4DCAD and Visualization in Construction

4D CAD and Visualization in Construction

4D CAD and Visualization in Construction: Developments and Applications Raja R.A. Issa Ian Flood William J. O’Brien University of Florida, Gainesville, USA

A.A. BALKEMA PUBLISHERS

LISSE /ABINGDON /EXTON (PA)/TOKYO

This edition published in the Taylor & Francis e-Library, 2005. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.”

Library of Congress Cataloging-in-Publication Data (applied for)

Cover design: Studio Jan de Boer, Amsterdam, the Netherlands

© 2003 Swets & Zeitlinger B.V., Lisse All rights reserved. No part of this publication 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 the prior written permission of the publishers. Although all care is taken to ensure the integrity and quality of this publication and the information herein, no responsibility is assumed by the publisher 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 Swets & Zeitlinger Publishers www.balkema.nl and www.szp.swets.nl

ISBN 90 5809 354 9 (Print Edition)

ISBN 0-203-97112-4 Master e-book ISBN

Table of contents

Foreword

VII

Benefits of 3D and 4D Models for Facility Managers and AEC Service Providers Martin Fischer, John Haymaker, Kathleen Liston

1

Beyond Sphereland: 4D CAD in Construction Communications Dennis Fukai

33

Fully Integrated and Automated Project Process (FIAPP) for the Project Manager and Executive F.H. (Bud) Griffis, Carrie S. Sturts

55

New Construction Management Practice Based on the Virtual Reality Technology Jarkko Leinonen, Kalle Kähkönen, Tero Hemiö, Arkady Retik, Andrew Layden

75

4D CAD and Dynamic Resource Planning for Subcontractors: Case Study and Issues William J. O’Brien

101

The Role of 4D Modeling in Trade Sequencing and Production Planning David Riley

125

The Link Between Design and Process: Dynamic Process Simulation Models of Construction Activities E. Sarah Slaughter

145

V

VI

Table of contents

Acknowledging Variability and Uncertainty in Product and Process Development Iris D. Tommelein

165

Application of 4D CAD in the Construction Workplace Richard J. Coble, Robert L. Blatter, Indrid Agaj

195

Virtually Real Construction Components and Processes for Design-For-Safety-Process (DFSP) Steve Rowlinson, Bonaventura H.W. Hadikusumo

211

The Potential of 4D CAD as a Tool for Construction Management Robert M. Webb, Theo C. Haupt

227

Virtual Reality: A Solution to Seamless Technology Integration in the AEC Industry? Raja R.A. Issa

243

Construction Management Pull for nD CAD Peter Barrett

261

Closure R.R.A. Issa, I. Flood, W.J. O’Brien

281

Index

285

Foreword

The final frontier of the application of information technology in construction is the job site. And perhaps there is no stronger technological link between the job site and the design office than the practice of 4D CAD, and for good reason. Implementing a 4D CAD system cuts to the very heart of issues that mankind has struggled with for centuries: linking space and time; turning concept into reality; ownership of knowledge; effective communication between business partners. Models have been used for centuries to explain what is to be built, and the promise of this new technology is exciting. Never before in history has man, without ever turning a single shovel of earth or driving a single nail, been able to virtually construct a bridge or building one piece at a time and link each step to a corresponding step on a schedule. Nonetheless, entrenched ways of working do not change overnight. At the time of writing, the flurry of activity with dot.com investments and startups has slowed to a trickle. It comes as no surprise to many industry observers that many of the quick fixes with technology promised by these startups failed to take hold. What did take place, however, was a demonstration that the construction industry continues to be ripe and ready for new technologies that will provide value. A problem still plaguing the industry is that many in it are not in sync with the technology. A growing body of workers is well versed in the ways of the computer, but novice at best in the ways of construction. Juxtapose that with a still significantly sized body of industry veterans that have yet to embrace technology. Arguments can be made in support of delays to widespread adoption of 4D CAD are that the technology is not evolved enough for the people, or that the people are not evolved enough to use the technology—both viewpoints are valid. Those assertions are, at a deeper level, the continuing story of mankind’s relationship with his tools. Using contractual methods such as design-build to force parties to work together, rather than technological tools, has met with some success. On the horizon, it could also be a driving force to help the growth of 4D CAD. Superficially, a simple schedule and a CAD model may qualify as 4D CAD. Yet to make the practice truly widespread, deeper issues such as trade sequencing, production planning and dynamic cost and resource planning need to be addressed. VII

VIII Foreword

No industry needs 4D CAD more than construction. Like NASA’s Space Program, and the developments borne from it that ultimately benefited mankind as a whole, 4D CAD is similarly poised to have a major impact on any industry, including construction, that struggles to simultaneously manage the scheduled creation of objects, be they airplanes, airports or air cleaners. When designers have a better grasp of scheduling and buildability issues facing contractors, projects will be better. Similarly, when contractors are more involved earlier in the process, buildability issues will be dealt with on screen or on paper, instead of in situ. And without the disputes, disagreements and the rampant adversarial relationship that so often plagues construction projects today, designers and constructors can better concentrate on what they want most, which is ultimately to build. Matthew Phair Senior Editor Engineering News-Record

BENEFITS OF 3D AND 4D MODELS FOR FACILITY MANAGERS AND AEC SERVICE PROVIDERS Martin Fischer1, John Haymaker2, Kathleen Liston2 1

Civil and Environmental Engineering and (by courtesy) Computer Science, Stanford, CA, USA 2 Civil and Environmental Engineering, Stanford, CA, USA

Abstract The first part of this paper presents an extensive list of benefits users of 4D models have realized and illustrates the benefits with specific examples from actual uses on a variety of projects. It illustrates how current business practices and project delivery approaches allow or do not allow facility owners to reap these benefits. All owners and AEC service providers (designers, general contractors, subcontractors) who have used 4D models to assist in understanding, analyzing and communicating a design and construction schedule have reported benefits from the use of these models. Owners have used 4D models to plan the construction of facilities that require significant phasing prior to contract award to verify the overall constructibility of a proposed design given the project timeline and available space. General contractors have used 4D models for overall and for detailed construction planning, to communicate scope and schedule information effectively to subcontractors and other parties, and to test the constructibility of the design and the executability of the schedule prior to committing resources to the field. The second part of the paper describes in detail the application of 4D models for construction scheduling and constructibility analysis on the Walt Disney Concert Hall in Los Angeles. It discusses the reasons for the use of 4D models on the project and details the technical challenges the 4D modeler had to overcome. Specific examples of the impact of the 4D model on the schedule are also shown. Keywords: 4D modeling, construction planning, case studies, benefits

INTRODUCTION Traditional construction planning tools, such as bar charts and network diagrams, do not represent and communicate the spatial and temporal, or 4D, aspects of construction schedules effectively. Consequently, they do not allow project managers 1

2

M. Fischer et al.

to create schedule alternatives rapidly to find the best way to build a particular design. Extending the traditional planning tools, visual 4D models combine 3D CAD models with construction activities to display the progression of construction over time. 4D models combine 3D CAD models with the project timeline (Cleveland, 1989). Systems linking 3D CAD models with schedule and other project information started to be developed in the mid-eighties (Kahan & Madrid, 1987; Atkins, 1988). Experience on many different types of projects (simple to complex, new to retrofit) has shown that combining scope and schedule information in one visual model is a powerful communication and collaboration tool for technical and non-technical stakeholders (Williams, 1996; Retik, 1997; Edwards & Bing, 1999). The 4D research team at Stanford University has tested the usefulness of visual 4D models in planning the construction of a hospital, the roof of a university building, a small commercial building, a Frank Gehry designed museum, a theme park, and a Frank Gehry designed concert hall. These cases have shown that more project stakeholders can understand a construction schedule more quickly and completely with 4D visualizations than with the traditional construction management tools. Since they understand the scope and schedule of a project better, the stakeholders can then provide input to the scope and schedule and the important interrelationships, and help improve the project design and schedule. We and other 3D and 4D practitioners found that project managers using 4D models are more likely to allocate resources (e.g. design time, client review time, management attention, construction crews) more effectively than those who do not use 4D models. Danhier et al. (1994) came to similar conclusions in their application of 4D models to the replacement of steam generators. 3D CAD is often seen mainly as a design tool. It should also be seen as a construction tool, since a detailed 3D CAD model mirrors the completed project in the computer. It affords a project team the opportunity to practice or rehearse the construction of a unique artifact virtually before building it in reality. Project teams need to decide what problems they want to resolve through the use of 3D and 4D models. The resulting purpose of the 3D and 4D modeling effort has implications on who needs to be involved in the modeling effort. Should the models help answer questions to overall site logistics, flow of work, or access to various parts of the project at various times? Or should the models help answer questions about the specific sequence of work for a group of subcontractors, the laydown spaces needed for particular activities, or the distance in time and space between succeeding work? For example, effective use of 3D and 4D CAD as a detailed construction tool has implications on the project delivery process, the output or deliverables of various parties, and the processes and organization of projects. If the 3D model is to mirror the real project in detail, the same organizations that build the project should build the model because they will have the biggest stake in the accuracy of the information in the model. It is also unrealistic to expect that a group of designers and

Benefits of 3D and 4D models 3

modelers has all the expertise about construction details necessary for a detailed 3D model. The experience of the 4D research group at the Center for Integrated Facility Engineering (CIFE) at Stanford University shows that including at least key subcontractors as design-build firms from the beginning of a project makes detailed 3D modeling more efficient and effective than including them later (Staub et al., 1999). It is difficult for designers to know to what level of detail they should model a particular part of a project, since they often do not benefit directly from accurate detailed 3D models that clearly show what needs to be built. We have found that the subcontractors, however, are very interested in having accurate, reliable, and well-coordinated detailed design information because they can leverage that information in material procurement and management, and in planning and scheduling. Building a 3D CAD model in this way leaves accountability for the correctness of the information in the 3D model with the firms who are best equipped to leverage the investment in building 3D models. Designers remain in charge of the overall design concept, and subcontractors can focus on streamlining the production of their part of a project. A detailed and well-coordinated 3D CAD model allows firms to prefabricate directly from the model and improves material management. In this way, 3D CAD models enable project managers to allocate and use material resources more efficiently. 4D models extend the usefulness of design information to the construction planning and construction phases. If 4D models are built during the design phase they can help provide constructibility feedback to the design team, and they can also help set priorities for design work so that the necessary material procurement and crew planning information for construction are available in a timely manner. In this way, 4D CAD models help project managers to manage the flow of work and the allocation of crews and space on construction sites better (Vaugn, 1996). The next sections introduce benefits companies using 3D and 4D CAD models have realized in more detail. The later sections in this paper give specific examples of uses of 4D models and describe the parties that participated in the modeling efforts and discuss the level of detail they found useful. Table 1 lists benefits of 3D and 4D models realized by companies using such models for design and construction. Engineers and managers from owner, design, and construction firms reported them at a workshop on the use of 3D and 4D models hosted by Walt Disney Imagineering (WDI) and CIFE in May 1999. The column on the left lists specific benefits users have reported, the two middle columns show who realizes the benefit and who has most control and influence over the information in the 4D model necessary to realize the benefit. The column on the right shows whether the beneficiary matches the controlling party. Y means that it does, N means that it does not and S means “somewhat”, i.e. the primary beneficiary has significant control over the information in the 4D model necessary to realize the benefit, but other parties have some influence as well. Rows where the party that realizes the benefit does not match with the party controlling

4

M. Fischer et al.

Table 1. Benefits of 4D models for owners, designers, general contractors and subcontractors*. Benefit

Realize

Control

Influence

R ⫽ C?

Reduce design time Reduce design effort Speed up evaluation of design Reduce time needed to model an alternative Improve evaluation of design (functional sensitivity analysis) Share work around the world (model-centric project teams) Eliminate design production work (CD) Increase and improve information available for early decision-making Reduce project management costs Improve evaluation of schedule Reduce number of change orders Increase number of alternatives studied Increase number of project stakeholders who clearly understand the project and who are able to provide input Shorten (simplify, streamline) permitting time and effort Increase concurrency of design and construction Reduce interest costs Reduce time to make a decision Obtain management decision, funding Reduce life-cycle costs Maximize value to owner Increase productivity of crews Reduce wasted materials during construction Reduce rework Create complete information to build from Improve (verify, check) constructibility Verify consideration of site constraints in design and schedule (sight lines, access, …) Avoid (minimize, eliminate) interferences on site Maximize off-site work (prefabrication) Increase schedule reliability Verify executability of GC and sub-schedules Shorten construction period Speed up evaluation of schedule Increase site safety Minimize in-process time in supply chain Shorten site layout/surveying time Improve site layout accuracy Reduce RFIs Improve portability of design Shorten design and construction period Improve learning and feedback from project to project Improve effectiveness of communication Bring new team members up to speed quickly Coordinate owner, GC and sub-schedules

D D D D D

D D D D D

O O O

S S S Y Y

D

D

D DO

O DO

GC GCS O O O

GC GCS D D D

O

D

O O O O O O S S S S SGC SGC

DGCS GC O O O SGCD S S S SD D D

SGC SGC SGC SGC SGC SGC SGC SGC SGC SGC SGCD SGCD SGCD SGCDO

D D SGC SGC SGC SGC SGC SGC SGC SGC D D SGCD O

SGCDO SGCDO SGCO

O SGCDO SGCO

Y SGC SDO D SGC

N S S S N N N N

SD D D D GCDO GCDO GCDO

S N S S S N S S S N N N

O D

N N S S S S Y Y Y Y N N S N

SGCD

N

D D DO O

* Keys: O ⫽ owner, D ⫽ designer, GC ⫽ general contractor, S ⫽ subcontractor.

Y

Benefits of 3D and 4D models 5

or generating the information are shown in bold. The assignments to who controls the data and who realizes a benefit assume a traditional project delivery process. Table 1 shows that many benefits that potentially translate into significant time and cost savings are unlikely to be realized with a traditional project organization because the party benefiting from the use of 3D and 4D models is not in control of the information necessary to realize the benefit. Benefits for designers It is commonly understood that a design documented with a 3D CAD model will most likely have fewer errors and coordination issues because the construction of the model by multiple designers forces and allows them to reconcile inconsistencies. Evaluation of a design in 3D is also faster than with 2D drawings because reviewers can more quickly understand the scope and status of the design. Workshop participants who have been using 3D CAD models for several years reported that, after an initial learning curve, the overall design effort and design time is less than with a process using 2D drawings. They use 3D even when the client asks for 2D drawings because design revisions are faster and need to be done only once (instead of updating plans, sections, elevations and details). A further benefit is the potential to eliminate construction documents. Most participants saw little value in most 2D construction documents currently produced by design firms. On many projects subcontractors complete a new set of shop drawings anyway, and in some cases subcontractors fabricate parts directly from the 3D CAD model with numerically-controlled machines. 2D construction documents and shop drawings appear to be rather useless on a project where the design is documented and shared with a detailed 3D CAD model. Designers involved in projects that used 3D models from design through construction reported that they saw an increased coordination effort during the design phase of the project followed by fewer requests for information during construction. Hence, designers were able to focus on the phase of the project they enjoy most. Benefits for owners Owners are, of course, the ultimate beneficiaries of better performance by designers and builders from the use of 3D and 4D models. The workshop participants noted, however, that owners can use 3D and 4D models themselves to speed up and improve decision-making and to involve many more stakeholders than traditionally possible. For example, WDI was able to get the input from about 400 stakeholders during the two-month pre-bid design and construction schedule review for the Paradise Pier portion of Disney’s California Adventure. They were holding meetings with groups of eight to ten people at a time in their Computer-Assisted Virtual Environment (CAVE). The groups could interactively review the proposed design and construction schedule from any perspective and quickly understand the design, schedule, and corresponding constraints (Fischer et al., 2001).

6

M. Fischer et al.

Benefits for builders All participants at the workshop who have labor risk on site reported that detailed 3D and 4D models greatly increase the productivity of crews and help eliminate wasted materials and resources. Even if all the other project team members are working with 2D drawings many subcontractors still elect to build a 3D model for their scope of work and for the related scope of work. If all the information is readily available, they can build the 3D and 4D CAD models to verify that no interferences exist and that they have all the information and materials available for construction. If the information to build a detailed 3D CAD model is not available it is far cheaper for an engineer in the office to figure out what exactly needs to be built than for a crew in the field. The 3D models also support automated quantity takeoff for material procurement to ensure that each crew has the appropriate amounts and types of materials for a given day or week’s work. General Contractors (GCs) and subcontractors benefit from smooth, safe, and productive site operations, since that contributes to the shortest and most economical construction period. If built from subcontractor and GC schedules 4D models help the construction team coordinate the flow of work and space use on site. Contractors usually produce phasing drawings for a project. Typically, they are done manually, which makes it difficult to communicate them to all the interested and affected parties in a timely manner when they can still be improved economically. It also makes updating of the phasing plans a chore. Furthermore, they are only produced in 2D and for a few snapshots in time, which makes it more likely that a potential interference between trades gets overlooked. Combining 3D models with schedules automatically produces 3D phasing drawings at the daily, weekly, or monthly level depending on the level of detail in the schedule and the 3D CAD model. Contractors can easily see who is working where on what and how the work proceeds over time and through the site. 3D phasing drawings automatically reflect schedule updates. In summary, all workshop participants found that 4D models communicate schedules much more effectively than the abstract bar charts used on most projects, which, in turn enables the benefits listed in Table 1. Songer et al. (1998) came to similar conclusions in their study.

EXAMPLES OF APPLICATIONS OF 4D MODELS AND CORRESPONDING BENEFITS Members of the 4D research group at CIFE have supported construction project teams in applying 4D models on their projects since 1993 and have used the insights gained from applying the 4D models in real world settings to drive research efforts. Fischer & Aalami (1996), Akinci & Fischer (1998), and Fischer et al. (1998) give examples of the use of observations on construction projects to

Benefits of 3D and 4D models 7

formulate research questions, help formalize specific knowledge, and test research prototypes. This section briefly discusses these applications of 4D modeling and summarizes the corresponding benefits. In our experience, 4D models offer benefits on simple and complex projects, on new construction and on retrofit projects, and at the detailed nuts and bolt level as well as for overall project phasing. Unless noted otherwise, we used the Bentley Schedule Simulator or an earlier version of this program to combine 3D CAD and schedule information. Most projects were modeled in 3D with AutoCAD, and the schedule information was mostly in Primavera’s P3 tool, although MS Project has also been used. These brief descriptions are followed by an in-depth case study.

1993–95: RECONSTRUCTION OF THE SAN MATEO COUNTY HEALTH CENTER Together with the GC, Dillingham Construction Company, CIFE researchers build 3D and 4D models to coordinate the overall master plan so that the five-year construction period interfered as little as possible with the operation of the hospital. The 4D model coordinated owner relocation and operations schedules with construction schedules, facilitated client input, eased relationships with the community, and became, according to the hospital director, the best fund-raising tool (Collier & Fischer, 1996). It allowed, for example, verification that hospital staff could always reach all parts of the hospital from any other part without leaving the hospital. A more detailed 4D model was used to verify the constructibility of the central utilities plant and to make sure that the design and schedule information were complete and well coordinated. The 4D model for the US $100 m, 320,000 sf project was built from about 25,000 3D CAD elements and 500 activities in about 1,000 hours. The 4D model to study the overall phasing of the project consisted essentially of the main architectural components of the project: walls, windows, doors, columns, slabs, roofs. The 4D model for the central utilities plant was more detailed and included the foundations, some of the stud walls, the equipment platforms and equipment and the major mechanical ductwork. Collier & Fischer (1995) provide detailed information on the 4D modeling effort required for this project. The 4D modeling effort was carried out in parallel to the GCs construction planning efforts. Hence the 4D modeling effort served to confirm the project manager’s thinking about his approach to construction. It also helped the project manager communicate how hospital operations and construction were going to coexist in close proximity for the scheduled five-year duration of the project. The paragraphs below explain how 4D modeling helped improve the construction schedule for the San Mateo County Health Center. Figures 1 to 8 show eight stages in the reconstruction of the San Mateo County Health Center project over the planned five-year construction period from May 1994 to June 1999. The first

8

M. Fischer et al.

Figure 1. May 1994. Site with existing buildings, Main Hospital at left, Clinics Building at right foreground, East Wing in between, Existing Utilities plant in middle, Aids Clinics at top right.

Figure 2. March 1995. As work on the new Central Utilities Plant proceeds as the critical activity, construction on the first half of the North Addition begins. Trailers for temporary office space are installed next to the Clinics building. The first portion of the Central Hub is under construction (shown in green in the center of the snapshot). The hospital operations are linked through the existing East Wing (shown in gray in the center of the snapshot).

Figure 3. October 1995. Central Plant is completed, North Addition interior work is being completed. Aids Clinics is now in trailers. Aids Clinics is demolished and work has begun on new Nursing Wing.

Benefits of 3D and 4D models 9

Figure 4. January 1996. Nursing Wing exterior shell begins construction as critical path activity. East Wing functions are moved to North Addition. Connector is built to link first half of Central Hub to Clinics Building. The first portion of the Hub is finished and can now serve as a link between departments. This makes it possible to demolish the East Wing (shown in green in the center of the snapshot) to make room for the second half of the Central Hub.

Figure 5. July 1996. East Wing has been demolished and new Clinics Building and second half of North Addition are under construction in its place. Nursing wing work has moved to the interior.

Figure 6. June 1997. Old Clinics Building has been demolished and cleared away. Construction of the new Diagnostics and Treatment Building has begun.

10

M. Fischer et al.

Figure 7. December 1998. Remodeling of existing hospital structure begins. Entire third floor and ancillary wings are removed.

Figure 8.

June 1999. The new San Mateo County Health Center.

snapshot from the 4D model (May 1994) shows a view of the hospital prior to construction (Fig. 1), and the last snapshot (June 1999) shows the 3D CAD model of the reconstructed hospital (Fig. 8). As can be seen from comparing these two models, the transformation of the hospital during the reconstruction period is very dramatic. Constructing new parts and renovating other parts of the hospital without interrupting hospital operations was a challenging task. Originally, the GC showed the construction period with a bar chart schedule based on a critical path network. As can be imagined by just looking at the “before” and “after” models, such an abstract representation of the flow and sequence of construction fails to uncover potential time-space conflicts between construction and operations. The six snapshots of the planned progress of the construction (Figs 2 to 7) taken from the simulation of the construction schedule show the relationship between construction activities and hospital operations more clearly. It is easy to create as many snapshots as desired. In these snapshots, building activities are shown in green (non-critical activities) and red (critical activities). Parts of the facility that are not under construction, i.e. where construction has not yet started or

Benefits of 3D and 4D models 11

where construction has been completed, are shown in gray and yellow (the original colors in the 3D model). The following example illustrates the usefulness of 4D models to support master planning. We have also tested the usefulness of 4D models for more detailed planning and found them to be extremely helpful for the coordination of contractors and for constructibility improvements (see the other case studies in this paper). The 4D model, from which the snapshots are taken, took about four person-months to build. Instead of focusing on all the activities that are shown in the snapshots, we would like to draw the reader’s attention to an improvement to the interaction between construction and operations the 4D modeling effort helped make (the snapshots show the improved version). Initially, the Central Hub (the building with the round yard in the center of the hospital), connecting all parts of the hospital, was to be constructed as one building at one time. However, if that had happened (as was confirmed with the first 4D simulation), construction would have cut hospital operations in half. It would have been necessary to put patients on gurneys, wheel them out the door and around the block to bring them from their rooms to radiology services. This was, of course, not acceptable for the hospital staff, and the designers and construction managers had to find a solution to maintain uninterrupted links within the hospital between all hospital departments at all times. They decided to cut the Central Hub in half, add a seismic joint in the middle, and build one section of the Hub in the early phases of the project, as seen in the March 1995 snapshot (Fig. 2).

1995: ROOF FOR HAAS SCHOOL OF BUSINESS, UC BERKELEY The 4D modeling effort on the San Mateo Health Center project was done as part of the early construction planning phase and informed the project management team about potential problems and opportunities for improvement. In contrast, we completed a small study of the applicability of 4D models to day-by-day subcontractor coordination after the work had been done. The advantage was that we knew why and in what way the construction of the roof had not been as efficient as possible. Misunderstandings between the architect, the GC and the roofing, stucco, and sheet metal subcontractors led to extra cost of about US $200,000 due to low productivity and rework. Together with the roofing subcontractor we developed 4D CAD models of the various design solutions and several construction sequences using keyframes produced with 3D Studio in less than 40 hours. The 4D model included all the parts including the main assembly pieces that needed to be installed on the roof. The model clearly showed the challenges and tradeoffs of the various design and schedule proposals and would have been helpful for the contractors to understand each others’ constraints. Fröhlich et al., 1997 show snapshots from this 4D model.

12

M. Fischer et al.

1997–99: SEQUUS PHARMACEUTICALS PILOT PLANT IN MENLO PARK The 4D model for this biotech project coordinated the mechanical, electrical, and piping (MEP) contractors’ day-by-day work. As a result there were no field interferences, no rework, higher productivity, only one contractor-initiated change order, no cost growth during construction, and 60% fewer requests for information than expected for this type of project (Staub et al., 1999). The GC also used the 3D model for automated quantity takeoff. The Stanford 4D group built the 4D models on this project with input from the GC and from the MEP subcontractors. The model was very detailed, including all components that needed to be installed for the scope of work of the MEP subcontractors down to 50 mm (2 inch) piping.

1998: MCWHINNEY OFFICE BUILDING, COLORADO The 4D model for this small commercial project allowed junior engineers to improve a CPM schedule developed by the project manager and superintendent or the GC (Koo & Fischer, 2000). The improvements could have saved about two weeks in project duration. This study was also done after construction was completed. It demonstrated that 4D models have the potential to make junior engineers productive contributors to getting a project built.

1998: EXPERIENCE MUSIC PROJECT (EMP), SEATTLE The 4D models for this project with extremely complex geometry visualized various schedule versions so that the owner representative, architect, and GC could more easily understand the repercussions of, for example, delaying a decision. 4D models also showed detailed construction sequences. Although the architect, Frank O. Gehry and Associates (FOGA), made the 3D models available to the GC, the 4D modeler needed to add significant construction detail to the 3D model to generate a realistic 4D visualization (Fischer et al., 1998).

1998–99: PARADISE PIER, DISNEY CALIFORNIA ADVENTURE A 4D model including staging and laydown areas allowed the owner’s construction planning team to verify that the project timeline requested in the bid documents

Benefits of 3D and 4D models 13

was aggressive but realistic. The 4D model became part of bid documents. The owner, WDI used the 4D models in pre-bid meetings with the invited GCs to explain the scope and challenges of the project. The winning bid came in slightly under WDI’s budget and proposed a schedule that was two months shorter. Throughout the owner’s construction planning effort, the owner used the 4D models on desktops and in a CAVE to support design and schedule reviews. WDI further leveraged its investment into the 3D model developed for the 4D model to check 3D sight lines and to simulate the rides. On this project, WDI and CIFE researchers collaborated to develop a prototype 4D tool that emphasizes ease of use and interactivity. The prototype allowed planners to work with the 4D model at several levels of detail and make changes to the 3D model and schedule in the 4D environment (Schwegler et al., 2000).

WALT DISNEY CONCERT HALL The rest of the paper describes our most recent involvement in 4D modeling efforts on an ongoing construction project.

THE PROJECT, PARTICIPANTS AND MOTIVATION FOR 4D MODELING The Walt Disney Concert Hall (WDCH), designed by FOGA, is the new 2,300 seat home of the Los Angeles Philharmonic Orchestra. Located in downtown Los Angeles, the US $240 m project incorporates complex architectural, structural, and acoustical requirements in a tight one-city-block site. The project is scheduled for completion in early 2003. Figure 9 shows a photo of the front entrance to the WDCH.

Figure 9.

Photo of the physical model of the Disney Concert Hall.

14

M. Fischer et al.

The architectural process undertaken by FOGA provides opportunities and challenges for the construction of 4D models to assist in the construction planning process. FOGA’s design process yields a highly developed 3D CAD product model, which is used extensively for dimensional control and fabrication in the construction process. This product model and the process model contained in the construction schedule prepared by M.A. Mortenson Company, the GC, with input from many subcontractors, provide the necessary elements to begin construction of the 4D model. The GC used the 3D and 4D models as communication tools to share project information with all project participants including architects, engineers, the GC, subcontractors, and the owner. John Haymaker from the 4D research team at Stanford University worked on site to help build the 4D models discussed below and to introduce the GC and key subcontractors to the 4D modeling process. He used the prototype 4D modeling software developed through the collaboration of the Research and Development group at WDI and researchers in the 4D CAD research group at the CIFE at Stanford University (Fischer et al., 2001). The complex project and a tight site made precise coordination of construction activities a very high priority. M.A. Mortenson saw the use of 4D visualization of the construction process as a valuable tool for accomplishing four project objectives: Schedule creation: 4D models help visualize schedule constraints and opportunities for schedule improvements through resequencing of activities or reallocation of work space. Schedule analysis: 4D models help analyze the schedule and visualize conflicts that are not apparent in the Gantt charts and CPM diagrams. Communication: Many participants join the project in midstream, and it is critical to bring new participants up to speed quickly. Team building: The GC’s project superintendent, Greg Knutson, felt strongly that it was very important to construct a team atmosphere, where people solve problems together. He realized that a shared, visual model to externalize and share project issues was a valuable team building tool. The following section details the project information that was available at the beginning of the 4D process. Subsequently the process undertaken to construct the 4D models and describe the 4D models constructed for the project is examined. We have also described some of the issues and challenges encountered in constructing the models. The final discussion focuses on how the GC used the models to accomplish the objectives.

AVAILABLE ELECTRONIC INFORMATION The interest in constructing the 4D models emerged in early 2000, as the GC mobilized to the site. At this point, the architect had already developed most of the

Benefits of 3D and 4D models 15

3D geometry, and the GC’s construction schedule had about 4,000 activities. This section describes the format and level of detail of the project information at the beginning of the 4D modeling process.

AVAILABLE 3D GEOMETRY The architect constructed the 3D models with CATIA. There are at least two reasons for the use of CATIA as the 3D modeling software. First, FOGA develops very complex geometry and considers the nature of the curves generated to be integral to the architectural design. CATIA uses NURBS-based curves and surfaces, which describe the curves mathematically, and therefore maintain a high level of accuracy. More traditional CAD packages for the AEC industry do not use NURBS, instead approximating the curves and therefore loosing the level of accuracy desired by FOGA. The second motivation for using CATIA is that the software handles very large, complex models. As described below, the architect modeled a great deal of the project in 3D, and the shear amount of information would overwhelm traditional AEC CAD packages. To reduce complexity, FOGA divides the 3D model into sub-models. First, FOGA divides the project geographically into “building elements,” as shown in Figure 10. Each building element is then further divided into models reflecting different significant building systems. Figures 11 to 17 show the different models available

Figure 10.

WDCH broken down by building element.

Figure 11.

Surface models for all building elements.

16

M. Fischer et al.

Figure 12.

Element 2 surface model.

Figure 13.

Element 2 pattern model.

for building element 2. Figure 11 shows all of the building elements’ surface models incorporated into one view. Figure 12 shows the surface model for element 2. The surface model contains everything that can be seen, from plaster, to glazing, to carpet, to wood panneling, etc. Figure 13 shows a pattern model. A pattern model describes any pattern in an element that is relevant for architecture or construction.

Benefits of 3D and 4D models 17

Figure 14.

Concrete model for element 2.

Figure 15.

Air and water barrier model for element 2.

Figure 13 shows the pattern of the stainless steel panels for the exterior of element 2. Figure 14 shows the concrete model, which models the structural and architectural concrete surfaces. Figure 15 shows an example of an air and water barrier model. The air and water barrier model defines the surface in space where the waterproofing systems should be placed. Figure 16 shows the structural wireframe model. This model defines a wire for each piece of steel in the building. The wire can symbolize centerline, top of steel, or bottom of steel. The steel detailer and the steel fabricator use this wire model as input and place the proper size member with each wire. The detailers detail all the connections in X-Steel or other detailing packages in 3D. The resulting detailed steel model, shown in Figure 17, is then re-imported into the CATIA model.

18

M. Fischer et al.

Figure 16.

Element 2 steel wireframe model.

Figure 17.

Detailed steel model.

Each 3D model consists of layers reflecting different sub-systems. Table 2 shows a partial listing of the layers. These layers are helpful for 4D modeling because they isolate certain scope information in the 3D model, which facilitates the identification of the appropriate geometric elements for a particular activity. However, frequently the layering organization is different from the organization of

Benefits of 3D and 4D models 19 Table 2.

A portion of the layer list.

Layer No. CATIA layer contents

Layer No. CATIA layer contents

General project data (1 Thru 10) 1 2 3 4 5 6 7 8

11 12 13 14 15 16 17 18 19 20 21

Project grid Column grid lines Property line Vacation envelope Project reference geometry Project workpoints CATIA construction geometry Existing construction Glazing assemblies (11 Thru 25) Skylight glazing Sloped glazing Vertical glazing Mullion wireframe (Center Line Mullion) Mullion Metal closure Trim Metal closure Panels Metal gutter Metal flashing Glazing anchor assembly Glazing boundary

Stone (46 Thru 55) 46 47 48 49 50 51 52 53 54 55

56 57 58 59 60 61 62 63 64 65

Vertical stone cladding Sloped stone cladding Stone coping Stone paving Stone base Decomposed granite Not used Not used Not used Not used Roof asssemblies (56 Thru 65) Roof membrane Type 1 Roof membrane Type 2 Roof hatch Roof drain Stainless steel gutter Expansion joint assembly Roof davit pedestal Roof assembly Type 3 Not used Not used

Metal panel assemblies (26 Thru 45) Miscellaneous exterior assemblies 26 27 28 29 30 31 32

Metal panel assembly condition Type 1 Metal panel assembly condition Type 2 Metal panel assembly condition Type 3 Metal panel assembly condition Type 4 Metal panel assembly condition Type 5 Metal panel assembly condition Type 6 Metal panel assembly air and water barrier

Layer No.

66 67 68 69 70 71

Not used Metal grill Metal grating Building maintenance equipment Building maintenance track Stain less steel clad door

86 87 88 89 90 91 92 93

96 97 98 99 100 101 102 103 104 105 106 107 108 109 110

126 127 128

the schedule, and the 4D modeler needs to reorganize the geometric information for the 4D model to fit the schedule organization (Fischer et al., 1998).

AVAILABLE SCHEDULE INFORMATION The GC created the construction schedule with Primavera’s P3™ software. At the start of the 4D modeling process in March 2000, the schedule contained about 4,000 activities. By Fall 2000, the schedule consisted of approximately 7,200 activities. The schedule divides the 3D project geometry into chunks that are relevant to an activity. Figure 18 shows the breakdown key for the activity ID in the

20

M. Fischer et al.

schedule. Activities are identified by building element, floor, area, and subarea, then by phase, system, component, and action. However, some activities do not fit easily into this breakdown. For example, steel installers like to break the steel into manageable chunks, called sequences, which are a grouping of steel that is selfsupporting and can be erected in a reasonable amount of time. These sequences often span more than one building element, or cover more than one floor. Even though it was useful to have one main way to organize the schedule (as shown in Fig. 18), many methods for decomposing the geometry and linking a scope of work to an activity are required to suit different types of work. Figure 19 shows

Figure 18.

Activity code key for defining activities and relating them to the 3D model.

Figure 19.

Organization of 3D model into levels, sequences and thirds.

Benefits of 3D and 4D models 21

the project broken into levels (red) and sequences (green). Figure 19 also shows the main potion of the Concert Hall broken into thirds (blue) as the GC organized some of the work in the main hall in this way.

4D MODELING PROCESS AND 4D MODELS Figure 20 maps the process for constructing the 4D models from the project geometry and schedule and shows the file formats used to translate between computer programs. Rhino3D™ proved to be very useful to import the NURBS-based geometry from CATIA, add names to the geometry, break up the geometry into relevant configurations for the respective activities, and convert the geometries to VRML. Named geometrical elements allow a 4D modeler to match geometry names to activity names quickly. We built four 4D models for the project. Figures 21 to 24 show a screen shot from each of these models.

Figure 20.

Process for constructing 4D models from 3D models and CPM schedules.

22

M. Fischer et al.

Figure 21. Steel, Concrete, and Exterior Enclosure model. This 4D model examines the overall sequencing for the major structural and enclosure activities. It shows the sequencing of steel and of structural and architectural concrete. It includes metal decking, roofing, glazing, and enclosure systems, such as metal cladding assemblies including secondary steel supports. Statistics: Number of 3D components: 340; Number of polygons: 515,000; Number of activities: 512.

Figure 22. Element 2 model. This 4D model goes into more detail for building element 2. It includes the interior work. The model includes interior stairs, elevators, fireproofing, and finishing systems. It shows mechanical and electrical activities by highlighting the floor slabs in the area of work. Statistics: Number of 3D components: 105; Number of polygons: 85,000; Number of activities: 185.

Benefits of 3D and 4D models 23

Figure 23. Interior hall model. The interior of the Concert Hall is a highly congested and complex space. All of the interior activities are squarely on the critical path. The model includes all the activities affecting this space: structural steel, concrete, plaster, wood finishes, mechanical, and electrical. The model also includes scaffolding. Statistics: Number of 3D components: 210; Number of polygons: 325,000; Number of activities: 667.

Figure 24. Detailed Hall Ceiling Model. In early 2001, we are constructing a fourth model to help with the detailed planning of the complex concert hall ceiling installation. Statistics: Number of 3D components: 180; Number of polygons: 520,000; Number of activities: to be determined.

24

M. Fischer et al.

CHALLENGES ENCOUNTERED WHILE BUILDING THE 4D MODELS The construction of the models posed a number of challenges related to the geometry, the schedule, and the linking of the geometry and the schedule. In our experience, such issues are quite common during the development of 4D models, especially when the 3D models are created without knowledge of the needs for 4D modeling and construction planning. Another reason for these issues is that the construction of a 4D model requires significant project scope and schedule information. Some of this information is precisely the information that project participants want to develop or refine through the 4D modeling process, and other information is simply not yet available because of resource or other constraints. A valuable contribution of the 4D modeling process is that the process makes it very clear where complete scope and schedule information exists and where additional thinking is needed.

GEOMETRY ISSUES Inconsistencies: The 3D models from the architect contained some inconsistencies. For example, an object that was on the plaster layer should have been on the gypsum board layer. Such inconsistencies create extra work during the linking of the schedule and the 3D model because the 4D modeler cannot easily identify, isolate, and show the scope of work for a particular activity in 3D. Lack of data: The surface model models only what is seen. In the case of a wood wainscot on a plaster wall, FOGA modeled the plaster only where the wood wainscot does not cover it. Even though there is plaster under the wood wainscot, it is not modeled. Hence, in those areas, the surface model does not provide 3D components that can be linked to activities. In addition, for some of the scope of work for steel erection the 3D models were also incomplete. The steel detailers took the wire models from the architect, and produced detailed 3D models from these wires. This process was time-consuming, and at the time of 4D model construction, the detailers had detailed only some of the steel for the main concert hall box. The rest of the steel had to have a 3D representation so that it could be seen during 4D model simulation. We created an algorithm to hang a simple rectangular shape on the wires to make the important information visible without overwhelming the software or the user. Level of detail: Sometimes there is too little detail in the 3D model. The steel 3D model came back from the steel fabricator all on one layer. However, one might want to split primary and secondary steel into two activities, which would make it necessary to have the primary and secondary steel on two layers. In addition, FOGA modeled just the surfaces for the metal skin. A metal skin system requires backing support and clips, which were not modeled, but need to be installed, and should therefore be reflected in the 4D model.

Benefits of 3D and 4D models 25

Figure 25.

Simple extrusion on wire.

Figure 26.

Mixture of two types of steel models.

Too much data: Sometimes there can be too much information, which slows down the computational processing of the 3D and 4D models. For example, the steel came back from the fabricator with all the bolts and holes modeled, but we did not need this information for the 4D models the GC wanted to create. Figures 25 to 27 show steel handled at two levels of detail for this project. The resolution of certain situations requires more detail, the resolution of others less detail. SCHEDULE ISSUES Inconsistencies: Just as the geometry can be inconsistent with the design intent, the schedule can also contain inconsistencies. For example, the schedule may call for a Concrete Masonary Unit (CMU) wall, whereas the geometry models a cast in

26

M. Fischer et al.

Figure 27. Steel model from fabricator.

place concrete wall. The inconsistency must be resolved, which, while valuable from the project standpoint, is time-consuming for the 4D modeler. Lack of data: Some geometry has no corresponding activity. Again, an activity may be required, but resolving this issue requires time and resources of the modeler.

ISSUES WITH LINKING OF 3D MODEL AND SCHEDULE Inconsistencies: Often, the geometry is defined in ways that conflict with the schedule. For example, the architect defined the geometry by building elements, but the GC places concrete and steel not by element, but rather according to steel sequence. The geometry had to be broken down and recombined a great deal to get a geometrical configuration to match the schedule. Other data: Cranes, laydown and staging areas, scaffolding, etc. are not part of the architect’s design model, but these elements play a large role on the construction site. We had to add these geometries to the 3D model. Figure 17 shows a crane we added to the 3D model to explore the spatial relationship of the crane and its location over time with surrounding work. Representation of activities with no geometry: Ductwork was not modeled in 3D on most of the project, but the GC was interested to know when and where ductwork was scheduled. A 4D modeler has to be sensitive as to the best way to communicate such activities, by perhaps attaching the activity to a floor slab (as we did), or ceiling framing.

Benefits of 3D and 4D models 27

USES OF 4D MODELS The 4D models supported M.A. Mortenson’s four objectives in the following way: Schedule creation: The GC used the 4D models to assist in planning the laydown areas for the enclosure contractor, to visualize overall project access at critical junctures in the project, to refine the interior and exterior scaffolding strategy, and to plan the installation of the complex ceiling of the main concert hall. Schedule analysis: The GC’s project management team used 4D models to discover several conflicts in the schedule which were not discovered in the CPMbased Gantt chart. Figures 28 to 30 show snapshops of the 4D models that show particular problems. Figure 28 shows a situation where a CMU wall was scheduled too early while steel was being erected directly overhead. Because the wall that is framed by the steel leans outward the steel erection requires shoring (not modeled), which would not only interfere with the construction of the CMU wall but also cause a dangerous situation. Figure 29 shows an Air Handler Unit (AHU) being installed too late after the steel is completely erected. There would no longer be the access necessary for the large AHU. After consulting with other project team members, the GC decided to leave some of the steel out to make it possible to slide the AHU into the structure at a later date. Figure 30 shows a conflict of scaffolding systems in the same area of the interior hall. The scaffold for the plastering of the walls will need to be removed before the ceiling scaffold can be erected. As a result of the schedule analysis through the 4D model of the interior construction the GC decided to consolidate the scaffolding contracts for the interior hall from three contracts to

Figure 28.

