopes along Alcosta Boulevard at the West Shaft site and on the hillsides at the East Portal. 2.3

Alluvial deposits

Recent alluvium consisting of variable mixtures of sand, silt, clay and gravel were present along the Alamo Creek channel and its large tributaries, and in the San Ramon Valley to the west of the West Shaft. Flat lying alluvial terrace deposits and gently sloping alluvial fan deposits consisting of medium stiff to stiff silt and clay with sand and occasional gravel were mapped along the margins of Alamo Creek and where tributaries intersected the main creek channel.

Figure 3. Subsurface conditions along the tunnel.

2.4

The Orinda Formation is an extremely weak to very weak sedimentary rock mass that generally consists of interbedded sandstone, siltstone, and claystone, with occasional layers of conglomerate. The sandstone, siltstone and claystone were typically thickly bedded to massive with beds ranging from one to greater than ten feet. The bedding orientation had a strike of about N30°W to N60°W and dip from 30°NE to 70°NE. The flatter dip angle was observed at the West Shaft area and the steeper dip angle was observed in the East Portal area. The four primary rock types that were encountered during geologic exploration and during tunneling are described below: Sandstone – The sandstone was generally finegrained, and also included both silty and clayey sandstone. The rock was highly to moderately weathered to a depth of about 60 feet and slightly weathered to fresh below that depth. Sandstone was encountered in the borings along the alignment in various length intervals ranging up to approximately 60 feet. The maximum true bed thickness was estimated to be about 42 feet. The unconfined compressive strength of the sandstone based on laboratory testing ranged from 7 to 275 psi with an average value of 83 psi indicating that the sandstone was an extremely weak to very

455 Copyright © 2004 Taylor & Francis Group plc, London, UK

Orinda Formation

09096-56.qxd

2/18/04

8:01 PM

Page 456

weak rock. Portions of the sandstone were only very weakly cemented and described as fraible. Siltstone – The siltstone included both sandy and clayey siltstone and was highly weathered to a depth of approximately 35 feet and moderately weathered to fresh below that depth. Siltstone was encountered in the borings in intervals ranging up to approximately 61 feet. The maximum true bed thickness was estimated to be about 43 feet. The unconfined compressive strength of the siltstone ranged from 31 to 523 psi with an average value of 186 psi indicating that the siltstone was an extremely weak to very weak rock. Claystone – The claystone rock type also included silty and sandy claystone. The claystone was highly weathered to a depth of at least 50 feet and moderately weathered to fresh below that depth. Claystone was encountered in borings along the alignment over intervals up to approximately 42 feet. The maximum true bed thickness was estimated to be about 30 feet. The unconfined compressive strength of the claystone ranged from 7 to 502 psi and averaged 114 psi indicating that the claystone was extremely weak to very weak. Conglomerate – With the exception of relatively thin (up to 6 inches thick) gravelly interbeds and occasional pebbles found within the sandstone, siltstone and claystone of the Orinda Formation, conglomerate was not encountered in the exploratory borings completed for this project. However, an approximately 10-foot thick conglomerate bed was observed in the field during geologic mapping along the bank of Alamo Creek about 900 feet north of the East Portal. This rock mass consists of a massive, moderately weathered conglomerate with rounded gravels and cobbles up to 12 inches in maximum dimension. The gravels and cobbles are composed of sandstone, quartzite, chert, and andesite that are embedded in a weakly cemented sandy matrix. Significant amounts of this rock type were not encountered during the tunneling. 2.5

Groundwater

Following drilling, seven of the borings were completed as open standpipe piezometers. Groundwater level measurements in the piezometers located in boreholes along the tunnel alignment indicated groundwater levels that fluctuated seasonally. Groundwater levels increased up to about 210 feet above the tunnel as it passed beneath the ridge (Figure 3). The permeability in the Orinda Formation encountered in the tunnel was relatively low, considering the results of hydraulic conductivity measurements in the boreholes (i.e., packer tests) which indicated values ranging from 2 106 to less than 1 107 cm/sec. Conglomerate beds and occasional shear zones were expected to exhibit higher rock mass permeability values. There was no significant groundwater inflow into the tunnel during construction except for

periodic minor inflows and in one location where there was some water that came in to the tunnel through a discontinuity. All groundwater collected during construction was pumped to a storage pond equipped with an overflow and oil skimmer. The clean overflow was discharged directly into the adjacent creek. 2.6

3 PORTAL DEVELOPMENT/SHAFT CONSTRUCTION Surface excavations were required for site development at the East Portal and West Shaft sites. The East Portal excavation required the construction of temporary cut slopes in both soil and rock. The West Shaft excavation required the construction of three soil nail retaining walls in a terraced configuration. In addition, a vertical shaft excavation was constructed in the level area between the base of the lowest soil nail wall and Alcosta Boulevard. East Portal development and West Shaft construction occurred in October and November, 2001.

456 Copyright © 2004 Taylor & Francis Group plc, London, UK

Ground characterization

The tunnel was divided into three sections for the purpose of summarizing the anticipated ground conditions (see Figure 3). Sections I and III were at the east and west ends of the tunnel where the cover above the tunnel ranges from about 20 to 135 feet and about 20 to 150 feet respectively. Section II includes the central portion of the tunnel and has a cover ranging from about 135 to 320 feet above the tunnel. Ground loads to be used for the design of the tunnel initial support system were developed for each section of the tunnel. Section I was anticipated to include 14 percent sandstone, 35 percent siltstone, 50 percent claystone and 1 percent conglomerate. Section II was anticipated to include 44 percent sandstone, 35 percent siltstone, 20 percent claystone and 1 percent conglomerate. Section III was anticipated to include 35 percent sandstone, 10 percent siltstone, 54 percent claystone and 1 percent conglomerate. The Geotechnical Report for the project described the anticipated ground behavior in detail for each rock type within Sections I, II and III (URS Greiner Woodward Clyde, 2001). In general, adverse ground conditions consisting of moderately to highly squeezing ground, swelling ground, and fast raveling ground were anticipated. Squeezing conditions were expected in the claystone and siltstone encountered in all three sections of the tunnel. Swelling ground behavior was also anticipated in the claystone and siltstone. Fast raveling conditions were anticipated primarily in the sandstone and conglomerate.

09096-56.qxd

3.1

2/18/04

8:01 PM

Page 457

East Portal

Tunnel construction was staged from the East Portal and the TBM was launched from this site (see Figure 4). The subsurface conditions at the East Portal consisted of topsoil/colluvium underlain by alluvial deposits and Orinda Formation claystone and siltstone. The topsoil/colluvium thickness in the vicinity of the portal excavation ranged from about 2 to 6 feet and generally consisted of medium stiff to hard silty clay. The alluvial deposits beneath the topsoil/ colluvium consisted of stiff clay and medium dense sand that ranged in thickness up to about 25 feet. Groundwater was generally encountered in the alluvial deposits. Beneath the alluvial deposits, extremely weak claystone and siltstone (with a compressive strength of less than 50 psi) was encountered. The rock was completely weathered in the upper 5 feet below the contact with alluvial deposits and highly weathered below that depth. The Contract Documents allowed the East Portal excavation to be constructed using temporary construction slopes which could be unsupported if they were sufficiently flat and would require excavation support if they were steeper than about 2:1 in soil and 1:1 in rock. The Contractor elected to construct the East Portal using steeper temporary cutslopes supported with shotcrete and rock bolts. The portal headwall was supported with shotcrete and 30-foot long soil nails. Groundwater was sumped and pumped during construction. Grading was done to create access for the muck trains to enter and exit the tunnel and for construction staging. 3.2

West Shaft

The West Shaft excavation was used to retrieve the TBM and to construct the West Shaft Transition Structure where the tunnel transitioned into the trunk sewer to the pumping station. The subsurface conditions at the West Shaft site included a 1- to 2-foot thick layer of medium stiff clay over Orinda Formation claystone. The claystone was highly weathered and extremely weak with an unconfined compressive strength up to about 30 psi. Groundwater was present about 20 feet below the ground surface. The excavation for the West Shaft extended approximately 25 feet below the existing grade and roadway surface. Above the shaft there was a terraced excavation, which met the existing cut slope adjacent to Alcosta Boulevard. The West Shaft was required to be a vertical shaft excavation in order to minimize disturbance to Alcosta Boulevard and the adjacent hillside. Excavation of the shaft was preceded by the construction of the three soil nail walls, which were terraced into the existing slope. The top soil nail wall was constructed first, followed by the middle wall and

Figure 4. Tunnel boring machine arriving at site.

Figure 5. West shaft constructed at base of soil nail wall.

then the bottom wall. Each wall was required to be excavated in two stages, first making a 4-foot maximum vertical excavation, placing 2 inches of shotcrete, and installing one row of soil nails. The sequence was repeated for the next stage of excavation. After the second row of soil nails was installed a second 2-inch thick layer of shotcrete was required to be applied prior to beginning excavation for the next wall. Composite geotextile drainage panels were installed on the cut face prior to applying the first layer of shotcrete to provide drainage behind the walls (see Figure 5). After the West Shaft was completed and the structure was constructed, a reinforced concrete wall was constructed in front of each soil nail wall and a stone facing was placed on the face of the concrete walls as architectural finish. 4 TUNNEL CONSTRUCTION 4.1

457 Copyright © 2004 Taylor & Francis Group plc, London, UK

Bid and award

The project was bid in June of 2001. Bids ranged from $12.3 million to $14.1 million with an Engineer’s estimate of $15 million. Mountain Cascade (MCI) was

09096-56.qxd

2/18/04

8:01 PM

Page 458

the low bidder with Walter C. Smith (WCS) as the tunneling subcontractor. The second lowest bidder protested claiming that MCI had a non-responsive bid. The Specifications required the tunneling subcontractor to have completed one or more projects involving 2,000 feet or more of tunneling a 6-foot diameter or larger tunnel in similar ground conditions. WCS had completed 6-foot or larger diameter tunnels in similar ground conditions aggregating a distance of well over 10,000 feet but none of the projects had a distance of 2,000 feet by itself. In July of 2001 the CCCSD Board held a hearing and determined that the bid was responsive and awarded the contract to MCI. On July 12, 2001 Notice of Award was issued to MCI at the bid price of $12,274,620. The total budget for the project including design, construction management, consultants, and construction was $19,239,000. Project and construction management was provided by CCCSD staff and a field office was established in San Ramon. 4.2

Schedule constraints

Due to the sensitivity of the project and problems with several very vocal residents, CCCSD wanted to complete the work along the residential streets as soon as possible and preferably before Thanksgiving. Therefore, an early completion incentive bonus of $150,000 was offered for the trunk sewer portion of the project. The work needed to be completed in the streets prior to Thanksgiving 2001 to receive the early completion bonus. On November 21, 2001, the day before Thanksgiving, MCI completed the trunk sewer portion of the project and earned the early completion bonus. 4.3

Tunneling milestones and challenges

In July, 2001, baseline vibration, noise and odor monitoring were conducted and a preconstruction damage assessment was performed. In August, 2001, the traffic control plan was approved by the City with numerous comments. On August 28, 2001, street excavation was started for trunk sewer construction. On September 11, 2001, the same day as the terrorist attacks in New York, the Contractor started excavation for a shaft to construct a 30-inch diameter microtunnel crossing in front of the pump station. Problems started as soon as the excavator bucket touched the ground at 8:30 am. Due possibly to the combination of the terrorist attack and her objection to the project, a nearby resident became hysterical and personally stopped the work. She then called the mayor, police department, public works department, fire department and City manager. This same resident complained almost daily throughout construction. She complained of roof damage prior to the contractor beginning excavation. She complained about the noise

level of 64 to 65 dB in front of her house. Noise levels were 43 to 46 dB inside of her house, well within normal levels. The Contractor and CCCSD offered her a hotel suite but she declined and continued with complaints and threats of a lawsuit throughout the project. On November 14, 2001, a tunnel preconstruction meeting was held at the site with CCCSD, construction management staff, Cal OSHA, the local fire department, unions, contractors, and designers. Portions of the tunneling machine started arriving that day (see Figure 4). The Contractor elected to use a refurbished Technicore VLC M 116 Earth Pressure Balance Tunneling Machine. The machine was 10 feet in outside diameter and 21 feet long. The machine was comprised of 3 sections including: a forward shell, stationary shell, and trailing shield. It had a weight of 66 tons with a 624 horsepower hydraulic transmission. The cutting head was bi-directional with carbide perimeter cutters, standard spade teeth, and carbide ripper teeth. It was powered by two 250 horsepower electric motors and driven by six hydraulic piston motors. The machine included ten 83-ton propulsion cylinders with a maximum thrust of 830 tons and a stroke of 66 inches. The cylinders were powered by a 75 horsepower electric motor. Twenty-inch diameter screw augers transferred muck to a 24-inch wide 100-foot long conveyor. The muck was then conveyed into muck cars that were hauled with a locomotive. The machine was equipped with methane gas detectors with auto shutoff. The machine had numerous mechanical and electrical problems throughout the early phase of the tunneling. On December 5, 2001, the TBM was launched and pushed in 12-feet. The following day, the machine had problems with leaking oil and grease lines and the machine was pulled out and disassembled for repair. On December 10, the TBM was reassembled (see Figure 6). On December 14, the Contractor tried to re-launch the machine but they had oil and grease problems again and a manufacturer representative from Technicore

Figure 6. Relaunching TBM.

458 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-56.qxd

2/18/04

8:01 PM

Page 459

was called to the site. Between December 20, 2001 and January 1, 2002, the Contractor redesigned and reinstalled the launching pad and trailing gear (over the Christmas and New Years holidays). On January 2, 2002, CCCSD held a special meeting with MCI and WCS regarding the schedule delays. On January 4, 2002, the Contractor completed laying the rail but one of the locomotives was inoperable. On January 5, 2003, the TBM had electrical problems and blew a fuse. Some rewiring was performed and on January 7 there was another electrical problem followed by a hydraulic pump malfunction on January 8. On January 10, the TBM was re-launched about 5 feet before having problems. On January 12, 36 feet of tunneling was completed. Steel ribs and timber lagging were installed full perimeter for initial support throughout the tunnel as shown in Figure 7. The production for the month of January totaled 514 feet in 15 work days or about 34 feet per day. In February, the production rate increased to about 49 feet per day and by the end of March, production was averaging about 68 feet per day. Once the initial start up problems were overcome the TBM performed effectively and the tunneling progressed steadily.

4.4

Pipe installation in tunnel

After completion of the tunnel the contractor demobilized the equipment at the East Portal site. The rail was then removed from the tunnel and two 12-inch bypass pipes were installed in the invert. An invert concrete slab encasing the two ductile iron pipes was poured between May 24 and June 26, 2002. The 60-inch reinforced concrete pipe was then installed in the tunnel using a rubber tire mounted pipe carrier as shown in Figure 8. The pipe carrier was fabricated by welding a steel beam to a forklift, with hydraulic cylinders to raise or lower the rubber tires for traveling in the tunnel. 4.5

Special requirements during construction

In accordance with Cal OSHA requirements and the requirements in the Contract Documents, a probe hole was advanced 20 feet ahead of the tunnel face to check for gas and to provide an indication of ground conditions ahead of the face. The probe hole was intended to identify limits of potentially difficult ground conditions, gas, or groundwater and to locate areas where tunneling could be more difficult. During construction, the Contractor requested that CCCSD eliminate the probe hole drilling requirement because it was difficult to perform the drilling and the schedule was very tight. CCCSD elected to maintain the requirement and the Contractor typically performed the probe hole drilling on the night shift. In order to control risk, the Contract Documents contained a Differing Site Conditions clause and provisions for a Disputes Review Board to resolve disputes. The DRB met monthly during construction of the tunnel and there were no differing site conditions or other claims.

5 CONCLUSIONS Figure 7. Tunnel was supported with ribs and lagging.

Figure 8. Pipe carrier.

The tunnel holed through on April 23, 2002, at 11:30 am as shown in Figure 9. The machine was no more than 6-inches off alignment, which was well within the allowable tolerance. The construction area at the East Portal has since been completely restored without any noticeable sign of the work. The area immediately above the West Shaft where the hill was cut for the soil nail retaining walls was also completely restored and landscaped (see Figure 10). After a somewhat difficult project start up due to litigation with the City and mechanical problems with the TBM, the tunnel was successfully constructed by MCI and WCS with no differing site conditions or claims and only minor change orders. The final constructed cost was $12.5 million. The project is considered a success by CCCSD, the Contractor, the

459 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-56.qxd

2/18/04

8:01 PM

Page 460

designers, and the developer. Homes are currently on sale in the Dougherty Valley at prices ranging from the high $400,000 for condominiums to well over a million for single-family homes. ACKNOWLEDGEMENTS

Figure 9. TBM hole through.

The authors would like to thank Central Contra Costa Sanitary District for permission to publish this paper. In addition to the contractor, there were several consultants working for CCCSD during construction of the project. Brown and Caldwell provided resident engineering services; EPC and Underground Construction Managers provided tunnel inspection; SOHA Engineers performed a preconstruction damage assessment; Wilson Ihrig and Associates performed vibration and noise monitoring; and URS Corporation performed engineering services during construction and assisted CCCSD with tunneling and geotechnical issues.

REFERENCES

Figure 10.

Restored and landscaped West Portal site.

Central Contra Costa Sanitary District Daily Inspection Reports, Project Number 5902. URS Greiner Woodward Clyde, 2001, “Geotechnical Interpretive Report,” Dougherty Valley Tunnel, San Ramon, California, May. Woodward-Clyde Consultants, 1996, “Preliminary Design Report,” Dougherty Valley Tunnel, San Ramon, California, February.

460 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-57.qxd

2/18/04

2:33 PM

Page 461

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

City of Los Angeles Northeast Interceptor Sewer Tunnel Z. Varley Parsons Brinckerhoff, Los Angeles, CA, USA

R. Patel City of Los Angeles, Bureau of Engineering, Los Angeles, CA, USA

J. McDonald Traylor-Shea-Frontier-Kenny, JV., Los Angeles, CA, USA

ABSTRACT: The City of Los Angeles, Bureau of Engineering Major Sewers Construction Group is managing the construction of the second of two large-diameter deep sewer tunnel projects as mandated by a stateissued Cease-and-Desist Order. The 8.5-kilometer Northeast Interceptor Sewer tunnel is being excavated with two soft-ground Earth Pressure Balance and one hard-rock Tunnel Boring Machines. The 2.4-meter insidediameter finished sewer will replace a length of the 80-year-old existing North Outfall Sewer, which will be rehabilitated in sections before ultimately becoming the backup service line.

mandated by the state-issued Cease-and-Desist Order (CDO), TSFK launched all three machines within a five-month period.

1 DESIGN 1.1

Project overview

City of Los Angeles Engineers, assisted by Jacobs Associates in the design phase, and by Parsons Brinckerhoff Construction Services as CM team members, are managing construction of the 8.5-kilometer Northeast Interceptor Sewer (NEIS) tunnel being built by the Traylor-Shea-Frontier-Kenny (TSFK), Joint Venture. With construction of the East-Central Interceptor Sewer (ECIS) already underway, the NEIS tunnel is a slightly smaller, slightly shorter section of a new sewer system replacing a major portion of the existing 88-kilometer North Outfall Sewer (NOS) originally built in the 1920s. Designed to tiein to the east terminus of the ECIS tunnel just east of the Los Angeles River, the NEIS tunnel runs roughly due north to a connection with several smaller sewers just south of the City of Glendale (see Fig. 1). The alignment includes three new mining shaft sites in addition to the existing tie-in shaft already built by ECIS, seven maintenance hole access shafts and a special access shaft for demolition of a conflicting water tunnel, the Narrows Gallery. Three Lovat-manufactured Tunnel Boring Machines (TBMs) will complete the mining, the first two being new Earth-Pressure-Balanced (EPB) machines, and the third a refurbished, partially-shielded hard-rock TBM. Due to the aggressive construction schedule

1.2

The NEIS tunnel is the second major section of sewer tunnel being constructed to replace the existing NOS, the 80-year-old backbone of the Los Angeles Sewer system. Initially constructed in the mid-1920s, the NOS has been in dire need of repair for several years due to deterioration of the mortar and clay tile tunnel lining, accumulation of debris in certain sections and in large part due to the increased effluent demand, which is well beyond the design capacity. In the early 1990s the Los Angeles Department of Public Works (LADPW) began design work on a replacement system for the NOS, leading to the construction of the North Outfall Replacement Sewer (NORS) tunnel, and preliminary design work for the ECIS and NEIS tunnels. In the late 1990s, the El Nino rainstorms caused widespread flooding as groundwater seeped into the existing NOS through cracks in the lining and deteriorated maintenance holes. The California Regional Water Quality Control Board issued a CDO prompting completion of design work for the ECIS and NEIS tunnels and forced the City to commit nearly $800 million toward a systemwide program of sewer construction and rehabilitation. (See Hanks et al., 1999 RETC)

461 Copyright © 2004 Taylor & Francis Group plc, London, UK

Background

09096-57.qxd

2/18/04

2:33 PM

Page 462

Figure 1. ECIS & NEIS alignment and NEIS geologic profile.

As various milestone dates for sewer operation were prescribed by the CDO, the City sought to augment its engineering staff with tunnel design and construction management consultants, and construction of the ECIS tunnel got under way in late 2000. As ground was being broken at several of the ECIS sites, design work for the NEIS tunnel was completed, and the construction contract was put out to bid in December of 2001. 1.3

Design requirements

The NEIS tunnel was designed as a 3.9-meter excavated diameter, pre-cast concrete segment supported, 8.5-kilometer long tunnel for the installation of an 2.4-meter ID PVC lined carrier pipe. With a constant 0.2% slope for the entire length of the sewer, and upstream to downstream flow ratios varying from 0.43 to 0.54 d/D, the expected design life is intended to be roughly 100 years. The pre-cast concrete bolted and gasketed doubly tapered segments were designed for assembly in six sections using two trapezoidal keys and with two separate ring lengths to accommodate 152-meter radius horizontal curves. Backfill grout injected through the segments will fill the annular overcut and small voids to prevent excess surface settlement during mining, and will transfer the ground surcharge loading uniformly

over the segment rings for permanent structural support. Lightweight cellular backfill grout will be placed in the annulus between the concrete segment liner and the carrier pipe. In addition fiberoptics conduits will run between maintenance holes outside the carrier pipe for future use, and will be embedded in the backfill grout between the segments and the carrier pipe. Maintenance holes were spaced roughly half a mile apart allow for utility service and sewer tunnel access. Three separate cast-in-place concrete drop structures will be constructed to drop surface sewer lines into the interceptor, and eventually feed into the ECIS sewer line. The tunnel alignment virtually parallels the existing NOS alignment from just east of Downtown at the Los Angeles River at 7th St., northward along Mission and San Fernando Roads to Eagle Rock Blvd., just south of the City of Glendale. Completion of the NEIS tunnel will allow for rehabilitation of the existing NOS, which will become a backup sewer line, and allow for the tie-in of sewage flows from the newly microtunneled Eagle Rock Interceptor Sewer (ERIS), as well as from the NEIS 2 project that will extend northwest of the Division shaft into NEIS (see Fig. 1). Approximately 60% of the tunnel mining is expected to be in “soft ground” below the groundwater table in soils varying from alluvial sands to lightly cemented

462 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-57.qxd

2/18/04

2:33 PM

Page 463

siltstones, with naturally-occurring hydrocarbons in some areas. The remaining mining is expected to be through silt- and claystone rock (see Fig. 1). Due to the likelihood of mixed-face conditions at the various soil interfaces and general apprehension due to ground settlement issues encountered during the tunneling of the MTA Redline Subway in Hollywood, two new EPB or SPB TBMs were specified for the southern two tunnel reaches, with a refurbished openface rock machine being allowed for the northernmost reach which is entirely in rock. Additionally, geotechnical instrumentation along the tunnel alignment was specified in order to monitor settlement both above the tunnel crown and at the surface. Permeation grouting was specified beneath Critical Structure foundations and areas of “LowRock Cover”. Concrete slurry walls were specified at the mining access shafts for temporary structural support during mining and drop structure construction, and to prevent the migration of potentially contaminated groundwater into the shafts.

Bid document preparation

Upon completion of the Final NEIS Design by the City’s Bureau of Engineering (BOE) and Jacobs Associates, NEIS Contract Bid documents were prepared with additional assistance from Parsons Brinckerhoff for technical accuracy and constructability. Issued on Dec. 19, 2001, the documents essentially mimicked the ECIS contract documents, with several improvements from the nearly two-year hindsight experiences from ECIS taken into account. Significant differences included retention of the responsibility for real estate Right-Of-Way acquisition for mining shaft sites and underground easements beneath private properties by the City, and revised Bid Item language for the definition and payment of permeation grout beneath critical structures. The NEIS documents also specified that the EPB TBM’s screw conveyor be installed in the lower quarter of the cutterhead chamber and without a muck ring, both improvements believed to enhance TBM performance. A final change was the potential to use a refurbished partially shielded rock TBM for the Upper Reach of NEIS, instead of a new EPB machine as was required for the Lower and Middle Reaches on NEIS and on all of the ECIS drives. 2.2

Additionally, strict adherence to the labor compliance requirements in regard to local hiring, D/W/MBE subcontractor participation and the Project Labor Agreement (PLA) similar to the requirements of ECIS was emphasized as a critical issue as far as the City’s Board of Public Works was concerned. 2.3

2 BID PURSUIT AND AWARD 2.1

1. EPB mining of “Gassy Tunnels” below the Groundwater Table (GWT), with as much as 30 meters of hydrostatic head (ECIS alignment was above GWT for all but the first 100 meters). 2. Permeation grouting beneath Critical Bridge Structures and Low Rock Cover Areas. 3. Excess water inflows in the Upper Reach rock, tunneling with an “open face” (not EPB) machine, specifically in regards to the Narrows Gallery whose exact location was not known at the time of bidding. 4. Naturally-occurring hydrocarbons, oil deposits, hydrogen sulfide and other heavy-metal contaminants in the soil and potential contaminated groundwater migration.

Bidding risks

From a bidding and constructability standpoint the following items were considered to be the major potential risks for completion of the project:

Three signed and sealed bids were opened on Feb. 19th, 2002 with submissions from the TSFK JV, Affholder/Ellmore, JV. and Obayashi/Shimmick, JV. The TSFK JV, essentially the same joint venture already constructing the ECIS project, with Traylor Brorthers Inc., as the lead instead of Kenny Construction, submitted the low bid of just over $162 million, about $2 million less than Affholder/Ellmore. Over the 90 days following the bid openings, the City reviewed the bids and investigated the legitimacy of a potential protest from Affholder/Ellmore, causing the bid bonds to be extended an additional 30 days. The protest was disallowed and on June 5th, 2002 the City issued Notice of Award. Notice to Proceed (NTP) was issued on July 23rd, 2002 and construction work on site got quickly underway. 3 CONSTRUCTION PREPARATION 3.1

Master Schedule and Critical Path

Due to the extremely tight schedule for construction mandated by the CDO, TSFK was tasked with submitting a Critical-Path-Method (CPM) Baseline Schedule with all three reaches mining simultaneously with concurrent pipe-laying and drop structure construction succeeding the mining. The closely concurrent work meant that the Critical Path could potentially move from one tunnel reach to another with any given delay, proving to be one of the very early schedule analysis challenges. Additionally, the Baseline Schedule had to be approved through the submittal process

463 Copyright © 2004 Taylor & Francis Group plc, London, UK

Contract award

09096-57.qxd

2/18/04

2:33 PM

Page 464

prior to mobilization payment, forcing rapid development of a detailed CPM schedule immediately after NTP. Several of the major “long lead-time” items and activities that drove the schedule are listed below. 3.2

Major items

3.2.1 TBM manufacturer Lovat Inc. of Toronto, Canada was contracted to deliver and launch two new EPB and one refurbished partially shielded rock TBM (see Fig. 2) within a 2month window according to the Baseline Schedule. The EPB machines required the installation of a compressed air bulkhead for the potential of up to 3-bars pressure for cutterhead access in saturated soils well below the groundwater table. Located just behind the tail shield and in front of the screw conveyor discharge, the bulkhead was designed to minimize the amount of compressed air required for maintenance and emergency operations in a sealed and pressurized tunnel environment. A combination of disc cutters and carbide-tipped picks were required for mixed face conditions and a screw conveyor with an intake in the bottom quarter of the cutterhead was specified for the EPBMs. The EPBMs also had all electric Variable Frequency Drive (VFD) motors, and probehole drill ports were installed for access through the stationary shield as required for “Gassy Tunnels”. The partially-shielded rock machine had electronically driven hydraulic drive motors and a belt conveyor with a guillotine gate. 3.2.2 Pre-cast Concrete Segment Plant The Traylor/Shea/Ghazi Pre-cast Plant in Palmdale, CA, which was initially setup for ECIS, had to procure smaller segment forms for the casting beds.

Figure 2. Lovat EPB and rock TBMs for The NEIS Lower & Upper Reaches.

3.2.3 Pre-cast Concrete Cylinder Pipe Plant The Ameron International Concrete and Steel Pipe Plant in Fontana, CA, which was also setup for ECIS, had to resize casting equipment for the smaller diameter carrier pipe sections. 3.3 3.3.1

Mining shafts

3.3.1.1 Slurry walls Soletanche Inc. was selected as the subcontractor and used a hydrofraise to accelerate construction of the 50-meter deep shaft walls through mixed ground soil conditions including alluvial sands, silts, clays and some rock (see Fig. 3). 3.3.1.2 Shaft excavation TSFK staged conventional backhole and scale pan shaft excavation to immediately follow slurry wall construction, and required completion of excavation prior to the delivery and launching of the TBMs. The excavation proceeded by alternating between north and south cells at the Richmond and Humboldt DualCelled Shafts (see Fig. 4). 3.3.2

Surface work

3.3.2.1 Utility work Doty Brothers Construction was selected to pothole and relocate utilities prior to maintenance hole and permeation grout drilling. 3.3.2.2 Maintenance holes Malcolm Drilling was scheduled to drill and case maintenance hole shafts prior to tunneling, with the larger diameter shafts in soft soil on the Lower and Middle Reaches being completed before the smaller diameter shafts in weak rock on the Upper Reach. Malcolm used a crane-mounted Kelly-bar auger and advanced welded oversized steel casings sealed with

Figure 3. Hydrofraise excavation at Richmond shaft.

464 Copyright © 2004 Taylor & Francis Group plc, London, UK

Major activities

09096-57.qxd

2/18/04

2:33 PM

Page 465

a cement/bentonite plug down to the tunnel springline in soft ground and just above the tunnel crown in rock (see Fig. 5). 3.3.2.3 Geotechnical instrumentation Group Delta w/C&C Drilling were selected to install multi-point borehole extensometers (MPBXs), piezometers and inclinometers adjacent to the mining shafts

and along the tunnel alignment in advance of the mining drives as was done on ECIS. 3.3.2.4 Permeation grouting Hayward Baker Inc. (HBI) was selected to complete both Chemical and Microfine-Cement permeation grouting in soils on the Lower Reach and in rock on the Upper Reach respectively. Drilling and grouting work was scheduled concurrently on the Lower and Upper Reaches in advance of tunneling, and included the Narrows Gallery, Rt-5/110, Rt. 101, Whittier Blvd overpasses and Low Rock Cover area on San Fernando Rd (see Fig. 6). 3.3.3

Figure 4. Richmond shaft excavation.

Mining

3.3.3.1 Soft ground The longer Lower Reach, with a more consistent soil profile advancing southward from the Richmond Shaft was scheduled for the first TBM launch and took advantage of access to both cells of the Richmond Shaft for equipment assembly. The shorter Middle Reach, with a mixed soil interface and curve much closer to the launch shaft was scheduled for the third TBM delivery. Both drives start out in a weak clay-like siltstone before transitioning into a more sandy material (see NEIS GBR & Fig. 1) within a zone of naturally-occurring hydrocarbons and hydrogen sulfide source rock. The drives were scheduled for initial two-shift production mining of 3-meters per day during startup and improving to 16-meters per day during full scale mining according to the original Baseline Schedule. 3.3.3.2 Hard ground Upper Reach mining through the Puente Unit 2 rock was also scheduled for 3-meters per day startup mining and 16-meters per day full production mining. The hard rock TBM was the second machine delivered and launched for the longest of the three reaches.

Figure 5. Maintenance hole #3 construction.

Figure 6. Permeation grouting underneath Rt. 101 freeway.

465 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-57.qxd

2/18/04

2:33 PM

Page 466

3.3.3.3 Ventilation and access Ventilation for all three reaches was designed for topside access shaft fans for the entire length of each reach, instead of using SMH risers with permanent traffic detours (as was the case on ECIS). Temporary inflatable emergency refuge chambers will be used to satisfy the CalOSHA requirement of emergency stations or access points every 1500 meters. 3.3.4 Carrier pipe laying This work was scheduled in the reverse direction from mining with ventilation and pipe section delivery from the mining access shafts and riser tie-in T-pipe sections at the maintenance holes. 3.3.5 Drop structures and maintenance hole risers Cast-in-place concrete drop structures and pre-cast concrete maintenance hole risers will be constructed after carrier pipe installation.

4 DESIGN CHANGES AND OBSTACLES 4.1

Slurry walls and hydrofraise

Figure 7. Lower Reach TBM exit seal.

Soletanche’s use of the hydrofraise led to several improvements on the initial design. 1. The shorter “fraise” panels allowed for two smaller diameter cells at the Richmond Shaft reducing the shaft excavation footprint (saving precious topside real estate) and reducing the quantity of soil excavated by roughly 35%. (See McKenna et al., 2003 RETC) 2. The reconfigured Humboldt Shaft eliminated horizontal cross-bracing for the second drop structure adit tunnel shaft and reduced the main access shaft sequential excavation duration by replacing shotcrete and lagging with rock bolt tie-backs. (See McKenna et al., 2004 NAT) 3. The machine cut through weak rock faster and without chisel eliminating “hard-digging” with clamshell and potential impact vibrations and extended the walls ten meters below the working shaft bottom elevation eliminating an optional 5-meter grout curtain for groundwater cutoff (see Fig. 7). 4. The environmentally cleaner operation with a desanding plant was an improvement over a wet clamshell spoils pile. 4.2

Mining drives and exit seal

The Richmond Shaft reconfiguration led to the launch of two EPBMs from one shaft, consolidating machine shop, backfill grout and foam generation plants, and changed the mining direction of the Middle and Upper Reaches from downhill to uphill. The uphill mining has the advantage of encountering

some of the “mining risks” sooner (i.e. Narrows Gallery, mixed soil face conditions and hydrocarbons) rather than later, when full production mining would be able to make up for early delays. (Every tunnel operation has a learning curve!) The Richmond Shaft was also constructed with cast-in-place concrete “square-off ” and thrust block walls, with a steel ring beam and rubber gasket exit seal mounted to the wall prior to mining (see Fig. 6). The square-off and thrust walls were formed perpendicular to the tunnel alignment and were used to launch the TBMs by pushing off on a flat surface as opposed to curved slurry wall panel. The exit seal allowed the TBMs to start mining through the concrete wall in EPB mode with a cutterhead full of foam, water and concrete debris, eliminating the need for jetgrouting of the soil just outside of the slurry walls. 4.3

The Contract documents added a $1 million contingency bid item almost “by accident” as design neared completion for the location and demolition of a 100year old abandoned water supply tunnel believed to conflict with NEIS tunnel alignment. According to long lost historical survey data, the Narrows Gallery tunnel would cross the NEIS Upper Reach tunnel just south of the Arroyo Seco creek at a depth of invert just within the new NEIS tunnel crown. The Contractor-proposed method for locating the Narrows Gallery included grouting in a grid pattern in the “best known” location after exploratory “hunt and

466 Copyright © 2004 Taylor & Francis Group plc, London, UK

Narrows Gallery

09096-57.qxd

2/18/04

2:33 PM

Page 467

peck” borings were drilled. Low-mobility and microfine-cement grout were used to seal out the potentially contaminated trapped water in the pipe. Probe drilling from the TBM face would also be used to confirm the tunnel location and water infiltration rate prior to mining. As of yet, there has been no complete confirmation as to exact location or condition (including possible cable reinforcing) of the Narrows Gallery, however investigative borings are being drilled at this time. TSFK has elected to not sink a shaft to locate the tunnel until the location is more accurately identified and determined to be too large an obstacle for demolition with the rock TBM. Geophysical surveys have not been used yet either, as the exact location for downhole spacing appears to cover too wide of an area to obtain meaningful results (see Fig. 8). 4.4

MH5 deletion and steel segment connections

SMH #5 was deleted by Change Order due to neighborhood mitigation concerns in a quiet residential area for $400,000 credit just after NTP. Prefabricated welded steel segments similar in shape to the pre-cast concrete segments were used at SMHs to allow for future torch-cutting for maintenance hole riser and conduit connections. Slightly longer than the liner plates used on ECIS, the steel segments allowed for a minimal production mining delay, and were also used to initially launch the machines from the shafts. 4.5

Gassy Rock Tunneling

Upon initial mining startup on the Upper Reach, several methane spikes through rock fissures and backfill grout ports caused the tunnel to be reclassified from a “Potentially” to “Gassy Tunnel”, with advanced probe-hole drilling required prior to tunneling. Extra labor and time for the probe drilling has been tracked as a Time and Materials Change Order. Minor TBM equipment reconfigurations were performed to accommodate additional ventilation lines added to dilute any future methane emissions. 4.6

Annular backfill grout

A two-part aggregate-less pumpable grout initially developed in Japan by Tachibana was used in the annulus between segments and TBM overcut, eliminating grout cars and centralizing the batch plant operations at the surface. A cement-bentonite-flyash mixture with a retarder was mixed at the injection nozzle with a sodium silicate accelerator. The twopart grout was pumped through the segment ports during mining, and generally remained viscous for about 20–30 minutes prior to initial set allowing for full extrados coverage from a single grout port. Grout injection alternated between rings from 10 o’clock to 2 o’clock for more complete extrados grout coverage.

Figure 8. The Narrows Gallery Tunnel.

5 PROGRESS AND SCHEDULE As of mid-November the construction work is roughly 50% complete, however the project remains several days behind the original baseline schedule due to mining startup delays. The critical path originally followed the Upper Reach mining activities. However, due to late delivery of the first TBM the path shifted briefly to the Lower Reach, then on to the Middle Reach as machine shop fabrication and delivery delays forced the third TBM to arrive on site about 3 months late. Probe drilling on the Upper Reach due to tunnel reclassification and investigative boring work for the Narrows Gallery has also slowed progress. Geotechnical instrument installation, carrier pipe and segment pre-casting activities are well ahead of schedule and all SMHs have been drilled, cased and covered. The Low Rock Cover area and Whittier Blvd. permeation grouting are currently underway but well ahead of schedule.

6 LESSONS LEARNED 6.1

467 Copyright © 2004 Taylor & Francis Group plc, London, UK

Benefits of probe drilling for gases

For EPB machines, gas detection occurred only at the screw conveyor discharge, at leaks through the cutterhead access door and through ground water sump

09096-57.qxd

2/18/04

2:33 PM

Page 468

Figure 10. Lower & Middle Reach backfill grout and foam batch plants at Richmond Shaft.

easy to use and provided sufficient annular coverage and settlement support in silts/clays and rock to date.

ACKNOWLEDGEMENTS

Figure 9. Middle Reach TBM shield hoisting at Richmond.

pump discharge and was not detected by probe drilling. The Upper Reach also did not discover any methane through drilling as all spikes were detected with the hand held monitor at the belt discharge point or at backfill grout ports behind the tail shield. TSFK has urged CalOSHA to revise the Tunnel Safety Orders to reduce or eliminate the requirement for probe drilling in advance of tunneling when using EPBMs. 6.2

Twin mining operation logistics

Delivery and assembly of two TBMs at one shaft site was hampered by the available real estate on site and delayed systems testing prior to shipment, thus causing additional effort to hoist the 3-section Middle Reach TBM cutterhead, stationary and tail shield into the shaft as one piece. (See Fig. 9) Centralized maintenance shop, backfill grout and foam batch plants, and joint segment storage should facilitate mining operations. (See Fig. 10) 6.3

The authors would like to thank the following for their assistance with the writing of this paper: Baron Miya, City of Los Angeles, BOE; John Critchfield, Parsons Brinckerhoff; Michael McKenna, Steve Dubnewych and Francis Fong of Jacobs Associates; and Michael Traylor and Brett Robinson of TraylorShea-Frontier-Kenny, JV.

REFERENCES City of Los Angeles, Bureau of Engineering. 2000. NOSECIS Conformed Contract Documents. City of Los Angeles, Bureau of Engineering. 2001. NEIS Conformed Contract Documents. Hanks, K., Fong, F., Edgerton, W. & Miya, B. 1999. City of Los Angeles Large-Diameter Interceptor Sewer Tunnels, Proceedings For The Rapid Excavation and Tunneling Conference 1999 pp. 308–329. McKenna, M., Traylor, D., Tarralle, B. & Itzig-Heine, E. 2003. Proceedings For The Rapid Excavation and Tunneling Conference 2003 pp. 368–382. McKenna, M., So, K., Krulc, M. & Itzig-Heine, E. 2003. North American Tunneling Conference 2004. TSFK, JV, 2002 NEIS Baseline Schedule.

Annular backfill grouting

After the initial systems setup and testing on first 200 segments installed, this grouting system proved to be

468 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-58.qxd

2/18/04

2:34 PM

Page 469

Session 4 Design/build risk

Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-58.qxd

2/18/04

2:34 PM

Page 471

Session 4, Track 1 SEM/NATM practices/prescriptive specifications

Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-58.qxd

2/18/04

2:34 PM

Page 473

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

NATM and its practice in the US Wern-ping (Nick) Chen & Hugh Caspe HNTB Corporation, Boston MA, USA

ABSTRACT: Since its creation over 40 years ago, the New Austrian Tunneling Method (NATM) has been constantly debated by tunnel designers and contractors around the world. Is NATM a tunnel construction method or a design concept? Does it have a solid theoretical background? Is it just a general tunnel design philosophy with an Austrian name? Is it applicable to soft ground tunneling? We have seen many NATM perspectives from tunnel practitioners in Europe. What is the perspective from the US tunnel practitioners? What lessons have we learned from NATM since it was introduced to the US in the 1970s? Its application in the US is implicit and slow. What are the contractual concerns involved in putting NATM into practice in the US? American may have applied the NATM principles without explicitly referring to its name for decades. This paper updates the NATM perspectives from tunnel practitioners in Europe and in the US and suggests resolutions to those NATM contractual concerns in the US tunneling industry.

1 INTRODUCTION Named by Rabcewicz around 1960s, NATM is also referred, by some tunnel practitioners, as the Observation Method, the Sequential Excavation Method (SEM), or the Sprayed Concrete Lining (SCL) method. Its excavation method usually divides the excavation face into heading, bench, and invert excavations, and the excavation technique is typically by roadheader or drill and blast. During tunnel construction, shotcrete, rock dowels, lattice girders, and grouted anchors are generally installed as initial supports in accompany with heavy geotechnical instrumentation. Its final lining can be another layer of shotcrete lining, single shell, or cast-in-place concrete lining. The NATM construction is flexible and can fit almost any tunnel geometry. The first tunnel in the US that is designed and constructed by the NATM principle is the Mt. Lebanon Tunnel in Pittsburgh in the 1970s. It origin design include conventional and NATM alternatives and the winning bidder selected the NATM alternative. Since then, several tunnels in the US have been designed and constructed using this method, including Washington Metro’s (WMATA) Wheaton Station and tunnels, for which NATM was proposed by the contractor through a value engineering change proposal process. NATM’s application in the US, primary for its cost-effectiveness, resides on the short non-circular transportation tunnels.

To adopt a new principle is always challenging. Nobody likes changes, especially when he/she performs well in his/her current practices. On the other hand, it is dangerous and risky in adopting a new principle without fully understand all of its associated requirements. This view fits well for the introducing of NATM to the US tunneling industry in the 1970s to 1980s. One may fully adopt and understand the NATM design philosophy, but may neglect its other required ingredients for success, such as contracting and construction practices. The following sections discuss the NATM principles, its contracting practice, its issues that are specific to the US, and possible resolutions to those concerns and issues. 2 NATM PRINCIPLES The fundamental principle of NATM is to mobilize the inherent strength of the ground around an opening during construction such that the designed initial support would be flexible and fully in compression as a ring. This ground-lining interaction behavior is a 3-D phenomenon as depicted in Figure 1. Though this fundamental principle is clear, arguments still exist among tunnel engineers since late 1970s about what is NATM and what are the NATM principles, even among experts in this field. However, most tunnel engineers would agree that NATM is a philosophy/ concept or an approach rather than a construction

473 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-58.qxd

2/18/04

2:34 PM

Page 474

in economic retaining structure designs. This application is known to retaining structures for civil engineers, though its principle is obviously applicable for underground excavations.

3 NATM CONTRACTING PRACTICE

Figure 1. Ground lining interaction: conceptual diagram (HSE, 1996).

method. It seems that argument about NATM is always around its application to soft ground and weak rock tunneling. Readers interested in the discussions of NATM principles should refer to publications by Miiler (1978), Golser (1978), Brown (1981), Kov’ari (1994), Wood (1994), Ayaydin (1995), Barton (1995), Sauer (1995), F B de Mello (1996), Scholey (1996), and ICE (1996). Major elements of the NATM philosophy are summarized below (Brown 1981):

• • • • • • •

The preservation of the ground inherent strength should be to the maximum extent possible, Controlled ground deformation, at an acceptable level, to allow the ground to develop its full inherent strength, Support system consisting of rockbolting (dowel) and a thin flexible shotcrete lining, The timing of the installation of the initial support is crucial and varies from case to case, The initial support will partially or completely represent the total support required, The tunnel unsupported length shall be as short as possible, and Design and supervisory engineers and the contractor’s engineer (all parties involved) must understand and accept the NATM approach and adopt a cooperative attitude to decision making and the resolution of problems.

When NATM was developing, the soil mechanics principle for retaining structures (which is similar to that of NATM) was also developed. Karl Terzaghi, one of the greatest geotechnical engineers in the modern history and a proponent and practitioner of the observational method, has pioneered, between 1920s and 1930s, the earth pressure theory to mobilize the inherent strength of the ground such that a retaining structure can be as flexible as practical to retain the earth behind it. This principle would result

The standard NATM practice would specify the initial ground supports into three to six classes, which depends on the complexity of the ground along a proposed alignment. Each support class defines its own support elements (such as rock dowels, spiles, lattice girder, shotcrete, and probing requirement), excavation sequence, and the length of each excavation advance to match the type of ground anticipated. In contract document, the length of each support class is pre-estimated, in unit price (including installation and material), for bidding purpose. The payment is made by actual length of each support class that is installed in the field and is agreed upon by both representatives from the owner and the contractor. If dispute occurs for the class of ground support to be installed, the contractor will make the ultimate decision. The owner can either withhold payment should a higher class of support be placed than he/she thinks is required, or order the contractor to place a more robust support class if he/she thinks the ground is going to be undersupported. Yet, ultimate safety of the opening and the construction workers belongs to the contractor and difference in the class of support often times becomes a claim. Other construction items, such as the handling and control of groundwater inflow and grouting, will also be paid in a similar mechanism. That practice is similar to current US practice for tunnels that are constructed in a two-pass system, which may include initial supports, of different classes to fit anticipated ground condition, and final lining. However, differences exist between the original NATM and the US practices, including:

•

•

474 Copyright © 2004 Taylor & Francis Group plc, London, UK

NATM philosophically extends its design spectrum into the construction phase. This is depicted from the quote by Golser (1978), “At the planning stage however only a preliminary analysis and design will form the basis for the contract documents. But as soon as construction is under way a continuous process of measuring and checking deformations and stress in situ will provide a guideline for the actually required strengthening and supporting measures and will enable the contractor to change and adjust his construction procedures.” This philosophy may vary slightly from practice to practice, but forms the fundamental NATM spirit. NATM promotes the concept of risk sharing, which relies only on its flexible contract practice, including

09096-58.qxd

2/18/04

2:34 PM

Page 475

various types of ground supports to fit actual ground condition encountered in the field, and the trust between the owner and the contractor. In the US, we conduct our tunneling practice within a sophisticate legal system. This has automatically put our designers, owners, and contractors in defensive seats since day one of a project. We have instituted risk management tools, such as Geotechnical Baseline Report (GBR), Escrow Bid Document (EBD), Dispute Review Board (DRB), Differing Site Condition clause, partnering, and contractor’s prequalification process to the extreme so as to prevent litigation. These tools are working at present, but new tools may be needed to fit the new construction related involvement of the designer during NATM construction. 4 NATM ISSUES IN THE US

•

•

Owner’s issues include:

• • •

As of any construction project, its execution from the beginning to the end relies on all parties involved, including the owner, the designer, and the contractor. Each party has its own issues in the development of NATM practice in the US. These issues are discussed below. Designer’s issues include:

•

•

Misapplication of NATM principles. Structural engineers normally perform tunnel initial support and final lining designs and geotechnical engineers perform tasks for geotechnical design parameters and ground behavior interpretations. Quite often, structural engineers are asking for the loads for support designs and forget the true physical phenomenon resulted from excavating a ground is the deformation. Stronger is better, a misconception of engineers. Stronger support elements result in relative high stiffness in comparison to that of the surrounding ground and attract more loads (less deformation) from the ground during excavation. It is not new to hear that I need double layers of reinforcements to satisfy the State’s seismic requirement. Uneconomical design by not taking into consideration the contribution of the initial support to the final lining. This is not a unique issue to the US as it is also a standard practice in UK and Austria. This approach is contrary to the original intention of the NATM, for which the initial support should always be part of the final lining for an economical design, as indicated by Golser (1978). It is this writer’s opinion that if the design has included the consideration that the initial support has mobilized the inherent ground strength, it should be considered as part of the final lining in theory. Reasons given for not taking the initial support into consideration are the deterioration of the shotcrete from

•

Very few owners have the experience or opportunity working on NATM projects, which results in the lack of NATM knowledge. Back analysis and support class revision during construction is not uncommon for NATM practice. Are owners willing to recognize the benefit and pay for the design? Owners have difficulty in finding its NATM field representatives who are knowledgeable of NATM designs, specifications, and the ground condition to be encountered and are willing to give control to other parties. The involvement of designer’s NATM field representatives during construction is normally not explicitly included in the design scope. Though it may be included in the construction phase service, the scope of Engineering Service During Construction (ESDC) is usually negotiated as a separate contract or as an amendment to the existing contract. At the time this new scope is negotiated, the design may have been completed. This may put the designer in an uneasy situation during design, which may result in uneconomical products (more robust more than required), since the engineer is worried that he/she may not have any involvement in the field to verify his/her design.