CMU wall (in dark green) scheduled too early.

28

M. Fischer et al.

Figure 29.

AHU (shown in red) scheduled too late.

Figure 30.

Scaffolds collide.

one contract. The 4D models supported the discovery of these (and many similar issues) during planning, well before construction started. Note though, that because of the physical and temporal interrelationships between many scopes of work an early detection of potential problems is essential to revise the design or schedule economically. For example, even though the AHU was not scheduled to be installed

Benefits of 3D and 4D models 29

Figure 31.

Collaboration in the Virtual Reality Cave.

for many months it was critical to identify potential AHU installation problems prior to work being released for steel fabrication to ensure that the right steel was installed (and not more). Communication: The GC used the 4D models in training sessions with as many as 40 people, where subcontractors, owners, designers, and the GC reviewed the models and discussed the strategy and constraints for erecting the project. Figure 31 shows a view of subcontractors in a meeting in the WDI CAVE. Team building: After a 4D review session ended, it was not unusual to have people from different subcontractors remain in the room for an hour or more beyond the scheduled meeting time to discuss issues and solutions to problems or questions identified during the meeting. The GC’s project superintendant mentioned that, in a tight labor market, where everyone is committed to too many projects, it is critical to get the attention and collaboration of the subcontractors focused on his project. Given the complexity of the project he wanted to make sure that the subcontractors put their creative energy into improving the construction of his project.

CONCLUSIONS Our applications of 4D models to construction projects have shown that 4D models help avoid or overcome many of the inefficiencies found on projects

30

M. Fischer et al.

today: congestion, out of sequence work, multiple stops and starts, inability to do detailed planning in advance, obstructions due to material stocks, etc. (Koskela, 1999). In all cases except on the Sequus Pharmaceuticals, the Experience Music Project, and the WDCH, the 4D modeling effort required the construction of a separate 3D model because the design had been done in 2D, or the 3D models were not up to date or incompatible with the 4D modeling tools. The schedule information could be used as it was, but often the project team decided to make the activities more detailed to see more detail in the 4D model. As can be seen from Table 1, for many of the benefits the generator of the information necessary for 4D modeling is not the same party realizing the benefits of 4D modeling. Hence, the realization of the benefits of 4D models on projects with a traditional design-bid-build approach often requires extra modeling work. However, the benefits a GC or a subcontractor can realize from 4D models still often outweigh the cost of building the necessary CAD models. On the Sequus project the owner avoided this extra work by awarding a design-build contract to a team consisting of Flad & Associates (architect), Hathaway-Dinwiddie (GC), Rosendin Electric, Paragon Mechanical, and Rountree Plumbing. This maximized the opportunity for each party to enter and maintain the information in the 3D CAD model necessary to realize the benefits. In summary, 4D models allow project stakeholders to work out many design and construction issues in the computer model before actual construction, maximizing project value to owners and making it more likely that the project will be completed as planned and designed.

ACKNOWLEDGMENTS We are indebted to many professionals and students who have been instrumental in making our 4D modeling efforts successful. We would like to acknowledge the following people in particular: Buddy Cleveland, Jerry King and Kent Simons for the technology support over the years; Jack Ritter, Tom Trainor and George Hurley for getting us started on the San Mateo County Health Center; Todd Zabelle and Greg Silling for sharing construction insights with us over the years; Melody Spradlin and everyone else from the Sequus project for going live; Jim Glymph, Kristin Woehl, and Dennis Sheldon from FOGA for the challenge, fun and excitement of applying 4D models on FOGA projects; Chris Raftery for letting us participate in EMP and Lisa Wickwire for keeping us current with project information on EMP; Ben Schwegler from WDI for his substantial financial and intellectual support and for co-hosting the workshop; and Greg Knutson, Derek Cunz, Jim Yowan, David Mortenson, David Aquilera, and Joe Patterson on the WDCH.

Benefits of 3D and 4D models 31

REFERENCES Akinci, B. & Fischer, M. 1998. Time–space conflict analysis based on 4D production models. In K.C.P. Wang (ed.), Proceedings of congress on computing in civil engineering: 342–353. Reston, VA: ASCE. Atkins, D.C. 1988. Animation/simulation for construction planning. Engineering, construction, and operations in space: Proceedings of space 88: 670–678. ASCE. Cleveland, A.B., Jr. 1989. Real-time animation of construction activities. Proceedings of construction congress I—Excellence in the constructed project: 238–243. ASCE. Collier, E. & Fischer, M. 1995. Four-dimensional modeling in design and construction. Technical Report, Nr. 101. Stanford, CA: Center for Integrated Facility Engineering (CIFE). Collier, E. & Fischer, M. 1996. Visual-based scheduling: 4D modeling on the San Mateo County Health Center. In J. Vanegas & P. Chinowsky (eds), Proceedings of the 3rd congress on computing in civil engineering: 800–805. ASCE. Danhier, B., Massonnet, A. & Verminnen, F. 1994. SG replacement problems anticipated and avoided with 3D CAD. Journal of Nuclear Engineering International 39(474): 16–18. Edwards, R. & Bing Z. 1999. Case study: 4D modeling and simulation for the modernization of Logan International Airport. Proceedings of the 1997 international conference on airport modeling and simulation: 8–27. ASCE. Fischer, M.A. & Aalami, F. 1996. Scheduling with computer-interpretable construction method models. Journal of Construction Engineering and Management 122(4): 337–347. ASCE. Fischer, M., Aalami, F. & Akbas, R. 1998. Formalizing product model transformations: case examples and applications. In I. Smith (ed.), Artificial intelligence in structural engineering: Information technology for design, collaboration, maintenance, and monitoring, Lecture Notes in Artificial Intelligence, 1454: 113–132. Springer. Fischer, M., Liston, K. & Schwegler, B.R. 2001. Interactive 4D project management system. The 2nd civil engineering conference in the Asian region, Tokyo, 16–18 April, 2001 (accepted for publication). Fröhlich, B., Fischer, M., Agrawala, M., Beers, A. & Hanrahan, P. 1997. Collaborative production modeling and planning. Computer Graphics and Applications, IEEE 17(4): 13–15. Kahan, E.T. & Madrid, X.H. 1987. Integrated system to support plant operations. Hydrocarbon processing symposium: 55–60. ASME. Koo, B. & Fischer, M. 2000. Feasibility study of 4D CAD in commercial construction. Journal of Construction Engineering and Management 126(4): 251–260. ASCE. Koskela, L. 1999. Management of production in construction: a theoretical view. In I.D. Tommelein & G. Ballard (eds), Proceedings of the seventh annual conference of the International Group for Lean Construction (IGLC-7): 241–252. Retik, A. 1997. Planning and monitoring of construction projects using virtual reality. Project Management 3(1): 28–31. Songer, A.D., Diekmann, J. & Al Rasheed, K. 1998. Impact of 3D visualization on construction planning. In K.C.P. Wang (ed.), Proceedings of congress on computing in civil engineering: 321–329. Reston, VA: ASCE. Schwegler, B., Fischer, M. & Liston K. 2000. New information technology tools enable productivity improvements. North American Steel Construction Conference, American Institute of Steel Construction (AISC), Las Vegas, 23–26 February: 11-1 to 11-20.

32

M. Fischer et al.

Staub, S., Fischer, M. & Spradlin, M. 1999. Into the fourth dimension. Civil Engineering 69(5): 44–47. ASCE. Vaugn, F. 1996. 3D and 4D CAD modeling on commercial design-build projects. Proceedings of computing in civil engineering congress: 390–396. ASCE. Williams, M. 1996. Graphical simulation for project planning: 4D-Planner™. Proceedings of computing in civil engineering congress: 404–409. ASCE.

BEYOND SPHERELAND: 4D CAD IN CONSTRUCTION COMMUNICATIONS Dennis Fukai M.E. Rinker, Sr. School of Building Construction, University of Florida, Gainesville, FL, USA

Abstract This study examines the fourth dimension as the product of a fundamental shift in a paradigmatic world view. This shift changes the normal way of “seeing” or visualizing the obviousness of the context of our everyday practices and leads in a different direction with a completely new vision of the processes antiquated by its transformation. This can be seen in the renaissance of closely held ideas that occurred in the change from an oral tradition to descriptive diagram, from diagram to written and reproducible text and twodimensional images, and from simple image to perspective drawings and photographs. These are “visionary” changes that triggered immediate and lasting displacements in our social and technical development. The shift from three to four dimensions in computer aided design (CAD) does not seem to have had this revolutionary impact even though its value and potential have been made quite clear by a number of researchers. As a consequence, this study explores the context of these new computational tools and how they might be used to enhance the communication process in construction. It suggests that computer mediated communications in construction might be better used to understand the process delineated by a model’s construction, where that model is developed as a preview of its construction, “built” according to the same methods and techniques anticipated in the actual project. Keywords: computer, visual, communications, construction, modeling

INTRODUCTION: “UPWARD, YET NOT NORTHWARD” In the 1880s, Edwin Abbott wrote a classic book about the idea of dimensions that has been reprinted more than six times (Abbott, 1952). Abbott was a schoolteacher writing about a society of objects that inhabited a land he called Flatland. There 33

34

D. Fukai

Figure 1.

A circle appears as a line when you live in Flatland.

were no solids in Flatland. Instead, objects existed in two dimensions and were viewed along their edge. This means geometric shapes had lengths and widths, but because they had no height, they always looked like a line. As shown in Figure 1, this is much like a circle first standing on its edge so you can see its threedimensional face and then laid flat on a table so that only the edge is visible. Unable to rise above this edge view meant circles always appear as a line in Flatland. This also meant that a square and a circle looked the same. To “see” the difference, one had to feel the edge of an object to know whether it had corners or curves. There was an educated class of citizens in Flatland that could distinguish those shapes without touching them because their eyes had been trained to see the object’s outer edge fade slightly into a fog or haze. However, ordinary people did not have the skill to visualize this subtle shift in dimension. “Space” was therefore flat on a two-dimensional plane and measured in the direction of the length and width of the edge of an object. This meant objects could move north, east, south, and west according to two axis, but it took great skill and a fundamental understanding of the nature of this two-dimensional “space” to move around. Abbott notes there was a “southward” attraction that could be felt in some regions of the plane, but since the edge of an object was always a line, any view of distance had no perspective. In other words, an object could be near or far, polygon or curved, or open or closed, as one moved from place to place. Everything looked the same and it took education and training to read two dimensions and understand the subtle variations of the forms it contained in order to navigate without getting lost. A sphere called the “stranger” came into Flatland. When it arrived, it introduced another dimension: up and down. From the Flatlander’s point of view, a sphere moving up and down through the edge view of the two-dimensional surface of Flatland appeared as a line that grows shorter and shorter until it disappears. This meant that a three-dimensional object like a sphere might look like any other circle or square, but it could actually rise above its two-dimensional plane and disappear as illustrated in Figure 2. No other object in the world view of Flatlanders could shrink, stretch, or disappear in and out of its restricted twodimensional view like a three-dimensional object. Even more amazing to the inhabitants of Flatland, was that the very notion of up and down meant that things that were once secret or hidden when viewed from their edge were now exposed when viewed from “above.” This meant the inside of a two-dimensional circle could be seen from this other dimension. Anyone able to see in this new dimension could therefore see things that were once considered

4D CAD in construction communications 35

Figure 2.

A sphere appears as a diminishing line when it moves through Flatland.

private and privileged. The idea of a new and elevated perspective changed the fundamental concepts of their world. It required a paradigmatic shift in their way of thinking that made all the “substantial realities” of their world view “appear no better than the offspring of a diseased imagination, or the baseless fabric of a dream” (Abbott, 1952). Of course, the leaders of Flatland refused to believe in the possibility of another dimension and prohibited all discussion of its existence throughout their land. The idea was simply too disruptive to consider because they were locked into their own restricted, but well-ordered, world view. To see another dimension meant highly skilled citizens had to change their way of seeing. This is not easy for anyone to do, not only because it is disorienting, but also because it calls for a perceptive displacement in a way of living that is not easy to accommodate.

TWO, THREE, AND FOUR DIMENSIONS? This is the same difficulty many students face when they first look at a complex set of two-dimensional construction drawings (Wilson, 1997; Wei & Gibson, 1998). The idea that a three-dimensional object can be projected onto a collection of two-dimensional planes is contrary to their view of their world. As shown in the example in Figure 3, plan-reading calls for training in order to “see” the shapes and images represented by the lines and symbols laying so narrowly defined on the surface of a piece of paper. Students eventually learn to read plans, but they only learn to see them in three dimensions after they have had a good deal of construction experience. In practice, the relationship of two-dimensional drawings to three-dimensional space in the design of buildings is less of a challenge. This is because most floor plans require little more than the ability to visualize a vertical extrusion of a collection of lines, certainly not much of a challenge for designers to draw, and even less challenging to visualize before construction. And when spaces are stacked one on top of another in multiple stories, they most often become a series of identical floors, nothing more than a vertical collection of the same extruded twodimensional spaces. The restrictions of our perceptions as designers and builders therefore seem to confine us to spaces that are relatively simple to draw, visualize,

Figure 3.

It takes experience to see in three dimensions.

36 D. Fukai

4D CAD in construction communications 37

and build. There are exceptions, but when the imagination of an architect or engineer produces something more complex and three-dimensional than the ordinary, the result is difficult to understand and usually more costly to build. The variation from the norm requires a higher level of interpretation and the result can be at once disturbing and exciting. This odd perceptive shift from three to two and back to three dimensions also constrains the potential of the buildings that we build. A designer imagines space in three dimensions, but then must translate that space to two and the builder must then take the two-dimensional drawings and transform the lines back to the third dimension.

FOUR DIMENSIONS AND MORE INFORMATION? This becomes even more disorienting when we introduce the idea of a fourth dimension. Now computers automate much of the information found on a twodimensional drawing. For some, any computer aided drawing (CAD) introduces a fourth dimension when it is correlated with computer-generated data not normally found in hand drawings (Wright, 1994; Vaugn, 1996; Cardone et al., 1999). However, others point to the obvious flaws in the assumption of this dimensional shift when applied to CAD documents (Shah & Wilson, 1988). If information is a broad category of “stuff” or “chunks,” it can in itself describe multiple dimensions of perception (Hofstadter, 1980). In the same way, written descriptions, perhaps associated with two-dimensional diagrams, can bridge multiple dimensions in a way that only a writer can describe. It follows then that a construction document, with the addition of a description or sequential representation that was already inherent in that document, does not necessarily approach a true fourth dimension (Shah & Wilson, 1988; Falcioni, 1999). For example, it is hard to argue that a time sequence image of two-dimensional images shown in Figure 4 represents three or four dimensions. Similarly, by associating the two-dimensional symbols for chairs, desks, lamps, and credenzas as informational “Blocks” in a CAD program, a furniture layout can be made to automatically sort, count, and list the quantities of the objects that have been inserted into the two-dimensional plan. Is this three or four dimensions? Of course, the apparently flat furniture plan has a very important third dimension. Walking through the spaces before the installations would show vertical dimensions of counters, furnishings, equipment, steps and stairs, and the height on the walls where signs, clocks, or coat racks are to be placed. There is no doubt that the representation of three-dimensional space and its associated information are well served by the automated references that can be generated by a database, however, it begs the question: do two-dimensional diagrams and symbols represent three-dimensional objects with the automated data defining a third dimension?

38

D. Fukai

Figure 4. Three phases for the installation of concrete in a runway construction represent four dimensions.

Most would agree that a two-dimensional diagram of a three-dimensional space is not in itself three-dimensional. A three-dimensional document must be drawn in three dimensions to visually break from a two-dimensional plane (Wei & Gibson, 1998). After all, there are no true vertical relationships in the average extruded floor plan, and furniture installations can most often be completed with no special visualization skills. In practice, the actual construction may not even follow the original plan, primarily because placement will be adjusted according to the way objects “fit” within the assembly. The perceptive reality of the plan will thereby be ignored once its diagrammatic representation is over shadowed by the actual space. In other words, once we see the real thing, everything else is a “baseless fabric of a dream” (Abbott, 1952). With the introduction of 3D CAD programs, construction models can now be built to meet this same perceptive challenge (LaCourse, 1990). These are models assembled from the three-dimensional pieces of a total structure. As shown in Figure 5, the combination of the assembly of a building as solids and the ability to view these solids from many viewpoints greatly enhances a constructor’s understanding of spatial relationships, construction details, and fabrication techniques. At the same time, it is difficult to actually construct an object from a threedimensional model even though, as shown in Figure 6, annotations can be added to explain the construction. This is because details about materials, dimensions, and specifications are still required to build the object on a construction site. To meet this challenge, some innovative practitioners have begun to use three-dimensional models to create two-dimensional drawings (Wilson, 1997). In AutoCAD2000, operators can use “layouts” to convert models to two-dimensional diagrams, annotate and dimension them for use in the actual construction. The perceptive shift to a three-dimensioned construction model therefore reverts by default to the standard two-dimensional drawing.

4D CAD in construction communications 39

Figure 5. Three-dimensional models bridge the perceptive gap left by two-dimensional drawings.

Figure 6. Three-dimensional models can be annotated, but it is difficult to show the kinds of layout dimensions necessary for actual construction.

40

D. Fukai

Figure 7. CAD–CAM models used in rapid prototype development embody time in the manufacturing process.

There are exceptions, of course. For example, objects like machined parts can be fabricated directly from the three-dimensional model using CAD–CAM and CNC equipment (Kamarani, 1999). As the result of this direct link between the computer model and the fabrication of a product, an industry of rapid prototypes and integrated design and manufacturing method analysis has emerged. This includes the notion that computer modeling can include a sequential analysis of time as a fourth dimension (Potter, 1998). The kind of equipment shown in Figure 7 shows how computer models can define both the shape, sizes, and sequences of the manufacturing process, as turns on a lathe, cuts on a milling machine, or holes drilled. This suggests that time is evident in the machine’s interpretation of the three-dimensional model. In other words, since the results emerge from the computer model through the mechanisms of a machine, the production “process” is defined by the construction document (Jerrens, 1999). In the same way, the plotted output of a three-dimensional model in a CAD program includes time as part of its documentation, if the process and sequence of the application of ink to paper for the resulting image was specifically defined as the output of the computer model. These time relationships between model and fabrication point to the importance of time in any description of the fourth dimension. Again, the idea raises the question: where is time in this relationship? Does time occur during the actual output as a production process? Or is it embedded in the informational description added to the representation of that process in the drawing or model itself? When we consider the construction of large objects like buildings and other engineered structures, the representation of time in a computer model becomes even more blurred. First, is an isometric or perspective drawn flat on a piece of

4D CAD in construction communications 41

paper or visible on a computer screen three-dimensional? And if “chunks” of information are added to the isometric drawings can it be said that this information represents another dimension? It seems logical that the result is simply an extension of the same information that was already resident in the drawing. Why then is the addition of time, in a similar image or visual composition, not simply another informational layer? How does time extend the underlying model into a revolutionary fourth dimension?

TIME AS A DIMENSION Defining the fourth dimension is a matter of dimensional relativity. Abbott wondered if the fourth dimension had anything to do with an unseen axis in threedimensional space. He saw a mathematical progression. If a non-dimensional “point” has a single point, a one-dimensional “line” has two points, a two-dimensional “square” has four points, and a three-dimensional “cube” has eight points; then the fourth dimension might be some extension of an object with 16 points. Abbott thought that within the perceptive paradigm of three-dimensional space the fourth dimension might be something he called “extra-solids” produced by “motion of the solids” and “double-extra solids” that result from the “motion of the extra-solids through space.” This analogy of solids in motion through space is interesting when we consider that Abbott was writing these words when Einstein was a child, and space and time relationships remained to be hypothesized and tested as a theory of relatively. It was with Einstein’s work that the idea of a three-dimensional world took on the fourth dimension of time. For Einstein, objects no longer simply exist as solids, instead they were part of a continuum of time, perpendicular to the space created by the juxtaposition of the mass of that object. Time is thereby measured by the movement of light and is affected or changed by mass to produce a series of “relativities” defined by the position of the observer. Non-physicists have taken the notion of time as the fourth dimension and used time in its simplest form to describe a sequence of passing events or phases. In this interpretation, time can be as simple as a series of photographs that capture a particular event. However, this does not seem like an elegant interpretation of the fourth dimension. Are a series of photos showing a sequential event in Figure 8, four-dimensional? Time can also be shown in a series of model images as the phases of an object’s production, operation, deterioration, and/or maintenance (see Fig. 9). In this model, “time” is embedded as a representation of the “motion” of an object in space. The result is that the fourth dimension in CAD has come to be understood as the visual representation of time in the form of images of the phases of the evolution of a three-dimensional model. Others argue we cross the threshold of a

42

D. Fukai

Figure 8.

A sequential series of photographs as a four-dimensional image.

Figure 9.

Time as the fourth dimension of a three-dimensional model.

4D CAD in construction communications 43

fifth dimension when layers of information are added to a four-dimensional model (Cardone et al., 1999). When we think of the complexity of Einstein’s theory, the idea that time can be represented in a three-dimensional model as the sequence or phases of a construction, does not seem to reach its full potential. After all, his theory is that the space–time relationship of the fourth dimension is affected by the mass of the solids. In fact, the energy produced by the relationship of mass and time is a derivative of this relationship. This means that the presence of the solid object must distort and constrain time according to the point of view of the observer. Accordingly, the space–time dimension is a continuum that includes the observer; it is the relative position of the observer within this continuum that defines the perceptive shift in that person’s view of both time and space. It could be said that a sequential series of three-dimensional images are nothing more than additional information about the same three-dimensional object, even when the motion associated with the sequence discloses a new perspective (Yamaguchi & Liu, 1998). Time existed or could have existed in the original image as an annotation, it is only its interpretation or visual representation that has been enhanced. This of course includes notes about movement, flow, and events, past or imagined, of the actual object in a photograph or its representation in threedimensional space. It seems then that relationship of time and space in 4D CAD may therefore play to the immediate application of sequential modeling without looking to its full potential. There is no doubt that understanding the evolution of an object, either to document its construction or to visualize some aspect of the design, is important, however, this may be a narrow view of a larger space–time relationship.

ONWARD THROUGH THE FOG Perhaps the fourth dimension is not northward or upward, but onward through the fog of uncertainty. Consider that if the relative position of the observer in a space–time continuum changes the perceptive results of both space and time for that observer, the true fourth dimension may depend on how the observer visualizes changes in the model over time. This would make change important in understanding the continuum of these dimensional boundaries. This is evident in the work of a number of researchers. For example, the Center for Integrated Facility Engineering (CIFE) used a construction method model (CMM) and expanded it into something called “Collaborative 4D CAD” (Aalami & Fischer, 1998). A CMM is a variation on a phased model that is used to analyze the construction process through visualization. It asserts the inclusion of a fourth dimension because its CMM analysis focuses on a process that relates planning and management decisions directly to the sequence of the building’s construction.

44

D. Fukai

Collaborative 4D CAD is a variation of this analytical model in that it includes the scheduling requirements and milestones which may change or constrain the construction process (McKinney et al., 1996). Its goal is to establish a “common ground” on which all designers can interact and enable them to better understand the design intent. This has been explored in a number of other variations in other disciplines (LaCourse, 1990; Potter, 1998; Johnson, 1999). In construction, interactive 4D CAD represents time as a schedule tied to sequential variations of a single representational model. In other words, the information on a schedule is correlated to the information in a CAD model to help visualize a sequence of events. The magic of this idea is that the model can be presented as a sequence of assemblies that directly parallel the critical points on the schedule. In theory, the model could be layered into a very detailed animation that would simulate the total construction, or at least one view of that construction. Of course, time-related visualization also includes the design, planning, and construction processes defined in Gantt charts, PERT diagrams, and/or CPM schedules. These diagrams abstract time as scaled lines and informational nodes, but do not always represent those abstractions with an image of the object’s construction. Instead, the image is implied through technical references to the construction documents. The image of the object is therefore resident in a schedule when interpreted by a trained and experienced “eye”. After all, construction managers have had to associate schedules with construction drawings and specifications in order to make the decisions required to build their projects for a very long time. With this in mind, if we build a three-dimensional model to represent the final design of an object and we learn to move around that object in a virtual space, it seems logical to assume that we have bridged the perceptive gap of a twodimensional diagram. There can be little argument against the perceptive value of a three-dimensional model. In fact, for those that can use the computer to construct these models, we know that immersion in this virtual environment can be an approximation of the real world (Johnshon, 1999; McKinney et al., 1999). However, the absence of motion and movement that were part of the model’s construction are not clearly represented in the three-dimensional form. In the end, one wonders if this embodied time and motion might be a prelude to the construction process (LaCourse, 1990; Kamarani, 1999).

THE CURVATURE IN TIME–SPACE FROM AN IMMERSIVE POINT OF VIEW The potential of improving project production and assisting managers in making decisions by helping them visualize and simulate a construction seems clear, but is there a deeper level of insight that might be derived from modeling space and time? For example, in a recent experiment with students at the M. E. Rinker,

4D CAD in construction communications 45

Figure 10.

Model developed in subcontracts by a novice team.

Sr. School of Building Construction at the University of Florida, several students constructed a building by dividing themselves into subcontracts (Fukai, 1996a). The students used ACAD R14 and had just completed about 20 hours of exercises in which they learned how to build solid models of concrete blocks, steel extrusions, concrete, and window assemblies. Using a common work point and “chalklines” to guide their assembles they were able to build the model shown in Figure 10. Perhaps not surprisingly for construction students, they began the work by breaking the building into its primary subcontracts: concrete and rebar, steel frame, masonry cladding, interior framing and finishes, and mechanical and electrical systems. They then used the workpoint and benchmark shown on the drawings and a common set of “chalklines” to locate the building and placed their assemblies from that workpoint according to the same process that would probably have occurred in the actual construction. The original idea was to simply construct this model in a team environment. Students understood how to merge files and were expected to divide the work, individually model components of the building, and then combine their files into a single three-dimensional model. Time limits were imposed to compress project delivery and set up the need for a coordinated effort to complete the construction. Setting a completion time contextualized communications and focused the interaction associated with the construction process on building the building as efficiently as possible. This included meetings, e-mail exchanges, presentations, a project website, and documentation of the construction process as daily logs, memorandums, and field notes that made up a loosely assembled construction information system. Interestingly, this data record documented the communications that occurred during the construction process as a chronology of interactions and the completed model became a graphical index to the related communication data. Sequencing the resultant model was impressive for three reasons (see Fig. 11). First, because it was a fully detailed construction model built by relative novices from the actual construction drawings in less than three weeks. Second, because

46

D. Fukai

Figure 11.

Subcontracts define the model construction and the modeling technique.

Figure 12. Coordination between team members parallel actual construction communications.

these students discovered a way to break down the model for this building that paralleled the way the building was actually constructed. Third, and most importantly, the students demonstrated that time, embedded in the construction of this model, was not in its final representation as the completed model, but in the interaction and communications that occurred during the modeling process. Like any building, “time,” as a function of the “mass” of the building, was thereby embodied in the components of the building as the energy expended in the virtual materials and labor. In other words, the pieces of the model constructed by these students contained the “process” from which it emerged. For example, as shown in Figures 12 and 13, the efforts to coordinate the assembly and exchange the kind of information the students needed to build this model meant that they also tested the quality of information on the construction drawings. The students found errors and omissions in the documents and had to send requests for information (RFIs) for clarification before they could proceed. Errors meant resolving alternate details and in some cases actually reworking the model according to change orders to compensate for the lost “time.”

4D CAD in construction communications 47

Figure 13.

The pre-construction model allows analysis of lifts to erect trusses.

THE FOURTH DIMENSION AS A COMMUNICATIONS PROCESS Most importantly, students had to communicate with each other to verify and coordinate their work. This goes beyond the idea of collaboration and visualization and more toward the directed and purposeful interaction that occur during the actual construction process. The two-dimensional drawings for the building were therefore critical to complete the virtual construction. This means there was no intent to achieve a predictable result, no neat representation of phasing, or preconception of assemblies intended to visually represent three-dimensional space. The modeling effort became a precursor to the actual construction process and the information that resulted was useful as a pattern of communications indexed by the individual pieces of the model. For example, in the virtual construction of this building, contact points and interfaces between the components of “subcontracts” had to be continually checked to make sure they fit together as shown on the drawings. Any error would be compounded in much the same way as the physical construction. This meant students had to calculate dimensions and the coordinates of assembly points to make sure prefabricated pieces would fit when they were “delivered” to the virtual job site. In some cases, this meant waiting until the assembly was complete before beginning prefabrication by transferring the project file to a “subcontractor” in much the same way the same subcontractors might move onto the job site to custom fit their portion of the work. Model construction therefore meant conflicts were discovered in

48

D. Fukai

much the same way they might be found in the field. In fact, many of the changes made to the model were visible in the completed building. These changes and the conflicts that were discovered in the pre-construction of this building were valuable as a precursor of the actual construction process, but the pattern of communications about the construction of this model emerged as the lesson learned from this experiment. The communications log showed that time and space were in fact changed by the construction. In the analogy of Einstein’s theory: time, or the warp of time, occurred as the “mass” of the construction model disturbed or changed the vector of assumptions made from a particular point of view. In other words, as one might suspect, a series of errors, omissions, and requests for information indicated the need for change orders that adjusted both scope and schedule. Though a simple experiment, and certainly not intended to be the kind of proof that Einstein sought for his hypothesis, it seems like this idea might in fact point in a different direction for what is more commonly thought to be 4D CAD. It also suggests a more a robust use of computer modeling, giving a “hands-on” experience from which to learn about constructing buildings while simultaneously previewing the quality of the construction documents. What is important is that the resultant model was 4D CAD, but unlike a sequential visual explanation like the series of images shown in Figure 14, time became the distortions and changes that were visible in the pre-construction communications. First, the modeling effort turned up errors and omissions in the construction drawings. This would be important to limiting requests for information, clarifications, and change orders that would have occurred during the actual construction. Second, the model helped to understand the quantities and methodologies associated with the materials and labor that would be part of the same process. In practice, this would help verify estimates and strengthen confidence in work plan simply because the construction process could be evaluated from the construction model rather than traditional two-dimensional drawings and specifications. Third, the resultant model provided an archive of graphical images indexed according to the pieces of the construction that would support future construction communications. This includes zoomed close-ups, three-dimensional details, time sequence or phased images, and two-dimensional projections used as shop drawings or field clarifications in both the building and the falsework, formwork, and special structures for the project. The pedagogical opportunities are equally exciting. One of the conclusions of the students involved in the experiment was that the experience had taught them a lot about constructing a building. They pointed out that they were also able to contextualize many of the things they had learned in other construction classes. For example, as shown in Figure 15, the construction of the running bond and reinforcing on a CMU wall is quickly modeled in a class assignment. If this is true, a similar 4D CAD experience for a specially designed building might be useful as a tool in construction education. The overall learning strategy

4D CAD in construction communications 49

Figure 14. Actual construction is enhanced by communications during the pre-construction modeling process.

Figure 15. Modeling a CMU wall teaches a lot about the actual construction of that wall.

50

D. Fukai

Figure 16. Construction model and physical construction can be used in combination to test ideas and planning.

of such a tool might include a construction project that allowed students to work in teams to complete the assembly of a virtual project. Students would then break the work down and interact with each other to build the virtual model within the time constraints. In these interactions, students would have to read the twodimensional plans, use hand-drawings to communicate their ideas in face-to-face meetings, build portions of the building, and transfer their ideas to other team members as email attachments. This could be supplemented by a full size version of the construction model to give students a hands-on feel for the actual construction (see Fig. 16). After completing the virtual construction they strengthened their plan-reading skills and their ability to visualize a three-dimensional object from two-dimensional plans. This suggests they could use hand-drawings and computer images to spontaneously visualize the construction. And perhaps most importantly they would be able to review the entire process to understand the context of their actions and how a similar pre-construction effort might be used for other projects.

PRE-CONSTRUCTION AS PRE-COMMUNICATIONS In an informal market study, the notion that buildings can be “pre-constructed” on a computer from a set of construction documents using the same labor and materials that would be used on an actual job site seemed to stir the interest of construction managers (Fukai, 1996a). In its simplest form this means prethinking a building prior to its construction; but it also means using computer modeling to

4D CAD in construction communications 51

improve job safety, simulate staging for difficult operations, coordinate complex lifts, discover errors and omissions in a set of drawings early in a project, analyze change orders, value engineering, and improve client communications in standard construction reports. It is important to emphasize that a pre-construction model is not the same as an architectural or engineering model. Architectural models help designers, builders, and owners visualize the spatial or aesthetic relationships in a building and are often used to illustrate phases or alternate finishes. Similarly, engineering models are programmed to illustrate the reaction of the structure to dynamically loaded forces placed on its frame or other structural elements. By contrast, a preconstruction model is an anatomically correct representation of all the pieces of the construction. This includes both the building and the falsework and formwork that will be required for its construction. Pre-construction models require a complete understanding of the construction process, including the methodology and techniques that will be unique to any particular project. In other words, the model’s assembly must include all the details of the construction and follow the same methods used in the actual production process. Thus the model becomes an instrument of communications, rather than visualization. To achieve this potential, it must be built in a way that allows clients to archive and manipulate images that will be useful during the project’s actual construction. This means the resulting model is not as important as the communications process that occurred during its virtual construction. To capture this process, the modeling effort should therefore be divided into subcontracts that parallel the work breakdown structure for the actual building. This delineates the responsibility for the pre-construction and sets up the context for communications. Using a common workpoint to coordinate the construction in much the same way it would be done on a job site, each piece of the building could therefore be identified according to its subcontract and placed on a distinctly controllable layer. This also allows rapid development among multiple team members and the ability to deconstruct and analyze the building piece by piece. Once the model is complete, it then becomes important to extract the data that it contains. One way to do this is to use something called a data-theater shown in Figure 17. The data-theater was introduced as a concept at the ACADIA (Association of Collegiate Schools of Architecture) convention (Fukai, 1996b). It is basically a computational “black box” that surrounds the model and acts as a graphical interface to a software engine. Clicking on the box initiates macros that deconstruct the model according to preset planes that surround and slice through its three-dimensional form. Images and diagrams can then be extracted from the model by a software engine to facilitate the actual construction project. This “engine” is a simple set of customized macros that does not involve high level programming skills. This is important because it will have to be “tuned” to each particular project. Another method is to use the model as a hypergraphic interface (Fukai, 1996b). Clicking on pieces represented in the interface zooms in on detailed “layers” of

52

D. Fukai

Figure 17.

A data-theater surrounds the model to annotate layout planes.

Figure 18. Two-dimensional construction drawings can be generated from the four-dimensional modeling process.

supporting visual information. For example, excavation, backfill, concrete, and reinforcing steel can again be represented in the model as subcontracts on different layers so that they can be displayed separately or in different combinations. The organization of these layers and the way they are “called” by the graphical links would allow users to dynamically deconstruct the model by moving toward ever more detailed representations of the construction. Both of these methods point to a graphical interface that leads to an infinite collection of dynamically generated “snapshots” of the pre-construction model.

4D CAD in construction communications 53

This means that there is no part of the building, before or after its construction, that will not have a graphical representation available for inclusion in daily or monthly reports, presentations to clients or subcontractors, websites, service manuals, and ongoing management and repair of the facility. In addition, many parts of the building will have detailed annotated two and three-dimensional layouts that can be used to direct the construction. These layouts are generated from the three-dimensional drawings (see Fig. 18).

CONCLUSION The challenge is to explore the modeling process in the context of an actual construction project in order to search for the revolutionary insight that must be part of what is truly the fourth dimension. Missing from a three-dimensional model is the communications that is embedded in the construction simulation. The perceptive shift is not in our view of the model, but in a sideways view at the communications associated with that model’s construction within its virtual environment. Extracting the graphical data associated with the flow of that communication provides the constructor with an archive of visual explanations that could thereby anticipate the interactions that will occur in the actual construction process. In this way, pre-construction becomes pre-communications, and the effort suggests a new graphical “intelligence” that might be added to support an industry of complex practices. In conclusion, if there is to be a radical shift in our world view as we move from three to four dimensions, perhaps it must occur in the kind of communications represented by the pre-construction modeling process rather than the model itself.

REFERENCES Aalami, F. & Fischer, M. 1998. Construction method models: the glue between design and construction. Proceedings of the 1998 international computing congress on computing in civil engineering: 376–378. Boston: ASCE. Abbott, E.A. 1952. Flatland, 6th edition. Dover Publications. Cardone, F., Francaviglia, M.& Mignani, R. 1999. Five dimensional relativity with energy as the extra dimension. General Relativity and Gravitation 31(7): 1049. Falcioni, J.G. 1999. Managing product life cycles. Mechanical Engineering, CIME, 121: 4 (editorial). Fukai, D. 1996a. A WORLD of data: an animated hypergraphic construction information system. Presentation to the Association for Computer Aided Design in Architecture, Tucson, AZ, October 1996. Fukai, D. 1996b. Real-world, real-time, real-fast: using a trainer in a computer mediated classroom. A presentation as a fellow to the Materials and Technology Institute of the Association of Collegiate Schools of Architecture, Berkeley, CA, 1996.

54

D. Fukai

Fukai, D. (unpublished paper). Current Research: Insitebuilders.com. M.E. Rinker School of Building Construction, College of Architecture, University of Florida. Hofstadter, D.R. 1980. Godel, Escher, Bach: An Eternal Golden Braid. Vintage Books. Jerrens, K.K. 1999. NIST’s support of rapid prototyping standards. IEEE Spectrum 36(2): 38 Johnson, D. 1999. Discuss and change models in real time. Design News May 3, 1999: 96–101. Kamarani, A.K. 1999. Direct engineering: toward intelligent manufacturing. Klumer Academic, Monograph. LaCourse, D. 1990. How solid modeling previews the future. Design News 46(10): May 10, 1990: 90–92. McKinney, K., Kim, J., Fischer, M. & Howard, C. 1999. Interactive 4D-CAD. Computing in Civil Engineering: 383–389. Potter, C.D. 1998. Process control CAD/CAM’s newest tool aims to oversee both the data and the methods used to design complex assemblies. Computer Graphics World 22(8): 69. Shah, J.J. & Wilson, P.R. 1988. Analysis of knowledge abstraction, representation and interaction requirements for computer aided engineering. Computers in Engineering, 1988 – Proceedings. Vaugn, F. 1996. three-dimensional and 4D CAD modeling on commercial design-build projects. Computer in Civil Engineering 1996: 390–396. Wei, D. & Gibson, K. 1998. Computer visualization: an integrated approach for interior design and architecture. McGraw-Hill. Wilson, J. 1997. AutoCAD: a visual approach. Autodisk Press. Wright, V.E. 1994. 4D CAD. Heating, Piping, Air Conditioning 56 (July 1984): 41–53. Yamaguchi, T. & Liu, H. 1998. Computational visualization of external and internal biological flows with fluid-wall interactions. Advances in Bioengineering 39: 127–128. American Society of Mechanical Engineers: Bioengineering Division.

FULLY INTEGRATED AND AUTOMATED PROJECT PROCESS (FIAPP) FOR THE PROJECT MANAGER AND EXECUTIVE F.H. (Bud) Griffis, Carrie S. Sturts Department of Civil Engineering and Engineering Mech, Columbia University, NY, USA

Abstract Fully integrated and automated project process (FIAPP) is an acronym suggested by the research committee of the Construction Industry Institute. A schedule linked to a threedimensional model (or 4D CAD) is a component of a larger FIAPP picture. This paper briefly describes 4D CAD in the context of the larger picture of FIAPP and the threedimensional computer model. The focus of our research has been on the use of threedimensional computer models for construction management. Furthermore, we focus on the industrial process or commercial power projects because they are routinely designed in three dimensions. This paper introduces some of the benefits of using three-dimensional models for construction and the conclusions developed in a three-year research project into FIAPP and 3D CAD that relate to 4D CAD development and usage. This discussion is illustrated with a case study project: the construction of an Air Separation Plant in Baytown, Texas. Finally, this paper discusses a possible course outline to teach project managers and project executives how to benefit from using FIAPP in the management of the construction project process. Unless project personnel actually have hands on use of the model, it loses its value as design is turned over to construction. Therefore the state of the art can only be advanced if project personnel feel comfortable operating in a FIAPP environment. Keywords: FIAPP, training, advantages, 3D CAD

BACKGROUND The process and power industries routinely use 3D CAD in design and construction; however, most building and heavy construction projects are still being designed and constructed using two dimensions. Given that contractors work from twodimensional drawings, unless the designers are proficient in three-dimensional design, there are probably few benefits to be gained from three dimensions for 55

56

F.H. Griffis & C.S. Sturts

relatively simple structures. The benefits of three-dimensional design in residential and commercial buildings have not been shown. On the other hand, the industrial process and commercial power sector of the architectural/engineering/ construction industry routinely use three-dimensional computer models for the design and construction of plants and facilities, and the benefits have been well documented (Griffis et al., 1995). Current research efforts are being made to integrate all aspects of the project process (preplanning, design, construction and start-up) using three-dimensional computer models and integrated databases. This paper will discuss 4D CAD and its role in the integration process and the functions of the integrated system in the construction process.

WHAT ARE THE BENEFITS OF USING THREE-DIMENSIONAL MODELS ON THE CONSTRUCTION SITE? The authors were involved with the Construction Industry Institute’s (CII) research in the use of three-dimensional computer models for construction management applications spanning from 1993 to 1995. The study consisted of three parts. First the researchers used questionnaires to investigate the perceived benefits and impediments to using three-dimensional models in the management of construction. Second, they performed statistical studies on 93 projects that used threedimensional models in the management of construction to varying degrees. Finally, the research team used a case study project to judge the reality of the statistics results. Some of the results are as follows: Most common usage • Checking clearances and access • Visualizing details from non-standard viewpoints • Using model as reference during project meetings Performing constructability reviews •

Greatest perceived impediments to the use of 3D in construction • Undetermined economic impacts • Inertia • Lack of trained people • Cost was perceived as an impediment only by non-users

Perceived benefits by users Reducing interference problems Assisting in visualization Reducing rework Improving engineering accuracy Improving jobsite communication

Differences between only 2D and “average” to “very good” use of 3D* • 5% reduction in cost growth • 4% reduction in schedule slip • 65% reduction in total rework

• • • • •

* Benefits were quantified by the statistical study (Griffis, 1988).