Contractor’s issues include:

•

• •

475 Copyright © 2004 Taylor & Francis Group plc, London, UK

the corrosive accelerator admixture that is added in during construction and the poor workmanship from nozzleman. Both these reasons need to be clearly reexamined in the advent of the latest technology developed in these areas. Designers may be reluctant to be owner’s field representative to avoid any potential adverse responsibility that may add upon them, especially when that work is not profitable.

Very few US contractors have NATM construction experience in the US. To quickly gain their competitiveness in the market, the contractor may turn to foreign engineers or firms who have NATM experience to build up their NATM strength. Though these engineers mostly are qualified to perform NATM works, they may not be allowed to perform engineer’s function under the jurisdiction of several States’ licensure requirement. Additionally, these engineers or firms may not familiar with US contractual requirements. It is difficult to find experienced NATM miners in the US. Consequently, the quality of the workmanship may be difficult to control. Contractor may impose safety provisions for higher ground support class than that is actually needed for the ground condition encountered. On the other hand, the contractor may down size the

09096-58.qxd

2/18/04

2:34 PM

Page 476

ground support class to catch up his/her schedule, which may result in safety issue. 5 POSSIBLE RESOLUTIONS Panacea does not exist in our current practice, but improvements can be made for NATM practice in the US. The followings are possible resolutions for issues and concerns raised in the previous section.

• • • •

•

•

•

Trust must exist among all parties involved in the process. All parties must understand the NATM principles, both technically and contractually. Include or negotiate, in the design phase, the engineering service during construction scope with the owner. Design build contract may minimize conflicts among parties, since designer works for the contractor in a single design/construction package. Also, the contractor is more willing to promote innovation for cost saving ideas. Utilize contractor’s prequalification process, which has been successfully used in many projects in the US; however, disadvantages also exist in this process as indicated by Edgerton (2002), “(1) It’s expensive and time-consuming – the process may take four to six months, (2) It can restrict competition which drives up bid prices, and (3) It can result in bid protests, from either direction.” Provide training for construction workers, but we must recognize training cannot replace experience. Questions that develop include: Who should provide the training and who should pay for such training? When shall the training be conducted? Who is responsible for the training? Once who is trained, does it mean he/she is qualified for the NATM work? Qualify NATM field engineers/representatives based on experience. For those engineers who are not licensed in a specific State, they should work under the supervision of engineers who are licensed.

6 CONCLUSIONS For the NATM practice to survive in the US, it must be cost-effective in comparison with other conventional

tunneling practices. Cost-effectiveness relies on an efficient execution process from all parties involved; therefore, a successful NATM project requires a fully cooperation and trust among the owner, the designer, and the contractor through flexible contracting practices both in the design and construction phases. It also demands the understanding of the fundamental NATM principles from all parties involved. The lack of NATM knowledge, both in theory and in contracting approach, among the US tunneling industry may be the major hindrance for the development of NATM practice in this country.

REFERENCES Ayaydin, N. (1995) “The Choice between NMT and NATM,” Tunnels & Tunnelling, January. Barton, N. (1995) “Updating the NATM,” Tunnels & Tunnelling, December. Brown, E.T. (1981) “Putting the NATM into Perspective,” Tunnels & Tunnelling, December. Edgerton, W.W. (2002) “Prequalification: Is it the Silver Bullet?” Tunnel Business Magazine, October. F B de Mello, V. (1996) “Fallacies in NATM/RSST Shotcrete Supported Tunnelling,” Tunnels & Tunnelling, July. Kov’ari, K. (1994) “Erroneous Concepts behind the New Austrian Tunnelling Method,” Tunnels & Tunnelling, November. Golser, J. and Mussger, K. (1978) “The New Austrian Tunnelling Method (NATM), Contractual Aspects,” Tunnelling in Difficult Ground, pp 387–392. Health Safety Executive (1996) “Safety of New Austrian Tunnelling Method (NATM) Tunnels.” Institution of Civil Engineer (1996) “Sprayed Concrete Linings (NATM) for Tunnels in Soft Ground.” Miller, L. (1978) “Removing Misconceptions of the New Austrian Tunnelling Method,” Tunnels & Tunnelling, October. Rabcewicz, L.V. (1964, 1965) “The New Austrian Tunnelling Method”, Water Power. Sauer G. (1995) “Updating the NATM,” Tunnels & Tunnelling, December. Scholey, J. (1996) “NATM in the UK: the Debate so far,” Tunnels & Tunnelling, September. Terzaghi, K. and Peck, R.B. (1948) “Soil Mechanics in Engineering Practice,”, John Wiley & Sons, Inc. Wood, A.M. (1994) “Can the Newcomer Stand up?” Tunnels & Tunnelling, September.

476 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-59.qxd

2/18/04

2:34 PM

Page 477

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

SEM/NATM design and contracting strategies Joseph Gildner Sound Transit, Seattle, WA

Gerhard J. Urschitz Dr. G. Sauer Corporation, Herndon, VA

ABSTRACT: The Sequential Excavation Method (SEM), also referred to as the New Austrian Tunneling Method (NATM) is becoming increasingly popular in the US for the construction of shafts, tunnels and other underground structures. Besides the economic competitiveness of SEM, advantages of the method include the outstanding flexibility in terms of geometric shapes and its adaptability to varying ground conditions, as well as the substantial reduction of impacts and disruptions to communities in urban areas. Even though used in the US since 20 years, it is still considered a new method grasping for acceptance and support in the engineering as well as the contracting community. To take best advantage of SEM and to best utilize its cost-saving potential strategies have to be developed that include establishing of contractual frameworks, procurement strategies and unit price contracts, employment of skilled and experienced designers willing to take responsibility for a prescriptive design, extending the designer’s responsibility and involvement into the construction period and allowing for a competitive bidding environment. Technical pre-qualification of contractors needs to be reviewed on a case by case basis and should only be required if varying ground conditions dictate frequent change of support means, application of more sophisticated additional support measures like SEM Toolbox items and utilization of complex excavation sequences with multiple headings. Some of the strategies extend beyond what is common practice, and therefore require flexibility and adaptability of owners as well as contractors to deliver a successful project. The Beacon Hill Light Rail Station in Seattle is a recent example where all these issues have been considered and implemented.

1 INTRODUCTION About 20 years ago, in the late 1970s and early 1980s the first applications of the New Austrian Tunneling Method (NATM) for tunnels in Pittsburgh and Washington, D.C. generated attention within the US consulting and construction industries (Martin, 1984). The new method to construct tunnels resulted in low bids either in a competitive bidding environment (bidding on two alternative construction methods) or as value engineering by contractors (Cavan et al., 1985). Due to the success of the NATM value engineering solution the Washington Metropolitan Area Transit Authority (WMATA) decided to use the method either as alternative or sole design on their upcoming contracts (Heflin, 1985). Since then a variety of projects have been designed and constructed using NATM, a method recently also referred to as Sequential Excavation Method (SEM), including transit and highway projects in Dallas, TX,

Boston, MA, Jersey City, NJ, New York and Seattle, WA. Modifications to the original concept were required to adjust to the domestic design and construction environment. These include the more robust base design, peer reviews, pre-qualification of contractors, strict requirements for construction management and supervision, the careful evaluation of proposed means and methods, application of risk management and the adjustment of the contracting strategy. For the method to be successful and to best utilize its flexibility and cost savings potential a series of requirements has to be fulfilled. Owners have to adapt and accept the concept of the method that is labeled as “innovative” and “new” by demonstrating willingness to reach beyond well established boundaries in terms of procedural and contractual frameworks. Skilled and experienced designers shall be employed to continue the optimization process and provide economic design solutions combined with a robust contractual framework. Finally, due to the lack of experience of most

477 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-59.qxd

2/18/04

2:34 PM

Page 478

development of the most economic, but of course safe design for all tunnel structures, considering all relevant issues, including geology, overburden, settlement sensitivity and technical feasibility. Reference is made to a series of publications where in-depth information on the principles of NATM/SEM for rock and soft ground are described in detail.

3 PROCUREMENT STRATEGY: DESIGNBUILD, DESIGN-BID-BUILD OR COST PLUS FEE? Figure 1. Typical NATM Excavation Sequence in soft ground.

American contractors especially with soft ground mining methods in urban areas and the importance of observing the ground and its movements, the construction has to be supervised in a way to assure that the design intent is conveyed into the construction phase.

2 SEM/NATM – A DESIGN PHILOSOPHY NATM was developed in the 1950s with shotcrete being an integral part to stabilize openings immediately after excavation of each round. Initially developed for rock tunnels, the method was advanced in theory and practice and adapted for soft ground tunnels in urban environments, with the first application in Frankfurt, Germany in 1968. Since then NATM was used throughout the world on many projects for transportation, water/wastewater conveyance or other purposes. Due to the modifications of the method to adjust it to the US market the term SEM is sometimes used instead of NATM. Such modifications include the development of “robust” design as a base case supplemented with toolbox items, instead of following the original idea of an observational method, where support measures are selected based on geology at the face and deformations of the ground and lining. However, the general principle of mobilizing the surrounding ground by developing the maximum selfsupporting capacity of the rock or soil remains unchanged. “Greater skill is needed to avoid (minimize) ground load than to resist it” (Rziha, 1872), a statement from a tunneling engineer of the 19th century is still relevant and valid today and separates experienced designers from inexperienced incumbents in the field of SEM design. It is not an achievement to turn soft ground into concrete before mining a tunnel and thereby incurring substantial costs. The challenge and satisfaction of a SEM designer shall always be the

One of the early decisions an owner has to make is to decide on the procurement strategy. Assuming that no participation in funding of the project by the contractor is required (i.e. Public Private Partnership), there are generally three contracting options available:

• • •

The aim of the procurement strategy is to achieve the optimum balance of risk and control, depending on the legal environment, where the project is located. The procurement route should ensure that design, construction, operation and maintenance are considered as a whole, and that the delivery team for all of these aspects works together (OGC, 2003). 3.1

Design-bid-build

The traditional design-bid-build contract, where a single contractor acts as the sole point of responsibility for the management and delivery of a construction project on time, within budget and fit for the purpose for which it was intended is the most widely used and suitable form of procurement for underground construction, especially where ground conditions are difficult, and the owner has high expectations on the quality of the final product. Compensation for the contractor is accomplished with lump sum or unit price payments or a combination of both, depending on the complexity and predictability of the project. SEM project experience has shown that some adjustments are required. Unit price contracts are generally preferable to maintain the flexibility and adaptability of the design to the actual ground conditions encountered. It was also found valuable to extend the services of the SEM designer into the construction period due to the detailed knowledge he has about ground, geotechnical and geo-hydrological conditions and the intensive design development required for a thought-through SEM design. The design philosophy and design intent are thereby conveyed into the construction phase.

478 Copyright © 2004 Taylor & Francis Group plc, London, UK

Design-Bid-Build Design-Build Cost plus Fee

09096-59.qxd

3.2

2/18/04

2:34 PM

Page 479

Design-build

Design-build can be defined as “using a single contractor to act as a sole point of responsibility to a public sector client for the design, management and delivery of a construction project on time, within budget and in accordance with a pre-defined output specification using reasonable skill and care” (OGC, 2003). There have been many attempts throughout various countries to make the design-build concept work for the construction of underground structures. While the approach seems reasonable and has been applied for many projects with success, underground construction has one major specific characteristic, namely the ownership of the nature and behavior of the ground. Differing site conditions can result in claims of large magnitudes with lump sum contracts due to the lack of detailed information about the contractor’s bid. Risks of the design-build contract include the transfer of too much or inappropriate risk, which might not be cost effective. Also, the clarity of the output specifications that might be based on assumptions that are inaccurate represent the potential for major cost increase in the case of changes or adjustments of these specifications. A strong and powerful construction management with a knowledgeable SEM support team on the owner’s side, and an experienced contractor are essential for successfully completing mined tunnel projects on a design-build basis. 3.3

Cost plus fee

This procurement option also known as time and material (T&M) contract is rarely used for the construction of mined tunnels due to the high degree of responsibility and risk that remains with the owner. A highly experienced team on all sides, the owner, designer and contractor are required for successfully completing projects using this approach. One successful application of this concept was the construction of seven cross-over caverns at the Exchange Place Station in Jersey City, NJ (Dinkels, 2003). The owner, the Port Authority of New York and New Jersey was under high pressure to rehabilitate and reopen the station affected by the collapse of the World Trade Center in New York and re-open train service between New Jersey and New York for the millions of commuters at a specified date. The schedule driven project did not allow any of the traditional contracting methods, therefore the cost plus fee contract was selected. Close coordination between the owner, the design team and the contractor combined with a very experienced and powerful construction management resulted in the on-time completion of this project. The owner requested from the contractor a fixed fee and a lump sum fee for its project management. A

Figure 2. Exchange Place Station project, Jersey City, NJ.

substantial incentive was provided to finish the project on or even before the specified completion date, and no penalty was foreseen for late completion. This system generated a team-like environment where every party had the same goal and focus. 4 SEM DESIGN PHASE Designing SEM tunnel structures is different from designing other structures due to the complexity of and the interaction with the surrounding ground, when considering stress redistributions and development of equilibriums after each excavation and support step. Excavation sequences, pre-support, face support and ground improvement methods have to be developed depending on the size of the structure and the ground conditions. A comprehensive subsurface exploration program is essential and becomes more critical the more complex ground conditions are. An essential part of the investigation is the determination of the hydrologic regime, which can comprise of multiple groundwater horizons. In some cases, especially when the ground has been affected by seismic or tectonic activities it will be impossible to determine the exact stratification of the various subsoil layers. The variability has to be taken into consideration during the design. While it is typical to develop excavation and support classes based on various available classification systems for hard rock, soft ground is usually more variable and results in mixed face conditions of some sort. Traditional classes are therefore difficult to define. A different concept is required to maintain flexibility but also to assure proper support during all phases of

479 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-59.qxd

2/18/04

2:34 PM

Page 480

Figure 3. SEM toolbox.

construction. Therefore the “SEM Toolbox” concept has been developed to cope with all, including the most adverse conditions. 4.1

The SEM toolbox concept

Developed for NATM tunneling in soft ground, the SEM Toolbox concept follows the requirement for multiple or selected pre-support, ground improvement and support elements to stabilize the open face of an excavated tunnel heading. The complete toolbox contains all elements to allow mined tunneling through virtually any ground (Sauer, 2003). For each project, in addition to standard support measures toolbox items are selected that are required for the expected ground conditions and contingency procedures. The toolbox items are defined in the contract documents in terms of material, installation/application procedure and installation/application criteria. Figure 3 shows a typical cross section and longitudinal section for an SEM tunnel during excavation. SEM Toolbox items are numbered and described below. Standard support measures typically include fiber reinforced flashcrete (9), reinforced shotcrete (2) and standard dewatering measures. SEM Toolbox items include the following: Geometry and Sequence: • Top Heading/Bench/Invert • Sidewall Drift • Dual Sidewall Drift Sidewall Drift Improvements: • Foundation for Sidewall (4) • Increase Bearing Capacity of Sidewall (6) Pre Support Measures (3): • Rebar Spiling (3a) • Grouted Pipe Spiling (3c) • Metal Sheets • Grouted Barrel Vault/Pipe Arch (3b)

Face Stabilization Measures: • Face Stabilization Wedge (1) • Pocket Excavation (10) • Face Bolts (8) • Reduction of Round Length Ground Improvement Measures: • Dewatering and Vacuum Dewatering (8) • Permeation Grouting, Fracture Grouting, Jet Grouting Annular Support Measures: • Additional Shotcrete • Soil Nails (5) • Temporary Invert (7) While being in conflict with the more traditional lump sum payment method, the SEM Toolbox concept requires the use of unit prices for the various items selected as part of the project toolbox, while standard support measures are still compensated with lump sums. Quantities for the toolbox items are estimated and form the basis for the bid. Standby quantities are defined which must be available on site ready to be applied at any time during excavation of the respective SEM tunnel. Actual quantities will vary and depend on the actual ground conditions encountered. Application and installation of the toolbox items shall be approved or directed by the Engineer. For the contractor it is important to have criteria for the installation/application of each of the toolbox items as well as the estimated quantity for each tunnel reach. For this purpose typical tables are prepared that define relationships between Ground Type and SEM Toolbox Item (see Table 1), and Tunnel Structure and SEM Toolbox Item Quantities, which should be part of the Geotechnical Baseline Report. 4.2

480 Copyright © 2004 Taylor & Francis Group plc, London, UK

Instrumentation and monitoring

Instrumentation and monitoring of ground movements and lining deformations is essential for successful SEM tunneling. Instrumentation comprises surface

09096-59.qxd

2/18/04

2:34 PM

Page 481

Table 1. Relation between Ground Type and SEM toolbox item.

Support elements Standard support measures Flashcrete Reinf. shotcrete Gravity Dewatering Probing Barrel vault

Stable Clays and Silts, Till

Clays and silts, Soft or Slickensided

Dry or Moist sands and silts

Wet sands and Silts

Comments/ Location

X X X

X X X

X X X

X X X

As required

X

X

X

X As specified

Toolbox items Rebar spiling

X

X (moist) X (dry)

Grouted pipe Spiling Metal sheets Face bolts Grouting

Crown X (stable) X (flowing)

X X

X

Soil nails

X

X

X

X

Additional shotcrete

X

X

X

X

Reduced round Length Vacuum Dewatering Jet grouting

X

X

X

Face stabilization Wedge Pocket Excavation

X

monitoring, ground instrumentation, and in-tunnel instrumentation, to monitor lining performance. Surface monitoring includes monitoring of surface settlements and building deformations, and can require surface settlement points, tiltmeters, crack monitoring devices and vibration monitoring, depending on the sensitivity of the overlying and adjacent structures. Instrumentation for the surrounding ground includes inclinometers, extensometers and deflectometers. The purpose of these instruments is to determine the loosening zone in the ground, and to observe movements of the tunnel face by monitoring the ground ahead of the excavation. In-tunnel monitoring includes convergence and lining deflection measurements, and pressure and stress monitoring by utilizing ground and shotcrete pressure cells. For the ground and in-tunnel monitoring the reading frequency and reporting procedures are essential. Timely reporting to allow interpretation of data is the

Crown/Invert Face Crown/Face/ Invert To cease deformations of Crown To stabilize or cease deformations Crown/Face

X

Crown/Face

X X

X

Crown/Face if required Face

X

X

Face

base for decisions at the face regarding the adequacy of the support, round length and sequence of installation. While monitoring of surface and ground deformation instruments is carried out from the surface and can also be done remotely, in-tunnel monitoring disrupts the excavation and support process significantly. Today, optical surveying methods are used that are less disruptive to operations but still require an instrument to be placed in the center of the drift or tunnel. The decision of who is carrying out the readings requires some consideration. Because it is a disruptive activity it is recommended to make it a responsibility of the contractor. This option allows the contractor to stage his work such that monitoring is a part of his excavation and support sequence. Two options are feasible under this approach, that the contractor either reads himself, with qualified survey personnel meeting or exceeding the specified

481 Copyright © 2004 Taylor & Francis Group plc, London, UK

Crown

09096-59.qxd

2/18/04

2:34 PM

Page 482

requirements, or that he employs an independent, qualified surveyor. When tunneling underneath sensitive buildings an owner might be more comfortable specifying an independent surveyor. The contractor is responsible to submit the raw and processed data to the owner in a pre-determined time frame, where raw data should be submitted immediately after they become available, and processed data should be available no later than 24 hours after reading. Alternatively, all surveying could be under the owner’s control by employing a surveyor to read the in-tunnel instruments. This approach requires close coordination between the surveyor and the contractor and often results in access problems to carry out the reading. Damaged survey points, muck piles and parked equipment are some of the obstacles to overcome. Therefore, this approach is generally not recommended. 5 PRESCRIPTIVE VS PERFORMANCE DESIGN

•

Unique in the American way of design and construction is the high degree of performance elements in a traditional final design package rather than providing a prescriptive base design, where contractors can propose alternatives and submit value engineering proposals. The concept behind this philosophy is that final decisions on means and methods of construction remain with the contractor. This should guarantee most economic solutions for mainly temporary construction elements where contractors have more knowledge or provide the flexibility to choose the most preferable solution based on the contractor’s experience. In case of SEM tunneling, designers are usually more knowledgeable in terms of ground conditions and behavior at the time of contract award due to the extensive exploration and research during the design phase and the resulting development of excavation sequences based on experience and structural analysis. Leaving the design of the temporary support, in this case the ground improvement, pre-support and initial shotcrete lining to the contractor is not reasonable in the short time frame allowed between advertising the project and start of construction. In addition, contractors can usually not refer to the extensive SEM experience designers provide due to their involvement in projects worldwide. To reduce risk of cost and schedule overruns the following elements of the SEM design should be prescribed:

• • • •

Figure 4. Risk vs. Degree of Prescriptiveness for Design of SEM Projects.

Excavation Sequence (top heading/bench/invert, sidewall drift or dual sidewall drift excavation) Advance Length (for each round) Ground Improvement (grouting, dewatering) Pre-support Elements (application of the SEM Toolbox)

• •

The prescription of ground improvement and presupport elements is only possible to the extent that the anticipated methods are specified and quantities are estimated, but final decision on location and application will depend on ground conditions encountered and will be determined in the field. Waterproofing is prescribed due to the many negative examples and leaking tunnels. So far, only the membrane waterproofing system has proven to provide watertight structures. Despite the high degree of prescriptiveness there is still flexibility for the contractor to optimize his operations on site, such as elements like the construction sequence (sequence when various tunnels are built), the advancement of split headings (only minimum distance is specified), utilization of its equipment, selection of wet or dry shotcrete and – with some limitations – the selection of SEM Toolbox items by proposing elements of choice (Sauer & Gold, 1989). 6 CONTRACTOR PREQUALIFICATION Prequalification of contractors for tunneling projects has become popular in the last 10 to 15 years to reduce risk. However, the process does not always provide the expected outcome. Problems encountered include that bidders are pre-qualified based on previous successful projects but the crews have left the company in the meantime, the lack of similar past work experience when new methods are employed, and the difficulty to provide resumes for key personnel sometimes a year or more ahead of the actual construction.

482 Copyright © 2004 Taylor & Francis Group plc, London, UK

Support Elements (flashcrete, reinforced shotcrete lining, lattice girders) Waterproofing Final Lining

09096-59.qxd

2/18/04

2:34 PM

Page 483

There is no question that bidders should be qualified to perform the work, but how can this be accomplished and what/who needs to be prequalified? Upon other criteria responsibility, financial viability and size of the prospective bidder as well as previous experience with similar projects, including key personnel should be evaluated. Tunnel projects are usually singular projects and only few agencies or owners have historic data on successful or unsuccessful contractors and procedures. Also, criteria that would eliminate a specific contractor based on a problem on a previous project are difficult to formulate and execute. It has to be recognized that Prequalification does not relief any owner of the responsibility to monitor and supervise construction (Brierly, 2003). In the US, there is still a lack of experienced contractors when it comes to large underground structures to be constructed in soft ground. The two schools of thought in this case are to qualify all contractors based on financial capability and responsibility, and have a strong, SEM experienced construction management team available to guide the contractor through the process, or to rely on expertise from abroad. In the first case a team-like environment has to be established that allows the construction management to train, guide and work closely together with the contractor heading for a common goal. Alternatively, the contractor could add SEM experienced key staff that would fulfill the same role and train and guide the contractor through the project. One example of a successful application of this arrangement was the widening of the Berry Street Tunnel in Pittsburgh, PA. The contractor with no experience in tunneling, shotcrete or NATM/SEM decided to propose a value engineering alternative utilizing NATM for the tunnel excavation and support. He was supported by experienced NATM engineers and superintendents from his designer. The Port Authority of Allegheny County accepted the proposal and the project was completed ahead of schedule and generated cost savings of $2 million (Garrett, 1998). In the case of the Beacon Hill project in Seattle the owner, Sound Transit decided to increase the prequalification requirements requiring the US contractors to most likely look abroad for the requested expertise. The decision was made based on the fact that the tunnels to be constructed would be the largest SEM soft ground tunnels to date in the US and the station construction is on the critical path of the entire project, the Link Light Rail – South Link. 7 CASE HISTORY – BEACON HILL STATION The above mentioned Beacon Hill light rail station is the most recent project, where large tunnels in soft ground are designed using SEM. During the design

Figure 5. Beacon Hill station.

development many of the above concepts were refined and applied. The complex geology, with fractured, inconsistent glacial deposits and multiple water horizons required the utilization of a series of excavation sequences, ground improvement and pre-support measures in order to provide a safe design. Details about the design of this underground station are provided in another paper of this conference (Laubbichler et al., 2004). It is important to mention that during the design the SEM Toolbox approach was further developed and allowed an economic design approach for the conditions at hand. A strict prequalification was carried out to select qualified and experienced bidders able to utilize the designed tools for a safe excavation. The approach for the instrumentation and monitoring is different for surface and in-tunnel instrumentation. For the in-tunnel instrumentation the contractor will be responsible for the installation, reading and reporting of results. Surface instrumentation will be provided and installed by the contractor and monitored by an independent surveyor. Due to the complexity of the project and the various new features in the design Sound Transit decided to extend the services of the SEM design team into the construction phase and thereby provide SEM inspection and support services.

8 SUMMARY AND CONCLUSIONS Despite the fact that SEM is still a new technology the benefits are recognized throughout the industry. Owners realize that SEM can provide cost effective alternatives to other traditional methods. In some cases, where traditional methods would fail, SEM is the only viable method available to build tunnels. Due to the need for infrastructure in many US cities the utilization of SEM will gain more popularity and become a standard, cost effective construction method. With more projects the experience of owners,

483 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-59.qxd

2/18/04

2:34 PM

Page 484

designers and contractors will increase and become more efficient and cost effective. It will and should be a method where new boundaries for technical feasibility in tunneling can be established. Some flexibility and adaptability is required when utilizing SEM the first time. This includes the evaluation of the contracting strategy, selection of experienced SEM designers, the utilization of the most economic excavation and support system and the use of a prequalification process if required. During construction of SEM tunnels, especially in difficult ground conditions the designers should be involved and part of the SEM construction support team. This is the only way that immediate responses to construction problems are guaranteed and adjustments can be made efficiently. The most advantageous contracting strategy has to be selected on a case by case basis. Experience shows however, that the traditional design-bid-build process is the preferable solution for complex, soft ground tunneling utilizing the sequential excavation method.

Cavan, B., Rhodes, G., Mussger, F., 1985. NATM Provides Improved Design and Construction Method for US Tunnel Projects. 1985 RETC Proceedings, Volume 2: 645–664 Garrett, R., 1991. NATM by any Other Name. World Tunnelling May 1991: 173–180 Laubbichler, J., Schwind, T., Urschitz, G., 2004. Benchmark for the future – The largest SEM soft ground tunnels in the United States for the Beacon Hill Station in Seattle, AUA Conference Proceedings. Heflin, L., 1985. WMATA Use of the New Austrian Tunneling Method for Lining and Support. 1985 RETC Proceedings, Volume 1: 381–391 Martin, D., 1984. How the Austrians cracked the hard American nut with NATM. Tunnels & Tunnelling December 1984 OGC (ed.), 2003. 06 Procurement and Contract Strategies. London: Office of Government Commerce Sauer, G., 1989. NATM Ground Support Concepts And Their Effect On Contracting Practices. 1989 RETC Proceedings: 67–86 Sauer, G., 2003. Ground Support and its Toolbox. Earth Retention Systems 2003 (Conference by ASCE, The Deep Foundations Institute, and ADSC). New York City

REFERENCES Brierly, G., April 2003. Final Thoughts on Contractor Prequalification. http://www.tunnelingonline.com

484 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-60.qxd

2/18/04

2:35 PM

Page 485

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Engineers, contractors, and soft-ground tunneling equipment W.H. Hansmire & J.E. Monsees Parsons Brinckerhoff Quade & Douglas, Inc., San Francisco and Los Angeles, California, USA

ABSTRACT: Major issues confronting soft-ground tunneling involve the equipment. Owners, the Owner’s engineer, the construction contractor, and the tunnel equipment manufacturer have different and often competing interests on a tunnel project. All parties feel they have some stake in what equipment should be used. In current practice, the selection of tunneling equipment is being driven not only by the contractor for considerations of geologic conditions and fundamental constructability, but by Owner and engineer-specified requirements that may be motivated by other factors, such as to reduce risk. A brief history of soft-ground tunneling equipment is presented in order to give a perspective of soft-ground tunneling practices of today. Analysis of past experience and the trends form the basis for presenting the realities of what could, and possibly should, be done in the underground construction industry with regard to tunneling in general and soft-ground tunneling equipment in particular.

1 INTRODUCTION In the business of tunneling, geologic conditions are sometimes erroneously assumed to solely define what can and cannot be tunneled successfully. In practice, however, the tunneling equipment, along with the skill and workmanship applied to operate such equipment, are often the real factors in achieving tunneling success. We present our views based on numerous years of experience on soft-ground tunneling equipment from various perspectives. The goal is to provide an understanding of where the industry is today and to point toward positive changes in the business for the benefit of all involved. 2 EVOLUTION OF SOFT-GROUND TUNNELING MACHINES In less than three decades, soft-ground tunneling machines have gone from very highly labor-intensive hand operations to (almost) horizontal tunneling factories, computer controlled. The first step in this evolution was actually taken almost two centuries ago by Sir Marc Brunel, who wrote, “The great desideratum (sic) therefore consists in finding efficacious means of opening the ground in such a manner that no more earth shall be displaced than is to be filled by the shell or body of the tunnel and that the work shall be effected with certainty” (Copperthwaite, 1906). In other words, never open

more than is needed, can be excavated rapidly, and quickly supported. Following that concept, he patented a circular shield (Figure 1) in 1818 that was described by Copperthwaite as covering “every subsequent development in the construction and working of tunnel shields.” Later (1824–1842) Brunel applied the same principles to the Thames Tunnel. In that tunnel, Brunel controlled the tunnel face by means of a rectangular shield built of cast iron and containing 36 cells; the cells stacked 12 wide and 3 high. Each cell was approximately 1 m wide by 2 m high and occupied by a single worker. Within the cell, the ground was supported by full breasting, using timbers 300 mm high by 75 mm thick (Figure 2). The faces were worked down from the top by excavating and resetting one board at a time at an advance of about 110 mm. It took 18 years (11 years of actual tunnel driving) to complete the approximately 370 m long, brick-lined

Figure 1. Sir Mark Brunel’s patent for circular tunnel shield (Copperwaith, 1908).

485 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-60.qxd

2/18/04

2:35 PM

Page 486

Figure 3. Digger shield with orange-peel breasting doors used to construct wastewater tunnel.

Figure 2. Brunel’s rectangular shield used to construct Thames Tunnel (Copperwaith, 1908).

tunnel, for an effective advance rate of 0.1 m/day. Numerous accidents and problems were overcome, including one in which the works were flooded and six miners were drowned. Isambard Brunel (the son) was swept away but survived. Ultimately the tunnel was completed, and it is still in service as part of the London Underground. If we fast forward, we find that nearly all softground tunnels driven in North America into the 1960s and early 1970s were driven using the basic concepts of the Brunel tunnel shield; viz. compartmentalized, face breasting with timber and lots of hand labor. This for certain applied to small diameter tunnels (under 3 m diameter), some work on BART in San Francisco, and other such as in Mexico City. Incidentally, working in extremely bad ground conditions in Mexico City, the second author showed in the early 1970s that 6 m diameter tunnels could be driven 22 m/day, with laborers who didn’t know they weren’t supposed to be able to do so. In ground conditions that required a higher level of support than the basic Brunel shield, compressed air was commonly used (actually from the mid 1800s into the 1980s). When used correctly, compressed air provided the needed support and allowed many tunnels to be completed that would otherwise not have been possible. Because of the decompression required and all the associated equipment and procedures, not to mention the potential hazards to the workers, e.g., the bends or even death, compressed air was clearly not an acceptable alternative to modern tunneling. Hence, except for special, last-ditch cases (such as obtaining working access to the front of a tunneling

machine for repair or obstacle removal) compressed air has largely been eliminated as a tunneling adjunct. Starting in the late 1960s and early 1970s, some mechanization began to be introduced into the basic shield as with the “Big John” digger. Originally used in a weak rock tunnel, Big John was a very robust, hydraulically operated excavator installed within a tunnel shield. With a new shield, Big John was subsequently used for the first soft-ground tunnel on WMATA in Washington, D.C. 1971. From this development in soft-ground tunneling equipment, the terminology “digger shield” emerged. Tunnels of typically smaller size are built with these open-faced shields using a wide variety of diggers that range from standard rubber-tired back hoes, to all sorts of customized hydraulically operated excavators, all with varying success. As excavation was being mechanized at the tunnel face, tunnelers were motivated to mechanize other key activities. There followed several tries to improve equipment for support of ground at the face. One example being the concept of “orange peel” face support and poling plates, illustrated by Figures 3–5. In these machines, the perimeter was intended to be supported by various configurations of pivoting or extending plates. The lower 90–120° of invert was supported by an earth plug, and excavation was by a digger mounted in the middle of the shield. Such machines too often met with poor results and were usually unsatisfactory for three reasons: 1. Ground loss occurred ahead and above the shield when retracting the doors or poling plates. Typically, the orange-peel doors could not be retracted in tune with the forward progress of the shield. Also when retracting the doors, the miner does not have access to deal with running ground. Thus the machine encouraged unwanted ground movement, rather than controlling it.

486 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-60.qxd

2/18/04

2:35 PM

Page 487

Figure 4. Digger shield with hydraulically operated breasting plates on periphery of top heading of shield used to construct transit tunnel.

Figure 5. Cross-section of digger shield shown in Figure 4.

2. Maintaining the right plug in the invert was always a headache. 3. Mounting the digger in the center created a “Catch 22”: if the ground movement in the center became excessive, the only way to stop it was to cram the digger bucket into the face. However, that made it impossible to excavate and move the shield forward because to do so meant the bucket had to be moved, allowing the face to fail. Shields with open faced wheeled excavators were another, early step in mechanization of soft-ground machines that have some things in common with their cousins the hard rock TBMs. In 1953, as an outgrowth of efforts to develop mechanized mining equipment, a “circular tunnel borer” was developed for tunnels at Oahe Dam in the state of South Dakota in the USA. As the precursor of what was to come, that machine successfully excavated large-diameter (7.85 m) tunnels in weak rock (shale). This was the

Figure 6. Earth pressure balance tunnel boring machine (EPB) (Lovat).

first step in the evolution of Robbins hard rock tunnel tunneling equipment that dominated hard-rock TBM tunneling world-wide for many years (Robbins, 1970). Wheeled excavators were used with success in firm ground conditions, but not so well in running or fast raveling ground conditions. As a turning point in global tunneling equipment development, softground tunnel shields equipped with wheeled excavators were exported to Japan. Further development of soft-ground tunneling machines was flat in the USA for many years. Japan, however, took a good idea, invested heavily in equipment development and within a decade or so exported vastly improved tunneling methods back to the USA in the form of pressurizedface tunneling machines. Thus as tunneling in the USA was sticking with traditional shield tunneling, the Japanese, Europeans (read that Germans), the UK, and Canadians were developing more “modern” machines – the earth pressure balance machine (EPB) and the slurry face machine (SFM). See Figures 6–9. At first-hand, these machines are similar in that they both have: 1. A revolving cutter wheel. 2. An internal bulkhead that traps cut soil against the face (hence, they are called closed face) that maintains the combined effective soil and water pressure and thereby stabilizes the face. 3. No workers at the face but rely on mechanization and computerization to control all functions, except segment erection (to date). 4. Precast concrete segments erected in the shield tail, with the machine advanced by shoving off the segments. The actual functioning of the machines, however, has some distinct differences: in the EPB the pressure is transmitted to the face mechanically, through the

487 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-60.qxd

2/18/04

2:35 PM

Page 488

Figure 9. Simplified cross-section of slurry face tunnel boring machine (SFM) (from Herrenknecht). Figure 7. Simplified cross-section of earth pressure balance tunnel boring machine (EPB).

Figure 8. Slurry face tunnel boring machine (SFM). (courtesy of Herrenknecht).

soil grains, and is reduced by means of friction over the length of the screw conveyor. Control is obtained by matching the volume of soil displaced by forward motion of the shield with the volume of soil removed from the pressurized face by that screw conveyor and deposited (at ambient pressure) on the conveyor or muck car. Clearly the range of natural geologic conditions that will result in suitably plastic material to transfer the earth pressure to the face and, at the same time, suitably frictional to form the “sand plug” in the screw conveyor is rather limited – generally only combinations of fine sands and silts. In contrast, the SFM transmits pressure to the face hydraulically through a viscous fluid-formed by

material cut in the face and mixed with slurry (basically bentonite and water). In this case the pressure transmitted can be controlled by means of pressure gages and control valves in a piping system. By this system a much more precise and more consistent pressure control is attained. The undesirable aspect of this system is the separation plant that has to be built and operated at the surface to separate the slurry from the soil cuttings for disposal and permit re-use of the slurry. Finding a site for the slurry separation that is satisfactory for the process and acceptable to the public can present interesting challenges. During the last half-decade great strides have been made in developing new families of conditioning agents that can be used in both types of closed face machines. These additives tend to blur the distinctions portrayed above and widen the range of applicability of both types of machines. Indeed, we predict that within one decade we will not be talking about the two types of machines but rather a new family of machines that will operate interchangeably and with equal efficiency as an open face wheel machine in stable ground or as a closed face machine (with conditioners) that will cut any type of soft ground. Herrenknecht, for one, is already moving ahead with development of this new breed of machine. Throughout all of this development, the role of the miner at the tunnel face is steadily being diminished. With any closed face machine, the miner is not doing any excavating or breasting of the face. The miner is operating machines that, unfortunately, can not always do the job as advertised.

3 WHERE ARE WE TODAY WITH SOFTGROUND TUNNELING EQUIPMENT? 3.1

488 Copyright © 2004 Taylor & Francis Group plc, London, UK

Owner’s perspective

Owners only use completed tunnels. Only of necessity do Owners become involved in constructing them. An Owner wants to get a project completed and

09096-60.qxd

2/18/04

2:35 PM

Page 489

in use for the least reasonable cost, within their budget, and within their schedule, which is usually the sooner the better. Most Owners do not regularly need to have tunnels built, and thus will typically engage a suitably experienced engineer to make a design that meets the Owner’s requirements. The Owner may also engage professionals to manage the contract for tunnel construction. On major tunnel projects, the Owner typically becomes somewhat experienced in tunnels and may engage their own engineering workforce to design and manage tunnel construction. In multi-tunnel contracts for major wastewater or transit projects, the experience of one tunnel project is typically used to guide the next. This experience often comes from problem projects, mistakes, bad luck, and bad press. In euphemistic terms, this experience is called “lessons learned”. Thus for various reasons, an Owner may get more intimately involved over time in the tunneling specifications and contract terms. One major reason comes from an Owner trying to maintain a positive public image. The public is not always happy with their community being dug up, even with the huge mitigation that tunneling typically affords the community in comparison to all other methods of construction. During construction, the community also rarely appreciates what the new project will afford. Thus, keeping the public happy is a major motivator to the Owner. Major projects are rife with the public being unhappy. Their perception is of construction taking longer, the work costing more than they were originally told, damage to property (or the possibility of), and a fear of the unknown dangers of tunneling. Any of these factors on their own can drive an Owner to take strong defensive measures in tunneling. A powerful motivation to be conservative comes to an Owner trying to avoid the specter of sinkholes opening up in the street. In the clinical terms of management, the forgoing items are generally called “risks,” and, typically, the Owner does not want to accept these risks. 3.2

Engineer’s perspective

The engineer is a businessman with a professional license to practice engineering. The engineer typically does the “design” of a tunnel to meet the Owner’s varying requirements. The most manageable part of the design is often the functional and structural aspects of tunnels, such as tunnel linings and shafts. The engineer’s role gets less clear when construction methods are factored into a design. Perhaps the most difficult part of the engineer’s job is to develop contract terms that manage, mitigate, and otherwise deal with the “risk” discussed in the previous section.

The heavy construction industry and tunneling in particular have come a long way toward sensible contracts that have to deal with difficult project conditions, of which only some are be geologic (Essex, 1997). We feel that one place the Engineer is squeezed to walk a straight professional line is to balance the requirements of an appropriate and efficient design that accommodates the risks that the Owner wants taken care of, with a design that is still sensible and constructible. The Engineer has a professional responsibility for safety of the public – but how far is it reasonable for an Engineer to meddle in the means and methods of construction for the sake of “safety”? 3.3

A quintessential businessman, the Contractor has no fixed worksite and continuously moves from project to project, and Owner to Owner, building whatever that Owner wants. As a business, a Contractor wants to make a profit that represents an acceptable return on the substantial investment of equipment and human capital. The Contractor has a huge responsibility for worksite safety, not just for physical injury to workers and the public, but by our definition here, safety also includes responsibility for protecting the environment and nearby facilities To the Contractor, the TBM is a tool for driving the tunnel. With notable exception, the Contractor would prefer that the specifications not go into great detail on the characteristics and/operation of the machine. Rather, contractors would rely on the TBM manufacturer to provide a machine capable of meeting the job conditions. The Contractor then integrates the TBM into a complete plan of the “means, methods, and procedures” to drive the tunnel. A conflict arises when the Engineer believes that additional specifications must be added to cover the special geologic or other conditions as well as the special concerns of the Owner. This leads the Engineer to make major additions to the specifications. Unfortunately, this writing of tunnel contracts with substantial tunnel machine specifications often has led to a struggle of wills among the owner, engineer, construction manager, and Contractor as to who is in charge, and who knows best how to tunnel. 3.4

Tunneling equipment manufacturer’s perspective

Equipment manufactures are in the business of selling a very specialized and often quite costly product. Tunneling equipment is manufactured to meet the construction demands of a society that is asking for tunnels to be build in increasingly difficult and challenging conditions. Manufacturers accordingly

489 Copyright © 2004 Taylor & Francis Group plc, London, UK

Contractor’s perspective

09096-60.qxd

2/18/04

2:35 PM

Page 490

strive to make machines that permit tunneling to be less costly and ever safer. Tunneling equipment has a prime role in tunnel project profitability. Faster tunneling, less maintenance, and greater machine utilization are all key to profit for the contractor. As the equipment manufacturer’s prime clients, construction contractors want equipment that meets their expectations of production (speed), maintainability, reliability, and safety. A machine that does not work is quickly eliminated in the competitive tunnel equipment business. Successful manufacturers are in a continuous process of upgrading equipment or developing new equipment to meet their client’s demands. This has been part of the continuous evolution of mining equipment for over a century. Improved worker safety has been a strong factor influencing the tunneling equipment business. A poignant example was voiced by one of the best known tunnel equipment manufacturers, Richard Lovat. When he received the Golden Beaver Award (from the construction industry organization The Beavers) in January 2003 in Los Angeles, Richard Lovat spoke of his goal to build better and safer machines to construct tunnels. His motivation came from first-hand experience with the hazards of hand-mined soft-ground tunneling. He has been very successful in achieving his goal. Finally, manufacturers are continually striving to build machines that can tunnel in a wider range of conditions. Considering the Brunel shield as a starting point, it took decades to get real improvement in tunneling technology. Other technologies had to come into play. Compressed air made tunneling in soft clay or in sand below the ground water table feasible, but the tunnel shield did not change much from that of a tunnel in free-air conditions. Technological advances in hydraulic equipment, controls, instrumentation, and computerization were all required to permit manufacture of the modern-day closed-face machine. A manufacturer that could build a machine that would tunnel through about anything mother nature had to offer would rule the world of tunneling equipment. However, it is our perception that no one manufacturer will accomplish this. Rather, competition between manufacturers will continue (and possibly accelerate) the evolution of a horizontal tunnel factory. In this factory all operations would be essentially automatic and computer-controlled, but with an operator on the surface who can override or stop the operation in an emergency.

perspectives. We have selected those issues which we feel deserve attention in this forum. It could also be said that we are particularly peeved about these issues, and know that so are many others in the business, but with possibly widely different views on the same issue. In a later section we summarize our views on where the tunneling industry should change.

4.1

Over recent decades, we have seen construction specifications for tunnel projects literally explode by an order of magnitude in length. Some of this explosion is unavoidable due the increased complexity of tunneling and tunneling machines. However, we feel that as these specifications get longer, often they are not getting any better. Some reflection on past practices is felt to be valuable in understanding how this problem has evolved. In earlier times before ubiquitous computers and word processing software, the Washington, D.C. Metro (WMATA) developed a single bound volume of General Provisions and Standard Specifications for construction projects that totaled about 550 pages in length. Where applicable, the engineer (called section designer) wrote Special Provisions for each design project. These Special Provisions modified or amplified the General Provisions and/or Standard Specifications as necessary to meet the conditions and needs of the specific tunnel design section. For the most part these Special Provisions were a few pages to as short as one-half-page long and much of that was devoted to measurement and payment. Admittedly they were sometimes too simple, and many became quite dated and never used (such as the specification for cast iron tunnel segments). The bottom line is that tunnel projects were built with between 500 and 600 pages of specifications. Compare that with a recent transit project (admittedly a bit more involved) that had approximately twenty thousand pages of specifications. So what has happened? Overall, much of the explosion is due to the specifications writers’ propensity to write longer and longer specifications. However, the writers are often the tunnel engineers themselves. We have several observations on what has been going on as follows:

•

4 ISSUES Several issues confront the tunneling business. These issues involve an infinite variety of geology, tunnel equipment, and players with widely different

490 Copyright © 2004 Taylor & Francis Group plc, London, UK

Tunnel specifications in general

Inadequate experience Experienced tunnel people rarely want to write specifications. A good life-long tunnel superintendent may have the right answers to good specifications, but it is difficult to translate that experience into tunnel construction contract specifications. The exception is in design-build contracts when contractor and engineer truly work together as a team.

09096-60.qxd

• •

• •

•

•

2/18/04

2:35 PM

Page 491

No textbooks on tunnel specifications Writing a specification cannot be learned from a text book. Learning requires real design and construction experience. When in doubt, put it in The inexperienced or timid take this course in writing tunnel specifications. It also happens when specification review comments are not successfully resolved with experienced leadership. More is better Not necessarily so. Somewhat of a corollary of the above two items. Shopping lists Engineers try to make shopping lists in specifications that endeavor to accommodate every possible condition and nuance that might apply to current tunneling equipment and/or conditions. Inadequate coordination within whole contract Not enough attention and recognition is given to truly integrating all elements of a tunnel contract. Everything has to fit together, general and special conditions, plans and specifications. It is common for good work in one part of a contract to be undone by conflicting requirements in another part. No common staring point Tunnel specifications are either copies of previous projects, or more frequently, a merging of several specifications under the misconception of being more comprehensive and, therefore, better. Specification language is the result of past experience – but what worked last time, is not necessarily right for the next project. The effort must be made to verify the purpose and need for every word and line that is put into a specification.

4.2

Tunneling equipment, lost ground, and settlement

Soft-ground tunneling, particularly in urban areas, causes a real concern for the risk of damaging overlying and adjacent utilities and structures because of ground settlement. The enduring questions are: What is possible? What is reasonable to expect? And how does an Owner get a tunnel built without problems caused by settlement? In recent times, the answer has been to write a long specification that tacitly requires the contractor to procure the latest and best available tunneling machine. In a low-bid selection of the tunnel contractor, this may levelize bids by preventing a contractor from using what is perceived as inappropriate equipment. But it does not guarantee success. One unfortunate and unrealistic expectation has been that soft-ground tunneling with EPBs or SFMs could be done with minimal or no ground settlement.

This modern and evolving tunneling equipment has made it possible to tunnel in a wider range of ground conditions with less risk. In particular, the risk of a major loss of ground at the tunnel face is drastically reduced. However, this equipment certainly has not eliminated problems with settlement. As with the oldfashioned open-faced shields, workmanship is always key to good tunneling and hence to minimizing settlement. So it is with tunneling with EPBs or SFMs: good workmanship is required to minimize settlement regardless of the equipment. By workmanship, we mean that the miners and operators of the machine, as well as the successive levels of supervision, understand how tunneling is related to settlement and how it occurs. It is possible to tunnel with minimal settlement, but it is difficult to achieve when the priority is on tunnel progress alone and possibly working with inexperienced crews.

4.3

It is our observation that workers in Japan and in Europe take a personal pride in doing the work effectively and properly, a pride that is less evident in the USA. This is attributed to the United States tunneling business not being large enough to sustain over decades of time a tunneling workforce with depth. Examples below illustrate what we have seen elsewhere. At the Grauholz tunnel in Switzerland we have seen the use of a slurry machine in disparate ground conditions but with impressive results. We do not know of all problems they may have had but the end result, the completed tunnel, was compared by one observer to the “inside of a watch”. Similarly, segment manufacturers evidenced a distinct pride of workmanship. We note that one-pass linings have been used world-wide for many years. On the other hand, the USA has suffered with use of the prejudicial term “junk segments” that has come into tunneling jargon. We feel that if the initial lining is deprecated as being “junk,” that kind of thinking can lead to lower standards for all tunneling aspects. Only in recent projects in the USA have precast segmental tunnel linings started to be built with pride and quality standards fit for permanent works. In a highway tunnel in Japan in 1972 we saw our first EPB machine. Although the EPB was crude by today’s standards, operations at the time of the visit were flawless. The operator was not even down in the hole, but rather in a white suit in an air conditioned trailer at the surface. We strongly suggest that monetary incentives be established for the workers for a job well done. As a simple example, consider control of settlement near a

491 Copyright © 2004 Taylor & Francis Group plc, London, UK

Workmanship

09096-60.qxd

2/18/04

2:35 PM

Page 492

tunnel. On a major transit project, it was suggested that a significant fund be established for two purposes:

• •

Pay for repairs Any real or perceived building damage would be restored back to, or better than, original condition. Bonus to Contractor’s workers Any remainder (i.e. funds not needed for repair) would be given to the Contractor’s workers as a bonus.