Fully integrated and automated project process 57

These benefits are impediments that may specifically apply to 4D CAD in varying magnitude; however further research should be conducted to isolate the benefits of 4D CAD for various project types and at different management levels. The case study project was used to actually perform cost estimates of the benefits as they occurred in the field. Those direct cost benefits exceeded those predicted by the statistical computer models. (Griffis et al., 1995).

WHAT IS FIAPP? We have come to the conclusion that the greatest benefits from the threedimensional computer model come from the integrated databases and not just from the three-dimensional computer model. Four-dimensional computer models are part of a larger picture. Researchers at CII have coined the term fully integrated and automated project process (FIAPP) to describe this bigger picture. FIAPP is not a software system per se, but an idea about the future computer data systems that will support a project from inception to start-up and beyond. FIAPP describes how information will flow automatically from one system to another, from one project participant to another, from owner to designer to fabricator to constructor. FIAPP and the relationship of the three-dimensional model to the other systems is still being debated. When we use the term FIAPP, we include all activities in the pre-project planning, the design, the procurement, the construction management, the start-up, and the operations and maintenance phases. In the ideal FIAPP (Fig. 1), all systems are integrated from payroll to job costing to scheduling to the design systems. Of most interest to construction managers

Project Mgt Systems

Design & Analysis Systems

3D Computer Model

Materials Mgt Systems

FIAPP F&A Systems

HR Systems

Figure 1.

Procurement Systems

The fully integrated and automated project process.

58

F.H. Griffis & C.S. Sturts

ENGINEERING DESIGN AND ANALYSIS

GRAPHICS 3D DATABASE

GEOMETRIC DATABASE

OTHER DATABASES

Procurement and Material Management

Project Management

Construction Field Operations

Figure 2. The three-dimensional computer model and its relationships with the design and construction process.

is a subsystem of FIAPP consisting of the three-dimensional model, the PM systems, the procurement systems and the material management systems. It is within these systems that we will consider 4D CAD. The three-dimensional computer model (Fig. 2) progresses from the design phase with output to the project management systems, construction field operations and the procurement and material management systems. These systems in turn provide feedback to the model and/or its associated databases. Many feel that the three-dimensional computer model is but one of the databases associated with FIAPP. Others feel it is the hub of the system through which the other databases are accessed. Figure 3 illustrates these two different integration schemes. For many projects, much of the design and procurement is initiated long before a detailed three-dimensional computer model is available. The design often starts with the process flow diagrams (PFDs) and the piping and instrumentation diagrams (P&IDs). The front end engineering design (FEED) generally consists of PFDs, preliminary P&IDs, multiple simulation cases and cost comparisons, detailed equipment data sheets, request for quotation responses, general equipment and piping layouts and early three-dimensional computer model reviews

Fully integrated and automated project process 59

DB 3

DB 1

3D Model

DB 2

3D Model

DB n

?

DB 1

DB n

DB 2

Figure 3.

Spoke concept versus the wheel concept of FIAPP.

Figure 4.

Piping planning.

DB 3

(Carell, 1999). Most of the piping, piping accessories, and major equipment items are developed in this stage and some of the procurement may be initiated (Fig. 4). This initial development supports the wheel model for FIAPP; however, there are those who feel that once an initial model is developed, it should be the center and organization format for the project data. FIAPP is still in development.

4D CAD AND FIAPP The following rendering is of the case study project (Fig. 5) upon which this paper is based. 4D CAD was not used in this case study project although it was tried.

60

F.H. Griffis & C.S. Sturts

Figure 5.

Baytown case study project.

There were numerous reasons that it was not used. This project was the project team’s first attempt to use three-dimensional computer models in the management of construction. The way in which the three-dimensional model was developed greatly inhibited the attempt to use 4D CAD. We found that 4D CAD was impossible to use with this model. The system was designed in Intergraph PDS and third party add-ins. The organization of the model was done so as to preclude separating the commodities and groupings into activities. First, the underground civil work was not modeled. (We have found this to be a shortcoming in most models and will recommend that civil work be modeled in the future.) The civil work could not be scheduled, however, from the model. Secondly, there is the issue of the plant design area relationship to the construction area. Plant design areas take form during front-end project planning. The areas are developed using associated equipment usually by the process section or by the process system. If areas are system based, they can overlap. Construction areas should use the same geographical boundaries as the plant design areas. Construction zones are used to identify physically defined subsets of design areas and are identified by design element attributes. Specific design element attributes can be used to identify the specific subcontractor work packages. Work packages are usually based on geographical boundaries and specify subcontractor data and system tie-in packages. Work is usually controlled by line list data and subcontract data by specific spool number. A back-to-front schedule establishes the design release by construction area or sub-area and is issued as a design area package (Hall, 1999). Finally,

Fully integrated and automated project process 61

the project team did not feel that the value of using 4D CAD on the project site would justify the effort of trying to make the model compatible with the 4D CAD software. The contractors required two-dimensional paper drawings. The work packages were put together from these drawings and scheduled by experienced engineers. The benefits of the 4D CAD could not be imagined. Based on this fact, we tried to anticipate what the benefits of using 4D CAD on a project this size could be if the model was developed in such a method as to facilitate its use. The following potential benefits were noted:

• If the model were developed early enough, a 4D CAD approach would have been of benefit to permit the Board of Directors to make financial decisions regarding the plant. • A 4D CAD approach could have been valuable tool for the marketing department in their effort to sell the product. • A 4D CAD analysis could have helped with the interface with the Exxon plant to which liquid oxygen was to be piped. • A 4D CAD modeling approach might have found some scheduling errors in the construction of the plant. However, there were no important scheduling delays encountered. Numerous conflicts were found and resolved in the field.

COURSE OUTLINE: USING FIAPP FOR PROJECT MANAGERS AND EXECUTIVES This course will acquaint the project manager or project executive with the use of three-dimensional computer models to enable an understanding of the evolution of the fully integrated and automated project process. The course will involve computer software. However, it will not be a programming course. It will require understanding of the software characteristics and the manipulation of the software. In this course, we will use mostly software marketed by Bentley Systems. This software is not the only available software that will do the job that we require and it may not be the absolute best software for what we are after. Nevertheless, it is easy to learn and robust enough to serve our purposes. The course is a one-semester course consisting of 45 contact hours. It uses a relatively simple (but not trivial) construction project as a semester project. Objectives of the course • To acquaint the project management professional or the project executive with the concept of FIAPP.

62

F.H. Griffis & C.S. Sturts

• To acquaint the project management professional or the project executive with the software in use, and to train him or her to have the facility to use the software in the management of construction. Session #1: Introduction • Discussion of the background of FIAPP. • Examples of some three-dimensional computer models. Session #2: Pre-project planning using 3D-FIAPP If a company is building a first of its kind facility, there will be no detailed threedimensional model in place during the pre-project planning phase. However, the state of the art is such that a conceptual model, populated by geometric objects, can be readily developed in the pre-project planning phase. This model can result in considerable payoff. As the model is developed, associated databases are populated. As business objectives are defined and market research and analyses are made, the preliminary model is populated with graphic and geometric data and external databases are defined and linked to the model. As the facility objectives and capacity demands are computed however, the model is developed further. Concepts of operation of the facility are adjusted by what-if trials with the model. The model helps address regulatory issues such as initial permitting, wetlands issues, and waste disposal issues. Depending on the project, the initial block model greatly enhances the public relations presentations and presentations to the funding approval body of the organization. While the process and facility planning is initially accomplished by line diagrams, the initial preliminary model quickly becomes a facilitator for equipment and vessel placement, transmission line routing, customer interaction, and transportation arrangement. As project execution begins, the preliminary model is used to develop the contracting strategy. As major items of equipment are modeled, fabrication contracts are drafted. Using the plan view of the model, the project geography is developed and contract procurement strategies are initiated. The project scope is set, initial discussions begin with potential contractor bidders and bid package scope development begins. The EPC contractor team uses the model to assure that each organization understands its scope and its interface with each of the other organizations. The model serves as a vehicle for discussions with environmental agencies, negotiations with offsite utility providers, and future operations and maintenance personnel. The owner and designers begin product development, and accumulate certification and testing procedures databases which they attach to the objects in the model. The state of the art with respect to hardware and software makes the development of a primitive block model exceptionally easy. Ideally, one starts with the same coordinate system as will be used in the design and construction of the project. All of the popular CAD software perform in true three-dimension space today. The initial project model is a relatively small model. The hardware systems can readily facilitate the model manipulation in real time. Projection systems that can handle

Fully integrated and automated project process 63

1280 ⫻ 1024 are currently on the market and available for project meetings and presentation and the technology will improve.

INTRODUCTION TO MICROSTATION: DEVELOP SMALL BLOCK MODEL IN MICROSTATION Session #3: Pre-project planning (continued) Introduction to PlantSpace Enterprise Navigator. Convert small block model from dgn file to jsm file. Use PlantSpace Enterprise Navigator to view the block model. Study features of navigator. Session #4: Introduction to databases Introduction to Microsoft Access. We propose to use Microsoft Access as the database driver since it is so readily available. There are other, more powerful programs such as Oracle that could also be used. In this session the concept of procurement databases is proposed. By the end of the preplanning phase of a project, the following databases will probably be initiated and population of data items begun:

• • • • • • • • • • • • •

major equipment lists, raw material vendors and other sources for product, bulk material vendors for construction product procurement, regulatory permit register, process preliminary design criteria, P&ID preliminaries, major equipment criteria, start-up procedure design criteria, labor with cost and productivity, quantities, cost estimate, schedule, work breakdown structure.

To begin learning Microsoft Access, a sample procurement table will be developed. The concept of object based programming is introduced. The access database will be linked to the Microstation by its Open DataBase Connectivity (ODBC) linking capabilities. Session #5: Introduction to project design using Microstation The object of this session is not to turn the student into a designer but to refresh the manager’s concept about how design in a three-dimensional world is accomplished. This is important because the naming conventions, the file designs, and other design organizational matters will affect the use of the tools on the construction project.

64

F.H. Griffis & C.S. Sturts

STRUCTURAL DATA Foundation Analysis & Design Structural Analysis Site Analysis

PROCESS DATA Steam list Process Summaries Process Analysis Process Optimization Process Flow Diagrams

Civil & Structural Engineering

Process Engineering 3D MODEL DATA Structural Steel Design Pipe Analysis Hanger Design Elasticity Analysis Seismic Analysis Isometrics Plant Layout HVAC Equipment Data Sheets

ELECTRICAL DATA Equipment Specs Load Tests Load & Fault Analysis Grounding Design Conduit & Cable Tray Design Substation Design Circuit Raceway Schedule

Figure 6.

Plant Design & Layout

ENGINEERING DESIGN & ANALYSIS

Mechanical Engineering

Control Systems

Electrical Engineering

GRAPHICS 3D DATABASE

GEOMETRIC DATABASE

P&ID DATA Equipment List Relief & Line Sizing Tie-in List P&ID Index

CONTROL DATA Logic Diagrams Loop Diagrams DCS/PLC Configuration Instrument Locations Instrumentation Design

OTHER DATABASES

Engineering design and analysis.

The design process varies from company to company and project to project. Considerable design usually takes place before the three-dimensional model is available. For instance, much of the equipment and many of the values and their attributes are developed directly from the piping and instrumentation diagram developments. These attributes must be entered as attributes to the threedimensional model as the model is being developed. Equipment fabrication orders may be placed before the model is developed. Figure 6 is an example of how various disciplines are coordinated within the design. Considerable thought must be given to the organization and design of the model and its associated databases early in the design stage. The current state of the art is such that there are no consistent data elements or object attributes that serve as standards within the industry. This is a gap in the state of the art that should be resolved through research in the future. As it now stands, each design organization designs its own data structure or uses one that is suggested by the design software. Figure 7 shows the milestones associated with the design process (CII, 11111 1997).

Fully integrated and automated project process 65

Detail Estimate Complete

Prelim Scope Def'n Complete

Design Development & Estimate Complete

Construction Documents Complete Detail Design Complete

Detail Integrated Project Schedule Complete

Figure 7.

Owner Review & Approval of Scope, Estimate, and DD Documents

Schematic Design Complete

Bid Documents Complete

Steps in the design process.

Session #6: Designing with Microstation (continued) Session #7: The application of three-dimensional computer models in procurement and materials management One of the most valuable benefits of the three-dimensional computer model in the management of construction comes from its use in procurement and materials management. As the design progresses, the process begins to populate the databases or object attributes. Procurement and materials management is the process of ensuring that the right materials are at the right place at the right time of a project. All activities necessary for the procurement; expediting, shipping, and receiving of all project materials, is included in this process. As the design objects are more refined, procurement can commence. Figure 8 shows some of the databases associated with the procurement and material management system. Many of these databases are being populated before the three-dimensional model is completed. For instance, piping and valve data will initially be developed from the P&IDs and the process flow diagrams. As design progresses, the objects are more and more defined. There are basically five types of procurement specified in a plant design. They are

• • • • •

bulk commodities, fabricated items, standard engineered equipment, specialized engineered equipment, services (general contractors and subcontractors).

In the course, each type of procurement is discussed in turn. Session #8: Applications of FIAPP in the construction phase Based on the research thus far, we are of the opinion that there is a large benefit to be accrued by having and using the three-dimensional computer model on

66

F.H. Griffis & C.S. Sturts

GRAPHICS 3D DATABASE

MATERIAL CONTROL DATA Bill of Material Items & Parts Material Packages MR's

MATERIAL SPEC. DATA Tag Numbers Stock Codes Specification Classes

GEOMETRIC DATABASE

OTHER DATABASES

INVENTORY DATA Stock Management Storage Receipts Issues

Material Requisitions

Material Takeoff Inventory Control Material Coding & Specs

Procurement & Material Management

Expiditing + Inspection

SUPPLIER DATA Bid Data Supplier Surveys Bid Tabulations

Supplier Evaluation

PURCHASING DATA Bidding & Sourcing

Subcontracts Purchase Orders Bid Tabulations Delivery Progress Expiditing

Figure 8. Procurement and material management.

the construction site. Previous research showed that cost growth and schedule growth could be improved and rework could be reduced by 65% if the threedimensional computer model was used on the construction site, not to its full extent, but with average use. This was a statistical result based on a sample of 93 projects. In the following paragraphs, we shall discuss how to make the maximum use of the model on the construction site using the current state of the art tools available. Figure 9 shows some of the databases and functions associated with construction and project management of an engineering, procurement and construction project. These functions will be discussed in relation to the detailed activities that take place on a construction site. Some of the functions generally associated with construction and project management are planning and project scoping, document control, cost estimating, cost control, productivity monitoring, scheduling, field material management, and construction field operations. As a point of reference,

Fully integrated and automated project process 67

MATERIAL DATA SCHEDULING DATA

Material Packaging Purchasing Expediting Inventory Conntrol Release to Field Material surplusing

Preliminary Schedule Detailed Schedule EPC Interface Schedule Construction Summary Schedule Engineering Summary Schedule

ESTIMATING DATA Order of Magnitude Estimate Preliminary Estimate Definitive Estimate

PERFORMANCE & PRODUCTIVITY DATA Contract Progress Committee Reports Labor Performance Cash Flow Quantity and Unit Rates

COST CONTROL DATA Operating Plant Site Cost Control Reports Operating Plant Work Order Control Reports Cost & Commitment Reports ACWP and BCWP Reports (by resource groups) Engineering Performance Reports Accounts Payable Payroll & Timekeeping

Figure 9.

Field Matl Mgt

Project Automation Plan Quality Assurance Plan Contracting Plan Project Execution Plan Design Scope Book

Scheduling

Cost Estimating

Productivity Monitoring

PLANNING DATA

Planning and Project Scoping

Construction and Project Management

Const Field Ops Document Control System

Cost Control

INSTALLATION DATA

DOCUMENT DATA

Visualization Rigging Installation of Piping Installation of Steel Installation of Equipment Welding Equipment Maintenance

Document Distribution ControlLog Initiate & Review Document Performance

GRAPHICS 3D DATABASE

GEOMETRIC DATABASE

OTHER DATABASES

Construction and project management.

we will use the case study project three-dimensional computer model as shown in Figure 10. Consider the following functions in this session preparation:

• • • • • • •

site mobilization (Fig. 11); mobilize facilities; provide construction utilities; establish safety and quality; obtain permits and licenses; submit project documents (Fig. 12); establish security (Fig. 13);

68

F.H. Griffis & C.S. Sturts

Figure 10.

Preliminary model.

Figure 11.

Baytown offices.

• • • • •

develop materials management plan; develop execution strategy; define training procedures; plan for major action items; develop work plan (Fig. 14);

Fully integrated and automated project process 69

Construction and Project Management

GRAPHICS 3D DATA

DOCUMENT DATA

Document Control System

GEOMETRIC DATABASE

Document Distribution Control Log Initiate & Review Document Performance

OTHER DATABASES

Figure 12. Document control.

Figure 13.

Security area in Baytown.

• execute labor management and construction (some execution management functions): – perform constructability reviews with design team, – define contract scope(s), – use as reference during project coordination meetings, – plan rigging or crane operations (Fig. 15), – check installation clearances and access,

70

F.H. Griffis & C.S. Sturts

ESTIMATING DATA

SCHEDULING DATA

Order of Magnitude Estimate Preliminary Estimate

Preliminary Schedule Detailed Schedule EPC Interface Schedule Construction Summary Schedule Engineering Summary Schedule

Definitive Estimate

Cost Estimating

Scheduling

Work Plan Development

Figure 14.

Work plan development.

Figure 15.

Rigging operations.

– – – – – – – – –

plan and sequence construction activities, plan survey control and construction layout, manage material layout yards or stockpiles, exchange information with vendors, locate installation points in field, track construction progress, estimate costs, visualize project details or design changes, record “as-builts” conditions,

Fully integrated and automated project process 71

PERFORMANCE & PRODUCTIVITY DATA Contract Progress Committee Reports Labor Performance Cash Flow Quantity and Unit Rates

Figure 16.

Productivity Monitoring

Monitor Schedule & Maintain Status

Productivity monitoring.

– – – –

train construction personnel, give safety briefings, plan temporary structures, turn-over design documents to owner; • monitor schedule and maintain status (Fig. 16); • establish design support; • issue progress reports: – submittals and document control management, – execute subcontractor management, – document QA/QC, – establish design support, – monitor schedule status/maintain schedule, – execute human resource management, – execute subcontract management and administration, – execute materials management and monitor status, – inspect test equipment, – issue progress reports, – monitor cost and budget status, – change management, – submittals and document control management, – process invoices, – return excess materials and receive credit. Session #9: Using the PlantSpace Enterprise Navigator on construction site Session #10: Using the PlantSpace Schedule Simulator on construction site Session #11: Using the PlantSpace Antimator on the construction site

72

F.H. Griffis & C.S. Sturts

Session #12: Integrating the process Session #13: Workshop to prepare individual projects Session #14: Course presentations Session #15: Course presentations

SUMMARY This paper was begun with the emphasis of showing the results of over a decade of research in the applications of three-dimensional computer models and their associated databases to the management of construction. We have shown that there is an average of 5% savings in cost growth, 4% savings in schedule growth, and 65% savings in rework. Through the paper’s development, several facts became obvious. First, in order that the tools be useful throughout the process, the model and associated databases must be used. In investigating many companies, the proponent for the tools were the information technology personnel and the design engineers that work with the systems all the time. Secondly, the management personnel generally used their computers for word processing and analytics, but did little with high level graphics engines. Therefore, the second part of this paper is an outline of a single semester course, consisting of 45 hours of classroom instruction with an objective to acquaint management personnel with the applications of project integration software. We found that with a little training, project personnel began to feel comfortable using the three-dimensional computer model navigation software. The proposed course outline makes an effort to development a comfort level among management personnel for integration software. If project management personnel use the tools in the pre-project planning, design, procurement and materials management, construction, start-up and commissioning phases of a project, unanticipated savings will occur.

ACKNOWLEDGMENTS The authors express appreciation to the National Science Foundation and the Construction Industry Institute for supporting the research leading to this paper. In addition, we appreciate the support of the construction industry personnel serving on two CII Research Teams, 106 (Chairman John Voeller of Black and Veatch) and 152 (Chairman John Berra of H.B. Zachary Inc.) The thirty team members

Fully integrated and automated project process 73

provided a great deal of the research associated with project. Brent Senseny of Air Products and Chemicals, Inc. provided liaison with the Baytown Plant case study project. Ron Zabilski, formerly of Stone and Webster Engineering Corporation was a co-principal investigator on the NSF project that concluded in 1989. Doctors Dan Hogan and Winqing Li used the projects mentioned in this paper as a partial fulfillment of the requirements for a Doctor of Philosophy degree at Columbia University.

REFERENCES Carell, R. 1999. Front end engineering design (FEED), CAD/CAE software strategies. Rebis Industrial Workgroup Software (unpublished paper). CII. 1997. Determining the impact of process of change on the EPC process. Construction Industry Institute Report IR125-2, July. Griffis, F.H. 1988. Benefits of using three-dimensional computer models in the management of construction. National Science Foundation Report. Griffis, F.H., Hogan, D. & Li, W. 1995. An analysis of the impacts of using three dimensional computer models in the management of construction. Construction Industry Institute. Research Report 106-11, September 1995. Hall, 1999. Personal correspondence and discussions, Bechtel Corporation, Inc.

NEW CONSTRUCTION MANAGEMENT PRACTICE BASED ON THE VIRTUAL REALITY TECHNOLOGY Jarkko Leinonen1, Kalle Kähkönen1, Tero Hemiö2, Arkady Retik3, Andrew Layden4 1

VTT Building and Transport, Finland Eurostepsys Ltd., Finland 3 School of the Built and Natural Environment, Glasgow Caledonian University, Glasgow, UK 4 Department of Civil Engineering, Division of Construction Management, University of Strathclyde, Glasgow, UK 2

Abstract This paper focuses on experiences of implementing 4D applications (3D building geometric data ⫹ time) to meet the needs of construction companies. The authors have been developing and experimenting with 4D applications based on virtual reality (VR) technology and its integration with other state of the art software. The objective of the paper is to provide understanding for balancing possibilities and challenges of this approach for construction planning and management. Several case studies with YIT Corporation, a Finnish construction company, have provided basis for many findings to be presented in the paper. This long-term co-operation has covered product modeling, software architecture design, web technology and, more recently, 4D together with VR. First the present problems originating from the current construction planning practice are discussed. This is followed by presenting the possibilities of construction management practice with the aid of 4D—what is needed and what can be achieved. The 4D approach has been demonstrated in the live building construction project, THK office building. The case study covers cost-benefit analysis of applying 4D for construction planning. Keywords: 4D, virtual reality, construction management, information technology, construction company

INTRODUCTION Building construction is about transforming the vision and views of a client into a building where the client’s needs and objectives are met. This process comprises 75

76

J. Leinonen et al.

many stages and phases, iterative activities, and can have large numbers of participants involved having massive information transfer and communication needs. Furthermore, the practical set up of the building construction processes can vary greatly from one project to another. A major reason for this is the technical differences between projects. Another main reason is the varying building construction skills, experience and knowledge of participants. This experience, skills and knowledge can vary greatly between different organizations involved and require special attention. It seems that in practice too many building construction processes are far from the most suitable processes. This results in clear and easily identifiable high costs, low productivity growth and poor quality, which, unfortunately, are internationally common features in building construction. Recent results from several studies show that deficient management and organization are main causes of these shortcomings (Koskela, 2000). Additionally, Koskela shows that the main sources for production defect costs are the design and engineering phases and the production management. Likewise, deficient production planning has proved to be one of the underlying reasons for the problems in the production management (Josephson & Hammarlund, 1996). Consequently, it often happens that the design phase in a building construction project is realized without sufficient constructability consideration. Also, it has been discovered that the methods for decision support are ad hoc and unsystematic. This is a root cause for many problems and disappointments in building construction projects (Laitinen, 1998). The common feature of the important findings presented above is that they all are related to building construction management and associated decision-making. In modern building construction, where power and responsibilities are shared between key partners, and, where many other stakeholders are involved, the operations management is a real challenge. The overall process from the early project start to hand over needs to be under improved control at all stages in order to maximize added value to client. This is principally a task for the construction management practice. Due to historical preoccupations regarding building construction planning and management, these processes have been very much project-realization-oriented. This means that practitioners tend to get involved in detailed planning and its decision-making in a too straightforward manner. However, in practice one can often encounter unclear or conflicting objectives, high levels of uncertainty relating to most estimates, communication problems between individuals, unrealistic opinions and a lack of creativeness, flexibility and consensus between various parties. New methods are needed particularly for reaching an improved level for building construction management and relating decision-making in an environment where numerous participants are working together (Turner, 1993; Barkley & Saylor, 1994). One of these new methods is the use of virtual reality (VR) technology for production planning and construction management. One of the benefits that can be gained from VR technology is the possibility for the 3D visualization of construction plans (Goldstein, 1995). These visualizations can

New construction management practice 77

enable an improved communication over the product and its construction processes (Ogata et al., 1998). It would be much easier to get all key individuals from partnering organizations involved in an improved way and, in particular, to take advantage of their experience and knowledge. The potential problems relating to the efficiency of the current design solution and construction plan can be more easily identified. More detailed benefits have been discussed by Alshawi (1996). VTT Building Technology has, together with YIT Corporation, been developing and implementing applications based on VR technology and 4D (3D ⫹ time) to improve construction management practice. Examples of the areas this long-term operation has covered are product modeling, software architecture design, web technology and 4D together with VR. The results from this co-operation form the basis for this paper. The paper presents different approaches for the implementation of the 4D model. In addition, it is demonstrated how 4D models based on the use of VR technology are created:

• Combining together product modeling technology, scheduling data and VR technology. • Having a specific building component library which is used step-by-step to build up a 4D model from drawings and scheduling data. • Additionally, the potential benefits of 4D models are analyzed using a live construction project as a test bench.

3D PRODUCT MODEL AS A STARTING POINT A building product model is an information base of a specific building (Björk, 1995). These product models can be used by different computer systems to create, edit, store, retrieve and check data about buildings. A leading principle, and promise, is to store all data relating to a specific building into its building product model. As a result the product model would be a common source for all project participants throughout the building life cycle to access building data (Fig. 1). Development of building product models is strongly facilitated by modern data modeling techniques and tools. A working building product model requires a product data model, which structures the information needed to describe a building. In a way, the product data model sets up a standard that forms a basis to integrate together various software tools used for different purposes, in different phases, and by different participants, during the building project life cycle. The building product model captures a lot of different kind of data from which the location and dimension data of each component are only a small portion. However, this data enables the graphical viewing of the resultant product model, or its parts, providing a significant means for communication and teamwork. The resultant product model

78

J. Leinonen et al.

Figure 1. Product model as a main source for accessing building data by various project participants.

works as a common platform for storing and accessing data for different purposes. Typical examples of various needs for using the data captured in the product model are quantity take-offs, cost analyses, scheduling, resource usage planning and procurement planning. Another important development principle is the minimization of data redundancy, i.e. the information is stored only once and the documentation and reports are produced from the product model when required. In this way the building product model captures the geometrical representation of the building items, e.g. superstructure, slabs, rooms, walls and doors. This data together with project scheduling data form the starting point for 4D applications (Fig. 2). The latest versions of Industry Foundation Classes (IFC) standardize the data structure of the geometrical representation of the building together with scheduling data. Product modeling technology can contribute strongly towards advanced partnering and co-operation in construction. Thus, in YIT Corporation, the product modeling technique is being used as a core technology for providing the basic enabling solution for supporting, forming and managing temporary networks of companies and their resources working in a building construction project. The application of this technology is the CoVe model builder (Laitinen, 1998). The CoVe model builder enables the use of heterogeneous data from various project partners for setting up a product model. The model builder CoVe was first developed for the modeling of residential buildings (Fig. 3).

New construction management practice 79

Figure 2. Using building product model as a means to integrate together various software and contributions by different partners.

Figure 3. CoVe model builder: interactive views over product model hierarchy and the resultant building under design activity.

80

J. Leinonen et al.

4D MODEL

3D + Time

3D + Time

"Automation approach"

"Linking approach"

"Lego approach"

Combination of 3D and schedule

Schedule and construction site model preparation using virtual reality model

Automated generation of 4D Model Interpretation of 3D Logic

Figure 4.

ProMo Te

3D + Time

THK Case

Different main approaches for the development of 4D applications.

INTEGRATING 3D AND TIME Integration of the geometrical representation of the building together with scheduling data has been a topic for many research and development efforts. Different approaches for this integration are as follows (Fig. 4):

• automated generation of 4D models; • linking geometrical 3D model with schedule; • “Lego approach”, i.e. developing a 4D model from building parts library. The automated generation of 4D models refers to computer applications where a reasoning engine, or engines together with construction operations knowledge, are able to interpret 3D geometric data and produce process plan showing how the building can be constructed. Examples of the results from these type of research and development efforts can be found in Aalami & Fischer (1998), Froese et al. (1997), Levitt et al. (1988) and Zozaya-Gorostiza et al. (1989). The second approach for the development of 4D models is the “Linking approach”. In this approach the schedule data, e.g. start time, finish time, actual start and actual finish, are linked to the building components in question. This approach requires that the schedule data can be assigned to the appropriate building components. Typically, an activity in a schedule can cover a piece of work relating to a set of building components. For example, the scope of the activity “Install façade elements for block A” is self-explaining and in a 4D model this activity needs to be linked to all façade elements in the block in question. Traditionally, and still rather often, CAD systems are used in building projects as a drawing tool. This

New construction management practice 81

means that designers are not defining objects, such as slab elements, beams, doors and windows, but they are merely producing a graphical representation of the building using primitive geometric symbols (lines, symbols). This does not provide possibilities for handling building component concepts, which is a burden for the development of 4D models. In the following section a research and development effort combining 3D product model and schedule data is presented. The resultant ProMoTe tool demonstrates how a 4D model can be used interactively to visualize data captured in a building product model. The third approach for the development of 4D models is the “Lego approach”. This approach relies on the interactive study and definition of the building construction model. The user takes advantage of building parts and resource libraries. These libraries are used for setting up a VR model of the planned building process. In this paper a THK Project case study is presented where this approach was used for the development of a 4D model.

ProMoTe: LINKING GEOMETRICAL 3D MODEL WITH SCHEDULE ProMoTe is a product model browser for accessing product model over the Internet (Hemiö & Hannus, 1999). It provides an effective way to improve the use of product data technology. ProMoTe retrieves information from various sources and adds it to the product data model. Product data technology (ISO 10303-1 IS, 1994) is proliferating gradually in the facility management and construction industry. The emerging IFC product data model standard (IAI, 1997) driven by International Alliance for Interoperability (IAI) is gaining wide support. The belief that product models are the foundation for information sharing in the future is becoming generally accepted. Data transfer in the future will be based more on sharing than sending (ISO 10303-11 IS, 1994; ISO 10303-21 IS, 1994). The main interest of this paper is the creation of VR models from product data. Creating virtual reality models with ProMoTe The user of ProMoTe can create VR models from the whole building model or from a part of it. VR models can contain references to other VR models, which enable the creation of hierarchical sets of models. These models can be created by request so that they are always up to date. If the structure of the model supports content hierarchy specifications, it enables the user to specify different kind of hierarchical VR models for different kinds of purposes in the project. Scheduling data is one essential part of the product model data that can be visualized using VR models. ProMoTe combines a 3D model of a building and scheduling data originating from project scheduling tool (Microsoft Project) into a 4D-VRML model (Fig. 5; VRML97). ProMoTe retrieves the scheduling data as an exported file from the scheduling tool and adds this to the original product data model.

Figure 5. Linking schedule and 3D model using ProMoTe.

82 J. Leinonen et al.

New construction management practice 83

Changes to the schedule are currently done within the scheduling tool. However, at present an effort is under way for enabling two-way data traffic. This is supposed to result in a system where the required work breakdown structure (WBS) can be created in the 3D-VRML model, i.e. a user can group related building parts to form schedule activities. Furthermore, changes to the schedule can be done either in the 4D-VRML model or in the MS Project schedule. Using Virtual Reality Browser as an interface to product data Virtual Reality Browser (VRB) can be used as an interface to product data, enabling selection of objects and triggering specified methods. Information about the selected object, the values of attributes defined for this specific object, can be browsed and the documents linked to selected object can be studied. Objects can be hidden from the VR model and their properties can be changed. For example, in visualization of construction schedule object hiding can be used to simulate construction of the building day by day. Changing the color of the object can be used for highlighting objects, which are in a conflict, or to visualize changes in the model by showing different versions of objects in different colors.

PROVIDED FUNCTIONALITY Scheduling Changing the number of items to be shown can be used to animate the construction process. The user can set the date when the status of the construction process is shown. A slide bar is used to visualize the construction of the building from day one to the end of the project. This is an efficient technique to examine the construction process and especially the space allocations during the process. The VR model can also be used for grouping the objects for scheduling. An example of schedule visualization is shown in Figure 6. Browsing the data relating to a selected object The values of attributes of selected instance can be browsed (Fig. 7). This enables an easy validation of the properties of the selected object. Properties are typically the material, size and location of the object. Documents (files) can be linked to selected instances so that all users can immediately see what additional information is available for this object (Fig. 8). These documents can be anything from the architectural specifications of a room to the maintenance log of an air conditioner. Version comparison Version comparison is an efficient way to compare two proposals or to view what has been changed in the model. Models to be compared are displayed in the same

Figure 6.

Construction schedule simulation.

84 J. Leinonen et al.

New construction management practice 85

Figure 7. Attribute value browser window for looking through the values of attributes recursively.

Figure 8.

Linking desired documents to a certain object in the VR model.

86

J. Leinonen et al.

VR model and the user can select the model whose geometry is to be shown. This enables switching between the different representations of the selected instance as well as changing between the whole models. It is also possible to show multiple representations of the same object at the same time. This is very convenient when the user wants to compare the alternatives. Experiences from case studies Users of ProMoTe have felt that the possibility to visualize and analyze a building and its construction in a platform independent of a particular CAD program can enhance significantly project planning and execution. VR models are a userfriendly way to access product data. The possibility to generate models by demand from the server using only selected objects makes it efficient also when the data is accessed through the web. Possibility to define hierarchical VR models using the content hierarchy specified by the user into the product model itself, makes it possible to generate different kind of hierarchical models for different purposes. Ability to simulate the construction of the building makes it easier to check the status of the work in a specified point of time. Connection to document management through the objects visualized in the model provides easy access to all documents liked to the model.

4D MODEL FROM BUILDING COMPONENTS AND RESOURCE LIBRARY Visualizing construction process Overview The intention in this effort was to show the construction of a building visualizing work progress on a day-to-day basis. The VR based model was devised and developed well in advance to the project completion. It employs a low detail simulation of main superstructure activities involving six basic elements: columns, beams, façade elements, concrete staircases, partitions and hollow slabs. Case study The model and simulation are based upon a real construction project in Helsinki. The operational schedules, element supply table, drawings and photographs from the site were used to construct the VR model. The simulation also incorporates the surrounding building site, including familiar site equipment. The case study addressed the construction process of the THK office building located close to downtown Helsinki. The premises are owned by The Pension Foundation of the Finnish Broadcasting Company and the main part of the building is rented by the Telecommunication Administration Center of Finland (THK). The

New construction management practice 87

Figure 9. Office building in Helsinki for which the VR process simulation model was developed.

YIT Corporation was the property developer and the main contractor. The total floor area of the building is 13,000 m2 and gross volume is 45,000 m3. The total time required for the completion of the project was 17 months including the site activities, which together with excavation took 14 months (Fig. 9).

METHODOLOGY The methodology has been built based on previous experience of Virtual Construction Simulation Research Group (VCSRG) at Strathclyde (Retik & Hay, 1994; Retik, 1997; Retik & Shapira, 1999). Though several projects had already been carried out by the Group, the intention of this project was not only to test existing methodology by applying it in the real environment, but also to evaluate, with the practioners involved, the benefits of the approach and technology to construction companies. The visualization was carried out in the following stages: Stage1: To create the VR based building model in its entirety and surrounding environment. This is the most time-consuming part of the visualization. However, the library of virtual elements and virtual plant available from previous projects facilitated and shortened the model development. Stage 2: To determine the work progress, i.e. how much of the building could be constructed each day. This was calculated from the time schedule produced by the project manager. Consequently, the building was then divided into groups. Each type of element was associated with a separate group. The amount of elements within a group is dependent on the above calculation. Stage 3: Once the above calculation and set up were completed, the program that controls the simulation was created. At the start of the simulation, the program

88

J. Leinonen et al.

Figure 10.

Division of stories to six sections (a view from the top).

renders the entire building invisible. The program then controls what groups are visible on a specified day. The building is constructed on a story-by-story basis in the technological sequence. Columns and the concrete staircases are erected first, followed by beams and then by hollow slabs. The façade elements are assembled last. This activity is performed in the bottom–top direction.

DEVELOPING THE MODEL Dividing the building The first six stories of the building are identical, so they are divided in an identical manner. Each floor is divided into six sections. This was determined by the calculation of the daily rate of construction for the building. The arrangement is shown below (see Fig. 10). The top three stories each have a different layout. The rate of construction is different for each story. This means that they have a different number of groups, but the principle remains the same.

CREATING THE ELEMENT GROUPS Not all elements are placed on the same day at a particular part of the building. This means that each type of element must be contained in a separate group. Figure 11 shows the first section of the ground story—columns, beams, staircase and hollow slabs. Note that the columns extend for three stories. Each type of element is contained in its own group, with the exception of the staircase, which is a single entity.

New construction management practice 89

Figure 11.

Section 1A of the building.

Figure 12.

Exploded view of Section 1A.

An exploded view of the same section is shown in Figure 12. The numbering convention is also shown. The group is split into three properties: the type of element, the story, and the section of the story. The area shown is on the first story, and in Section “A”. Note that the group for the hollow slabs sits on top of the other groups. However, it will occupy the same space as the groups for the first section of the second story. Each section of every story shares this arrangement of groups. In general, the groups for the façade elements do not occupy the same space as the other groups. They are normally placed along the sides of the other groups. However, this is not the case for the first section of the building (Section “A”). This is because that area of the building has an irregular shape. It is best to ensure that

90

J. Leinonen et al.

different groups do not intersect one another, so a special grouping arrangement was created. The top three stories also contain some unique grouping arrangements.

Creating the elements The simulation shows only a few of the different types of elements found in buildings. For the sake of graphical simplicity, the majority of elements shown are merely simple cuboids: beams, columns, hollow slabs and façade elements. These elements make up the bulk of the building. They can be modeled accurately, but the simulation would run much slower due to the increased amount of facets on screen. The dimensions of the elements were taken from the plans. Some liberties, however, were taken to increase the modularity of the building, and hence decrease the complexity of it. The first elements to be placed in the world were the columns. The first column to be placed formed the origin of the building. This is at the bottom, left corner of Figure 10. The beams were simply placed between the columns. The hollow slabs were given the same thickness of 200 mm. These elements were initially modeled as individual slabs, with the dimensions taken from the plans. Yet, it soon became apparent that the simulation would run much faster if the slabs were simplified. Thus, they were changed into long strips of slabs. These could not be allowed to share the same space as the columns, so shorter strips were used to fill in the gaps. The façade elements were also simplified for the simulation. They are not made of individual panels, as indicated by the plans. Larger panels replace many small ones. The thickness of the panels was set at 300 mm. The concrete staircases started off as four cuboids grouped together as a square. A group enclosed these objects, and they formed a wall group. This was recently changed to a dedicated shape, created with the shape editor. This was done to decrease the number of objects in the world, and hence speed up the simulation. Once the first story was completed, it was then taken apart. This was in order to create the groups required for each element. Each type of element was temporarily displaced by a set distance from the building. A group was placed around them, and then the group containing the elements was placed back into the correct position. Figure 13 shows how the columns were temporarily moved, so that their groups could be created. The other element groups have already been created and moved back to the correct location in the world. The groups are then checked for any improper child/parent adoptions (a hierarchy should be defined between different objects). Sometimes, when a group is moved back to its correct position, it adopts one of the other groups. The groups must be completely orphaned. An element group must only contain the appropriate type of element. Once the first story was checked, it was simply cloned and the next story was placed above the current one. The groups in the new story were renamed to follow the numbering convention previously explained. The first six stories were created

New construction management practice 91

Figure 13.

Creating column groups.

this way. The top three stories were also created in a similar way, but with different numbers and sizes of groups.

THE SIMULATION PROGRAM The code for the simulation can be classified into three different types: the main code, the icon code and background code. They are all interconnected, by passing counter values to each other. The main code controls the simulation of the building. It decides what sections of the building to display on a given day. The main code also decides what icons and instruments are enabled on the screen display. The code is attached to the “Anchor” object. This is disguised as a portacabin in the VR world (from viewpoint 4, it is visible in the bottom right corner of the screen). The main code is by far the largest code in the simulation. The icon code is used to pass control data to the main program. It is also used to control some of the vehicles in the world, both directly and indirectly. Many of the icons are unavailable at certain times during the simulation. Their availability is controlled by the main code. The icons process the SCL User Functions. The background code is attached to various objects in the world. Some of the objects, like parts of the tower crane, communicate with the icon code and the main code. The cars that travel on the roads around the world have their own predetermined paths. These are controlled by their own code. They have no connection to the icon code, or the main code.

92

J. Leinonen et al.

Figure 14.

Time-based simulation.

USING THE MODEL Once the simulation scenario is built in, the construction process can be simulated day-by-day (see Fig. 14) by pressing the forward or backward arrow icons on the screen (see left corner, first image, of Fig. 14). The user can stop simulation anywhere or select to “jump” to any day he/she wants to explore the project at any particular time. The exploration can be done interactively by walking through or by changing views (see Fig. 15).

New construction management practice 93

Figure 15.

Exploring the project.

FUTURE DEVELOPMENTS This application has proved to be an excellent communication means for different purposes, e.g. to run project demonstrations to authorities or other important interest groups, and to run planning meetings participated by different subcontractors. There is still much scope for refining this model. There should be an increase in the amount of activities shown, e.g. internal walls, roofing, windows and the external glass canopies. Future work could focus on making the building site more active and additional plant machinery could work away in the background. This has already been experimented on an another project. On the YIT Project, the

94

J. Leinonen et al.

Figure 16.

Monitoring the project (superimposed manually).

plant machinery was made controllable rather than self-automated. Yet, it should be possible to combine the two ideas, so that the machines will be computer controlled, but can be taken over by the user at any time. The materials placed on the building site can also be simulated accurately, if storage area data was made available as a function of time. Materials could be delivered to a certain location, on a certain day, and then disappear as the building takes shape. It is also possible to use the model during the construction stage for progress monitoring purposes. If remote access to the site is available (as described and demonstrated in Retik et al., 1997), real as-built pictures from site could be superimposed with VR as-planned pictures as demonstrated in Figure 16, so the project progress could be judged visually.

BENEFIT ANALYSIS Construction production planning aims to find out the means to achieve and maintain the most efficient site production. The site engineer needs to consider various options and choices, and do trade-off analyses between contradictory targets like keeping minimum buffer between jobs and allocating same time and adequate space for each work. 4D models can help the decision-making by offering support to production planning challenges as explained in Table 1. The results presented in Table 1 are based on site managers’ and engineers’ estimates. In this survey the site managers and engineers from the THK case project

New construction management practice 95 Table 1.

Usage and benefits of 4D modeling in building construction.

Challenges/questions

4D model usage

Results

Are the schedules realistic and feasible?

Easy to explain to architect, structural and service engineers when the designs and drawings are needed. Easy to show to decisions makers the dead-lines of decisions concerning design choices. Easy to show around the (virtual) real premises to customers earlier. Customers can start designing the interiors sooner. Bottlenecks are easier to notice in advance. More efficient resource management. Dangerous places easier to find out. Better-organized site. Tower crane size is based on the actual loads and locations of the loads. Taking advantage of novel technology. Right locations for storage areas easier to detect. Material flows can be analyzed in VR.

Efficient procurement with complete designs and drawings. Less waiting in the site. Materials delivered JIT.

How could I more effectively market my project?

How should I allocate resources and how can I avoid production bottlenecks? Have the safety factors on site considered properly? What are the appropriate tower crane size, location and capacity? How can I brighten up the image of my company? How can I more efficiently analyze the logistics of the site?

Premises rented earlier. Less rebuilding due to changes.

Less reworking. Less rush at the end of the project. Less unnecessary resources. Safer site. Better working conditions. More satisfied workers. Right tower crane size, location and capacity. Image of the front-line company. Efficient material handling. Field factories located correctly. Storage places located correctly.

were familiar with the 4D model that was developed for their project. Based on that knowledge they estimated the possibilities and potential benefits when a fully operational 4D system would be used in a project comparable to THK building project. Furthermore, this survey covered the magnitudes of the potential benefits. The potential benefits were estimated using ranges (minimum, most likely, maximum) and possible correlation between benefit items identified. After this, the overall benefits were analyzed using @RISK software package. Figure 17 and Table 2 present the summary of the results of the benefit analysis. The results demonstrate that one can expect considerable advantages from the use of operational 4D system in building construction. It seems that the biggest potential for benefits lies in the reliability of the schedule, bottleneck identification and resource management. The results are in line with Akinci et al. (1997).

96

J. Leinonen et al.

Figure 17.

Benefit sum distribution.

Table 2. Variances of benefit items. Benefit

Minimum

Most likely

Maximum

Reliability of the schedule Marketing and selling the space in the building Bottlenecks and resource management Tower crane selection Logistics Safety in the site

$14,492 $7,429 $19,343 $96 $4,840 $2,558 $70,342

$27,205 $11,111 $41,688 $3,111 $10,556 $5,290 $98,961

$41,810 $15,619 $68,192 $6,851 $17,193 $8,626 $132,758

DISCUSSION The main reason for many severe problems in the construction projects originate from the individual’s misconception of reality during project planning. Often two-dimensional drawings are the only main source of background material available for planning purposes. Yet, particularly in the case of multi-story buildings, it is essential to possess an ability to visualize in 3D in order to prepare successfully the necessary plans. The results gained propose that with the help of VR technology one can reach more advanced and concrete construction management within the early stages of the construction project. It is likely that this can contribute towards identifying and solving any possible problems before the actual building construction. VR technology is developing fast and continuously all the time. There are expectations that the next “killer” application (like word processors) of information

New construction management practice 97

Figure 18. Augmented reality demonstration. A telepresence video stream from construction site and VR image combined (project in September 1997).

technology will be based on the VR technology. Progressive companies can already start to apply the current VR technology and gain advantages from it. However, widespread applications are set to emerge in the near future. Augmented reality Augmented reality is combination of a 4D model and a telepresence stream. Telepresence is a technology that provides a possibility to feel present in a remote location, usually with the help of a digital video camera. Applying augmented reality, a 4D model and telepresence image can be viewed together and synchronized in a way that the user can easily compare the planned schedule and actual performance (Retik et al., 1997). In addition, the 4D model could be used as a pointing device for telepresence equipment. To receive additional information in some detail on site the user only needs to point the place on a 4D model. The telepresence equipment would then provide the appropriate video stream from site or via model details where further information can be accessed, i.e. some documents linked to the details, see Figure 18.

CONCLUSIONS The main efforts for integrating 3D and time can be classified according to three different approaches. These are automated generation of 4D models, linking a 3D model with schedule and the “Lego approach”. The authors’ main conclusions from

98

J. Leinonen et al.

the case studies on linking 3D model with schedule and applying the “Lego approach” are as follows:

• Product data access using VR models can cover both viewing and product data editing. This can considerably improve the possibilities of construction practitioners to take advantage of modern information technology. • Integration of 4D scheduling with the companies’ working practice enables new improved planning processes having clear potential for significant benefits. The benefits are mainly due to improved communication making it possible to get other necessary people, e.g. subcontractors, material suppliers and representatives of client, involved in schedule preparation and analysis. It is obvious that first widespread 4D applications in construction shall cover linking the 3D product model with scheduling data and possibilities for viewing the planned construction process. For reaching this the research and development efforts should be more focused on case studies and piloting where construction practitioners try to apply the prototype tools by themselves. Detailed virtual construction sites are an interesting research topic from which some very useful new planning and analysis techniques can arise to be used in practice, e.g. machinery selection and usage planning, site logistics planning and analysis of hazardous situations.

REFERENCES Aalami, F. & Fischer, M. 1998. Joint product and process model elaboration based on construction method models. In B.-C. Björk & A. Jädbeck (eds), The life-cycle of IT innovations in construction—Technology transfer from research to practice. Proceedings of CIB78 conference, 5–8 June, Stockholm. Sweden: Royal Institute of Technology. Akinci, B., Staub, A. & Fischer, M. 1997. Productivity and cost analysis based on a 4D model. In R. Drogemuller (ed.), Information technology support for construction process reengineering, CIB Publication 208. Australia: James Cook University of North Queensland. Alshawi, M. 1996. Virtual reality: future implication on construction. Proceeding of the second international conference in civil engineering on computer applications research and practice, Vol. 2, Bahrain, April: 789–795. Barkley, B.T. & Saylor, J.H. 1994. Customer driven project management. USA: McGrowHill Inc. Björk, B.-C. 1995. Requirements and information structures for building product data models, VTT Publications 245. Espoo: Technical Research Centre of Finland. Froese, T., Rankin, J. & Yu, K. 1997. Project management application models and computer-assisted construction planning in total project systems. International Journal of Construction Information Technology 5(1): 39–62. Goldstein, H. 1995. Is virtual reality for real? Civil Engineering June 1995: 45–48. IAI. 1997. IFC object model for AEC projects. IFC Release 1.5 Model Reference Documentation, Final Version. Washington, DC: IAI Publ.

New construction management practice 99 ISO 10303-1 IS. 1994. Product data representation and exchange—Part 1: Overview and fundamental principles. ISO TC 184/SC4, Geneva. ISO 10303-11 IS. 1994. Product data representation and exchange—Part 11: Description methods: The EXPRESS language reference manual. ISO TC 184/SC4, Geneva. ISO 10303-21 IS. 1994. Product data representation and exchange—Part 21: Implementation methods: Clear text encoding of the exchange structure. ISO TC 184/SC4, Geneva. Josephson, P. & Hammarlund, Y. 1996. Kvalitetsfelkostnader på 90-talet—en studie av sju byggnadsproject (Quality defect costs during ninetees). Del 1, Report 49, Chalmers University of Technology, Gothenburg, Sweden. Koskela, L. 2000. En exploration towards a production theory and its application to construction. Espoo, Finland: VTT Publications. Laitinen, J. 1998. Model based construction process management. Ph.D. Dissertation, Royal Institute of Technology, Stockholm, Sweden. Levitt, R.E., Kartam, N.A. & Kunz, J.C. 1988. Artificial intelligence techniques for generating construction project plans. Journal of Construction Engineering and Management 114(3): 329–343. Ogata, S., Kobayashi, I. & Fukuchi, Y. 1998. Application of virtual model to achieve consensus for construction project. Proceedings of the first international conference on new information technologies for decision making in civil engineering. Ecole de technologie supérieve, Université du Québec, Montréal, Canada: 1217–1226. Retik, A. & Hay, R. 1994. Visual simulation using VR. Proceedings of the 10th ARCOM conference, Loughborough, September: 537–546. Retik, A. 1997. Planning and monitoring of construction projects using virtual reality. Journal of Project Management (APMF) 3: 28–32. Retik, A., Clark, N., Fryer, R., Hardiman, R., McGregor, G., Mair, G., Retik, N. & Revie, K. 1997. Mobile hybrid virtual reality and telepresence for planning and monitoring of engineering projects. Proceedings of the 4th UK virtual reality interest group conference, Brunel University: 80–89. Retik, A. & Shapira, A. 1999. VR based planning of construction site activities. Automation in Construction: 671–680. Turner, R.J. 1993. The handbook of project-based management. McGraw-Hill Book C. VRML97. 1997. The virtual reality modeling language. ISO/IEC 14772-1:1997. http:// www.vrml.org/technicalinfo/specifications/vrml97/ Zozaya-Gorostiza, C., Hendrickson, C. & Rehak, D.R. 1989. Knowledge-based process planning for construction and manufacturing. USA: Academic Press Inc.

FURTHER READINGS APM. 1996. Body of knowledge. United Kingdom: Association for Project Management. Hemiö, T. & Hannus, M. 1999. Implementing browsing tool for EXPRESS schemata and STEP data. Proceedings of product data technology Europe 1999, Stavanger, Norway. http://cic.vtt.fi/rtetjh/Promote/ (papers).

4D CAD AND DYNAMIC RESOURCE PLANNING FOR SUBCONTRACTORS: CASE STUDY AND ISSUES William J. O’Brien M. E. Rinker, Sr. School of Building Construction, University of Florida, Gainesville, FL, USA

Abstract Even in a world with widespread use of 4D CAD, changes in schedule and scope will remain a salient feature of construction projects. This paper discusses an extension of 4D CAD techniques to improve subcontractor cost and resource planning in a dynamic environment. As a basis for multi-project development of 4D CAD, this paper presents a detailed case study of a subcontractor’s resource management challenges, and from this develops a conceptual model for multi-project management at the subcontractor level. This empirical and theoretical groundwork supports a critique of current 4D CAD tools that are limited to a single project perspective. 4D tools must be extended to integrate detailed planning on individual projects into the multi-project resource allocation process. An extension to a multi-project environment will allow subcontractors to better plan for changes in schedule and scope as well as accommodate them when they do occur. The proposed extensions have additional benefits. They allow detailed cost accounting of the ramifications of changes and hence improve the ability to negotiate compensation for those changes. More broadly, they allow evaluation of project schedules from the perspective of subcontractor ability to adjust to changes. This promises applications of 4D CAD that promote both efficiency and flexibility. Keywords: 4D CAD, subcontractors, construction costs, resource planning, cost planning

INTRODUCTION It is no great insight to note that how a subcontractor’s management operates its resources is a key factor in firm success. Nor is it a great insight to note that subcontractors work on multiple projects concurrently. Yet relatively little literature 101

102 W.J. O’Brien

exists to help subcontractors manage their resources beyond the productivity improvement literature for individual activities. There is a complete lack of tools that relate a subcontractor’s resource allocation across projects to cost (O’Brien & Fischer, 2000). Costing methods follow the assumptions of network scheduling methods, focusing on costs of individual activities and of single projects. No allowance is made for interaction between projects. Yet subcontractors emphatically do manage their resources from a multi-project perspective, taking a variety of actions to optimize resource allocation across projects in the face of changing conditions. Costs to subcontractors in a multi-project context occur because of changes in project schedule and scope. If changes did not occur, subcontractors and suppliers could predictably allocate their capacity (resources) to projects and they would not suffer any capacity related costs. (Further, bidding would ensure an efficient allocation of capacity across firms (see O’Brien et al., 1995 for further discussion)). However, changes in schedule are a common occurrence on construction projects. The causes of such changes are numerous and well cataloged in the construction literature: weather, owner directed changes, problems with permitting, unexpected soil conditions, materials delays, accelerations, rework that affects schedule, coordination difficulties, etc. In one sense, existing 3D and 4D CAD tools are an attempt to control the incidence of changes on construction sites. 3D/4D tools improve designs, reduce the incidence of time and space conflicts on construction sites, improve the materials flow, and improve schedule reliability. Thus 3D/4D tools may reduce the number of changes on project sites and therefore improve subcontractors’ ability to plan their resource allocation across projects. But it is the nature of projects to be dynamic; while generating one set of improved capabilities, 3D/4D tools do not remove the changes in project schedule and scope from influences such as weather. Moreover, the capability of 3D/4D tools in design may make owners more rather then less likely to initiate changes, particularly in a business environment that demands speed and flexibility. Changes to projects are given, and thus subcontractors will continue to face multi-project resource management challenges. Unfortunately, existing 3D/4D tools lack a multi-project perspective, having been developed in a single project context. These tools do not aid a subcontractor in its resource management and may in some cases hinder it. For example, many 4D models make assumptions that resources are fixed or that there is costless flexibility in assigning resources. Other 4D models that automate schedules assume that methods direct resources, and assign subcontractor resources based on methods and schedule needs. Yet for many subcontractors, it is the availability of resources that influences the choice of methods and staffing needs on a job site. Thus the real-world considerations of subcontractors may be in direct conflict with the virtual dictates of a 4D model. This paper discusses extensions of 3D/4D modeling methods to subcontractors’ multi-project resource management needs. A case study of the progressive

4D CAD and dynamic resource planning for subcontractors 103

subcontractor Pacific Contracting provides details of the constraints and considerations a subcontractor makes when making resource allocation decisions in a multiproject environment. From the case findings, a conceptual basis for multi-project resource management and costing is presented, allowing definition of several questions pertinent to subcontractor resource management. These provide a foundation for examination of current capabilities in 3D/4D CAD and discussion of extensions. In particular, there needs to be development of 4D tools that automatically assess the affects of resource re-allocations on project performance and ripple effects across projects.

PACIFIC CONTRACTING PART ONE: RESOURCE MANAGEMENT PROBLEMS The Pacific Contracting case study presented in this section is an intimate description of how a subcontractor manages its business, particularly with regard to management policies in relation to site conditions and resource allocation. Such a view of subcontractors is largely absent from the construction literature. One explanation for this lack of literature about subcontractors is the arm’s length relationships that exist between subcontractors and general contractors. Subcontractors are there to provide a contractual service. Hence, the extant literature on procurement and contracting serves as a substitute for a more specific literature about subcontractors. Discussion of the relative merits of contracts and incentives can be found in Abu-Hijleh & Ibbs (1989), Ashley & Workman (1986), Griffis & Butler (1988), Stukhart (1984) and Uher (1991). This literature does provide a useful way to structure and understand relationships and, of course, contracts are a necessary aspect of business practice. However, none of this literature explicitly considers the production choices facing subcontractors or their internal cost drivers. Some literature provides a macro-level view of the conditions facing subcontractors (Gray & Flanagan, 1989; Bennett & Ferry, 1990) or focuses on the contractor–subcontractor relationship (Hinze & Tracey, 1994) rather than on internal subcontractor policies. While none of this literature details resource management policies at the operations level, there are several points of agreement between the literature and the Pacific Contracting case study that suggests Pacific Contracting is a common example of its genre. One point common among all the authors is that subcontractors are commonly subject to changing conditions with poor management by general contractors. In a review of subcontracting in the United Kingdom, Bennett & Ferry (1990: 271) note that: … under construction management contracts, and to a greater extent under management contracting, the specialists are just thrown together and told to sort things out by themselves.

104 W.J. O’Brien

Another point of agreement is that subcontractors generally try to take matters into their own hands in attempt to better control their fate under the vagaries of changing site conditions and schedule. Hinze & Tracey (1994) concluded in a United States-based study of the contractor–subcontractor relationship that subcontractors often act autonomously of general contractors to smooth execution of their project obligations. Gray & Flanagan (1989) note that subcontractors are increasing their involvement in the design stage to improve productivity and shield themselves from disruption on-site due to poorly coordinated design; this is a strategy Pacific Contracting makes aggressive use of. More generally, Bennett & Ferry (1990) describe how successful subcontractors actively maintain a core set of skills and capabilities that they protect by established boundary control systems. Such boundary systems buffer production workers from disruption, something Pacific Contracting attempts to do in its planning process. Unfortunately, no amount of buffering can fully shield a subcontractor’s resources from varying site conditions affecting productivity or from changes in schedule. As such, the Pacific Contracting study below details not just the strategies used to avoid disruption but also how the firm responds to changes in site conditions and schedule which affect resource productivity and deployment.

OVERVIEW A small roofing subcontractor working primarily in the California Bay Area, Pacific Contracting bills approximately US $5 m a year. As a young and growing firm, Pacific Contracting (Pacific) has sought to use technology and best-practice management methods to improve productivity and quality and therefore enhance market position. Currently, Pacific has formed an alliance with a major general contractor to become the general’s preferred supplier of roofing on the West Coast. Pacific’s long-term strategic plan is to become a “skin contractor” that organizes and controls the erection of the exterior walls and roof of a building. In full realization of its growth plans, the firm will provide engineering services for design, detailing, and value engineering; these engineering services will support and improve the construction capabilities of the firm and its suppliers. Pacific Contracting’s strategic plan is largely a response to conditions of site construction. In their work, Pacific’s management has observed that interaction between trades on-site is often poorly coordinated and that design information is often incomplete and/or the project is difficult to build as designed. These problems delay or otherwise interrupt the progress of work on-site, increasing direct costs for Pacific as well as decreasing its ability to predict and level demand for resources across jobs, leading to a series of indirect costs. As a roofing firm, Pacific is often in the position of doing its work after other firms, subjecting them to the aftereffects of the problems of others. At the same time, roofing is usually a

4D CAD and dynamic resource planning for subcontractors 105

critical path item because of the need to make the building watertight for interior work; this places Pacific under enormous pressure to complete their work on-schedule even if there were delays caused by others.

RESOURCE MANAGEMENT PROBLEMS The reality of construction sites is that trades closely follow each other. If they are poorly scheduled or if production slows, Pacific crews can be kept waiting, working unproductively in a start-stop fashion, subject to the progress of others. Pacific has also found that because of pressures to complete the overall project as quickly as possible, general contractors will attempt to get subcontractors to start on-site as early as possible. This leads to a condition known as “trade stacking” where multiple trades follow each other closely and, like dominoes, if one trade falls prey to difficulty, all are affected. In particular, Pacific finds itself subject to both poor scheduling and work area coordination among trades. Start and end dates for scheduled work are uncertain, and site conditions are not conducive to productive work. These two problems are intimately linked and comprise the biggest challenge to profitable allocation of Pacific’s resources. Pacific has a fixed group of workers and site superintendents. To maintain productivity, projects should be scheduled such that all workers are fully employed at all times. The implication of this is that workers should finish one project and move directly to another. Unproductive site conditions, because they cannot be fully anticipated and may vary over the course of a project, lead directly to uncertain end dates (and hence, uncertain start dates for the next project), making it difficult for Pacific to maintain constant and level use of its work force.

RESPONSES TO RESOURCE MANAGEMENT PROBLEMS: INFLUENCES AND CONSIDERATIONS Pacific maintains a number of degrees of freedom in the deployment of its (finite) resources. However, these are subject to a number of constraints and considerations, including site conditions, the nature of the work underway, worker skills, and the availability of resources. Profitability considerations, relationship management, contractual conditions, and upcoming project commitments also constitute meaningful constraints on Pacific’s actions. These various influences constitute a complex environment that Pacific manages heuristically, seeking its most profitable allocation of resources in an environment that changes daily. Major influences on Pacific’s resource allocation and profitability are detailed below, together with policies that Pacific has adopted to manage in response to these influences.

106 W.J. O’Brien

Site conditions Poor site conditions comprise the biggest threat to Pacific’s worker productivity and profitability. Ideally, when Pacific begins to work on a project, all prerequisite work has been completed, materials are available and accessible, access routes and work areas are clear and clean, and hoisting equipment is readily available. The most important of these is the completion of all prerequisite work. If work is incomplete, Pacific’s work is impacted in three ways: first, through the necessity of work arounds; second, through increased coordination requirements with other subcontractors; and, third, through diminished work area. In the first case, if work has not been performed, then Pacific must work around the unfinished areas, lowering worker productivity in the first pass and in the necessary second pass to complete the affected area. Such work arounds also make meeting completion dates difficult, which can lead to conflicting demands on resources as well as increased liabilities. This leads to the second problem of unfinished prerequisite work: other subcontractors must make multiple passes, requiring increased coordination between those subcontractors and Pacific. Because other subcontractors have already completed a first pass, they may not readily have resources available for a second pass and will keep Pacific waiting. Further, on subsequent passes, subcontractors may damage materials that Pacific has installed. This is especially worrisome given the delicacy of roofing membranes and the liability Pacific assumes in the case of leaks. The third difficulty of incomplete prerequisite work is altered or diminished work areas. Proper sizing of work areas gives Pacific the ability to sequence and structure work for maximum productivity. In the simplest case, Pacific follows the work of others in a linear fashion and there must be a large enough buffer between trades such that there is no interference from random fluctuations in production rate (and, hopefully, a large enough buffer to provide some protection against work stoppages). More commonly, Pacific inherits an area of work from a preceding subcontractor. When finished, Pacific releases that area to a following trade and moves on to the next area. Size and sequence of work areas or packages are often determined by the general contractor’s plan for completion of the facility. Where possible, Pacific will negotiate with the general to sequence and size work areas in an optimal manner for Pacific’s worker productivity. About 80% of Pacific’s work is negotiated in the sense that the overall and the particulars of schedule are discussed. Pacific plans its work very carefully; after the building is divided into multiple work areas or work packages, a bill of materials is determined for each. Pacific then sequences its work within each work area, determining the order in which to lay down the multiple layers of membranes or surfaces, the flashing around the roof edges and penetrations, etc. Resource requirements are determined from the volume and type of work in the work packages, the sequence of the work to be done, and the completion dates of each package. Reference is also made to Pacific’s overall resource commitments. This aggressive planning serves to maximize worker productivity in each work package. However, work areas smaller than planned for

4D CAD and dynamic resource planning for subcontractors 107

make it difficult for Pacific to sequence its work effectively in that area. This is exacerbated if the physical space allotted to the crew is too small for them to be productive; below a certain size Pacific finds that crews just cannot be efficient. Nature of the work and worker skills The severity of resource and productivity problems relating to site conditions is closely related to the nature of the work at hand and the skill of Pacific’s workers. For example, some kinds of work are very sensitive to sequencing concerns while others are not (this extends to setting up an area for work as well as performing the work). Likewise, some kinds of work are very flexible in staffing requirements while others are not. Laying tile, e.g. can be done in small teams and is an activity that can be rapidly scaled up or down. Laying roofing membranes is a less flexible activity where there is a definite crew size and may also be physically constrained, e.g. a specific area must be covered in a single pass to ensure continuity. One strategy that Pacific uses to balance demand for resources is to shift them between projects; thus rather than work at lower productivity due to smaller work area, Pacific may pull its workers from one project and place them on others. Its ability to do this, of course, is determined heavily by the nature of the work underway. Together with current backlog available to be worked, the type of work determines the ability of a project to absorb (or loan) resources. In practice, Pacific shifts workers (and less frequently, other resources such as equipment) between projects on a weekly and sometimes daily basis. Currently, site superintendents and project managers determine resource re-allocations. They determine the project needs for the week and allocate resources, fine tuning the plan and responding to changing conditions daily. Of course, Pacific’s ability to do this, beyond being constrained by the mix of work underway, is also constrained by worker skills. Pacific employs full-time a core group of union workers. These workers are highly skilled and reliable, assets that Pacific has difficulty finding in the labor market. Pacific is reluctant to release core workers to the union hall as other firms may hire and keep them. Of course, not all of Pacific’s work force is equally skilled. Some workers can handle certain types of technologies while others cannot; this limits the ability to shift workers between projects. Similarly, certain workers work well in teams but not apart; these comprise discrete resource restraints not directly related to the nature or technological requirements of the work. Acquiring temporary workers from the union hall also poses some problems as they have not had the experience working together that Pacific’s core workers have had; further, as the skills of temporary labor are questionable, Pacific will try to employ the most experienced (hence, most expensive) workers. Learning effects are also important. In practice, Pacific has observed that workers quickly become productive on a job once they learn the general layout (where to get materials, etc.). If Pacific has done its job planning the work and communicating the design requirements clearly, maximum productivity is quickly obtained for all but the most complex operations. However, if workers are constantly switched

108 W.J. O’Brien

between projects, morale goes down and productivity is decreased as workers constantly must relearn the job conditions as well as the work to be done. As some switching is necessary even in the best conditions, Pacific will try to maintain a core group on the project to act as a project memory that can quickly educate workers who come late onto the project. Thus when Pacific must completely pull off a project, the cost of lost learning is greatest. Resource availability Pacific’s greatest problem in resources is the availability of workers. Skilled workers are scarce and thus Pacific is limited to a fixed pool of workers which it employs fulltime; Pacific will only lay off core workers if it expects a long period where it works below capacity. In this sense, core work force capacity is a fixed cost to Pacific. Equipment is rented where possible; if equipment is rented, it is generally available in the market. In cases where Pacific owns or must buy the equipment for a project, there are problems when project demand for resources changes. This is the same problem of having a fixed core of workers and variable demand, leading to capacity conflicts. In such cases, Pacific will choose to rent extra equipment at its expense, or try to share its existing resources (either through overtime and switching equipment between projects or at the expense of completion dates for project work packages if not for entire projects). Availability of materials is somewhat more complicated and is partially a function of the relationships Pacific has cultivated with its suppliers. On any given project, Pacific will have between 2 and 10 different suppliers (15 suppliers maximum), including equipment suppliers. Pacific maintains sole source alliances with 10 different suppliers in return for which it gets a discount; however, Pacific does endeavor to remain aware of market prices. Lead-times for materials are generally 30 days. With this lead-time, materials are generally available although a supplier will occasionally miss its delivery date. More troublesome is the shifting demand from projects for resources. Accelerated projects will often violate leadtime requirements (delayed projects increase Pacific’s inventory carrying cost). If lead-time requirements are violated, Pacific can pay a premium to obtain materials. However, the cost of this premium has been increasing as suppliers have been moving from a make-to-stock to a make-to-order system; this has made it more difficult for suppliers to accelerate delivery. Part of Pacific’s reason to cultivate good relationship with suppliers is to obtain preferential treatment when the suppliers are capacity constrained. Another option that Pacific has is to transport materials between projects—in those cases where there is a backlog of materials on one project, Pacific can transport at its own expense materials to the needy project. Deliveries (possibly expedited to avoid new shortages) are then made to the project borrowed from. Another option is to start slowly (use a small crew) on the project with the materials shortage, postponing the need for materials until they can be delivered. Typically, Pacific uses a combination of these alternatives to cope with materials shortages.

4D CAD and dynamic resource planning for subcontractors 109

Payment, contractual conditions, and relationship management Management of resources is also subject to business considerations apart from the physical constraints of site conditions, worker skill, and resource availability. These business considerations—payment and profitability, contractual commitments, and relationship management—inform the resource decisions Pacific makes in response to changing physical conditions. These considerations have different timeframes. In the near term, the specifics of payment can have a great impact on how Pacific chooses to allocate resources among current projects. Like any firm, Pacific must closely watch cash flow and tries to get paid as soon as possible. Projects often have different billing dates; e.g. project A may require subcontractors to submit invoices on the 15th of the month while project B requires invoices on the 30th. Pacific is able to submit invoices for completed work areas or as a percentage of work complete. In either case, Pacific is likely to shift resources to project A in the week preceding the 15th to maximize work completed and the amount that can be billed. Similarly, Pacific is likely to shift resources to project B in the week preceding the 30th (effectively slowing down work on project A). Another consideration is the terms of payment. Some contractors pay promptly and others pay slowly; depending on cash flow needs Pacific may shift resources to the project run by the general contractor that pays more quickly. Terms for materials payment can also influence how Pacific allocates resources, especially if Pacific only gets paid when those materials are installed as opposed to delivered to site. A further nuance comes from the terms that Pacific has with its suppliers; these terms are often longer than the terms that Pacific has negotiated with the general contractor. This is especially true of larger suppliers who have a low cost of capital compared to smaller firms such as Pacific. These larger firms are often commodity suppliers who compete on price; one way to “buy” the job is to extend payment terms. Pacific has seen instances of terms net 180 days. While Pacific does act opportunistically to shift resources among jobs to better manage its cash flow, it is limited by the balancing factor of switching costs as described above. Contractual commitments also constrain its actions. As with most subcontracts, Pacific is usually liable for delays that it causes. Shifting resources between projects can expose Pacific, should this shifting lead to delays (either immediately or at some future date). Contractual commitments for completion often extend to parts of a project as well as the whole project; Pacific will be asked to commit to a specific schedule to make areas of a building watertight. This further constrains Pacific’s ability to allocate resources without subjecting them to liabilities. Thus Pacific’s short-term actions of daily or weekly re-allocation of resources are often informed by medium-term commitments for project completion. Relationship management for the long-term is also an influence on short-term actions. Pacific maintains a series of relationships with general contractors, other subcontractors, and suppliers. Pacific has favorite generals that it works with on a repeat basis; these tend to pay on-time, respond favorably to subcontractor requests, and do a good job of planning and managing the project. Desiring to maintain and

110 W.J. O’Brien

improve the relationship, Pacific will allocate resources favorably towards the projects run by those contractors. Similarly, Pacific must maintain good working relationships with subcontractors. As subcontractors work closely with each other, often negotiating hand-offs directly when the general contractor does not take an active role in on-site workflow management, Pacific must maintain its commitments to other subcontractors lest it develop a bad reputation and be subject to the whims of others. Suppliers have less immediate influence on Pacific’s actions onsite; as noted above, Pacific does maintain good relations with key suppliers to obtain favorable pricing and shipment dates. For its part, Pacific treats its suppliers well with repeat business and prompt payment.

CONCEPTUAL BASIS FOR MULTI-PROJECT MANAGEMENT Pacific Contracting is not unique in its resource management difficulties or in its responses to those difficulties. Birrell (1980) appears to be the first author to explicitly note that subcontractors autonomously control resources allocation and face challenges balancing demand across projects. He advocated that construction managers should attempt to level resource utilization on the project level to ease subcontractors’ resource management challenges. Building on this work, O’Brien & Fischer (2000) found that subcontractors generally do have finite resources that they shift between projects to optimize productive use of those resources. From empirical research, we know that:

• Subcontractors shift resources frequently (up to daily) among projects. • Shifting of resources occurs not just due to changes in project schedule, but also • • • •

due to poor site conditions that lower productivity (resources are assigned to projects allowing full productivity). The fundamental unit that subcontractors shift resources from and to is not the project but the work areas (work packages) that projects are composed of. There are multiple classes of resources (both labor and equipment) and some of these resources do not shift as easily as other resources. There are constraints in shifting labor because of training, union rules (jurisdiction and prescribed balances between apprentices and journeymen), and morale/ team work. Project completion dates and liability must be considered, but “soft” considerations of relationship building may also play a role.

The empirical evidence that subcontractors have finite resources that they dynamically allocate across projects suggests a need for a framework within which to make resource allocation decisions. Unfortunately, little research exists to guide subcontractor management in their decisions. O’Brien & Fischer (2000)

4D CAD and dynamic resource planning for subcontractors 111

critique existing scheduling and costing methods such as the network scheduling and the time–cost tradeoff technique (Fondahl, 1961; Antill & Woodhead, 1990) as incapable of representing the true costs of subcontractors. Existing techniques only treat projects individually, and do not consider interaction between projects. Only a few new methods for considering subcontractor resource management from a multi-project basis are beginning to emerge. Choo and Tommelein describe a resource constraint database to aid subcontractor space coordination on-sites (Choo et al., 1999; Choo & Tommelein, 2000a, b). Their database allows manual entry of multi-project resource allocation. It detects resource conflicts and is directed towards helping subcontractors maintain a level use of resources; it lacks a decision support component relating to cost and does not address fluidity of resource allocation across projects. O’Brien (2000a) describes a set of parametric models that relate site conditions, resource allocation, and productivity on a work area. These models can be used to support multi-project resource allocation decisions by providing detailed assessment of individual work areas, but do not provide a unified framework to assess multi-project costs. A starting point for the development of framework to address multi-project management concerns of subcontractors is to view the allocation of their resources across time to projects. This is shown in Figure 1, where projects are shown as (numbered) blocks that occupy a certain portion of capacity over time. The representation and discussion around Figures 1 and 2 was first published in an earlier form in O’Brien et al. (1995). Capacity is a manufacturing concept that can be loosely defined as the maximum productive output of a firm’s resources (see Gershwin, 1994 for a more general discussion). Capacity utilization in Figure 1 is shown as an aggregate of the firm’s total set of productive resources. This is an abstract representation. A more realistic one would be to show a capacity utilization/time plane for each resource. Each resource works on a project for a period of time, then moves to the next project (or is idle). As firms usually work on each project with a set of resources, however, we can view capacity utilization in Figure 1 as the set of resources working on each project over time. As we know from case evidence, this set of resources working on a project can change (e.g. workers can be shifted from one project to another). Yet for any given moment in time, Figure 1 is an accurate representation of how resources are currently deployed. In this

Figure 1. Capacity utilization by a firm’s projects over time (sample projects numbered 1–8).

112 W.J. O’Brien

sense, 100% capacity utilization (the dashed bar) represents the capacity constraint of the limiting resource for that moment in time. At one instant, the limiting resource may be equipment; as the firm moves to new projects, the limiting resource may be labor. Case evidence suggests that for subcontractors labor is more likely to be the limiting resource than is equipment. Figure 1 depicts the capacity utilization of a firm working on projects over time. This representation is similar to some manufacturing representations of scheduling jobs (projects) on machines (see, e.g. Pinedo, 1995). However, there are substantive differences in that construction projects are unique and hence it is difficult to represent them as jobs that follow a stochastically predictable production demand. Also, the history of site production can have implications on schedule. Despite differences from manufacturing production models, there are some basic understandings from the operations literature that can be used to understand the link between capacity utilization and cost. Basic queuing theory indicates that, in a stochastic system, as capacity utilization increases, the waiting time in the queue increases non-linearly, becoming infinite at 100% utilization (Gross & Harris, 1985). In actual manufacturing environments, this manifests itself as congestion effects where backlog increases, consequently increasing cost due to delays, carrying excess inventory, etc. (Askin & Standridge, 1993). Banker et al. (1988) report a strong non-linear relationship between costs and increasing capacity utilization when firms are working near capacity. Consider the projects shown in Figure 1: if Project 3 requires more resources these can only come from Projects 2 or 4 or from finding a way to increase the total available resources. Should resources be drawn from Projects 2 or 4 to support Project 3, this might cause Projects 2 and 4 to be late and/or have follow-on effects on later projects. (Follow-on or ripple effects may affect both the firm in question and other firms that are dependent on it meeting schedule on the affected projects.) Similarly, should schedule change such that Project 5 in Figure 1 starts late there will be resource problems with Project 6. If a firm is near full capacity utilization, responding to a change involves significant costs of reallocating resources (possibly delaying other projects) and/or costs of adding overtime. Here, costs are highly non-linear and may not be continuous. This relationship carries even if there are no dramatic schedule changes; if a firm is working near 100% capacity the natural small variations and problems that manifest themselves on any project can lead to large costs as there is no reserve. With this understanding, it is possible to show a basic relationship between costs and capacity utilization when firms are working near their capacity. Figure 2 shows a cost surface projected above the capacity utilization/time plane. Cost represents total cost of output. The cost surface is only meant to show a general relationship in the relevant region of high capacity utilization (if continued to zero capacity, the cost surface would be u-shaped to reflect the fixed and overhead costs of the firm). Actual shape of the cost surface will depend on the mix of production the firm is undertaking at the time, and the firm will experience only the

4D CAD and dynamic resource planning for subcontractors 113

Figure 2. Relationship of cost of output to capacity utilization when the firm is operating near full capacity.

costs indicated along the expenditure path of Figure 2 (which generally follows capacity utilization). As we can see from Figure 2, the performance of a subcontractor stems from its ability to productively allocate its resources across all the projects it works on. Thus, to operate, a subcontractor’s management must ask, answer the question “How do we best allocate our resources?” on a regular basis. This translates to several questions worthy of future research:

• • • • • • • •

Which project should we borrow resources from? Which project should we deploy idle resources to? What are the ripple effects of a resource re-allocation and what do they cost? How do we value the flexibility of resource deployment on a project when bidding? What is the most profitable mix of projects? How does production technology affect our ability to accommodate changes? How should we negotiate a schedule with a contractor? How much reserve capacity should we have to accommodate changes?

These questions are used as a basis for discussion of extensions to 4D CAD models below.

PACIFIC CONTRACTING PART TWO: 3D/4D CAD AND RESOURCE MANAGEMENT Pacific Contracting’s approach to multi-project resource management difficulties is to minimize to the causes of changes to plans. Following the Ballard & Howell’s

114 W.J. O’Brien

(1998) philosophy of shielding production, Pacific takes a three-pronged approach to reducing the incidence of changes: 1. To better control those factors that influence its costs, Pacific seeks to “own” as large a portion of the job as possible. As noted, it plans to become a full-service “skin contractor” where it contracts for the design and construction of the outer surface of the building. This includes taking control of detail design and design coordination so that work is never held up for lack of information. Pacific also attempts to control a large number of the technologies and trades that regularly interfere with its work by conducting the work directly through its own work force or indirectly through second-tier subcontractors who work for Pacific. While not yet a full-service “skin contractor,” Pacific does what it can to control the factors affecting its own work, including taking on design coordination for its work and related trades and aggressive review of schedule with general contractors. 2. Pacific maintains a policy under the “Last Planner” system of Ballard & Howell (1998) not to work on any project or work area of a project until certain preconditions are met that allows full worker productivity. These preconditions include contract complete, all prerequisite work finished, enough work area for the workers to work, all materials available, crew available, design complete, and production sequence planned. Pacific uses these preconditions as a checklist for each work area on a project; unless each condition is met, Pacific will not work on the area. The checklist is also used as a planning tool. Updated weekly, it gives a status report about project readiness and is used to direct efforts towards making projects and work areas workable. Ideally, the checklist is completed early for each work area, creating “workable backlog” so that if an unexpected change happens, Pacific can shift work from one work area to another with little lost productivity. 3. Pacific inputs all design drawings related to the roof/skin into a 3D CAD model. This insures that the design is complete and coordinated (supporting its first approach of coordinating trades that may interfere with its operations). Pacific uses the 3D model as part of a materials management program; it uses the 3D drawing to count parts and obtain exact dimensions and quantities. It then uses this data to order all supplies; where it can, it combines orders across jobs to obtain bulk discounts or economize on waste. The 3D model also allows material needs to be determined exactly not just for projects but also for work areas, allowing supplies to be directed managed and directed to those work areas (this directly supports the “Last Planner” checklist). A sample bill of materials drawn from a 3D model is shown in Figure 3. Pacific extends the 3D model to 4D for both trade coordination and production planning. Trade coordination in terms of sequencing and interference checking is a traditional use of 4D (see, for example, Thabet & Beliveau, 1997; Akinci & Fischer, 1998; and other papers in this volume), and Pacific uses the 4D model to check contractor schedules and make suggested improvements. It also uses the

Figure 3.

Sample Pacific Contracting bill-of-materials drawn from 3D model.

4D CAD and dynamic resource planning for subcontractors 115

116 W.J. O’Brien

Figure 4.

Sample Pacific Contracting assembly drawing sent to field workers.

4D model to sequence its own work and those of any firms that work for it. With regard to production planning, Pacific is more innovative. More than any other firm, it uses detailed 3D models to create an installation sequence. It pre-plans all operations in a 3D/4D model, using the model to create a series of isometric drawings sent to field. These isometric drawings replace the traditional 2D design and shop drawings that are used in the field. Figure 4 shows a detailed assembly drawing, and Figure 5 shows stages 1– 4 in an 11-stage installation process. Pacific has found that this level of detailed production planning nearly eliminates time wasted in the field due to questions about design intent. This level of planning also allows Pacific to plan for efficient operations, using the 4D model as the basis for simulating production. While expensive to create and manipulate 4D models at the level of detail shown in Figures 4 and 5, Pacific has found its investment to be profitable. In the words on one Pacific manager, “It is far cheaper to hire one engineer than to have a crew waste time in the field.”

CURRENT DIRECTIONS AND LIMITATIONS IN 4D CAD Pacific’s use of 3D/4D tools is a response not just to design problems but also to resource management difficulties. Pacific tries to ensure that the design is

4D CAD and dynamic resource planning for subcontractors 117

Figure 5.