This was not tried because the owner stated that they were hiring only qualified contractors and were not going to pay them twice. We believe that had the workers known they would receive bonuses of thousands or tens of thousands of dollars they would have been motivated to double and/or re-double their efforts to control or eliminate settlement. We are convinced that equitable implementation of such an incentive scheme to achieve quality construction goals, in this case minimal settlement, is possible. Contractors should act to meet the performance requirements of the contract. The engineer should not have to tell a contractor in a specification to “prevent lost ground” or “use good workmanship.” 4.4

Are EPBs and SFMs always better than open-face hand-excavated tunnel shields

We feel that the old-fashioned tunnel shield has undeservedly gotten a bad reputation and is not viewed as a viable alternative. Most major projects now require more sophisticated tunneling equipment. It is true that such new equipment have made tunneling better in many ways. But the old ways are not always wrong. The current solution to tunneling under a sensitive structure often is to immediately assume that at least an EPB is essential to control ground loss and eliminate the risk of an unacceptable ground loss at the heading. In some situations we think otherwise. EPBs and SFMs are made to achieve tunneling as rapidly and efficiently as possible; Contractor profit is very much linked to tunneling progress. The equipment is also made to minimize ground loss and settlement, particularly at the tunnel face. How much settlement occurs is affected by how skillful the miners are, what is called good workmanship. However, large tunnels require big machines, and it is not possible to have workmanship to make up for what may be unavoidable with the equipment. In dry ground conditions, we feel that tunneling under sensitive structures might be better done with a very short, hand-mined, open-faced tunnel shield. A short shield is easy to steer and a short ring of precast concrete tunnel lining would be used. Miners would work in small headings, hand excavating and if necessary, fully breasting the face in increments as the shield is advanced, much like Brunel did almost

200 years ago. For a short tunnel, this may make sense, and we believe that minimal settlement can be achieved for the limited conditions assumed. However, we are not suggesting that this be an alternative routinely to be considered for typical long tunnels. Our intent is to highlight the fundamental linkage among ground conditions, workmanship, and tunneling equipment and how successful tunneling does not of necessity mean the use of “modern” machines. 4.5

Tunnel construction contracts have evolved through a painful process of tunneling calamity, crisis, and claims. A turning point in tunnel contracts occurred in 1974 with publication of “Better Contracting for Underground Construction” (USNCTT, 1974). In turn this led to the two editions of the ASCE/UTRC-sponsored report on “Avoiding and Resolving Disputes in Underground Construction” (ASCE 1989, 1991), which in a sense formalized guidelines for implementing some of the recommendations of 1974. The most recent product of this evolution is the ASCE “Guideline for Geotechnical Baseline Reports” (Essex, 1997). If a geotechnical baseline could be defined as a set of truly independent conditions, the tunneling industry might have achieved salvation. However, the success of a tunnel project is not necessarily assured even if the geotechnical conditions are found to be just as baselined. Tunneling equipment and workmanship invariably play a role. The links among means, methods, tunneling equipment, workmanship, and geologic conditions cannot be ignored. Thus, instead of the baseline concept being just geologic, GBRs are mucked-up with design information and assumptions regarding tunneling equipment and construction means and methods. As with specifications, the lengths of GBRs have tended to increase dramatically with time, so that they typically are now several times longer than they were just a few years ago. It is time that we recognized that, just as their name states, they are baseline reports and not intended to be geotechnical data reports, geotechnical interpretative reports, or design summary reports. The latter types of report, or relatives thereof, may be written and included as contract documents when deemed appropriate, but should not confuse the intent and purpose of a geotechnical baseline. When we look at the content of typical GBR, we clearly see the results of a committee. (Note the writers also had substantial participation and input over the years). The intent and purpose is clearly noted, but as GBRs are being implemented, many are going astray. Some examples:

• •

492 Copyright © 2004 Taylor & Francis Group plc, London, UK

Geotechnical Baseline Reports (GBR)

Words like “concise” and “may” are being used – neither are appropriate Specification requirements are given.

09096-60.qxd

• •

2/18/04

2:35 PM

Page 493

Lists of do’s and don’ts are given. Assumptions on construction means and methods are given.

It is time that we make GBRs truly baseline documents. Specifications go in specifications; the lists and assumptions go into other documents, e.g., design summary reports. 4.6

Safety

The issue of safety is always a concern in tunnel projects. Traditionally, the Owner, Designer and Contractor have accepted the Engineers Joint Contract Documents Committee (EJCDC) Standard General Conditions of the Construction Contract which states: “ENGINEER will not supervise, direct, control or have authority over or be responsible for CONTRACTOR’s means, methods, techniques, sequences or procedures of construction, or the safety precautions and programs incident thereto, or for any failure of CONTRACTOR to comply with Laws and Regulations applicable to the furnishing or performance of the Work.” Similarly the EJCDC General Conditions state that the Contractor is to be “solely responsible for the means, methods, techniques, sequences and procedures of construction.” This delineation of responsibility and liability has generally been accepted for the last decade and a half for design-bid-build projects. However, we observe that an updated standard form of Agreement should be developed. This update should address at least four questions: 1. How are responsibility and liability addressed when the Designer sees something in the field that is perceived to be a safety concern? 2. How are responsibility and liability addressed for a design-build contract where the Designer is now a part of the Contractor’s team? 3. How are responsibility and liability addressed for the common situation of the Construction Manager reviewing, approving, or accepting required submittals by the contractor? And, does this answer change depending upon whether the Construction Manager is a third party or of the same firm as the Designer? 4. How are responsibility and liability addressed when the Construction Manager requires (or even suggests) changes in Contractor’s submittals?

would foster another major step in improvement of tunneling practices. 5.1

Every project will have its own, special needs and some customization of specifications is normal. But, all soft-ground tunneling projects have many similarities and should start from the same common point. The emphasis and effort of the designer should then focus on the project-specific items that make the project unique (or non-standard). As a starting point, we suggest writing a national tunneling “standard,” or “guideline.” We do not desire to quibble about what term is used, but want to emphasize the intent and purpose. Notably, this has been done in the in the UK (British Tunneling Society, 2000). As the writers of that document note, it took a great deal of effort to complete, and should not be considered definitive or absolute but nevertheless is a good place to start. It would seem reasonable to do something similar in the USA. 5.2

The following is a suite of changes that individually have merit, but when implemented as a whole, they

5.3

Geotechnical baseline

Return to the core contractual purpose of a “geotechnical baseline.” We need to implement this concept as a fully integrated element of the tunnel contract. We suggest the shorted terminology of Geotechnical Baseline (GB) to more correctly reflect the intent. This means getting rid of the “R,” and adjusting the content of the many supporting documents (GDR, GIR) in a tunnel design accordingly. Finally, make these baselines short. As a guideline, not a rule, we suggest the GB should be no more than ten pages consisting of succinct text, summaries, tables, and bullet items. Incentive achievement of quality goals such as limiting tunneling settlement

We feel strongly that good incentives (read that payments in dollars) for the workers will go farther toward meeting quality goal than will voluminous

493 Copyright © 2004 Taylor & Francis Group plc, London, UK

Stop writing tunnel equipment specifications

We feel that Owners and engineers should not prescribe the tunneling equipment. Tunneling is very specialized work and requires very specialized equipment. We feel that good, experienced contractors bidding on a tunnel contract are the best suited to select the equipment. In turn, Contractors have to do their part: accept the responsibility and work not only within the letter of a contract but also in good faith regarding the Owner and the public to construct tunnels successfully.

5.4

5 WHERE SHOULD THE TUNNELING INDUSTRY BE GOING

Tunnel specifications in general

09096-60.qxd

2/18/04

2:35 PM

Page 494

contract documents. Therefore, we challenge the AUA to sponsor a working group to develop and promote “Guidelines for Incentivizing Underground Projects.” In parallel with re-directed GBs, we are convinced that acceptance and implementation of strong incentives into underground contracts will be returned to the Owners many times over. 6 CONCLUSION Better, more cost-effective tunneling will require change. As a direction for the achievement of that goal, we have set out the foregoing thoughts. It is acknowledged that these recommendations and opinions are not all-inclusive. However, they represent “talking points” that we put forward in an effort to trigger discussion and reflection. We are trying to open the discourse towards finding ways to improve our total system of delivering tunnel projects to our clients (most often the taxpayers who pay for, and use, the tunnels we build). ACKNOWLEDGEMENT We have worked on many transit, highway, water, and wastewater tunnel projects, both in the United States and abroad. The opinions stated herein reflect an amalgam of that experience, and not one specific

project. We are grateful for having had the opportunity to work over the many years with the leaders in the tunneling business that were teachers and mentors.

REFERENCES ASCE. 1989. Avoiding and Resolving Disputes in Underground Construction, Technical Committee on Contracting Practices, Underground Technology Research Council, 24 p  appendices. ASCE. 1991. Avoiding and Resolving Disputes in Underground Construction, Technical Committee on Contracting Practices, Underground Technology Research Council, 2nd ed. British Tunnelling Society (2000). Specification for tunnelling, Thomas Telford Ltd., 144 p. Copperthwaite, W.C. 1906. Tunnel Shields and the Use of Compressed Air in Subaqueous Works, Van Nostrand, 389p. Essex, R.J, ed. 1997. “Geotechnical Baseline Reports for Underground Construction,” the Technical Committee on Geotechnical Reports of the Underground Technology Research Council, ASCE, 40 p. Robbins, R.J. 1970. The “Robbins” Moles – Status and Future, Rapid Excavation – Problems and Progress, Proceedings of the Tunnel and Shaft Conference, Minneapolis, May 1968, D.H. Yardley, ed., SME/ AIMPE, pp. 272–295. US National Committee on Tunneling Technology (USNCTT). 1974. “Better Contracting for Underground Construction,” National Academy of Sciences.

494 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-61.qxd

2/18/04

5:31 PM

Page 495

Session 4, Track 2 Transit oriented development – making the case for going underground

Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-61.qxd

2/18/04

5:31 PM

Page 497

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Atlanta West Area Combined Sewer Overflow Storage Tunnel and Pumping Station R.C. Divito Hatch Mott MacDonald, Cleveland, Ohio, USA

W. Klecan Jordan Jones & Goulding, Atlanta, Georgia, USA

G.D. Barnes City of Atlanta Department of Watershed Management, Atlanta, Georgia, USA

ABSTRACT: The West Area Combined Sewer Overflow (CSO) Storage Tunnel is one of the largest and most ambitious CSO Tunnel Projects underway in the US today and when completed in late 2007 it will be the largest tunnel system in the southeast region. This project is designed to meet water quality standards by capturing overflows occurring annually at three of the City’s CSO control facilities that would be diverted into an 8.3 mile long, 24-foot diameter rock tunnel, capable of storing 150 million gallons. A pumping station will be constructed at the end of the tunnel to lift the stored flow to a new CSO treatment plant prior to discharge. This paper presents a general technical description of the project and geotechnical investigations. Construction is expected to start mid-2004.

1 INTRODUCTION 1.1

Project overview

The project site is located within the city limits of Atlanta, Georgia, as shown in Figure 1. The City of Atlanta is under an EPA Consent Decree to bring its CSO system into compliance with federal and state water quality regulations by November 2007. The City submitted a plan for meeting these requirements to the US Environmental Protection Agency (EPA) and Georgia Environmental Protection Division (EPD) and received authorization in July 2001 to proceed with its plan. This project is a major component of the City’s plan to control combined sewer overflows (CSO) and facilitate compliance with the Clean Water Act. Specifically, CSO’s within the west side watersheds will be controlled by the proposed facilities. This project will result in the protection, preservation, and revitalization of water resources within the City of Atlanta and neighboring communities. The Atlanta Wastewater System Improvement Program Management Team performed preliminary design of the proposed project. Final design of the proposed project facilities was performed by the JDH Joint Venture under the direction of the City of

Atlanta Department of Watershed Management. The JDH Joint Venture is comprised of Jordan Jones & Goulding, Hatch Mott MacDonald, and Delon Hampton & Associates. 1.2

The project facilities were designed to divert, store, convey, and pump CSO’s for treatment prior to discharge to the Chattahoochee River to meet water quality standards. The facilities were designed to provide a capture and conveyance capacity of 177 million gallons (MG), a storage capacity of 150 MG, and a pumping capacity of 85 million gallons per day (MGD) expandable to 100 MGD. In addition, the project facilities were also designed to provide emergency overflow relief at the northernmost downstream end of the tunnel in the event of overfilling of the tunnel. The overall project facilities consist of 8.3 miles of partially-lined 24-ft finished diameter CSO storage tunnel, three tunnel junctions (plus two adit junctions for future tunnel connections), four 40-ft finished diameter construction shafts, three CSO flow drop and intake structures (including diversion structures, drop shafts, and de-aeration chamber facilities), four

497 Copyright © 2004 Taylor & Francis Group plc, London, UK

Project facilities

09096-61.qxd

2/18/04

5:31 PM

Page 498

Figure 1. Project facilities plan.

vent shafts, one 66-ft finished diameter pumping station and shaft, and one 24-ft finished diameter overflow tunnel and shaft. These facilities are arranged in plan and profile as shown in Figure 1 and Figure 2, respectively. 2 PROJECT FACILITIES DESIGN Project facilities are arranged as shown in Figure 1. The tunnel has been subdivided into two discrete alignments, designated as the Clear Creek Tunnel (Contract A) and the North Avenue Tunnel (Contract B). The Clear Creek Tunnel consists of a 3.9-mile, partially-lined 24-ft finished diameter storage tunnel, ranging from 140 to 309 ft in depth below ground surface, as shown in Figure 2, and is designed to store and convey CSO from the existing Clear Creek and Tanyard CSO Facilities. It includes the Rockdale Construction Shaft, Clear Creek Construction Shaft, Clear Creek CSO Intake (easternmost upstream intake), Tanyard CSO Intake, and Tanyard Tunnel and Junction. All deep underground tunnel facilities and

structures are located on the tunnel centerline alignment. The 307-ft deep Rockdale Construction Shaft is the westernmost downstream shaft designed with a 40-ft finished diameter and will provide for tunnel boring machine (TBM) launching, tunnel muck removal, and construction and post-construction access. The 153-ft deep Clear Creek Construction Shaft is the easternmost upstream shaft designed with a 40-ft finished diameter and will provide for tunnel boring machine (TBM) removal, access and includes the hydraulic flow intake drop shaft. The Clear Creek CSO Intake and Tanyard CSO Intake consist of intake channel diversion structures and vortex-type tangential flow drop structures, as shown in Figures 3 and 4. The Clear Creek CSO Intake structures further consist of a 15.5-ft finished diameter drop shaft (designed as an integral structure with the Clear Creek Construction Shaft permanent lining), 27.3-ft finished diameter de-aeration chamber, 7.4-ft finished diameter chamber vent shaft, and a 10ft finished diameter tunnel vent shaft, as shown in Figure 4. The Tanyard CSO Intake structures further consist of a 12.8-ft finished diameter, 152-ft deep

498 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-61.qxd

2/18/04

5:31 PM

Page 499

Figure 2. Project facilities profiles.

Figure 3. Tanyard CSO intake structure plan.

drop shaft, 22.7-ft finished diameter de-aeration chamber, and a 6.2-ft finished diameter chamber vent shaft. The Tanyard Tunnel consists of an 800-ft, 11.3-ft finished diameter conveyance tunnel designed to connect the Tanyard CSO Intake with the Clear Creek Tunnel. The construction site for the Tanyard CSO Intake structures is limited to one acre because of residential buildings to the west and the I-75 highway to the east, as shown in Figure 3. The North Avenue Tunnel consists of a 4.4-mile, partially-lined 24-ft finished diameter storage tunnel, ranging from 129 to 314 ft in depth below ground surface, as shown in Figure 2 and is designed to store and convey CSO from the existing North Avenue CSO Facility and Clear Creek Tunnel. It includes the R.M. Clayton Construction Shaft, North Avenue Construction Shaft, North Avenue CSO Intake (southernmost upstream intake), Rockdale Connecting Tunnel and Junction, Overflow Tunnel and Junction, R.M. Clayton Overflow Shaft, and R.M. Clayton Pumping Station. The R.M. Clayton designated structures are all located at the R.M. Clayton Water Reclamation Center (WRC). All deep underground tunnel facilities and structures are located on the tunnel centerline alignment. The R.M. Clayton Construction Shaft is the northernmost downstream construction shaft designed

499 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-61.qxd

2/18/04

5:31 PM

Page 500

Figure 4. Clear Creek CSO intake structures.

with a 40-ft finished diameter and a depth of 195 ft. This shaft will provide for tunnel boring machine (TBM) launching, tunnel muck removal, access, and includes the hydraulic flow intake drop shaft. The North Avenue Construction Shaft, designed with a 40-ft finished diameter and a depth of 214 ft, will provide for tunnel boring machine (TBM) removal and construction and post-construction access. The North Avenue CSO Intake consists of an intake channel diversion structure and vortex-type tangential flow drop structure. The North Avenue CSO Intake structure further consists of a 10.7-ft finished diameter drop shaft (designed as an integral structure with the North Avenue Construction Shaft permanent lining), 24-ft finished diameter de-aeration chamber, and a 10-ft finished diameter chamber vent shaft. The Rockdale Connecting Tunnel consists of a 100-ft, 14-ft finished diameter conveyance tunnel designed to connect the Clear Creek Tunnel with the North Avenue Tunnel. The Overflow Tunnel consists of an 800-ft, 24-ft finished diameter conveyance tunnel designed to allow overflow relief for the West Area CSO Tunnel system in the event of a flow intake control gate failure. The Overflow Shaft is designed with a 24-ft finished diameter and a depth of 129 ft. The Pumping Station

is designed with a 66-ft finished diameter and a depth of 183 ft.

3 OPERATIONAL REQUIREMENTS AND HYDRAULIC DESIGN 3.1

The West Area CSO Storage Tunnel and Pumping Station has been designed according to operational requirements intended to meet water quality standards. To meet those requirements, the operational requirements of the project facilities include a capture and conveyance capacity of up to 177 MG with storage availability of 150 MG, complete tunnel dewatering within two days of the filling event, air release during tunnel filling, and provisions for operational maintenance and inspection. Based upon historical records, it is anticipated that overflows in excess of the tunnel storage capacity will be limited to an average of four per year. Under these requirements the project facilities were designed for peak inflow rates of 5,071 cubic feet per second (cfs) at the Clear Creek CSO Intake, 3,171 cfs at the Tanyard CSO Intake, and 2,017 cfs at

500 Copyright © 2004 Taylor & Francis Group plc, London, UK

Operational requirements

09096-61.qxd

2/18/04

5:31 PM

Page 501

the North Avenue CSO Intake for a total peak inflow of 10,259 cfs (ie., an estimated average equivalent peak flowrate of a 25-year, 12-hour storm event). 3.2

Hydraulic modeling

During final design, hydraulic modeling studies of the operational performance of the project facilities were performed by the Iowa Institute of Hydraulic Research (IIHR). These studies consisted of numerical modeling for the design of the intake structures, as well as overall system performance during filling of the tunnel. Numerical modeling was performed based on the specified construction shaft and storage tunnel diameters, shaft and junction locations, shaft overflow elevations, previously-described for the project facility configurations, and previously-described operational requirements. The hydraulic modeling studies included evaluations of transient pressures and surges, confirmation of system capacity, air release rates at shaft locations, overflow estimation at intake locations, required CSO Intake Structure configurations (diversion structure, vortex drop shaft and inlet, de-aeration chamber and vent shaft), flow velocity and internal pressures, and friction loss coefficient. Results from the hydraulic modeling identified a significant water level rise or surge height up into the shafts of approximately 20 ft due to transient wave pressure coupling effects. Coupling of transient waves between nearby shafts was found to increase water level surge heights within the shafts. To mitigate this effect, permanent shafts in close proximity near junctions were prohibited (ie., two permanent shafts could not be located in the vicinity of the Rockdale Junction). A downstream emergency overflow tunnel and shaft, both with a 24-ft finished diameter, were found to be required to mitigate transient wave surge effects and avoid flooding at the R.M. Clayton WRC. Due to rapid tunnel filling rates (ie., complete filling within three hours of the start of the design storm event), it was determined that flow instrumentation would be required at the intakes, within the tunnel, and at the pumping station. Sluice gates at each intake will control the rate of filling of the tunnel. 4 GEOLOGIC SETTING 4.1

schist, amphibolite, and quartzite, and by both large and small igneous intrusions. These crystalline rocks have a widely varying content of the minerals quartz, feldspar, amphibole, mica, and accessory minerals such as garnet, epidote, and sillimanite. Many geologic terms are used to describe these rocks in a scientifically-precise manner, but from an engineering perspective, these rocks are similar. Exception for the geotechnical features noted in the following discussions, the crystalline rocks are strong, hard, and dense. The crystalline rocks of the Winder Slope District were apparently sedimentary or volcanic rocks deposited between 300 and 1,100 million years ago. These original deposits were intruded, folded, and metamorphosed during several periods of crustal movement. These rocks, which underlie the Atlanta area, are in the form of a doubly-plunging syncline (i.e., a structural basin with an outcrop pattern resembling an eye). The syncline is elongated in the northeast direction and is about 56 miles long and 25 miles wide. The exposed rocks are also layered and folded on a scale that is observable in a road cut and in a hand specimen. This thick and repetitive sequence of rocks has been subdivided into a number of formations to facilitate geologic mapping. The northern part of the North Avenue Tunnel will be located within the Brevard Zone which is a major regional zone of deformation that extends from Alabama to Virginia. Major earthquake activity associated with the Brevard Zone ceased about 200 million years ago and there are no active faults with measurable displacement in Georgia. The Brevard Zone is 1 to 3 miles wide and contains crystalline rock that was once sheared by fault movement, but is now hardened and well-healed. In the Atlanta area, the Brevard Zone is a topographic lineament that for some distance controls the Chattahoochee River. 4.2

Structural geology

Based on the results of structural geologic mapping and rock core discontinuity orientation, the structural geology can be defined for engineering purposes. In general, the structural geologic conditions are widely variable with high angle joints predominantly oriented northeast-southwest and northwestsoutheast and sub-horizontal foliation dipping toward the southeast.

Regional geology

The project is located within the Piedmont Physiographic Province, which is underlain by metamorphic and igneous rocks. The project lies mainly in the rolling hills of the Winder Slope District and extends to the eastern bank of the Chattahoochee River and into the Brevard Zone. The Winder Slope District is underlain by “crystalline rock,” including gneiss,

5 GEOTECHNICAL INVESTIGATION AND FINDINGS 5.1

Geotechnical investigation techniques for the project consisted of conventional subsurface exploration,

501 Copyright © 2004 Taylor & Francis Group plc, London, UK

Investigation techniques

09096-61.qxd

2/18/04

5:31 PM

Page 502

soil and rock laboratory testing methods, and geological desk studies. Subsurface exploration methods included standard penetration test (SPT) soil borings, rock core borings, water pressure (packer) testing in rock, pump testing in rock, borehole geophysical discontinuity logging, manual rock core discontinuity logging and orientation analysis, field structural geologic mapping of surface bedrock exposures, and geologic lineament analyses. Geologic lineaments are approximately linear topographic features that reveal locations of dominant geologic features, such as faults, shear zones, softer rock layers and other anomalous subsurface geologic structures or characteristics. Standard penetration test soil borings were performed at construction shaft locations, drop shaft locations, diversion structure locations, the R.M. Clayton Pumping Station, and at selected locations along the tunnel alignments to characterize overburden soil stratigraphy and engineering properties. Rock core borings were performed at 101 locations along the tunnel alignments to characterize bedrock stratigraphy and engineering properties. The rock core borings were oriented vertically with the exception of eight borings that ranged 30 to 45 degrees in inclination from vertical. Inclined borings were performed at geologic lineament locations to investigate associated characteristics of potentially dominant geologic features, such as faults and shear zones. Based on the number of rock core borings performed, the average rock core boring spacing was 1,000 ft. Rock core boring samples were logged for total recovery, rock quality designation (RQD), joint count (fracture frequency), and weathering index in accordance with International Society for Rock Mechanics (ISRM) recommendations. Water pressure or packer testing in rock was performed in most of the completed rock core boring boreholes. This testing was performed to evaluate the effective permeability of the rock mass near and along the boreholes and in the vicinity of the tunnel. Double packer equipment with 20-ft testing intervals was used for water pressure testing. Pump testing in rock was performed at locations where water pressure testing results yielded high effective rock mass permeabilities (typically greater than 103 cm/sec). This testing consisted of moderate to high flowrate groundwater extraction pumping of wells installed in rock with simultaneous monitoring of groundwater table drawdown in nearby observation wells and piezometers. The results of pump testing provide insight into the water storage or recharge capacity of the rock mass relative to groundwater inflow estimates for tunnel construction. Borehole geophysical logging was performed using borehole groundwater temperature logging and acoustic reflectivity techniques to detect groundwater

flow and rock mass discontinuity location and orientation data. Rock core discontinuity orientation was performed using manual joint and foliation angle measurement methods combined with borehole geophysical discontinuity orientation data. Rock core discontinuity orientation was performed for all rock core borings. Field structural geologic mapping consisted of the identification of surface bedrock exposures or outcrops and measurement of the strike and dip of observed discontinuities, including joints and foliation. Geologic lineament analyses were performed using historical topographic mapping data and aerial photography combined with rock core boring and testing data to assess the characteristics of dominant geologic features, such as faults, shear zones, and other anomalous subsurface geologic structures. Such characteristics include feature width or thickness, orientation, degree of disturbance, and infilling material type and consistency. Geotechnical laboratory testing consisted of soil and rock testing for the determination of soil index properties and compressibility and rock strength, anisotropy, deformation, abrasivity, boreability, and mineralogical characteristics. Rock laboratory testing was performed by the Colorado School of Mines Earth Mechanics Institute of Golden, Colorado, and the SINTEF Civil and Environmental Engineering Soil and Rock Mechanics Laboratory of Trondheim, Norway. Rock laboratory testing was performed to determine specific rock engineering properties, including uniaxial compressive strength, Brazilian tensile strength, point load index, unit weight, elastic modulus, Poisson’s ratio, Cerchar Abrasivity, Punch Penetration, Drillability Indices, and mineralogical content.

5.2

The project geologic stratigraphy generally consists of fill material and naturally-derived overburden soil, up to 50-ft thick in places, underlain by transition zone material ranging from 10 to 20-ft thick over bedrock. The fill generally consists of anthropogenic materials, including miscellaneous debris and natural soils. The overburden soils consist of fine- to coarse-grained alluvial soils and residual soils. The transition zone materials consist of residual soil to partially weathered rock. Bedrock consists of very strong and hard mylonite, biotite gneiss, schistose gneiss, granitic gneiss, and migmatitic gneiss. Groundwater was encountered at widely ranging depths below ground surface, including depths above and below the top of bedrock.

502 Copyright © 2004 Taylor & Francis Group plc, London, UK

Soil and rock stratigraphy and groundwater

09096-61.qxd

2/18/04

5:31 PM

Page 503

6.2

6 ENGINEERING DESIGN AND CONSTRUCTION

Tunnels, chambers, and junctions

6.1.1 Design As described previously, shaft structures for this project consist of air vent shafts, hydraulic flow drop shafts, construction and post-construction access shafts, an overflow shaft, and the pumping station shaft. Vent shaft finished diameters range from 6.2 to 10.0 ft. Drop shaft finished diameters range from 10.7 to 15.5 ft with one 12.8-ft finished diameter stand-alone shaft (ie., not constructed integral with a construction shaft). The construction and post-construction access shaft diameter is 40.0 ft. The Pumping Station Shaft and Overflow Shaft finished diameters are 66.0 ft and 24.0 ft, respectively. Diversion structures, including approach channels and inlets, range from 10.7 to 25.0-ft finished width and 30 to 80-ft depth below ground surface. Shaft final lining and diversion structures are designed to be constructed of conventional castin-place reinforced concrete.

6.2.1 Design The Clear Creek Tunnel and North Avenue Tunnel, including the R.M. Clayton Connecting Tunnel and Overflow Tunnel, are designed for 24.0-ft finished diameter final cast-in-place concrete (plain or reinforced) linings, with the exception of areas of the North Avenue and Clear Creek Tunnels where massive, competent, intact bedrock is encountered and does not require final lining for structural support and/or infiltration and exfiltration control. The Tanyard Tunnel is designed for a 11.3-ft finished diameter final cast-in-place concrete (plain or reinforced) lining. The Clear Creek, Tanyard, and North Avenue Deaeration Chambers are designed for final cast-inplace reinforced concrete linings with 27.3-ft, 22.7-ft, and 24.0-ft finished diameters, respectively. The R.M. Clayton (24 ft to 24 ft), Rockdale (14 ft to 24 ft), Tanyard (11.3 ft to 24 ft), North Avenue Adit (12 ft to 24 ft), and Clear Creek Adit (12 ft to 27.3 ft) junctions are designed for final cast-in-place reinforced concrete linings.

6.1.2 Construction Shaft and diversion structure excavations in soil are anticipated to be performed by means of conventional heavy mechanical hydraulic excavating equipment. Initial support of shaft and diversion structure excavations in soils can be supported using concrete secant pile walls, concrete diaphragm walls, steel sheet pile walls, steel soldier pile and lagging walls, steel liner plate and ribs (for circular shafts only), steel ribs and timber lagging (for circular shafts only), and steel casing (for small circular shafts only) to be selected by the constructor, except where site-specific subsurface conditions restrict the use of some support systems. These initial support systems must be used with internal and external bracing and support for stability. External dewatering as needed to control groundwater for underpinning-type support methods. Vent shaft and drop shaft excavations in rock are anticipated to be performed using conventional drilland-blast techniques, reverse circulation shaft drilling, raise boring, or a combination of these methods. Construction and post-construction access shaft, pumping station shaft, overflow shaft, and diversion structure excavations in rock are anticipated to be performed using drill-and-blast techniques. Initial support of shaft and diversion structure excavations in rock include steel dowels with welded wire mesh-reinforced shotcrete. Shafts without in-take diversion structure penetrations can also be initially supported with the final cast-in-place concrete lining installed concurrently with shaft excavation advance.

6.2.2 Construction The North Avenue and Clear Creek Tunnels, with the exception of the starter and connecting tunnels and also the Overflow Tunnel and Tanyard Tunnel, will be excavated by means of a tunnel boring machine (TBM). All remaining structures will be excavated by means of drill-and-blast techniques. In addition, the North Avenue Tunnel De-aeration Chamber and Clear Creek De-aeration Chamber may be excavated by means of drill-and-blast techniques to expedite construction and to facilitate enlargement of the Clear Creek De-aeration Chamber. Initial support for tunnel and chamber excavations in rock include three types corresponding to anticipated ground conditions. These types include: 1) four rock bolt patterned rock reinforcement with supplemental steel channels and mesh and spot bolts, 2) six rock bolt patterned rock reinforcement with supplemental steel channels and mesh and spot bolts, and 3) full circle expanded steel ribs with steel mat lagging. Initial support for junction excavations in rock include additional patterned rock bolt reinforcement. Rock bolts will consist of steel bar double corrosion-protected mechanical bolts with expansion shell anchors and polyethylene sheath. All rock bolts will be cement grouted in-place, after installation and tensioning, to provide double corrosion protection. Double corrosion protection for rock bolts is required due to partial lining requirements and for the rock bolts to provide long-term support of unlined tunnel. A typical tunnel concrete lined cross section is shown in Figure 5.

6.1

Shafts and diversion structures

503 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-61.qxd

2/18/04

5:31 PM

Page 504

contributions. The authors also wish to thank Larry Williamson and Michael Vitale of Hatch Mott MacDonald, and Refik Elibay of Jordan Jones & Goulding for their contributions to the development of this paper.

REFERENCES Atlanta Wastewater System Improvement Program Management Team 2002a. City of Atlanta Remedial Measures Plan CSO Pre-Design Report for the Consolidated Storage Tunnel Component. Atlanta, Georgia. Atlanta Wastewater System Improvement Program Management Team 2002b. City of Atlanta Remedial Measures Plan Geotechnical Report for the Consolidated Storage Tunnel Component. Atlanta, Georgia. JDH Joint Venture 2003. Geotechnical Data Report for the West Area CSO Storage Tunnel and Pumping Station. Atlanta, Georgia.

Figure 5. Typical concrete tunnel lining cross section.

ACKNOWLEDGEMENTS The authors wish to thank their joint venture partner, Delon Hampton & Associates for their project design

504 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-62.qxd

2/18/04

2:36 PM

Page 505

Session 4, Track 3 Analysis and design

Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-62.qxd

2/18/04

2:36 PM

Page 507

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Consideration on machine data and load in TBM excavation for tunnel support selection N. Isago & H. Mashimo Incorporated Administrative Agency, Public Works Research Institute, Ibaraki, Japan

W. Akagi & H. Shiroma Japan Highway Public Cooperation, Tokyo, Japan

ABSTRACT: Application of Tunnel Boring Machine (TBM) for constructing road tunnel is considered as the rational and economical tunneling method because rapid excavation and reduction of support lead to cost reduction. To establish the method of ground estimation and selection of proper support in TBM excavation is desired. The relation between support pattern and tendency of machine data variances excavated by TBM were examined, and load acting on support was calculated and compared with rock quality and support pattern. That some of machine data had relation with support pattern and that load acting on support was decided by the characteristic of ground when the ground condition was poor while it was not when the ground condition was good were found in this study.

1 INTRODUCTION Rapid tunnel construction by mechanical excavation method is thought to be effective to reduce construction cost for long road tunnel because it can restrain the looseness of ground without dividing section and leads to reduction of tunnel support. Actual achievement up to now using Tunnel Boring Machine for road tunnel excavation in Japan was limited to that of evacuation tunnels or pilot tunnels which had small section area. In addition, the reason why TBM was adopted less was restriction of weak and changeable geological condition. Pre-construction phase investigation method was examined all over the world (Nilsen (1999)), however, construction phase investigation and observation method like ground estimation should be established especially in as handy a method as possible for weak ground condition including. Furthermore, some support patterns was shown for TBM tunnel excavation (Scorali (1995)), however, more proper support design method on the basis of the grasping load value acting on support by TBM excavation should be proposed to widen the application of tunneling method with TBM not limited to very hard rock tunnel and disseminate as the normal method in tunneling. There were many machine data acquired in construction phase. Comparison and examination of relation between the tendency of its data variance and

grade of geological condition with the data from the pilot or evacuation tunnels by using TBM excavation in Japan were carried out and the parameter suitable for the evaluation of ground condition was selected in this study. Furthermore, numerical analysis using frame model on the basis of the results of in-situ strain measurement data from tunnel support to calculate the load acting on support was carried out. The tendency of load value in terms of ground condition and rock quality was also discussed.

2 RESEARCH METHODOLOGY 2.1

Tendency of machine data variance comparable acquired in each tunnel was examined in terms of support pattern, which was relating to the grade of ground conditions for principal parameter acquired from the machine data in excavation. Table 1 shows the tunnels examined in this study concerning their diameters and kinds of main rock material. The tunnels used for the pilot tunnels or evacuation tunnels of road tunnels were constructed by TBM. The number of tunnel was fifteen (named from Tunnel A to O) and their diameters ranged from

507 Copyright © 2004 Taylor & Francis Group plc, London, UK

Tunnel specification and analysis method concerning TBM machine data

09096-62.qxd

2/18/04

2:36 PM

Page 508

3.5 to 5.0 meter. Tunnel lengths were approximately from 1.7 to 4.2 km. Table 2 shows the definition of the classification of rock quality on the basis of rock kind using in design stage in Japan. This table means that the quality of rock as relative property was only presented, not as the absolute property such as uniaxial compressive strength and fracture status in rock. Table 3 shows the machine data of TBM used in this analysis. Various machine data were selected to examine the correlation between machine data which was thought to be useful for support design and support pattern, that is, the characteristics of ground. Real machine data were much volume because of being

Table 1. Tunnel diameter and kinds of main rock in this study. Tunnel Diameter (m) Kinds of main rock material A B C

3.5 5.0 5.0

D

5.0

E

5.0

F

5.0

G

5.0

H

5.0

I

5.0

J K L M

5.0 5.0 4.5 4.5

N O

4.5 4.5

Tuff, Basalt, Rhyolite Andesite, Conglomerate Andesite, Tuff breccia, Sandstone, Gravel, Conglomerate Sandstone, Mudstone, Conglomerate, Alteration strata of sandstone and mudstone Sandstone, Alteration strata of sandstone and mudstone Sandstone, Alteration strata of sandstone and mudstone Tuff, Limestone, Chert, Clay stone, Gabbro, Dolerite Tuff, Chert, Clay stone, Gabbro, Dolerite Hornfels, Granite, Alteration strata of sandstone and mudstone Granite Granite Andesite, Tuff, Rhyolite Andesite, Tuff breccia, Tuff, Conglomerate, Alteration strata of sandstone and silt Aplite, Granite porphyry, Tuff Granite porphyry, Rhyolite

Table 2. Definition of classification of rock quality. Rock quality

Kind of rock

Hard rock

Chert, Hornfels, Granite, Limestone, Gabbro, Dolerite, Aplite, Granite and porphyry Basalt, Rhyolite, Andesite, Sandstone, Conglomerate and Clay stone Tuff, Tuff breccia and Mudstone

Medium-hard rock

Soft rock

acquired in the same little time step while continuing excavation. The machine data was calculated within one construction cycle at first. Then average machine data was acquired using all of one-cycle machine data as long as support pattern was same. The value as the representative machine data of a certain support pattern in a tunnel was defined and was examined in this analysis. Table 4 shows the main specification of tunnel boring machine in each tunnel. Both machine types, that is, shield and open type were used. Most of support patterns in each tunnel site had been separately regulated. Classification of support type and contents were summarized in Table 5 and used in the analysis of machine data. Support pattern B was used in good ground condition, while pattern L was in bad one. Distance of steel-arched support, which averagely ranged from 0.75 m to 1.50 m, and influence of the material for reinforcement such as temporary wood panel for preventing falling debris from roof and invert segment used with jack advance in bad ground condition were neglected. Table 3. Machine data for analysis in this study. Machine data

Net advance rate Unit average speed of thrust jack Cutter rotation speed Rotation number of cutterhead (rpm) Thrust force Pressure of thrust jack multiplied by jack area and number Cutter torque Resistance to cutterhead Advance energy (Thrust force/Excavation area) Rotation energy ((Cutter torque Cutter rotation speed)/(Net stroke speed Excavation area)) Unit volume Sum of the energy by excavation energy advance and rotation divided by excavation volume (Advance energy + 2 Rotation energy) Table 4. Machine specification. Full Maximum Length Weight power torque Tunnel Machine type (m) (ton) (kW) (ton m) A B C D E, F G, H I J, K L, M N O

508 Copyright © 2004 Taylor & Francis Group plc, London, UK

Content

Full shield Full shield Full shield Full shield Full shield Open Open Open Open Revised Open Full shield

9.20 9.75 11.1 11.2 11.2 13.9 13.2 15.4 12.3 13.0 8.85

170 358 390 380 300 200 213 300 192 196 310

750 1618 1300 1100 1100 1160 1020 1400 1240 1110 1618

101 258 300 250 250 110 130 257 115 155 200

09096-62.qxd

2.2

2/18/04

2:36 PM

Page 509

this study as the following equation:

Analysis method of load acting on support

Strain gages were attached on steel arched support in some sections from tunnel A to O. Measurement results of strain change were arranged and the stresses at the edge of steel arched support were calculated. Frame model which simulated the support was adopted and the stress occurred at the edge of the beam element was calculated by the model in analysis. Figure 1 shows the outline of modeling. Constant distribution load Pv in vertical and the same value load from bottom were set. Furthermore horizontal load Ph considering load increment D (: unit volume, D: tunnel diameter) in proportion to depth and K ratio was also set. That is, horizontal load Ph at tunnel roof was equivalent to K Pv and the load at tunnel bottom was K (D  Pv). The comparison of stresses from in-situ measurement and numerical analysis was done. The difference of stress between result from in-situ measurement and numerical result at measuring point (the number varied from tunnel) was multiplied by itself and the summation of all the difference at each section was calculated. That is, the error function was defined in

Table 5. Classification of support pattern in this study. Support pattern

Content

B CI CII

No support 2 cm shotcrete 2 cm shotcrete, Steel arched support (H-Beam:100 mm) 3 cm shotcrete, Steel arched support (H-Beam:100 mm) 3 cm shotcrete, Steel arched support (H-Beam:125 mm) Simple liner (not much stiffness) Steel segment and invert

DI DII L-1 L

(1) where  error function; *i  stress at the edge of support by in-situ measurement; i  stress at the edge of support by numerical analysis; and n  number of measuring points (varies in each tunnel, from 3 to 7). Equation 1 was thought to be the function of vertical load Pv and K ratio. Arbitrary K ratio was given and Pv was calculated when would be minimum. Then various K ratios were given and Pv was also calculated at the same process. Finally K ratio and Pv were decided when would be minimum among each conditions of K ratio. Pv was thought to be equivalent to the pressure acting on support and loads were evaluated by Pv/D. K ratio was set in every 0.1 step from 0 to 1.5 and Pv/D was calculated in every 0.03 when Pv/D was under 0.1 and in every 0.15 when Pv/D was over 0.1. No support weight was considered in numerical analysis because measurement of strain started after steel arched support set. Table 6 shows the kind of rock and support pattern used in analysis that strains were measured in these sections. Main composition of tunnel support was shotcrete and steel arched support, or simple liner or steel segment and invert as shown in Table 5. Support element in analysis was supposed as beam element and stiffness and area of shotcrete was neglected when using the combination of steel arched support and shotcrete like support pattern CII, DI and DII. It means only steel arched support bear the load from ground in these support patterns. Young’s modulus, area and section Table 6. Rock kind and support pattern in measurement. Tunnel Kind of rock (Support pattern) A

Vertical load Pv Spring for ground reaction K Pv

B C D E G H

K (Pv + γD) Horizontal load

Figure 1. Outline of analysis modeling.

N O

509 Copyright © 2004 Taylor & Francis Group plc, London, UK

Tuff (CII), Rhyolite (CII (2 sections), L) Tuff including sand (L) Tuff including basalt (L), Fault fracture zone (L) Tuff including basalt (L), Fault fracture zone (L) Sandstone (CII) Alteration strata of sandstone and mudstone (DII), Sandstone (CII, DII) Clay stone (DI (2 sections), DII (2 sections)), Green tuff (CII), Phyllite (CII, DI), Chert (DII) Chert (DI (2 sections), Gabbro (CII), Dolerite (DI), Phyllite (DI (2 sections), Clay stone (CII, DI (2 sections)) Granite porphyry (DI), Tuff including rhyolite (CI, CII), Fault fracture zone (L) Rhyolite (L (2 sections))

2/18/04

2:36 PM

Page 510

moment of inertia were set from the specification of steel arched support. The load quantity was presented in the value of unit length in tunnel longitudinal direction. Main beam was only regarded as the structural material in steel segment or simple liner and specification of section was set. Poisson’s ratio of steel was used as 0.3. Spring for ground reaction in radial direction was only considered and the one in tangent direction to the support was neglected. Spring was set in full circle as shown in Figure 1 and stiffness was considered 0 in tensile situation. Ground reaction modulus kn was firstly calculated and sprin stiffness was acquired by multiplying kn with the area that spring would bear. Spring stiffness was calculated referred from the standard of cut and cover tunneling method in Japan (1996) as follows:

120 Net stroke speed (mm/min)

09096-62.qxd

A B C D E F G H I J K L M N O

100 80 60 40 20 0

B

CI

CII DI DII Support pattern

L-1

L

Figure 2. Relation between net stroke speed and support pattern. 14 Cutter rotation speed (rpm)

(2) where kn  ground reaction modulus to the radial direction to support (kN/m3);   compensating value considering the calculation method of deformation modulus and load condition; E0  deformation modulus of ground (kN/m2) and D  tunnel diameter (m). E0 was the same value as the Young’s modulus from uniaxial compression test using the in-situ material. The joint of steel segment and simple liner was supposed as enough stiffness.

A B C D G H I J K L M N O

12 10 8 6 4 2 0

B

CI

CII DI DII Support pattern

L-1

L

3 RESULT OF ANALYSIS 3.1

Figure 3. Relation between cutter rotation speed and support pattern.

Analysis result of machine data

The relation between net stroke speed or cutter rotation speed and support pattern were examined. As for thrust force and cutter torque, it was thought better to eliminate the influence of tunnel diameter so the relation between advance energy or rotation energy and support pattern were checked. Firstly, the speed related items of TBM performance were focused to discuss the relation with support pattern. Figure 2 shows the relation between net stroke speed and support pattern. The speed was thought to be relatively low when the ground condition was both good and bad. However, there were some cases that the speed increased in bad ground condition. It was thought that it might take time in hard excavation of high rock strength in good ground condition. In bad ground condition there are either easy or cautious excavation situation because of low rock strength. It was found that there was the correlation between net stroke speed and support pattern in the ground conditions except support pattern B and L. However, it was difficult to discuss the relation in site about using net

stroke speed for ground evaluation because the speed may be influenced by other construction conditions. Figure 3 shows the relation between cutter rotation speed and support pattern, except tunnel E and F because it could not be calculated. The speed decreased when the ground condition became bad from this figure for all of tunnels, however, the change range of the speed was small when each tunnel was separately discussed. It was thought to be difficult to discuss the support pattern using this item, also because there are also large differences for each tunnel. The relation between net stroke speed or cutter rotation speed and support pattern was confirmedly found, however, large difference at each tunnel was also admitted. The other items should be used at the same time to evaluate ground condition and support pattern. Secondly, the energy related items of TBM were focused to discuss the relation to support pattern. Figure 4 shows the relation between advance energy and support pattern. Advance energy decreased when

510 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-62.qxd

2/18/04

2:36 PM

Page 511

12000

400 300 200 100

B

CI

CII DI DII Support pattern

L-1

8000 6000 4000 2000 0

L

B

Figure 4. Relation between advance energy and support pattern.

the ground became bad, that is, the support was stiff. To find the difference between support pattern L and L-1 was difficult. This reason was thought that the excavation was done not only by TBM itself but also by shield jacking force. Distribution of the average value of advance energy was found approximately two times considering all of tunnels at each same support pattern. Figure 5 shows the relation between rotation energy and support pattern. Rotation energy decreased when the ground condition became bad and the tendency was thought to be same as that of advance energy. Distribution of the average value of rotation energy was found approximately three times considering all of tunnels at each same support pattern. In addition, there was less distinction in rotation energy than in advance energy. The multiple number of rotation energy was bigger than that of advance energy. Support selection should be done considering this difference when these items were focused. Figure 6 shows the relation between unit volume excavation energy and support pattern. Unit volume excavation energy was acquired on the basis of the values both advance and rotation energy. The value of rotation energy was much larger than that of advance energy so unit volume excavation energy was thought to be dependent on the rotation energy. The tendency of rotation energy and unit volume excavation energy was found to be almost same. The average value of unit volume excavation energy at each support pattern was found approximately three times as well as the rotation energy. From the results of each machine data and energy variance, it was found that some machine data variances had correlation with support pattern tendency. Next the tendency of machine data combination in the same tunnel would be discussed. Figure 7 and 8 show the relation between cutter torque and thrust force in tunnel O. Figure 7 means the

CI

CII DI DII Support pattern

L-1

L

Figure 5. Relation between rotation energy and support pattern. 70000 A B C D G H I J K L M N O

60000 50000 40000 30000 20000 10000 0

B

CI

CII DI DII Support pattern

L-1

L

Figure 6. Relation between unit volume excavation energy and support pattern. 800 700 B

600 500 400

CI

300

CII

200 DI

100 0

0

2000

4000 Thrust force (kN)

6000

8000

Figure 7. Relation between cutter torque and thrust force in soft rock quality in tunnel O.

rock quality was soft, while Figure 8 was hard shown in the classification of Table 2. The center of ellipse was equivalent to the average value and the length of ellipse was standard deviation for each machine data in

511 Copyright © 2004 Taylor & Francis Group plc, London, UK

A B C D G H I J K L M N O

10000

Unit volume excavation energy (kN/m2)

0

Rotation energy (kN r/m2)

A B C D E F G H I J K L M N O

Cutter torque (kN m)

Advance energy (kN/m2)

500

09096-62.qxd

2/18/04

2:36 PM

Page 512

900

800 CI

600

CII

500 400 300 200 DI

100 0

2000

4000 Thrust force (kN)

6000

Cutter torque (kN m)

CI

600 500 DII

400 300 200 100 0

0

2000

4000 Thrust force (kN)

500

DII

400

B

DI

300 200

0

2000

4000

6000

8000

Thrust force (kN)

CII DI

600

0

900

700

700

8000

Figure 8. Relation between cutter torque and thrust force in hard rock quality in tunnel O.

800

CI

100

L 0

CII

800

B Cutter torque (kN m)

Cutter torque (kN m)

700

6000

8000

Figure 9. Relation between cutter torque and thrust force in soft rock quality in tunnel F.

Figure 10. Relation between cutter torque and thrust force in hard rock quality in tunnel F.

excavation energy have mutual relations with support pattern. However the value of machine data themselves were different in different tunnels such as approximate two times difference in advance energy and three in rotation energy. Therefore, characteristic of machine and rock quality might be considered for selection of support and other items should be added if needed. Rotation energy and unit volume excavation energy was thought to be unsuitable for selection of support in good ground condition of hard rock and thrust force itself also should be examined when the support pattern was decided, while all of three data had good correlation with support pattern in all of the ground condition in soft rock quality and bad ground condition in hard rock quality. 3.2

these figures. The almost linear relationship between cutter torque and thrust force for soft rock was found from Figure 7. There was also linear relation from CII to DII, however it was hard to admit clearly linear relation between B and CII from Figure 8. Thrust force was increasing as the ground became good, however, the value of cutter torque did not change so much. It was thought that the selection of support with cutter torque was difficult because of less difference in the good ground condition of hard rock. As a result, it means the rotation energy or unit volume excavation energy might be unsuitable for selection of support in good ground condition of hard rock quality. Figures 9 and 10 show the relation between cutter torque and thrust force in tunnel F for soft rock and hard rock quality. The same tendency was acquired as well as the case of tunnel O, despite support pattern C in soft rock and B in hard rock were small volume of data and the tendencies were a little bit different. As a result from the analysis of machine data, advance energy, rotation energy and unit volume

512 Copyright © 2004 Taylor & Francis Group plc, London, UK

Results of numerical analysis

Load acting on support was examined by numerical analysis. Steel arched support in one section of support pattern CI was set to measure strain and calculate load, despite adoption of steel arched support in normal section of support pattern CI was not. Figure 11 shows the average, maximum and minimum values of Pv/D when the error function acquired from the relation between the stress from in-situ measurement and one from numerical analysis was minimum. Increase tendency of the load acting on support was shown while the ground condition became bad. The maximum value of Pv/D equal to 0.9 was acquired and the load height equivalent to approximately tunnel diameter acting on the support was shown. The difference of load values between the group of support pattern CI and CII and the group of DI, DII and L was clearly found. Figure 12 shows the average, maximum and minimum values of K ratio in the same way. No large difference of K ratio distribution at each support pattern was found. K ratio from 0 to 0.5 was the most case.

09096-62.qxd

2/18/04

2:36 PM

Page 513

Figure 13 shows the values of Pv/D classified by rock quality classification on the basis of table-2 and support pattern over C, that is, CI and CII and under D, that is, DI, DII and L. Size of circular in the figures means the frequency of analysis results. Pv/D in the ground with the support pattern over C distributed under the value 0.2 and was no big difference, while Pv/D in the ground with the support pattern under D had large values. In terms of rock quality, Pv/D was found to have larger value in soft rock quality. It was found that the maximum load in the support pattern over C was decided regardless of rock quality and that the maximum load in the support pattern under D should be examined considering the rock quality.