Sample Pacific Contracting installation stages drawing sent to field workers.

complete, is coordinated with other disciplines, and is buildable. To optimize use of its resources and in conjunction with its “Last Planner” inspired resource assignment process, Pacific sequences its field production through the use of detailed isometric drawings and also uses 3D models to create a detailed bill-of-materials. This use of 4D and production shielding is an effective way to optimize operations on individual projects and to shield these operations from disruptions, and has been very profitable for Pacific. However, despite the success of Pacific’s approach, it does have limitations in that it stems from a single project perspective: all shielding and 4D planning occurs for individual projects and there is no explicit framework to consider resource management and production planning from multi-project perspective. This is particularly troublesome under conditions of uncertainty, something that Pacific is subject to even with its planning. Thus there is a need to extend 4D techniques to a multi-project perspective. Before considering multi-project resource planning and costing extensions to 4D CAD, it is worthwhile to review the development and current state-of-the-art of 4D tools (beyond that practiced by Pacific). Current research directions in 4D CAD are logical descendants of earlier project centric research and practice with 3D CAD models. The development of 3D CAD models in the late 1980s was a

118 W.J. O’Brien

significant advance in the development of construction modeling technology. The ability to construct a virtual prototype before field construction began dramatically decreased interference among major systems (e.g. piping, mechanical, and structural systems) and increased the speed and quality of design review. Other early uses of 3D CAD models included limited studies in constructability. On one project, 3D CAD models and renderings were used in coordination meetings to discuss trade sequencing (Griffis et al., 1990). Research activities associated with that project explored the link between traditional simulation of construction activities and simulation of construction in the context of the 3D model (Griffis et al., 1991). This work quickly led to the development of 4D CAD approaches to modeling both the facility and the construction process. Early 4D CAD efforts manually integrated schedule information with 3D models to represent the planned state of facility construction at fixed points in time. Later 4D CAD research efforts automated the link between scheduling software and 3D models to create more flexible and dynamic 4D representations of construction progress. Research has also focused on generation of automated (4D) schedules by reasoning about the 3D design and construction process information (e.g. Darwiche et al., 1988; Thabet & Beliveau, 1994, 1997). Current application and research frontiers in 4D CAD include detailed work planning (Riley, 2000) and coordination of multiple trades in a dynamic and uncertain project environment (Akinci & Fischer, 1998; Tommelein, 2000). Research has also begun to incorporate construction costs in 4D CAD models. Staub-French & Fischer (1999) review the practical needs of cost, schedule and scope integration and outline an approach to cost planning at the activity and object level in a 4D environment. Staub-French & Fischer’s (1999) research perspective can be shown in Figure 6, where cost is a third link in a triangle integrating cost,

Figure 6. Current state of 4D CAD research—linking design, schedule, and cost in the context of a single project.

4D CAD and dynamic resource planning for subcontractors 119

scope (design), and schedule in the context of a single project. They identify several impediments to effective integration of project information, in particular different levels of aggregation in the use and generation of design, schedule, and cost data. Through case examples, Staub-French and Fischer also show how the standard construction accounting and control methods and existing software are often too rigid to accommodate dynamic reasoning about and representation of construction methods. Building on the work of Fischer & Aalami (1996), they propose a conceptual schema that will accommodate both different uses and different levels of aggregation of construction information in a unified format that will allow integration of cost, schedule, and scope on a project. While creating increasingly more powerful tools, research in 3D/4D CAD has stemmed from a single project perspective. Existing 3D/4D tools do not directly support a multi-project viewpoint (a perspective this author has called “5D CAD” in to distinguish it from single project approaches (O’Brien, 2000b)). Based on the discussion above, a multi-project or 5D CAD tool should support decisions about cost, time, and resources at the firm level. While more research needs to be performed developing integrated costing models to provide a decision support framework for multi-project resource management, several extensions to 4D techniques are possible. Let us consider them in the context of the resource management questions presented above. Consider the first three questions:

• Which project should we borrow resources from? • Which project should we deploy idle resources to? • What are the ripple effects of a resource re-allocation and what do they cost? These are intimately related and concern the daily operational decisions a subcontractor must make about where and how to deploy resources. Should there be a problem or acceleration on one project, the first question asks which of the subcontractor’s other projects should resources be borrowed from. This is not a simple question as moving resources from one project to another may simply solve one crisis by creating another. Many subcontractors do shift resources, and many subcontractors do seem to always be running late. Similarly, deploying idle resources to a project may not help that project. If the subcontractor completes work early on one project, which project should it allocate those now idle resources to? Ideally, the subcontractor will maintain a level use of resources (perhaps with some spare capacity from a queuing perspective (Hopp & Spearman, 2000)), and should there be changes, the subcontractor will seek to minimize the cost or consequences of any ripple effects. Thus, for the first three questions, what is needed is an extension of 4D CAD that allows the ripple effects of resource re-allocations to be modeled. This can be accomplished manually today insofar as we can assess productivity on a project for a given resource level. With some knowledge of the overall schedule for each project (especially with regard to float and space) and any penalties for delays (and incentives for early completion), it is possible to model the impact of a given

120 W.J. O’Brien

resource re-allocation on each project. The overall impact of any proposed change can thus be assessed by individually evaluating each affected project. Of course, this is extremely cumbersome to do manually, and should a resource re-allocation generate significant further re-allocations (i.e. large ripples) it may be impossible to manually enumerate all the possibilities. What is needed is a 4D tool that allows automatic exploration of the impact of resource re-allocations on a project. With this, we also need a tool that tracks the impact across projects (including ripple effects), allowing exploration and evaluation of changes. This appears to be a significant research challenge, although some groundwork has been put in place by the automatic space and conflict evaluation work of Akinci & Fischer (1998) and Thabet & Beliveau (1997), and the resource tracking work of Choo et al. (1999). The second set of questions concern themselves more with planning than with operations:

• How do we value the flexibility of resource deployment on a project when bidding? • What is the most profitable mix of projects? • How does production technology affect our ability to accommodate changes? These questions are seemingly unrelated, but consider that a subcontractor’s management knows there is a high probability of resource re-allocation. In this case, the management would like to have a set of projects where re-allocation of resources is not costly. Conceptually, one set of projects will be more or less flexible in accommodating changes than another set of projects. This flexibility directly relates to profitability, and thus subcontractor management would like to assess the value resource flexibility on projects both individually and as a set. To a certain extent, some subcontractors already practice choice about project mix; e.g. one steel fabrication and erection subcontractor studied likes to work on one large project and several small ones at any given time (O’Brien, 1998). Such a mix provides the firm with flexibility in meeting changes while keeping near full use of its productive resources. Valuing flexibility is not simple. Influences on the cost of shifting resources include not just schedules and any associated penalties, but also the technologies involved. In a simple sense, a subcontractor cannot shift from one technology to another if there are different equipment and skills involved. In a broader sense, construction technologies can accommodate changes in resource level with different levels of impact on productivity (see O’Brien, 1998, 2000a for more discussion and examples). With knowledge of the impact of a technology on productivity, conceptually it is possible to extend the 4D/ripple-effect technologies envisioned above to provide detailed knowledge about the value of flexibility for a given set of projects. Essentially, the question is the same: what is the impact of a change? Of course, in the early planning stages less detail will be known about the project(s) than in the operational stages. It is unclear if the same set of technologies

4D CAD and dynamic resource planning for subcontractors 121

that will allow detailed assessment and enumeration of changes for improvement in operations will work in a less information rich environment. It is likely that the data from the 4D technologies envisioned would have to be put into some sort of options framework (Trigeorgis, 1996; Brennan & Trigeorgis, 2000) that can accommodate uncertainty. The final questions directly concern short-term operational planning rather than accommodation of changes after they occur (questions 1–3) or with pre-bid or investment planning (questions 4–6):

• How should we negotiate a schedule with a contractor? • How much reserve capacity should we have to accommodate changes? If the first three questions (above) concern themselves of what to do after a change occurs, these final questions concern themselves with planning for those changes. With some knowledge of the value of a mix of projects and the influence of schedules and incentives on that value, subcontractors can better address the detailed negotiations with contractors about cost/price, schedule, materials, etc. Similarly, with a solid plan of projected resource assignment to projects over time, subcontractors can better negotiated a detailed schedule of work area hand over dates on individual projects. With regard to extensions to 4D tools, it is likely that a combination of the tools described above will provide subcontractors with the requisite knowledge about what to do. New development that is needed is a better visualization tool to support negotiations between the contractor and the subcontractor(s). As for schedule changes, improved knowledge about the impact of possible changes on cost and schedule can guide the subcontractor in making choices about reserve capacity. Subcontractors can choose to keep some resources in reserve for a given mix of projects (type and amount of resources determined by the extended 4D tools), or they can choose to make strategic investments in new resources (either rental or new permanent investment). As with schedule negotiation, some combination of the tools envisioned above will provide the necessary decision support.

CONCLUSIONS Methods and models in 4D CAD have many benefits to construction projects, not least an ability to resolve design and construction conflicts before they occur in the field. Efforts in 4D CAD can be seen as a way to reduce the incidence of costly changes on construction projects. However, even in a world with widespread use of 4D CAD, changes in schedule and scope will remain a salient feature of construction projects. A core competency of subcontractors is the ability to adjust to these changes in a low cost manner. The principal difficulty that subcontractors have in adjusting to changes is maintaining a level use of their finite resources. Any change

122 W.J. O’Brien

in schedule and scope on one project can cause changes in resource allocation to other projects. This creates a complex interaction between projects and costs that cannot be captured in the single project perspective of 4D CAD models. The questions above developed from the multi-project perspective of subcontractors are very different from the questions asked of single project 4D models. A multi-project perspective does not invalidate current 4D research but does suggest augmentation of and parallel development with 4D research efforts. A multiproject decision support model (“5D”) needs detailed information about individual projects, in particular the likelihood of changes in schedule and scope and uncertainty in production progress, feasibility of alternative work plans, and the flexibility of a project in loaning or absorbing resources. The work of Riley (2000) in developing detailed work plans in a 4D environment, Akinci & Fischer (1998) in generating production alternatives in an integrated fashion, and Tommelein (2000) in understanding the affects of uncertainty on production are useful in a multi-project context. Similarly, Staub-French & Fischer’s (1999) schema for cost integration should provide a partial basis for reasoning about multi-project costs. Multi-project resource allocation models also provide information to improve decision making in a single project context. That subcontractors can and do shift resources among projects (O’Brien & Fischer, 2000) suggests that production reasoning in 4D CAD models based on static models of resource allocation will give inaccurate results. A multi-project model can improve the accuracy of project schedules, and provide improved support for cost negotiations when there are changes in schedule and scope. More broadly, the dynamic resource planning extensions to 4D CAD promise development of methods to plan projects from the perspective of subcontractor ability to respond to changes, allowing us to generate project schedules that are efficient and flexible.

REFERENCES Abu-Hijleh, S.F. & Ibbs, C.W. 1989. Schedule-based construction incentives. Journal of Construction Engineering and Management 115(3): 430–443. ASCE. Akinci, B. & Fischer, M. 1998. Time–space conflict analysis based on 4D production models. In K.C.P. Wang (ed.), International computing congress, Boston, 18–21 October: 342–353. ASCE. Antill, J.M. & Woodhead, R.W. 1990. Critical path methods in construction practice. New York: Wiley. Ashley, D.B. & Workman, B.W. 1986. Incentives in construction contracts. Report 83-5, Construction Industry Institute. Askin, R.G. & Standridge, C.R. 1993. Modeling and analysis of manufacturing systems. New York: John Wiley and Sons, Inc. Ballard, G. & Howell, G. 1998. Shielding production: essential step in production control. Journal of Construction Engineering and Management 124(1): 11–17. ASCE. Banker, R.D., Datar, S.M. & Kekre, S. 1988. Relevant costs, congestion and stochasticity in production environments. Journal of Accounting and Economics 10(3): 171–197.

4D CAD and dynamic resource planning for subcontractors 123 Bennett, J. & Ferry, D. 1990. Specialist contractors: a review of issues raised by their new role in building. Construction Management and Economics 8(3): 259–283. Birrell, G.S. 1980. Construction planning—beyond the critical path. Journal of the Construction Division 106(CO3): 389– 407. ASCE. Brennan, M.J. & Trigeorgis, L. 2000. Project flexibility, agency, and competition: new developments in the theory and application of real options. New York: Oxford University Press. Choo, H.J. & Tommelein, I.D. 2000a. Interactive coordination of distributed work plans. In K.D. Walsh (ed.), Proceedings of construction congress VI: Building together for a better tomorrow in an increasingly complex world, Orlando, Florida, 20–22 February: 11–20. ASCE. Choo, H.J. & Tommelein, I.D. 2000b. WorkMovePlan: database for distributed planning and coordination. In J. Barlow (ed.), Eighth annual conference of the IGLC, University of Sussex, Brighton, United Kingdom, 17–19 July: 12 pp. Choo, H.J., Tommelein, I.D., Ballard, G. & Zabelle, T.R. 1999. WorkPlan: constraintbased database for work package scheduling. Journal of Construction Engineering and Management 125(3): 151–160. ASCE. Darwiche, A., Levitt, R. & Hayes-Roth, B. 1988. OARPLAN: generating project plans by reasoning about objects, actions, and resources. AI EDAM 2(3): 169–181. Fischer, M. & Aalami, F. 1996. Scheduling with computer-interpretable construction method models. Journal of Construction Engineering and Management 122(4): 337–347. ASCE. Fondahl, J.W. 1961. A non-computer approach to the critical path method for the construction industry. Technical Report 9, The Construction Institute, Department of Civil Engineering, Stanford University. Gershwin, S.B. 1994. Manufacturing systems engineering, Englewood Cliffs, NJ: Prentice-Hall, Inc. Gray, C. & Flanagan, R. 1989. The changing role of specialist and trade subcontractors Ascot: The Chartered Institute of Building. Griffis, F.H. & Butler, F.M. 1988. Case for cost-plus contracting. Journal of Construction Engineering and Management 114(1): 83–94. ASCE Griffis, F., O’Brien, W. & Bronner, P. 1990. Columbia construction research: the applications of three-dimensional computer models in construction. Architectural and Engineering Systems. Griffis, F.H., Rubinson, D., O’Brien, W., Retailleau, S. & Zabilsky, R. 1991. Productivity applications: 3D models and simulation. Proceedings of construction congress ’91: 247–252. ASCE. Gross, D. & Harris, C.M. 1985. Fundamentals of queueing theory. New York: John Wiley & Sons, Inc. Hinze, J. & Tracey, A. 1994. The contractor–subcontractor relationship: the subcontractor’s view. Journal of Construction Engineering and Management 120(2): 274–287. ASCE. Hopp, W.J. & Spearman, M.L. 2000. Factory physics. New York: McGraw-Hill. O’Brien, W.J. 1998. Capacity costing approaches for construction supply-chain management. Ph.D. Dissertation, Stanford University. O’Brien, W.J. 2000a. Multi-project resource allocation: parametric models and managerial implications. In J. Barlow (ed.), Eighth annual conference of the IGLC. University of Sussex, Brighton, United Kingdom, 17–19 July: 11 pp. O’Brien, W.J. 2000b. Towards 5D CAD—dynamic cost and resource planning for specialist contractors. In K.D. Barlow (ed.), Proceedings of construction congress VI: Building together for a better tomorrow in an increasingly complex world, Orlando, Florida, 20–22 February, 2000: 1023–1028. ASCE.

124 W.J. O’Brien O’Brien, W.J. & Fischer, M.A. 2000. Importance of capacity constraints to construction cost and schedule. ASCE Journal of Construction Engineering and Management 125(6): 366–373. O’Brien, W.J., Fischer, M.A. & Jucker, J.V. 1995. An economic view of project coordination. Construction Management and Economics 13(5): 393–400. Pinedo, M. 1995. Scheduling: theory, algorithms, and systems. Englewood Cliffs, NJ: Prentice Hall. Riley, D. 2000. The role of four-dimensional (4D) models in trade sequencing and production planning. In K.D. Walsh (ed.), Proceedings of construction congress VI: Building together for a better tomorrow in an increasingly complex world, Orlando, Florida: 20–22, February, 2000: 1029–1034. ASCE. Staub-French, S. & Fischer, M. 1999. The practical needs of integrating scope, cost, and time. Proceedings of the 8th international conference on durability of building materials and component, May 30–June 3, 1998: 2888–2898. Vancouver, Canada: NRC Research Press. Stukhart, G. 1984. Contractual incentives. Journal of Construction Engineering and Management 110(1): 34– 42. ASCE. Thabet, W.Y. & Beliveau, Y.J. 1994. Modeling work space to schedule repetitive floors in multistory buildings. Journal of Construction Engineering and Management 120(1): 96–116. ASCE. Thabet, W.Y. & Beliveau, Y.J. 1997. SCaRC: Space-constrained resource-constrained scheduling system. Journal of Computing in Civil Engineering 11(1): 48–59. ASCE. Tommelein, I. 2000. Impact of variability and uncertainty on integrated product and process development. In K. Walsh (ed.), Proceedings of construction congress 6, special program of 4D CAD and visualization, Orlando, Florida, 20–22 February, 2000: 969–976. ASCE. Trigeorgis, L. 1996. Real options: managerial flexibility and strategy in resource allocation. Cambridge, Massachusetts: The MIT Press. Uher, T.E. 1991. Risks in subcontracting: subcontract conditions. Construction Management and Economics 9(3): 495–508.

THE ROLE OF 4D MODELING IN TRADE SEQUENCING AND PRODUCTION PLANNING David Riley Department of Architectural Engineering, Penn State University, University Park, PA, USA

Abstract A new class of planning tools has been developed through advancements in CAD and Scheduling software integration. 4D modeling provides a mechanism to visualize elements of 3D CAD models based on associated schedule intervals. This technology allows project managers to evaluate construction plans for time and space conflicts between operations and building elements. The use of 4D modeling for planning project logistics and evaluating project schedules is evolving rapidly. Recent research explores the 4D modeling of work spaces and material movement. Planning such spaces can be highly challenging when multiple sequence options and complex networks of prerequisite work exist. This paper discusses the use of 4D modeling for detailed trade sequencing and production planning for construction. Case studies of the sequence planning process are used to demonstrate how visualization and “what-if” tools could be used to improve the planning process. Conceptual methodologies are presented for modeling construction work spaces to support production planning, e.g. work, storage, and paths which are necessary to perform useful modeling and simulation of a dynamic work environment. The goal of this research is to identify appropriate applications for 4D modeling in the sequence planning process, and to make recommendations for the development of future 4D planning tools. Keywords: 4D CAD, CAD, coordination, planning, visualization

INTRODUCTION The integration of time with 3D spatial data permits the simulation of the dynamic work environments found on construction projects. 4D modeling provides a mechanism to simulate building elements and work spaces in a manner much more appropriate than traditional static layout plans and schedules. The benefits of 4D modeling 125

126 D. Riley

have been demonstrated through experimentation, and include the identification of potential conflicts between building elements and work spaces, safety hazards created due to proximity of construction activities, and the visualization of construction plans by crews (McKinney & Fischer, 1998). This paper is divided into two sections: Part One, which focuses on the planning process and potential benefits of 4D modeling to production planning; and Part Two, which focuses on modeling issues and the implementation challenges of 4D modeling for production planning. As a guide to the reader, each part is briefly introduced. Part One of this paper discusses the implementation of 4D modeling in an environment that typically resists large investments in detailed planning, and the potential impacts of 4D modeling on the creation of work sequences and production plans in construction. Consideration is given to the pragmatic issues facing construction planners such as the perceived lack of time for detailed planning. In an effort to address these issues, existing planning processes are examined to identify the most likely project attributes that lend themselves to useful applications of 4D modeling. Part Two of this paper explores the detailed modeling issues needed for the inclusion of physical work spaces, storage areas, and material paths as 3D objects in a 4D analysis of a construction project. Using experience from case studies and space planning efforts (described in more detail in Riley & Sanvido, 1995), attributes and properties for the modeling of construction spaces are defined. Next, the primary inputs and outputs of the planning process are discussed to demonstrate the role of these properties in the 4D planning environment. Finally, the relationships between the planning environment and the level of detail at which 4D modeling should be performed are discussed.

PART ONE—4D MODELING FEASIBILITY 3D computer models are rarely produced for building construction projects. Many contractors are quick to assume that the cost of CAD operators makes this type of investment too costly. On projects with complex geometric configurations of systems, however, 3D models are the most effective tool for examining the design of building elements. Examples of such spaces are mechanical rooms, interstitial spaces, and ceiling plenums. Many contractors recognize the value of 3D models in such areas, and invest in CAD operators for this purpose. 4D models provide an added dimension to planning, and permit the impact of work in place and available work space to be evaluated in addition to conflicts between building components. As building owners continuously demand more aggressive building schedules and computer modeling tools continue to become more affordable and easy to use, the development of 3D and 4D models will become more feasible. This paper explores the integration of 4D modeling into existing planning practices to help identify its most practical and feasible applications in construction.

4D modeling in trade sequencing and production planning 127

4D MODELING AND DESIGN COORDINATION A 4D model is typically created by associating schedule data with the elements of a 3D model. Since the most common use of a 3D model is for design coordination, it is logical that 4D models are likely most feasible on projects that also require extensive coordination between technical building systems and construction operations. In recent years the process of coordinating the design of mechanical, electrical, and plumbing (MEP) systems has intensified. Building systems are more complex and specialized, and the role of specialty contractors in the design process has increased. As a result, many contractors have recognized that a large investment in coordination, usually with CAD tools, is necessary to avoid costly delays due to conflicts. This coordination process offers the most likely opportunity to effectively generate and benefit from 4D models.

4D MODELING AND THE PRODUCTION PLANNING PROCESS A major challenge to construction managers is the conceptualization of how work crews, equipment, and materials will compete for limited available space during a construction project. 4D modeling provides a tool to visualize available space during the course of a construction project. Successful planning practices were investigated by Riley (1994) to determine how the management of space was used to create productive work environments. The results of this study are used to provide a framework for the integration of 4D modeling into the production planning process. The following five requirements of plans were identified: Provide spatial information: The decisions about how to use spaces on the site are made almost every day. A 4D model should provide spatial information to answer the following key questions:

• How should an activity be sequenced so that it does not interfere with the work of other activities or block paths? • Where is space available for a given amount of time for work, material or path related use? • When can materials be brought to the job without interfering with the work of activities, needing to be relocated (double handled), or having to be moved long distances between material and storage spaces? Balance project needs: 4D planning should consider three interdependent factors: productivity, quality, and safety. It should address all trades concurrently with an overall productive work environment as its primary objective.

128 D. Riley

Increase in detail as needed: 4D planning needs to increase in detail as a project progresses. A planning process should identify areas where additional planning is needed. Communicate the plans: For planning to be useful, it must be successfully communicated to the project participants. 4D models provide an excellent mechanism to communicate plans, as demonstrated by experimentation (McKinney & Fischer, 1998). Involve project participants in planning: The expertise and knowledge of specialty contractors is valuable in developing creative solutions to space planning problems. The value of planning will be proportionate to the inclusion of specialty contractors in the development of 4D models and resultant production plans. Case study 1: wastewater treatment plant Two case studies are presented to demonstrate the benefits of 4D modeling in the sequence planning process. The construction of a wastewater treatment expansion was analyzed by creating a 4D model of the major construction phases. The project included excavation, foundation, structural concrete, equipment, and piping for a new 35,000 SF treatment tank (Fig. 1). A 3D model of the facility was generated using FormZ™ software for the purposes of design coordination. The model was later linked to a summary schedule to permit the sequence of construction to be visualized. As a result, four key sequence issues were detected (Fig. 1):

• The routing of new utility lines across travel paths: Trenching and utility line construction threatened to cut off travel paths and disrupt the movement of materials and equipment on site. A 4D model of the site permitted an analysis of operations that require movement around the site, and the scheduling of trenching activities during intervals that minimize disruptions to this movement. • Equipment positions for the mat-slab foundation placement: A critical activity in the schedule of the project was the placement of the mat-slab foundation using

Figure 1.

Snapshot from 4D model of treatment plant construction sequence.

4D modeling in trade sequencing and production planning 129

two simultaneous concrete pumping operations. The 4D model permitted an evaluation of equipment placement and travel lanes needed for concrete delivery. • The phased construction of concrete walls to permit access for equipment deliveries: Completed concrete walls can block paths needed for delivery and placement of pumps and treatment equipment. At the same time, it was considered optimal to complete as many concrete walls prior to delivery of equipment to minimize risk of damage to equipment. A 4D model allowed path planning for equipment deliveries and the subsequent sequence of wall construction. • The integration of pipe spool deliveries with concrete wall construction: To minimize field welding of pipe spools and fittings it is advantageous to deliver large spools of shop-welded pipe. Completed concrete walls can make the delivery of large spools difficult or impossible. The 4D model of pipe spools and walls permitted an evaluation of which walls to leave out in order to accommodate pipe spool delivery and construction. This case demonstrates the value of a 3D model created for design coordination to the production planning process. By integrating the 3D model with schedule data, it was possible to identify several key spatial relationships affecting the construction sequence, and modify the construction plan accordingly. Case study 2: multi-story construction The construction of a five-story luxury apartment building was evaluated to explore the role of 4D modeling in the detailed sequencing of multiple trades on individual building floors. This project was first visited during the structural phase, as multiple trades prepared to follow shoring removal to complete interior rough-in and finishes. The general contractor had prepared a schedule for the work of each trade on successive floors. Figure 2 illustrates an excerpt of the two week per floor schedule

Figure 2.

Initial plan developed by general contractor.

130 D. Riley

for each trade. This schedule allowed for only one trade on a floor at a time, and demanded the completion of one floor every two weeks. No plan for a specific work direction was identified for the floors, as trades were expected to follow the removal of concrete shoring to begin their work sequence on each floor. Several problems existed with the initial plan. The MEP contractors on this project believed they would not be capable of completing the large floor plate in two weeks with their available workforce. Also, the lack of a defined work direction meant that crews of different trades would be forced to establish work sequence amongst themselves. Finally, no defined plan for loading materials onto the “L”-shaped floor plate had been established, allowing for walls and work in place to potentially obstruct material delivery and movement. Using a 2D CAD drawing of the floor plate and assigning a workflow through defined work spaces, the plan was evaluated more closely. By graphically reviewing the work on each floor in stages, it was determined that a phased construction of walls would permit open travel lanes for material handling activities necessary after the original scheduled wall construction. Also, by assigning a work direction on each floor, it was possible to manage how crews moved through each floor. This permitted multiple trades to occupy a single floor and remain out of each other’s way, and at the same time allow each trade to have more time on each floor. Figures 3 and 4 illustrate the proposed sequence of trades across a representative floor, and the associated work spaces needed for each trade. In addition to the revised sequence of wall construction, the major change to the original plan essentially reduces the “batch size” of work for each trade from the entire floor plate, to one wing of each floor. The characterization of construction work activities in terms of available space permits quantities of repetitious work to be grouped into desirable “batches” and adds detail to a work plan that goes beyond activities in the CPM schedule. The relationships between sequence, batch size, and buffers to 4D modeling methodologies will be discussed in more detail in Part Two of this paper. Figure 5 illustrates the revised plan. The new schedule is no longer sequential, as now multiple trades are planned to occupy the same floor at the same time. The durations of the MEP activities, however are extended and thus more realistic. As the plan was implemented, it was modified slightly to accommodate the shoring removal, however the trades were able to complete their work with minimal interference problems. This case provides a classic example of the threshold of current planning and visualization. The contractor was unwilling to schedule more than one trade on any given floor at a time, however had not accounted for the fact that several trades did not have the capacity to complete each floor at the stipulated rate. An examination of the flow patterns of specific crews across building floors was needed. By integrating flow patterns and the phased construction of core and demising walls into a dynamic model of the construction process, it was possible to generate a more feasible production plan. While the duration of the overall floor construction

4D modeling in trade sequencing and production planning 131

Figure 3. Location and sequence of work space assigned for HVAC (1), core wall framing (2), and in-wall plumbing (3).

Figure 4. Location and sequence of work space assigned for demising wall framing (4), in-wall electrical (5), insulation (6), and gypsum wallboard (7).

Figure 5.

Revised plan produced with visualization model.

132 D. Riley

was equal to the original plan, the individual activities were planned at a more realistic pace. While performed in this case with 2D CAD drawings, the combination of CAD and schedule data proved highly effective in developing a productive work sequence. Part One of this paper presents several issues facing the acceptance of 4D planning on construction projects. Two case studies provide illustrative examples of how 4D modeling and visualization of construction work can be used to design productive work plans on a wastewater treatment project, and a multi-story building project. Part Two addresses the modeling issues that must be considered when developing 4D models for production planning, with a specific emphasis on the modeling of work spaces, material paths, and storage areas.

PART TWO This section explores the inclusion of physical work spaces, storage areas, and material paths as 3D objects in a 4D analysis of a construction project. Using experience from case studies and space planning efforts (described above and in more detail in Riley & Sanvido, 1995), attributes and properties for the modeling of construction spaces are defined. Next, the primary inputs and outputs of the planning process are discussed to demonstrate the role of these properties in the 4D planning environment. Finally, the relationships between the planning environment and the level of detail at which 4D modeling should be performed are discussed.

MODELING OF CONSTRUCTION WORK SPACES Several efforts have been made to characterize the space required for construction work. Tommelein & Zouein (1993) related need, timing, and location as three necessary attributes of spatial resources. Thabet (1993) defined two parameters to compare space demand and availability, and developed a technique to monitor space demand by assigning activities to defined zones in a facility. Riley (1994) defined 12 unique activities performed by crews that require space, and a method to evaluate construction sequences for potential interference problems. Zouein (1995) developed MoveSchedule, a tool that alleviates space conflicts by changing activity durations and resource use, and also defined how space gets used and freed up as resources are consumed. This paper focuses on the necessary modeling characteristics of detailed crew level work spaces and places an emphasis on defining how these properties can be integrated with a realistic planning process. The development of detailed 4D models is feasible only through advancements in construction method modeling systems which permit a more detailed

4D modeling in trade sequencing and production planning 133

representation of how crews perform work than traditional CPM construction schedules. Aalami et al. (1997) demonstrate the advantages of detailed method modeling with the Construction Method Modeler (CMM), and how this technique permits the automated generation of detailed models of construction work sequences. This paper discusses issues related to the extension of the CMM to include models of construction work spaces at an equal level of detail, and assumes a planning technique will require a 3D model to be evaluated for conflicts at selected intervals. Adjustments could then be made to construction sequences or positions of stored materials and paths.

ATTRIBUTES OF CONSTRUCTION WORK SPACES The first step in defining attributes of construction work spaces is the definition of a modeling format and language to reference the different types of work spaces and their respective properties. Four key space needs by crews are physical work space, storage areas for materials, paths for material movement, and access points for unloading materials onto building floors. These four spaces will be the focus of this discussion, as they are most intensely related to production planning of multiple crews. The detailed modeling of discrete spaces, such as hazard areas, will be addressed later in a discussion of the types of potential conflicts that may exist between work spaces. Figure 6 illustrates an example of the spaces required to install two overhead fixtures. Each fixture requires a work space and a storage area at some nearby location. An access point and clear unloading space will also be needed. Connecting these spaces are paths between unloading and storage areas, and between storage areas and work areas. Paths between work spaces may also be required. For simplicity in this example, it is assumed that all work spaces will be rectilinear in shape. Property definitions Three categories of properties are used to describe construction work spaces as 4D objects: physical properties, which describes size, location, and density of work spaces; temporal properties, which associate the spaces to schedule data; and inherited, which associates spaces with product model objects and schedule activities. Table 1 provides a decomposition of these properties and more specific definitions for each. Properties of 4D construction space objects vary between different types of space usage. The number and types of spaces modeled may vary depending on the level of detail desired in the planning process, as will be discussed later. Initially however, it can be assumed that work spaces would be included in the 4D planning process. Certain material intensive properties may warrant storage spaces to be

134 D. Riley

Figure 6. Table 1.

Examples of spaces needed to install fixtures 1 and 2. Properties of 4D construction work spaces.

Properties Physical Size (x, y, z) Position (x, y, z) Density (1.0–0.1)

Description Length, width, and height dimensions. Sizes may be determined from a method database or assigned by a planner. 3D position of the base centroid, or endpoints of paths. Positions may be assigned by a planner or inferred from positions of related objects. Measure of object’s ability to share space. If the density of a space in the model exceeds 1.0 due to overlapping objects, a potential conflict may exist.

Temporal Start date Date in which object becomes active, or occupied. End date Date in which object becomes inactive, or unoccupied. Status (active, inactive) Determines if space is occupied during a selected time interval. Buffer Determines end date based on preset lag time. Inherited Object name Product model object with which the space is associated. Activity name CPM schedule activity of the product model object from which start dates of spaces are inferred.

4D modeling in trade sequencing and production planning 135

included, as well as considerations for paths. For example, electrical crews installing overhead fixtures or cable trays require space for ladders or scaffolding and perhaps large carts of conduit and materials. Work spaces and paths for this activity would differ from an electrical crew installing wall conduit or fixtures, where ladders and large material components would not be needed. Activities such as gypsum board installation are very material intensive, requiring an emphasis to be placed on paths for rolling carts of material or the use of pallet jacks. Among physical properties, the size of space objects is clearly a critical attribute, and one that will require more detailed definition for different types of crews. The position of work space objects varies depending on the nature of work required to install different materials. Figure 7 illustrates several different types of work space positions for different types of work, e.g. unit work, linear overhead work, linear work, and vertical work. Another variable physical property of space objects is the density of the space needed, which provides a mechanism to assess potential conflicts. Hypothetically, numerical values between 0.1 and 1.0 could describe the ability of a space object to share a portion or all of its physical space with other space objects concurrently. If the sum of overlapping spaces equals 1.0 or greater, a potential conflict may exist. The use of density values permits a conflict analysis tool to perform automated reasoning about different types of potential interference problems. For example, it might be acceptable for paths of low density to be shared by 2–3 different activities, however the work areas would most likely be considered high density to avoid conflicts between crews. Potential density ranges will be addressed in more detail later with a discussion on types of potential conflicts between activities. The temporal properties of different types of spaces are vital to the 4D modeling environment. At selected intervals, certain spaces will be occupied by crews or stored material. During these intervals, the space would be defined as having an “active” status. The start and end dates at which spaces are considered to have active/inactive status are related to respective product model objects and the associated activities in the CPM schedule. For example, an activity in a CPM schedule might be defined as “install 1st floor bath fixtures”. This activity would require a collection of work spaces in each bathroom area on the 1st floor, and a related storage area for bath fixtures. For these spaces to be representative of how crews use space, the status dates for individual spaces may require a shorter time interval from the actual activity dates. For example, if the case above referred to a hotel with 50 bathrooms and a two-week duration, it would be desirable to model the progress of a crew across a floor during the two-week activity duration. This would require that the work areas become “active” and “inactive” over the course of the two-week duration according to some defined sequence, as opposed to making all 50 of the modeled work spaces “active” for the entire two-week duration. Work spaces become active when objects are scheduled to be installed and inactive after the activity is completed. Buffers that extend the active period of a

136 D. Riley

Figure 7.

Examples of types of objects and associated work spaces.

work space may be used to accommodate clean-up time, etc. And, more importantly, to assist in production planning as discussed in Howell et al. (1993). It should be noted that the variable granularity for activity status described above, combined with the concept of buffers that extend the active status for activities,

4D modeling in trade sequencing and production planning 137

represent the two key variables needed for detailed production planning of multiple crews, which is the focus of this paper. Recall the example in Part One of this paper, which discussed the multi-story building project. In this case, the time interval of two weeks for an activity to take place on a building floor was insufficient to represent how crews actually moved across the floor plate. By modeling individual work spaces that were “active” in a defined sequence, it was possible to demonstrate the feasibility of two or more crews occupying a floor at the same time. This analysis made it possible to reduce the “batch size” of work that each crew performed before completed work areas were made available to subsequent crews. It was also necessary, however, to allow crews to proceed with a buffer of available work space between them to allow for uncertainty in the rate of progress for each crew. Tommelein et al. (1999) discuss the impact of uncertainty on the sequencing of multiple crews. If spaces for material unloading, movement, and storage are included, active and inactive times will take place at intervals surrounding respective object activity start and end dates, and also require that material deliveries be included as separate activities in the CPM schedule. For example, unloading areas and paths to storage would only be active at times immediately prior to and during material delivery dates. Storage areas would be active from the dates materials are delivered, until the respective activity is completed. Provisions may also be made for these spaces to diminish in size over time as described by Zouein (1995). Inherited properties may include a multitude of attributes from product models and schedules. At a minimum, however, the object name will indicate the product model object with which the space is associated, and the activity name will indicate the CPM schedule activity of the product model object from which start dates of spaces are inferred. This discussion presumes that a 4D model of the facility is already completed, and thus permits work spaces to inherit properties from the associated building component and the schedule activity. The next section will discuss in more detail the expected inputs and outputs to the planning process.

INPUTS AND OUTPUTS OF THE SPACE PLANNING PROCESS For space planning to be effective, it must be viewed as an investment of planning resources. For this investment to yield a return, it must be performed judiciously, on activities and projects which offer the opportunity to increase production with detailed adjustment to batch size and work sequences. To minimize the investment, the planning process should be efficient to perform and provide immediate, usable results. Existing space planning applications and construction projects were investigated, Riley & Sanvido (1997), to determine a realistic set of inputs and outputs to a space planning process in a 4D environment.

138 D. Riley

Inputs to planning Space planning may be considered as a technique to evaluate scheduling or sequencing alternatives to determine if spatial conflicts exist between different trades. Judgment must then be made as to the severity of potential spatial conflicts and potential course of action. For this reason, it is assumed that a 4D model of the facility to be analyzed is a required input to the type of detailed planning described in this paper. This 4D model consists of a 3D product model of the facility in which all objects to be planned are related to activities in a CPM construction schedule. Automated input: elements of 4D space planning that may be automated

• A 3D model of each work area to be considered in the planning process. Ideally, the development of a 3D product model should include the modeling of work spaces, in appropriate size, for each object in the model. An object oriented 3D modeling environment would permit these work spaces to be predefined for various components of a building and generated concurrently with the 3D model. Figure 7 illustrates various work areas that could be associated with product model objects for unit work spaces in isolated locations (A), overhead sections of ductwork, pipe, or cable tray (B), linear wall assemblies, in-wall plumbing, and electrical or wall finishes (C), or vertical sections of pipe risers, elevator shafts, and ductwork, etc. (D). • A property database for work spaces and associated spaces. The inherited properties of construction work spaces described earlier must be included as attributes of work space objects in the model. User input: elements of 4D space planning required by user

• A sequence in which model objects that are associated with unique construction activities become active would need to be determined by the planner. For example, a single schedule activity: “Install 4th floor light fixtures” requires further detail to define the order in which fixtures are installed. Subsequent duration adjustments to schedule activities would then adjust the rate at which fixtures would be installed and the rate at which a crew would move through the work spaces for that activity. It should also be noted that this sequence would most likely be changed to adjust work direction and flow rates. • Assigned positions of material access points and storage areas for discrete or multiple work areas. These positions are project specific, and cannot be inferred from the position of objects in the product model. From these locations the distance and proximity relationships of material paths between access, storage, and work spaces may also be calculated automatically. It would be advantageous to define access points for material loading and waste removal, and allow paths between these points and storage locations to be inferred by a planning tool.

4D modeling in trade sequencing and production planning 139

• Lead times or fixed dates for material delivery to storage spaces which permit the timing of material movement and required storage space to be calculated from the CPM schedule information of respective objects in the product model. For example, light fixtures may be loaded onto building floors two weeks prior to installation. Using lead times permits material loading dates to shift along with schedule changes, while fixed dates might be determined by manufacturing or logistical constraints on specialty materials. Planning outputs The ultimate product of 4D modeling and space planning should be a construction plan that is free from disruptive spatial conflicts. The automated detection of potential conflicts between work space, storage areas, and paths of different crews represents the primary goal of the 4D modeling process because it permits complex and long duration work sequences to be evaluated and re-evaluated after adjustments are made. Additional benefit would also be observed if automated reasoning about the severity of potential conflicts were possible. For example, a planner may choose to evaluate only specific types of conflicts or only those that occur for a planning interval of one week or more. As discussed earlier when the concept of density was introduced as an attribute of 4D objects, it might be possible for particular activity spaces to occupy the same location concurrently with little or no negative impact to production. Some materials may be stacked and some paths of low density may be shared by more than one activity. Rules for evaluating the severity of detected spatial interferences based on their durations and density would also be beneficial, and is the subject of current research. Table 2 identifies six types of spatial conflicts that would be beneficial to detect, and a range of suggested densities that may be used to assess the severity of the conflict. A “full” density range indicates that the type of conflict should be identified and resolved at all density levels of the spaces involved. A “variable” density range indicates that a conflict will only be identified between those types of spaces if the sum of the densities of those spaces exceeds an unacceptable level. Table 2.

Potential spatial conflicts between Activitya and Activityb.

Conflict type

Represents conflict between

Density range

(1) Worka–Workb Storagea–Workb Patha–Workb Storagea–Storageb Patha–Pathb Patha–Storageb

(2) Activitya and Activityb work spaces Activitya storage and Activityb work space Activitya path and Activityb work space Activitya and Activityb storage areas Activitya path and Activityb path Activitya path and Activityb storage area

(3) Full Full Variable Variable Variable Variable

140 D. Riley

Conflicts are identified only between spaces for different activities, e.g. storage space of gypsum board and work space for sprinkler pipe. It is assumed that conflicts between work, material, and paths for the same activities could be resolved by the crew performing the work and should therefore be ignored during planning. This convention would also loosen the detail needed for modeling work spaces, and more likely permit the automatic generation of work spaces based on a 3D product model. The detail in modeling work spaces can be expanded to include more robust descriptions of how space is used by activities, for example staging areas, hazardous areas, etc. Current research is focused on the use of 4D production models that include more specialized representation of work spaces, thus permitting a vigorous analysis of conflicts. Akinci & Fischer (1998) categorize such conflicts as those creating constructability problems, safety hazards, productivity problems, and potential damage to work in place.

IMPACT OF PLANNING ENVIRONMENT ON MODELING DETAIL Most planning efforts require a judgment to be made on the level of detail that must be included in the development of a realistic plan. When planning construction sequences, too little detail may result in critical elements of a work sequence being overlooked, or in a lack of allowances for uncertainty. Conversely, too much detailed planning can become tedious and may exceed an appropriate level considering inherent unexpected events and actions outside the control of management. Four aspects of 4D production models planning provide opportunities to adjust the level of detail in the planning process:

• Planning interval: A time frame that is planned and evaluated individually, e.g. hours, days, weeks. • Space usage: Spaces that should be modeled, e.g. work, storage, prefabrication, etc. • Activity type: Crew level operations that warrant planning. • Work zone: Specific areas of a facility that are likely candidates for congestion. Planning interval A precise model of the construction work environment is perhaps an unrealistic goal since work crews typically plan at daily or hourly intervals. However, it is reasonable to assume that significant progress can be made in most crew level operations in a week (five-day). Sequence planners are particularly interested in this progress, as it determines work space that may be freed up for sequential trades to move into and perform work. In addition, a one-week look-ahead is a realistic time frame for crew foremen to be comfortable when planning, as they

4D modeling in trade sequencing and production planning 141

often attend weekly coordination meetings. To support this planning, it is recommended that 4D modeling of works paces be performed with one-week planning intervals. Recent research on the Last Planner approach (Ballard & Howell, 1994) also utilizes one-week intervals to measure the effectiveness of planning. Space usage The use of a one-week planning interval suggests that the number and type of spaces modeled must also be limited to an appropriate level of detail. For example, spaces that are occupied for only a day or less might be omitted. Riley (1994) observed through case studies that unloading, storage, and work areas, as well as paths that connect these spaces, are the most commonly occurring spaces used by crews, and should thus be the focus of detailed planning efforts. As discussed above, robust planning environment should, however, permit unique spaces such as prefabrication areas or hazardous areas to be modeled on a discretionary basis by a planner.

CHARACTERIZATIONS OF HOW ACTIVITIES USE SPACE Observation of construction operations in progress has shown that storage and unloading spaces tend to occupy unique spaces for the longest time intervals and occur in positions that are determined by crews or management. Work areas occur in intervals that vary from minutes to several days or even weeks, and are complicated by the need to include buffers that allow for uncertainty. The positions of work areas are typically determined by the design, rather than by choice. Paths for materials are often unoccupied but must remain free from obstruction for discrete movement of materials, equipment, or workers. Which route the path will take is determined by the positions of work areas and assigned positions of storage and unloading areas. Activity type Constraints placed upon crews and materials during the phases of a building project vary. For example, excavation rates are largely determined by equipment capacities, while physical constraints such as weather and gravity affect the erection of the superstructure. It is limitations of space, however, that often become the driving factor during the enclosure and finish phase of a building. For this reason, the activities that take place during these phases are most likely to benefit from 4D modeling. In addition, work during these phases can often be completed in more than one direction or sequence, requiring more decisions during planning as well as more alternatives to sequence work. A 4D model developed for production planning would be beneficial to evaluate such alternatives. Investigations into work space congestion have identified MEP, and fire protection trades as the most vulnerable to interferences from congestion (Riley & Sanvido

142 D. Riley

1997). This study also noted the dramatic impact that the construction of wall studs have on available space. In particular, the completion of demising walls, tends to break up remaining work areas and cut off travel paths. Perimeter enclosure work, may also cut off access to spaces for loading of materials into the building. In summary, it is recommended that 4D modeling of construction operations focus on the following types of crews and materials: HVAC Electrical Plumbing Fire protection Carpentry Curtain wall

Hangers, ductwork, VAV boxes, related insulation; Overhead and in-wall cable tray, conduit, wiring, and fixtures; Hangers, overhead and in-wall pipe, fixtures, and testing; Hangers, overhead and in-wall pipe, fixtures, and testing; Layout of top and bottom track, wall studs, wall finishes; Perimeter studs, mullions, precast, masonry, and glazing.

Work zones Work zones can be used to prioritize the development of 4D models for production planning. A work zone is a separable area of a facility in which a series of crews perform sequential activities. Often a unique work zone exists in the core, perimeter, and open floor areas of a building. Separate work zones may also be defined by the geometry of a building floor, for example the two legs of an “L”shaped building might be defined as separate zones. Large floor plates may be partitioned into smaller, more manageable work zones by a planner to reduce the batch size of work between crews. Finally, the density of materials needed may warrant planning in particular areas of a facility, such as mechanical rooms, interstitial spaces, and plenums.

CONCLUSIONS This two part paper has examined the use of 4D modeling for production planning and sequence planning in construction. Part One proposed how feasible applications of 4D planning might be identified, and presented two illustrative case studies to demonstrate the benefits of 4D modeling to sequence planning. Part Two provided an organized description of the necessary attributes of 4D models for production planning, and suggested techniques to automate 4D model generation and analysis, as well as limit planning to appropriate detail. Trial implementation of 4D modeling on construction projects has yielded significant benefits to trade sequencing, and production planning. Currently the confined use of 3D models in building construction places limitations on the feasibility of 4D modeling. However, increased demands on MEP design coordination, and the development of advanced 3D modeling tools for these systems are

4D modeling in trade sequencing and production planning 143

eminent. As 4D modeling tools continue to evolve, and the creation of 3D models for construction coordination becomes more popular, more construction projects will have the opportunity to benefit from this tool. 4D modeling represents a technique to accurately model the dynamic construction work environment. Part Two of this paper has attempted to define attributes of construction work spaces as a first step towards utilizing existing 4D modeling tools to evaluate work sequences for potential conflicts between crews, stored materials, and paths. The pertinent issues that must be considered for this type of planning to be feasible are the detail of planning to be performed, the effect of the planning environment on the modeling methodology, and the types of activities that may warrant this detailed space planning. Using planning experience and case study observations as a reference, this initial discussion summarizes these key issues. Future research on the 4D modeling of construction work spaces is clearly needed to link these attributes more directly with appropriate planning tools, and to develop libraries of these attributes for specific types of crews, materials, and alternative methods of work.

REFERENCES Aalami, F., Haddad, Z. & Fischer, M. 1997. Improving project control by automating detailed scheduling. Proceedings of the construction congress V, Minneapolis, MN: 430–437. Akinci, B. & Fischer, M. 1998. Time-space conflict analysis based on 4D production models. International Computing Congress, K.C.P. Wang, ASCE, Boston, October 18–21: 342–353 Ballard, G. & Howell, G. 1994. Implementing lean construction: stabilizing work flow. Proceedings of the 2nd annual meeting of the International Group for Lean Construction, Pontificia Universidad Catolica de Chile, Santiago: 101–110. Reprinted in Lean Construction. Howell, G., Laufer, A. & Ballard, G. 1993. Interaction between subcycles: one key to improved methods. Journal of Construction Engineering and Management 119(4): 714–728. New York: ASCE. McKinney, K. & Fischer, M. 1998. Generating, evaluating and visualizing construction schedules with 4D-CAD tools. Automation in Construction 7(6): 433–447. Riley, D.R. & Sanvido, V.E. 1995. Patterns of construction space use in multi-story buildings. Journal of Construction Engineering and Management 121(4): 464–473. New York: ASCE. Riley, D.R. 1994. Modeling the space behavior of construction activities. Ph.D. Dissertation, Penn State University, University Park, PA 16802. Riley, D.R. & Sanvido, V.E. 1997. Space planning for mechanical, electrical, and fire protection trades in multi-story construction. ASCE construction congress V, Minneapolis, MN, October 1997: 102–109. Tommelein, I.D., Riley, D.R. & Howell, G.H. 1999. The parade game: impact of work flow variability on succeeding trade performance. Journal of Construction Engineering and Management 119(2): 266–287. New York: ASCE.

144 D. Riley Tommelein, I.D. & Zouein, P.P. 1993. Interactive dynamic layout planning. Journal of Construction Engineering and Management 119(2): 266–287. New York: ASCE. Thabet, W.Y. & Beliveau, Y.J. 1993. A model to quantify work space availability for space constrained scheduling within a CAD environment. Proceedings of the 5th international conference computer civil and building engineering: 110–116. New York: ASCE. Zouein, P.P. 1995. MoveSchedule: a planning tool for scheduling space use on construction sites. Ph.D. Thesis, Civil and Environmental Engineering Department, University of Michigan.

THE LINK BETWEEN DESIGN AND PROCESS: DYNAMIC PROCESS SIMULATION MODELS OF CONSTRUCTION ACTIVITIES E. Sarah Slaughter MOCA Systems, Newton, MA, USA

Abstract Recent research at MIT has developed a theoretical framework and specific methodologies, resulting in computer-based process simulation models for 12 selected construction processes, to systematically assess the construction process impacts of design and technology alternatives. Specifically, the research takes two distinct approaches. First, the research allows the explicit linking of the design process to the construction processes through the focus on the specific components and subsystems in the built facility. The second approach considers the system and inter-system relationships, both spatially and operationally, throughout the construction phases. The combination of these two approaches provides a means through which the impacts of particular designs and processes can be analyzed with respect to a specific system and for the project as whole, considering the primary, secondary and tertiary impacts. The research can have significant implications for improving the efficiency of the construction of facilities and the performance of these completed assets. In particular, the application of the methodologies developed in this research can improve the robustness of facility design and technology selections through the explicit evaluation of multiple alternatives for a specific project and its objectives. It can also provide a common basis of analysis for design and construction organizations, to collaborate and reconcile design and construction objectives. Keywords: dynamic process simulation, design/technology innovation, assessment of design and process

INTRODUCTION While many people refer nostalgically to the days of the masterbuilder, when a single person directed the design and construction of great human works, others 145

146 E.S. Slaughter

point practically to the current complexity of facilities and the greater speed with which they must be delivered. The implication is that, to achieve the current required performance levels, the activities of design and construction must become increasingly specialized and often clearly separated from each other. However, this separation between design and construction often appears to add significant delays and inefficiencies to the creation of constructed facilities. In particular, it often delays or eliminates the consideration of design and technology innovations that could significantly improve the functions of the completed facility, or improve the construction process performance through reducing duration or cost or improving worker safety. The link between design and construction is strained not only because of the specialization of the fields, but also because of the different organizational membership of the participants. Each design specialist (e.g. architects, structural engineers, and MEP designers) is often in a separate organization, and is also usually in markedly different organizations from the general and specialty contractors engaged to construct the facility. As a consequence, the designers cannot specify or determine the specific processes used to accomplish the design. Indeed, in many cases, the designers may explicitly exclude process specification, to avoid the associated liability and risk. During the training of designers, the focus is often significantly more on the development and analysis of designs within a specific area than on the processes required to accomplish those designs. In these cases, the designers could not realistically specify the construction means and methods needed to accomplish the designs, since they lack experience and expertise in these areas. Recent research at the Massachusetts Institute of Technology, funded by the National Science Foundation, directly addressed the relationship between the design and the construction process through the development of a theoretical framework and specific methodologies to systematically assess the construction process impacts of design and technology alternatives. Specifically, the research takes two distinct approaches. First, the research allows the explicit linking of the design process to the construction processes through the focus on the specific components, subsystems, and systems in the built facility. The second approach considers the inter-system relationships, both spatially and operationally, throughout the construction phases. The research created a modeling system to analyze design and process alternatives for each specific system and for the project as a whole. As a change in design or process is introduced, the primary, secondary, and tertiary impacts of that change can be traced throughout the specific systems, and among the multiple systems that constitute a facility. The modeling system can provide a common basis of knowledge for designers and contractors to collaborate and to improve the reconciliation of design and construction process goals. The results of the research were recently exclusively licensed by MIT to MOCA Systems to provide a commercial software system to the design and construction industry. The commercial system builds upon the theoretical basis developed in the research to provide a tool that architects, engineers, construction managers (CMs)

Dynamic process simulation models of construction activities 147

and specialty contractors can use to estimate the time, cost, and worker safety impacts of specific design and construction process alternatives for their projects.

BACKGROUND The traditional link between the design process and the realization of that design in construction has been through the cost and duration estimation. These estimates are primarily based upon previous experience, both within the specific company and throughout the industry (through various aggregated cost estimate resources). Members of the design and construction team draw upon their experience on past projects and their specific expertise to estimate the resources that will be needed to complete the project within a specified time. During the early conceptual development phase of the project, the team often estimates the cost based upon the area for each usage (e.g. square foot costs for an emergency room for a hospital). This general level of estimates provides a generous range (e.g. US $120 to 175 per square foot) that reflects the uncertainty of the project specifics at this stage. The construction duration is likewise estimated based upon experience with previous projects, and reflects the facility owner’s time objective. The uncertainty is reduced somewhat as the design is developed. At specific intervals, the design team will deliver the hardcopy or computer-based drawings to the CM or general contractor (GC) for adjustments in the early duration and cost estimates, with a resulting range in the estimates that continues to reflect the uncertainty in the design specifics. Depending upon the project, the designers and the CM/GC generally provide the completed design drawings to specialty contractors to obtain their cost estimates. The specialty contractors in the US are generally responsible for procuring the material and components, and so their cost estimates are based upon the purchase price of these materials and their estimate of the labor costs for the installation. The specialty contractors submit their estimates on the expected time and cost of their particular portion of the project to the GC or CM, and these estimates often become the basis for the contractual relationship between the specialty contractors and the GC. The CM or GC then compiles the specific system estimates, and creates a cost estimate and schedule for the project as a whole, using their experience and expertise to evaluate the requirements within and among the different system-specific processes. The link between the design and the actual construction is, therefore, a fairly tenuous connection. It rests upon the assumption that the average aggregated cost derived from previous projects can be used to predict with confidence the cost of a new project. It also assumes that the logical relationships between the construction processes have not changed significantly. In addition, because the duration of a process and the productivity of the resources are estimated with respect to all of

148 E.S. Slaughter

the activities associated with the system as a single quantity, and often aggregated across the complete project, specific design or process changes cannot easily be tracked to fully estimate their relative impacts. These assumptions can limit the accuracy of the resulting estimates for complex projects, and can inhibit the analysis of design and technology alternatives, particularly innovative systems that are new to the organization. Cost estimating systems To reduce the variance and improve accuracy, some firms have internally developed their own cost estimating systems that use historical cost data from previous projects (Peltz, 1996). The advantage of these internal cost databases is that they provide organization-specific data, reflecting the capabilities and past costs for specific types of projects. Many of these systems build upon the industry standard work breakdown structures, such as the Construction Specification Institute’s multi-digit classification code. The classification codes, however, often focus upon the physical elements after they are installed (such as plumbing tree assemblies) rather than the specific construction processes that were required to transform and customize the standard components for the specific project. The disadvantage of relying upon these internal databases is that they cannot accurately predict the costs for an activity with which the organization does not have experience. For example, a project that requires a different structural system from previous projects could not rely upon organizational experience, and would have to be estimated using the general industry cost databases or other generalized source. To complicate matters further, the extent to which this generalized data would need to be changed, and in what dimensions, for application to the specific project would not be known by the organization. New construction methods or equipment would be even more difficult to estimate using these systems, since these estimating databases are primarily arranged by system rather than construction process and could not easily accommodate specific process changes. The aggregation of the costs across a whole system, and the aggregation of the material and labor costs, can become increasingly inaccurate as the markets quickly change. Several cost estimating approaches start with a general count of major components for a system (e.g. number of bathrooms), and then multiply this count by a constant number that represents the assumed ratio between the material and labor costs (e.g. a multiplier of 2.5 for the plumbing system assumes that the labor cost will be 1½ times the material cost). Unfortunately, all material costs associated with a specific system do not rise uniformly, so this approach can obscure the overall cost impacts of specific cost increases. In addition, the ratio between the material and labor costs may not be uniform across all geographical areas, and can certainly change dynamically over time. The use of historical data at this aggregated level can actually increase the financial risk for the owner and project team for complex projects where the design and construction phases may span several years.

Dynamic process simulation models of construction activities 149

Duration estimating systems Estimates of the project duration often rely upon the general and organizationspecific data on resource productivity, coupled with the logical sequence between processes. The specialty and general contractors use their experience and expertise to plan the mobilization and deployment of their resources to accomplish the required tasks, while recognizing that certain tasks must precede or follow other tasks. For instance, the structural frame must be erected before the exterior enclosure units can be installed. The most common method in construction to identify and plan a project’s logical sequences and overall duration is through Critical Path Method (CPM) scheduling (Callahan et al., 1992). In this method, the basic construction processes are organized by their logical sequence, and these processes are then linked in a logical network that represents the sequential and parallel relationships. The duration for each process is estimated, and the overall duration is calculated by summing the duration for each system along the pathways through the network. The critical pathway is the series of sequential and parallel linked activities that has the longest duration. The advantage of this method is that it provides a strong mapping of the general tasks associated with the project, and can provide critical milestones that indicate the relative progress of the project, compared to the planned schedule. The crucial disadvantage of CPM with respect to new designs and technologies is that the aggregation of the construction processes into their general grouping can obscure changed logical relationships between specific aspects of the processes. For example, a CPM schedule would not be able to distinguish between two service design alternatives, such as one that has multiple shafts versus another that has a single centralized utility shaft, even though these alternatives have different requirements for access to the working surface by the different trades. In addition, the spatial factors can significantly influence the logical sequence of these activities, and can be only incorporated into the duration estimation through certain “rules of thumb” or heuristics. For example, a GC may decide to start installation of the curtainwall panels after the first five floors of the structure are erected, expecting that this lead-time for the structure is sufficient to ensure both worker safety and progress on the project as a whole. To capture more accurately the simultaneity of the construction processes and the importance of physical location on the progression of tasks, some firms break down some of the processes into specific spatial or other groupings (e.g. by floor or subcontractor). This disaggregation of processes can modify the sequence of tasks, and the overall duration is recalculated for the new sequence. The duration can also be adjusted by either limiting the resources for a specific set of activities or overall (for instance, on a constrained site, only a certain number of laborers can be accommodated) or increasing the resources for each specialty contractor. While these changes can better reflect the actual progress of construction on a project, they are often difficult to formulate accurately, and must be recreated for each project. In addition, the changes to the process itself that may be involved with

150 E.S. Slaughter

a design or technology innovation may not be effectively incorporated into these modified CPM schedules. Thus, current project management techniques estimate the project cost based upon previous projects and in-house expertise, and separately estimate the duration of the project using the logical sequences of tasks, with duration estimates based upon past projects, as embodied through the CPM scheduling approach. The list of materials and components to be procured for the project are generated after the design is completed (in whole or for each section) and, in the US., is generally the responsibility of the specialty contractors who will actually perform the work. These techniques respond to the unique requirements for each project, in the design, site, and resources, through attention to common units across projects, modified for each project. Unfortunately, these cost and duration estimating systems rely upon highly concentrated expertise to accurately predict the duration and cost for specific projects. In addition, the current duration and cost estimation techniques often require a significant commitment of internal resources to labor intensive activities to generate the estimates for a specific design and set of construction means and methods. These high labor requirements often preclude their application to the evaluation of alternatives, and instead are used to focus attention on the planning and control stages of the project. New systems: 4D CAD and queuing-based simulation Recent research focuses on improving the link between the design and the construction processes. The visual representation of the construction sequence conveys both the spatial relationships among the components being assembled and the overall project progress. Using a computer-aided design (CAD) file for the specific project, the designers can understand the spatial implications of their decisions, and the construction planners can demonstrate the stages to both specialty contractors and clients (Fischer & Aalami, 1996). Certain aspects, such as interference of components, can also be examined through the CAD animation programs (Stouffs et al., 1993; Vanegas & Opdenbosch, 1994). Other analysis systems have sought alternative methods to improve the efficiency and the effectiveness of resource utilization, and the duration estimation for construction projects. Several methods exist using queuing-based simulation. The objectives of these analyses are to optimize the use of the resources and to decrease process duration. In these models, the specific tasks required for a process are linked to the required resources. The capacity of these resources defines the rate at which the specific tasks are performed. The duration for the process time is based upon a probability distribution of times for each activity with the associated resources (Cheng & O’Connor, 1993; Vanegas et al., 1993; Alkoc & Erbatur, 1997; Shi & AbouRizk, 1997; Chehayeb & AbouRizk, 1998; Oloufa et al., 1998; Senior & Halpin, 1998). Other queuing-based simulation models are being used to explore alternative resource allocation schemes, site alignment, and other changes in activity

Dynamic process simulation models of construction activities 151

relationships (Ioannou & Martinez, 1996; Tommelein, 1998). Emerging simulation models using Petri nets and neural networks are also being used to chart the progress of construction activities (Chao & Skibniewski, 1995; Wakefield & Sears, 1997; Shi, 1999). Using these visualization and simulation models, the construction planners can explore different combinations and distributions of equipment and other resources, similar to the analyses performed in many manufacturing facilities. These analyses can also explore the impacts of increased or decreased resources on overall progress, as well as the sensitivity of the process to spatial relationships among the design elements. The advantages of these models are that they represent the flow of tasks and activities and their associated resources, and can significantly improve the understanding of the factors effecting construction progress. The disadvantages are that they often have to make assumptions on certain aspects of the construction activities, such as the distribution of the production times, which can limit their applicability for actual construction projects (Schexnayder, 1997). In addition, the task/resource specification must often be recreated for each new task, as well as each new project. New process simulation systems In several industry sectors, the production process specifics are increasingly recognized as the primary factors defining the economic and technical feasibility of each new product offering. Simulation of these production processes can provide a powerful tool to assess the full context of design and process changes, and to evaluate alternative inputs, processes and methods (Glasscock & Hale, 1994). In these dynamic process models, the focus of modeling is on discrete events. For example, the location and time at which specific inputs are introduced into the process and diverted through the process can determine both the total output quantity and the rate of output. The objectives of these simulation models are to better understand the complex interactions between the specific processing stages and their outputs, and to explore the full system impacts of changing either the inputs or processes. Dynamic process simulation has clear applicability to the construction industry. Construction processes are complex and dynamic, and the outputs from one stage can be seen as inputs for subsequent stages, in the transformation and aggregation of components and systems into a built facility. Process simulation could provide significant benefits to the construction industry, since the construction activities define both the performance of the construction project and of the completed facility. The modeling system could provide a means to examine the primary, secondary, and tertiary impacts associated with the introduction of changes into complex inter-dependent processes. However, the context of the construction industry differs significantly from the chemical process industries, and the processes themselves do not share many similarities. Unlike chemical processing, construction resources can be assigned to many different types of tasks, at many different stages in the processes rather than being

152 E.S. Slaughter

dedicated to a specific set of activities. In addition, unlike the continuous processing that is required in chemical plants, construction activities are performed within a project, with a defined beginning and end, to accomplish specific objectives. Each construction project is by its nature unique, consisting of new combinations of components, systems, and resources to create the facility. The industries also differ in their performance measures of the processes. In general, process and manufacturing industries (and their simulation tools) focus on the process time per unit and overall process throughput (i.e. the rate at which the units are processed through the complete cycle). In contrast, the performance of a construction project is judged by the cost and duration needed to realize the facility design. This research developed a construction process simulation system to address the shortcomings of existing estimating techniques, and to provide a tool to directly calculate the time, cost, and worker safety impacts of design and technology alternatives. Taking a different approach than the current research in 4D CAD and queuing-based simulation, this research is strongly complementary to the ongoing efforts in those areas, providing a different view on the link between design and construction, focusing particularly on the assessment of specific alternatives within a particular project. The research built upon the process modeling approach used in the chemical industry, and extended the theoretical framework to accommodate the specific requirements of the construction industry.

RESEARCH APPROACH Recent research at MIT has developed the theoretical basis and specific methodologies to systematically assess the construction process impacts of design and technology alternatives. The objectives of the research were to characterize construction processes by system and material, and to assess design and technology alternatives within systems and across systems in the project as a whole for their impacts on duration, cost, and safety (Slaughter & Eraso, 1997; Slaughter, 1999). The specific research methodologies culminated in the creation of specific computer-based dynamic process simulation models that can be used to evaluate specific designs and technologies for particular projects. This research resulted in the characterization and model development for 12 system and material-specific construction processes. The models are stored in a “library” of system and material-specific processes, and can be accessed and used for any size or type of construction project that employs the process (Table 1). The four general systems are the structure, exterior enclosure, services, and interior finish. The inter-system links are incorporated in a meta-model for the whole project, with status and information tracking across each of the systems to represent overall progress.

Dynamic process simulation models of construction activities 153 Table 1.

Construction process models completed or in process.

System

Material-specific model

Structure

Steel Cast-in-place concrete Light wood framing Precast concrete panels Glass/metal curtainwall HVAC (heating/ventilation/air conditioning) Hot water heating Plumbing Fire protection Electrical Interior walls Suspended ceiling

Exterior enclosure Services

Interior finish

PROCESS FLOW -Tasks -Sequences -Unit Worked Upon -Decisions

PROJECT SPECIFICS -Design Attributes -Resources -Production Rates -Site Conditions

SIMULATION MODEL

PROJECT DYNAMICS -Simultaneous tasks -Shared resources -Constraints - Logical - Technical - Regulatory

Figure 1.

Conceptual framework for construction process modeling system.

The modeling system captures the common aspects of each construction process (i.e. tasks and associated resources and the related production rates by specific entity) while remaining completely adaptable to the elements that differ between projects (i.e. design attributes, number and character of resources), and the dynamic aspects of each project (e.g. site management strategies) (Fig. 1). Each process model is applicable across all types and designs of projects and can be immediately applicable to a new project or design without recreation or respecification of the tasks and their resources, and the sequence of tasks. It also captures the aspects of the design configuration of the components, subsystems, and systems that affect the process.

154 E.S. Slaughter

Structure of the models The models are configured to use the detailed design attributes for a specific project as input. The modeling system uses this data to chart the path of each component through each relevant task in the process according to its characteristics, and to map the progress achieved in each subprocess by logical sequence and location to ensure progress for the process as a whole. Status information by subprocess, process and location is also transferred across the system-specific processes to capture the dynamics of the whole project. The output for the model is the duration of each subprocess and process by location, the duration-based costs of the on-site resources, and an index measuring the exposure of workers to dangerous conditions. The resources are pulled from the pool of resources as each unit progresses through the tasks and the resources become available. It is also possible to set the priority level of the task or set of tasks to establish the distribution order of the resources. Establishing the resources with each task or small set of tasks ensures that relative progress is maintained across all tasks and subprocesses. The structure of the models is scalable, being able to accurately model small and large construction projects. The meta-model of the whole project is also scalable, and can accommodate either a few or many specific processes. The modeling system is also extendable, since additional processes can be added to the meta-model. The simulation results are the time taken to perform each task and each sequence of tasks based upon the availability of resources and the specific design. These results can be aggregated by system-specific process and across the project as a whole to directly estimate the project duration. The on-site resource costs, which include the labor and equipment, are calculated directly from these time estimates. The duration-based cost estimates calculate the cost of having those resources on site to perform the tasks. The index of the exposure of workers to dangerous conditions is also directly calculated from the task duration. The OSHA-identified causes of worker injury are matched to the exposure of the workers to each injury cause for each task. The index scales the exposure of the workers to each injury cause by the amount of time workers are performing the tasks (Slaughter & Eraso, 1997). The modeling system clearly links design and technology alternatives to the details of the processes, to most effectively represent all of the project performance impacts. The models can also incorporate organization-specific knowledge and expertise, and provide a realistic basis in which to evaluate alternatives for new processes. Methodology The research methodology consists of two parts. The first portion is the development of a systematic methodology to characterize complex construction processes. The second portion of the methodology is the development of a set of compatible, consistent dynamic process simulation models. Construction process knowledge is usually not documented in written or pictorial form, but is rather obtained through experience, including hands-on training and

Dynamic process simulation models of construction activities 155

participation in numerous projects. Therefore, creating a detailed characterization of a construction process, including identification of each task in its proper sequence with required resources and related production rates could not rely upon existing documentation. The methodology to characterize the construction processes was developed to ensure the accurate representation of the on-site activities. The characterization methodology relied upon actual field data and in-depth interviews with personnel at design and construction companies, similar to the methodology employed in other research (Thomas et al., 1990). The process characterization was then translated into the dynamic process simulation model. In-depth interviews with personnel in general and specialty contractor companies, as well as designers, owners, and other knowledgeable parties, provided data for the specification of relevant design attributes, common types and quantities of resources throughout the US, and related production rates. They also provided critical expertise in assessing the validity of the completed models. Over 75 companies were involved in this stage of the research, contributing to the initial characterization of the process, and verifying that the process characterization is complete and accurate. In addition, over 100 construction sites were visited to conduct direct field observations (Table 2). The process characterization was translated into a computer-based process simulation model. The objectives of the model development were to explore the representation of the tasks and activities, including the ease by which the models can be modified to represent a specific project, and to accurately model the design and technology alternatives. During the research at MIT, the simulations were run using commercially available simulation software, SimProcess™, which was developed to model business processes. SimProcess is a hierarchical, discrete event simulation package, which provides basic simulation functions (e.g. gate, split, or join) for easy assembly into subprocesses. However, because this software was primarily developed to represent cyclic processes on standard items, the development of the construction process modeling system required significant modifications of the simulation environment in multiple areas to better represent the distinctive characteristics of construction activities.

Table 2. Construction sites visited (1993–1998). Type of facility

Number of sites

Institutional Office Residential Retail Industrial Other Total

24 29 21 10 6 9 102

156 E.S. Slaughter

The model results for a specific building design were sent to industry experts to assess the accuracy of the results, and in every case the model results were within 1–5 % of industry duration and cost estimates. The specific system design for each construction process was also defined to reflect a representative building type, and was reviewed by relevant designers. The development of each process simulation model has taken 18–24 months, including the estimation of results for the prototype building, and calibration to accurately reflect actual project performance. These results become the baseline against which selected scenarios are compared, to analyze the relative system-level impacts of alternative designs and technologies. The scenarios can also be used to analyze the attributes of specific proposed design and technology innovations, to identify certain characteristics that may increase or decrease the expected benefits and to influence the development of more appropriate construction innovations. In addition, the scenarios can be used to explore potential complementary aspects of alternatives, when a combination of innovations may provide greater benefits than the sum of the individual innovations. Design and process links The modeling system provides explicit links between the design and the construction process at many different levels. These levels include the specific components, their subsystem and systems, and the specific configuration of the components and systems spatially and temporally for a particular project. It also provides a common basis in which the design and construction team can plan and coordinate the construction activities, and reconcile the design and construction objectives. As clients increase their requirements for high quality facilities obtained at a reasonable price and within shorter durations, design and construction professionals increasingly recognize the incentives to work together. The design is linked to the process through the type and quantities of the components, and also through their configuration. For instance, if a pipe run is a straight line between the riser and the fixture, there are fewer connections than a pipe run with multiple bends. As a result, the straight run will take less time to place and connect, which will influence the progress of the process as a whole. In some cases, the progress on that particular location in the building may be critical to the progress for subsequent processes, and the time to place and connect the pipe lengths for the run can directly influence the overall project performance. The spatial relationships of the design elements also influence the overall time, cost, and worker safety of the project. For example, for a floor of an office building, the restroom facilities could be centralized in one area, or separated into several zones. These design alternatives could be used for one, several, or all floors, and the relative costs and time for each design alternative will differ, depending upon the design alternative selected for each system for each floor and the relationship to the contiguous floors. Since the plumbing must be installed before the interior finish can be placed for those rooms, and the interior finish must be complete before the fixtures are installed, the design layout of the facilities determines the

Dynamic process simulation models of construction activities 157

rate at which the restroom facilities are completed for each floor, which in turn determines the completion time for the project as a whole.

LINKED DESIGN AND PROCESS CHANGES: SIMULATION RESULTS The duration, cost, and safety results are calculated for a specific project, for each design alternative or set of alternatives. Examples of the representation of the links between the design and process are discussed here for several different systems (i.e. structural system, exterior enclosure system, and services system) with the simulation results to demonstrate the applicability of the research approach to actual construction projects. The baseline for the analysis is a five-story office building, with a footprint of 30.5 by 38 m (100 by 125 feet), with the floor-to-floor height of 3 m (10 feet) and a bay size of 7.6 by 7.6 m (25 by 25 feet) (Fig. 2). The structural system described here is cast-in-place reinforced concrete columns and beams with a two-way slab of 200 mm (8 inches) (Carr, 1998; Slaughter & Carr, 1999). The exterior enclosure system analyzed here is glass curtainwall system (Attai, 1997; Slaughter, 1997). The service system described here is a domestic plumbing system, including hot and cold potable water with a drain, waste, and vent outflow (Murray, 1999). Cast-in-place reinforced concrete structure results Significant on-site resources are required to construct a cast-in-place reinforced concrete structure. While this structural type performs very well under extreme loads (such as seismic conditions), it does take many direct labor hours to fabricate 7.6 m

Columns

Beams

7.6 m

Service Core

Figure 2.

Floor plan for prototype building.

158 E.S. Slaughter

the reinforcing steel bar cages, build the formwork, cast the concrete and wait for it to cure, and then strip the formwork and finish the surfaces. Certain innovations in material, equipment, and process have significantly changed the construction process for this system (Carr, 1998). In addition, changes in the design of the system, including the components, connections and configuration, can alter the construction process and the time and cost to construct it, as well as the exposure of workers to dangerous conditions. The simulation model of CIP reinforced concrete structure was run for the prototype building, assuming standard components, connections, and configurations. Among the standard methods included in this simulation was the use of prefabricated slab and beam formwork. The model results estimated that this building structure would take 45 days to complete to the last pour, with an additional 10 days to final full-strength curing, at a cost of approximately US $220,000 (including direct and indirect costs, but excluding the GC overhead) (Table 3) (Fig. 3) (Carr, 1998). Table 3.

Project results for cast-in-place concrete structure for prototype building. Standard method, design

Precast concrete members as stay-in-place forms

Duration Last pour Complete Total cost (US $) Danger index

45 days 55 days 217,491 1.0

44 days 54 days 132,173 0.58

4 3

Pour beam, Slab Beam, Slab rebar Column rebar

1

Improvement 2% 39% 42%

Finish Strip Pour beam, Slab Beam, Slab rebar Column rebar Finish Strip Pour beam, Slab Beam, Slab rebar Column rebar Finish Strip

2

Floor

5

Performance

Finish Strip Pour beam, Slab Beam, Slab rebar Column rebar Finish Strip Pour beam, Slab Beam, Slab rebar Column rebar 0

10

20

30

Days

Figure 3.

CIP concrete structure for prototype building.

40

50

60

Dynamic process simulation models of construction activities 159

Throughout the CIP concrete industry, interest has been increasing on the potential savings that may be available from use of prefabricated and/or stay-in-place formwork. Therefore, the design and process specifics for the prototype building were modified to analyze this alternative. Specifically, the structural slab and beam members are redesigned to include precast concrete panels and members with imbedded reinforcing steel. The precast concrete elements are supplemented by field-placed reinforcing steel and concrete to reach the full depth and strength required to meet the design load requirements. The primary impact of the use of the precast concrete members (as stay-in-place forms) is to eliminate the stripping of the formwork. However, because this activity is performed simultaneously with other construction activities, the reduction in process duration is minor, at approximately 2% reduction in duration (Table 3). Interestingly, the major time reduction is primarily due to the elimination of placing steel reinforcing bars in the bottom layer of the members. The secondary and tertiary impacts of this alternative are much more significant, however. Since the stay-in-place formwork eliminates several steps, the formwork crew can be re-assigned to other activities, and can complete those activities in less time (and, therefore, for almost 40% lower cost). This reassignment of workers to different tasks, and particularly the elimination of the stripping stages, significantly reduces the exposure of the workers to dangerous conditions (by 42%), since they are no longer required to move large panels of formwork from under the cast slabs and beams during the formwork stripping stage. These impacts from the design alternative of precast concrete members as stayin-place formwork would not be readily apparent without the detailed analysis link of the design to the specific process activities and resources, available through the process simulation models. Domestic plumbing service system results Domestic plumbing systems, which supply hot and cold water and remove the waste, are required for all occupied facilities. The supply and drain/waste/vent system form a closed loop system, and the point of transfer between the supply and return systems is the plumbing fixtures (e.g. sink, toilet, or appliance). The vertical supply units are called the risers, the vertical return units are called stacks (and usually rely upon a gravity-based flow), and these subsystems each have horizontal piping to and from the fixtures. The vent element is included with the return system to equalize the pressure in the system to allow the waste to flow freely down the stacks. Certain building codes require different configurations of vent elements for different types and configurations of fixtures, and these venting requirements can consist of several separate stack elements, as well as the activities needed to install and connect them. The plumbing design for the prototype building has one major shaft for the stacks and risers. Each floor has two restrooms placed in close proximity to the shaft, each with three sinks and four toilets and/or urinals and a drinking fountain in the

160 E.S. Slaughter

hall. The simulation model of plumbing installation was run for this building design, and the model results estimate that it would take 36 days, at a cost of approximately US $105,000 (including direct and indirect costs, but excluding the GC overhead) (Table 4) (Fig. 4) (Murray, 1999). One design alternative is to install a mechanical aeration device within the waste return system to actively equalize the pressure in the waste pipes and stacks. This device could eliminate the venting units for a medium-sized occupied building. Since the vertical venting units are usually integrated with the waste stacks, and the horizontal venting loops are usually only a small portion of the horizontal pipes, the primary impact of this design alternative is minimal, with small reductions in the plumbing installation process. The secondary and tertiary impacts of this alternative are quite significant. Because the installation and connection of the venting units is usually performed in close physical proximity to the other vertical and horizontal piping elements, the elimination of these units can considerably simplify the placement and connection of the remaining units. This simplification can reduce the time and cost of the process significantly, allowing the workers to progress more efficiently for each Table 4.

Project results for plumbing system for prototype building.

Performance

Standard Method, design

Aerator

Improvement

Duration Total cost (US $) Danger index

36 days 104,976 1.0

30 days 89,100 1.0

17% 15% 0%

Figure 4.

Plumbing results for prototype building.

Dynamic process simulation models of construction activities 161

floor and throughout the building. For the complete building, the mechanical aerator reduces both the construction duration and the on-site costs by over 15% (Table 4). Although the aerator does eliminate several physical components and the activities required to install them, it does require placing and connecting the mechanical aerator itself, and so does not reduce the exposure of workers to dangerous conditions. Glass curtainwall exterior enclosure results Design alternatives in one system can also alter the performance and progress for other systems. The exterior enclosure system must provide protection from the weather, including wind and rain, and it is usually installed as soon and as quickly as possible, to create a weatherproof environment for the installation of the services (e.g. electrical, plumbing, and heating systems). A curtainwall exterior enclosure system does not carry any of the dead or live loads of the building, including its own weight, and instead is hung off of the building structure. Glass curtainwall panels can be hung from the structure, and usually incorporate transparent elements that function as windows within the panel units. A common configuration is for each panel to span from floor to floor, and these panels are usually positioned and installed from the interior of the building (rather than from an external crane or platform). The simulation model for glass curtainwall installation was run for the prototype building, assuming a curtainwall system with panels measuring 3.0 by 1.8 m (10 by 5 feet), and weighing approximately 1,000 kg (2,200 pounds). The model results estimate that this building’s glass curtainwall system would be completed in 54 days, at a cost of approximately US $102,000 (including direct and indirect costs, but excluding the GC overhead) (Attai, 1997) (Table 5). A design and construction process alternative that changes both the specific design of the panels, as well as the construction means and methods, is a panel lift shuttle system (Attai, 1997). This system uses a platform that surrounds the building periphery. The glass curtainwall panels are fitted into this shuttle platform, and each panel is attached to the adjacent panels. The shuttle, which contains a full floor’s worth of panels, is then lifted into location (where each floor is set starting from the top of the building working down), and the ring of panels is then attached to the building frame. This alternative requires redesign of each panel, particularly in its connections to contiguous panels and its connection to the building frame. It also introduces a new piece of construction equipment onto the site (the shuttle platform).

Table 5. Project results for glass curtainwall exterior enclosure system for prototype building. Performance

Standard method, design

Shuttle system

Improvement

Duration Total cost (US $) Danger index

54 days 102,000 1.0

32 days 98,370 0.5

41% 45% 51%

162 E.S. Slaughter

The primary impact from this alternative is to reduce the time required to place and connect each panel, since the work is now performed on the ground within the shuttle platform rather than singly for each floor for each panel. The caulking and external sealant for the panel connection is also performed while the panels are on the ground within the shuttle system, rather than being performed on a swinging scaffold on the building face after erection. The secondary impacts from this innovation are the improved efficiency of the workers, since the tasks are performed more quickly, and the significant reduction in their exposure to worker conditions from the ground performance of the work. This alternative for the glass curtainwall system reduces the duration by over 40%, reduces the cost by 45%, and reduces the worker exposure to dangerous conditions by over 50%. The shuttle platform, therefore, can have significant impacts on the speed and cost of the curtainwall erection, but it can also change the overall performance of work. Specifically, because the shuttle system requires the curtainwall installation from the top of the building working down, it therefore requires that the structural erection be complete before the curtainwall erection can be initiated. In contrast, many projects prefer to initiate the exterior enclosure erection while the structure is still being erected, to reduce overall project duration. This alternative, and the other design and process alternatives analyzed in this research, would need to be evaluated for each specific project to test the direct and indirect implications of these selections with respect to the design specifics and project dynamics. The theoretical framework and specific methodologies developed in the research provide a context in which the analysis of these alternatives links directly the building design and the activities to realize the design. The modeling system tracks the primary, secondary, and tertiary impacts of design and construction processes changes through each building system and construction process, and monitors the inter-relationships between the construction processes, both spatially and functionally, to calculate overall project progress.

CONCLUSIONS The theoretical framework developed in this research provides a unique capability to analyze the time, cost, and worker safety impacts of design and technology alternatives. It explicitly links the design to the processes needed to realize that design, and tracks the inter-dependency in the performance of the construction activities both spatially and operationally. It strongly complements existing techniques in the construction industry that link the design with the construction process in cost estimation, in scheduling, and in the allocation of resources. The MIT research builds off of developments in cost estimation, scheduling, CAD visualization, and queuingbased simulation models, and employs key aspects of process modeling from other industries to represent the dynamic aspects and unique attributes of construction

Dynamic process simulation models of construction activities 163

processes. The general objectives of the research are to improve the efficiency and effectiveness of the construction of facilities, specifically through improving the link between design and the construction process. The immediate potential link of design to construction is use of the modeling system by designers, CMs, and general and specialty contractors to analyze design and technology alternatives to improve their planning and control capabilities. These companies could use the models for a specific project to analyze their processes and understand the nature of their activities and the impact that design and technology alternatives would have on their operations and costs. As the design and process links improve, the variation and uncertainty associated with each project could diminish, thereby reducing the actual and perceived risk associated with most construction projects. In addition, the new links between design and process can create an environment in which the costs and benefits associated with each design and technology alternative are better understood and can be explicitly distributed among the project participants. Innovations in design and technology, as well as new design configurations, can be assessed directly for their impacts on the construction process for each specific construction project. These analyses provide a critical feedback system for design and construction organizations to learn across projects to improve their internal competencies. It also provides a means to assess and improve the designs, components, systems, and process of construction projects, as well as to improve the efficiency of the design and construction processes.

REFERENCES Alkoc, E. & Erbatur, F. 1997. Productivity improvement in concreting operations through simulation models. Building Research and Information 25(2): 82–91. Attai, L.M. 1997. Simulation to assess exterior enclosure innovations. Master of Science, Massachusetts Institute of Technology, Cambridge, MA. Callahan, M.T., Quackenbush, D.G. & Rowings, J.E. 1992. Construction project scheduling. New York: McGraw-Hill. Carr, M.N. 1998. Simulation to assess cast-in-place concrete construction innovation. Master of Science, Massachusetts Institute of Technology, Cambridge, MA. Chao, L.-C. & Skibniewski, M.J. 1995. Neural network method of estimating construction technology acceptability. Journal of Construction Engineering and Management 121(1): 130–142. Chehayeb, N.N. & AbouRizk, S.M. 1998. Simulation-based scheduling with continuous activity relationships. Journal of Construction Engineering and Management 124(2): 107–115. Cheng, M.Y. & O’Connor, J.T. 1993. Simulation analysis of process piping construction. Automation and robotics in construction X. Elsevier Science Publishers: 519–526. Eraso, M. 1995. Methodology for the economic assessment of construction innovations: simulation of structural steel erection. Master of Science, Lehigh University, Bethlehem, PA.