1

Pv/γD

0.8 0.6 0.4 0.2 0

CI

ICI

DI

DII

L

(5)

(9)

Support Pattern Section number

(1)

(10)

(11)

Figure 11. Load acting on support from numerical analysis.

Comparison and examination of relation between the tendency of machine data and grade of geological condition with the data using TBM excavation in Japan were carried out. Furthermore, numerical analysis using frame model was done and the load acting on support was calculated. The results were acquired as follows:

1

K ratio

0.8 0.6 0.4 0.2 0

4 CONCLUSION

CI

CII

DI Support pattern

DII

L

Figure 12. K ratio acting on support from numerical analysis.

1 Sample number 3 0.8

Pv/γD

0.6

(1) The correlation between some machine data and support pattern was found when each tunnel was separately discussed. However, the difference of values was also admitted in the same support pattern when all of tunnels were compared at the same time. The combination of items which reflect rock and/or machine characteristic in addition to machine data should be examined to decide support patterns. (2) Linear correlation between thrust force and cutter torque was less to be admitted in good ground condition of hard rock quality. (3) Maximum load acting on TBM tunnel support was found to be approximately the weight of rock height equivalent to the tunnel diameter. K ratio among 0 and 0.5 was also acquired. (4) The less influence of rock characteristics to the load in ground condition applied as support pattern over C was admitted, while the influence in support pattern under D was admitted.

0.4

5 FUTURE PROBLEMS 0.2

0 Hard and Medium Soft Hard and Medium Soft hard rock rock hard rock rock Support Pattern Support Pattern Support Pattern Support Pattern over C over C under D under D

Figure 13. K ratio acting on support from numerical analysis.

Correlation between machine data in TBM excavation, support pattern and ground condition will be checked by adding the machine characteristics and detailed rock quality. Calculated load and other rock characteristics such as the result of Schmidt hammer test or point load index should be also done. Comparing load value acquired in this study, that is, by TBM excavation with the one acquired by NATM excavation should

513 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-62.qxd

2/18/04

2:36 PM

Page 514

be achieved in the same way. Finally, the standard of ground estimation and support design method in Japan will be proposed. REFERENCES

Scorali, F. 1995. Open-face borers in Italian Alps, World Tunnelling, November 1995. 361–366. Japan Tunneling Association, 2000. TBM handbook in Japan (in Japanese) Japan Society of Civil Engineers, 1996. Japanese standard for cut and cover tunnel. (in Japanese)

Nilsen, B. & Ozdemir, L. 1999. Recent developments in site investigation and testing for hard tock TBM projects, 1999 RETC proceedings, Ch. 39: pp.715–731.

514 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-63.qxd

2/18/04

2:37 PM

Page 515

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

A durability design for precast concrete segments for tunnel linings Gustav Bracher Sika Tunneling & Mining, Widen, Switzerland

Dennis Wrixon American Commercial Inc., Bristol, VA

ABSTRACT: Today, most major tunnel projects specify a design life of 100 years or more. The consequence thereof is a further development in concrete technology, especially for the production of tunnel lining segments in TBM tunneling. This paper highlights the proposal of a systems approach to the task of producing ultra-durable precast concrete tunnel lining segments. The system comprises a High Performance Concrete (HPC) having a Chloride Penetrability 1000 Coulombs and a recently developed epoxy dispersion coating, immediately applied to the warm surface of the concrete after demoulding. Application of the coating eliminates early age surface micro-cracking, resulting in improved concrete durability. The synergies of steam cured HPC and the coating system provide significant benefits to both the precast segment producer and the project client.

1 INTRODUCTION In large underground projects TBM tunneling is continuously gaining more importance. It allows the construction of tunnels without interference and disturbance to existing structures. Especially in urban areas with sensitive infrastructures, mechanised tunneling is often the preferred method. It is also selected in areas of minimum overburden, in soft grounds, and in areas of ground water to avoid interactions or settlements at the surface. TBM tunneling technology has improved consistently over the years. In parallel, concrete technology continues to develop and has recently taken a further step forward with respect to the production of tunnel lining segments, following the development of a complete system design. Although the following article concentrates on single shell tunneling, the findings can be applied for double-shell linings in aggressive environments as well (e.g. Adler tunnel). As is the case with most concrete production, the durability of tunnel lining segments is frequently the result of a compromise between production requirements and proper curing methods. Few products provide a better example of this trade-off than steam cured, precast concrete segments. In order to gain multiple uses of moulds in a 24-hour period, steam curing techniques are employed by precast producers to achieve high early compressive strengths and enable demoulding of segments within several hours after

casting. Steam curing itself has been shown to be detrimental to the permeability (durability) of concrete. It is generally accepted that the thermal stresses induced in the surface layer of concrete, after removal from the steaming chamber, can cause severe microcracking [1,2]. Micro-cracks allow the penetration of deleterious substances and expose more surface area to attack, resulting in concrete deterioration which, in turn, causes more cracking and an increase in permeability. It has been suggested that micro-cracks greater than 80 microns in width will cause an increase in the water permeability of concrete [3]. The steam curing process does not actually “cure” the concrete, but rather it accelerates the hardening process by heating to temperatures up to 70°C, in an environment of 100% relative humidity. After removal from the steam chamber, the core of the concrete section retains its temperature relative to the surface of the concrete which quickly loses heat as it tries to reach equilibrium with the surrounding ambient temperature. The consequence of this temperature differential is that moisture is rapidly lost from, and large thermal gradients are set up in, the surface layer of the concrete, causing micro-cracks from drying and thermal shrinkage. If wet curing of the concrete is required after steaming, it is usually specified that the curing is not applied until the temperature of the concrete has cooled sufficiently to eliminate the possibility of thermal

515 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-63.qxd

2/18/04

2:37 PM

Page 516

shock. Obviously, the application of thorough wet curing techniques for several days to large quantities of precast concrete segments is a task that is not relished by concrete manufacturers. It requires a large covered area in which to store and cure the segments. In addition to these plant requirements, “double handling” and movement of the segments will be required, causing a greater risk of damage to the units. In the event that a protective coating is required to be applied to parts or all surfaces of the segments, then further handling and surface preparation operations will be necessary. If the protective coating required is a solvent based material, then the manufacturer will have the logistical problem of trying to reduce the surface moisture of a wet cured (saturated) concrete down to a level which is acceptable for the application of a solvent-based primer.

The ingress of such materials is significantly reduced by the incorporation of High Performance Concrete (HPC) in segments and may be totally eliminated by providing a waterproof barrier to the extrados with a suitable coating material. 2.2

If the outer surface of the concrete, below the water table, becomes saturated, a water vapour gradient is set up between the dry interior and the saturated area. As water evaporates from the internal face, more water is drawn through the concrete and, upon evaporation, salts are deposited at a point close to the saturation front. Concrete Lining Water Table

2 CORROSION MECHANISMS IN TUNNELS LINED WITH SEGMENTS

Deposition

There are, primarily, three different types of chloride induced corrosion mechanisms in tunnels lined with concrete segments. A fourth method of penetration, by air-borne chloride ions in marine and coastal environments, is also thought to contribute to corrosion, particularly in railway tunnels where the “piston” effect of moving trains creates high air pressures that force chlorides and other deleterious materials into the lining. 2.1

Transpiration

Normal penetration

Groundwater will penetrate the external face of tunnel linings relatively quickly, even in concrete of low permeability and low absorptivity. In deep tunnels, the rate of ingress can be alarmingly fast, due to high pressure heads from groundwater. The groundwater itself may contain various concentrations of chlorides, sulphates and other deleterious materials.

Evaporation

Saturated Zone

Flow

These salts, such as chlorides and sulphates, can build up their concentrations and cause corrosion of reinforcement and attack the concrete paste matrix. The ingress of substances contributing to the mechanism of transpiration can, again, be significantly reduced by the use of HPC and the provision of a barrier type coating to the external surface of the segments. 2.3

Leaking joints and cracks

The sealing of segment joints relies upon compression seals, or gaskets. In spite of all the good intentions of Water Table

Lining Lining

Water seepage through leaking segment joint

Penetration Under Hydrostatic head

516 Copyright © 2004 Taylor & Francis Group plc, London, UK

Deposition

Evaporation

09096-63.qxd

2/18/04

2:37 PM

Page 517

designers and contractors, gaskets can become misaligned during the erection of segments in the tunnel. A consequence of this is that a newly constructed and lined tunnel may well have several leaking joints. If leaks occur through cracks or joints in the concrete lining, the groundwater will run over the inside face and be absorbed into the concrete. The water evaporates off and leaves behind deposits of chlorides and other deleterious materials. Prolonged cycles of wetting and drying (analogous to the tidal zone in a marine structure) will lead to the build up of very high concentrations of chlorides and sulphates. In the case of chlorides, back diffusion into the concrete from the internal face of the lining will occur, with the resultant corrosion of reinforcement and spalling of the concrete cover zone. Experience in recent years has shown this type of corrosion mechanism to be the fastest acting and most damaging. Corrosion damage to concrete linings in tunnels less than 10 years old has been encountered in some Asian countries. The use of a low permeability HPC in tunnel segments is extremely helpful, but, with the very high levels of surface chloride contamination (circa 5% by weight of cement) that can be built up through re-absorption from leaking joints, it cannot provide a solution on its own. The best way to keep re-absorption (and back diffusion) to a minimum is to apply a barrier coating to the intrados and sides of tunnel segments. To be effective, however, the coating must be maintained throughout the life of the structure.

parameters must be considered before deciding upon a possible solution. The provision of a HPC (containing condensed silica fume) in the segments is an obvious answer. Not only are the benefits of low permeability realised, but also improved resistance to sulphate and acid attack, and higher early and ultimate compressive strengths are achieved. The application of a barrier type coating to all surfaces of the tunnel segments is also necessary. The problem is not to find a suitable coating material with the required barrier properties, but to provide a coating that acts in synergy with the concrete, making a very good concrete even better. The logistics of segment production also have to be considered. The wet curing of segments after demoulding is difficult to achieve successfully, unless suitable plant and manpower resources are allocated to the process. The successful application of solvent based epoxy and other coatings to concrete is possible, but attention and effort have to be paid to substrate preparation and reducing the surface humidity of the concrete to levels low enough to receive these materials without the possibility of blistering and delamination. Often, coating manufacturers recommend that concrete should be a minimum of 28 days old before it is able to receive a coating. This could mean that a concrete manufacturer has at least 4 weeks of its production in temporary storage, prior to coating, and will have to lift and handle the segments several times before they can be delivered to site. Considering the above parameters, it can be seen that the ideal coating system to be combined with HPC precast concrete would have the following properties:

2.4

• • • • • • •

Air-borne chlorides

In marine and coastal environments, the penetration into concrete tunnel linings of air-borne chlorides must be considered. These salts are driven through tunnels by ventilation systems and the movement of trains and road vehicles. Once the temperature of the surface of the tunnel lining falls below the dew point, moisture will condense on the surface of the concrete and the chlorides will go into solution. As the dew evaporates, deposits of chlorides will be left on the concrete surface and build up in concentration, due to cyclic wetting and drying. In rail and metro tunnels, high air pressures caused by the “piston” effect of moving trains tend to “drive” chloride ions into concrete capillaries and accelerate the processes of penetration and diffusion. The application of a barrier type coating to the inside face of the lining to provide durability should be considered. 3 A SYSTEMS APPROACH TO PROVIDING DURABILITY The problem of corrosion in precast concrete segments for tunnel linings is a complex one and several

• • • •

4 A SOLUTION TO PROVIDING 100 YEARS DESIGN LIFE In 1993, prior to the construction of its Rail 2000 and AlpTransit projects, the Swiss Federal Railway Organisation (SBB) launched a research program for the evaluation of concrete mix designs with high durability for tunnel linings [4]. Sika AG was consulted

517 Copyright © 2004 Taylor & Francis Group plc, London, UK

Impenetrable to water High resistance to chloride and sulphate attack Low oxygen diffusion properties Good bond to substrate Fire resistant Anti-carbonation properties Able to act as a curing membrane at very early concrete age Able to be applied to concrete after demoulding and moisture vapor permeable Easy application, with little or no surface preparation Fast curing to enable early delivery to final outside storage Non toxic, solvent-free product.

09096-63.qxd

2/18/04

2:37 PM

Page 518

at the beginning of this program and was involved in the production of over 80 different types of concrete mixes for investigation. One particular aspect of this research related to the durability of steam cured, precast concrete segments for linings. It was acknowledged that the durability of such elements would have to be improved by a combination of better overall curing and a protective surface coating. In addition to the concrete mix evaluation, several types of water based epoxy coating materials were developed according to the above mentioned requirements. Based upon the results and recommendations of the program, SBB decided to use a system of HPC (containing 7.5% of a silica fume based admixture by weight of cement, and a 28-day minimum compressive strength of 60 MPa) for the production of the segments for the 4.3 km long, 12.58 m diameter, bored section of the Adler tunnel, near Basle. The tunnel passes through some severe geological conditions, including an area of swelling gypsum keuper, requiring the segments to have an exceptional resistance to sulphate attack. In addition to the selected HPC, it was decided that a protective coating would also be necessary and a water based epoxy dispersion was chosen to be applied to the tunnel segments immediately upon their demoulding after steam curing. An extensive 2-year testing program for sulphate resistance was carried out with excellent results. Microscopic analysis of coated and uncoated samples of segment surfaces also showed that the early application of the coating completely eliminated the surface micro-cracking in contrast to the uncoated concrete elements. It was also decided to test the coated concrete surface for its abrasion resistance to scouring and scratching from the TBM tailskin sealing brushes (steel wire), with the result that, even at maximum pressure, the steel brushes only abraded the coating to a depth of 1/10 of its thickness. During 1996/97, segments for the Adler tunnel were successfully manufactured in the contractor’s precasting yard. The specified concrete strength of 60 MPa was achieved between 5 and 7 days and, since the segments had already been coated, the contractor was able to install them in the tunnel at about 10 days of age.

at Taywood’s London plant. The moulds used were from the Copenhagen Metro project and the same concrete mix had already been used in production of segments for the Channel Tunnel. Three segments had all faces coated within half an hour of demoulding. The other three segments were not coated and cured and stored as would be in normal production. The object of the program was to compare the difference in various technical performance parameters between coated and uncoated concrete samples cored from the six segments. All testing was conducted by Taywood Engineering Limited (TEL), at its London testing laboratories. In addition to the testing of several penetrability characteristics, fire testing of the coating material for flame spread (BS 476) and smoke and toxic gas emission was undertaken at the Warrington Fire Research Station, in the United Kingdom. On its own, the concrete exceeded the minimum requirements for water permeability and oxygen diffusion [6] that were specified to provide a 120-year design life on the Jubilee Line Extension project in London. Early age application of the coating was anticipated to improve upon the already excellent hardened properties of the concrete mix. 5.1

Results

The concrete mixture used by Taywood Precast in the testing program was as follows: OPC: 310 kg/m3, Pulverized Fuel Ash: 128 kg/m3, 20 mm aggregate: 581 kg/m3, 10 mm aggregate: 454 kg/m3, granite crushed fines: 781 kg/m3, Sikament FF: 5.85 lt/m3, W/B ratio: 0.34. The segments were cured in a steam tunnel for 18 hours, with a maximum temperature of 53°C and a maximum concrete temperature controlled to 60°C. Compressive strength

18 hours 7 days 28 days

35 Mpa 62 Mpa 78 Mpa

The following results were obtained from the testing program: Water vapor transmission (BS 7374; 1990)

5 SIKA/TAYWOOD PRECAST/TAYWOOD ENGINEERING TEST PROGRAM Despite the successful experience from the Adler tunnel project, Sika decided that a further independent investigation of the epoxy coating material and its benefits in tunnel segment production was required. In January 1998, a comprehensive testing program commenced at Taywood Precast Ltd in the United Kingdom [5]. In this program, 6 tunnel segments were cast and steam cured in a normal production schedule

Control concrete Coated concrete Equivalent air layer thickness

Mean flux 0.367 gm/m2 (24 hours) 0.5 m

For “breathable” anti-carbonation coatings, a widely accepted criteria is that the equivalent air layer thickness should be 4 m in order to minimise the possibility

518 Copyright © 2004 Taylor & Francis Group plc, London, UK

Mean flux 0.545 gm/m2 (24 hours)

09096-63.qxd

2/18/04

2:37 PM

Page 519

of moisture vapour induced blistering. Thus, as found, a low water permeability is compatible with a relatively high vapour diffusion coefficient.

Coated concrete

Water Permeability (TEL Method) – 10 bar pressure for 4 days

Coated concrete

Coeff. of permeability avg. 1.05 1013 m/s Coeff. of permeability nil

Typically, a good quality concrete would be expected to have a water permeability coefficient of 1.0 1012 m/s – as found for both uncoated control samples. For both coated specimens, no water penetration could be detected after a 4-day test period. Chloride ion bulk diffusion Control concrete Coated concrete

Coeff. of Cl diffusion avg. 7.84 1013 m2/s Coeff. of Cl diffusion avg. 3.36 1013 m2/s

The results obtained for the coated specimens (after a 28 day immersion period) demonstrate a distinct reduction in chloride ingress. Chloride penetrability (ASTM C1202) Control concrete Coated concrete

Mean Mean

653 Coulombs 282 Coulombs

All samples are within the range quoted in ASTM C1202 as having a very low chloride ion permeability. However, for the coated samples, the total charge passed under the test conditions was approximately half of that determined for the control concrete. Limpet pull-off adhesion tests Average bond strength between concrete substrate and coating 1.87MPa. No failure of coating or coating/substrate adhesion. The good quality adhesion may be attributed to the penetration of the resin from the coating into the concrete substrate.

Micro-cracking (Petrographic analysis) Micro-cracking (10 m) was found originating from the surfaces in the uncoated specimens. No surface micro-cracks were found on the coated specimens. Carbonation (Petrographic analysis) An average carbonation depth of 0.45 mm was noted in the control concrete specimens, with a maximum depth of 1.3 mm. No carbonation was noted in the coated specimens. Even at the lower film thickness, the coating has effectively arrested the carbonation process. Fire testing (BS 476, BS 6853 & Other) The epoxy coating achieved a Class 1 rating for flame spread, when tested in accordance with BS 476:Part 7. Smoke emission tests, in accordance with BS 6853, showed no flaming of the surface. Testing for combustion gases showed to have “Conventional Index of Toxicity” (CIT) value of 8.261. These results show that, in the event of a fire in a tunnel, no special precautions need to be taken due to the presence of the coating. The investigated coating material even exceeds the requirements contained within the specification issued by MTRC (Hongkong) for the interior surfaces of railway-stock. Bonding of gasket An additional full-scale experiment was carried out in order to determine, whether the epoxy coating could be used to bond an EPDM gasket to the segment without the use of a contact glue and to integrate the fixation of the gaskets into the working process before placement in the final storage area. The inspection two months later showed excellent adhesion between the gasket and the concrete.

6 CONCLUSIONS With the experience from the production of the segments from the Adler tunnel and the additional results from TEL with the newly developed water based epoxy coating, the Land Transport Authority (LTA) of Singapore decided to apply this coating to one of their

519 Copyright © 2004 Taylor & Francis Group plc, London, UK

Coeff. of O2 diffusion avg. 1.87 108 m2/s Coeff. of O2 diffusion avg. 1.15 108 m2/s

The coating has a measurable and beneficial barrier effect.

Penetration 6.3 mm Coeff. of permeability 9.7 1014 m/s Penetration nil

No water penetration was observed with the coated samples.

Control concrete

Control concrete Coated concrete

Water permeability (DIN 1048) – Upto 7 bar pressure for 4 days Control concrete

Oxygen diffusion tests

09096-63.qxd

2/18/04

2:37 PM

Page 520

contracts for the construction of the new North-East Line from the Metro of Singapore, as an alternative to the originally formulated specifications by coating the segments with a coal/tar epoxy product. The achieved results with the Sika system design of producing highly durable tunnel lining segments, were very successful, so that for the next metro line extension (Marina Line) the specifications will be changed to cure and coat the segments on 6 sides with a water based epoxy dispersion immediately after demoulding to achieve, in combination with a High Performance Concrete, the required durability of 100 years design life. The Sika system has also been chosen for the Metro Caracas Tunnel in Venezuela. At the CTRL (Channel Tunnel Rail Link) project in the UK, the Sika product has been used as the segment gasket glue.

[2] Hooton, R.D. et al., “Influence of Silica Fume on Chloride Resistance of Concrete”, International Symposium in High Performance Concrete, New Orleans, October 1997. [3] Shah, S.P. & Wang, W., “Microstructure, Micro-cracking, Permeability and Mix Design Criteria of Concrete”, Fifth International Conference on Structural Failure, Durability and Retrofitting, Singapore, November 1997. [4] Bracher, G., “The Technology of the Production of Tunnel Lining Segments with the Aspect of High Durability Under Aggressive Conditions and Optimized Production Facilities”, Sika A G, November 1997. [5] Taywood Engineering Ltd, Technology Division, “Technical Report 1304/98/10256 – Assessment of Sikagard 65W for use with Precast Concrete Units”, London, July 1998. [6] Varley, N., “Concrete Tunnel Linings at London Bridge”, Concrete, February 1998.

REFERENCES [1] Brugger, M., “Curing Precast Concrete Tunnel Lining Segments, Concrete International, 1994, pp. 51–54.

520 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-64.qxd

2/18/04

8:01 PM

Page 521

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Design and construction of the Lindbergh Terminal Station, Twin Cities, Minnesota E.E. Leagjeld, B.K. Nelson, C.R. Nelson, D.L. Petersen, R.L. Peterson & B.D. Wagener Minneapolis, Minnesota, USA

ABSTRACT: The Lindbergh Terminal Station was constructed beneath one of the nation’s busiest airports with minimal impact to the airports facilities or services. This unique project uses cavern design and construction technology developed with several similar projects constructed in the Minneapolis-St. Paul metropolitan area. Successful project completion demonstrates the economic and technical feasibility of constructing large underground excavations in heavily developed metropolitan areas.

1 INTRODUCTION

2 SITE GEOLOGY

The $675.4 million Hiawatha LRT Line will connect three of the Twin Cities most popular destinations – downtown Minneapolis, Twin Cities International Airport and the Mall of America in Bloomington. Currently under construction, the segment of the line at the airport will run in twin-bore, 1.4-mile-long tunnels, and includes one underground and one aboveground station. The Lindbergh Terminal Station is the fifth major cavern constructed in the Twin Cities using the Platteville limestone to form the flat roof, and with the cavern excavated in the Glenwood shale and St. Peter sandstone. Design calculations ranged from simple linear-arch calculations for a single rock beam to complex structural models incorporating multiple rock layers separated by seams, vertical jointing, rock reinforcement and nonlinear sandstone behavior. Instrumentation data has been invaluable in assessing the location and timing of rock reinforcement, and in inferring the location of roof beams. As of November 2003, the Hiawatha light-rail project is approaching 90 percent complete and is on time and budget. Passenger service is anticipated to begin in April 2004, serving the segment of the line from the Minneapolis warehouse district to just north of the airport. The airport to Mall of America segment will come online in December 2004. Twelve of seventeen stations have been completed along the 12-mile line. The only underground station, at the Lindbergh Terminal, is the subject of this paper.

2.1

The Twin Cities metropolitan area, including the Lindbergh station site, is underlain by nearly 1,000 feet of sedimentary rocks of early Paleozoic age. These gently dipping to near-horizontal rocks form the Twin Cities structural and hydrologic basin. A mantle of glacial and post-glacial deposits covers the area. The Lindbergh station site is under the approach apron for the terminal parking payment kiosks. Under the apron pavement is about 7 feet of soil, followed by the following three bedrock formations pertinent to station construction.

2.2

Platteville limestone

The 30-foot thick Platteville formation consists of several dolomitic limestone and dolomite members. Steep to vertical joints are common. The upper, more weathered limestone is also more fractured. Many fractures are tight, allowing little water movement, whereas others transmit water readily. Bedding in the limestone is nearly horizontal. Open bedding planes in the limestone are commonly water bearing.

2.3

Glenwood shale

At the base of the Platteville limestone is the thin (2-feet to 5-feet thick) Glenwood shale, which overlies

521 Copyright © 2004 Taylor & Francis Group plc, London, UK

General

09096-64.qxd

2/18/04

8:01 PM

Page 522

the St. Peter sandstone. The shale consists of a series of beds of soft, argillaceous, and sandy shale grading upwards to harder, shaley, dolomitic layers. The contact between the Glenwood and the Platteville limestone is distinct but irregular over a zone of zero to two inches. There is no parting plane between the shale and limestone. The shale is transitional over a few feet into the St. Peter sandstone. 2.4

St. Peter sandstone

The St. Peter is a 150-foot thick massive sandstone unit composed of fine-grained to medium-grained, uniformly graded quartz sand. The formation is bedded and cross-bedded, and closely-spaced, weak bedding and cross bedding planes can result in very weak sandstone. Layers with variations in strength and hardness are common. Irregularly spaced joints are present in the sandstone. Most are steeply dipping (more than 70 degrees). The unconfined compressive strength of the St. Peter sandstone typically ranges from 0 psi to 1000 psi, and a realistic average is 500 psi.

was also used to remove Glenwood shale from the undersides of the Platteville limestone roof, necessary because of the absence of a parting plane. The station roof is about 30 feet below grade, consisting of about 7 feet of soil and 23 feet of Platteville limestone. The station is 39 feet high, making the invert approximately 69 feet below grade. As is typical of transit stations, many excavations intersect the station excavation (refer to Figures 1 and 2). Existing and planned site features, including utility excavations, pedestrian bridge foundations, and buildings constrained the station and access layout. Two TBM bores intersect the north and south endwalls of the station. The north shaft is located just outside the station proper in the northwest corner, while the south shaft is located within the station near the south endwall. The principal user access to and from the station is via the airport Transit Center connection, which exits the station from the west wall.

3 STATION GEOMETRY Figures 1 and 2 illustrate the station plan and cross section, respectively. The 530-ft long excavation is roughly rectangular in cross section, with flat roof and floor, and slightly curved sidewalls. The excavation span is 57 feet at the roof and 63.5 feet at the wall midheight. A total of 48,000 cubic yards of shale and sandstone were excavated for the station, about 13,000 cubic yards by the TBM bores and the remainder by roadheader, loaders and excavators. The roadheader

Figure 2. Lindbergh station cross section.

Figure 1. Lindbergh station plan.

522 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-64.qxd

2/18/04

8:01 PM

Page 523

4 DESIGN AND GROUND SUPPORT 4.1

Other flat-roofed caverns in the Twin Cities

The Lindbergh Terminal Station is the fifth major cavern in the Twin Cities using the Platteville limestone to form the flat roof, with the cavern excavated in the Glenwood shale and St. Peter sandstone. The development of flat-roof design methodology began in 1976 with the first NSF-sponsored 50-foot wide research cavern constructed at the University of Minnesota. The designs have been continuously refined to create the 57-ft span and 40-ft high Lindbergh Station cavern. Other major flat-roofed caverns include the Minneapolis East Interceptor valve chamber, the University of Minnesota Civil Engineering Building and the Elmer Andersen Library. 4.2

Flat-roof cavern design

The first step in designing large-span, flat-roofed caverns is characterization of the location and strength of horizontal seams in the roof rock. These features control how the roof rock subdivides into separate beams as the cavern is excavated. Next, the rock quality and deformability of the various roof beams must be determined. The beam thickness and rock modulus determine the beam stiffness, which also influences how beams interact (e.g. a thin, low stiffness beam will tend to sag away from a stiffer overlying beam, or will sag onto and load a stiffer underlying beam. Finally, the rockbolts must be selected, including type, length, spacing, bar size, and prestressing. Typical design criteria are beam compression and tensile stresses, interbeam shear forces and roof deflections. Station roof rockbolts consisted of: 1. Hollow-core, prestressed, cement grouted, 1-3/8 diameter, 10 ft-long rockbolts were installed on a 5 5 or 5 7 pattern. 2. Solid, untensioned, resin-grouted 1-3/8, 12 ftlong, all thread, rockbolts were installed on diagonals of the hollow core rockbolts. 4.3

footing bearing on the sandstone and through a shotcrete closure, similar to an inverted footing, placed between the panel and limestone roof. The wall panels also support a suspended solid precast roof over the entire station and hollow core floor plank in the ancillary areas. To reduce the grouting pressures and minimize panel thickness and weight, lightweight cellular grout was specified for contact grouting. A 12-inch thick panel was originally specified, but reduced to 10-inches during construction by increasing the amount of reinforcement and reducing the lightweight cellular grout density from 50 pcf to 38 pcf. 5 CONSTRUCTION SEQUENCE AND SCHEDULE The station excavation was performed in two main phases, as depicted in Figure 3. The first phase consisted of a 17-foot high top heading. The top heading was started in April 2002 after the first TBM tunnel passed through the station. The top heading invert was within two to three feet of the tunnel liner crown. Installation of rockbolts was specified to follow immediately behind the top heading excavation. Excavation of the second TBM tunnel through the station limits occurred in August 2002. The top heading excavation was completed by September. Excavation of the bench and removal of the tunnel segments was not started until December 2002, with the completion of the second TBM bore. Figure 4 shows the initial phase of TBM liner and bench removal at the station’s south end. Bench and tunnel ring removal was limited to approximately 70 feet beyond the last installed and contact grouted precast wall panels. Sections of footing were cast with a high early strength concrete mix, which allowed placement of wall panels on top of the footings in as little as three days. Panels were installed in groups of either four or five panels.

Sandstone wall design

The sandstone walls provide support for the overlying limestone roof. To maintain the sandstone pillars integrity and provide a finished concrete surface, curved precast wall panels were erected and contact grouted. Curved precast panels were chosen due to the installation speed and immediate support provided. The panels were analyzed as arch structures and designed to handle two primary loading conditions, pillar support and confinement, and contact grouting pressures. The thrust generated in the wall panels was transmitted through a cast-in-place

Figure 3. Lindbergh station excavation sequence.

523 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-64.qxd

2/18/04

8:02 PM

Page 524

Figure 4. Lindbergh station bench and TBM liner removal at south end of station.

Once the panels were properly aligned, the space between the panel top and limestone roof was filled with shotcrete. A bulkhead was placed on the panel end and the annulus between the panel and rock was contact grouted in two stages. The first grouting stage consisted of lightweight cellular grout being placed to within two to four feet of the limestone roof. The remainder of the annulus was filled with a normal weight pearock grout. A typical sequence of footing installation through placement of normal weight grouting required as little as 11 days.

6 ROOF BEHAVIOR 6.1

top heading first followed by bench removal and wall trimming. Bench removal made rockbolt installation very difficult and expensive, so it was desirable to make all roof rockbolting decisions before bench removal passed the area to be rockbolted. 6.2

Limestone quality

Figure 5 illustrates the location of vertical joints mapped in the underside of cavern roof during station excavation. The figure illustrates that the northern most segment of the station roof had more joints than elsewhere. These joints were also wetter than elsewhere – more than 50 percent of roof water inflow came from the north 10 percent of the roof.

General

From the start, the station design and construction philosophy was to adjust roof rockbolting based on observed conditions and instrumentation. A base level of rockbolting was shown in the contract documents, with additional rockbolt quantities included. As noted above, the station was excavated from north to south,

6.3

524 Copyright © 2004 Taylor & Francis Group plc, London, UK

Roof instrumentation

The station excavation was monitored by a variety of methods. Inclinometers were installed adjacent to the shafts, connection and station prior to the start of excavation. Ten inclinometers were installed, with depths ranging from 30 to 100 feet.

09096-64.qxd

2/18/04

8:02 PM

Page 525

Figure 5. Limestone joint mapping in Lindbergh station.

Figure 6. Lindbergh station roof deflections.

Multi-point extensometers were also widely used. Vertical extensometers were installed from the surface prior to the start of excavation, and also from within the excavation immediately behind the top heading excavation face. The array of eleven vertical extensometers, located at approximately 50-foot spacings on the station centerline, provided bed separation and elevation change data. Grids of PK survey points were installed on the asphalt pavement above the excavation to provide additional roof monitoring. The 127 points allowed determination of roof sag, pillar compression and aided in the determination of when and where additional rock reinforcement was required. 6.4

Roof deflection

Figure 6 illustrates surface deflections along the station centerline, with north to the left and south to the

right. Three distinctive behaviors are apparent: 1. Deflections near the north and south endwalls are small, due to the support provided by the endwalls. 2. Maximum deflections occur from about 100 feet to 200 feet from the north endwall. This area is slightly north of and adjacent to the Transit Center excavation. The additional roof jointing present at this location also influences deflections. 3. Intermediate deflections occur from about 250 feet to 450 feet from the north endwall. Figure 7 illustrates typical results from vertical inclinometer S2, located just outside the west wall of the station, just north of the Transit Center. The sawtooth pattern of deflection results from the formation of rock beams arching to span the excavation. At the cavern wall, beam arching produces horizontal compression at the bottom of the beam and horizontal tension at

525 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-64.qxd

2/18/04

8:02 PM

Page 526

the top of the beam. Hence, the bottom of the beam tends to deflect outward more than the top of the beam. The approximate location and thickness of the rockbeam can be inferred from the sawtooth pattern.

7 WALL BEHAVIOR Stability of the station sidewalls was identified as a potential problem during project design, due to several factors: 1. The wall height of 40-feet was greater than any previous project. 2. Passage of the TBM slightly undercut both sidewalls and also was expected to create a loosened zone above and to the side of the tunnels. 3. The Transit Center excavation forms an acute angle with the station wall. As a result, the contract documents required that the curved precast wall panels be installed within 70 feet of bench removal. In areas of good sandstone, this was increased to 110 feet. Figure 8 shows the installation of curved precast wall panels following the bench and TBM liner removal. During construction, wall sloughing was limited to areas of stress concentration and in the vicinity of a soft seam. The areas of stress concentration were all on the west wall of the station: near the north shaft, and north and south of the Transit Center excavation. The soft seam was between the north shaft and Transit Center. These areas were instrumented and reinforced with shotcrete, rebar, rockbolts. The reinforcing measures were effective in controlling sloughing. 7.1

Eleven single- and multi-point extensometers were installed in the sandstone walls to measure convergence

Figure 7. Limestone inclinometer.

Figure 8. Lindbergh station precast wall panel installation.

526 Copyright © 2004 Taylor & Francis Group plc, London, UK

Wall instrumentation

09096-64.qxd

2/18/04

8:02 PM

Page 527

of the walls and monitor movement on sandstone joints. Vibrating wire transducers were attached to three multi-point wall extensometers, which will allow for long term monitoring of wall movements. Convergence points were also installed in arrays adjacent to the horizontal extensometers to provide additional wall movement data. Based on data collected from the horizontal extensometers and convergence points, the walls showed little or no movement.

educational and transportation facilities across the Twin Cities. With the completion of this project, the technology for designing and constructing large span roofs has been advanced. Caverns have become progressively wider, longer and higher and placed in close proximity to critical facilities and structures.

8 CONCLUSION

The authors gratefully acknowledge the Metropolitan Airport Commission and its staff for permission to publish this paper; the project contractor ObayashiJohnson Brothers Joint Venture for their skill and dedication to a well-constructed project; and the other design team members (HNTB, Hatch Mott MacDonald, Hammel Green & Abrahamson, American Engineering Testing) for fostering a professional design environment.

The Minneapolis-St. Paul metropolitan area geology is ideal for creating large-span Platteville limestone roof/St. Peter sandstone caverns. The widespread presence of these formations will allow for future developments in many areas. The technology used to develop this type of excavation has contributed to the construction of municipal,

ACKNOWLEDGEMENTS

527 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-65.qxd

2/18/04

8:02 PM

Page 529

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Design and impact of the Beacon Hill Station exploratory shaft program Christopher Tattersall Design Manager for Beacon Hill Project, Hatch Mott MacDonald, Mississauga, ON

Tomas Gregor Tunnel Design Specialist, Hatch Mott MacDonald, Mississauga, ON

Michael J. Lehnen Resident Engineer for Beacon Hill Test Shaft, Hatch Mott MacDonald, Seattle, WA

ABSTRACT: To help reduce construction risks for the up-coming Beacon Hill Tunnels & Station contract, the project Owner, Seattle’s Sound Transit, made a decision to construct a test shaft into main soil strata at the location of the future Beacon Hill Station. In addition to evaluating the performance of the ground from the point of the station constructibility using the Sequential Excavation Method (SEM), the construction of this shaft provided an opportunity to evaluate the numerical modeling approach used in the design of the station excavation, prior to the start of the station construction. Observations made during the test program resulted in a fundamental change in the selected construction methodology of the station shafts. The results also indicate that it is feasible to use 3-dimensional structural analysis to predict behavior of SEM-constructed shafts in soft ground.

1 INTRODUCTION The Central Puget Sound Regional Transit Authority (Sound Transit) is proceeding with construction of the Central Link Light Rail Project, a new light rail transit line extending 14 miles southwards towards SeaTac Airport, from Convention Place at the north end of the existing Downtown Seattle Transit Tunnel in the center of the City of Seattle. The one-mile Beacon Hill Tunnels and Station located just south of the downtown area will be the only tunneled portion in this initial segment. In 2000, a joint venture team of Hatch Mott MacDonald and Jacobs Civil Inc., (HMMJ) were awarded a contract for the final design of the Beacon Hill Tunnels (D710) segment of the project. Dr. G. Sauer Corporation was awarded a sub-contract by HMMJ for the design of the large platform and concourse excavations. 2 STATION DESCRIPTION The underground station, shown on Figure 1, will consist of twin shafts and a complex configuration of vehicle, pedestrian and ventilation tunnels. The invert of the platform tunnels will be 156 ft below ground

surface. Platform and connector tunnels will be 550 ft long each and spaced 140 ft apart, center to center. Access from surface to platform level will be provided by four high speed elevators from the West station head-house. 3 GEOLOGICAL CONDITIONS Based on extensive drilling and testing, Shannon & Wilson (S&W) – Sound Transit’s D710 geotechnical consultant – determined that the running tunnels, and station shafts and tunnels, will be constructed through an extremely complex sequence of glacially overridden deposits consisting of very dense and hard clay, silt and sand, gravel and cobbles. Multiple ground water levels were detected in granular deposits, typically due to perched groundwater overlying clay and till units. A large number of geologic units were identified over the depth and length of the project. In order to simplify the large number of unit descriptions and reduce complexity of the geologic profile, Shannon & Wilson grouped geologic units into five major ground types summarized as follows. Soft to Very Stiff Clay and Silt – These deposits generally consist of soft to very stiff, silty clay and clayey silt with variable amounts of sand and gravel

529 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-65.qxd

2/18/04

8:02 PM

Page 530

West Headhouse

East Headhouse

Main Shaft

East LVA Ancillary Shaft East TVA

N/B Platform Tunnel We LVA West Damper Chamber

Connector Tunnel Patform Cross-Adit S/B Platform Tunnel Concourse Cross-Adit Platform Cross-Adit West TVA

Connector Tunnel Station General Arrangements

Figure 1.

Figure 2.

530 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-65.qxd

2/18/04

8:02 PM

Page 531

with some medium dense to dense clayey sand. This unit occurs only in the upper five feet of the shaft. Till and Till-Like Deposits – These deposits consist of a dense to very dense or hard mixture of silt, sand and gravel and varying amounts of clay. These units range in consistency from extremely dense to very hard to a consistency of very soft rock or lean concrete. Cobbles and boulders are common in these units and water-bearing silt and sand lenses may be present, but typically are not hydraulically connected to the regional groundwater regime. Very Dense Sand and Gravel – These deposits consist of non-glacial (interglacial) fluvial deposits of very dense sand, gravelly sand, and sandy gravel and lenses of gravelly cobbles. Without dewatering or ground improvement, these granular soils will flow into the shaft or tunnel excavations or shaft bottoms will boil or blowout as the excavation proceeds below the groundwater level. Dewatered granular soils will still ravel and run unless fully supported. Pre-support may be required to stabilize these soils and to limit ground losses. Very Dense Silt and Fine Sand – These deposits consist of non-glacial (interglacial) lacustrine deposits of very dense to hard silt, fine sandy silt, silty fine sand, and clayey silt. The soils that have some cohesion will behave similarly to, but somewhat poorer than the Very Stiff to Hard Clay described below. Without dewatering or ground improvement, the cohesionless portions of these soils will become unstable and will flow into the excavation, or the shaft bottoms will boil or blowout as the excavation proceeds below the groundwater level. These deposits will be difficult to dewater or grout. Very Stiff to Hard Clay – These deposits consist of glaciolacustrine deposits of very stiff to hard clayey silt and silty clay. The Very Stiff to Hard Clay has been a good tunneling and shaft excavation material above or below the groundwater table, in previous tunnel and shaft excavations throughout the Seattle area. The hard consistency and cohesive nature of the clays make them relatively easy to excavate with conventional soil excavation equipment. When unfractured, their intact strength properties promote good standup time in tunnel faces and in open cuts. Except for seams or lenses of sand and silt, the clay will not yield appreciable quantities of water and are stable for short periods in the presence of water. Slickensided fractures and shear zones are expected to be encountered in the Very Stiff to Hard Clay, especially in the high plasticity clays. All of these ground types are likely to contain variable amounts of cobbles and boulders. 4 RATIONALE FOR TEST SHAFT PROGRAM Cross-correlation between boreholes was found to be quite challenging, indicating a very complex

stratigraphy over the station site. In particular, the sand and gravel, and silt units, both potentially problematic soils, were difficult to isolate. The results of in-situ and sample tests also raised questions as to the behavior of these particular soils during tunneling. Similarly, it was perceived that the magnitude of this uncertainty could lead to the perception of a high level of risk being associated with the design; risk that would lead to high bid prices for the construction. To mitigate this, Sound Transit agreed to implement a test shaft program – located within the station Main Shaft – with the following goals: 1. Explore the stratigraphic nature of the ground. 2. Verify mechanical properties of the soil units. 3. Verify the accuracy of the numerical modeling to be used for the station design. 4. Observe the behavior of the soil units during SEM construction. 5. Test effectiveness of support measures proposed, including dewatering, permeation grouting, spiles and grouted forepoling. 6. Demonstrate to the contracting community that this method of excavation was feasible, and provide a measure of rates of progress that were achievable.

5 TEST SHAFT DESIGN The shaft was designed 18 feet in excavated diameter and 148 feet deep, with a 10” shotcrete lining. The shaft behavior prediction and the lining forces were obtained using the state-of-the-art geotechnical software FLAC3D 2.1 by ITASCA Consulting Group Inc. The Mohr–Coulomb failure criterion was used to evaluate stresses in the soil elements. It was assumed that the lining was water permeable and that dewatering and the construction process would remove water pore pressures from the soil within the limits of the model. Two three-dimensional models were created. Each FLAC3D model consisted of approximately 60,000 brick shaped zones, representing the soil, and over 3800 structural members, representing the shaft and adit lining. The geometry of the models was first generated. Then the grid was assigned appropriate material properties, loaded with initial loads, and brought to equilibrium. For each excavation step, a portion of the soil representing the shaft or tunnel excavation step was removed and the structural elements representing the lining were placed into the model. Once a new equilibrium was reached, forces and displacements were saved and the process was repeated for the next excavation step. The first model had the shaft and adit lining modelled as linear elastic plates of the actual lining thickness. However, since the linear elastic model does not allow for release of tensile stresses in concrete by

531 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-65.qxd

2/18/04

8:02 PM

Page 532

cracking, high tension developed in some portions of the lining model, mainly due to local bending. However, based on the geometry of the structure, it is reasonable to assume that the membrane hoop compression forces in the lining are the primary loadcarrying mechanisms. The correctness of this assumption has been demonstrated in practice for the primary lining where, from experience, very small amounts of reinforcement have been used in the shotcrete. In order to model this assumption a modification to the lining thickness and elastic modulus was made for the second model. The model lining thickness tm was assigned as t/10, where t is the lining actual stiffness. The modulus of elasticity of the lining in the model was taken as Em  10E, where E was the modulus of the elasticity of concrete. Therefore while the axial load stiffness Em tm  Et was unchanged, whereas the flexural stiffness was significantly reduced i.e. Em tm3  Et3/100 and resulted in very small bending moments and true membrane loads in the lining. The results proved that force equilibrium could be found without excessive deformations with this greatly reduced flexural stiffness of the lining. While the first model was used for estimation of the ground displacements, the second model was used for the analysis of the lining structural strength. In-house developed spreadsheets were used to obtain bending moment-axial (M–N) concrete section capacity curves. Pairs of axial forces and moments are required to establish the relationship of load to the concrete capacity curve. However, the directions of principal moments and principal axial forces do not generally coincide. Therefore, once the directions of the principal forces were established the moments and axial forces were paired up according to which two of the principal directions were the closest. Figure 3 shows the shaft elevation and its relationship to the future Main Shaft of the station. 6 TEST SHAFT CONSTRUCTION The Frank Coluccio Construction Company (FCCC) of Seattle, WA was the pre-qualified low bidder, and was awarded the C710.04 – Beacon Hill Exploratory Test Shaft & Tunnels construction contract for $1.76 million. Notice to Proceed was issued effective February 19, 2003. S&W – who provided geotechnical instrumentation monitoring and geological mapping for the Test Shaft Program – installed surface leveling points, inclinometer casings, piezometers, inductance probe rings and a pump test well in the Lower Sand Layer while FCCC was mobilizing and performing some required preparatory work. Based on information in the contract geotechnical report, and the preliminary results of the pumping test,

FCCC proposed the installation of two additional surface wells – one screened in the Upper Sand Layer and one in the Lower Sand Layer to supplement the one installed by S&W – to facilitate dewatering of these water-bearing strata. This was intended to provide a flexible, phased approach, with additional wells being added as necessary as construction progressed. Also, this would allow for shaft construction to commence as scheduled on April 14, 2003, since the installation of any more than two wells would have delayed shaft work, due to the constraints of the small site. FCCC had approached ST and HMMJ about using an auger drill to expedite construction in the competent upper clay and till materials. Although this was a fullface excavation – rather than the specified SEM – this proposal was accepted based on the fact that the ground would be exposed for evaluation and mapping prior to installation of the lining. This method was utilized in the upper portion of the shaft above the Upper Sand Layer, and again between the Upper and Lower Sand Layers; equating to approximately half the 18-foot excavated diameter portion of the shaft. SEM was used in the remainder of the 18-foot excavated diameter portion of the shaft. Prior to reaching the Upper Sand Layer, it was decided to temporarily suspend shaft excavation and install an array of vacuum dewatering lances to supplement the one well pumping in this stratum. Piezometric data had indicated that groundwater levels had dropped, but not to the full depth of the 12-foot thick sand layer. Eleven well points, equally spaced around the circumference, were installed and operated from inside the shaft. This system proved effective, and excavation subsequently proceeded smoothly through the Upper Sand Layer. Piezometric data indicated that the Lower Sand Layer, which was believed to be approximately 10 feet thick, had about 25 feet of water head on it. The two surface wells that had been pumping on this layer had lowered the groundwater level by 15 feet; however, the remaining head within the sand was not responding to the surface wells. The decision was made to again install an array of vacuum wells to supplement the surface wells. 20 well points were installed circumferentially around the inside shaft perimeter, and an additional five above the planned upper adit. Drawdown of the groundwater did occur after operation of this system. However, almost immediately after excavation resumed, the variability of the Lower Sand Layer was revealed. Many silt and clay lenses and layers were contained within the sand, which trapped perched water and impeded drainage. Progress through this layer was very slow, with various combinations of additional vacuum well points, steel sheeting and grouting employed to stabilize the sand. Based on risk, schedule and budget considerations – in particular to complete the shaft in time to incorporate

532 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-65.qxd

2/18/04

8:03 PM

Page 533

Figure 3.

any changes deemed necessary to the station design – the decision was made in late August to forego construction of the two adits and complete the shaft to depth utilizing a smaller diameter auger drill and casing

pipes to evaluate the underlying silt layer. The investigation of the silt layer, and the uncertainty of the behavior thereof, were prime considerations during planning of the Test Shaft program.

533 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-65.qxd

2/18/04

8:03 PM

Page 534

The final auger drilling operation consisted of the initial installation by vibratory hammer of an 8-foot diameter, 24-foot long casing through the remainder of the sand to ensure sidewall stability, and the subsequent auger drilling and 6-foot diameter casing advancement, also by means of the vibratory hammer. The auger drilling and casing were advanced in approximately 5-foot lengths, and manned inspection and mapping via a basket lowered into the augered hole were performed where ground conditions dictated. This was completed on August 29, 2003. Finishing works consisting of grouting around the 6-foot diameter casing, removal of the 8-foot diameter casing, installation of stairs and windows and the shaft cover, were subsequently performed by FCCC. Demobilization was completed on October 8, 2003.

Actual vs. Predicted Inclinometer for SB351 - ID#4 09/05/2003 B-Axis/Y'-Axis Displacement (in) -0.3 0

-0.2

-0.1

0

0.1

0.2

0.3

20 40 60

Depth (ft)

80 100

Depth of Shaft Excavation

120 140 160

7 GROUND DEFORMATION AND COMPARISON WITH MODEL

180

The test shaft enabled the design team to compare the ground displacement prediction obtained from the 3-D numerical model with the field measured ground displacements. This way the design team could verify whether the ground parameters used for the numerical modeling of the station construction were reasonable representation of the ground properties. Five inclinometers were installed in the vicinity of the test shaft by S&W. These were located primarily to measure ground movement due to construction of the two test adits radiating from the shaft; that is at some distance from the shaft wall. These trial adits were not constructed and the displacements measured by these instruments were very small, since ground displacements tend to decrease with increasing distance from the shaft. Figure 2 shows the plan location of the inclinometers with respect to the test shaft and planned adits. Figures 4 through 7 show the comparison of the measured and the predicted displacements for two of the inclinometers. It is important to note that the nominal sensitivity of the inclinometers is 0.1 inch/100 ft of length in the primary direction (A or X direction) and 0.3 inch/100 ft of length in the other direction (B or Y direction). The magnitude of the measured ground displacement values was less than the published precision tolerance values for these instruments. This can be clearly seen from the values shown for Inclinometer SB-353, Figures 6 and 7. The B axis is tangential to the shaft and the displacements are expected to be zero. The measurements show some small, non-zero displacements. Inclinometer SB-351, (Figures 4 and 5) is closest to the shaft and as expected it showed the largest displacement values. Although the measured values are in the same order of the magnitude as the equipment

220

200

Actual

Figure 4.

Actual vs. Predicted Inclinometer for SB351 - ID#4 09/05/2003

-0.3 0

A-Axis/X'-Axis Displacement (in) -0.2 -0.1 0 0.1 0.2

20 40 60

Depth (ft)

80 100

Depth of Shaft Excavation

120 140 160 180 200 220 Actual

Figure 5.