164 E.S. Slaughter Fischer, M.A. & Aalami, F. 1996. Scheduling with computer-interpretable construction method models. Journal of Construction Engineering and Management 122(4): 337–347. Glasscock, D.A. & Hale, J.C. 1994. Process simulation: the art and science of modeling. Chemical Engineering 101(11): 82–89. Ioannou, P.G. & Martinez, J.C. 1996. Comparison of construction alternatives using matched simulation experiments. Journal of Construction Engineering and Management 122(3): 231–241. Murray, J. 1999. Simulation to assess innovations in the installation of plumbing and fire protection systems. Master of Science, Massachusetts Institute of Technology, Cambridge, MA. Oloufa, A.A., Ikeda, M. & Nguyen, T.H. 1998. Resource-based simulation libraries for construction. Automation in Construction 7: 315–326. Peltz, C. 1996. Model estimating in design-build. Construction Business Review 6(2): 48–50. Schexnayder, C.J. 1997. Analysis of earth-moving systems discrete-event simulation: discussion. Journal of Construction Engineering and Management 123(2): 199. Senior, B.A. & Halpin, D.W. 1998. Simplified simulation system for construction projects. Journal of Construction Engineering and Management 124(1): 72–81. Shi, J. & AbouRizk, S.M. 1997. Resource-based modeling for construction simulation. Journal of Construction Engineering and Management 123(1): 26–33. Shi, J.J. 1999. A neural network based system for predicting earthmoving production. Construction Management and Economics 17(4): 463–471. Slaughter, E.S. 1997. Computer-based process simulation of construction activities. Construction Congress. Minneapolis, MN. ASCE, New York. Slaughter, E.S. & Eraso, M. 1997. Simulation of structural steel erection to assess innovations. IEEE Transactions on Engineering Management 44(2): 196–207. Slaughter, E.S. 1999. Assessment of construction processes and innovations through simulation. Construction Management and Economics 17(3): 341–350. Slaughter, E.S. & Carr, M.N. 1999. Dynamic process simulation model of cast-in-place concrete process. Structures Congress. New Orleans, LA. ASCE, New York. Stouffs, R., Krishnamurti, R., Lee, S. & Oppenheim, I. 1993. Construction process simulation with rule-base robot path planning. Automation and robotics in construction X. Elsevier Science Publishers: 495–502. Thomas, H.R., Horner, R.M. & Smith, G.R. 1990. Procedures manual for collecting productivity and related data of labor-intensive activities on commercial construction projects: Structural steel. Pennsylvania Transportation Institute, University of Pennsylvania, State College, PA. Tommelein, I.D. 1998. Pull-driven scheduling for pipe-spool installation: simulation of lean construction technique. Journal of Construction Engineering and Management 124(4): 279–288. Vanegas, J.A., Bravo, E.B. & Halpin, D.W. 1993. Simulation technologies for planning heavy construction processes. Journal of Construction Engineering and Management 119(2): 336–354. Vanegas, J.A. & Opdenbosch, A. 1994. Using simulation and visualization technologies to strengthen the design/construction interface. 1994 winter simulation conference: 1137–1144. Wakefield, R.R. & Sears, G.A. 1997. Petri nets for simulation and modeling of construction systems. Journal of Construction Engineering and Management 123(2): 105–112.

ACKNOWLEDGING VARIABILITY AND UNCERTAINTY IN PRODUCT AND PROCESS DEVELOPMENT Iris D. Tommelein Construction Engineering and Management Program, Civil and Environmental Engineering Department, University of California, Berkeley, CA, USA

Abstract Four-dimensional (4D) models describe product geometry in three dimensions and process time in one dimension. Most 4D models in use today do not explicitly reflect that the values taken on by geometric and temporal variables may not be known exactly, but can vary for a variety of reasons such as human indecision or physical tolerances. Moreover, practitioners involved in product and process development for the architecture/engineering/construction industry need to reason about numerous other variables in addition to those regarding time and space. As a result, 4D models support their endeavors only in part. This paper describes sources of variability and uncertainty in product- as well as process-definition. It then argues for developing and using representations that lend themselves to studying the impact of variability and uncertainty on the integrated product- and process-development process. Reasoning is biased by the conceptualization and representation people choose to adopt. People’s problem-solving abilities are biased in a similar way. A world in which no variation or uncertainty is recognized gets modeled deterministically, often by means of expected values. Models based on averages are unrealistically optimistic. Single values reflecting early commitment tend to lead to process iteration. This paper therefore argues for the creation and study of extended 4D models, referred to as 4D⫹, that explicitly represent alternative sets, interdependence, variability, and uncertainty. Examples illustrate what features are necessary in 4D⫹ representations to make it possible to perform concurrent engineering and to manage the production system that supports integrated product and process development. Keywords: integrated product and process development, 4D modeling, lean construction, supply chain management, design, production planning, simulation, interdependence, uncertainty, AEC, concurrent engineering, set-based design

INTRODUCTION Four-dimensional (4D) models describe product geometry in three dimensions and process time in one dimension. Though generally well understood, 4D refers 165

166 I.D. Tommelein

to a range of different conceptualizations and associated models. This will already be clear to the reader who has reviewed the various papers of this book.

HISTORIC BACKGROUND 4D models are the product of a long evolution in computer modeling. While early computer input and output was type-based, the need for and benefits of providing graphical output soon became apparent. Three-dimensional (3D) viewing was made possible by “Evans and Sutherland (who) demonstrated a head mounted stereo display as early as in 1965” (mentioned in Issa et al., 1999) in order to create an environment of virtual reality. Vector-based “graphical” monitors were created to avoid the jagged lines displayed on low-resolution type- and raster-based monitors. Still facing limited computational power in the 1970s, graphics programmers were challenged by the desire to provide hidden-line elimination, which was necessary for realistic image rendering. As early as the 1950s when computers became available, algorithms were programmed to solve numeric problems. Pioneers in civil engineering soon performed structural analysis and other decision support as well as automation tasks using computers (Levitt, 1995; iii–v). Likewise in the 1950s, neural nets were conceived of (though not much researched due to lack of funding) and list-processing languages based on the mathematics of rule-based logic were implemented to support non-numeric processing. The 1960s were a prolific time in terms of development of new computing practices. Programmers in the field of Artificial Intelligence (AI) ambitiously developed a “general problem-solver” called GPS (Simon, 1996), object-oriented programming (OOP) languages came about, and the Internet was established. Soon Xerox PARC played a key role in prototyping new ways for people to interface with their computers (e.g. the SmallTalk OOP language and the mouse were invented there). The notion of blackboard systems developed in the late 1970s (Nii, 1986a, b). Their problem-solving ability relied upon distributed, complementary yet competing knowledge sources, and thereby allowed for problem solving to go on simultaneously at different levels of abstraction. Alternative blackboard architectures emerged in the 1980s (Hayes-Roth, 1985; Durfee, 1988). They are the predecessors to agent-based and later web-based systems. The early 1980s saw the birth of personal computers, which not only made computing more accessible to the masses but also stimulated the development of local area networks to ease communication and data transfer. This started an era for integration of stand-alone programs (e.g. Tommelein, 1995a) so characteristic of our fragmented industry (Howard et al., 1989). As for computer modeling in architecture/engineering/construction (AEC) practice, large engineering and construction firms such as Bechtel purchased software and hardware capabilities in the mid-1980s in order to develop their proprietary

Variability and uncertainty in product and process development 167

3D capabilities and then extend it into a WalkThru™ environment, a direct predecessor of today’s 4D models. Other companies such as Stone&Webster tailored off-the-shelf computing environments used in the automobile industry (i.e. Catia) to their own needs. Today’s models allow for parametric design (e.g. Revit, 2001). Processors with ever-increasing speeds and better display technologies at increasingly competitive prices have since flooded the market with computing capabilities. Further miniaturization, from desktops to laptops, then on to palmtops and cellular telephone devices, has made computing today truly ubiquitous. The explosion in popularity of the world wide web and OOP languages such as Java, allowing for distributed, collaborative problem solving using applets, has spurred the development of applications that could only be dreamt of a few years ago. AEC researchers, teachers, and practitioners have barely begun to scratch the surface of what is computationally possible with today’s hardware and software. As an industry, we are all too often conservative and focused on individual projects. Project economics using traditional yardsticks have only on occasion been favorable with respect to promoting the use of cutting-edge practices, including innovative management practices or adoption of the latest computer technologies. Today’s computer capabilities are rarely—if at all—holding us back from improving, or—better even—radically reinventing our work processes. The benefit/cost ratios for adopting new practices are shifting, thanks to not only current market pricing and availability of hardware and software, but more importantly due to increasing demand by project owners for added complexity and reliable performance of the facilities that are designed and built. Accordingly, spending more time on prototyping parts of an AEC facility (or the entire facility), prior to construction yields advantage internally to the involved organization and may provide the organization with external competitive advantage. It is within this context that 4D modeling plays an important role. The true power of 4D modeling as compared to 3D CAD is that by extending geometry with time, the opportunity is created for planning. Planning means giving consideration to alternative means and methods, alternative durations of activities, and alternative times and sequences in which to perform them. The planning task that involves space is not an easy one, however. Most resources considered in traditional planning are scalar (e.g. labor and staffing, equipment, materials, money, and time). Depending on the purpose for which the model will be used, space may or may not be represented adequately as a scalar. Space may be abstracted crudely to be linear (e.g. as in line-of-balance methods, see for instance Harris & Ioannou, 1998), two-dimensional (2D) (e.g. as in layout planning, see for instance Tommelein, 1989; Thabet, 1992; Tommelein & Zouein, 1993; and others), and at best to be 3D. Thirteen topological relationships exist between two one-dimensional (1D) intervals, as needed to model time and linear space (Allen, 1984). One can restrict the meaning of existing English words such as “before”, “during”, “starts”, etc. to express those relations. Similarly, 132 or 169 topological relations exist between two rectangles (not to mention arbitrary shapes) in 2D space, yet the English

168 I.D. Tommelein

language does not provide 169 single words to uniquely name each one. Obviously, inventing new words for each specific purpose would remedy the apparent shortage, but that is in most cases not necessary (Tommelein, 1989: 45). Thus, the complexity of representation increases accordingly, AEC modelers must carefully assess the consequences of creating more detailed representations. Nevertheless, the integration of space–time considerations makes it conceivable to simultaneously develop AEC products and production processes. Odeh (1992), Tommelein et al. (1994), Dzeng & Tommelein (1993), Dzeng (1995), Aalami (1998), and Akinci et al. (1998) are some of the researchers taking steps in that direction. This simultaneous development may lead to solutions that are far superior to those obtained by independent pursuit.

USE OF 4D MODELS 4D models are serving a myriad of purposes today:

• Visualization during design, construction, and marketing: depicting model components to scale or in some graphically-abstracted way (e.g. a pipe may be shown by its center line) so as to verify their geometry and location relative to other components; tagging components in order to track design updates and changes; computing spatial conflicts in order to avoid trade interference; and realistically rendering the model in order to enable a (prospective) owner to “see” the facility prior to purchase approval. • Study of product alternatives: allowing product components to be replaced by others so one can evaluate the impact thereof on clearances, access, line of sight, etc. • Assembly sequencing: studying alternative means for assembling the components and bringing them into their final position so that the construction process may be performed more productively. • Facilities management: linking geometrical shapes to data pertaining to component and system specifications, as-built dimensions, materials origin and processing steps (e.g. pharmaceutical facilities that require FDA validation or are subject to other regulatory inspection may have to be documented in great detail, including which worker performed which weld), warranty documentation, maintenance data, costs, etc. Using today’s 4D modeling tools, most of these applications are developed interactively. A person enters data into the computer and then applies judgment, possibly after receiving computer-generated evaluation feedback regarding the quality of the model at hand. Computer feedback may answer questions such as: Do these two components intersect? Is there at least a 50 cm clearance for the worker to get around and weld that connection? Is there sufficient space for the door to open all the way?

Variability and uncertainty in product and process development 169

Researchers have attempted to automate some of these tasks. Automation promises not only to take away a burden otherwise placed on the system user but, more importantly, it makes it possible to systematically develop and explore alternatives that a user may not consider otherwise. In order to avoid brute-force generate-and-test methods, the exploration of alternatives may be based on assumptions as is done in assumption-based truth maintenance systems (ATMS) (e.g. Levitt & Kunz, 1985). Assumptions help prune the search tree to be explored at any one time and provide points to backtrack to, should the pursuit of a specific line of reasoning be unsuccessful. Alternatively, the exploration of alternatives may rely upon a set-based representation of alternatives and be driven by postponed- rather than early-commitment strategies (e.g. Tommelein, 1989; Tommelein et al., 1991). Generalized set-based approaches have since proven to be very powerful (Lottaz et al., 1999; Sobeck et al., 1999). The early ambition for completely automating human “expert” tasks has since given way to much more realistic human–machine interaction. Many models used in research and practice today exploit computer strengths that augment human capabilities, rather than aiming altogether at eliminating human involvement in problem solving (e.g. Tommelein, 1989).

EXTENDED 4D MODELS FOR INTEGRATED PRODUCT–PROCESS DEVELOPMENT Practitioners involved in product and process development for the AEC industry need to reason about numerous other variables in addition to those regarding time and space. Moreover, most 4D models in use today do not explicitly reflect that the values taken on by geometric and temporal variables may not be known exactly, but can vary for a variety of reasons such as human indecision or physical tolerances. As a result, 4D models support practitioners’ endeavors only in part. To make 4D models more useful in AEC practice, they must be augmented with a multitude of other non-spatial and non-temporal data, such as materials characteristics (e.g. density, weight, conductivity) as needed in engineering design; reference to who is involved in design, construction, and maintenance as needed for organizational design and contracting purposes; component and production costs; and supply chains. As Smithers (1989) pointed out, CAD software was not designed to efficiently store non-geometric data. CAD therefore is inadequate in supporting product (or process) design data. Rather than appending non-geometric data to a geometry and creating what he called “a decorated geometry” that like an overloaded Christmas tree stands the risk of toppling over, he suggested that a database be developed first and that geometry be one form of output derived from it. Along the same line, Voeller (1996) cogently argues in favor or data-centered thinking, rather than thinking mainly about the graphical depiction of a facility.

170 I.D. Tommelein

The data-centered database could then efficiently serve numerous purposes above and beyond what 4D models can support, such as:

• Performance testing: applying a behavioral model (e.g. to simulate load distribution in a structure or the heat dissipation in an enclosed space) so one can judge the quality of a configuration. • Supporting conceptual design, detailed design, and procurement processes: providing catalogs of simple descriptive, parameterized, or full 3D CAD components that can be selected for inclusion in a design, with links to vendors and delivery terms pertaining to the supply of those components. For instance, Sadonio et al. (1998) presented a system that supports design for procurability. This paper presents a case for adding data that describes product and production features, as well as interdependence, uncertainty and variability both in the spatiotemporal definition of the product as well as in the product-development process. By acknowledging variations exist, the augmented, 4D⫹ model can be used to support other tasks, such as concurrent engineering with set-based design as well as reliable production planning.

AEC TASKS 4D models support part of the AEC product- and process-development process, which include tasks such as design, analysis, construction planning, and project management. A task is “the process of using a particular problem-solving method to solve an instance of a particular problem class” (Hayes-Roth et al., 1987). Problem classes have characteristic inputs and outputs. In order to solve a particular problem, the problem-solver will use one of several problem-solving methods, which in turn require the application of mechanisms (Balkany et al., 1993) to perform a sequence of operations that transform inputs into outputs. Symbolic problem-solving tasks may require numeric or non-numeric computation. Examples are algorithmic tasks, for which solvable mathematical equations are available, and knowledge-based tasks, for which search-based problem-solving methods are available (Tommelein, 1995b). In modeling the AEC process, the following tasks may be identified: Selection: applying deductive reasoning to pick one (or several) from a set of elements based on some criterion that may distinguish elements from each other, but that is not necessarily limited to numerical ordering. Evaluation: selecting elements to yield an ordered list of alternatives. Classification: matching data to a fixed set of solution classes (Stefik, 1995). Heuristic classification comprises data abstraction, heuristic match, and solution refinement (Clancey, 1985).

Variability and uncertainty in product and process development 171

Configuration: constructing a new artifact by selecting parts from a fixed set of parts (catalogued in a library or given as input) and interconnecting them so as to meet given specifications. Parts must be well characterized with respect to their function and relationships among each other (Mittal & Frayman (1989) define configuration more formally). Arrangement assembly (3D) or layout (usually only 2D): determining the topological positions or geometric coordinates of a fixed set of parts relative to each other, while meeting adjacency, distance, and other spatial constraints between them. Planning: defining tasks or activities and sequencing them in time. This may include assigning durations and resources to activities. Scheduling: calculating activity start and finish times, floats, etc. as well as resource histograms to aid in further analysis for the construction plan. Design: synthesizing a new artifact or system from scratch so as to meet given specifications. This may include but need not be restricted to the aforementioned tasks. 4D modeling is useful in several of these tasks. Nevertheless, 4D modeling ability is limited and other tools must be brought to bear when one aims at covering the entire AEC product- and process-development process. This article first describes sources of variability and uncertainty in product as well as process definition. It presents an example that illustrates the impact variability and uncertainty may have on the integrated product- and process-development process. It presents an exercise in problem solving that contrasts point-based with set-based reasoning. It concludes by recommending the adoption of 4D⫹ models to support the simultaneous development of AEC products and their production process.

VARIABILITY AND UNCERTAINTY IN PRODUCT AND PROCESS DEFINITION Product-development processes, which integrate design and construction, are notoriously difficult to manage because they are plagued by numerous uncertainties, including human indecision, tolerances, and unforeseen circumstances (e.g. Forrester, 1961; Crichton, 1966). Making explicit what uncertainties exist, how large they are, and where they may manifest themselves is a first step towards engineering a product and process that will be least impeded by them. It will help in deciding which sources of uncertainties should be tackled to reduce that uncertainty vs. which ones should be allowed to remain.

PRODUCT UNCERTAINTIES Depending on the stage of development of a design and on the level of abstraction adopted by an observer, several sources of uncertainty can be articulated. At the

172 I.D. Tommelein

product level, components and their assemblies are the subjects of interest. Examples of product uncertainties are:

• Configuration: a key step in architectural- and-engineering design is deciding

–

•

• •

which parts to include in a design. Parts will evolve from a conceptual or onedimensional (1D) specification at the start to a fully-specified geometry at the end of the design process. In configuration design, parts may be selected from an a priori defined set according to their function. In other kinds of design, shaping those parts is part of the design process. Shaping or selecting parts and defining their spatial arrangement and connectivity (e.g. allowing load, fluid, or current transfer) is what design is all about and it is driven by human decision-making based on constraints and preferences. Dimensional tolerances: stochastic variation relative to the design dimensions of a product or an assembly is described as “tolerance”. Individual parts may be manufactured with little dimensional variation but field operations usually are subject to greater variation. In order to achieve a tight fit, design and construction methods must rely on using filler materials when gaps remain (e.g. elastomers such as caulking to seal joints), trimming excess materials (e.g. carpet to fit a room’s dimensions), or overlaying materials to bridge openings. If not managed properly, tolerances may compound problems as design and construction progress (Tsao et al., 2000). Dimensional variation (degeneration): Tools, dies, and forms, etc. may wear out during their use so that gradual dimensional variation is introduced while shaping the product. Location and layout: as mentioned, deciding on the location and connectivity of parts relative to one another and in the context of a larger whole is a key task in design. Various space planning techniques exist to allow for flexibility so that decisions can be postponed or de-coupled from other decisions. For instance, one may zone a space by dedicating areas for specific uses, so as to ease coordination of space use (e.g. specialty contractors may do this for trade coordination where sprinkler piping, HVAC duct, electrical conduit, and other piping each have their designated layer).

PROCESS UNCERTAINTIES At the process level, design or construction activities and their resources are the subjects of interest. Resources generally-speaking denote people, tools and equipment, materials, money, space, and information. The modeling methodology adopted may be simulation (e.g. Tommelein, 1998). An activity requires resources as input when it starts, engages those resources during its entire duration of

Variability and uncertainty in product and process development 173

execution, and outputs the same or other resources when it finishes. Resources may be generic or characterized. At this process level, uncertainties pertain to:

• Scope of work: What work is to be performed is not necessarily stated clearly in

•

•

•

–

•

contract documents. Scope gap and scope overlap are big issues in subcontract coordination. In addition, a contract’s scope may change during construction to accommodate an owner changing their mind; to correct design mistakes; or to deal with unforeseen site conditions, new building regulations, availability of superior materials, etc. Duration and timing: Duration gauges the amount of time elapsed from start to finish of an activity. Start and finish events each mark a point in time. These probabilistic though measurable quantities provide a way in which to abstract what goes on during construction, and also describe how successor activities may be affected when the timing and duration of their predecessors are uncertain. Quality: Variation in quality may be the result of activities being executed by workers with varying skill levels, using different methods and subject to changing environmental conditions, etc. Inspection will determine which variation in quality is acceptable and whether or not rework will be necessary. Resource assignment: Project-level planners in general tend to ignore the specific assignment of resources to activities. In contrast, process planners—those at the construction site who organize and perform work—must plan for the allocation of resources (i.e. assign resources and sequence their use). Workers who need to install unique materials with specific tools and equipment better know what task is ahead of them, so they can plan how and where the work will be done and make sure all that is needed will be available when needed (Ballard & Howell, 1998). When allocation planning is done in advance of activity execution, opportunities exist to optimally choose which activities to perform first and when. How much in advance of execution this planning process should take place is a function of the complexity of the work to be performed and the uncertainties associated with that work and the process it is part of. Note, however, that even the best plans may fail when uncertainties manifest themselves during process execution, so good process design must include means to recover from those failures. Flow path and sequencing: It may not be a priori clear in what sequence work is to be performed, what routing is to be taken when handling materials, etc. Such decisions may have to be postponed and made during construction, when the relevant decision variables take on specific values, or they may have to be decided on stochastically at that time.

4D⫹ REPRESENTATIONS FOR PRODUCT AND PROCESS DEFINITION 4D models can support a significant part of the AEC product-development process. Their use typically starts when design has been substantially completed

174 I.D. Tommelein

and concerns for product visualization and assembly sequencing step into the limelight. Early design development and consideration for uncertainties pertaining to configuration, layout and location, timing, dimensional tolerances, procurability, constructability, etc. generally remain ill supported by them. Tools for truly integrated product and process development must cover a wider time span, starting from the definition of requirements and conceptual design, and allowing for contractor input early on (e.g. Gil et al., 1999). A more detailed articulation of flow combined with conversion issues (most tools today are conversioncentric) as well as variability are in order as they are key to process development (Koskela, 1992). These, in turn, place a different demand on the expressiveness of the tools being used. 4D models must be augmented to explicitly represent alternative sets, product component features in addition to 3D geometrical features, variability, and uncertainty, this approach is termed 4D⫹. On the product side, set-based representations of alternative component choices and configurations provide the opportunity for implementing least- or postponedcommitment strategies. Tommelein et al.’s (1992) SightPlan system created layouts for temporary facilities using a 2D bounded interval representation (bounded intervals are one way of representing sets). SightPlan could mimic human decisionmaking using early commitment, but improved its performance by taking advantage of least- or postponed-commitment using the available computing power. These strategies can be further leveraged during concurrent design (Kusiak, 1993; Koskela, 1997; Koskela & Huovila, 1997). Lottaz et al. (1999) provide an excellent example at the interface of structural design and HVAC system design. Set-based representations make it possible to check multi-discipline configuration alternatives and avoid iteration in the design cycle. In manufacturing, extended product models also capture the impact tolerances may have on assembly processes. On the process side, discrete-event simulation tools (Halpin & Woodhead, 1976; Hapin & Riggs, 1992; Halpin, 1993; Law & Kelton, 1991; Martinez, 1996) provide the expressiveness needed to represent and study the impact of flow, conversion, variability, and uncertainty on a production process. Using simulation symbols, one can model the entire product-development process, including the necessary activities and resources (human as well as others) but also the supply chains that merge at the construction site. For instance, Tommelein (1998) describes off- and on-site work pertaining to the piping function (Fig. 1). This model makes it possible to study what impact variability in process times may have on activity completion (compare the StagedSpool line in Fig. 2 with that in Fig. 3) and how uncoordinated sequencing may impact project completion (compare the AreaDone line in Fig. 2 with that in Fig. 3). Another model (Tommelein et al., 1999) shows how tightly-linked production stations starve (they are unable to produce) when subjected to upstream variability. Alternative process designs, for example, processes that de-couple interacting sub-cycles (Howell et al., 1993), can then be studied. Such studies are essential when one sets out to integrate product and process development and to successfully manage the corresponding

Variability and uncertainty in product and process development 175

AStart INIT 1

LINK LABELS AD Area Done CR Crew DA Data DT Design Team DW Drawing FB Feedback OF Off-site Work ON On-site Work PS Pipe Spool WA Work Area

Start OffSite OF1

OffSiteWork 0

GENERATE 150 ASpec

DA1

ATextFile

CHARACTERIZED RESOURCES TYPE SUBTYPE ATextFile ASpec AGraphicFile ACutSheet AMaterial ASpool ASpace AnArea ADataPiece ADatum COMPOUND RESOURCE AnAreaDone AStart Start INIT 1 OnSite

Specs

ON1

DA2

DT1 Design 1

GENERATE 4 ACutSheet

Design Team

ADesignTeam INIT 1

FieldWork 85

GENERATE 15 AnArea

DT2

WA1

DW1

DW3 AGraphic CutSheet File

ASpace

Update

4 CR

PS6

Transport Normal[3,1]

CONSOLIDATE 10

PS7

Staged Spool AMaterial

ASpace

8 PS

PS4

2

PS 5

CR

Await Transport

CR6 GENERATE 40 ADatum FB1

CR

CR7

1

Rework Pertpg[3,5,14]

AnInstallCrew INIT 1

Install Crew

Work Area Ready

A4

CR

Defect Spool

FabCrew

PrereqWork 10 WA3

GENERATE 3 1 ASpool PS AMaterial

ADataPiece

PS2 90%

AFabCrew INIT 20

% 10

CR5

Prep Crew

AMaterial

W

PS1

3

Fabricate Pertpg[3,5,14]

APrepCrew INIT 1

Feedback

GoodBad

WA2

2 FB

DW2

DW4

Work Area

Install 10

PS9 SpoolInArea A AD1 W A5

Area Done

CR8

AnAreaDone

Figure 1.

Piping function for off- and on-site work.

Figure 2.

Impact of variability in progress time on activity completion (Case 1).

176 I.D. Tommelein

Figure 3.

Impact of variability in progress time on activity completion (Case 2).

production system. Computer simulation allows for easy and inexpensive exploration of alternatives that would be prohibitive to investigate otherwise.

4D⫹ EXERCISE TO SUPPORT INTEGRATED PRODUCT AND PROCESS DEVELOPMENT OF A KITCHEN An example exercise will illustrate where 4D fits within an AEC system where product and process development are integrated. The detailing of a simplified problem-solving process, starting from design and ending with construction, helps to stress the need for representing many kinds of data in addition to spatio-temporal data. It provides concrete examples that raise issues and questions to be addressed in the integrated product- and process-development effort. Consider how a product, for example a kitchen, may get specified and how its associated development process unfolds. A kitchen is chosen because it will be familiar to the reader in terms of its functional requirements and naming of constituent parts, which are mainly appliances and spaces. Issues raised in this example are illustrative of those encountered during integrated product and process development of many other AEC facilities, such as hospitals, wafer fabrication plants, etc. Major phases in the development process of an AEC facility are 1) articulating the owner’s needs and other requirements, 2) conceptual design, 3) detailed design, 4) procurement, 5) fabrication, 6) supply, and 7) construction and final assembly. The preparation of contract documents is purposefully left out from this process; so are the selection of which party will be involved at which phase in the process and various approval steps. This is done to keep the example simple, not to diminish the importance of these phases or steps.

Variability and uncertainty in product and process development 177

ARTICULATING NEEDS AND REQUIREMENTS A crucial task in the development process is to articulate the owner’s needs and other requirements (e.g. permitting). These requirements are the drivers for the product-development process, against which alternatives will be evaluated. A kitchen is characterized as a space that provides the following functionality: space, refrigeration, heating, waste disposal, light (unless daylight suffices). This functionality essentially describes spaces needed to store ingredients for cooking (the ingredients themselves are supplied separately), preparation, the cooking process itself, and clean-up afterwards. The owner has identified the functional requirements specified below and expressed values (denoted by a number in brackets next to each requirement, where a high number corresponds to great value) that reflect appreciation for meeting those requirements. Some functional requirements are “hard,” which means that they must be met: without them, the owner will not accept the design. Others are “soft,” which means that they are optional though not necessarily without value. Nevertheless, without them, the owner will still accept the design.

FUNCTIONAL REQUIREMENTS OF KITCHEN OWNER

• Storage shelf space: at least 7.5 m2 • Counter space: work area at least 2 m2 • Heating – Four burners—narrow range [4] or wide range [5] – Conventional oven [4] – Optional—grill [2] – Optional—microwave [2] • Waste disposal – Washing and rinsing—two basins [6] or one basin [5] – Dishwasher [4] – Fume and smoke removal [6] – In-sink disposal [0.2]—nice to have but no great value attached to it • Refrigeration [5] • Light—will be handled after other functions have been satisfied

178 I.D. Tommelein

Figure 4.

Kitchen floor plan.

• Budget not to exceed US$… • Design and construction time not to exceed … days. Assume that a room of 3 m long and 2.40 m wide is available to accommodate these requirements (Fig. 4). To allow for easy access, all shelving and appliances have to be placed on either one of the two sides, leaving an access path of about 1.10 m between them, across the entire space. Owners generally have at least a basic understanding of the functional requirements of the facility being developed; sometimes they are true experts. For owners to have such understanding is necessary because they must specify these requirements and provide associated value assessments. More often than not, however, they will count on support from the designer and builder before completing these specifications, as designers and builders can bring a wealth of detailed and up-to-date domain knowledge regarding products and processes into the development process.

CONCEPTUAL DESIGN The designer, familiar with numerous other kitchen designs solutions and knowledgeable about what is available on the market, may provide the owner with a functional hierarchy of kitchen appliances and spaces from which units can be selected for inclusion in the design. This hierarchy, depicted in Figure 5, shows that some appliances fulfill a single function (e.g. oven, trash compactor) whereas others fulfill multiple functions simultaneously (e.g. oven, range, and grill may be built into a single unit). Most appliances in this hierarchy are of fixed dimension, only a few remain to be sized. Provided that the owner is willing to work with those standard-size

Figure 5. Functional hierarchy for kitchen components.

Variability and uncertainty in product and process development 179

180 I.D. Tommelein

Figure 6.

Counter and shelf space depths.

components, the design problem, if it were to only involve appliances, is therefore simply a problem of selecting parts. The design problem involving counter and storage space has more degrees of freedom and is thus more typical of true design in terms of the number of alternatives that may have to be investigated prior to deciding on a solution. Even so, the designer may wish to stick to standard sizes of cabinets and shelving in order to keep production costs low. Figure 6 illustrates that the shelf space below the counter will most likely be 64 cm deep, and the shelf space suspended above it 32 cm deep. Typical heights are also shown. In this example, these dimensions will (not coincidentally) comfortably accommodate any built-in appliances. As for interconnecting these parts, a few spatial constraints exist between them. Several constraints are listed below; they express common practice, but preference rather than hard requirements. The location of most appliances is dictated by their use, not by their functionality. Stove burners, for instance, are mounted at counter top level to allow for easy reach, but would still function if mounted higher or lower than that. Similarly, an oven may be installed at eye level or under a range or counter. Appliances usually need hook-ups to power, water, gas, and ventilation. Because the connection “ports” for these are rather generic, this example problem requires more thought about layout than about configuration. An exception to this may be the fume hood whose location must be configured closely with an exhaust chimney. Similarly, the location of the sink must be configured jointly with the location of the remainder of the plumbing system. For instance, it is not uncommon for a designer to keep laundry rooms or bathrooms close to the kitchen

Variability and uncertainty in product and process development 181

and to align all “water functions” vertically (e.g. one bathroom above the other) throughout a building.

DIMENSIONAL AND SPATIAL CONSTRAINTS BETWEEN PARTS

• Two shelves 46 cm vertically separated can be accommodated under the counter and two shelves 23 cm vertically separated at eye elevation.

• A range or a range-with-grill is about 10 cm high, so two layers of shelf space • • • •

or an oven can be accommodated underneath of it. A sink is about 30 cm high, including space for the basin itself as well as drainage plumbing. Only one layer of shelf space can be installed underneath of it but there would not be enough space to accommodate a (built-in) dishwasher. A microwave oven can be integrated into the fume hood or it may be placed on open counter space. The fume hood must cover at least the area of the appliance where fumes and smoke may be generated, namely ovens, ranges, and grills. Built-in units such as ovens, dishwashers, and trash receptacles usually have counter space above them. Refrigerators are so tall there practically can be have one layer of shelf space above them.

The functional hierarchy depicted in Figure 1 is by no means exhaustive. Other components may exist though not be so readily available on the market (need to show procurement lead times for each item). Yet others may be custom fabricated (e.g. fume hood is shown to be made to any size). The conceptual design stage will start off by the designer jointly with the owner identifying (in approximation) the required space (which is a given here), then identifying appliances and spaces needed to fulfill the requirements, and attempt to fit them into the space. In this example, it is assumed that the available space is well defined at the start of design. This assumption is realistic in situations where the owner has access to historic information regarding what is “usually adequate,” but does not necessarily have a specific idea of what functionality will be needed. In the high-tech industry, for instance, a facility may be sized based on reference to existing facilities, standard designs, or prototypes, though manufacturing tools to be housed in the facility may still be under development when design details are being worked out. As always, any assumption carries risk. In this case, should the owner wish to include an unusually large tool or add tools that are not part of a typical facility, the available space may not suffice to accommodate them all. Before reading on, the reader may find it interesting to play the role of designerbuilder and solve this design problem by hand. While doing so, please take note of the problem-solving steps you take. Given the aforementioned specifications, the problem certainly is not over-constrained, that is, at least one solution exists.

182 I.D. Tommelein

ALTERNATIVE APPROACHES FOR CONCEPTUAL DESIGN Several methods can be followed to generate a conceptual design solution for this problem. Depending on the method being pursued, the solutions will be different. Early commitment Many designers follow a so-called early commitment approach. This approach most likely is also the one the reader pursued. Early commitment can mean several things. The problem solver may first choose appliances one at a time from the functional hierarchy and try them against the functional requirements. This will eliminate the range-and-grill, for instance, as it does not provide four burners. Similarly, the requirements do not mention any trash disposal other than a sink and dishwasher, so the bin and compactor may be eliminated from consideration as well. The chest freezer is not specified either. A designer may question, however, if these omissions are not an oversight on the part of the owner. Using this method, the problem-solver may find that many appliances, when taken individually, can meet the requirements. The next step is to fulfill each requirement separately and then verify if in combination they allow for sufficient storage shelf space and counter space to remain. For instance, the range & oven, the wide-range & oven, and the wide-range & grill & oven all meet the requirements for heating, though the first scores 8, the second 9, and the third 11. The maximum amount of upper-shelf space Umax one could have in this kitchen is equal to the kitchen length * depth of shelf * 2 layers * 2 sides of the room or Umax ⫽ 3 m * 0.32 m * 2 * 2 ⫽ 3.84 m2 Similarly, the maximum amount of lower-shelf space Lmax one could have in this kitchen is Umax ⫽ 3 m * 0.64 m * 2 * 2 ⫽ 7.68 m2 This yields a total shelf space of 11.52 m2. The maximum amount of counter space Cmax one could have in this kitchen is Cmax ⫽ 3 m * 0.64 m * 1 * 1 ⫽ 3.84 m2 Assuming the designer chooses the option that scores the highest, the space occupied by that appliance can be calculated based on an assumed position in the kitchen, and the remaining upper-shelf space, lower-shelf space, and counter space calculated. The next step may be to decide on means for waste disposal. Should the user proceed in this manner and for each function choose the high-end option (the largest appliance that meets the requirements), then ultimately the space constraints will be violated (in this particular problem). Backtracking to consider

Variability and uncertainty in product and process development 183

lower-end options instead then is necessary in order for a solution to be found. This iteration between “generate and test” and then “backtrack when failure is encountered” is characteristic of early commitment strategies. As mentioned, however, other early commitment strategies exist. Another designer, less keen on encountering failure and being forced to backtrack, may start off by choosing the low-end for every option. The process will then reveal whether or not a solution even exists, given the appliances shown in the hierarchy. In a subsequent iteration, lower-end options can then be substituted with higherend options, so as to reach greater value for the owner. Still, iteration is the name of the game. The early-commitment method is a logical one to pursue as the problem-solver can choose individual parts before moving on to the next choice. It is often the only method available when memory capacity is limited, as is the case when the problem is solved by a person without the use of computational support tools. Postponed commitment An alternative to early commitment is to postpone commitment until a later time, when other considerations can be added into the decision-making process, and iteration may be avoided (at least to some extent). A postponed-commitment problem-solver may recognize that two sets of solutions exist to meet the heating requirements. One set was already identified, namely any appliance chosen from {(range & oven)(wide-range & oven wide-range & grill & oven)} could do the job. Alternatively, any appliance combination chosen from {(range & oven) (wide-range & oven)} could also do the job. These two sets combined enumerate all possibilities. A microwave could be added to any of these configurations. At this point in problem solving, no selection of any specific solution is made. Instead, the postponed-commitment problem-solver will move on to the next requirement and identify the corresponding solution set for it. This process continues until all individual sets have been identified. The next step then is to achieve consistency among steps, where consistency means that elements in any of the sets may be eliminated because no solution across sets meets the space requirements. The following tables with calculations make this reasoning clear. Table 1 shows the space requirements for each appliance in terms of their width (w), depth (d), and height (h). In addition, a calculation was made as to how much space the appliance would take away if it were chosen for inclusion in the design. U denotes upper-shelf space occupied by the appliance, L lower-shelf space, and C counter space. This calculation required some judgment in terms of where the appliance would most likely be positioned. For instance, it would be impractical to have a range at upper-shelf space height, so U ⫽ 0 for any appliance that includes a range. Table 1 now makes additional calculations possible. For instance, one may characterize each set by an upper- and lower-bound space requirement. Six columns in

184 I.D. Tommelein Table 1.

Space requirements for appliances in functional hierarchy.

Microwave Conventional oven Wide conventional oven Range & oven Wide-range & oven Wide-range & grill & oven Range Wide-range Range & grill Dishwasher Sink Wide sink Double sink A Double sink B Refrigerator Wide refrigerator Two-door refrigerator

w [m]

d [m]

h [m]

U [m2]

L [m2]

C [m2]

0.46 0.54 0.76 0.54 0.76 1 0.54 0.76 0.76 0.6 0.5 0.6 0.9 0.9 0.72 0.82 0.92

0.30 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.76 0.76 0.76

0.40 0.6 0.6 0.9 0.9 0.9 0.1 0.1 0.1 0.6 0.3 0.3 0.3 0.3 1.64 1.64 1.64

0.32 0 0 0 0 0 0 0 0 0 0 0 0 0 0.2432 0.2752 0.3072

0.32 0.742 1.024 0.742 1.024 1.331 0 0 0 0.819 0.346 0.41 0.602 0.602 0.973 1.101 1.229

0 0 0.371 0.512 0.666 0.371 0.512 0.512 0 0.346 0.41 0.602 0.602 0.486 0.55 0.614

Table 2 illustrate such calculations for the first heating set from {(range & oven)(wide-range & oven)(wide-range & grill & oven)}. This calculation reveals that the lower-end choices leave sufficient space, whereas the higher-end choices do not. Further investigation will reveal what set combinations yield acceptable space use and example calculations are shown in the right-most nine columns of Table 2. Only the last three of these columns present sets of appliances that in any combination yield a solution. Admittedly, calculations like these are tedious to perform but this is exactly where computers are most useful. They can perform calculations and keep track of sets, and thereby enable people to focus on other decision-making steps. Note that in this example, the alternative possible configurations can be enumerated by hand should the designer wish to do so. In more complex design cases this will certainly not be the case.

DISCUSSION The advantage of the postponed-commitment method is that it keeps track of all possible configurations that meet the requirements. No solutions are prematurely rejected. While applying postponed commitment to define the product, the designerbuilder should also consider production process issues, such as procurement (e.g. availability and lead times), fabrication, supply, and construction and final

0.31 0.31

0.24

3.84 3.60

Total

Maximum space available Remaining space

7.68 4.80

2.88

0.97

3.84 2.64

1.20

0.49

3.84 3.53

0.00

0.24

0.35

Refrigerator Wide refrigerator Two-door refrigerator

0.35

0.00

0.00

0.00

0.00

0.37

Sink Wide sink Double sink A Double sink B

0.82

0.74

0.00

0.00

Dishwasher

Conventional oven Wide conventional oven Range & oven Wide-range & oven Wide-range & grill & oven range Wide-range Range & grill

7.68 3.70

3.98

1.23

0.60

0.82

1.33

L

U

C

U

L

Upper bound

Lower bound

Alternative Set 1

Table 2. Calculation of set properties.