534 Copyright © 2004 Taylor & Francis Group plc, London, UK

Predicted

Predicted

0.3

09096-65.qxd

2/18/04

8:03 PM

Page 535

were embedded into the walls of the shaft at six different levels. The relative movements of the points within each arrays were measured at different stages of the shaft excavation using tape extensometers, and the displacements were compared to the predicted displacements from the model. The comparison was performed at two of these arrays, at elevations 212 and 192 ft. At elevation 192 ft the difference between the comparable measured and the predicted displacements were, at the last reading, less than 0.005 on the absolute value of about 0.05. At elevation 212 the difference between the comparable measured and the predicted displacements were, at the last reading, less than 0.015, on the absolute value of about 0.05. It should be noted that the predicted displacement values represent array displacements at the time when equilibrium of the soil structure interaction system was reached. This means the displacements stop changing with time and the shaft lining is fully converged. The results of the comparison show that though the measured displacements were close to the published sensitivity of the measuring equipment, they confirm that the assumed geotechnical values used for the analysis of the underground structures reasonably reflect the actual ground behavior.

Actual vs. Predicted Inclinometer for SB353 - ID#3 09/05/2003 -0.3 0

B-Axis/X-Axis Displacement (in) -0.2 -0.1 0 0.1 0.2

0.3

20 40 60

Depth (ft)

80

Depth of Shaft Excavation

100 120 140 160 180 200 220 Actual

Predicted

Figure 6.

Actual vs. Predicted Inclinometer for SB353 - ID#3 09/05/2003 -0.3 0

A-Axis/Y-Axis Displacement (in) -0.2 -0.1 0 0.1 0.2

8 DESIGN REFINEMENTS RESULTING FROM TEST SHAFT LESSONS LEARNED

0.3

20 40 60

Depth (ft)

80 100

Depth of Shaft Excavation

120 140 160 180 200 220 Actual

Predicted

Figure 7.

accuracy, the measured displacement values are of a similar magnitude and shape as the predicted values. In the design of the Test Shaft, a series of three permanent anchors, arranged into a triangular array,

Figure 3 shows the mapped ground stratigraphy found along the walls of the shaft. It can be seen that the soil layering is very complex with a large number of lenses of noncohesive material within the layers of cohesive material. This lack of ground layer continuity explains the ineffectiveness of dewatering implemented during construction in the Lower Sand Layer. It is also obvious that it will be virtually impossible to clearly define the extent of each material over the footprint of the station. As a result of the test shaft construction experience, the SEM method of construction for the station shafts was replaced by a slurry wall construction technique. This technique is much less sensitive to the encountered type of ground and does not require dewatering to stabilize the sand layers. Because the Main Shaft construction is on the critical path, it was felt that the risk could be substantially reduced by constructing this shaft utilizing slurry walls. Also, a ground improvement program by jet grouting at the locations of shaft breakouts and over the length of the tunnels believed to be in non-cohesive soils will be implemented.The remainder of the station tunnel construction will proceed using SEM since the majority of the tunneling is in clay and till layers. The test shaft confirmed that

535 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-65.qxd

2/18/04

8:03 PM

Page 536

these types of soils performed satisfactorily when excavated. 9 CONCLUSIONS The test shaft program proved that in complex soil conditions a test shaft provides an excellent means of obtaining information on soil behavior and stratigraphy. The previously stated objectives of exploring the stratigraphic nature of the ground, verifying the mechanical properties of the soil units, and observing the behavior of the soils during SEM construction were achieved, although a modified approach was implemented for investigation of the bottom third of the shaft, in the area of the silt layer. The effectiveness of the support measures in the various soil units was determined, except for the grouted forepoles that were scheduledto be installed above the two test adits, which were not constructed. In particular, the difficulty of

dewatering the complex Lower Sand Layer was discovered. This would have been necessary for the Main Shaft to be successfully constructed by SEM. The test shaft construction provided an opportunity for a comparison of the predicted ground movement and the field measured movements, in order to verify correctness of the assumed field ground mass properties. Despite the fact that the two adits were not constructed as planned, the results obtained indicate that the assumed ground properties are a reasonable representation of the field values, and they increased confidence in the predictions made by the numerical analysis for the station project. Finally, it is concluded that SEM is a viable approach to construction of the station tunnels in the complex Beacon Hill stratigraphy. However, to mitigate risks associated with construction of the station shafts through the variable and unstable sand layers, slurry walls will be utilized for these structures, with ground improvement provided at breakout locations.

536 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-66.qxd

2/18/04

3:03 PM

Page 537

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Comparison of the predicted behavior of the Manhattan TBM launch shaft with the observed data, East Side Access Project, New York V. Nasri STV Incorporated, New York

W.S. Lee & J. Rice Parsons Brinckerhoff, New York

ABSTRACT: The Bellmouth excavation is one of the first packages constructed by East Side Access (ESA), which will allow Long Island Rail Road (LIRR) to access to Grand Central Terminal. This excavated site will become the main access point for the construction of the ESA Manhattan segment including bored tunnels and major carven construction in rock. It is also the starting point for the main cut-and-cover construction in the Queens segment. This open-cut is excavated within a congested urban area, abutted by a major thoroughfare, and both subway and elevated transit structures. Existing contaminated groundwater plumes are present in the rail yard near the excavation site. This paper discusses the design to address these concerns and the issues that arose during construction. The instrumentation program implemented to monitor the effectiveness of the design and the data recorded are also discussed.

1 INTRODUCTION The LIRR is the busiest suburban commuter railroad in the U.S., serving about 269,000 passengers on 740 trains a day in the year 2000. Its capacity to bring additional passengers to Penn Station is severely restricted due to physical limitation at Penn Station and the anticipated ridership growth of the other station occupants, Amtrak and New Jersey Transit. This space restriction, and the final eastside destination of a large number of riders, provided impetus for an East Side Terminal for LIRR (Figure 1). East Side Access will provide 24 peak-hour trains into the new terminal increasing LIRR service to Manhattan by about 109,000 passengers. The planning of ESA dates back to the 1960’s when the Metropolitan Transportation Authority, the parent company of LIRR, authorized New York City Transit (NYCT) to construct 8,400 ft of tunnel under the East River. The upper level of this two-level, four-track tunnel, completed in 1970’s, are being used by NYCT subway trains. The two lower level tracks will be used for ESA trains. One major challenge of the ESA project is to minimize impacts on the overlying/surrounding operating facilities, which include NYCT subway and elevated lines, Amtrak’s Sunnyside Yard, LIRR Main Line tracks

(also used by Amtrak’s North East Corridor service) and roadway bridges crossing the alignment. Another challenge is not to disturb the groundwater regime, which may in turn disturb known contaminant plumes within Amtrak’s property. The Bellmouth Excavation (Construction Contract CQ026) is one the first ESA construction packages constructed to facilitate the construction of the Manhattan segment. 2 PROJECT DESCRIPTION The ESA tunnel construction starts with an open-cut excavation adjacent to the existing 63rd Street Tunnel Bellmouth in Queens (Figure 2). The excavation is located south of the recently completed NYCT 63rd Street Connector tunnel. After completion of Contract CQ026 the open-cut excavation and surrounding site will be transferred to successive Manhattan tunnel contractors for personnel and equipment access, material delivery, and muck removal through the existing 63rd Street Tunnel. The excavation will remain open until the Manhattan tunnel and cavern construction is completed. It is expected that the excavation will remain open for 8 to 10 years. The final Manhattan tunnel and cavern contract will construct the tunnel segment

537 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-66.qxd

2/18/04

3:03 PM

Page 538

Manhattan

Project Site

within the open-cut, connecting to the completed segments east of Northern Blvd. In accordance with ESA commitment to the New York State Department of Environmental Conservation (NYSDEC), dewatering is not allowed outside of the project area. Slurry walls were selected to provide support of excavation and groundwater control. The slurry walls are keyed and sealed into bedrock to cut-off groundwater inflow minimizing disturbance to the groundwater regime. The new slurry walls are connected to the south side of existing slurry walls constructed during the NYCT 63rd Street Connector Project. The new and existing slurry walls, an above-grade NYCT ventilation structure to the west, and a new sheet pile wall with effective grout sealing to the east will form the groundwater cut-off system. The boundary of this cut-off system is shown on Figure 2.

3 GROUND CONDITIONS 3.1

Queens

Figure 1. Location plan.

The subsurface information at the project site is derived from 15 geotechnical borings drilled at the construction area and from other borings drilled in adjacent ESA areas. Standard split barrel samples and undisturbed samples were collected by fixed piston and Denison/Pitcher samplers. Laboratory testing was performed on selected soil and rock samples.

Figure 2. General excavation support plan.

538 Copyright © 2004 Taylor & Francis Group plc, London, UK

Site investigation

09096-66.qxd

3.2

2/18/04

3:03 PM

Page 539

Site geology

The project site is located near the physiographic province boundary between the Manhattan Prong and the Atlantic Coastal Plain. The site is underlain by metamorphic rock, which is covered by Pleistocene glacial and inter-glacial deposits as well as postglacial deposits. Bedrock consists mainly of gneisses and schists with pegmatite sills and dikes scattered throughout the rock mass. The most common rock type is granitic gneiss comprised principally of quartz, biotite and feldspar. The last glacial advance into the Metropolitan New York area is the source of most of the soil deposits and the present surface morphology. 3.3

Interpreted ground conditions

The glacial deposits above the bedrock can generally be divided into three groups, mixed glacial deposits, glacial till, and outwash/reworked till deposits. Each group is subdivided into several strata. Stratification is generally complex, and significant variations in the thickness and location of the individual units are common. Boundaries between strata are not clearly defined in many cases due to considerable inter-layering of the glacial materials, particularly in the mixed glacial deposits. Under such environment, different processes of deposition occur during cyclical periods of advance and retreat of the ice front. Prior deposits are re-worked and new materials are deposited. Overlying the glacial deposits is a layer of man made fill. The stratigraphy of the project area beginning at the ground surface is presented below. Stratum 1 – Fill: It consists of a heterogeneous mixture of sands, with silts, gravels, cobbles and miscellaneous debris such as brick fragments, wood, concrete and rubble. Some of the fill is believed to be excavated glacial materials from strata 2 and 3. Strata 2 to 4 – Mixed Glacial Deposits: The group contains glacial deposits, lacustrine deposits and glacio-fluvial deposits. Because of the nature of the depositional process, these soils exhibit a high degree of inter layering and spatial variation. Stratum 2 is predominantly granular with nonplastic to low plastic fine materials. Stratum 3 contains more fines than stratum 2. Stratum 4 contains significant amount of fines, mainly non-plastic silts to clays of low plasticity and occasionally sand and gravel. Stratum 5 – Glacial Till/ Reworked Till: These soils consist of heterogeneous mixture of sand, silt and gravel, mostly without a cohesive binder. A large number of boulders were encountered in this layer. Mixed with the till are outwash materials that are predominantly sand with little fines.

Figure 3. Inferred geological profile along slurry wall.

Stratum 6 – Decomposed Rock: This stratum consists of very stiff to hard silts, clays and sands. The materials exhibit relict structures of the parent rock. Decomposed rock was found in localized areas, generally thin but it was found to be 8m thick in one area. Stratum 7 – Bedrock: Bedrock at the site is predominantly fine to coarse grained, unweathered to moderately weathered strong to very strong gneiss and schistose gneiss with RQD generally between 50 and 100 percent. The stratigraphy along the slurry wall is shown in Figure 3. 3.4

The groundwater at the site is within an unconfined aquifer formed in the glacial deposits directly overlying the bedrock. The primary source of the regional groundwater is precipitation, which averages about 48 inches a year. Actual precipitation reaching the water table is less than 50 percent because of surface development and drainage characteristics. The regional groundwater flow is northwest toward the East River. However, the groundwater from the site flows primarily southwest towards the Newton Creek.

4 SUPPORT DESIGN 4.1

Site constraints

The CQ026 work site is in a tight urban setting closely surrounded by dense, mostly commercial properties. Approximately 25 feet to the northwest of the excavation is a NYCT ventilation building. The 41st Avenue roadway is approximately 25 feet to the west of the proposed excavation. Also to the northwest is the five-story Newcomers High school 210 feet away. The closest residential area is approximately 500 feet to the west. To the southwest is the one-story Bank of New York building. At the closest point this building is approximately 65 feet from the slurry

539 Copyright © 2004 Taylor & Francis Group plc, London, UK

Hydrogeology

09096-66.qxd

2/18/04

3:03 PM

Page 540

walls. About 80 feet to 120 feet to the southeast of the slurry walls and adjacent to the Bank of New York, is a two-story building. Beyond 120 feet of the excavation within the same Lot, there is a 10-story building. Immediately to the east lies Northern Boulevard. Besides the NYCT IND subway and BMT elevated lines within its limits, Northern Boulevard itself is a multilane roadway that extends from eastern Queens to the Queensboro Bridge. The roadway is a major east-west route in Queens carrying heavy traffic volumes in each direction during peak hours. The NYCT IND Queens Blvd Line subway and BMT Astoria Line elevated structure at Northern Blvd are located east of the open-cut excavation. The bottom of the proposed excavation is approximately 50 feet below the invert of the subway structure. The subway structure is approximately 72 feet wide and 25 feet high. The elevated structure is supported by concrete piers cast integral with the subway structure supported on piles. The two closest piers are about 9 feet away from the slurry wall along Northern Blvd. The west curb line of Northern Blvd and the west wall of the subway structure are approximately 17 feet away from the slurry wall. The open-cut excavation exposes the southern edge of the recently completed NYCT 63rd Street Connector T1/T2 subway connecting the upper level of the 63rd Street Tunnel with the Queens Blvd Line. The western end of the T1/T2 subway structure within the open-cut excavation sits atop of the existing LIRR structure, which in turn is resting on bedrock. The eastern most part of the T1/T2 subway structure within the open-cut rests on backfill. The excavation line is approximately 12 feet below the base slab of the T1/T2 structure. This results in an unbalanced earth and hydrostatic pressures on the T1/T2 structure during excavation. A support system maintains the stability of this structure while continuous dewatering for the duration of ESA construction relieves the hydrostatic pressure. The construction impact on these important facilities must be kept to a minimum to maintain normal transit and vehicular services levels throughout construction. Because of the close proximity of the NYCT transit structures and the existence of potential contaminants outside of the project site, dewatering outside of excavation is not allowed. Therefore, to facilitate excavation below groundwater, a watertight excavation wall system is required. Based on analyses, it has been concluded that if the base of a slurry wall is keyed into bedrock with a permeability, k, less than 1 105 cm/s or in rock with a higher k, but grouted to achieve similar permeabilities, groundwater drawdown can be limited to two feet from the initial groundwater levels, with seasonal variations taking into account. With a drawdown less than two feet, the impact to the groundwater regime is expected to be minimal.

4.2

Slurry walls are to be used as the primary excavation support and groundwater cutoff. The slurry walls are to be constructed around the southern perimeter of the excavation and connected to the existing 63rd Street Connector slurry walls, along with grouting and a sheet pile wall to the east to form a continuous watertight retaining structure (bathtub). The slurry walls are to be keyed approximately six inches into rock. The minimum height of the slurry wall is approximately 50 feet at location A. The wall height increases towards the eastern end of the excavation and reaches about 85 feet at location L. Borings are to be drilled 15 feet into rock at 10-foot spacing below the slurry walls for rock quality/permeability determination. For k values higher than 1 105 cm/s determined by Packer testing, cement grouting is to be performed to create an effective groundwater barrier in the rock mass. The stability of the slurry walls are to be maintained by tiebacks and bracing. To maximize open space for future contractors working in this area, the use of internal bracing is restricted to areas shown on Figure 2. A finite element program, PLAXIS, was used for the analysis of all support walls, both slurry and sheet pile walls. To avoid significant movements of Northern Blvd and thereby the transit structures, at rest earth pressures are used for the design of Wall O-L. This wall has a thickness of four feet. Other slurry walls have a thickness of three feet where active pressures were applied. Final design of the slurry walls and support of excavation was completed by the contractor. The final slurry wall design used steel reinforcing bars and soldier piles for end stops. The maximum/allowable deflection of the slurry wall adjacent to Northern Blvd is not to exceed 0.75 inch at any depth and at any stage of construction of the wall. Rock anchors are to be used as the slurry wall tiebacks due to high loads. The maximum vertical spacing between tiebacks is limited to 15 feet to avoid excessive wall deflection. At its closest point, the slurry wall is about 28 feet from the Bank of New York property to the south (parking lot area). To limit construction easements within the Bank of New York Property, the tiebacks are to be installed at 45 degrees from horizontal. The tiebacks are double corrosion protected and have a minimum of 15 feet for both the stressing length and the bonded length. Most of the tiebacks are located below the groundwater table and all tiebacks were field-tested. 4.3

Sheet piles, tiebacks/bracing and grout sealing

Sheet piles are installed to support the soil (fill) above the existing subway structure during excavation. The

540 Copyright © 2004 Taylor & Francis Group plc, London, UK

Slurry walls and tiebacks

09096-66.qxd

2/18/04

3:03 PM

Page 541

stability of the sheet piles is to be maintained by a combination of internal bracing and tiebacks. For the sheet pile walls to be installed directly above the existing T1/T2 subway, a vibratory hammer is to be used for driving sheet piles to a precise elevation. Tiebacks are to be used for Sheet Pile F-G to provide a working area in front of the LIRR bulkhead free of obstruction for future contracts. To avoid inducing additional stresses acting on the T1/T2 subway north wall, the bond lengths of the tiebacks were installed on the north side of the existing northern 63rd Street Connector slurry wall. This involved coring through the approximately 2.5 feet thick existing reinforced concrete slurry wall. At the eastern end of the excavation, a watertight sheet pile wall has to be placed over the T1/T2 subway structure roof to prevent groundwater flow into the excavation area. Because of the fines content (from 2 percent to as much as 20 percent) of the existing backfill, chemical grouting is required to seal the interfaces between sheet pile and the existing structure, and between the sheet piles and the existing slurry walls. Chemical grouting on both sides of the T1/T2 subway adjacent to the sheet pile wall is also required. All chemical grouting has to be completed prior to commencement of dewatering operations. Since the open excavation will remain open for more than 8 years, the life span of the grout is to be at least 12 years. 4.4

NYCT tunnel support

The project requires excavation south of the NYCT T1/T2 subway to rock. This creates an unbalanced force from the north side and may cause instability of the subway structure. Except at tunnel segment F-G, internal bracing is to be used to stabilize the subway structure. These struts are to support the tunnel structure only. The maximum allowable vertical and lateral movement for all NYCT subway structures is 0.5 inch. The review level is at 0.25 inch. A section showing the sheet pile wall, T1/T2 subway, and the new and existing slurry walls is presented in Figure 4.

Figure 4. Section at Sta. 117245.

5 INSTRUMENTATION PROGRAM A geotechnical instrumentation program was implemented to monitor ground and structure movements, peak particle velocity during rock blasting, and groundwater drawdown. Instrumentation includes:

• • • • • • •

6 CONSTRUCTION Construction ran into many unforeseen problems from the onset. A noise barrier wall along 41st Avenue, 29th Street, and the west end of 40th Road was required as the first contract element per the Final Environmental Impact Statement. The wall required more treated timber than locally available impacting the contractor’s schedule. Abandoned spread footings were grabbed and spun by the augur during wall foundation shaft drilling in several locations, collapsing the shafts and requiring time-consuming backfilling around the completed piers. The buy America clause in the contract posed a problem for sheet pile procurement. None of the local suppliers had the required pile size in stock and the quantity did not warrant a new rolling by steel mills. A sufficient quantity of used material with acceptable defects was eventually procured. The material was inspected by the contract design team and accepted for use. The fill material on top and north of the T1/T2 subway contained boulders, grouted material, fines, and steel pile sections. This created delays during sheet pile driving and even damaged some sections of sheet piles. The numerous boulders caused delays during slurry wall construction. Nested boulders, including boulders nested along the top of the bedrock were common. The rock probing along the slurry wall alignment specified by the contract was used to aid in determining between nested boulders and actual bedrock. The bedrock drops in some locations with a slope as great as 2 on 1. This lead to isolated locations of slurry wall panels terminating two feet above the rock. Several secondary panels were out of alignment with the primary panels. This can be attributed to crane operators inexperienced in slurry wall construction which requires a significant amount of “feel” when it comes to boulder removal, cobbles, panel alignment and sidewall cleaning. The slurry wall excavation difficulties resulted in some quality control problems such

541 Copyright © 2004 Taylor & Francis Group plc, London, UK

Ground surface monitoring points Inclinometer (in soil) Inclinometer (in slurry wall) Probe extensometer Structure monitoring point Tieback load cell Automated total station (within NYCT tunnels)

09096-66.qxd

2/18/04

3:03 PM

Page 542

as the panel misalignment already mentioned, numerous surface voids with exposed rebar, slurry wall/rock interface voids where only grouted soil was found as excavation exposed the bottom of the slurry wall, and gaps in the closure panels between the new slurry walls and existing 63rd Street Connector slurry walls. All of these defects impacted the contractor’s schedule. On multiple occasions the excavated material from the slurry wall trenching operation included detectable levels of volatile organic compounds that were not found during the environmental testing phase of design. While not hazardous material, the VOC containing material still requires separate storage, handling, and disposal that is made more difficult by the lack of an established pay item. Before the start of production dewatering groundwater sampling results were submitted by the contractor to the New York State Department of Environmental Conservation as required to procure a Long Island Well Permit. The samples were presented improperly, incorrectly indicating a pH level of 11. This resulted in the NYSDEC requiring a pH treatment system to balance water before discharge into the combined sewer. As production dewatering removed water from the entire site, the combined effluent pH never reached the NYSCE maximum allowable level of 10. During dewatering operations a bacteriological slime was encountered near the two deep wells used to lower the groundwater behind the T1/T2 subway. This slime aggressively clogged the well pumps requiring multiple pump replacement and several well flushing with acid. While iron oxide deposits have been encountered on similar projects, this material was only 0.8% iron oxide, with 22% Magnesium Hydroxide and 30% Organic Material. This material caused reduced pumping capacities for the deep wells, requiring the contractor to institute continuous pumping operation for several weeks to compensate. The reduced effectiveness of the dewatering allowed groundwater to flow into the open-cut, hampering excavation activities. Tieback drilling encountered difficulties when drilling in the local till material. The dense sand contains minimal fines and many tiebacks are completely under the groundwater table. The clean sand washed into the excavation through the drill holes in the slurry wall at several locations, leading to localized sinkholes and groundwater drawdown. Eventually, the tieback driller had to leave the drill casings in the holes and procure more robust seals to withstand the hydrostatic pressure. The impact of boulders was also underestimated requiring more drilling time and tool replacement than anticipated. Despite the difficulties during drilling, all of the tiebacks were successfully load tested with no rejections. Boulders were encountered during excavation that measured up to 14 feet in a horizontal plane. These

boulders were approximately 5 feet thick; the same boulder dimension range recorded in the project borings at the site. During rock blasting operations, the proximity of the operating NYCT T1/T2 subway hampered the size and charge configuration of the shots. Close coordination between the Resident Engineer and NYCT helped to overcome these delays. 7 ANALYSIS OF THE SLURRY WALLS For the slurry wall system, a finite element analysis using specialized geotechnical software PLAXIS was performed. The effects of stage construction including excavation and support installation, soil plasticity, soil wall interface, and water flow on the soil wall interaction were considered using PLAXIS. Some results for the tied-back slurry wall OP (Figure 2) are given here. The geometry of the excavation, its support system, surface traffic surcharge, and different soil layers are shown in Figure 5. Four rows of rock anchors are installed at an angle of 45° at the following distances from top of the excavation: 13, 30, 47, and 58 ft. The spacing above was selected to get relatively close values for the diaphragm wall moment at support levels and also to harmonize the support reactions. This results in a more optimized design for wall and its lateral support system. The anchors will be grouted into the bedrock. The diameter and the length of the grout hole in the bedrock are determined based on the prestressed force of the anchors and the allowable bond stress between the bedrock and the grouting. The distance between adjacent columns of anchors is 12.5 ft. 24 tendons of 15 mm diameter prestressing steel strands (ASTM A416, Grade 270) are used for each anchor. The anchor cross section is 5.2 in2 and its ultimate strength is 1406 kip. The first anchor from ground surface is prestressed to 475, the second to 600, the third to 800 and the last one to 783 kip, which are lower than 60 percent of their ultimate strength. As expected, stress concentration occurs around the grout bodies and in the lower edge of the wall, therefore local mesh refinement is applied in these areas. The model is extended sufficiently in lateral and vertical directions to minimize the boundary effects. Roller boundary condition is assumed in lateral and bottom boundaries and the excavation is accomplished in five stages. In reality, there is a complex three-dimensional state of stress around the grout body. Although the precise state of stress and interaction with the soil cannot be modeled with this 2D model, it is possible in this way to estimate the stress distribution, the deformations and the stability of the structure on a global level. The diaphragm wall is modeled as a beam and the interface elements are used around the beam to model

542 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-66.qxd

2/18/04

3:03 PM

Page 543

Displacement of top of the Slurry wall PQ (75 ft from Point Q) Displacement, in

1.20 Observed Predicted

1.00 0.80 0.60 0.40 0.20 0.00 0

50 100 150 Days - Starting from April 29, 2003

200

Figure 5. Analysis model. Figure 6. Comparison between observed and predicted displacements.

Table 1. Properties of soil layers. Layer

, kcf

E, ksf





C, ksf

K0

K ft/day

Fill Marine deposit 2 El. 298-316 Marine deposit 3 El. 290-298 Marine deposit 4 El. 268-290 Till El. 258-268 Till El.254-258 Rock

0.125 0.125 0.125 0.125 0.135 0.135 0.17

325 761 891 1476 1767 1861 99638

0.3 0.3 0.3 0.3 0.3 0.3 0.3

30 30 30 0 36 36 0

0 0 0 1.6 0 0 108

0.5 0.5 0.5 0.85 0.7 0.7 0.5

25.51 25.23 5.1 1.56 26.93 26.93 0.07

soil-structure interaction effects. The beam elements used to model the walls are fully permeable. Therefore, the interfaces around the wall must be used to block the flow through the wall for groundwater calculations. The properties of soil layers used in the analysis are listed in Table 1. The initial water pressure is generated on the basis of a horizontal general phreatic line and then the initial earth pressure field is generated using the K0 values. The calculation consists of 9 consecutive stages of soil excavation and tieback installation. The excavation stages include the de-watering on the excavation side and involve a steady state groundwater flow analysis to calculate the new water pressure distribution. Closed flow boundary condition is used for bottom and also for the vertical side in front of the excavation. On the vertical side behind the excavation, the ground water head remains constant at its initial value. Mohr-Coulomb elastic perfectly plastic constitutive law is used to simulate the plastic behavior of the materials. Figure 6 compares the predicted and observed displacements of top of the slurry wall PQ at a point approximately 75 ft from Point Q. Unfortunately, the exact tieback prestressing date was not recorded. Therefore, this has been approximately evaluated from the measured data based on the change in displacement direction. Relatively good agreement or a kind of acceptable trend can be seen to a certain point, but

later the curves diverge likely due to the sinkholes created by some construction problems in the area. 8 INSTRUMENTATION DATA Partly due to construction problems encountered, most of the movement data collected are not at the level of quality normally expected. More importantly, however the significant facilities that required protection do not appear to have moved. It is believed that the surrounding buildings, roadways, subway and elevated structures did not move more than 0.25 inch. An interesting observation due to the sinkholes developed behind the slurry walls PQ. At a point on top of the slurry wall approximately 25 feet from Point Q, the wall moved inward as the excavation progressed. It reached a peak of about 0.5 inch before the first tieback was placed. The top of wall then started to move in the opposite direction, toward the soil. The top of wall continued to move toward the soil after the first sinkhole was observed (Figure 7). It has since then stabilized, after approximately 120 days, close to 1.2 inches in the opposite direction. Figure 8 presents the typical movements measured within the NYCT IND subway, using the automated total station system. Little movement was recorded in the vertical direction and in the directions of excavation. Quite a bit of

543 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-66.qxd

2/18/04

3:03 PM

Page 544

Displacement, in

Displacement of PQ Slurry Wall (25 ft from Point Q) 0.60 0.40 0.20 0.00 -0.20 -0.40 -0.60 -0.80 -1.00 -1.20 -1.40

May 20, First Tieback installed June 27, First Sinkhole Observed

0

50

100

150

200

Days - Starting from April 29, 2003

Figure 7. Top of slurry wall movement. Figure 9. Bellmouth shaft during construction.

Figure 9 presents a picture of the shaft under construction. 9 CONCLUSIONS Various aspects of the design and construction of the ESA Queens TBM shaft were discussed in detail. The difficulties encountered during construction of this structure and also its instrumentation and monitoring plans were explained. Following this experience, the authors recommend that in order to make better use of monitoring data, the design engineers should be directly involved in implementation, reading and interpretation processes and should be allowed to supervise the entire instrumentation and monitoring activities.

Figure 8. Typical IND subway movements.

noise was recorded in the direction of the tunnel axis. This is due to the orientation of the line of sight – there are only 2 to 5 degrees between some of the reference points along the tunnel axis, making the calculations somewhat inaccurate. Also note that some of the baseline movements were not set at zero. Groundwater drawdown outside of the bathtub is generally less than 2 feet, indicating the slurry walls served as an effective groundwater cut-off system. There were occasions where large drawdowns were recorded but these were due to tieback installation difficulties and localized slurry wall defects and were temporary.

REFERENCES Nasri, V. and Schabib, J. 2002. Comparison of different Alternatives used for Design of the Launch Wall of the Queens Cut and Cover Tunnel, East Side Access Project, New York. International Conference of Urban Underground Space: a Resource for Cities, Turin, Italy, November 14–16, 2002. Nasri, V., Jafari, R. and Wone M. 2003. East Side Access Project in New York, hard rock and soft ground tunneling. 12th Panamerican Conference on Soil Mechanics and Geotechnical Engineering, MIT Cambridge, MA, June 22–25, 2003.

544 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-67.qxd

2/18/04

3:04 PM

Page 545

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Drop shafts – selection principals Jozef F. Zurawski & Edmond Petrossian Dawn Engineering, Inc., Hazlet, New Jersey, USA

ABSTRACT: Many interceptor and Combined Sewer Overflow (CSO) projects in the US require deep tunnel installations for conveyance and/or storage of effluent to minimize overflows of untreated sewers. Near surface sewers must be connected to these deep facilities requiring special drop arrangements. This paper provides overview of hydraulic and construction issues that designers must consider when making the selection of the drop structure type. It also discusses major design parameters of systems, which must be addressed before the final selection of drop structure geometry can be made. Hydraulics of drop arrangements, ground conditions, tunnel flow conditions, drop capacities, air release and odor control are examined as some of the factors entering the final design selections. Innovative adaptation of proven designs in the US and Japan are also discussed.

1 INTRODUCTION In the continual effort to keep our waters clean, we need storage and/or conveyance systems to minimize CSOs. Since these facilities must intercept most of existing sewers, they are compelled to be constructed deeper underground. In some cases, the storage can be provided at near surface where most of connections can be collected by gravity without any major changes in system. However in most cases, especially in large municipalities, the system ends up as a deep tunnel system, 50, 100 or even 200 feet below ground surface. In such situations, the conveyance of the effluent from the existing near surface facility to a deep CSO tunnel will require a special structure to drop the flows. These structures are referred to as drop structure or drop shaft. The function of the drop structure is to carry the effluent flow to the deep facility under controlled hydraulic conditions, controlled air entrainment, without damage to the facility, with long life expectancy and low maintenance cost.

2 PRELIMINARY EVALUATION 2.1

Design flows and system hydraulics

The usual facility planning effort for a new CSO system starts with computer hydraulic modeling of the existing sewers. This includes calibration of the computer

model to depict the actual hydraulic conditions, identify problem areas, and CSO events. In order to reduce CSO discharges to receiving waters, improvements or additions are modeled into the system. This may or may not include an additional treatment facility. Since most of the existing treatment plants are not utilized to their full capacity during dry weather flows, an overflow storage facility for later treatment often becomes the logical choice. Computer hydraulic modeling should also be performed to identify surges in storage tunnels. The results of surge analyses are used to determine the high water elevations in shafts, necessity to use larger shafts or tunnels to prevent overflow from shafts, and to determine impact forces on system components due to water momentum. However, it is almost impossible to determine by computer modeling the optimum locations for air release points. Once the need for storage is identified, two choices exist: near or at surface storage such as ponds or large shallow underground tanks, or deep storage tunnel system with pumping facilities. In most cases, due to area and/or grade restrictions, a deep conveyance/storage tunnel is the only answer. This in turn dictates the need for drop structures. The system computer hydraulic modeling will establish the flow rates at connections (drop structures) as well as the flow characteristics within the storage/conveyance tunnels. The hydrographs of the CSO tunnel system and the drop structure flow will establish the air removal requirement.

545 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-67.qxd

2/18/04

3:04 PM

Page 546

Figure 2. Typical plunge pool drop.

Figure 1. Inflow hydrograph and tunnel filling rate.

2.2

Air removal

As the effluent is dropped 50, 100 or 200 feet, air will be entrained into the deep storage facility. The excess air without proper removal will deprive the facility of storage space, cause air pressure buildup and release of odors. Improperly designed systems have been known for blowing manhole covers many feet into the air. Figure 1 shows the inflow hydrograph and the filling rate of the tunnel and shaft. The critical zone for air removal is when the water rises above 70% on the tunnel depth. For lower depths, there may be enough space at the tunnel crown for the air to be dragged to the next shaft for venting. For the case shown on Figure 1, the water will fill the tunnel while the drop shaft still contributes substantial volume. As a result deaeration is needed. Some drop structures rely on mixing air in the shaft to enhance energy dissipation as well as to minimize cavitation. However, these systems require larger deaeration chambers and air vents (vent shafts) to release air before the flow enters into the tunnel. 2.3

Energy dissipation

Originally, energy dissipation was provided by plunge drops and baffle (cascading) drops. The plunge drop relied on a large pool of effluent to absorb the drop energy with a restricted outlet sized to control the level of the drop pool. To create a permanent pool, often the pool bottom was

Figure 3. Typical baffle drop.

constructed below the tunnel invert elevation. See Figure 2. The baffle drop relied on short cascading effluent drops down to the tunnel in a series of steps. See Figure 3. Later “boot chambers” (Chicago, Rochester) or circular chambers (Milwaukee) were developed by hydraulic modeling to provide answers for energy dissipation and air release control. The modeling helped in identifying internal pressures, impact forces and geometry of the structures. More recently, vortex drops were developed to reduce the impact of flow as it hits the bottom of the shaft. 2.4

Once the hydraulic conditions of the storage/ conveyance tunnel and drop performance is identified, a preliminary selection of the drop structure can be

546 Copyright © 2004 Taylor & Francis Group plc, London, UK

Initial geometry selection

09096-67.qxd

2/18/04

3:04 PM

Page 547

made. Issues such as air release at desired locations, flow rate, and height of the drop will be the first guidelines for the selection process. The geometry should be selected from proven modeled information for the flow rates falling within well-documented ranges. They often fall into the following categories: a. b. c. d.

Simple drop Jet flow drop Plunge pool drop Vortex drop (with or without deaeration chamber) (Milwaukee, Cleveland) e. Radial drop with plunge pool (boot type: Chicago, Rochester) f. Vortex inlet with helicoidal drop (Japan). There are other vortex types such as spiral, etc. with various inlet configurations. It must be recognized that the geometries are not based on empirical formulas and as a result, modification of the modeled geometry or extrapolation for flow rates outside of the modeled limit must be avoided. Any such modification should be made with the assistance of an experienced hydraulic laboratory. One common drop arrangement not listed above is a baffle drop. The writers consider this type of a drop a poor selection. No proven modeling record exists for baffle drops. It entrains air, generates hydrogen sulfide, it is difficult to maintain and is almost impossible to inspect. 2.5

Hydraulic modeling

Figure 4. Vortex drop with deaeration in tunnel.

connections to the tunnel may require special segments, internal secondary lining, and ground improvements (i.e., freezing, jet grouting). In such cases, cost comparisons are required to select the least costly alternative. These may include: – Use of helicoidal drop that consists of a single connection to tunnel. – Incorporating drop arrangement and air release into tunnel shaft excavation. Figure 4 shows such an arrangement, in this case used for a shaft located at the upstream end of a tunnel. 2.7

Most of proven geometries are based on hydraulic laboratory modeling. Most recognized types with excellent air removal performance are types “d”, “e” “f ” (see above). Should a modification be needed by specific design or site constraints, hydraulic modeling must be used to develop geometry of the modified structure, before making any changes. Examples: Using tunnel as a deaeration chamber (Cleveland – scheduled for modeling) Multiple connections to a single helicoidal drop shaft (helicoidal – Japan).

Since the CSO storage/conveyance system is designed for certain rainfall occurrence, diversion structures and overflows must be provided to handle the excess effluent volume for larger storms. The excess flow is disposed of via overflows provided in the existing sewer/interceptor, or diversion structures, or overflow outlets at shafts. The diversion or overflow structure is an integral part of the overall system, and usually its design dictates the maximum quantity of flow allowed into the drop structure or tunnel system. 2.8

2.6

Ground conditions

One of the major factors in decision making on drop types is the geotechnical setting of the CSO system. For example, a vortex drop with a deaeration chamber may prove to be economical in rock or good soft ground conditions, but it may become costly in poor soft ground conditions. In poor soil conditions, costly ground improvements will drive up the cost of deaeration chamber and air release piping. Tunnels can be successfully constructed in poor ground conditions using one pass concrete segments. However,

System evaluation

A cost-effective system evaluation should be made to determine the optimum configuration and dimensions of drop structures/shafts/tunnel. In some cases it is beneficial to construct a large diameter tunnel with simpler drop structures that provide energy dissipation only (i.e., more air entrainment is allowed). In other cases, it is more cost effective to use a smaller diameter tunnel with more sophisticated drop structures (i.e., less air entrainment is allowed). Such cost comparisons should be prepared by persons experienced in underground construction cost estimating.

547 Copyright © 2004 Taylor & Francis Group plc, London, UK

Diversion structures

09096-67.qxd

2/18/04

3:04 PM

Page 548

3 TYPES OF DROP STRUCTURES 3.1

Simple drop

Simple drops consist of vertical pipes discharging into a horizontal channel at the tunnel level. The flow enters, without guidance, into the vertical pipe through a simple tee connection. The flow drops in the vertical pipe, clinging to the pipe wall and in the process entrains some air into the water. The drop energy is dissipated in the horizontal channel before being discharged into the tunnel. See Figure 5. The design is based on extrapolation of parameters used in designing vertical plumbing stacks in high rises (Dawson & Kalinske 1937, Steele 1977, Wyly & Eaton 1592). In tunnel projects, the vertical drops are sized for a 25% full flow condition. Tests made on building stacks have indicated that higher flows can cause considerable noise and vibration. Air is entrained into the vertical drop, a portion of which is vented through the pipe open core area and the remaining portion is entrained into the tunnel. Air entrainment may be as high as 50% of the effluent volume. The diameter of simple drop vertical pipes commonly varies between 15-inches and 24-inches. The 24-inch diameter with a discharge capacity of 29 cfs is generally the maximum size used for tunnel connections. In general, simple drops are cost effective, and are used for small flows, where the entrained air would not substantially affect the capacity or performance of the receiving tunnel or storage system. 3.2

– They usually do not require a separate deaeration chamber. – They have a compact design and as such are easier to locate within congested project sites than more complex drop structure arrangements, especially for smaller flow rates. – They are relatively small, inexpensive, easy to construct, to operate and maintain. In fact, a completed jet flow drop structure resembles a round, square or rectangular manhole. 3.3

Free jet drop

Free jet drop is an option for drops with less than 40-foot depths (Design Guidelines-Rochester 1981,

Figure 5. Typical simple drop.

“Green Book”). Free jet flow occurs when the flow from the near-surface incoming sewer is allowed to flow freely into the structure without being obstructed by walls, or other appurtenances. If the flow were allowed to impact a chamber wall or other obstruction, the dispersion of the flow would cause more air to be mixed in. It has been found that for less than 40foot drops the jet flow does not disperse, and its relatively small surface area makes it difficult for air to mix with the water; i.e., air entrainment is minimized (15% to 25%). See Figure 6. Sometimes, outlets are provided with restricting features to cause water to back up within the structure, and create a water cushion for the incoming flow. The pool artificially reduces the depth of the drop, requiring shorter lengths for the structure. In addition, air entrainment is reduced, since the water falls a shorter distance. The water cushion provides protection to the bottom slab from impact of falling water and heavy objects, thus minimizing structure vibration. The advantages of free jet drops are as follows:

There are different versions of plunge flow drops. They usually consist of an elbow inlet directing the

Figure 6. Typical free jet flow.

548 Copyright © 2004 Taylor & Francis Group plc, London, UK

Plunge flow drop

09096-67.qxd

2/18/04

3:04 PM

Page 549

flow radially through a drop shaft. The flow passes through a deaeration chamber before entering the tunnel. Numerous modeling has been performed on plunge flow drops by the University of Minnesota’s St. Anthony Falls, Hydraulic Laboratory for projects in Chicago and Rochester, NY (Anderson & Dahlin 1975, Dahlin & Wetzel 1982). A plunge flow drop example is shown in Figure 7. To improve energy dissipation, Plunge Flow Drops are designed to entrain a large quantity of air. A deaeration chamber located at the tunnel level vents the entrained air before the flow enters into the tunnel. Energy is dissipated in a pool created at the bottom of the drop. Usually the pool is created behind a weir located upstream of the adit to tunnel. The weir includes an opening at the bottom to allow the pool to drain during low flows. Plunge flow drops occupy large volumes, they are odd shaped, and are very costly to construct, especially in soft ground. 3.4

Vortex drops

The inlets to vortex drops are designed to impart an angular motion to the flow before it enters the drop pipe, creating a helical path for the flow down the pipe. Several types of inlet arrangements have been investigated and used for vortex drops, such as spiral, scroll, siphonic, tangential, and helicoidal. Compared to radial inlets of plunge flow drops, vortex drops entrain appreciably less air.

Figure 7. Typical plunge flow drop.

This presentation addresses only the tangential and helicoidal types of vortex drop structures, which are more commonly used in tunnel projects; the former in the US and the latter mainly in Japan. 3.4.1 Tangential inlets Figure 8 depicts a conventional vortex drop that includes an approach channel, a tangential inlet, a drop pipe, and a deaeration chamber. Numerous modeling has been performed on conventional vortex drop structures by the University of Iowa, Iowa Institute of Hydraulic Research for projects in Milwaukee and Cleveland (Jain & Kennedy 1983, 1984). The approach channel should be long enough so that flows are laterally uniform as it reaches the tangential inlet and be used for metering purposes. The bottom slope and side slope of the tangential inlet is predetermined by modeling; therefore, other angles or interpolated angles must not be used to design tangential inlets. Although there is less air entrained compared to a plunge flow drop, it is high enough to warrant deaeration. The deaeration chamber reduces air concentration down to between 1% and 2% before the flow enters the tunnel. An air vent located on top of the deaeration chamber releases entrained air back into the atmosphere or it is recirculated back to the tangential inlet. The location of the air vent is also determined by physical modeling. 3.4.2 Helicoidal drop Numerous modeling has been performed on helicoidal drops, by the University of Iowa, Iowa Institute of Hydraulic Research (Jain et al. 1993, Kennedy & Jain 1987). Currently there are no helicoidal drops installed in the US. However, helicoidal drops have been used in Japan since 1996, and manufacturers have already

Figure 8. Typical tangential inlet, vortex drop and deaeration chamber.

549 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-67.qxd

2/18/04

3:04 PM

Page 550

started fabricating pre-packaged units. These are made of stainless steel or Fiberglass Reinforced Plastic Mortar (FRPM). It is interesting to note, that these prepackaged units are manufactured for smaller flows also, where commonly simple drops are used. We anticipate that in the future helicoidal drops will be commonly used in the US also, especially in congested urban areas, where construction of tangential inlets interferes with existing utilities or space is at a premium. Different types of helicoidal drops have been modeled and manufactured. Some include continuous ramps from top to bottom of the drop pipe; some others consist of three segments: upper and lower segments with helicoidal ramps, and the middle segment with no helicoidal ramps. Figure 9 depicts a helicoidal drop that includes an inlet pipe, an upper segment with 2 ramps, a middle segment with no ramps, a lower segment with 6 ramps, and an outlet pipe. The flow is spun around the drop pipe in the upper ramps; some air is entrained during the drop, but is deaerated by the lower ramps. For shorter drops, the upper and lower ramps may overlap, and therefore a continuous ramp is provided for these cases. Some units are designed with one upper ramp, and less than 6 lower ramps, etc. If designed properly, helicoidal drops can reduce air concentrations down to between 1% and 2%.

c. Acceptable level of air entrainment. d. Number of connections at each shaft and the possibility to combine some or all of them into a single drop structure. e. That large on-line shafts will accept drop structure(s); or off-line shafts should be constructed and be connected to the tunnel with adits. f. That adequate easements are available to accommodate drop structures and their connections. g. The types of drop structures and drop arrangements that will assure that system will perform as it is intended and is cost effective. h. At this level additional hydraulic laboratory modeling may be identified to be completed before the final design is undertaken. This would include structures that have unique connection arrangements or fall outside of flow rate limits previously modeled.

4 DECISION TIME

The preliminary evaluation will usually narrow down the field of viable drop shafts to a couple of alternatives. In some cases, there is only one viable alternative i.e., a free jet drop for less than a 40 foot drop with allowance for some air entrainment; or simple drop for small flows and allowance for some air entrainment. In other cases, for larger flows and deeper drops, a choice should be made between a plunge flow drop, a tangential inlet, and a helicoidal drop. They should be ranked based on a range of parameters that affect their construction cost and operational effectiveness.

At the end of preliminary system design the following issues should have been determined:

5 AIR RELEASE AND ODOR CONTROL

a. Diameter of tunnel, plan and profile. b. Diameters of on-line shafts.

It is almost impossible to identify the shaft(s) where significant air releases would occur. In urban areas, air release and associated odor can be a major problem. To minimize odor, the air released from deaeration chambers are recirculated back into the drop shaft. However, since air is dragged into the system through the non-submerged cross section of the tributary interceptors, there always will be excess air that is released by shafts, somewhere along the tunnel. The air drag will cause excess air release even at low flow conditions. The odors are sometimes controlled by air treatment facilities. In some cases, the systems are designed with odor treatment facilities; or provisions are made to install such facilities in the future as their need becomes obvious as the system is put in operation.

6 CONCLUSION

Figure 9. Typical helicoidal drop.

The design of a drop structure is not carried in the vacuum and there is no “cookbook” approach for its design. Different system components such as tunnel, extent of acceptable air removal, number of

550 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-67.qxd

2/18/04

3:04 PM

Page 551

connections, depth of drop, ground condition, diversion structures, etc., influence shaft geometry and design. Several revisions are usually made in system components, including shafts, until an optimum configuration is selected. However, well-planned and designed systems work well. In some cases it will take 10-20-30 years to complete the system, which should operate without major maintenance for the next 50 to 100 years. During the development and construction period, several consultants and contractors may be involved in different phases of the project. The originally designed level of performance for air removal, surge and flow criteria must be maintained until system completion. When different degrees of air removal are designed by different engineers, the system may not perform as intended, and/or odor, capacity and air release problems start to appear in many places. Therefore, it is advisable that a master plan be prepared for the system and future connections be made in accordance with the master plan requirements, and where the system is broken down to several design sections, engineers must follow the master plan requirements. When additions are made in the future, the system should be re-evaluated in order not to alter the operation of the system, and ensure that the new work will improve the operation rather than hinder and destabilize it. It would also be advisable to update the computer model of the system every 5 or 10 years to account for new connections and flow changes. ACKNOWLEDGEMENTS

members: Larry Colella, Jennie Shareshian and Allen Parker, Jr. for their contribution to this presentation.

REFERENCES Anderson, A.G. & Dahlin, W.Q. 1975. Dropshafts for the Tunnel and Reservoir Plan, Metropolitan Sanitary District of Greater Chicago, St. Anthony Falls Hydraulic Laboratory. Dahlin, W.Q. & Wetzel, J.M. 1982. Rochester Dropshafts Model Studies, St. Anthony Falls Hydraulic Laboratory. Dawson, F.M. & Kalinske, A.A. 1937. Report on Hydraulics and Pneumatics of Plumbing Drainage Systems-I, U. of Iowa. Design Criteria & Guidelines for the Combined Sewer Overflow Abatement Program, Rochester Pure Water District, 1981, “Green Book”. Jain, S.C., Hayden, W.S. & Ali, M.A.M. 1993. Novel Truncated – Ramp Drop Structures, Iowa Institute of Hydraulic Research. Jain, S.C. & Kennedy, J.F. 1983. Vortex-Flow Drop Structures for the Milwaukee Metropolitan Sewerage District Inline Storage System, Iowa Institute of Hydraulic Research. Jain, J.C. & Kennedy, J.F. 1984. Hydraulic Design of SWI Drop Structures, City of Cleveland, Iowa Institute of Hydraulic Research. Kennedy, J.F. & Jain, S.C. 1987. The Helicoidal Ramp Drop Shaft, Iowa Institute of Hydraulic Research. Wyly, R.S. & Eaton, H.N. 1952. Capacities of Plumbing Stacks in Buildings, U.S. Dept. of Commerce, Building Materials and Structures Report 132.

The Authors would like to acknowledge the help and assistance of the Dawn Engineering, Inc. staff

551 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-68.qxd

2/18/04

3:05 PM

Page 553

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Stability evaluation and numerical modeling Exchange Place Station Improvements J.F. Lupo Golder Associates Inc., Lakewood, Colorado, USA

M.F. McNeilly Golder Associates Inc., Newark, New Jersey, USA

ABSTRACT: The Port Authority of New York and New Jersey (PANYNJ) initiated the Downtown Restoration Program (DRP) to restore commuter train service to the former Port Authority Trans-Hudson (PATH) World Trade Center Station following the tragic events of September 11, 2001. Part of this effort, the Exchange Place Station Improvements Project, consisted of the design and construction of new platform extensions, multiple cross-over tunnels, and converting the Exchange Place PATH Station from a “through” station to a “terminal” station configuration. Assuring the stability of the project’s new platform extensions and crossover tunnels presented several technical challenges: construction in low-strength Manhattan Schist rock, shallow ground cover with large excavated spans, close proximity of existing building foundations, and unknown construction details/records of the original (early 1900s) tunnel structures. This paper presents the approach used for the stability analyses in the design process, the development of model parameters and boundary conditions, and model verification.

1 INTRODUCTION

2 PROJECT DESCRIPTION

The Exchange Place Station Improvements Project is part of the Port Authority of New York and New Jersey (PANYNJ) Downtown Restoration Program (DRP) to restore commuter train service to the former Port Authority Trans-Hudson (PATH) World Trade Center Station, following the tragic events of September 11, 2001. In general, the project consisted of the design and construction of new platform extensions and multiple cross-over tunnels, converting the Exchange Place Station from a “through” station to a “terminal” station configuration and restoring commuter service to the Jersey City, New Jersey waterfront district. Construction of the proposed cross-over tunnels required development of up to 18-m (60 ft) wide spans with thin crown pillars, 9 m (30 ft) average. In addition, thin rib pillars, 3 m (10 ft), were to be developed in areas that are overlain by existing 5- to 16story buildings. This paper presents the approach used to assess the stability of the project’s new platform extensions and cross-over tunnels, the development of model parameters and boundary conditions, and model verification.

The Exchange Place Station Improvement Project was one of three components of PANYNJ’s overall DRP (the other two being the rehabilitation of the twin Hudson River tunnels and the reconstruction of the former WTC PATH Station). See Figure 1. The project involved substantial improvements to the existing Exchange Place PATH Station located in

Figure 1. Downtown restoration program layout.

553 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-68.qxd

2/18/04

3:05 PM

Page 554

beneath the overburden and moderate to good quality rock at tunnel level. Average rock cover thicknesses were about 9 m (30 ft). 3.1

Figure 2. Existing tunnel and cross-over tunnel plan.

Jersey City, Hudson County, New Jersey, and required that the station be re-opened by July 2003. The station and connecting tunnels were constructed during the late 1890s and early 1900s, and the station became operational sometime around 1908. The station is connected to 5 tunnel structures providing service to tracks E, F, G, H and L. At the east end of the station, Tunnels E and F provide service under the Hudson River to New York City. At the west end of the station, Tunnels E and F provide service to Hoboken, New Jersey and midtown Manhattan; Tunnels G and H provide service to Jersey City, Harrison, and Newark, New Jersey. Tunnel L, located at the west end of the station, was used as a lay-up track. As part of the Exchange Place Station Improvements Project, a total of 6 new cross-over tunnels were to be developed to connect the existing subway tunnels and allow the station to re-open by July 2003. Development of these new cross-over tunnels was accomplished by a combination of excavation and bulkhead infills. See Figure 2 for the layout of the project. 3 SUBSURFACE CONDITIONS As part of the project, a geotechnical investigation program was conducted to evaluate the underlying soil and rock mass characteristics, which included the drilling of oriented and non-oriented boreholes, geomechanical logging of collected rock cores, coring of the existing tunnel concrete liner, tunnel-level structural mapping (after removal of pre-existing concrete liners), and material testing. In general, this geotechnical investigation program established the following:

• • •

Overburden ranges from approximately 4.5 to 7.5 m (15 to 25 ft) thick, and consists of fill and native silty-clay, sand, silt and gravel materials. Groundwater was observed in the overburden at depths of 2.5 to 3.5 m (8 to 12 ft) below ground surface. Bedrock underlying the site is comprised of Manhattan Schist. Rock, and rock quality tends to vary with depth, with poorer quality rock directly

The dominant structural feature observed in the rock mass is foliation. Foliation was observed to be gently to moderately dipping and ranged from sub-horizontal (less than 5 degrees) to as much as 40 degrees. The direction of foliation varies from northeast to northwest. Sub-vertical, vertical and steeply dipping fractures were uncommon in the coreholes, and no measurable dip directions and angles were recorded for any steep joint planes observed in these coreholes. However, a series of steep joints were observed and measured at the tunnel level, as tunnel excavation advanced.

4 STABILITY ANALYSES The general approach for the project’s requisite stability analyses consisted of evaluating the stress conditions and assessing the effectiveness of proposed rock reinforcement elements in two-dimensions for selected design cross-sections. In particular, these stability analyses were conducted using the finite element program Phase2 (Rocscience 2002, version 5.0) and the distinct element code UDEC (Itasca 1998, version 3.0). The Phase2 model was used to evaluate the state of stress and changes in the stress field around the existing tunnels and new cross-over tunnels by modeling the rock mass as a continuum. The purpose of these models was to evaluate the potential for failure in excavation crowns and thin rib pillars. In addition, the models were used to assess stress changes within the existing tunnel concrete liner. The UDEC model was used to assess the effectiveness of proposed rock reinforcement elements on crown stability in tunnel sections with large spans, 17 m (55 ft), with shallow cover in a highly jointed and foliated rock mass. 4.1

Material properties

Representative material properties were developed for overburden, Manhattan Schist, and existing tunnel liner concrete materials. See Table 1 for a summary of the material properties used in the stability analyses. Values presented in Table 1 include both mean and minimum values, based on a review of available data. Stability analyses were completed for both mean and low values to assess the variability in stress and stability over the potential range of material properties.

554 Copyright © 2004 Taylor & Francis Group plc, London, UK

Rock discontinuities

09096-68.qxd

2/18/04

3:05 PM

Page 555

associated material properties were assumed based on typical values.

Table 1. Summary of analysis parameters. Model parameters Description Rock Mass Classification Rock Quality Designation (percent) Rock Mass Rating Geologic Strength Index Q-Rating Rock Mass Classification from RMR Description of Rock Mass Rock Mass Characteristics Unit Weight (kilograms/cubic meter) Unconfined Compressive Strength (megapascals) Modulus of Elasticity (megapascals) Poisson Ratio Hoek-Brown m Parameter Hoek-Brown s Parameter Joint Friction Angle (degrees) Joint Cohesion (megapascals) Joint Shear Stiffness (megapascals/ meter) Joint Normal Stiffness (megapascals/ meter) Overburden Characteristics Unit Weight (kilograms/cubic meter) Internal Friction Angle (degrees) Cohesion (megapascals) Modulus of Elasticity (megapascals) Poisson Ratio Water Conditions Water Elevation within Overburden (meters) Height of Water above Tunnel Invert Existing Concrete Lining Characteristics Unit Weight (kilograms/cubic meter) Unconfined Compressive Strength (megapascals) Modulus of Elasticity (megapascals) Concrete Backfill Unit Weight (kilograms/cubic meter) Unconfined Compressive Strength (megapascals) Modulus of Elasticity (megapascals) Rock Bolt Parameters Bar No. Grade (MPa) Pre-tension Load (kiloNewtons) Length (meters) Spacing (meters)

Mean values

Low values

90 64 74 15 II Good

40 33 43 2.7 IV Poor

2,723 34.5

2,403 13.8

2,068 0.3 3.95 0.056 23 0.05 22.9

1,379 0.3 1.31 0.0018

29.7

1,922 30 0 149.6 0.3

1,602

91.44 0

6

2,402 31

2,243 20.7

2,758

2,068

4.1.2 Manhattan Schist Material properties of Manhattan Schist were derived from observations from rock cores, core logging, and laboratory uniaxial compression (UCS) and direct shear tests. Corehole log data provided values for Rock Quality Designation (RQD) and Total Core Recovery (TCR). Collected cores were visually inspected to assess joint spacing and joint roughness. The methods presented in Barton (1976) were used to assess the rock joint properties. Laboratory test data and core log information were used to derive a Rock Mass Rating (RMR) for the Manhattan Schist using the methods presented by Bieniawski (1976). Calculated RMR values for the schist ranged between 33 (poor rock) to 64 (good rock), depending on whether mean or low values are used. Corresponding Rock Tunneling Quality Index (Q-rating, after Barton et al. 1974) ranged between 2.7 and 15. Rock mass strength parameters were derived using the Hoek-Brown strength criteria (Hoek et al. 1998). Observations of the Manhattan Schist rock core indicate that the rock mass is highly foliated, with numerous joints along foliation planes. Some joints were also noted that crossed foliation planes. To assess the influence of foliation and jointing on the stability of the cross-cuts, core samples were tested in direct shear along existing joints/foliation planes to develop joint shear strength parameters. Each core sample was tested in direct shear both parallel and perpendicular to the foliation plane to assess affects of observed crenulations on shear strength. Direct shear test results indicated that joint shear strength was not dependent on orientation to foliation. 4.1.3 Concrete liner and backfill Material properties for the existing tunnel concrete liners were derived from UCS testing on core samples taken from the liner. Material properties for the new concrete backfill used in the cross-over tunnel design were developed using American Concrete Institute (ACI) empirical equations.

2,402 17.2 1,930 9 517.1 133.4 3.7–4.6 1.5

4.1.1 Overburden Overburden soils generally consist of silty-clay, siltysand and gravel deposits. Since overburden soils play a relatively minor role in tunnel and cross-cut stability,

4.1.4 Ground support The stability analyses included resin grouted rock bolts for ground support. Rock bolts were modeled as No. 9 (Grade 75) bolts, and were modeled with a pretension set at 40% of the bar yield strength. Rock bolt lengths vary depending on excavated span lengths and type of support (pre-support versus final support). The finished tunnel support system consisted of steel fiber reinforced shotcrete (SFRS) with lattice girders spaced on 1.5-m (5-ft) centers. However, SFRS

555 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-68.qxd

2/18/04

3:05 PM

Page 556

and lattice girder support was not included in these models to assess stability prior to finishing the tunnels. This approach should yield conservative results as the tunnel stability is enhanced with the application of SFRS and lattice girders. 4.1.5 Field stress conditions Gravitational field stresses were used to induce stress into the model. Site-specific stress ratios [Ko (horizontal-to-vertical)] were not available, so the stability analyses included a range of Ko varying from 0.5 to 2.0. In addition to gravitational stresses, surface loads from the overlying multi-story office buildings were incorporated. Foundation loads from these buildings were incorporated into the model as either mat foundations with or without pile support, depending on the building design. Building loads were applied in the models as a distributed load along the footprint of the foundation. 4.2

Phase2 stress models

The modeling approach for the Phase2 stress analyses consisted of developing a model to approximate state of stress conditions prior to excavation of the new tunnel cross-overs. After these initial conditions were developed, sequential excavations and/or concrete backfills were introduced into the model to simulate anticipated construction sequences, and rock bolts were added during construction in areas requiring ground support. Figure 3 presents the final stress analysis results from one design cross-section, which shows the development of a 17-m (55-ft) span cross-over tunnel. The stress results are presented in terms of strength factor, which represents a ratio of rock strength to stress. As shown in Figure 3, a “halo” of low stress ratio developed in the crown of the large span, and rock bolting was added to the crown to provide support. In addition, rock bolts were added in the thin rib pillars, where areas of high stress occurred. Results from these stress analyses indicate that the imposed incremental change in stress due to cross-over tunnel excavation is relatively small compared to the in situ rock strength. Compressive stresses that develop in the rib pillars are moderate, but are well supported by adjacent backfills and/or excavation support. 4.3

Figure 3. Phase2 stress analysis output.

UDEC analyses

Two-dimensional UDEC analyses were conducted on select design sections to evaluate the stability of large excavated spans with shallow cover in a highly jointed and foliated rock mass. Conditions evaluated with the UDEC model consisted of the construction of a 17-m (55-ft) wide span. Rock mass jointing consisted of

Figure 4. UDEC model output.

3- to 4.5-m (10- to 15-ft) sub-vertical joint spacing and 0.6-m (2-ft) sub-horizontal joint spacing. Joint spacings were selected based on data from the core logs and tunnel-level joint mapping. Figure 4 presents output from the UDEC model for a 17-m (55-ft) span with rock bolt support. The UDEC model results indicate that the rock bolt support design for the new cross-over tunnels will provide a stable crown until lattice girders and SFRS support elements are applied.

5 CONCLUSIONS Design of the new cross-over tunnels for the Exchange Place Station Improvements Project required the development of up to 18-m (60-ft) wide excavated spans in areas of relatively shallow rock cover. Design analyses considered span development with thin crown pillars, 9 m (30 ft) average, in areas overlain by existing 5- to 16-story buildings. To support the cross-over tunnel design effort, stability analyses were conducted using both continuum and discrete element models, and these models were used to assess tunnel stability. The models considered and incorporated anticipated construction sequencing (excavation, backfilling, bolting) to evaluate changes in stress fields and to identify critically stressed areas.

556 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-68.qxd

2/18/04

3:05 PM

Page 557

Results of these stability models were used to verify rock bolt design, construction sequencing, and define pre-support. ACKNOWLEDGEMENTS The authors wish to acknowledge that the success of this project was attributed to genuine partnership between the owner and the various designers and contractors, and all are congratulated for re-opening of the Exchange Place Station on-schedule. The authors also wish to thank PANYNJ for providing the artwork for Figures 1 and 2, reviewing this paper, and allowing publication for this conference. Lastly, the authors would like to thank their colleagues who worked on the Exchange Place Improvements Project and reviewed and provided comments to this paper.

REFERENCES Barton, N.R. 1976. The Shear strength of rock and rock joints, Int. J. Rock Mech. Min. Sci. & Geomech. Abstr. vol. 13: 255–279. Barton, N.R., R. Lien, and J. Lunde. 1974. Engineering classification of rock masses for the design of tunnel support, Rock Mech. 6(4): 189–239. Bieniawski, Z.T. 1976. The Geomechanics classification in rock engineering design In Proc. 4th Int. Congress on Rock Mech., Montreax (ISRM), vol. 2: 41–48. Rotterdam: Balkema. Hoek, E., P.K. Kaiser, and W.F. Bawden. 1998. Support of underground excavations in hard rock. Rotterdam: Balkema. Itasca. 1998. UDEC User’s Manual, Minneapolis MN Itasca Consulting Group. Rocscience. 2002. Phase2 User’s Guide. Toronto, Ont.: Rocscience Inc.

557 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-69.qxd

2/18/04

3:05 PM

Page 559

Session 4, Track 4 Conventional underground construction

Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-69.qxd

2/18/04

3:05 PM

Page 561

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Tunnel and shaft construction for the Pingston Hydro Project Bruce Downing & Zach Vorvis Golder Associates, Burnaby, B.C., Canada

Graham Rawlings Graham Rawlings Consultants, Vancouver, B.C., Canada

Paul Kemp Canadian Projects Limited, Calgary, AB., Canada

ABSTRACT: The Pingston Hydro Project is a 45 MW hydropower project developed by private sector power development companies. The project involves 3.5 km of low and high pressure tunnels, a 450 m shaft, and 1225 m of steel liner. The tunnels are at internal pressures up to 480 m head, and the gross head at the powerhouse is 590 m, making the head at this plant one of the highest in North America. The paper will discuss construction of the tunnels and shaft, in-situ stress testing and determination of the steel liner length, high pressure plug design and construction and related design and construction issues.

1 INTRODUCTION AND PROJECT DESCRIPTION The Pingston Hydroelectric Project is a run-of-river hydroelectric project located approximately 60 km south of the town of Revelstoke in British Columbia, Canada (Figure 1) on the western shore of Upper Arrow Lake. Construction commenced in March 2001 and was substantially completed in March 2003 with commissioning in May 2003. The initial project development stage was for 30 MW capacity; powerhouse expansion to increase capacity to 45 MW will be completed in May 2004. Power from the project is sold to BC Hydro as part of the Green Power Generation procurement process. The project is one of the first projects to be developed and commissioned under the BC Hydro GPG program. Pingston Creek drains from the eastern edge of the permanent icefields of the Monashee Mountains and flows southwards parallel to Upper Arrow Lake (Figure 2); it is separated from the lake by Pingston Ridge. The creek is diverted via a headpond at El. 1035 m to the powerhouse at lake level. With a gross head of 590 m (maximum 480 m head in the tunnel), the Pingston facility ranks as one of the highest head plants in North America.

Various conveyance arrangements were evaluated for the project including: • low pressure tunnel through Pingston Ridge; long shallow buried penstock down east slope of Pingston Ridge to Powerhouse;

Figure 1. Project area location.

561 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-69.qxd

2/18/04

3:05 PM

Page 562

Figure 2. Plan of project area.

Figure 3. Section through conveyance system.

• •

low pressure tunnel through Pingston Ridge; shaft; high pressure tunnel; short shallow buried penstock to Powerhouse; long inclined tunnel; short shallow buried penstock to Powerhouse.

The second option was found to offer the best combination of cost and minimization of geotechnical risk from slope instability, which was shown to be present on the east slope of Pingston Ridge. A section through the conveyance system through Pingston Ridge is shown in Figure 3. It consists of a 1835 m low pressure tunnel (Upper Tunnel) connected to a 1450 m Lower Tunnel by a 454 m 70° inclined pressure shaft with 630 m of shallow buried penstock leading to the surface powerhouse. The 2.4 m by 2.7 m tunnels were excavated by drill and blast and the 2.1 m diameter shaft was raise-bored.

Key issues of the project included design, construction and selection of the location of the 480 m head Lower Tunnel Plug, design and construction of the plug in the Upper Tunnel and assessment and prediction of the hydrogeological impacts of the project. Tunneling conditions proved to be variable, with poor conditions being encountered in the shaft, Lower Tunnel and downstream part of the Upper Tunnel. Conditions proved to be particularly challenging in the shaft, where substantial amounts of lining were installed following completion of raise boring.

2 GEOLOGY The project area lies entirely within strong highly anisotropic metamorphic rocks of the Shuswap

562 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-69.qxd

2/18/04

3:05 PM

Page 563

Metamorphic Complex. Rock types include quartz mica feldspar gneisses, locally schistose, with minor amphibolites and some interfoliated quartzites. Quartz veining and pegmatites are common through the sequence. Uniaxial rock strengths generally lie within the range of 50 to 100 MPa. The project lies within the lower plate of the eastward-dipping Columbia River Thrust Fault which lies along Upper Arrow Lake and is a dominant regional feature. Faults and shearing parallel to this fault are common. Similarly, the foliation in the rocks dips to the east at 20–30 degrees, approximately parallel to the fault. Initial investigations, which were primarily focused towards the surface penstock option, were carried out by geological mapping along the proposed corridor with limited drilling along the alignment of the low pressure tunnel. Subsequent investigations were directed towards the Lower Tunnel with some additional deeper drilling. More detailed reconnaissance of surface features was carried out, suggesting the likelihood of shallow post-glacial sagging of the slopes dipping towards the lake. An assessment of the geotechnical risks to a surface penstock were evaluated particularly in relation to the experience of other British Columbia hydropower projects in the Columbia valley (Moore, 1999), and at Wahleach in the Fraser valley (Moore et al, 1992). It was concluded that the underground option would be adopted in order to avoid these risks. The high cost of drilling through the 400 m of cover over the pressure tunnel, and the fasttracking required to meet the power purchase deadlines determined that further investigations, including in situ stress measurement for the steel liner design, would be carried out during construction. From installed piezometers and surface observations, it was concluded that the water table was close to ground surface. The tunnels were expected to intercept groundwater flow from the high ground on Pingston ridge to the north of the tunnel which would be directed parallel to the strike of the main structural features towards the tunnel. Thus it was anticipated that local high inflows could be expected during construction where faults or shears parallel to the foliation, or major joint systems, were intercepted. Groundwater discharges at the surface were masked by the veneer of sagged ground of higher than average rockmass permeability. 3 CONSTRUCTION 3.1

General

The tunnels were excavated from three portals (Figure 3) – the Upper Tunnel was excavated downgrade from the intake at the west end (completed intake structure shown in Figure 4), and upgrade from the east portal which would be used later during operation

Figure 4. Intake structure and headpond.

for maintenance access. The Lower Tunnel was excavated upgrade from the outlet portal, located approximately 110 m in elevation above the powerhouse. Construction of the Lower Tunnel and the Upper West Tunnel began in April 2001. The Upper East tunnel advance started in July 2001. The Lower Tunnel excavation was completed in late April 2002 and the Upper Tunnel breakthrough occurred in early May 2002. The raise bore was excavated from a large cut-out in the south side of the Upper Tunnel, 250 m from the east portal. The pilot hole was drilled at 70° to the east, 445 m down to the Lower Tunnel for a period of 7 days starting March 16, 2002 and the reaming carried out over 39 days starting May 14, 2002. The rock types and structural attitudes encountered during excavation of the tunnels were similar to those predicted. However, shear zones parallel to foliation were more extensive and groundwater inflows were higher than anticipated. Concerns prior to construction were that the surficial sagged zone either might be much deeper than the mapping suggested, or that it would be underlain by a basal shear plane. Tunnel excavation did not show any open or relaxed zones indicative of deep disturbance. No major through-going structure was recognizable in the three headings; the Upper Tunnel, shaft and Lower Tunnel. A very weak and weathered shear zone which required considerable support was encountered in the Upper Tunnel and was further encountered in the shaft where it was associated with several vertical faults. This major shear zone separated the weaker, more fractured and faulted rock to the east from the more competent and less tectonically disturbed rock encountered in the western drive. The weakness of the foliation planes, even where not sheared, and the overbreak above the spring line, resulted in the need for pattern bolting in the crown throughout the Lower Tunnel. In the Upper Tunnel foliation dips were flatter, and there was more amphibolite

563 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-69.qxd

2/18/04

3:05 PM

Page 564

in the sequence thus the pattern bolting requirement was less.

60

3.2

40

50

The tunnels were all driven by drill and blast. The Upper East and Lower Tunnel drives were excavated using tracked methods. Drilling was carried out using “Long-Toms”, with mucking by track mounted overshot muckers and Hagglund cars. The tunnels were excavated at a 2% downgrade to the east, with a 100 m section at the east end of both tunnels excavated at 0.5% grade to assist the braking of the locomotives at the tunnel exits. Rounds were mostly taken in 1.8 m lengths, as determined primarily by the mucking operation. In general, two Hagglund cars were coupled in order to allow a full 1.8 m round to be mucked in one trip. Conditions in the Lower Tunnel were characterized by moderate to high seepage inflows and changeable ground conditions. The Lower Tunnel was very wet and changed quite abruptly from good to poor conditions. Total seepage inflows into the Lower Tunnel reached a maximum of almost 6000 l/min during tunnel excavation, and gradually reduced to approximately 3000 l/min. Typically, initial “flush-flow” seepage inflows would reduce with time, although there were some areas where sustained high inflows occurred. The Upper East tunnel was characterized by weathered and sheared rock, resulting in relatively slow production, and low to moderate seepage inflows. Little seepage was encountered in the first 300 m where drained conditions with local perched water tables existed. Beyond this, up to 3000 l/min was flowing from the tunnel during construction, much of this from the tunnel floor or lower walls. Rounds were taken in 1.2 to 1.8 m lengths, depending on the face and drilling conditions. The Upper West Tunnel encountered better geological conditions in comparison with the eastern drives. The permeability was very low due to the wider joint spacing and absence of shearing. Foliation was generally tight and dipped into the face throughout the drive. Rounds were taken almost exclusively in 2.4 m lengths. The Upper West Portal is located at a relatively high altitude (El. 1035 m) and it was expected that heavy snowfall, resulting in difficult access and portal working conditions, would curtail the extent of this drive. Also, given the 2% downward slope of this drive, accumulated drill/wash water and seepage water at the face was expected to inhibit progress. Excavation to approximately 550 m was planned, however due to slow progress from the Upper East, access to the west was maintained throughout the winter and the tunneling from this portal continued through to breakthrough at 1180 m (990 m from the east portal). The entire drive was accomplished with multi-stage scoop mucking

(%)

Tunnel excavation and lining

30 20 10 0 Lower

Upper East

Upper West

Total

Figure 5. Tunnel support encountered in each heading.

with re-muck bays used to store the previous round located approximately every 350 to 400 m. The rock encountered in the tunnels was mapped and classified on the Rock Mass Rating (RMR) system. Rock support was designed according to the following five classes:

• • • • •

Class I, RMR rating 86 to 100 (Very Good) – no support required; Class II, RMR rating 71 to 85 (Good) – spot bolting only; Class III, RMR rating 61 to 70 – 1.8 m pattern bolts at 1.5 m centres on crown, spot bolting as required on walls; Class IV, RMR rating 41 to 60 – 1.8 m pattern bolts at 1.2 m centres on crown, spot bolting as required on walls, followed by 50 mm thick fibre reinforced shotcrete on crown and shoulders; Class V, RMR rating 40 – 1.8 m pattern bolts at 1.2 m centres on crown and upper half of walls, followed by up to 150 mm fibre reinforced shotcrete.

The distribution of the support classes, as mapped during tunnel excavation, is presented in Figure 5. The contrast between the tunneling conditions on the east and west side is apparent; approximately 80% of the Upper West drive was in Class I and II, while less than 30% of the eastern drives were within these rock classes. 3.2.1 Rock support during tunnel excavation Rock support was provided principally with rock bolts and, where required, steel fibre reinforced shotcrete. In areas of high seepage water inflow “flash-set” plain shotcrete (approximate set time of 30 seconds) with welded wire mesh reinforcement was used. In areas of high water inflow, most of the holes drilled for bolt installation produced significant amounts of water, making it difficult to push resin cartridges into the holes. In these situations, Split Sets were installed as a means of temporary support, allowing work to proceed. The Split Sets also provided drainage to the rock so that rock bolts could be installed later to meet the permanent support requirements.

564 Copyright © 2004 Taylor & Francis Group plc, London, UK

Class I Class II Class III Class IV Class V

09096-69.qxd

2/18/04

3:05 PM

Page 565

In the Upper East tunnel, some problems were encountered with rock raveling in the back of the tunnel, causing up to 4 m of overbreak. Much of this was caused by a combination of close sub-vertical jointing and thin clay coatings or weak micaceous partings. Spiling was installed in 4 to 5 m lengths across the width of the back in these areas to prevent further raveling. Through a section of particularly sheared and wet ground in the Lower Tunnel (ch 440 to 580 m), five groups of 4 to 6 steel sets each, with wooden lagging and blocking were installed for temporary support. 3.2.2 Final rock support Upon completion of tunnel excavation, a program of permanent support installation was carried out. For the unlined sections of the upper and Lower Tunnels (ch 0 to 1850 in the Upper Tunnel and 1250 to 1450 in the Lower Tunnel), permanent support was required to prevent long term erosion or deterioration of the rock. In the Lower Tunnel where a free standing penstock was to be installed (ch 0 to 1225), rock support was required to provide protection to the pipe. The project was based on a minimum 50 year design life. Final rock support in the Upper Tunnel consisted primarily of local application of shotcrete on the arch and walls and locally applied concrete on the invert. Two concrete liners were installed in the Upper Tunnel in areas of weak weathered shear zones. The 30 m long liners were poured utilizing 2.4 m diameter liner plate culverts. For the Lower Tunnel, in the area of steel sets, final support was provided by applying a minimum 150 mm reinforced shotcrete lining within the steel sets. This arch was designed to carry all rock loads so that the sets were no longer relied upon for support. An extensive program of contact grouting was carried out to fill void spaces behind the shotcrete arch where lagging and blocking had been installed. Local areas of invert lining were installed to support locations of poor rock which were judged to be susceptible to long term erosion. Lining was also installed locally to provide a suitable surface for transporting equipment for future tunnel maintenance. An option to place concrete over consolidated tunnel muck was considered but rejected following a costing exercise. Upon completion of tunnel lining, an extensive tunnel cleaning program was carried out to minimize the risk of damage to the turbines from particles and to minimize requirements for future cleaning of the rock trap. 3.3

Raise bore excavation and lining

Drilling of a 12 inch pilot hole was the first stage of the raise boring. Careful logging of the cuttings was

carried out to gather as much information as possible about the ground conditions. Circulation losses and observation of the cuttings indicated that poor ground conditions were present near 260 to 290 m depth. Since the potential for excessive leakage from the conveyance system during operation had been identified as a significant concern, a constant head test and subsequent falling head test was conducted on the blind pilot hole (prior to breakthrough into the Lower Tunnel) to determine the permeability of the rock surrounding the shaft. The data was used to assist development of a hydrogeological model of the system. A down-the-hole survey of the pilot hole was undertaken in order that the Lower Tunnel could be steered towards an intersection point. To drain the pilot hole in a controlled manner (static pressure in the pilot hole was approximately 40 bar) a remotely operated “bar and arm drill” was used to drill through the final 3 to 5 m of rock from the face of the Lower Tunnel into the pilot hole. After this drill-through, seepage water from the shaft area through the pilot hole was measured at a steady 1700 l/min. A 2.1 m diameter reamer head was attached to the end of the pilot hole drill string and was pulled up at an average rate of 12 m/day to complete the shaft excavation. Observations during reaming confirmed the instability of the ground over the 260–285 m zone. A minimum shaft diameter of 1.8 m was required to keep the flow velocities below 4 m/s during operation, based on water flows for the expanded 45 MW project. Excavating the shaft to the 2.1 m reamed diameter allowed for up to 15 cm of liner thickness through the disturbed ground. The cuttings were mucked from the Lower Tunnel by overshot mucker and Hagglund cars, ensuring that there was always an opening at the bottom of the shaft to allow seepage water to flow and preventing choking of the base of the raise bore. Upon completion of the reaming, it was clear that some support would be required in the shaft, but the extent and type of support was uncertain. A remote survey of the shaft was undertaken to gather data for planning the support work. To conduct the investigation, a transport vehicle was built to carry a wideangle borehole camera down the shaft. The buggy was built by welding together two used bicycle frames at an angle, allowing the wheels to ride along the shaft walls at 90°. The buggy was lowered down the shaft with an electric winch and steel cable, and the camera returned a signal to the top of the shaft through 500 m of cable. The video feed was monitored in realtime and recorded. The survey provided sufficient information to delineate two main areas of concern, from 170 m to 190 m and from 260 m to 300 m. The camera survey also showed that almost all the 1700 l/min seepage inflow was coming from the 270 m to 290 m section of the shaft.

565 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-69.qxd

2/18/04

3:05 PM

Page 566

Design of the main section of shaft liner (260 to 290 m) primarily focused on constructability, with a number of options considered for constructing the liner in this area of high water ingress. In terms of design, the shaft liner was primarily required for stability purposes. A key early decision was that the liner would not be required to prevent or reduce leakage from the shaft. Preliminary hydrogeological modeling had indicated that the predicted rate of leakage from the shaft (and the upper and Lower Tunnels) was not anticipated to lead to unacceptable impacts on the groundwater regime, in particular to stability of the eastern slope of Pingston Ridge. Also, the anticipated loss of water was considered acceptable economically with respect to the project power generation. The elimination of the requirement for designing to limit leakage greatly simplified the design and construction of the liner, which was of great importance given the difficult construction conditions and the impact on schedule. Based on these requirements, the selected liner design consisted of a lightly reinforced, 15 cm (minimum) thick full circumferential liner (Figure 6). The liner was designed to be leaky to minimize required concrete thickness and reinforcement. Furthermore, high pressure consolidation grouting was not carried out, with only a modest program of low pressure grouting being required. A number of alternative methods for gaining access to the shaft to carry out the lining work were considered, with an Alimak raise climber eventually being selected. The contractor for the shaft lining (J.S. Redpath of Sudbury) selected a slip-form lining system that progressed from the top to the bottom, thus eliminating any requirements for temporary support (as required for a bottom-up approach) in this area. The liner was poured in 15, 2 m long slip form sections. Vertically, the forms were split 0.5 m from the bottom with the

Figure 6. View of upper terminus of main shaft liner.

lower section called the “curb” and the upper, 1.5 m section, called the “main”. The 0.5 m curb section was poured, followed immediately by the installation of the main form panels, which would rest inside the poured concrete surface of the previous section. The main pour would then proceed through a small window in the formwork located at the highest point on the hanging wall. The concrete was intermittently vibrated using a pneumatic form vibrator. Concrete was delivered from a batch plant located at the top of the shaft through a 75 mm diameter steel slickline bolted to the footwall. Six threaded ports for grouting and for drain hole drilling were fitted into each of the pour sections. Control of the 1700 l/min leakage between 270 m and 290 m was a key concern of the contractor and design team. Large, individual seepage sources were collected in steel boxes and pipes fixed to the walls of the shaft and routed through three 2 inch diameter steel pipes that were carried to the bottom of the liner, cast into the north wall. Diffuse water sources were controlled using geotextile drainage fabric (Nilex NuDrain), which was pinned to the rock surface and fed into a drainage collection system. This drainage system proved to be extremely effective in controlling inflows and minimizing damage to the concrete. Grouting was carried out on completion of the shaft lining and was commenced by grouting the drainage system from the discharge pipes at the bottom of the liner. Following grouting of the drainage system, contact grouting was carried out through the grout hole rings (six holes each). Once a ring of grout holes were drilled (each 3 m long from the inside of the liner), a network of injection lines were attached to the six ports and grout was injected one hole at a time, proceeding clock-wise around each ring. Maximum injection pressures which were slightly above hydrostatic groundwater pressures were maintained. Once one hole was completed, the injection valve was closed, maintaining the injection pressure in the hole, and the next injection valve was opened to start grouting of the next hole. A second grouting pass of the holes was generally attempted to see if any more grout could be injected. Grouting proceeded from the bottom to the top of the liner and proved to be effective as observed inflows were progressively channeled upwards in the liner. Once all the grouting was complete, the seepage stopped almost completely for 1–2 days while pressure increased behind the liner. With increasing pressure, seepage began to reappear, generally at the locations of construction joints. As indicated previously, it was decided that the liner was to be designed to leak in order minimize hydraulic particularly during watering and dewatering. Drainage holes were drilled upon completion of grouting. Many of the holes encountered significant amounts of water, which decreased with time, and within a week, the pressure

566 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-69.qxd

2/18/04

3:05 PM

Page 567

stabilized at a constant seepage flow, similar to that before grouting. In addition to the main concrete liner, 12 additional, more localized zones required concrete lining and numerous areas required bolting. The contractor elected to pour concrete in these areas rather than use shotcrete. These areas were all supported with 15 cm of reinforced concrete in partial or full circumference sections. Grouting was not required in these areas. 3.4

Plug design and construction

A detailed description of the design and construction of the two tunnel plugs is beyond the scope of this paper, however a summary of critical design and construction issues is provided in this section. 3.4.1 Lower Tunnel plug A plug was required in the Lower Tunnel to serve as the upstream terminus of the steel penstock (Figure 7). Selection of the plug location and design of the plug was to be determined during tunneling, based on observed ground conditions and in-situ measurements taken during construction. Based on cover considerations, it was anticipated that adequate confinement would be available near ch 800. Upon reaching this

location and identifying a suitable plug location, a hydrojacking testing program was carried out. The testing results indicated that the hydrojacking pressures were lower than what was considered acceptable by the design team. It was decided that an alternative location would be sought further upstream in the tunnel where greater depth of rock cover would be available. Due to poor rock conditions, the next suitable plug location was not identified until ch 1225. A second hydrojacking testing program was commenced upstream of the proposed plug location. Results from this testing indicated that hydrojacking pressures were still lower than required by the design and, in fact, were similar to the values measured at ch 865, where rock cover was approximately 100 m less. Further testing at ch 1420 m, under still higher cover indicated that hydrojacking values were not increasing. The stress measurements presented a significant challenge to the design team. It was clear that even if the plug was located at the extreme upstream end of the Lower Tunnel (at the location of greatest rock cover), the measured hydrojacking values were still insufficient to meet the design criteria. While a detailed discussion of the stress measurements, analyses and design deliberations is beyond the scope of this paper, it was eventually concluded that a lower factor of safety could be tolerated. Information considered in reaching this conclusion included the very good rock conditions, (indicated by very wide joint spacing, an absence of shearing or faulting, and very low hydraulic conductivity), hydrojacking values from the Upper Tunnel and precedence from other projects. In order to achieve a high level of confidence in the construction and subsequent performance of the tunnel plug, a number of key features were incorporated into the plug construction,

•

• •

Figure 7. Lower Tunnel plug – upstream end.

The plug was completed over a period of approximately 2 months with completion near the end of October, 2002.

567 Copyright © 2004 Taylor & Francis Group plc, London, UK

Self compacting concrete (SCC) was used for the entire plug. This eliminated the requirement to gain access into the extremely confined areas for concrete vibration and produced a highly workable, high strength and low shrinkage mix which was highly suitable for this application. This also allowed larger pours and fewer construction joints; The plug was constructed and poured in four equal 6 metre sections. Each section was poured in a single continuous pour, thereby completely avoiding longitudinal construction joints; Ultrafine grout was used for high pressure consolidation grouting. Highly stable, low viscosity grout mixes with extremely good penetration properties were achieved using mixes prepared with ultrafine grout.

09096-69.qxd

2/18/04

3:05 PM

Page 568

200

1200

Water Levels - Target

1000 800

100

Leakage - Predicted

600 Leakage - Actual

50

400 0

20-Mar

19-Mar

18-Mar

17-Mar

16-Mar

15-Mar

14-Mar

13-Mar

12-Mar

200

11-Mar

-50

Elevation (m)

Flow (l/sec)

150 Water Levels - Actual

0

Date

Figure 8. Upper Tunnel plug – upstream end showing door hinge on left

Figure 9. Tunnel filling curve – predicted and actual response.

• 3.4.2 Upper tunnel plug The Upper Tunnel plug was to be designed to allow passage of maintenance equipment into the tunnel for cleaning the rock trap and carrying out maintenance in the Upper Tunnel and shaft (Figure 8). A door, of minimum size 1.8 m (wide) by 2.0 m (high) was required for access of the equipment. After consideration of a number of alternatives, a hinged rectangular door, fitted inside a steel frame and liner was selected for the design. The static head acting at the plug was 5 bar. The plug length was 6 m, with the upstream 2 m of the plug being steel lined, and the downstream 4 m consisting of reinforced concrete. A significant challenge in constructing the plug was to minimize disruption of access into the Upper Tunnel, where a continuous supply of materials was required for permanent support activities to continue in the shaft and Upper Tunnel. This required a modular construction sequence, with much of the plug construction being carried out in late December 2002, when work in the shaft was halted for 10 days.

4 CONVEYANCE FILLING As discussed previously, the impact of the underground excavations on the existing groundwater regime during operation was considered to be a key concern. It was necessary to understand the potential impact of the project on existing groundwater conditions – particularly as slope stability might be affected – and to understand the potential losses of water available for power generation. To achieve this, a hydrogeological model was developed and calibrated using available data from:

•

in-situ testing carried out during the pre-construction geotechnical investigations and during construction;

Since much of the pre-construction investigations were carried out at shallow depths, emphasis was placed on collecting information during construction which could be used to calibrate the model. The tests conducted on the pilot hole for the raise bore (discussed previously) were particularly important in helping to calibrate the model. Filling of the conveyance system was also seen to be a useful opportunity to collect information on the performance of the system and for comparison with the hydrogeological model predictions. A carefully developed tunnel filling procedure was considered essential in order to minimize damage to the tunnel liners, plugs and rock and also to carry out tests to assess the performance of the hydrogeological model. A filling schedule was developed with a duration of 8 days. The schedule included rapid filling of the lower penstock and slower, more controlled filling of the Lower Tunnel, pressure shaft and Upper Tunnel. Control of the filling rate was achieved using controls on inflows via a gate valve at the intake, and with fine adjustments achieved by spilling excess water at the powerhouse using needle valves on the turbines. Regular measurements of total system inflows (intake inflow minus discharge at the powerhouse) and system pressure (measured at the powerhouse) allowed the measurement of total system seepage losses against system pressure. These values were compared at regular intervals to predictions developed using the hydrogeological model. For ease of comparison, all values were plotted, allowing immediate comparison against predictions (Figure 9). At a number of key steps, system inflows and outflows (at the intake and powerhouse) were closed and the response of the conveyance system measured. These tests provided an accurate measurement of leakage

568 Copyright © 2004 Taylor & Francis Group plc, London, UK

using measurements of groundwater inflows into the tunnel which were collected at regular intervals during tunnel excavation.

09096-69.qxd

2/18/04

3:05 PM

Page 569

losses from the tunnel and shaft (measured by the falling level of water in the shaft/tunnel). As shown in Figure 9, the conveyance filling showed a close agreement between measurements and predictions for the filling of the Lower Tunnel and shaft. Filling of the Upper Tunnel, showed a divergence between prediction and measurements during the initial hours of filling. Upon completion of filling of the Upper Tunnel, a system test showed the model underpredicted seepage losses but subsequent tests showed a reducing disparity. This disparity is considered to be due to re-saturation of the surrounding rock, drained during the two years of underground excavations. Long term (steady state) predictions were found to have an acceptable level of accuracy. Shortly after the successful completion of conveyance filling, a malfunction occurred in one of the main valves at the powerhouse. The repair of the valve required immediate dewatering of the tunnels and shaft. Tunnel dewatering was successfully completed, which allowed an early inspection of the tunnels and base of the shaft. The inspection revealed that the liners, the tunnel plugs and the rock withstood the demands of watering and dewatering extremely well, with no visible signs of damage.

• •

• •

Difficult concrete lining required in a 30 m section of sheared and faulted ground in the inclined shaft; Groundwater modeling based on in situ tests and assessment of tunnel construction inflows were carried out to assess level of potential leakage against acceptable losses and the potential for slope instability; Monitoring during conveyance filling to check against the model results showed good correlation; Inspection after early dewatering showed the absence of any rock instability or deterioration and no plug problems.

The project was developed jointly by Canadian Hydro Developers of Calgary and Brascan Power of Toronto. Design and project management was carried out by Canadian Projects Limited (CPL) of Calgary; Golder Associates provided tunnel design and construction inspection services and materials QA/QC services; EBA Engineering provided design and inspection services for the Diversion weir; and Graham Rawlings Consulting Ltd acted as geological advisor. AMEC Earth & Environmental provided rock support design services under a contract with Thyssen Mining and Construction Ltd (TMCC). TMCC carried out tunnel construction and raise boring, while JS Redpath installed all shaft linings.

5 CONCLUSIONS The significant aspects of the underground works carried out as part of the Pingston Hydro project are as follows:

• • • • •

The project is a very high head development located close to a geologically complex area; Quartz gneisses were uniformly present throughout the underground works with consistent eastdipping foliation; Foliation shearing and major jointing in much of the Lower Tunnel, shaft and part of Upper Tunnel resulted in high water inflows; Hydrojacking tests in the lower power tunnel showed that in situ stresses were unusual with minimum stress being much less than overburden; Design decision made to accept low FOS against hydrojacking at the plug upstream of the 1225 m steel liner because of the good rock and tight conditions; very careful design and construction of plug was needed;

ACKNOWLEDGEMENTS The authors wish to thank Canadian Hydro Developers, Brascan Power and Canadian Projects Limited for permission to publish this paper and for the opportunity to work on the project.

REFERENCES Benson, R.P. 1987. Design of Unlined and Lined Pressure Tunnels. Canadian Tunnelling: 37–65. Moore, D.P., Ripley, B.D. and Groves, K.L. 1992. Evaluation of Mountainslope Movements at Wahleach, in Geotechnique and Natural Hazards, Bitech Publishers, Vancouver, Canada, pp 99–107. Moore, D.P. 1999. Rock Slopes and Reservoirs. In “Slope Stability and Landslides”, Proc, 13th Vancouver Geotechnical Society Symposium, Vancouver, pp 1–18.

569 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-70.qxd

2/18/04

3:06 PM

Page 571

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Shoal creek raw water intake and pump station construction on Lake Lanier Donald Ackerman, Randy Wiek & Ryan Gutridge W. L. Hailey & Company, Inc., Nashville, Tennessee, USA

ABSTRACT: Due to an ever increasing need for water in the greater Atlanta Metropolitan area, a new raw water intake and pump station is being constructed on Lake Lanier just north of the Atlanta, Georgia area in Gwinnett County. The new intake will have a capacity of 240 MGD, with expansion capability to 330 MGD. This paper will provide an overview of each portion of the Project’s construction phases, including shaft excavation, tunnel construction, safely tying the tunnel into the lake tap shaft, potential water inflows while mining the tunnel 140 below the lake, and construction of the intake and pump station.

1 INTRODUCTION W.L Hailey & Company, Inc of Nashville, Tennessee was awarded the project in June of 2001 with a contract value of $29,623,000. The Notice to proceed was given on August 6, 2001 and mobilization began shortly thereafter. The scheduled contract time was 29 months. The Shoal Creek Raw Water Intake and Pump Station is being built adjacent to Buford Dam Road in Gwinnett County on Lake Lanier. The new facility is being constructed for the Gwinnett County Department of Public Utilities and will serve the present and future needs of their customers. The Project was designed by Parson Engineering Science of Norcross, Georgia. Figure 1 shows a general layout of the project that includes a 140 deep land shaft 80 in diameter that transitions to 20 diameter down to the tunnel invert, a 126 diameter 65 deep lake shaft drilled in 70 of water and a 10 finished diameter modified horseshoe

Figure 1. Shoal Creek Intake profile.

tunnel connecting the two shafts. This system, in conjunction with a new pump station and dual 72 pipelines transmission lines, will supply 240 MGD of raw water from Lake Lanier to two water filtration plants for treatment and distribution.

2 SHAFT CONSTRUCTION The main shaft for the pump station consists of an 85 foot diameter shaft excavated in rock to a depth 59 vertical feet that transitions to a 20 foot diameter shaft for an additional 81 feet. The major concern with excavating a shaft this large was whether to shoot the shaft in quarter sections, halves or blast the entire diameter at one time. Another issue effecting this decision was that the engineer stipulated a very stringent vibration restriction of 0.5 inches/second at any adjacent residential structure or at Saddle Dike No. 3. After much thought and planning, it was ultimately decided that these restrictions could be met by shooting the entire shaft diameter at one time. The 85 foot diameter shaft was shot in four lifts each ranging from 11 to 15 vertical feet. Ground support consisted of #8 bolts 10 feet long placed on 5 foot centers. Following the initial round of blasting for the 85 shaft, modifications to the blast design were made to minimize the cap rock size and to gain better fragmentation for excavating purposes. Figure 2 shows the plan view of the drill hole layout that was utilized. Each round consisted of 127 trim holes, 40 buffer holes and 96 production holes. The average powder factor was 1.95 pounds per cubic yard with a smooth wall

571 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-70.qxd

2/18/04

3:06 PM

Page 572

85 ft

PlanView

Loaded Center Hole

8Open Cut Holes Angled In

Angled In

Blastholes

Perimeter Holes

which the shaft would be drilled. Pre-construction borings and underwater surveys indicated that at the location of the intake approximately 4 vertical feet of silt and 12 vertical feet of highly weathered rock needed to be removed in order to reach sound rock. An American HC-80 crane was placed on a 80 40 Rendrag barge platform and excavated the overburden with a clam shell. A full depth impervious silt boom was in place around the work site and the spoil material from the excavation was wasted within the boom. A small amount of subaquious blasting was required. Underwater overpressures were controlled with the use of a bubbler curtain. The blast area was surrounded with a 160 120 rectangular mass of air pipes with forty-six thousand 1 mm holes. Air compressors with a combined output of 8400 CFM drove this system and its usage resulted in no recorded fish kills outside of the air curtain. 4 INTAKE SHAFT EXCAVATION

Figure 2. Drill layout.

Figure 3. Loading an 85 foot shaft round.

factor of 0.16 pounds per square foot. Production holes were 3.5 inches in diameter and bulk emulsion was used. With this loading scheme, the maximum peak particle velocity encountered was 0.45 inches/second. The geology in the shaft consisted of moderately weathered to weathered biotite gneiss that transitioned to granitic gneiss and mica schist. At the shaft, the transition zone was approximately 10 until the biotite gneiss was encountered. Ground water inflow into the shaft was minimal despite being approximately 140 feet below the static water level of Lake Lanier. 3 UNDERWATER BENCH EXCAVATION The first step in constructing the intake shaft was excavation of a 36 36 bench with slopes at a 2:1 within

The original contract documents called for a template and secant piling system to be installed prior to shaft excavation. A Value Engineering proposal was made and accepted by the Owner to replace the template and piles with a concrete anchored starter casing which allowed the shaft to be drilled with a Wirth Pile Top Drill Rig. The barge configurations for this stage included a 90 80 tee shaped Rendrag assist crane platform with a Manitowoc 888 and a 40 100 Flexifloat jack-up platform on which the Wirth PBA 928 Drill Rig sat as shown in Figure 4. A 13 diameter by 20 long by 1 thick starter casing was drilled with the Wirth Rig 1 into rock. Vertical measurements were taken from the casing to the deck of the barge at 4 points at 90 degree angles from one another to ensure the casing was plumb prior to drilling. Concrete was then placed inside and outside of the casing to anchor it and to allow for a flat surface to begin the drilling work. The Wirth rig was outfitted with button style bits and drilling began on May 22, 2002. The rock formations throughout the shaft consisted of layers of granitic gneiss and mica schist running at 45-degree angles. The angle of the formations combined with constant changes between hard and soft rock resulted in production rates ranging between 1–2 feet per shift. A low downpressure of 50,000 lbs was necessary to keep shaft within the plumbness requirements of .1 foot out for every ten vertical feet. The 12–6 diameter shaft was drilled to a depth of 65 feet which took it to the invert of the tunnel. Upon completion, a boretack was run through the drill string confirming that the shaft was within 0.65 of plumb. A Phantom HD2 ROV was used to inspect the shaft prior to liner installation to ensure that the shaft was clean and acceptable.

572 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-70.qxd

2/18/04

3:06 PM

Page 573

Figure 4. Intake construction barge layout.

Figure 6. Intake schematic.

Figure 5. Setting the intake liner.

5 LINER INSTALLATION The liner selected for the project consists of a 10 inside diameter by 50 long by 1 thick steel pipe with 3/4 hemispherical domes on the top and the bottom and four gusset plates which rest on the starter casing. The top dome has a flanged connection that will be placed on top of the intake structure at the end of the project. The lower dome had a welded connection that was removed once the tunnel tie in to the lake tap was complete. A series of eight 3 grout tubes of varying length were installed around the perimeter of the liner. With the shaft drilling complete, the starter casing was trimmed to grade and checked for plumb. The intake pipe was towed to the worksite, raised vertical with the crawler crane, and water was pumped into the pipe until it reached 1000 lbs of negative buoyancy as shown in Figure 5. The section was then lowered into the shaft and shims were installed between the gusset plates and the starter casing to ensure the liner was plumb. The domes were fabricated with 4 and 2 flooding and venting valves which were utilized to flood the pipe prior to grouting. 6 GROUTING Figure 6 shows a schematic of the intake liner and the various grouting stages. 200-psi flowable fill was

Figure 7. 10 Finished diameter conveyance tunnel.

placed in the bottom of the shaft and filled the area from the invert of the tunnel to a point 5 feet above the crown. This material was used since this area would be mined out once the tunnel intersected the intake shaft. The annular space between the shaft and the liner was grouted in two lifts using 5000-psi neat cement grout introduced through the grout tubes. The mix consisted of Type I cement, anti-washout, retarder, and water reducer. The grout was specially designed for underwater placement and the 1000 feet pumping distance between the on site batch plant and the intake site. 7 TUNNEL CONSTRUCTION The conveyance tunnel for this intake is a 10 foot finished horseshoe shaped tunnel with a cast-in-place concrete liner. The original bid document called for a shotcrete final lining but was later changed to cast-in-place. The tunnel is 625 feet in length and

573 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-70.qxd

2/18/04

3:06 PM

Page 574

approximately 54 feet below the lake bottom. Conventional drill and blast methods were utilized to drive a 12 foot horseshoe tunnel. Excavation was performed on a one shift per day operation achieving an average of 10 feet of advance per day. No significant ground support was required and water inflows were less that 25 gpm when the excavation was complete. A Gardner Denver Two Boom Jumbo was utilized for drilling and a Wagner ST2 was used for mucking. The final step of the tunneling operation included removal of the flowable fill below the intake liner and removal of its lower dome. The geology in the tunnel consisted granitic gneiss and mica schist. There were concerns of rock weathering or fractured conditions at the intake location but this was encountered. Provisions were provided in the contract documents to grout this section, but very little grout was actually required. 8 PUMP STATION The intended use of the overall facility is to supply raw water to both an existing and a new water filtration plant. Four 2500 HP and four 700 HP vertical turbine

pumps will drive the conveyance system with a dedicated 72 discharge line running from the pump station to each of the filter plants. 4600 lf of 72 cement lined welded steel pipe was laid on this project. The 16,000 SF pump station is below grade with a small electrical building being the only structure visible from the lake. To complete the intake structure, the tunnel and wet well were flooded by means of the flooding and venting valves installed on the top dome of the intake. The top dome was then removed and a 10 diameter by 6 diameter tee structure was mounted to the intake liner and twin Tee-style screens were then installed.

9 CONCLUSION The successful completion of the Shoal Creek Intake and Pumping Station Project means that Gwinnett County Georgia will have an ample water supply for the foreseeable future, as it is one of the fastest growing communities in the nation. The design of the lake tap and sub-surface pump station required minimal environmental impact during construction with an esthetically pleasing product left for the community.

574 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-71.qxd

2/18/04

3:07 PM

Page 575

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Design and construction of shafts at the San Roque Project Mark Funkhouser, Richard Humphries, Wayne Warburton & James Daly Golder Associates Inc., Atlanta, Georgia & Lansing, Michigan, USA

Edward O’Connor Washington Group International, Inc., New York, New York, USA

ABSTRACT: Construction of the San Roque Multipurpose Project in the Philippines has recently been completed. It is one of the world’s largest Build-Operate-Transfer/Design-Build hydropower projects. The seven shafts constructed at the project are described in this paper. They include a 23 m diameter, 100 m deep surge shaft, a 10 m diameter drop shaft, two large gate shafts, two small air vent shafts and a 28 m wide by 80 m long by 48 m deep shaft for the powerhouse. A variety of designs, excavation methods, rock support types, and lining types were used for these shafts. The design-build method of construction offered a unique opportunity for the designers and constructors to work as a team to develop the most efficient and cost effective methods to complete shaft construction in a very tight project schedule. This paper describes the design, excavation and rock support methods used for the shafts.

1 INTRODUCTION The San Roque Multipurpose Project is one of the world’s largest Build-Operate-Transfer projects recently constructed. The project will provide hydroelectric power, irrigation water supply, and flood control for a large portion of Luzon. The project was constructed on the Agno River in Pangasinan Province in northern Luzon Island, Philippines. Construction of the project began in 1998, and power generation started in 2003. Underground work for this project consisted of 8.5 km of tunnels and shafts. Figure 1 shows the overall arrangement of the project with the shafts and related features identified. Figure 2 shows the relative sizes of the various shafts. The majority of shaft excavation was performed by the raise bore and slash method. The two vent shafts were excavated by raise boring alone, and the uppermost approximately 15 m of the surge shaft was conventionally sunk. Typically, the shafts were initially supported with a combination of shotcrete and rockbolts, and then lined with castin-place concrete. The vent shafts were unlined at the time of excavation, and then lined with steel lining concreted in place. The surge shaft was initially supported with cast in place concrete rings and shotcrete, and then lined with cast in place concrete. The shafts were excavated in conjunction with the upper and lower power tunnel, the irrigation tunnel,

Figure 1. Plan view of San Roque project site.

and the grouting galleries. Excavation, support, and lining of the underground facilities were usually performed 20 hours per day, six days per week for the approximate 3-year construction period. The work force for underground construction consisted of up to

575 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-71.qxd

2/18/04

3:07 PM

Page 576

250 miners and 9 superintendents. Full service crews accompanied the equipment from Tamrock and Atlas Copco/Wagner. In addition to the tunneling equipment, shaft work was performed with the following equipment: blast hole and rockbolt hole drilling was performed with Tamrock Commando drills, mucking was performed with a Shaff mucker or Caterpillar excavators, raise boring was performed by Skanska, and shotcreting was performed with a Normet Spraymec, a Shotcrete Technologies robotic arm and by hand-held nozzle. All shotcrete and concrete was batched on-site and delivered to the respective shaft collars via transit trucks and then transferred to shotcrete machines or inside formwork via slicklines.

2 GEOLOGY Rock at the project site is predominantly andesitic volcanic breccia with localized diorite intrusions. In the southern part of the site there is a younger, fine to very coarse grained conglomerate that unconformably overlies the volcanic breccia. The conglomerate played a significant role in dictating excavation and support methods in the surge shaft and other excavations on site. The degree of cementation of the conglomerate varied greatly. Locally, the rock was little more than boulders with interstitial sand, silt, and clay, while in other exposures the conglomerate was so well cemented that during excavation, the boulders would break before dislodging from the matrix. The rock mass is typically weathered to several tens of meters below original ground, and has closely spaced joints, with a typical spacing of 20 mm to 200 mm. Shear zones were encountered during excavation of several of the shafts. These zones typically have several meters of highly jointed rock on either side of a thin clayey gouge layer. With the exception of the vent shafts, each of the shafts had an exploratory boring along centerline. During excavation of the shafts, the exposed rock mass was mapped and logged by site geologists.

3 ACCESS SHAFT 3.1

Figure 2. Size of the various shafts.

576 Copyright © 2004 Taylor & Francis Group plc, London, UK

Access Shaft design

The Access Shaft for the control gates of the Low Level Outlet (or irrigation) Tunnel is close to upstream end of the tunnel as shown on Figure 1. The shaft has an excavated diameter of approximately 6.5 m as shown on Figure 2 and is approximately 83 m deep. The gate chamber is located at the bottom of the shaft and is approximately 15 m wide by 19 m long by 16 m high. Rock support for the Access Shaft was based on three categories of support for three ranges of rock quality. Category III support, intended for the collar area and zones of rock with Q values less than 0.1, consisted of W160 11 kg (W6 25 lb) steel rings at 1.2 m spacing and 150 mm of shotcrete. The shotcrete was specified to be placed with at least 50 mm of fiber reinforced shotcrete as an initial support layer. The remaining 100 mm was specified to be plain (unreinforced) shotcrete to encase the steel rings. The shotcrete was required to be locally thickened to encase the steel rings. Category I and II rock support was designed for rock mass quality ranging from Q values of greater than 1.0 and between 0.1 to 1.0, respectively. Category I rock support was specified to be spot positioned rockbolts and 25 mm of fiber reinforced shotcrete. Category II rock support was specified as 2.5 m long

09096-71.qxd

2/18/04

3:07 PM

Page 577

rockbolts on a 2 m by 2 m pattern and with 50 mm of fiber reinforced shotcrete. Rockbolts for both categories were 32 mm diameter, 400 MPa yield strength, resin grouted, epoxy coated bars. The shotcrete for both categories was specified to have a 28-day compressive strength of 34 MPa, and also had a prescribed high early strength gain so that “immediate” rock support could be achieved during excavation. The final shaft lining consists of 300 mm thick reinforced concrete. 3.2

Access Shaft construction

Access Shaft excavation commenced after the Irrigation Tunnel was advanced the initial 100 m from the upstream portal to just beyond the shaft centerline. The raise bore drill was set up on the shaft centerline, and a 0.3 m pilot hole was drilled down to the tunnel. The raise bore cutter head was installed and used to ream the pilot hole to 1.2 m in diameter. Cuttings were mucked out of the upstream portal of the irrigation tunnel. Raise bore drill rates were approximately as anticipated and no difficulties with hole stability were experienced during raising. The shaft collar was constructed by excavating approximately 3 m deep in the near surface weathered rock with a hydraulic excavator. Reinforced concrete was placed behind formwork in this area. The remaining portion of the shaft was excavated by slashing to a diameter of approximately 6.5 m in 2 m lifts. A Tamrock Commando was used for blast hole and rockbolt drilling, and a Timberland 27 tonne stiff leg derrick was used for hoisting materials and equipment. The work deck was operated on a separate hoist system. During drill and blast slashing, advance averaged 1.6 m/day. The upper approximately 20 m were supported with Category III support. The remaining portion of the shaft was primarily supported with Category II support. The support for each round typically consisted of fiber reinforced shotcrete and pattern rockbolts. Based on the typical Q values mapped after each round, support could have been by spot rockbolts, but some clay coated joints and persistent overbreak resulted in collective conservatism during support selection. During excavation, the rock quality was logged after each round. A summary of the Q values and RMR mapped during the excavation is as follows:

Minimum Maximum Median

Q

RMR

0.75 17 4.3

27 89 47

Excavation of the Gate Chamber was difficult from the standpoint of logistics. The chamber is above a portion of the tunnel that was initially excavated, so the only access for the excavation and support was

through the relatively small Access Shaft. Initially the chamber was advanced as a central top heading driven to the north out of the Access Shaft. The central top heading was made just tall and wide enough to allow installation of two rows of 4 m long rockbolts in the crown. Two side headings were then advanced at each side of the top heading, and an additional two rows of rockbolts were installed. The initial bench lift was advanced full width south to north. Then the bottom lift was drilled and blasted in one round to drop the “floor” in to the open tunnel below. The chamber rock support was completed off the muck pile and advance downward. The final lining of the Access Shaft was reinforced concrete. The shaft and chamber were required to be “dry” structures, and consequently were designed to withstand full hydrostatic pressure at full reservoir level.

4 POWER TUNNEL GATE SHAFT 4.1

The Gate Shaft for the Power Tunnel, shown on Figures 1 and 3, is near the upstream end of the tunnel. The shaft has an excavated diameter of approximately 12.6 m, as shown on Figure 2, and is approximately 84 m deep from the collar to the crown of the Power Tunnel. The excavated cross section of the Power Tunnel is enlarged below the shaft for the gates and associated structures. Rock support for the Gate Shaft was designed similar to the rock support for the Access Shaft. Category III support, intended for the collar area and zones of poor rock, consisted of W250 22 kg (W10 49 lb) steel rings at 1.2 m spacing and 150 mm of shotcrete. At least 50 mm of fiber reinforced shotcrete provided the initial support layer with the remaining 100 mm being plain shotcrete to encase the steel rings. Category I and II rock support was designed for rock mass quality ranging from Q values of greater

Figure 3. Power Tunnel profile.

577 Copyright © 2004 Taylor & Francis Group plc, London, UK

Gate Shaft design

09096-71.qxd

2/18/04

3:07 PM

Page 578

than 8.0 and between 0.3 to 8.0, respectively. Category I rock support was specified to be spot positioned rockbolts and 25 mm of fiber reinforced shotcrete. Category II rock support was specified as 4 m long rockbolts on a 1.75 m 1.75 m pattern with 50 mm of fiber reinforced shotcrete. Rockbolts for both categories were 32 mm diameter, 400 MPa yield strength, resin grouted, epoxy coated bars. The final lining of the Gate Shaft consists of 600 mm thick reinforced concrete. 4.2

Gate Shaft construction

The Timberland 27 tonne stiff leg derrick was relocated from the Access Shaft to the Gate Shaft collar. Gate Shaft construction was performed in a similar manner as the Access Shaft with some modifications for the larger size. Pilot hole drilling and reaming, and collar construction were similar to that at the Access Shaft except the raise bore was 1.8 m diameter and the collar was more substantial for this larger diameter shaft. Also there were walls constructed, to form a “muck bin,” inside the Power Tunnel to contain the falling raise bore cuttings so that personnel access could be maintained in the Power Tunnel downstream of the Gate Shaft. After completion of the raise bore, slashing to full diameter commenced by drill and blast. A Tamrock Commando was used for blast hole and rockbolt drilling. A Shotcrete Technologies robotic arm mounted on a tripod was used for shotcrete application. A twolevel work deck was operated on a separate winch system. The upper 6 m were supported with a modified Category III support, and then the balance of the shaft was supported with Category II support – fiber reinforced shotcrete and 4 m long pattern rockbolts. Lift heights were typically 3 m. The Category III support was modified during construction and consisted of 250 mm thick cast concrete behind steel plate lagging and rings. The average drill and blast slashing rate was 1.8 m/day. Excavation of the Gate Shaft went well until the lower few meters, where remedial work was needed to prevent instability of the final shaft perimeter and the crown of the Power Tunnel. A steel lining was concreted in the raise bore to support the poor rock zone. Excavation of the lower most several meters of the shaft was deferred until other work activities were completed in the Power Tunnel downstream of the Gate Shaft. Then the shaft excavation was completed without difficulty. The steel and concrete lined interval of the Raise Bore was blasted out with the last lift. This was a tall lift because the muck pile needed to be large enough to fill the tunnel below to provide a work bench for rock support installation around the bottom of the shaft, and concrete placement.

5 POWER TUNNEL DROP SHAFT 5.1

The Drop Shaft, shown on Figures 1 and 3, is approximately at the mid point of the Power Tunnel. The shaft has an excavated diameter of approximately 10 m and is approximately 65 m deep. The excavated cross section of the drop shaft at the upper and lower elbows was significantly larger than the minimum specified cross section and significant rock support was required in these areas. The three categories of rock support for the Drop Shaft were designed to be similar to the Power Tunnel support. Category III support, intended for zones of rock with a Q value less than 0.5, consisted of W200 14 kg (W8 34 lb) steel rings at 1.2 m spacing and 300 mm of shotcrete. The shotcrete was specified to be placed with an initial support layer of at least 50 mm of fiber reinforced shotcrete. The remaining 250 mm was plain shotcrete to encase the steel rings. Category I and II rock support was designed for rock mass quality ranging from Q values of greater than 10.0 and between 0.5 to 10.0, respectively. Category I rock support was specified to be spot positioned 4 m long rockbolts and 25 mm fiber reinforced shotcrete. Category II rock support was 4 m long rockbolts on a 2 m 2 m pattern with 50 mm of fiber reinforced shotcrete. Rockbolts for both categories were 32 mm diameter, 400 MPa yield strength, epoxy coated bars. 5.2

Drop Shaft construction

Construction of the Drop Shaft was challenging for several reasons. The sequence of drilling the pilot hole and then pulling the raise bore had to be coordinated with the excavation and support of both the upper and lower Power Tunnels because both tunnel headings had to be completed before the shaft work could begin. The upper Power Tunnel heading was stopped a few meters short of the Drop Shaft centerline and then the lower Power Tunnel was completed to approximately 5 m upstream of the Drop Shaft centerline. The raise bore pilot hole was drilled from ground surface to the lower elbow, and then the 1.8 m diameter reaming head was installed. The raise bore was pulled from the lower elbow up to near springline of the upper Power Tunnel. The reamer was then lowered back down, the head removed, and then the drill string removed from above. Several weeks were required to assemble and erect the platform and work decks in the upper Power Tunnel at the top of the Drop Shaft. Hoists were operated from the ground surface (through the steel cased raise bore pilot hole) and from in the upper Power Tunnel.

578 Copyright © 2004 Taylor & Francis Group plc, London, UK

Drop Shaft design

09096-71.qxd

2/18/04

3:07 PM

Page 579

Drill and blast slashing went as planned and rock support was typically Category II support. Drilling for blast holes was with a Tamrock Commando drill. Lifts were typically 2 m with rock support installed on a lift-by-lift basis. The sinking rate during drill and blast slashing averaged 1.7 m/day. The final lining was 600 mm thick reinforced concrete. Placement logistics and sequence was dictated by schedule and coordination with continued work in the lower Power Tunnel. Concreting was started in the lower elbow, where massive timber formwork was required, and then progressed to the shaft where concrete was placed from bottom to top with a 2 m high 8.5 m diameter steel form. Concrete placement in the upper elbow, which included contact grouting through the hoist cable from ground surface, was the final work activity in the Drop Shaft. 6 SURGE SHAFT 6.1

Surge Shaft design

Design of the Surge Shaft was challenging because of the large diameter and the poor rock conditions anticipated in the upper portions where the Surge Shaft had to be excavated through approximately 45 m of the poorly cemented conglomerate. The Surge Shaft is shown in profile on Figure 4. Several iterations of initial Surge Shaft rock support design had been completed when underground excavation through the conglomerate indicated that behavior of the conglomerate would be better than indicated by the near zero RQD in the exploratory

boring at the shaft location. Using experience from previous excavation in the conglomerate, site staff and the design offices were able to develop an initial rock support design that was beneficial to the project because of ease of installation. The final design of the rock support for the Surge Shaft consisted of 400 mm thick concrete rings, 1.2 m high, spaced 2 m vertically. The gaps between concrete rings were shotcreted with weep holes to reduce build-up of hydrostatic pressure. 6.2

Surge Shaft construction

Prior to excavation, the perimeter of the surge shaft was grouted from the surface to reduce potential water inflows. Because the ability of the conglomerate to standup during raise boring and then also stay open during shaft sinking was seriously questioned, the upper half of the Surge Shaft was sunk conventionally with muck being hoisted to the surface. When the bottom of the shaft was in better quality rock, the raise bore machine was lowered into the shaft, the pilot hole was drilled, and a 1.8 m diameter hole raised. The balance of the Surge Shaft was slashed to approximately 23 m diameter by drill and blast in 2 m lifts. The limited lift heights were used for the entire depth of the Surge Shaft because the conglomerate rock quality was never that good, and there was a significant shear zone encountered in the volcanic breccia. Blast hole drilling was completed with Tamrock Commando drills. Mucking was completed with a Caterpillar 330 hydraulic excavator. A summary of the Q values and RMR mapped during excavation follows:

Minimum Maximum Median

Q

RMR

0.34 6.3 1.9

16 61 33

7 GALLERY VENT SHAFTS A vent shaft was constructed at each dead end of the two upper grouting galleries. The shafts were raise bored to 1.2 m diameter, and then 0.9 m diameter steel pipes were lowered into the shafts and grouted in place. No instabilities were experienced. Pilot hole drilling and raise boring drilling rates advanced rapidly with muck excavation from the grouting galleries. 8 POWERHOUSE SHAFT 8.1 Figure 4. Surge Shaft excavation.

The San Roque Powerhouse is housed in a 48 m deep shaft that is 28 m wide and 80 m long. The design is

579 Copyright © 2004 Taylor & Francis Group plc, London, UK

Powerhouse Shaft design

09096-71.qxd

2/18/04

3:07 PM

Page 580

was installed from the top down as the excavation progressed. The three tailrace tunnels were excavated first followed by excavation of two raise bores from the outer tunnels. Powerhouse shaft excavation was completed using the drill and blast method with mucking through the raise bores and tailrace tunnels. Tamrock Ranger drills were used for the drilling and rock bolt installation and Caterpillar 320 and 330 hydraulic excavators were used for moving the muck to the raise bores. 9 CONCLUSIONS The seven shafts at the San Roque project varied in cross-section from 2 square meters for the vent shafts to 2,240 square meters for the powerhouse shaft and varied in depth from 48 m to 90 m. The shafts required a wide range of designs and excavation planning. The main lessons learned are:

Figure 5. Powerhouse rock support.

• effectively an underground cavern without the roof, as shown in Figure 5. The powerhouse was located in an intrusive diorite zone. The diorite is generally better quality than the rock at the rest of the site but there are persistent joint sets that dip into the upstream and downstream walls of the excavation. These joint sets formed wedges that had to be supported with rock bolts that extend beyond the lowest joint planes as shown in Figure 5. The rock support for the walls typically consists of 46 mm diameter Grade 1000 MPa epoxy coated, untensioned steel bars with 200 mm of reinforced shotcrete. The rock bolts were installed at a spacing of approximately 2 m by 2 m. Additional support for earthquake loading is provided by the substructure concrete and by the roof beams. 8.2

Powerhouse Shaft construction

• •

REFERENCE Humphries, R.W., O’Connor, E., Gertler, L., Warburton, W.L., Daly, J.J., and Funkhouser, M.R., “Rock Engineering at The San Roque Multipurpose Project” Waterpower XII, 2001

The powerhouse was excavated as a very large shaft using the raise bore and slash method. Rock support

580 Copyright © 2004 Taylor & Francis Group plc, London, UK

The raise bore and slash method provides an efficient and effective excavation method for a wide range of shaft sizes; Careful planning and scheduling is required when the raise bore and slash method is used to prevent blocking construction as access is required to the top and bottom of the shafts; The design-build method proved very effective as designs for the shafts could be adjusted as the construction progressed and experience was gained with the rock conditions and labor practices in the Philippines.

09096-72.qxd

2/18/04

3:07 PM

Page 581

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Ten years’ experience using roadheaders to bore tunnels for the Bilbao Metro J. Madinaveitia IMEBISA, Bilbao, Spain

ABSTRACT: Much of the Bilbao Metro runs underground. The Metro’s rock-bored tunnels often pass not far below blocks of housing. Roadheaders were used in excavation with high safety levels thanks to their ability to limit vibrations and control rock and earth response while working. After more than ten years’ work, we have become acquainted with a range of roadheaders and reached a number of conclusions about the best way of working with them.

1 INTRODUCTION The city of Bilbao is located in northern Spain, overlooking the Bay of Biscay, less than one hundred kilometers from the French border. The port of Bilbao has received vessels of more than 500,000 DWT and its industrial belt provides a home for, among others, steel mills, shipyards and chemical and petrochemical works. Bilbao sits in a narrow valley more than twenty kilometers long and barely five kilometers wide, the intervening space totally occupied by districts and peripheral towns of mid to high population density. By the late 1980s, traffic congestion was chronic and apparently incurable on the access roads from the coast and the urban towns along both banks of the river that flows through the center of the city before running down to the sea. As the construction of new roads with improved connections and access did little to palliate the problem, it was decided to study a mass rail transport system, the Metro. In response to the shape of the valley, the Metro route is itself Y-shaped. The two upper branches of the Y provide a service to the towns on either bank of the river, while the lower common stem section cuts underneath the city of Bilbao, its main districts, the financial center and the medieval quarter before heading beyond the city to districts upstream of the river. (Photo 1) During construction work, the Metro has had to cut three times beneath the river in navigable zones and has gone a further three times over the river upstream from Bilbao beyond the reach of the tidal flow. For

Photo 1. Aerial view of Bilbao.

every single crossing new techniques were used. All of them have been described in other papers mentioned in the reference section at the end. What distinguishes the Bilbao Metro is its functional nature, with identical stations that are extremely simple in design. As coined by British architect Norman Foster, the theme is that the station cavern is the heart of the system. The entire activity of the system is housed in a single 22,000-cubic-metre space hollowed out from the rock. There are no hidden spaces or long passageways, everything is linear and simple, with stations as close as possible to the surface and direct access galleries from the canopies that define and signal access points to the stations from street-level. Construction work and, above all, the experience gained in the use of roadheader machines is the subject of the present paper.

581 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-72.qxd

2/18/04

3:07 PM

Page 582

2 GEOLOGY AND GEOTECHNIC 2.1

Rock as modeling material

When design work for the Bilbao Metro began, the engineers were faced with two major challenges. First we wanted to build Metro stations very close to the surface to make access pleasant and attractive for passengers; second we had to bore the tunnels under a mature, consolidated city with buildings of all ages and types. Help was provided by the Bilbao subsoil, consisting of a rock base with just a two- to five-meter layer of loose soil above. So the most important part of our work would concentrate on boring the rock and doing so in the most tranquil manner possible. We had to analyze the best way to bore the tunnels while affecting the natural equilibrium of the terrain as little as possible. The solution came in the shape of roadheaders, which enable users to model boring sections gradually and, at the same time, protect and reinforce the terrain as and when required to restore the original equilibrium, thereby improving overall safety conditions. Preliminary studies of the terrain were important and from them it was possible to forecast good rock performance, the rock’s relative homogeneity and the relative importance of the broken zones and the presence of elements distorting the equilibrium of the base rock. The influence of water, or the absence of it, in the different strata and materials appearing in the probes was also analyzed in detail. 2.2

•

•

•

Characteristics

In geological terms, on the large scale the zone of Bilbao is set on a tectonic line limited to the north by an accident affecting the upper Cretaceous base known as the Durango fault and to the south by another accident called the Bilbao fault, with signs of dextral gashing. The recovery of continuous cores in probe campaigns enabled us to identify three perfectly differentiated sections:

•

Photo 2. In front of tunnel.

•

Cretaceous deposits (upper Aptian – lower Albian) making up the rock through which most of the tunnels are bored. Clearly marine in origin, of a carbonaceous basin with fine (marl and marly lime) sediments and occasional chalco-arenaceous bands. The dark colors are due to the presence of pyrites probably spread in a reductive environment in shallow waters linked to the open sea as suggested by the frequent traces of ammonites. Two sections can be clearly distinguished: massive calcareous, lutites, with sandy swathes that occasionally intersperse levels of sandstone with cloudiness structure and marls with marly limestone with frequent schistosity. (Photo 2)

582 Copyright © 2004 Taylor & Francis Group plc, London, UK

The rock mass strength of these materials varies between 30 and 50 Mpa in the central area of Bilbao, rising to between 50 and 70 Mpa in the hills on the left bank of the downstream section of the river. One very important feature of these marls (known locally as cayuela) is that they weather rapidly when subjected to the action of air, sun and water. To prevent this from occurring during the works process, excavated zones were protected with a layer of sprayed concrete added as soon as possible after excavation. Quaternary deposits leading to relatively unimportant coverings that, during construction work, only affect the meters closest to the surface of the access galleries or the elevator shafts. In some zones on the right bank of the river the Metro tunneled through these quaternary deposits. This required a different typology from the usual rock-bored tunnel, a false tunnel being used. Volcanic rock signaling its presence though generally weak bodies (measuring from a few centimeters to little more than a meter) of tabular geometry (diabase discontinuities). These are structures injected in the general rock mass and which settled into the materials of the lower Cretaceous. Their presence during boring gave rise to two types of problems. If they are unaltered, their tremendous hardness causes additional wear and tear on the cutting elements and reduces excavation performance. The rock mass strength of these materials exceeds 100–120 Mpa. But if these discontinuities have been attacked by seeping water, they may well decompose, giving rise to an almost liquid texture which in turn may lead to collapse and chimneys, with the accompanying risk to the equilibrium of the zone they are located in.

09096-72.qxd

2.3

2/18/04

3:07 PM

Page 583

Characterization

The sequence of materials found at the tunnel faces when boring was classified following the criteria of Bieniawski in his Rock Mass Rating (RMR) in four grades, used as a reference in defining the characteristics of the support required in each case. Rock type A Rock type AB Rock type B Rock type C

RMR  60 60  RMR  45 45  RMR  35 35  RMR

Initially only three types of rock were defined but during boring the need to establish an intermediate stage between rock types A and B became clear, so that safety levels could be maintained without increasing the thickness of the sprayed concrete. Reducing the time the rock was left unprotected was an important move and influenced the quality of the support achieved. 2.4

Works in loose terrain

Areas close to the river, particularly on the right bank, had to be built in loose earth, silts and sands, which meant that a different construction system was used. False tunnels were built in the shelter of concrete diaphragm walls, some of which went more than 20 meters deep. The problem needing to be solved in such cases had to do with the position of the underground course of the river and the way the construction of transversal walls facilitating the damming of water might influence its contention. This phenomenon might have led to variations in the water table upstream that could have affected the foundations of the oldest buildings. 3 TUNNELS AND STATIONS 3.1

Tunnel geometry

If the tunnel is close to the surface and the layout of the streets allows, tunnels are built in false tunnel sections. This situation occurred very rarely in Bilbao and in general caused serious problems affecting the city’s services. In economic terms, the sections done using false tunnels were more expensive than the rock-bored tunnels, specifically owing to the replacement of the services affected. When the tunnels went deeper and the route cut through rock or underneath buildings, bored tunnels were designed. The tunnels of the Bilbao Metro are single-section for two tracks. Basically, they measure 8.30 meters high by 10.20 meters wide. The height is conditioned

by the overhead power line, unavoidable in the openair zone in rural areas near the coast on the right bank. The width was conditioned not only by the need to maintain dynamic clearances for two train units passing each other in opposite directions but also by the decision to build two service corridors, one on either side of the tunnel. These corridors serve a twofold purpose. They enable passenger evacuation if a train stops accidentally inside the tunnel and also provide protection for the service cables and tubes fitted underneath. This type of single tunnel has the advantage of reducing the work done at street level to a few, specific points. They also coincide with station areas and can be integrated into the corresponding works zones, something that requires detailed treatment of the possible effects on road traffic, the people living nearby and environmental conditions. The final section measures 62 square meters. 3.2

The stations of the Bilbao Metro were the result of a long process of reflection and dialogue between two teams of professionals. A team of architects directed by Norman Foster put forward its best ideas on the functionality and final simplicity of the design, while a team of engineers provided its knowledge of the terrain and construction techniques. The team of architects gave priority to two features: the station as a single-space cavern housing all services and making the accesses as immediate as possible. A further condition was that the finishes should not be an addition to the works but should be a structural part of the lining. The solution used fulfils all these requirements. A cavern station is bored with a transversal section of a little over 220 square meters and a length of around one hundred meters, to assure a platform length of 90 meters. Accesses are straight and in general very short. When someone decides to take the Metro and approaches the canopy (popularly known as a “fosterito”, after their designer, Norman Foster) at street level he comes to an escalator that takes him down to the distributor or mezzanine level. Located inside the station cavern, the distributor hangs from the vaulting like a swing to maintain the sensation of lightness achieved, despite being underground. From there, broad flights of steps give the passenger access to either platform. The entire station is finished in prefabricated concrete panels measuring 1.20 meters by 2.40 meters. The perimeter of each panel forms a groove which, when put into place on the walls, forms a system of joints. And it is this system of joints that the architects used to break up the general surface configuration

583 Copyright © 2004 Taylor & Francis Group plc, London, UK

Station concept

09096-72.qxd

2/18/04

3:07 PM

Page 584

and provide a visual accompaniment for passengers as they enter the Metro. Appearing first in the access galleries, the system of joints maintains its continuity throughout the remaining spaces. The transversal joints, generally vertical, help to avoid any sensation of claustrophobia typical of closed spaces. Once finished, station dimensions are as follows:

• • • • •

width at platform level distance between platform edges width of each platform height free from rail level to mezzanine level height free from mezzanine level to a tunnel crown level

16.00 m. 6.00 m. 4.85 m. 5.50 m. 4.00 m. Photo 3. Sprayed concrete equipment.

3.3

Construction process

A variant of the New Austrian Method developed while work was in progress was used in the execution of the Bilbao Metro tunnels and station caverns. Knowledge of the local rock and its response to the excavation phases designed allowed us to adjust the method to achieve the best results. To make the best use of the bearing capacity of the rock it is vital to use boring procedures that minimize the damage done to the rock mass. This is done using roadheaders. As is done to reduce the time the bored sections of rock are exposed to the action of the wind and water, concrete is sprayed as soon as the work sequence permits. (Photo 3) With these premises in mind, the tunnels are bored in two sequences, heading and bench. The heading has a section of about 40% of the total section and in each case is adjusted to the potential of the roadheader available. As far as is possible, once the roadheader is at the face, we try to ensure that it is not displaced laterally, boring from a single position thanks to the range of its arm. We have found that this way performance improves and cuts are achieved with greater precision. Boring the station is much more complex not only because it is bigger but also because the crown of the excavation gets much closer to the foundations of the buildings above. The stations of the Bilbao Metro are as close as possible to the surface to attract passengers and also, with its bowl shape, to facilitate the recuperation of the system’s power. Trains climb a ramp as they arrive at the station to facilitate braking and run down a ramp on leaving to facilitate acceleration without recourse to greater consumption. Station construction is divided into five phases. The first is the most complex, consisting in boring a 30square-meter pilot gallery, the crown of which coincides with the crown of the definitive cavern. This is basically a wide diameter probe, which enables us to

identify visually all the elements that will appear in the future excavation. In Bilbao, we use the same roadheader for the pilot gallery that is to be used in the rest of the work, although limiting its boring action to the minimum dimensions compatible with rock clearance. In the upper part a reinforced beam is constructed and anchored that will finally form part of the support section. The rest of the walls are protected with sprayed concrete and whenever jointed zones appear and there is a risk of a slip they are reinforced with fiberglass bolts. Fiberglass is used to limit the difficulties for subsequent roadheading, as in general these zones have to be excavated in the following phases. Once the information supplied by the pilot gallery has been studied, excavation work on the following phase begins. This involves boring the shoulders of the tunnel alternately, on one side and then the other. This phase, which takes the excavated section to roughly 50% of the total, requires complete support treatment. Depending on the class or type of rock identified at each face the thicknesses of sprayed concrete vary between 15 and 30 cm. The concrete is sprayed on in several runs or layers and is reinforced with metal fiber. When there is a risk of blocks forming, the sprayed concrete is reinforced with one or two layers of electrowelded meshing. The third excavation phase is the simplest and involves excavating a trench in the center of the cavern. Damage to the rock is minimized by leaving berms or ledges untouched on either side of the trench and, in general, additional support measures do not have to be considered. In any case, as we know at all times what the rock response will be, from the automatic measurements taken by extensometers placed in the interior and exterior of the cavern and from the regular manual measurements taken, it is possible to reinforce any insufficiently stable sections with new bolts or extra sprayed concrete as required.

584 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-72.qxd

2/18/04

3:07 PM

Page 585

we had then to select the most suitable one from those available on the market when work began. There were three types of selection criteria:

•

•

Photo 4. Full excavation cavern.

• The fourth phase of excavation involves eliminating the tunnel haunches, which also obliges us to give continuity to the support. The detailed knowledge of the rock response available to us at this point of the works means further relatively trouble-progress on execution can be made to finish with the fifth and final phase, which consists in excavating the countervault that has been damaged by the passage of heavy machinery and beginning the process of concreting for the lining ring. (Photo 4) 4 EQUIPMENT USED 4.1

Possibilities

Three possible rock-boring systems were analyzed for the tunnels of the Bilbao Metro. The characteristics of the rock were suited to excavation with explosives, but, being an urban project with high population density overhead, the inevitable vibrations caused by the explosions could well annoy the local people, and the procedure was ruled out. Tunnel Boring Machines (TBM) were ruled out because of the extra cost involved in the additional excavation required, around 30% extra, to enlarge the tunnel section from the strictly necessary to a circular section all round for clearances. So it was decided to use roadheaders to bore the tunnels, particularly in view of their greater flexibility and limited effects on the rock mass. The rock in the subsoil of Bilbao reacts very well to the precise cutting that roadheaders are capable of, while greatly limiting the range of the vibrations caused during operational phases. 4.2

Selection criteria

Once it was decided that the roadheader was the machine for the job (two being used in each tunnel, working from either end and meeting in the middle),

4.3

Machines used

The Bilbao Metro has been constructed in sections awarded to different construction companies, which means that a wide range of roadheaders has been used. While each company chose the roadheader in line with the criteria mentioned in the preceding paragraph, it also added its own technical (prior experience) and economic (using its own or leased machinery) perspective. 4.3.1 Common section (1990–1997) Just over ten (10) kilometers of tunnel were bored for the shared section of the Bilbao Metro. This section cuts from the west under the successive late 19th and early and mid 20th century city expansion schemes to the medieval quarter (founded in the year 1300) and continues eastwards under the hill at Begoña and beyond. In this route it crosses underneath the river three times, although only once was the bored tunnel section maintained, thanks to the presence of healthy rock and to the fact that the distance between sections enabled us to recover station level after the underriver descent. Five independent works were carried out using the following boring equipment. (see Table 1) In the common section, up to ten (10) roadheaders of five (5) different makes were used. Roadheader weight varied between 40 and 76 tons while head power varied from 100 to 225 KW. The Paurat roadheader bored a tunnel almost two kilometers long, including two cavern stations, with hardly any problems. The performance of the Alpine roadheaders greatly depended on their weight, as the ATM-70 (Photo 5) performed very well and worked virtually without breakdowns, while the ATM-50 model performed poorly. Despite being one of the lighter machines used, the Eickhoff performed satisfactorily with few breakdowns. The other machines

585 Copyright © 2004 Taylor & Francis Group plc, London, UK

The rock, its characteristics and, in the first place, its rock mass strength, this being a basic datum for deciding the type and weight of the machine required. In Bilbao the rock is very massive which means that the tenacity of the rock, rather than the joints or the stratigraphy, is decisive. The tunnel geometry, in two respects. One, the transversal section, which is worked better if the machine chosen can execute the excavation phase without having to move from its initial position, and two, the ramp or ascending gradient where the machine weight component is limited and with it the effectiveness. The operator, who has the capacity to strike up a dialogue with the rock and get the best results through his roadheading skill and technique.

09096-72.qxd

2/18/04

3:07 PM

Page 586

Table 1.

Table 2.

Model

Type

Power (Kw)

No. Units

Weight (T)

Paurat E-208 Alpine ATM-70 Alpine ATM-50 Westfalia FLGT–100 Westfalia WAV-200 Eickhoff ET-250 L Salzgitter STM-160H

Milling Ripping

225 200

2 2

76 62

Ripping

100

1

40

Ripping

100

1

40

Ripping

200

1

74

Ripping

200

1

40

Milling

160

2

45

Model

Type

Power (KW)

No. Units

Weight (t)

Paurat E-208 Paurat E-242 Alpine ATM-105 Eickhoff ET-350L Eickhoff ET-450L Icutroc ATM-105/10 Paurat E-134

Milling Milling Ripping

300 300 300

1 1 2

110 120 100

Milling/ Ripping Milling

250

1

90

300

1

110

Ripping

300

1

120

Ripping

300

1

70

Photo 5. Alpine roadheader ATM-70.

Photo 6. Icutroc roadheader ATM-105/10.

used had problems with the peripheral equipment, mostly the material loading elements, which spoilt their boring performance. The experience can be summed up as follows:

Eight (8) machines in all, heavier than the ones used in the common section, in line with the experience gained in the previous work. It should be noted that daily advance performance improved in comparison with the common section, except for the Paurat E-134, which was clearly not heavy enough. Special mention must be made of the Icutroc ATM105/10 (Photo 6) which included a number of improvements in the cutting head with sprayed water cooling in the bits and the capacity to modulate its attack velocities depending on rock characteristics. The machine’s manageability and flexibility also improved to the point where it is now well ahead of the other machines used to date on tunnel boring for the Bilbao Metro.

• • • •

Machine weight is decisive. Good performances are out of the question if the machine is not heavy enough for the characteristics of the rock to be bored. The longitudinal cutting system, milling, is suitable for tackling diabase discontinuities and intrusive rock of greater abrasiveness and strength than the marls. The transversal cutting system, ripping, is better for the theoretical tunnel section, reducing excess excavations. The machines incorporating peripheral elements such as loaders, belts, etc. cause stoppages for breakdowns unconnected with excavation operations.

4.3.2 Line 2 (1995–2002) The following roadheaders were used for the work on the eight (8) kilometers bored so far on Line 2: (see Table 2)

5 FACTORS TO BE CONSIDERED IN LIGHT OF THE EXPERIENCE IN BILBAO 5.1

It is clear that the first thing to think about when planning rock boring is the characteristics of the rock

586 Copyright © 2004 Taylor & Francis Group plc, London, UK

Rock characterization

09096-72.qxd

2/18/04

3:07 PM

Page 587

Photo 7. Ripping system.

Photo 8. Bad visibility, bad efficiency.

involved. Its characteristic rock mass strength is decisive for selecting both the equipment and the work system. However, rock abrasiveness and, as Bilbao has proved, tenacity need to be carefully analyzed; being relatively high it impeded rock breakage from hammering, absorbing much of the machine’s power without fracturing. This phenomenon was very important in Bilbao, where the rock is massive, without joints and with very little water.

this respect, the use of the Icutroc with a direct bit cooling system confirmed the advantages of ventilating the work area correctly. In general many contractors are reluctant to constantly move ventilation equipment to put them as close as possible to the work point, which means that ventilation at the face gradually deteriorates. Visibility becomes worse, as does the bit cooling action, as dust concentrated in the zone gets into the water. (Photo 8)

5.2

5.4

Cutting head type

The choice between the longitudinal milling head or the transversal ripping head led to many tests at natural scale without us being able to introduce an evident improvement for either type. As stated above, the milling or longitudinal system apparently achieved better progress with slightly lower consumption of bits in the case of intrusive rock of greater abrasiveness, although a precision cut was more difficult to achieve. To adjust better to the theoretical sections and not over-excavate in the Bilbao marls, the ripping system was clearly more effective. (Photo 7) 5.3

Importance of weight and ventilation

Using heavier machines for boring the tunnels of the Bilbao Metro’s Line 2 was a direct consequence of our experience in boring the common section. The machines used in the earlier work weighed less than 75 tons, while the ones in use on Line 2 weigh anything up to 100 tons. The weight, rather than the power, of the machine proved decisive, although it is in general true that heavier machines are also more powerful. One interesting experience in boring the Bilbao Metro is the substantial influence on machine performance of the immediate environmental conditions and, in particular, bit ventilation and cooling. In

However, in the light of our experience in Bilbao, it is quite clear that the operator, the man in charge of the operation, is the most decisive factor in boring operations that combine high performance with lower levels of bit wear and tear. If the operator adapts to the machine and achieves a constructive dialogue with the rock, then the operation will be a success. The rock permanently transmits its “vibrations” and if the operator can read them correctly and act in consequence performance can easily surpass anything up to double the average figures achieved by other operators using the same machine on the same rock and in similar work conditions.

6 CONCLUSION The execution of twenty (20) kilometers of tunnel in very massive, highly tenacious marly rock of mid to high rock mass strength was a major challenge within the works on the Bilbao Metro. This was underground work in an urban area, mostly under a wide range of buildings of different ages, many of them centenarians several times over. Boring rock in Bilbao has been a job of precision modeling thanks to the use of highly flexible roadheaders capable of adapting to the conditions of

587 Copyright © 2004 Taylor & Francis Group plc, London, UK

Dialogue between operator and the rock face

09096-72.qxd

2/18/04

3:07 PM

Page 588

very healthy, largely unjointed rock with barely defined stratification. Although the weight of the machine is important, the skill of the operator, his ability to “understand” the language of the rock being bored, is ultimately decisive. REFERENCES N.B. 1995. Under river tunnel construction in the Bilbao metro. Paper presented at Stuva-Tagung’95, published in Congress Minutes. pp. 73–77. Alba Fachverlag. GmbH. Stuttgart. N.B. 1996. Cavern Station Concept for the Bilbao Metro. Paper presented at ITA’96. Washington. Pub. pp. 339–344. Rotterdam. Balkema.

N.B. 1998. Discrete Supports in Submerged Tunnels. Paper presented at IABSE Colloquium. Stockholm. Sweden. Pub. pp. 367–372. IABSE. Zurich. Madinaveitia, J. 2000a. Túneles urbanos perforados en roca: el caso del metro de Bilbao. Ingeotúneles. Libro 3: Ingeniería de Túneles. Chap. 16. pp. 521–556. Madrid. López Gimeno. Madinaveitia, J. 2000b. La maquinaria en las obras del metro de Bilbao. In “Tunelmaq 2000”. Universidad de Madrid. Escuela de Ingenieros de Minas. Madrid. N.B. 2001. Rock responsing during construction of cavern station at Bagatza. Paper in IS-KYOTO 2001. pp. 249–252. Rotterdam. Balkema. N.B. 2002. Wet mix sprayed concrete for support in the Bilbao metro. Paper presented at 4th Symposium on sprayed concrete. DAVOS. N.B. Papers included in the bibliography are available on the IMEBISA web page.

588 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-73.qxd

2/18/04

3:09 PM

Page 589

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Rio Piedras Project, San Juan, Puerto Rico Brian Fulcher Project Manager, Kiewit Construction Company, Omaha, Nebraska

Niels Kofoed Project Engineer, Kiewit Construction Company, Omaha, Nebraska

Paul Madsen Tunnel Engineer, Kiewit Construction Company, Omaha, Nebraska

Matthew Bartlett Field Engineer, Kiewit Construction Company, Omaha, Nebraska

ABSTRACT: This paper discusses the challenges to complete the construction of two large underground metro stations and connecting tunnels in the Tren Urbano system. The work included the installation of specialized mechanical and electrical systems, extensive architectural finishing as well as coordination with systemwide power and controls. A unique feature of this project was the simultaneous design and construction of the structural/civil/architectural/mechanical/electrical works by KKZ/CMA separately from the power and control systems under the management of the Owner, its consultants and other contractors. Completion of this project comes after almost six years of continuous design and construction in a dense urban environment. The project was confronted by many challenges that could only be met with the combined efforts of the Joint Venture team in close cooperation with the Owner and its Systemwide Contractor.

1 OVERVIEW OF THE PROJECT Phase I of the Tren Urbano System located in San Juan, Puerto Rico is a 17.2 km route designed to connect suburban areas of Bayamon, Guaynabo, and Rio Piedras to commercial areas in Hato Rey and Santurce. There will be 16 stations located in commercial and residential areas having the greatest need for mass transportation. Future expansion of the system will connect to the government center in Minillas, the international airport in Isla Verde, Old San Juan and to the extensive suburban areas in Carolina. The system utilizes electrically-powered trains running on an alignment with at-grade, elevated and underground sections. See Figure 1. The Rio Piedras alignment section is approximately 1,500 meters long and was the only section to be constructed entirely underground. It is located in the suburban district of Rio Piedras and runs under Ponce de Leon Avenue through a densely built-up area and in front of the University of Puerto Rico. Beginning at Highway PR-3 and descending underground from an elevated section, the alignment runs north to Highway

PR-17 before emerging at-grade and connecting to the adjoining alignment section to bridge over the highway. See Figures 2 and 3. Two underground stations were included in this Design-Build Contract to service patrons in Rio Piedras and the University of Puerto Rico (UPR). Considerable quantities of additional works were also included in the contract for upgrades to the local community and aging utility systems. All work was originally scheduled to be completed in 44 months following Notice-to-Proceed in April 1997. 2 OWNER AND DESIGN-BUILD CONTRACTOR 2.1

The project Owner is the Puerto Rico Highway and Transportation Authority (PRHTA) who established a Tren Urbano Office (TUO) to manage the construction of the transit system. Tren Urbano administered the Rio Piedras Design-Build contract with the assistance of a General Management, Architectural and

589 Copyright © 2004 Taylor & Francis Group plc, London, UK

Project Owner

09096-73.qxd

2/18/04

3:09 PM

Page 590

Rio Piedras Project

Figure 1. Aerial photo of the Tren Urbano System running through suburban San Juan from Bayamon to Santurce. The Rio Piedras Project is the only underground portion of the system and is located adjacent to the University of Puerto Rico.

Engineering Consultant (GMAEC). All Alignment Section contracts were set-up as Design-Build projects. Additionally, TUO had a concurrent Systemwide Design-Build Contractor responsible for numerous scopes that included, power, communications, train control and signaling as well as the supply of the transit cars. This contractor was Siemens Transit Team (STT). 2.2

Design-Build Team

The Rio Piedras Project, Alignment Section 7 for the Tren Urbano System was awarded to a Joint Venture composed of Kiewit Construction Company (sponsor) based in Omaha, NE, Kenny Construction based in Wheeling, IL and H.B. Zachry Company, based in San Antonio, TX along with CMA Architects and Engineers, based in San Juan, PR. CMA teamed-up with Sverdrup Civil Inc. (now Jacobs Engineering) for the station design, Jacobs Associates for the tunnel design and Woodward Clyde Consultants (now URS) for the considerable geotechnical, instrumentation and monitoring aspects of the project. Anil Verma Associates from Los Angeles was retained as the project architect for the stations. The Design-Build Team became known as KKZ/CMA for the Rio Piedras Project of the Tren Urbano System. 2.3

Subcontractors

Kiewit-Kenny-Zachry/CMA (KKZ/CMA) engaged several well-established subcontractors to assist with

the design and construction of specialized portions of the work. The extensive mechanical-electrical systems were subcontracted to Bermudez & Longo, SE, while A.H. Beck Foundation Co. performed all pile drilling. An on-site concrete batch plant was installed and operated by RDG Material Resources. Waterproofing was supplied and installed by WISKO America. Soletanche Bachy performed fracture and compensation and compaction grouting for building settlement control. Over the course of the project, several dozen other local and off-island subcontractors were involved with the project. 3 DESIGN-BUILD CONTRACT CONDITIONS 3.1

590 Copyright © 2004 Taylor & Francis Group plc, London, UK

Contract features

This Design-Build contract had many unique and often challenging features. TUO initially engaged Daniel Mann Johnson & Mendenhall (DMJM), Fredrick R. Harris, Eduardo Molinari y Asociados and Barrett & Hale Engineering to prepare 30% design drawings for the Bid Documents. Bid specifications were also prepared to the extent that they would become part of the Contract. There was no baseline Geotechnical Design Summary Report since at the time of the bid, this document was to be prepared by the successful bidder. The contract documentation was voluminous and clearly required reading as a whole to fully understand the scope for the design and construction. It was not

09096-73.qxd 2/18/04

Copyright © 2004 Taylor & Francis Group plc, London, UK

3:09 PM Page 591

591

Figure 2. Overall plan layout of the stations and tunnels for the Rio Piedras Project over a distance of 1.5 km.

Figure 3. Geological profile of the ground conditions for the stations and tunnels on the Rio Piedras Project.

09096-73.qxd

2/18/04

3:09 PM

Page 592

initially anticipated that there would be many substantial changes to the original design concepts initially described in the Bid Documents. It was later realized that numerous and far-reaching changes were needed and requested by the TUO as well as its Systemwide Contractor as the design and construction of the stations and tunnels progressed. The contracting conditions were extensive and well suited to the complexities, risks and scope of the project. They included the following:

• • • • • •

Disputes Review Board (DRB) Escrowed Bid Documents (EBD) Differing Site Conditions clause (DSC) Changes and Delay clauses Partnering agreement Owner Controlled Insurance Program (OCIP).

3.2

Contract Schedule

Following the bid and award of the Contract, KKZ immediately mobilized to the site. An aggressive and effective early job start-up was essential to meet the Contract Schedule as well as all Interim Milestone Dates. The Rio Piedras Design-Build Contract was performed concurrently with another significant Design-Build contract for all Systemwide Elements and Interface Control between the project and Tren Urbano’s system operating facilities. This relationship proved to be a significant challenge and one that may be reviewed on future similar work where there are complex, inter-linked and inter-dependent systems designs and facilities that need close coordination to fulfill the project schedule.

that 33% of the value was assigned to underground work whereas 67% was reserved for station design, construction, utilities and restoration. 3.4

Progress payments

Payments under the Contract were initially linked to a “cost-loaded” CPM schedule. This was subsequently adjusted in light of the many significant changes to the scope and the corresponding project CPM schedule that was initially very complex. It contained five critical paths, each leading to an Interim Milestone Completion Date. Overall, the Rio Piedras Project was very large and extensively intertwined with complex and challenging issues. To a great extent, the work including design was to be completed under compressed schedule conditions with no room for indecision and unforeseen delays.

4 DESIGN-BUILD PROJECT SCHEDULE

The Contract was bid as a fixed price with a Schedule of Values for the major items only. In that it was a Design-Build contract, no quantities for measurement and payment could be initially presented in the Bid Documents. Table 1 below lists the major items of work as well as the corresponding Contract Price breakdown (and percentages) associated with each. From a further breakdown of this summary, it was apparent

The Rio Piedras Contract contained five Interim Milestones and a Final Project Completion Date. Table 2 below summarizes the scopes and durations for the design and construction required to meet these dates. Under the terms of the Contract, the Owner could assess Liquidated Damages for delays to each and every Interim Milestone Date. The Contract Schedule was considered aggressive in light of the type and quantities of work under contract. This project, like many in the Tren Urbano system, was the subject of numerous changes in scope. Some were in response to design changes and others were due to additional features built into the facilities that were not originally anticipated. Delays to the performance of the work and impacts to the project schedule will be described below. Overall, KKZ’s project CPM Schedule could be characterized as a “fair weather” schedule and one that needed constant and careful attention to adjust with revised scopes and logic as well as periodic as-built data updates in order to be a useful tool for the overall management of the Contract.

Table 1. Contract Price breakdown summary.

Table 2. Contract Schedule Interim Milestones – Planned Dates.

3.3

Contract Price

Description

Price

Percent

Milestone

Duration

Dates

Management and administration Quality control and assurance Utilities and street restoration Hazardous materials allowance Design: stations & tunnels Construction: stations & tunnels Total

$ 28,712,743 $ 2,220,000 $ 19,279,640 $ 200,000 $ 16,895,520 $ 158,292,097 $ 225,600,000

12.7 1.0 8.5 0.1 7.5 70.2 100.0

Guideway completion Rio Piedras Station structural UPR Station structural Rio Piedras Station architectural UPR Station architectural Contract final completion

1,068 Cal Days 1,160 Cal Days 1,160 Cal Days 1,343 Cal Days

21 Mar 00 21 Jun 00 21 Jun 00 21 Dec 00

592 Copyright © 2004 Taylor & Francis Group plc, London, UK

1,343 Cal Days 21 Dec 00 1,343 Cal Days 21 Dec 00

09096-73.qxd

2/18/04

3:09 PM

Page 593

5 CIVIL SCOPES OF WORK The Rio Piedras Project can be described a complex and very challenging project. It required the considerable and combined talents of the Joint Venture design and construction staffs as well as beneficial input from the Owner, Tren Urbano, the local community and the University of Puerto Rico. The subsurface site conditions were not well evaluated prior to the bid. There was a potential for Differing Site Conditions, Changed Conditions and Extra Work. The Owner recognized these conditions in the Contract Documents. Rio Piedras community is a densely populated and built-up residential, commercial and university area, dating back several hundred years. It has variable ground conditions, uncharted utilities and building foundations and a high water table to further complicate the working conditions. In response, KKZ’s design and construction staffs utilized many traditional and innovative methods to complete the construction and produce a high quality, distinguished facility. 5.1

5.1.1 Stacked drift tunnels The stacked drift tunnels were easily the most challenging and riskiest portion of the tunneling operations on the Rio Piedras Project. They were designed to form the supporting arch over the station and were constructed in onerous conditions confined by the surface site restrictions that allowed little space for equipment while imposing critical requirements for building settlement control. The initial grouting gallery was started in October 1997. This was used to drill and inject compensation grout for 15 drifts that would be excavated under the congested built-up portion of suburban Rio Piedras.

Tunnels

Tunneling was a major element of the scope of the work followed by installation of elaborate below-grade concrete structures for the stations and entrances. The tunneling operations were substantial, as will be briefly described below, and required over two years to complete.

Figure 4. Interior of a stacked drift tunnel with mining crew.

STACKED DRAFT TUNNEL

MEZZANINE LEVEL

ARCHITECTURAL FINISH

BAYAMON R/T

BAYAMON L/T

Figure 5. General arrangement of a stacked drift tunnels.

593 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-73.qxd

2/18/04

3:09 PM

Page 594

Figure 7. Installing the concrete and waterproof lining. Figure 6. General arrangement of NATM tunnels.

All 15 drifts were excavated and backfilled with concrete by February 2000. Mass excavation of the core of the station began immediately afterwards and was completed in June 2000. See Figures 4, 5 & 14.

• • • • • • •

Stacked drifts: nominally 3.0 3.0 150 m long (15 each) Grouting gallery: nominally 2.5 2.5 150 m long (one only) Initial support with ribs and timber lagging; later shotcrete was used for rib lagging Backfilled with concrete with interconnected bolted steel ribs. Ground conditions were sandy silts with clay Dewatering with wells (limited coverage) Fracture and compensation grouting for building settlement control.

5.1.2 NATM tunnels Four NATM tunnels were constructed for the project. Two of the NATM tunnels connected the south end of the Rio Piedras Station with the adjoining alignment section (after going through lengths of cut and cover tunnels). The two other NATM tunnels dead-ended at approximately 90 m from the station – to be connected in a future Phase II expansion of the Tren Urbano System – to the Carolina suburbs. All NATM tunnels were constructed with a top heading and bench excavation sequence through silty sands and some clay. Initial support was installed using forepoling, shotcrete, lattice girders and wire mesh. The final lining (to a horseshoe section) was completed with cast-in-place concrete. It is noteworthy that these tunnels have different geometrical sections whether on a tangent or curve portion of the alignment. See Figures 6 and 7. Details of the tunnel conditions are listed below.

•

Two tunnels: nominally 6.5 3 6.5 3 110m long each (37m2 sectional area)

Figure 8. Hole-through of first EPBM tunnel.

• • • • • •

5.1.3 EPBM tunnels Twin EPBM tunnels were constructed between the Rio Piedras Station and the University of Puerto Rico Station, a distance of 430 m each. They have a serpentine layout and were excavated with a 6.450 m (21-2) diameter Lovat EPBM TBM and supported with a single-pass precast concrete segmental, bolted and gasketed tunnel lining. See Figures 8 and 9. Details of the tunnel conditions are listed below. The final hole-through was in August 1999 followed by the installation of cast-in-place concrete

594 Copyright © 2004 Taylor & Francis Group plc, London, UK

Two tunnels: nominally 6.5 6.5 90 m long each (37 m2 sectional area) Initial support with forepoling, shotcrete, lattice girders and wire mesh Final lining completed with cast-in-place reinforced concrete Ground conditions were sandy silts with some clays Dewatering with wells (limited coverage) Compensation grouting for building settlement control.

09096-73.qxd

2/18/04

3:09 PM

Page 595

inverts and walkways. A hand-mined, NATM supported cross-passage was excavated also, followed by the installation of fire-life safety systems. The Systemwide Contractor installed rail in 2002.

• • • • •

Two tunnels: nominally 6.450 m diameter 430 m long each Supported with single-pass precast concrete, bolted and gasketed liner Cross-passage between tunnels – one only Ground conditions were sandy silt with clays Compensation and compaction grouting.

5.1.4 Cut and cover tunnels Four cut and cover tunnels were constructed for the project. Two tunnels run south from a connection with the NATM tunnels located south of the Rio Piedras Station. They emerge from below grade to a portal and make a connection with the adjoining alignment section south of Highway PR-3. They were each 6.0 m wide 7.0 m high (interior).

These tunnels also convey the 13.2 kV power transmission system to the station, for traction power and station operations. Two other tunnels run north from the UPR Station. They also emerge from below grade to a portal and make a connection with the adjoining alignment section south of Highway PR-17. These tunnels also convey the 13.2 kV power transmission system to the station for traction power and station operations. See Figures 10, 11 and 12.

• • • • • • •

South – two tunnels: nominally, 6.0 m wide 7.0 m high 130 m long each North – two tunnels: nominally, 6.0 m wide 7.0 m high 280 m long each Phased construction needed to suit traffic management and temporary bridging needs Support-of-Excavation using drilled piles, struts and walers and timber lagging Ground conditions were sandy silts with some clays Dewatering with wells as needed Instrumentation for ground settlement monitoring and building settlement control.

5.2

Figure 9. Completed EPBM tunnel with rail & utility systems.

Rio Piedras Station

5.2.1 Excavation The Train Room of the Rio Piedras Station was excavated with a top heading and bench (236 m2 sectional total area) operation that removed 35,400 m3 of spoil in spring 2000. Once excavated it was one of the world’s largest underground openings of this type ever constructed. This resulted in a cavern 19 m wide by 16 m high by 150 m long with from 3 to 5 meters of cover to the foundations of the overlying buildings. See Figures 13 and 14 that illustrate the arrangement of the stacked drifts to form the opening of the station

Figures 10, 11 and 12. Forming and pouring the north cut and cover tunnels from the UPR Station area to Highway PR-17.

595 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-73.qxd

2/18/04

3:09 PM

Page 596

Table 3. Rio Piedras Station – Shaft quantity summary. Description

South Shaft

North Shaft

General layout Main shaft dimensions Entrance dimensions Nominal depth Overall footprint Shored wall area Excavated quantity CIP concrete quantity

‘L’ layout Trapezoidal 155 m L 27 m W 49 m L 21 m W 27 m L 27 m W

49 m L 9 m W

25.9 m 1,300 m2 4,900 m2 45,000 m3 14,700 m3

24.4 m 1,360 m2 3,400 m2 35,500 m3 10,000 m3

Figure 13. Layout for stacked drifts in Rio Piedras South Shaft.

Figure 15. Multi-level Support-of-Excavation (SOE) system.

Figure 14. Rio Piedras Station Train Room excavation.

Train Room. A structural mud mat was poured sequentially with advance of the lowest bench excavation throughout the station. The Rio Piedras Train Room could only be accessed from the large shafts excavated at each end. Existing urban development prevented other alternate access. These shafts were formidable excavations and required significant quantities of Support-ofExcavation. They were designed to enclose the below grade portions of the station entrances as well as the essential service rooms needed for station operations and for systemwide power distribution. The project CPM Schedule for the Rio Piedras Station started and ended with these shafts – known as the North and South Shafts. Table 3 lists the principle quantities for the construction. Figure 15 illustrates the size and conditions under which these shafts were constructed. 5.2.2 Concrete work A heavy invert slab (2.5 m thick) was placed followed by CIP exterior walls, platform walls and slab. Substantial interior wall and slab work was needed in the

adjacent service rooms located in the North and South Service Rooms adjoining the Train Room. These service rooms required 24,700 m3 concrete and in excess of 36 months to complete (to grade) before the follow-on mechanical, electrical and architectural systems could be installed. Largely due to the schedule delays experienced in the excavation of the Stacked Drifts for the Train Room, both the North and South service room areas were severed from the concurrent mining operations for the Train Room. Work in these areas progressed independently in an effort to mitigate delays. 5.2.3 Station systems The Rio Piedras Station was designed with considerable mechanical and electrical systems in support of the station operations, traction power and fire-life safety systems. Mechanical/electrical systems for Platform Edge Doors were installed and included an extensive chilled water system and HVAC ducting throughout the station. A 13.2 kV power substation was also incorporated into the design with provisions for the future expansion. See Figure 16.

596 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-73.qxd

2/18/04

3:09 PM

Page 597

Figure 16. Interior Rio Piedras Train Room under mezzanine. Figure 18. Initial station excavation stage and SOE system.

power substation with a corresponding 480-volt secondary distribution to station equipment. A traction power substation was also included. 5.2.6 Station work schedule Work on the Rio Piedras Station began with the Support-of-Excavation system in December 1997. All structural work below grade was completed in fall 2001 with the placement of mezzanine slabs in the Train Room and the roofs of the South and North Shafts (to grade). This achievement fulfilled the Interim Milestone in the Contract Schedule. Extensive architectural finishing followed for the next 20 months until the entire station was completed in spring 2003.

Figure 17. Mezzanine level in Rio Piedras South Shaft.

5.2.4 Station finishing The station was designed with virtually all public areas covered with designated finishes that included modular porcelain enamel-coated wall panels, perforated stainless steel ceiling panels, ceramic floor and wall tiles, granite stair treads, stainless steel trim, handrails and cover plates. See Figure 17. This station was air-conditioned. HVAC was a major portion of the re-designed station cooling system and required considerable logistics to install as the civil work was underway. Similarly, the escalators had to be installed concurrently with the civil works – to provide access for installation of modular units in pre-engineered wells. See Figures 16 and 37. 5.2.5 Systemwide interfaces TUO’s Systemwide Contractor (and subcontractors) undertook many scopes of work within and around the station following the completion of the civil works. Various areas were progressively turned over to Siemens Transit Team (STT) for the installation of their systems, equipment and facilities. The most essential was the installation of the 13.2 kV primary

5.3

5.3.1 Excavation The Train Room and service rooms for the UPR Station were excavated as open cut operations supported with soldier piles, pipe struts and timber lagging. Excavation for the adjoining cut and cover tunnels leading north from the station to the North Portal were similarly excavated in a continuous operation that burrowed under numerous utility systems and two major roadways. No tiebacks were used in the Support-of-Excavation system. The excavation removed a total of 176,000 m3 of spoil for the station and for the North Cut and Cover Tunnels. Once the southern extremity of the UPR station was excavated, a Lovat 6.450 m (21-2) diameter EPBM TBM was installed and began excavating the Bayamon Right, then Bayamon Left tunnels to connect to the Rio Piedras Station. All tunnel mucking was performed from the UPR Station excavation. See Figures 18 and 19 that illustrate the scope of station excavation and the extensive Support-of-Excavation system utilized. Once the station box had been excavated, it measured 245 m long 20 m wide 20 m (average depth).

597 Copyright © 2004 Taylor & Francis Group plc, London, UK

University of Puerto Rico Station

09096-73.qxd

2/18/04

3:09 PM

Page 598

Figure 20. Train Room exterior walls and restruts in place.

Figure 19. Aerial view of UPR station excavation operations. Table 4. UPR Station – Shaft quantity summary. Figure 21. Mezzanine level and platform slab in UPR Station.

Description

South rooms

North rooms

General Layout Main shaft dimensions Entrance dimensions Nominal depth Overall footprint Shored wall area Excavated quantity CIP concrete quantity

Rectangular box 245 m L 20 m W 27 m L 27 m W

Curved box cut 280 m L 14 m W 49 m L 9 m W

20 m 4,900 m2 9,800 m2 98,000 m3 30,300 m3

15 m 3,900 m2 8,400 m2 78,000 m3 10,800 m3

Building settlement control was a major concern considering the proximity and the condition of several older structures located within the grounds of the University of Puerto Rico, immediately adjacent to the excavation. The UPR Station was located entirely below grade and could be accessed for construction from the surface throughout its entire length. This large open cut excavation also allowed for servicing the construction of the EPBM tunnels leading from the south end. These service rooms were formidable structures to complete. They were designed to enclose the below grade portions

of station entrances, essential service equipment needed for station operations as well as for systemwide power distribution. The project CPM Schedule for the UPR Station started and ended with the excavation and concreting of the service rooms. Table 4 lists the principle quantities for the UPR Station and the connecting North Cut and Cover tunnels. 5.3.2 Concrete work A heavy invert slab (2.5 m thick) was placed in the station and service room areas followed by exterior walls, platform walls and slabs. Substantial interior wall and slab work was also needed in the adjacent Service Rooms. These service rooms required 12,000 m3 concrete and in excess of 30 months to complete (to grade) before the follow-on mechanical electrical and architectural systems could be installed. Largely due to the schedule delays from the EPBM tunnels, the North Service rooms and Train Room areas were severed from the ongoing tunneling operations. See Figures 20 and 21.

598 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-73.qxd

2/18/04

3:09 PM

Page 599

Figure 24. UPR Station Train Room finishing in progress.

Figure 22. Station chiller plant and piping system.

Figure 23. Station emergency exhaust fan. Figure 25.

5.3.3 Station systems The UPR Station was designed and built with considerable mechanical and electrical systems in support of the station operations, future traction power substation and fire-life safety requirements. A Change Order directed additional work for the installation of mechanical/electrical systems for Platform Edge Doors (PEDs) that included an extensive chilled water system and HVAC ducting throughout the station. See Figures 22 and 23. 5.3.4 Station finishing The station was designed with virtually all public areas covered with designated finishes that included modular porcelain enamel-coated wall panels, perforated aluminum ceiling panels, ceramic floor and wall tiles, granite stair treads, stainless steel handrails and trim. This station was air-conditioned. HVAC was a major portion of the re-designed station cooling system and required clever logistics to install as the civil work was underway. Similarly, the escalators had to be installed concurrently with the civil works – to provide access for modular units into pre-engineered wells. See Figures 24 and 36.

5.3.5 Systemwide interfaces TUO’s Systemwide Contractor (and subcontractors) undertook many scopes of work within and around the station following the completion of the civil works. The most essential was the installation of the 13.2 kV primary power substation with a corresponding 480-volt secondary distribution to KKZ installed equipment. The most intrusive was the installation of the train rail and third rail through the station. See Figure 25. 5.3.6 Station work schedule Work on the UPR Station began with the Support-ofExcavation system in early 1998. All structural work below grade was completed in mid-2001 with placement of the mezzanine slabs in the Train Room and the roofs of the South and North Service Room/ Entrances (to grade). This achievement fulfilled the Interim Milestone in the Contract Schedule. Installation of extensive mechanical/electrical systems and architectural finishing followed for many months until completed in spring 2003.

599 Copyright © 2004 Taylor & Francis Group plc, London, UK

480-volt electrical distribution and load center.

09096-73.qxd

2/18/04

3:09 PM

Page 600

6.2

6 GROUND TREATMENT OPERATIONS Ground treatment for many areas of station and tunnel excavation were substantial and highly engineered to meet the construction requirements that included building settlement control and for improving tunneling conditions. Ground treatment was utilized in the following areas – see Table 5. 6.1

Ground treatment subcontract

KKZ engaged Soletanche Bachy, a specialist grouting subcontractor to design and implement specific grouting programs for ground settlement control. The majority of the ground treatment programs were performed as fracture and compensation grouting and placed over the Rio Piedras Station Train Room throughout excavation of the stacked drift tunnels. See Figure 26. These ground treatment programs were performed concurrently with a continuous dewatering program in the same area. See Figure 27.

Compensation grouting program – RP Station

The compensation grouting program – primarily over the Rio Piedras Station – required the installation of numerous arrays of angled grout pipes designed to suit the drilling and grouting equipment and access restrictions and installed in such a manner to provide for variable and immediate grout injection for building settlement control as a reaction to ground behavior while tunneling was underway. A four-phase compensation grouting program was performed and included the following. See Table 6. 6.3

Compensation grouting program – tunnels

Another compensation grout program was performed over the NATM Tunnels and a portion of the EPBM

Table 5. Ground treatment summary. Description

Scope of work

Access by

Rio Piedras Station

Compensation grouting Fracture grouting Compensation grouting Compaction grouting Fracture grouting Compensation grouting

gallery gallery shaft/surf surface surface shaft

EPBM Tunnels Cross passage (1) NATM Tunnels

Figure 27. Grouting arrangement for Rio Piedras Station.

Figure 26. Section through the stacked drifts for RP Station.

600 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-73.qxd

2/18/04

3:09 PM

Page 601

Tunnels using shallow shafts for access in each case. In both cases, arrays of horizontal grout pipes were installed above the tunnel alignments to provide for variable and immediate grout injection for building settlement control as a reaction to ground behavior while tunneling was underway. The compensation grouting programs were specifically targeted to arrest potential settlement of designated buildings located over the tunnel alignments. These buildings included both historical and poorly supported multi-storey structures. See Figures 28 and 29. The NATM Tunnels were situated under an historical wood frame building that although resistant to some building settlement, was particularly important to the local community. Other adjacent buildings, however, had very little ability to withstand even small settlements and had questionable and completely unrecorded foundations.

range of specifications, guidelines and rationale for the selection of materials in virtually all areas of the project. It frequently allowed for the use of several different types of materials. The initial design commitments were critical – to meet the schedule – and to provide economical solutions while meeting the Design Criteria throughout the project. The design team within the KKZ Joint Venture was working to the same project schedule as the construction operations. 7.2

Community involvement

Architectural compatibility was essential and took considerable time to achieve. The local community as well as TUO personnel was deeply involved throughout this portion of the design development program. The key areas were the station headhouses. See Figures 30 and 31 that illustrate the early and the final renderings

7 DESIGN DEVELOPMENT & MANAGEMENT 7.1

Design criteria

The Design Criteria for the project was incorporated into the Contract Documents and included a wide Table 6. Compensation grouting – phases of work. Phase

Description

Phase 1 Pre-conditioning grouting Phase 2 Pre-lifting grouting Phase 3 Compensation grouting Phase 4 Post-grouting

Purpose and Goal to stiffen the soil to produce a slight heave real-time grouting of soil to arrest bldg settlement

Figure 29. Grout access shaft for the NATM tunnels.

Figure 28. NATM tunnels general arrangements.

601 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-73.qxd

2/18/04

3:09 PM

Page 602

Figure 30. Early design for Rio Piedras North Headhouse. Figure 32. Chilled water system for platform edge doors.

Figure 31. Final design concept for RP North Headhouse.

of the same structure for the Rio Piedras Station North Headhouse. In this case, the community and also local university architectural students were heavily involved with the design development that materially changed the size, shape and overall appearance of the headhouse and the corresponding substructure. KKZ’s project architect, Anil Verma, prepared alternate designs for review, modification and ultimately, approval. At the UPR Station there was a similar occurrence. 7.3

Design concurrency with STT

The Rio Piedras Project was designed concurrently with an extensive systemwide design development scope performed by TUO’s Systemwide Contractor, Siemens Transit Team (STT). The scope of the systemwide contract was very far reaching and touched virtually all areas of the Rio Piedras Project. The integrated systems required close coordination. See Figure 32 for a view of the chilled water plant. 7.4

Figure 33. Rio Piedras Station Train Room early rendering.

Design completion

The complete design for the project had to include many changes desired by TUO as well as those needed to suit the site conditions. As a result, equipment and

Figure 34. Rio Piedras Train Room near completion.

materials needed to finish portions of the work were delayed pending the completion and/or modification of the design. The majority of the work was constructed with the benefit of a complete, consolidated and comprehensive design with specifications. See Figures 33 and 34 above that show an early rendering and the near final condition of the Rio Piedras Station Train Room.

602 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-73.qxd

2/18/04

3:09 PM

Page 603

Figure 35. Final artist’s rendering of Rio Piedras Station Southwest Entrance Headhouse.

8 DESIGN-BUILD ISSUES AND IMPACTS 8.1

Figure 36. Rio Piedras Station at under mezzanine level.

Scope of the design task

The Rio Piedras design team was comprised of CMA, Sverdrup Civil with Anil Verma (project architect), Jacobs Associates and Woodward Clyde Consultants. All but CMA had their principle design offices located on the United States mainland and were, therefore, initially detached from the center of activities on site. KKZ recognized the essential need to establish and staff a design office at the site and for the first eight months following Notice-to-Proceed, senior engineers from the design team member companies assembled a core staff on site. 8.2

Architectural design integration

The design review process faced additional challenges when TUO invited both the University and the local community to comment on architectural aspects of the design. The approach was justified by TUO in an effort to yield a socially and politically responsible transit system. KKZ responded with several design alternatives and subsequent iterations on the selected alternatives. This process was repeated for all station headhouses. The final architectural product has been well received by the community and is consistent with the local architecture. It is distinctive yet well balanced. See Figure 36. 8.3

Schedule for the completion of design

Throughout the design process, the project experienced wholesale changes in architectural, structural, electrical, and mechanical scopes. These changes required the design management team to spend considerable effort revising schedules, pursuing adjustments to the Contract Price, and mitigating impacts as a result of the consequential effects of abandoned designs, redesigns, and additional design. Meanwhile, the design team had to perform their work within the parameters and work sequence defined by the project CPM Schedule. The design team was involved with the initial development

Figure 37. Rio Piedras Station HVAC systems in Train Room.

of critical elements of the project CPM Schedule since the early activities in the schedule – on the Critical Path – were virtually all design-related. Key dates were established for the production of deliverable products. 8.4

9 RECOMMENDATIONS FOR FUTURE WORK The issues mentioned in this paper are in some ways unique to the Rio Piedras Project. This was the first phase of the transit system in the city, with a Design/ Build contract model, and a separate Systems contract. The Rio Piedras Project was an important section of the Tren Urbano system in San Juan, PR. It is the only

603 Copyright © 2004 Taylor & Francis Group plc, London, UK

Systems design integration

The concept of contractor design management was even more critical because the systems design was performed concurrently with the station and guideway design. Many design changes were implemented only when the details of the train, power, and communication systems became available from the Systems’ designer. See Figures 37 and 38.

09096-73.qxd

2/18/04

3:09 PM

Page 604

underground section and as such, required a large capital expenditure. 9.1

Owner and consultants

TUO retained an experienced consultant (GMAEC) for review of KKZ’s technical submittals and the emerging design in relation to the Design Criteria. The project was bid based on 30% complete design drawings. Inevitably, there were differences of opinion as to the scope and the intent of the 30% drawings that were also the basis of the Contract. 9.2

Contractor’s designer

ACKNOWLEDGEMENTS

The success of the Design-Build team owes a lot to the experience and level of cooperation between the contractor and its designer. A strong and functional relationship was essential for the timely and troublefree performance of the work, particularly for work that was complex, full of risks and in this case, for an Owner involved in the work for the first time. KKZ’s designer fully understood and embraced the risks of the work and was able to fulfill and react promptly to the variances encountered over the course of the project. Exceptional communication skills were needed. To this end, the designer and contractor personnel were able to identify and address the needs and meet the challenges. The designer and KKZ’s staffs had a clear understanding of the common goals for the project. Moreover, they had the same sense of urgency and cost sensitivity necessary to produce the required work products on time that were so essential for economical construction. The designer had to understand and work to the same project CPM Schedule as KKZ and to provide information that was consistent with the needs of the construction activities. 9.3

is substantially complete at this time. It has been a challenging project and one that has fully utilized the combined talents of the Joint Venture Design-Build team. Many “lessons learned” can be drawn from the project. KKZ has completed the work and delivered to Tren Urbano and the people of Rio Piedras and Puerto Rico, a well-designed and high quality finished product that meets the Design Criteria and the contract requirements. It is visually appealing and nestles neatly and discretely into the existing neighborhood along historical Ponce de Leon Avenue in the Rio Piedras suburban area.

Systemwide changes and concurrent design

The scope of the Systemwide elements and Interface Control was substantial and integrated into the scope of the civil, structural, mechanical, electrical and architectural work under the Rio Piedras contract. The two contracts were performed concurrently and with little staggered lead-time. Owing to the unrelated delays experienced on the Rio Piedras contract, the concurrent Systemwide design and construction was able to stay abreast of progress.

Tren Urbano has been a strong supporter of the project since commencement in 1997 and has made important contributions to this paper. The KKZ Joint Venture Design-Build team that included an integrated staff from Kiewit (sponsor), Kenny, Zachry, CMA, Sverdrup, Jacobs Associates, Woodward Clyde Consultants and Anil Verma committed to the completion of this complex project six years ago. It is now a reality and underlines the scope of this commitment and the combined talents, energy and creativity needed to meet the enormous engineering and construction challenges encountered.

REFERENCES Fulcher, Brian, J.E. Carlson M. Bartlett 2003, Rio Piedras Project Completion, Proceedings of the 2003 Rapid Excavation and Tunnelling Conference, New Orleans, LA, sponsored by SME, pages 746 to 769. Gay, Michael, G. Rippentrop, W.H. Hansmire, V.S. Romero 1999, Tunnelling on the Tren Urbano Project, San Juan, Puerto Rico, Proceedings of the 1999 Rapid Excavation and Tunnelling Conference, Orlando, FL, sponsored by SME, pages 621 to 644. Morrison, James A., P.H. Madsen, S. Carayol 1999, Ground Control Program for the Rio Piedras Project, Tren Urbano Program, San Juan, Puerto Rico, Proceedings of the 1999 Rapid Excavation and Tunnelling Conference, Orlando, FL, sponsored by SME, pages 415 to 435. Romero, Victor S., P.H. Madsen 2001, Design and Construction Performance of a Large Diameter Tunnel Constructed in Soft Ground by the Stacked Drift Method, Proceedings of the 2001 Rapid Excavation and Tunnelling Conference, San Diego, CA, sponsored by SME, pages 199 to 219.

10 SUMMARY AND CONCLUSIONS Phase I of the Tren Urbano system intends to start revenue service in late 2003. The Rio Piedras Project

604 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-74.qxd

2/18/04

8:03 PM

Page 605

North American Tunneling 2004, Ozdemir (ed) © 2004 Taylor & Francis Group, London, ISBN 90 5809 669 6

Devil’s Slide Tunnels Y. Nien Wang, Bill Hughes, Hugh Caspe HNTB Corporation

Moe Amini Caltrans

ABSTRACT: This paper describes the major design parameters and processes employed for this state of the art tunnel in a high seismic zone. It includes as a case study the design of a twin transportation tunnels in the County of San Mateo, California. With an estimated overall project construction cost of $260,000,000, the facility is currently under final design stage.

1 INTRODUCTION In 1983, ground movement at Devil’s Slide caused Highway 1 to be closed for an extended period of time which caused significant economic loss to the surrounding coastal communities. A number of options were considered for bypassing the slide area, and the Martini Creek inland bypass was selected for design. Subsequent litigation resulted in a Devil’s Slide Tunnel Study being carried out in 1996. The results of the study, the Feasibility Report, indicated that a tunnel alternative was competitive with the inland bypass. The final decision was left to the citizens of San Mateo County, who approved the tunnel alternative in the November 1996 election. After further environmental studies and legal challenges were processed, a final environmental document was prepared in 2001, and the preferred alternative was selected to proceed with final design and construction. 2 PROJECT DESCRIPTION The Devil’s Slide Tunnels consists of a separated twolane road, one lane in each direction, which passes through twin tunnels and over twin bridges and then connects with the existing non-separated, two-lane road at each end. The length of the entire project is approximately 1900 meters, made up of four major project sections which are described in the following paragraph, moving from south to north. (1) Operations and Maintenance Center (OMC) area, the 250-meter-long south approach roadways,

including the South Rock Cut, extend to the tunnels’ South Portals; (2) Twin tunnels are 1250 meters long and extend to the North Portals near the south abutments of the twin bridges; (3) Twin bridges are 275 and 300 meters long respectively and span the Sham-rock Ranch valley; (4) North approach roadways then rejoin the existing highway. The horseshoe-shaped tunnels are 9 meters wide and 6.8 meters high and approximately 18 meters apart. There are 10 cross passages for emergency egress, and 16 jet fans in each tunnel for ventilation. The tunnels have a lighting, fire protection, and operation and control systems. An Operation and Maintenance Center will be located south of the tunnels at the excavation Disposal Area, normal operation of the tunnel will be controlled remotely from the Caltrans Oakland Traffic Management Center (TMC). The arched three-span twin bridges’ decks are 7.8 meters wide and are about 38 meters above the valley floor. The bridges will use the cast-in-place segmental balanced cantilever construction method to avoid environmentally-sensitive areas. The cross section of the tunnels is shown in Figure 1 below. Each tunnel has a vertical clearance of 4750 mm and provides a single 3.6 meter wide traveled way, two shoulder areas (2.4 meter and 0.6 meter wide), and two 1.2 meter wide sidewalks, for a total width of 9.0 meter. The ventilation jet fans are placed in the crown of the tunnel and cross passage is provided about every 120 meters to provide routes of escape for fire incidents. The oval shape of the tunnels is the best shape for accommodating the resulting stress field in the rock due to excavation and liner stresses.

605 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-74.qxd

2/18/04

8:03 PM

Page 606

Figure 1. Tunnel cross sections.

Figure 2. Fault map at tunnel location.

3 GEOTECHNICAL CONSIDERATIONS The project site is located along the western edge of the Coast Ranges geological province of California, which extends from Oregon southward to the Transverse Ranges in the Santa Barbara area. This province is characterized by northwest-southeast-oriented geomorphic and structural trends (mountains, valleys, faults), the existence of two geologic terrene existing side-by-side, and major northwest-southeast trending faults. The major active regional faults in the site region are the San Andreas Fault to the east and the San Gregorio fault-zone located offshore just 3 Km to the west.

These intervening inactive faults as shown in Figure 2 below are identified as faults B through D. A fifth fault, A, is an internal feature within the granitic rock. Fault B, which forms the boundary between the crystalline basement and the overlying section of clastic sedimentary rocks, is referred to by Pampeyan (1994) as the San Pedro Mountain Fault. Faults A and B dip moderately (35 to 50 degrees) to the north and are probably cut off at depth by the steeply north dipping fault C. Fault B is the only mappable fault that was intersected by core borings. Faults A, C, D and E are all clearly evident in natural and road cut exposures along the seaward face of the San Pedro Mountain. Faults B, C, D and E are all mappable inland at least

606 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-74.qxd

2/18/04

8:03 PM

Page 607

water elevations ranged from 60 to over 150 meters above mean sea level and, except for the portal areas, the ground water elevation is above the tunnel, and rises to approximately 110 meters above tunnel grade in the central part of the alignment. In general, ground water flows northward, north of the San Pedro Mountain divide, and southward, on the south side of the divide. It is important to note that the estimated ground water surface shows a depression in the vicinity of Fault B (Pampeyan’s (1994) San Pedro Mountain Fault), suggesting that the fault may act as a ground water drain. There are borings completed in the 1996 investigations, but the recent site exploration program includes 23 soil/rock borings for the tunnel, portals and South Rock Cut, 10 soil borings for the disposal area, and several test pits (trenches). A number of in-situ tests are also performed recently, which include a Pressuremeter Test, an Optical Televiewer, a Packer Test, and Downhole P&S Logging and Inclinometer tests. 4 TUNNEL Figure 3. Fault map for project site.

as far as the tunnel alignment by contrast in structure and stratigraphy across their traces. Devil’s Slide is located within a seismically-active area, which generally consists of a group of looselyclustered, small-magnitude events (M  3) in proximity to San Pedro Mountain (events are generally in the 5 to 10 km depth range). This area also contains the San Andreas Fault which is about 8 km to the east; the active Hayward Fault which is 36 km to the northeast, the potentially-active Seal Cove Fault which is 2.4 km offshore to the west and the active San Gregorio Fault which located offshore is about 3 km to the west. The largest historical earthquake in this region was the magnitude 7.8 earthquake which ruptured the San Andreas Fault to the east in 1906. Figure 3 is a fault map of the Northern California region with the Devil’s Slide Tunnels site shown on the map. The San Andreas Fault to the east and the San Gregorio fault to the west are the dominant seismic sources for the site. Due to their high activity rates and close proximities, the two faults contribute most of the seismic hazard for the project. All other faults shown on the map have also been considered in the hazard analysis, even though they contribute very little to the site’s overall seismic hazard. Random area sources have also been included in the analysis, which take into account smaller, unknown faults that may exist at the project site. Ground water elevations were taken at boreholes along the tunnel profile which showed that ground

The primary support of the Devil’s Slide Tunnel’s rock mass consists of steel fiber reinforced shotcrete, lattice girders, rock bolts, and spiles. These support elements are adaptable to the actual rock conditions based on the geotechnical monitoring program. Therefore, the shotcrete thickness, as well as the length and the number of rock bolts, will vary along the length of the tunnels. Shotcrete provides the initial structural support and seals the rock surface against weathering thereby achieving an interaction between the rock mass and the lining. The shotcrete is reinforced with steel fibers, which reduces shrinkage and increases the ductility of the material thereby increasing the resistance against fallouts of rock blocks. Lattice girders are used as templates and additional reinforcement, and are fully embedded in shotcrete. In addition to shotcrete, rock bolts are used to increase the structural stability of the initial lining. The exact type of rock bolts will be defined in final design. The final lining is a cast-in-place concrete lining, which is reinforced to withstand the effects of earthquakes. For construction reasons, a minimum thickness of 350 mm is anticipated in order to obtain an adequate concrete cover when two layers of reinforcement are installed. A fire load is also included in the design of the final lining. The planned drainage systems for the Devil’s Slide Tunnel include separating the formation (groundwater) and road drainage so that water can be drained off safely and continuously. This separation is necessary because the road surface runoff is contaminated and requires treatment, while the formation drainage can be directly discharged into the local drainage channel.

607 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-74.qxd

2/18/04

8:03 PM

Page 608

Two types of cross-passages are included in the design of this tunnel. They have been provided for pedestrians for escape/emergency purposes and therefore a clearance envelope of 2.4 m 2.4 m is sufficient. In addition, the cross-passage at the midpoint of the tunnels has been enlarged to allow access for small emergency vehicles. The 120 meter cross-passage spacing was used in these Tunnels. It is based on several considerations and studies which include the Computational Fluid Dynamics (CPD) analysis, the considered one-meter-per-second average walking speed for a pedestrian under stress and allowed twominute escape time for life safety in tunnels. The spacing satisfies the safety requirements of NFPA 101 Life Safety Code, which requires a maximum 122 meter distance to a safe area, and the results of the CFD modeling. It would allow pedestrians to reach a cross passage before they would be consumed by smoke of fire. The New Austrian Tunneling Method (NATM) also known as the Sequential Excavation Method (SEM) is considered to be the most viable method for excavating the Devil’s Slide Tunnels. The concept of the NATM is to activate the load-bearing capacity of the surrounding rock mass and then integrate it into an overall ring-like support structure. Additional support elements are then used which can be adapted to specific site conditions. Where poor conditions occur, deformable support elements will be introduced to allow stabilization of the opening by applying the minimum required support resistance, as derived from the ground reaction curve. Based on the site investigations, either shotcrete, rock bolting and steel arches, or combinations of these support elements and their sequencing; the required support is adapted to the various rock types. Various support categories are predefined in accordance to the expected rock mass behavior during excavation. Geotechnical monitoring and interpretations of the readings are continuously performed to check the effectiveness of the chosen support and to allow for any adjustments to support measures. When stress distributions have ceased, the final cast in-place concrete lining is installed. Therefore, the final lining will be initially loaded by its self-weight. NATM allows for excavation using the heading and bench method, which is supplemented by an invert excavation in areas of poor ground. Avoiding the multiple top heading and multiple bench sequences will result in a higher productivity rate of excavation and eliminates temporary supports at faces which would have had to be subsequently excavated. For the Devil’s Slide Tunnels, the approach is to drive the heading independently of the subsequent bench/ invert excavation. Driving the heading independent from the bench/invert excavation allows the contractor to optimize means and methods for the excavation and the installation of the initial support. It is possible

to drive the heading over the full tunnel length before starting the bench/invert excavation, except in areas where a temporary invert is needed in order to achieve ring closure for the support of the ground. For this project, it is anticipated that the bench/invert excavations will closely follow the top heading. The Devil’s Slide Tunnel design begins with the establishment of geological data in sections along the tunnel with consistent characteristics, and follows with a summary of geological series that demonstrate similar mechanical behavior. Further, the boundary conditions such as virgin stresses, size, shape, and orientation of the opening all are taken into account in order to establish possible failure mechanisms and the behavior of the opening. Because different failure mechanisms require different support measures, as well as modes of analysis to design the support measures, support categories are established that are applicable for the various types of opening behaviors. During the final design, criteria for the application of support categories during construction will be established. Tunnel support measures consist of elements such as shotcrete, rock bolts, steel arches, and spiles, which act as temporary support. Eight possible support categories (as shown in Table 1 below), based on the behavior of the opening, have been developed in the preliminary phase. The grading of categories is based on a comparison of the support resistances of each support element. For the various support categories the length of each category is limited depending on the behavior type and sequencing of support measures. In the final design, the number of support categories will be reduced to save construction cost. Structural design of the initial support depending on the type of support measure, different structural Table 1. Support categories in relation to behavior types. Type Failure mode

Support measures

1

Sealing shotcrete Local rock bolts Shotcrete Systematic rock bolting Steel arches Shotcrete lining Systematic rock bolting (grouted) Steel arches Spiles Shotcrete lining Systematic rock bolting (grouted) Steel arches Spiles Face bolts

2

3

Fracturing induced by stresses

4

Progressive failure induced by stresses

608 Copyright © 2004 Taylor & Francis Group plc, London, UK

Stable opening surface slabbing Fracturing along discontinuities

Support category I II II III IV V VI

VII VIII

09096-74.qxd

2/18/04

8:03 PM

Page 609

models will be used. Typical models and the corresponding structural elements are listed below:

• • •

Wedge theory Block sliding theory Finite Element model

for rock bolts for face bolts for shotcrete design/ overall behavior

For final lining design, a two-dimensional racking and ovaling study of the tunnel final liner, as shown in Figure 4, was performed to gain an understanding of the Tunnel’s performance when subjected to static and seismic loading. The liner section studied, for a drained tunnel and weak ground condition, includes a continuous tunnel liner with an invert arch. For a drained tunnel and competent rock conditions, the invert arch will be eliminated. However, elimination of the invert arch would likely not adversely affect the seismic performance of the liner, due to its increased flexibility. A computer program (ADINA) was used to investigate the seismic performance of the tunnel. Due to the close proximity of the San Gregorio Fault, the Ground Motion Report determined this fault would produce greater ground deformations at the Tunnel site than the larger, but more distant, San Andreas Fault. The primary task of the analysis was to study deformations imposed on the liner as a consequence of seismic wave propagation which assumes a fault normal attack on the liner and provides upper bound analysis results through the rock medium surrounding the liner. To accomplish this task, pseudo-static time history analysis was performed. This analysis consisted

of stepping the structure statically through the fulltime history earthquake ground displacement record to capture the maximum deformations imposed on the liner from seismic wave propagation through the rock. The ground input motion records used for this analysis consisted of two orthogonal displacement time histories in the vertical and transverse directions relative to the Tunnel section.

5 SOUTH PORTAL The South Portal is located in a steep (60 to 70 degree) slope with sparse vegetation. On the east side of the portal is a creek with very low normal flows, but intermittent high flows during rainstorms. There is an existing drainage basin in the approach to the portal that presently has an inlet structure that conducts surface drainage flows beneath Highway 1 in a corrugated metal culvert. This area will be backfilled and provided with appropriate drainage catchments and culverts to construct the new highway approaches for the north and south bound lanes. After rock mechanics analyses were conducted to determine the removability of finite blocks (i.e. wedges) isolated from the rock mass by a continuous series of faces, the seismic stability of each of the key blocks was evaluated with pseudo-static and time history analyses. For the time history analyses, Newmark’s sliding procedure, modified for movement on multiple planes (multiple joint surfaces), was adopted. Due to the critical nature of the portals, the rock motions were scaled up for lifeline level seismic criteria.

6 NORTH PORTAL

Figure 4. Typical tunnel cross section (undrained condition).

The North Portal will be located on the west flank of a north-facing ridge, with the roadway/tunnel alignment approaching the hillside at a skewed angle. The Feasibility Report indicated the presence of an existing landslide in the area of the North Portal, but subsequent examination of aerial photographs has redefined the surface expression of the landslide into two separate slide masses, which encompass roughly the same area as the single-slide that was mapped initially. The North Portals and the south abutments of the bridges are located within the landslide areas. The twin bridge structures will be located immediately north of the North Portal. The area between the bridge abutments and the tunnel portal is very limited and is required for circulation of vehicles between the northbound and southbound lanes and access to the Equipment Chamber between two tunnels. The geologic setting at the North Portal is more complex than the South Portal because the North Portal

609 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-74.qxd

2/18/04

8:03 PM

Page 610

structures will be built in sedimentary rocks such as sandy siltstone, sandstone, and shale that have been folded, faulted, and sheared extensively. One inclined boring, P-7-96, which was completed in the vicinity of the North Portal, indicates that there is approximately 8 meters of very poor rock, underlain by poor-to-fair rock, which is based on information from the Rock Quality Designation (RQD). In addition, several shear zones were encountered deeper in the boring. Generally speaking, the bedding planes at the North Portal dip into the hillside in a “favorable” direction relative to the anticipated cut slopes. It should be noted that no measurements were taken for the joint sets that were oriented perpendicular to bedding. These joints indicate that they contain occasional clay infilling, and appear to be polished in other places. Seismic performance of the slope at the North Portal was evaluated using the rock motions developed for the Devil’s Slide Tunnels. To be consistent with the “lifeline” design criteria adopted for the portals, the non-life line rock motions have been scaled-up by 35percent and were used for the deformation analyses. 7 OMC BUILDING The Operations and Maintenance Center (OMC), located 330 meters south of the South Portal, will consist of a one-story 1367-square-meter building. For security, the OMC will have a locking gate/fence at the entrance to the site, High-Pressure Sodium (HPS) lights, Closed Circuit Television (CCTV) surveillance cameras, and intrusion alarms on the doors and windows. The OMC’s heating system will be electric, since there is no natural gas line readily available. Water for the building will be provided from the North Coast County Water District, but the type of sewage disposal system needs to be further evaluated in final design once the amount of excavated material from the portals and South Rock Cut are determined, as these volumes affect the area available for on-site disposal. The OMC building will be placed several hundred feet east of Route 1 with bermed areas to the east, south, and west of the building and suitable landscape materials which will inhibit it being viewed from the highway, as well as providing screening for the westerly setting sun. A high earth berm is proposed for the west and south sides to screen the building from the highway, and lower berm is being considered for the east to screen access paving on the east side of the building. There will be direct visual access from the hiking trail to the east of the site, until the landscape matures after 10–15 years, but consideration will be given to plant selection which will also help aid in restricting access to the OMC site.

8 VENTILATION SYSTEM The jet fan type longitudinal system is used for this configuration of tunnels because they have been successfully tested and proven to work, based on the results of the final Memorial Tunnel Fire Ventilation Test Program. Similarly jet fans are installed in the 1250meter-long double bore Cumberland Gap Tunnels in Tennessee, and are also used in tunnels of comparative length all over the world. The use of jet fans for the Devil’s Slide Tunnels is acceptable to both Caltrans and FHWA and their effectiveness has also been validated by a Computational Fluid Dynamics (CFD) analysis. The design heat-release-rate produced by a vehicle fire is used to design the ventilation system used during fire fighting operations. This ventilation is based on 20 MW fire, which is the equivalent of a single truck or bus being consumed by fire and 50 MW fire, which is the equivalent of a gasoline or propane trucks being consumed by fire. However, currently it is Caltrans policy not to permit any vehicles with high fuel loading enter into the tunnels. This restriction is acceptable to Caltrans and local agencies for safety purposes and also because alternate routes are available for commercial and tanker truck vehicles. Ventilation of the tunnel for the control of heat and smoke during fire-fighting conditions varies greatly due to such factors as fire size, tunnel grade, tunnel crosssection, and direction of airflow. The velocity of air for smoke control is predicted using the methodology developed from studies conducted by the U.S. Bureau of Mines to determine the “critical velocity” at which the buoyant effect of the hot gases is overcome by longitudinal airflow. A refined analysis using “Solvent”, which is a commercially available CFD computer program for analyzing tunnel fires, confirmed that with the assumption that the jet fan above the fire is not operable, that the critical velocity was exceeded by the thrust from the operating jet fans which will be able to control the movement of smoke and heat in the tunnel. The CFD computation and model also showed that the proposed cross-passage spacing was acceptable for allowing the public to safely evacuate the tunnel during build-up of the design fire emergency. Jet fans are located at the crown of the tunnel above the roadway. For tunnel sections with pairs of fans, one fan is located above the roadway and the other above the roadway shoulder to minimize the disruption to traffic when fan repair or replacement is required. No portal buildings are required for the Tunnel’s jet fan systems with the exception of the chambers that will house the electrical equipment and controls. For jet fan ventilation, the tunnel opening is set flush with the hillside with only the tunnel portal structure extending beyond the hillside for rock-fall protection.

610 Copyright © 2004 Taylor & Francis Group plc, London, UK

09096-74.qxd

2/18/04

8:03 PM

Page 611

9 TUNNEL LIGHTING The luminance of the interior surfaces of the tunnel portals must be higher to compensate for the high background luminance of sky and pavement. The design light levels will also take into account the anticipated medium traffic volume and the worst conditions for northbound and southbound traffic will be between 1:00 and 2:00 p.m. Luminaires in the tunnels will be mounted directly on the ceiling and their wiring will be routed through conduits and terminated at junction boxes. These junction boxes will be made of cast iron for embedded applications and stainless steel for surface-mounted applications and the conduits will be galvanized rigid steel. No PVC, fiberglass, or other plastic materials are allowed in the tunnels due to poisonous by-products that are released during a fire. Emergency lighting conduits will be embedded in the concrete lining, with non-emergency lighting conduits surface-mounted. Since the tunnels are long and straight, their exit portals will be visible from the approach to the entrance portals. Or more precisely, the luminance of surfaces beyond the exit portal will be visible from the approaches. The exit portals will be a small part of the field-of-view because of the length of the tunnels, but will provide directional guidance, traffic, and daylight if conditions permit. 10 TUNNEL OPERATION SYSTEMS The controls design for the tunnels will provide manual operational capability to local emergency services personnel, to local tunnel operator-directed monitoring and control, and to Caltrans’ Oakland TMC remotely directed monitoring and operational control capabilities. State-of-the-art, microprocessor-based tunnel controls will be provided to allow for the safe and

efficient operation of ventilation, lighting, incident/ intrusion detection equipment, and other systems within the tunnels. Surveillance and monitoring systems in the tunnels provide a safe environment for the traveling public, assist operations staff in detection of incidents, and assist emergency response staff when responding to an incident. Systems elements are used to notify drivers of potential problems ahead, to detect and warn overheight vehicles, to monitor emissions in the tunnel, to visually confirm problems, to detect fires in the tunnel, and to provide emergency communications for drivers as well as emergency personnel. The Tunnels Control System (TCS) will consist of redundant pairs of Programmable Logic Controller (PLC) chassis at each of three locations: one pair in the south portal equipment chamber, one pair in the north portal equipment chamber, and one pair in the operations and maintenance center. Field monitoring instruments will be interfaced to the TCS at remote input/output (I/O) panels located within each crosspassage.

11 ENVIRONMENTAL CONSIDERATION There are wetlands and red-legged frogs in the Devil’s Slide Tunnels North approach area. Building the bridge using a segmental box girder erection method will minimize the intrusion into the Shamrock Ranch valley. New bridges will be used for access to the North Portal construction and a new waterline will be routed from Pacifica up Highway 1. The drainage channel on the west side of the North Portal and on the east side above the south portal is in the environmental sensitive area. However, using portal designs that avoid the channel will be accommodated into the final design.

611 Copyright © 2004 Taylor & Francis Group plc, London, UK