3.84 1.96

1.88

0.61

0.60

0.00

0.67

C

L

C

3.84 7.68 3.84 3.53 4.01 2.11

0.31 3.67 1.73

0.31 1.23 0.61

0.00 0.60 0.60

0.00 0.82 0.00

0.00 1.02 0.51

U

Alternative heating L

C

3.84 7.68 3.84 3.56 3.83 2.02

0.28 3.85 1.82

0.28 1.10 0.55

0.00 0.60 0.60

0.00 0.82 0.00

0.00 1.33 0.67

U

Alternative refrigeration L

C

3.84 7.68 3.84 3.60 4.27 2.24

0.24 3.41 1.60

0.24 0.97 0.49

0.00 0.60 0.60

0.00 0.82

0.00 1.02 0.51

U

Alternative heat and refrigeration Variability and uncertainty in product and process development 185

186 I.D. Tommelein

assembly. After considering these, the product solution sets will likely be further narrowed, but one or several solutions will remain if the problem is solvable. By contrast, should such issues be considered after the design had been locked in to a single product solution, little flexibility would have remained and any additional constraint might then lead to failure and force another cycle of design iteration. The likelihood for additional considerations being revealed after design “completion” is great. During design, owners will likely learn about and become interested in options they did not think of initially, when they learn about new functional capabilities being desirable and/or available on the market. When choices such as “Which of two double sinks is best?” leaves the owner indifferent, issues such as availability may prevail. Owners may also express different product and process preferences regarding life-cycle concerns at the time the system in its entirety has been configured. A design that is further refined to include component choices and geometry can be represented in 3D CAD with timing information added to reflect manufacturing, supply, and construction assembly sequencing. Questions to be answered using the 4D model may include: Once all parts have been selected and dimensioned, where will they fit within the kitchen space provided? Upon delivery and installation, will all appliances fit through the door opening to the kitchen? Questions to be answered using an extended 4D model may pertain to numerous other issues. Questions to be answered during production planning may include: How well can the owner articulate the requirements and preferences? Are these likely to change during the design-build process? What are the lead times for getting the various appliances and materials to site? How will deliveries be made? What are the logistics of staging and moving materials about the site? What is the availability of skilled labor? In what order will the trades (cabinet makers, electricians, plumbers) proceed to build this kitchen? How long does it take to install cabinets and shelving, rough-in plumbing; rough-in electrical; install and hook-up each appliance; finish plumbing, finish electrical, finish tile counter tops? When is a component or a system hooked up so it can be tested for operation? Questions to be answered throughout the product-development process may include: Does each individual selection meet the owner’s specifications? What configuration best meets the owner’s needs?

Variability and uncertainty in product and process development 187

The set-based representation, like the one presented here to extend the usefulness of 4D models, enables AEC practitioners to consider alternatives in products while taking process issues into account. It not only yields better solutions, but also avoids needless iteration. Other industries have adopted this approach (e.g. Ward et al., 1995; Sobek et al., 1999).

CONCLUSIONS This study has presented sources of variability and uncertainty in product as well as process definition. A world in which no variation or uncertainty is recognized typically gets modeled by means of single numbers, often representing expected values. These values are mathematically speaking most likely to occur but in reality have an almost zero likelihood of actually occurring. Singe values reflecting early commitment tend to lead to process iteration. Systems that are subject to variability and uncertainty in terms of processing times exhibit phenomena such as “starvation” that lead to detrimental performance. Models based on averages are unrealistically optimistic. Systems that are based on unique selections at each design step, tend to cause substantial iteration and rework when conflicts are detected. This chapter has therefore argued for the creation and study of 4D models that explicitly represent variability and uncertainty as well as sets of alternatives, referred to as 4D⫹. Examples illustrated how 4D⫹ representations make it possible to better manage the production system that supports integrated product and process development. People’s ability and ease with which they can solve problems depends on the representation that is being used. Researchers and practitioners need to expand their conceptualizations of AEC systems so as to allow for integrated product and process development, and then develop representations and problem-solving methods to enable us to really tackle the problems face.

ACKNOWLEDGMENTS Several ideas presented in this chapter were refined during much-valued discussions with Glenn Ballard, Carlos Formoso, Hyun Jeong Choo, Marcelo Sadonio, Nuno Gil, Cynthia Tsao, Jan Elfving, Nadia Akel, Michael Whelton, and Yong-Woo Kim. Research leading to the development of CADSaPPlan was funded by grant CMS-9622308 from the National Science Foundation (NSF). On-going research on integrated product and process development is being funded by grant SBR9811052 from NSF. All support is gratefully acknowledged. Any opinions, findings, conclusions, or recommendations expressed in this chapter are those of the author and do not necessarily reflect the views of NSF.

188 I.D. Tommelein

REFERENCES Aalami, F.B. 1998. Using construction method models to generate four-dimensional production models. Ph.D. Dissertation, Department of Civil and Environmental Engineering, Stanford University, CA. Akinci, B., Fischer, M. & Zabelle, T. 1998. Proactive approach for reducing non-value adding activities due to time–space conflicts. Proceedings of the 6th annual conference of the International Group for Lean Construction (IGLC-6): Guaruja, Brazil, August:12 pp., available at http://www.ce.berkeley.edu/⬃tommelein/IGLC-6/ Allen, J.F. 1984. Towards a general theory of action and time. Artificial Intelligence 23: 123–154. Balkany, A., Birmingham, W.P. & Tommelein, I.D. 1993. An analysis of several design systems. Journal of Artificial Intelligence in Engineering, Design, and Manufacturing 7(1): 1–17. AI EDAM. Ballard, G. & Howell, G. 1998. Shielding production: essential step in production control. Journal of Construction Engineering and Management 124(1): 11–17. ASCE. Clancey, W.J. 1985. Heuristic classification: 68 pp. Knowledge Systems Laboratory Report Nr. KSL-85-5, Stanford University. Crichton, C. 1966. Interdependence and uncertainty. A study of the building industry: 83 pp. London: Tavistock Pubs., Tavistock Institute. Durfee, E.H. 1988. Coordination of distributed problem solvers. Kluwer Academic Pubs. Dzeng, R.-J. 1995. CasePlan: a case-based planner and scheduler for construction using product modeling. Ph.D. Dissertation, Department of Civil and Environmental Engineering, University of Michigan, Ann Arbor, MI, September, 300 pp. (also Report No. UMCEE 95–22). Dzeng, R.-J. & Tommelein, I.D. 1993. Using product models to plan construction. Proceedings of the 5th international conference on computing in civil and building engineering, Anaheim, Calif., 7–9 June: 1778–1785. ASCE. Forrester, J.W. 1961. Industrial dynamics. M.I.T. Press. Gil, N., Tommelein, I.D., Miles, R.S., Ballard, G. & Kirkendall, R.L. 1999. Integrated product–process development model to support design-build. In M. Hannus, M. Salonen & A.S. Kazi (eds), Concurrent engineering in construction: challenges for the new millenium: 367–376. CIB Publication 236. Proceedings of the 2nd international conference on concurrent engineering in construction (CEC 99), Espoo, Finland, 25–27 August. Organized by CIB TG33 and VTT Building Technology. Halpin, D.W. 1993. Process-based research to meet the international challenge. Journal of Construction Engineering and Management 119(3): 417–425. ASCE. Halpin, D.W. & Riggs, L.S. 1992. Planning and analysis of construction operations: 381 pp. New York: Wiley-Interscience. Halpin, D.W. & Woodhead, R.W. 1976. Design of construction and process operations. New York: John Wiley and Sons. Harris, R.B. & Ioannou, P.G. 1998. Scheduling projects with repeating activities. Journal of Construction Engineering and Management 124(4): 269–278. ASCE. Hayes-Roth, B. 1985. A blackboard architecture for control. Artificial Intelligence 26: 251–321. Hayes-Roth, B., Garvey, A., Johnson, M.V., Jr. & Hewett, M. 1987. A modular and layered environment for reasoning about action. Report No. KSL-86-38, 63 pp. Knowledge Systems Lab., Computer Science Department, Stanford University, CA, April.

Variability and uncertainty in product and process development 189 Howard, H.C., Levitt, R.E., Paulson, B.C., Pohl, J.G. & Tatum, C.B. 1989. Computer integration: reducing fragmentation in the AEC Industry. Journal of Computer and Civil Engineering 3(1): 18–32. ASCE. Howell, G., Laufer, A. & Ballard, G. 1993. Interaction between sub-cycles: one key to improved methods. Journal of Construction Engineering and Management 119(4): 714–728. ASCE. Issa, R.R. (ed.), Fernando, T., Naidoo, S.R. & Toledo Santos, E. (section eds) 1999. State of the art report virtual reality in construction. CIB TG24, January, available at http://www.bcn.ufl.edu/tg24/final/ Koskela, L. 1992. Application of the new production philosophy to construction. Technical Report 72, 75 pp. CIFE, Stanford University, CA, September. Koskela, L. 1997. Re-engineering, concurrent engineering, lean production: what is the ideal antidote for the construction industry’s ailments? In S. Mohamed (ed.), Proceedings of the international conference on construction process re-engineering, Gold Coast, Griffith University, QLD, Australia, 14–15 July: 293–301. Koskela, L. & Huovila, P. 1997. On foundations of concurrent engineering. In C. Anumba, & N. Evbuomwan (eds), Paper presented at Concurrent engineering in construction CEC 97; 1st international conference, London, 3–4 July: 22–32. London: The Institute of Structural Engineers. Kusiak, A. (ed.) 1993. Concurrent engineering: automation, tools, and techniques. John Wiley & Sons. Law, A.M. & Kelton, W.D. 1991. Simulation modeling and analysis, 2nd edition: 759 pp. McGraw-Hill, Inc. Levitt, R.E. 1995. Foreword. In I.D. Tommelein (ed.), Expert systems for civil engineers: integration issues: 247 pp. Expert Systems and Artificial Intelligence Committee, Technical Council on Computer Practices. ASCE. Levitt, R.E. & Kunz, J.C. 1985. Using knowledge of construction and project management for automated schedule updating. Project Management Journal 16(5), December. Lottaz, C., Clément, D.E., Faltings, B.V. & Smith, I.F.C. 1999. Constraint-based support for collaboration in design and construction. Journal of Computer and Civil Engineering 13(1): 23–35. ASCE. Martinez, J.C. 1996. STROBOSCOPE state and resource based simulation of construction processes. Ph.D. Dissertation, Civil and Environmental Engineering, University of Michigan, Ann Arbor, MI, 518 pp., available at http://www.strobos.ce.vt.edu/ Mittal, S. & Frayman, F. 1989. Towards a generic model of configuration tasks. Proceedings of the 11th IJCAI, August: 1395–1401. Nii, H.P. 1986a. Blackboard systems: the blackboard model of problem solving and the evolution of blackboard architectures. AI Magazine 7(2): 38–53. Nii, H.P. (1986b). Blackboard systems: blackboard application systems, blackboard systems from a knowledge engineering perspective. AI Magazine 7(3): 82–106. Odeh, A.M. 1992. CIPROS: Knowledge-based construction integrated project and process planning simulation system. Ph.D. Dissertation, Department of Civil and Environmental Engineering, University of Michigan, Ann Arbor, MI. Revit. 2001. www.revit.com Sadonio, M., Tommelein, I.D. & Zabelle, T.R. 1998. The LAST DESIGNER’S databaseCAD for sourcing, procurement, and planning. Proceedings of the Computing Congress 98: 364–377. ASCE. Simon, H.A. 1996. The sciences of the artificial, 3rd edition. Cambridge, Mass.: MIT Press. Smithers, T. 1989. AI-based design versus geometry-based design or why design cannot be supported by geometry alone. Computer-Aided Design 21(3): 141–150.

190 I.D. Tommelein Sobeck, D.K., II, Ward, A.C. & Liker, J.K. 1999. Toyota’s principles of set-based concurrent engineering. Sloan Management Review Winter: 40(2): 67–88. Stefik, M. 1995. Introduction to knowledge systems. Morgan Kaufmann Publishers. Thabet, W.Y. 1992. A space-constrained resource-constrained scheduling system for multi-story buildings. Ph.D. Dissertation, Civil Engineering Department, Virginia Polytechnique and State University, Blacksburg, VA, 644 pp. Tommelein, I.D. 1989. SightPlan—An expert system that models and augments human decision-making for designing construction site layouts. Ph.D. Dissertation, Department of Civil Engineering, Stanford University, Stanford, CA, 203 pp. Tommelein, I.D. (ed.) 1995a. Expert systems for civil engineers: integration issues: 247 pp. Expert Systems and Artificial Intelligence Committee, Technical Council on Computer Practices, ASCE. Tommelein, I.D. (ed.) 1995b. Introduction: knowledge-based systems and models for integration: 1–39 in Tommelin 1995a. Tommelein, I.D. 1998. Pull-driven scheduling for pipe-spool installation: simulation of a lean construction technique. Journal of Construction Engineering and Management 124(4): 279–288. ASCE. Tommelein, I.D., Carr, R.I. & Odeh, A.M. 1994. Assembly of simulation networks using designs, plans, and methods. Journal of Construction Engineering and Management 120(4): 796–815. ASCE. Tommelein, I.D., Levitt, R.E. & Hayes-Roth, B. 1992. Site layout modeling: how can artificial intelligence help? Journal of Construction Engineering and Management 118(3): 594–611. ASCE. Tommelein, I.D., Levitt, R.E., Hayes-Roth, B. & Confrey T. 1991. SightPlan experiments: alternative strategies for site layout design. Journal of Computer and Civil Engineering 5(1): 42–63. ASCE. Tommelein, I.D., Riley, D. & Howell, G.A. 1999. Parade game: impact of work flow variability on trade performance. Journal of Construction Engineering and Management 125(5): 304–31. ASCE. Tommelein, I.D. & Zouein, P.P. 1993. Interactive dynamic layout planning. Journal of Construction Engineering and Management 119(2): 266–287. ASCE. Tsao, C.C.Y., Tommelein, I.D., Swanlund, E. & Howell, G.A. 2000. Case study for work structuring: installation of metal door frames. Proceedings of the eighth annual conference of the International Group for Lean Construction (IGLC-8), Brighton, UK, 17–19 July. Voeller, J. 1996. Data-centered thinking. Editorial. Journal of Computer and Civil Engineering 10(1): 1–2. ASCE. Ward, A.C., Liker, J.K., Cristiano, J.J. & Sobeck, D.K., II. 1995. The second toyota paradox: how delaying decisions can make better cars faster. Sloan Management Review Spring: 43–61.

FURTHER READING Alarcon, L. (ed.) 1997. Lean construction: 497 pp. Rotterdam: A.A. Balkema. Anumba, C.J. & Evbuomwan, N.F.O. 1997. Collaborative working in construction—the need for effective communication protocols. Proceedings of the 4th congress on computer in civil engineering: 89–96. ASCE. Autodesk 1997. AutoCAD release 14. User’s guide. San Rafael, CA: Autodesk, Inc.

Variability and uncertainty in product and process development 191 Ballard, G. & Howell, G. 1994. Improving performance behind the shield. Proceedings of the 2nd annual conference of the International Group for Lean Construction, Santiago, Chile, reproduced in Alarcon, 1997. Bechtel. 1988. 3D design systems users manual, Overview of Bechtel 3D design system. Version 1.1, Bechtel Power Corp., Gaithersburg, MD. Beliveau, Y.J., Williams, J.M., King, M.G. & Niles, A.R. 1995. Real-time position measurement integrated with CAD: technologies and their protocols. Journal of Construction Engineering and Management 121(4): 346–354. ASCE. Bernold, L.E. & Salim, Md. 1993. Placement-oriented design and delivery of concrete reinforcement. Journal of Construction Engineering and Management 119(2): 323–335. ASCE. Bernold, L.E. & Treseler, J.F. 1991. Vendor analysis for best buy in construction. Journal of Construction Engineering and Management 117(4): 645–658. ASCE. Choo, H.J. & Tommelein, I.D. 1999. Space scheduling using flow analysis. In I.D. Tommelein (ed.), Proceedings of the 7th annual conference of the International Group on Lean Construction (IGLC-7), Berkeley, CA, 26–28 July: 299–311, available at http://www.ce.berkeley. edu/⬃tommelein/IGLC-7/ Choo, H.J. & Tommelein, I.D. 1999. Parade of trades: a computer game for understanding variability and dependence. Technical Report 99-1, Construction Engineering and Management Program, Civil and Environmental Engineering Department, University of California, Berkeley, September, available at http://www. ce.berkeley.edu/⬃tommelein/parade/parade.pdf Choo, H.J. & Tommelein, I.D. 2000. Interactive coordination of distributed work plans. ASCE, Proceedings of the construction congress VI, Orlando, Florida, 20–22 February:10 pp. (in press). CPI. 1998. Collaboration in the building process. The Collaborative Process Institute, http://www.dprinc.com/collabor.html The Economist 1999. Tomorrow’s internet. The Economist November 13: 23–26. Edlinger, P. 1997. Importance of data exchange standards for process industry. Editorial. Journal of Computer and Civil Engineering January: 6–7. ASCE. Elzarke, H.M. & Bell, L.C. 1995. Object-oriented methodology for materialsmanagement systems. Journal of Construction Engineering and Management 121(4): 438–445. ASCE. Fischer, M. & Froese, T. 1996. Examples and characteristics of shared project models. Journal of Computer and Civil Engineering 10(3): 174–182. ASCE. Fisher, N. & Yin, S.L. 1992. Information management in a contractor. A model of the flow of project data. 260 pp. London: Thomas Telford. Goldratt, E.M. & Cox, J. 1986. The goal. Croton-on-Hudson, NY: North River Press. Hammer, M. 1990. Reengineering work: don’t automate, obliterate. Harvard Business Review July–August: 104–112. Jin, Y. & Levitt, R.E. 1996. The virtual design team: a computational model of project organizations. Computational & Mathematical Organization Theory 2(3): 171–196. Lin, K.-L. & Haas, C.T. 1996. An interactive planning environment for critical operations. Journal of Construction Engineering Management 122(3): 212–222. ASCE. Maxim, B., Balkany, A., Birmingham, W.P., Darr, T.P., Runkel, J.T. & Tommelein, I.D. 1992. Prototyping knowledge-based design systems in an object-oriented environment. The Society for Computer Simulation and Object Management Group. Proceedings of the international conference on object-oriented manufacturing systems, ICOOMS ’92, University of Calgary, Canada, 4–6 May: 55–59. Oglesby, C.H., Parker, H.W. & Howell, G.A. 1989. Productivity improvement in construction. New York: McGraw-Hill Inc.

192 I.D. Tommelein Proof. 1995. Using proof animation, 2nd edition. Annandale, VA: Wolverine Software Corp. Riley, D.R. 1994. Modeling the space behavior of construction activities. Ph.D. Dissertation, Department of Architectural Engineering, Pennsylvania State University, University Park, PA, 296 pp. Riley, D.R. & Sanvido, V.E. 1995. Patterns of construction-space use in multistory buildings. Journal of Construction Engineering and Management 121(4): 464–473. ASCE. Riley, D. & Sanvido, V. 1997. Space planning for mechanical, electrical, plumbing, and fire protection trades in multi-story building construction. In S. Anderson (ed.), Proceedings of the construction congress V, Minneapolis, MN: 102–109. ASCE. Rojas, E.M. 1997. Developing web-centric systems for collaborative engineering. Proceedings of the 4th congress on computers in civil engineering: 232–239. ASCE. Rother, M. & Shook, J. 1998. Learning to see: value stream mapping to create value and eliminate muda, Vol. 1.1, Oct. Brookline, Mass.: The Lean Enterprise Inst. Schmenner, R.W. 1993. Production/operation management: from the inside out, 5th edition: 825 pp. Prentice Hall. Schrage, M. 1997. The real problem with computers. Harvard Business Review September–October: 178–188. Stukhart, G. 1995. Construction materials management. Marcel Dekker, Inc. Stumpf, A.L., Ganeshan, R., Chin, S. & Liu, L. 1996. Object-oriented model for integrating construction product and process information. Journal of Computer and Civil Engineering 10(3): 204–212. ASCE. Tommelein, I.D. 1991. Site layout: where should it go? Proceedings of the Construction Congress 91: 632–637. ASCE. Tommelein, I.D. & Ballard, G. 1997a. Coordinating specialists. Technical Report No. 97-8, Construction Engineering and Management Programme, Civil and Environmental Engineering Department, University of California, Berkeley, CA. Also in Proceedings of the 2nd international seminar on lean construction, Sao Paulo, Brazil, 20–21 October. Organized by A.S.I. Conte, Logical Systems, Sao Paulo, Brazil. Tommelein, I.D. & Ballard, G. 1997b. Lookahead planning: screening and pulling. Technical Report No. 97-9, Construction Engineering and Management Programme, Civil and Environmental Engineering Department, University of California, Berkeley, CA. Also in Proceedings of the 2nd international seminar on lean construction, Sao Paulo, Brazil, 20–21 October. Organized by A.S.I. Conte, Logical Systems, Sao Paulo, Brazil. Tommelein, I.D., Dzeng, R.J. & Zouein, P.P. 1993. Exchanging layout and schedule data in a real-time distributed environment. Proceedings of the 5th international conference on computing in civil and building engineering, Anaheim, California, 7–9 June: 947–954. ASCE. VRML. 1995. The virtual reality modeling language, Version 1.0 specification. May 26, available at http://www.virtpark.com/theme/vrml/, visited 7/16/99. Womack, J.P. & Jones, D.T. 1996. Lean thinking: banish waste and create wealth in your corporation: 350 pp. New York: Simon & Schuster. Zabelle, T.R. & Fischer, M.A. 1999. Delivering value through the use of threedimensional computer modeling. CEC 99—Proceedings of the 2nd international conference on concurrent engineering in construction, CIB TG33 and VTT, Espoo, Finland, 25–27 August.

Variability and uncertainty in product and process development 193 Zouein, P.P. 1995. MoveSchedule: a planning tool for scheduling space use on construction sites. Ph.D. Dissertation, Civil and Environmental Engineering Department, University of Michigan, Ann Arbor, 308 pp. Zouein, P.P. & Tommelein, I.D. 1999. Dynamic layout planning using a hybrid incremental solution method. Journal of Construction Engineering and Management 125(6): 400–408. ASCE. Zouein, P.P. & Tommelein, I.D. 2001. Improvement Algorithm for Limited Space Scheduling. Journal of Construction Engineering and Management 127(2): 116–124. ASCE.

APPLICATION OF 4D CAD IN THE CONSTRUCTION WORKPLACE Richard J. Coble1, Robert L. Blatter2, Indrid Agaj1 1

M.E. Rinker, Sr. School of Building Construction, University of Florida, Gainesville, FL, USA 2 Gresham Smith and Partners, Jacksonville, FL, USA

Abstract 4D is the new breed of CAD, which accounts for the fourth dimension, time. A 3D drawing of a project is linked with the schedule forming an integrated 4D model that becomes a powerful tool for construction managers. It is also a tool that can be used at the foreman level. The concept of requiring foreman to equate 3D drawings with time should be the goal of every construction manager. Time is the basis on which contracts are awarded, disputes originate, and the projects success is judged. A foreman must understand the intent and end result of a segment before it can be constructed. If the foreman can be helped through this process with visualization tools, the issue becomes one of cost vs. value analysis. Current computer technology is making its way into the hands of foreman through the use of portable computers, PDAs, and wireless communications. The acceptance of computer technology by foreman is on the rise with many companies making it a requirement for employment. This acceptance is what makes 4D CAD desirable for field application in the construction industry. Presently, by the use of still photography and other information and communication media, 4D CAD usage at the foreman level can be a reality in the near future. Keywords: 4D CAD, construction foreman, computer applications, change order, field applications

INTRODUCTION The techniques utilized today in turning an owner’s need and/or idea into a physically standing and usable structure typically involve graphing the complex construction process into a bar chart and/or CPM schedule. This representation communicates the sequence of activities over time but does not provide a 195

196 R.J. Coble et al.

relationship to the physical objects represented by those activities. Construction managers, owners, and designers have to visualize the relationship of the schedule with the actual physical product and rely on experience to make appropriate planning decisions. Visualizing the transition from a bar chart to a finished product in the field is the hardest for the owner of the facility. A need exists for a comprehensive tool that allows architects, engineers and contractors to simulate and visualize construction sequences as part of an interactive exercise. 3D graphic models are useful in visualizing spatial relationships of parts of facilities/projects. Since a CAD model provides the basis for a common language between all parties, adding time to a 3D model creates a visual simulation of the construction process—4D CAD. With 4D CAD design and construction planning alternatives can be assessed realistically within the context of space and time. Simultaneous modeling of temporal and spatial aspects of scenario can optimize and justify the conscious decisions that jeopardize or hinder the completion of many construction projects. Industry and academia have been exploring this medium to better understand how it can be used effectively during the planning process. 4D CAD is a tool that supports the 4D nature of engineering and construction. This tool must capture and dynamically manage the interaction between project components and resources over time, visualize these interactions and support real-time interaction of users with the 4D model. This tool also encourages the communication, approval and improvement of construction schedules by various parties, such as construction managers, clients, designers, subcontractors and community members. (http://gaudi.stanford.edu/ 4D-CAD/INTRO-4DCAD.HTML)

TODAY’S INDUSTRY TRENDS The implementation success of 4D CAD in the construction industry is greatly dependent and driven by the different trends and factors that govern this industry. These factors include:

• • • • • • •

Percentage of work done by subcontractors. Large companies’ volume vs. total industry volume. Percentage of large companies that perform design-build work. Percentage of large companies that perform CM-at-risk work. Percentage of negotiated vs. hard-bid work. Number of change orders and RFIs. CAD literacy among construction field management.

The fluctuations of these percentages, and the increase in the number of change orders and RFIs, will mean an increase in the need of construction companies to

4D CAD in the construction workplace 197

implement and fully adopt 4D CAD. It would be expected that the 4D CAD methods would be put to use by the large construction companies first. The projects these companies are involved with are of a size and complexity that can absorb the initial added design costs that come with 4D CAD. Furthermore, it is these companies that can afford further training and CAD literacy lessons for their field managers and foremen. 4D CAD applications in the construction workplace will become a reality for mid-size companies when the experiences of large companies account for better project delivery method. The above factors are the major force pushing for the implementation of 4D CAD in today’s efforts for minimizing time and cost at project delivery by the construction companies. Analyzing these factors in further detail will provide more accurate reasons why this implementation is the door to the future of project management and delivery.

SUBCONTRACTOR WORK It is estimated that in general, subcontractors do about 80 to 85% of the work in commercial projects. With the percentage of work subcontracted-out by the general contractors becoming higher, the amount of coordination and planning (between the general contractor and the subcontractor) required to complete a project is increasing. The high demand of coordination efforts is a compelling reason that calls for new and untraditional solutions to project management. The major advantage of 4D CAD is the rate at which communication interference diminishes. The enormous coordination needed between the general contractor and the subcontractor will be the reason that will initiate a widespread use of 4D CAD. The level of interaction and coordination between the general contractor and the subcontractor usually determines if a project will be delivered on time, under budget, and contractor initiated change order free. Implementation of 4D CAD will undoubtedly contribute to the lessening of these communication barriers, and the improvement of project schedule coordination. These, in turn, will be major factors in “ironing-out” the scheduling imperfections, which will eventually lead towards a timely project delivery. In the future, general contractors will need training in CAD software so that they will be able to manipulate the models and interpret how they should be drawn. In addition, contractors will need to provide input to designers so that the CAD objects are drawn in a way that supports automated quantity takeoffs. Finally, general contractors are likely to become the keepers of the models, accepting information from the designer and parceling it out to the subcontractors. This flow of information will continue throughout the project as design changes are incorporated and propagated. (Fischer, et al., 1999)

198 R.J. Coble et al.

In addition, subcontractors will become more active in the early phases of the design as the architect and engineer develop the specifications and schematics that form the basis for subcontractors’ design. The subcontractors’ detailed designs will still need to be approved by the architect and engineer through the shop drawing process, but subcontractors will need to develop CAD modeling capabilities to benefit from this involved process. Also, as subcontractors become more active in the design, they will become better able to assist the general contractor in coordinating the work of subcontractors throughout project delivery. (Civil Engineering, May 1999)

LARGE COMPANIES’ VOLUME VS. TOTAL INDUSTRY VOLUME In 1997 there were a total of approximately two million contractors in the United States alone. Of these firms, only 667,089 were not self-employed. These companies include “establishments primarily engaged in the construction of buildings and other structures, heavy construction (except buildings), additions, alterations, reconstruction, installation, and maintenance and repairs” (US Census Bureau). Also included are companies involved in land preparation, demolition, and construction management. Of these, companies with 1–9 employees accounted for roughly 82% (543,753 companies) of all construction companies not self-employed. That means that only about 18% of construction companies with at least one employee, or just over 6% of all construction work is done by companies with 10 or more employees. In fact, only 55,645 companies, or 3% of all construction companies, employ 20 or more people. Construction put in place in September 1999 was estimated at US $700.1 billion. Of this, US $540.3 billion was performed within the private works sector while US $159.8 billion was generated on public works projects. It could be assumed that companies with more employees are doing more work. This would lead to the conclusion that only 6% of all construction companies account for a significant percentage of construction in place. While the medium and large firms constitute only 6% of all construction firms, these firms are doing most of the work. It is important to understand the differences between these firms. Among the medium and large firms, there are only a few really large companies. One-third of 1% of all construction companies (5,684) employ more than 100 people and only 77 of those companies employ more than, 1,000 people. Engineering News Record (ENR) annually reports the top 400 contractors based on total revenues of the previous year. The top 400 contractors of 1999 range from US $78.7 million to US $9.7 billion annual revenues and most have at least 100 employees. The total volume produced domestically by those companies in 1998 was approximately US $127 billion. The total value of construction in place in 1998 equaled US $665.446 billion. That means that ENR’s

4D CAD in the construction workplace 199 Table 1. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Top 15 contractors for 1999. Bechtel Group Inc., San Francisco, Calif. Fluor Daniel Inc., Irvine, Calif. Kellogg Brown and Root, Houston, Texas CENTEX Construction Group, Dallas, Texas The Turner Corp., New York, NY Foster Wheeler Corp., Clinton, NJ Skanska (USA) Inc., Greenwich, Conn. Peter Kiewit Sons Inc., Omaha, Neb. Gilbane Building Co., Providence, RI Bovis Construction Corp., New York, NY McDermott International Inc., New Orleans, La. Raytheon Engineers and Constructors, Cambridge, Mass. J.A. Jones Inc., Charlotte, NC Jacobs Sverdrup, Pasadena, Calif. Morrison Knudsen Corp., Boise, Idaho Black and Veatch, Kansas City, Mo. PCL Enterprises Inc., Denver, Colo. Structure Tone Inc., New York, NY The Clark Construction Group Inc., Bethesda, Md. The Whiting-Turner Contracting Co., Baltimore, Md.

Source: 1999 ENR top 400 contractors.

top 400 contractors of 1999 (one third of 1% of medium and large firms) generated 19% of the total construction in place in 1998. The statistics in Table 1 contribute towards further understanding of the impact of large companies in the overall construction industry. These companies, constituting only 6% of the total number of construction companies, will be the testing ground for 4D CAD. The relative small number of companies that have to buy into the 4D CAD idea, and start incorporating it in their pre-construction and construction phases, increases the probability of 4D CAD becoming a valuable tool in the construction workplace.

COMPANIES THAT PERFORM DESIGN-BUILD WORK The in-house capability of design-build companies is the breeding ground for a future full-bloom 4D CAD. Implementation of 4D CAD will further lessen the communication barriers between owners, architects, and construction managers. The structuring of a design-build company accounts for these barriers by trying to eliminate them. Furthermore, the structure of a design-build company is set in such a way that implementation of 4D CAD fits and further enhances the working relationships within the company. This makes possible the review of the project’s

200 R.J. Coble et al.

schedule while being designed, which is attainable since both teams, the architect and the contractor, work together for the same company. The possibility of the contracting side of the project reviewing the way it is going to be put together while in the design phase, gives design-build companies an enormous edge over the traditional design-bid-build method of building a facility. Both teams (designers and builders) will be working simultaneously on a project throughout the design phase. The contractor side will see how the building will be put together and at the same time will alert the designer of any necessary changes that need to be made. These changes can be related to the projects schedule or constructability. This process will enable the minimization of field changes and will give the contractor a better view of the schedule before the project has started, and as a result, improve the possibility of a timely delivery of the project.

PERCENTAGE OF LARGE COMPANIES THAT PERFORM DESIGN-BUILD WORK Most of the top design-build companies listed in Table 2 are among the top firms (according to the ENR’s “Top 400 Construction Companies”) listed based on total revenue. Table 2. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Top 20 design-build firms for 1999. Kellogg Brown and Root, Houston, Texas The Turner Corp., New York, NY Bovis Construction Corp., New York, NY Structure Tone Inc., New York, NY Skanska (USA) Inc., Greenwich, Conn. DPR Construction Inc., Redwood City, Calif. Foster Wheeler Corp., Clinton, NJ Gilbane Building Co., Providence, RI Fluor Daniel Inc., Irvine, Calif. Morse Diesel International Inc., New York, NY Chicago Bridge and Iron Co., Plainfield, Ill. Stone and Webster, Boston, Mass. Peter Kiewit Sons Inc., Omaha, Neb. The IT Group, Monroeville, Pa. Marnell Corrao Associates Inc., Las Vegas, Nev. Morrison Knudsen Corp., Boise, Idaho Parsons Corp., Pasadena, Calif. The Haskell Co., Jacksonville, Fla. Ryan Cos. US Inc., Minneapolis, Minn. The Austin Co., Cleveland, Ohio

Source: ENR’s 1999 top 100 design build firms.

4D CAD in the construction workplace 201

As mentioned earlier, the key in making 4D CAD a viable tool for construction management will be the initial ability of this new technique to infiltrate through the “old ways” of management techniques and become a popular and efficient way of overlooking and managing a construction project. The best chance 4D CAD has in becoming the new tool in construction management is to start introducing itself to the large companies first. It is these companies that can absorb the initial cost and adjustment that comes as a result of implementing this new technique for the first time. The nature of the projects these companies are involved with provides a good ground for the testing of 4D CAD. These projects are large in size and usually innovative management techniques is what determines the timely and “within budget” delivery of them. The introduction of 4D CAD in these types of projects performed by the large design-build companies will improve the communications among the different levels of parties involved in the project. The obvious result will be a better understanding by the owner of how the future facility will progress through the building phase. In addition, 4D CAD implementation in project delivery will further improve the flow of communication between the design and construction personnel within the design-build company. The 4D technology will prove more beneficial if it is implemented further on the job site by making the 4D model available to the project manager, field engineer, and foreman. When starting to analyze the numbers, the percentage of top design-build firms that are part of the overall top construction firms is relatively high. These companies are the ones that cover most of the large project design-build work in the country. For 4D CAD this means that only a small percentage of the design-build companies have to implement it in their system of project design and management. This relative small percentage of companies will be the testing ground for the new tool, but at the same time this small percentage accounts for most of the large design-build projects currently being built. This way the 4D CAD implementation will be realistically an attainable goal, since only a relatively small number of companies that will use it will cover the majority of the testing ground (the large design-build projects). COMPANIES THAT PERFORM CM-AT-RISK WORK The sixth largest firm in 1999 was Foster Wheeler Corporation of Clinton, New Jersey, producing US $3.072 billion in revenue. Nearly 72% (US $2.204 billion) of its revenues were generated abroad. Foster Wheeler is another company that specializes in industrial projects. They completed 79 industrial projects in 1998. Of those, 38% were performed CM-at-risk. Industries served by Foster Wheeler include oil and gas field development, chemicals, petrochemicals and polymers, pharmaceuticals and fine chemicals, petroleum processing, power generation, cogeneration, and resource recovery.

202 R.J. Coble et al.

The Turner Corporation of New York, New York, ranked fifth with revenues of US $3.699 billion. It performed US $54.3 million of construction internationally. Turner completed 100 buildings in 1998. Its expertise is strictly in buildings, be it pre-construction, construction management, or general contracting. Unlike other giants in the industry, which perform relatively small amounts of construction under CM-at-risk contracts, Turner performed 71% of its 1998 revenues as a CM-at-risk. Gilbane Building Company of Providence, Rhode Island, ranks ninth and earned US $2.248 billion. All revenues were domestic. Of this work, 51% was performed CM-at-risk. This company is involved in all major construction and real estate markets: industrial, institutional, and commercial. This includes education, criminal justice, healthcare, public assembly, aviation, life sciences, corporate, government/ public service, library, and infrastructure. At number 10, Bovis Inc of New York, New York, earned US $2.213 billion in 1998. Although Bovis is an international firm, New York, is their base of operations in the United States. That means 100% of those revenues were in the United States. Bovis provides the following services: construction management, design/ build, project management, general contracting, build/operate/transfer, and systems/ consulting. Of those services, work performed CM-at-risk accounts for 87%. The types of projects built include shops, offices, factories, schools, hospitals, airports, arenas, and theaters. A glimpse at CM-at-risk operations may lend some insight. Two of the top 10—Turner and Bovis—obtain more than two-thirds of their revenues through CM-at-risk. These firms are earning around nine times the industry average per employee while the others are only doubling. All the above statistics help in better understanding the role of the top building firms in today’s construction industry. It is this small percentage that performs a large portion of the total construction volume in this country. The volume of business these companies cover is the perfect stage for the new 4D CAD because it requires testing only on a small number of companies. These firms are able to accommodate the new methods without feeling the cost of the initial expenses. When comparing Table 1 with Table 3, among these top firms, most of them perform a sizable portion of their work as CM-at-risk. Managing projects as a CM-at-risk company means that the actual work has to be subcontracted away. This is when 4D CAD proves to be a most valuable tool. The extensive amount of coordination needed in such a project more than justifies the use of 4D CAD. If the CM firm has complete 4D models than they can make them available to the major subs involved in the project. This will require full willingness to cooperate from the subs, but these large firms can outline that as one of the items in the contract between them and the subcontractors. The use of 4D CAD than will greatly reduce of amount of miscommunication among the parties, which will in turn produce a timely delivery of the project.

4D CAD in the construction workplace 203 Table 3. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Top 20 CM-at-risk firms for 1999. Bechtel Group Inc., San Francisco, Calif. Fluor Daniel Inc., Irvine, Calif. McDermott International Inc., New Orleans, La. Kellogg Brown and Root, Houston, Texas Jacobs Sverdrup, Calif. Raytheon Engineers and Constructors, Cambridge, Mass. Black and Veatch, Kansas City, Mo. ABB Lummus Global Inc., Bloomfield, NJ Opus Group of Companies, Minnetonka, Minn. Foster Wheeler Corp., Clinton, NJ Chicago Bridge and Iron Co., Plainfield, Ill. Stone and Webster, Boston, Mass. Peter Kiewit Sons Inc., Omaha, Neb. The IT Group, Monroeville, Pa. Marnell Corrao Associates Inc., Las Vegas, Nev. Parsons Corp., Pasadena, Calif. Morrison Knudsen Corp., Boise, Idaho Kraus-Anderson Construction, Minneapolis, Minn. Perini Corp., Framingham, Mass. Austin Industries, Dallas, Texas

Source: ENR’s 1999 top 100 CM firms (at risk).

NEGOTIATED VS. HARD-BID WORK Among the large construction firms, most of them perform a certain quantity of negotiated work. Furthermore, the amount of negotiated work performed by the top construction firms is on the rise. “Centex”, for example, claims that 80% of their business is repeat business and negotiated work. The nature of the process of a negotiated project’s delivery accommodates the use of 4D CAD. Throughout the design and construction period the owner of the project follows the proceedings very closely. This closeness to the project demands a better method of communication between the professional parties and the owner. 4D CAD, being this new method, will greatly facilitate the interaction among the parties, and furthermore, help the owner visualize the way the project is going to be built. The amount of construction firms that perform negotiated work is small, which makes the effort to introduce these firms to 4D CAD easier. These firms, even though in small number, perform a large quantity of the total negotiated work in the country. By implementing 4D CAD in these construction firms, therefore, a large portion of negotiated contracts and work will be performed by using 4D CAD. This, along with the other factors affecting the spread of 4D CAD, will be the foundation for the future of this new tool in the construction industry.

204 R.J. Coble et al.

4D CAD AND CHANGE ORDER DOCUMENTATION 4D CAD is an ideal way to document changes in job site conditions. In the following section of this paper is shown how actual job photographs can be integrated to accomplish a clear and accurate picture of what has occurred to necessitate a change. Additionally, a clear representation is depicted as to the depth and breadth of the change. By having this clear representation, produced in four dimensions, conflict resolution is taken from the initial cause to an analysis of cost and process. Using this methodology, owners gain an understanding of why their projects cost more and take longer to complete. At the same time, designers can decipher the cause of the conflict and how the resolution will impact associated costs. The example presented in this paper shows clearly an error by the design team. A lack of coordination resulted in a change that necessitated a time and cost delay to the project. This could have been a major stumbling block for the project if a 4D representation had not been used to illustrate the problem and present a solution. The programs used to create this 4D example are readily available and ideal for the issue being discussed. This type of descriptive representation is also appropriate for multiple changes. Manipulating the simulation to fit actual field photographs is not a difficult process. Other areas that could be examined using this tool include mechanical and electrical conflicts. Case study: Girder conflict The space being constructed was a fitness room with an exposed ceiling structure and glass walls on two sides overlooking a nearby lake. The room was intended to

Figure 1. Job site photo of kicker penetrating soffit and window system (not yet installed).

4D CAD in the construction workplace 205

have a clear height of 9 feet and the window system was to match this height using a 3-foot module. The project designer did not convey this information to the structural consultant. As the project was being constructed, a joist girder traversing the space was found to have a bottom angle at 8⬘-6⬘⬘. This girder passed through the window system to its support column (see Fig. 1). To alleviate this problem, the bottom angle of the joist girder was cut back behind the impending window system. Because this made the girder unstable, kickers had to be introduced to brace the bottom angle laterally (Figs 2 and 3). The roof deck and substrate were examined to determine if the attached kickers would present any uplift problems. After determining that there would be no critical uplift stresses, the steel contractor removed the required girder angle and braced the member with kickers. This example illustrates a problem that is easily solved requiring minimal rework and cost to maintain the architectural intent of the project. An element of time was added to demonstrate to the parties involved how minimal the impact would be on the project (Fig. 4). By incorporating the project schedule, showing the progression of applicable line items, the total projects duration change would be apparent. This exercise in computer technology served many purposes. First and foremost, it allowed participants in the project not familiar with construction technique or terminology to understand the problem and the resolution immediately. This may seem inconsequential until one realizes that the people frequently responsible for financing the project usually have little construction experience. The second item accomplished with this virtual explanation is an interpretation of the sequence and process required to complete the change. Steps to procure the

Figure 2.

Section of girder to be removed from conflict with window system.

206 R.J. Coble et al.

Figure 3.

Kickers to brace girder after removal.

Figure 4.

Project schedule outlining changes generated by girder conflict.

4D CAD in the construction workplace 207

materials, schedule the labor, and install the necessary components are mapped into the video. The project schedule associates and generates the requisite line items and incorporates them to produce and adjust the time scale.

TECHNOLOGY AND PROCESS The tools and skill necessary to produce 4D CAD simulations are readily available and require moderate training to use. Table 4 is a simplified process chart outlining the applicable tools and techniques. Table 4 shows the four main components needed to develop a project. Solid modeling is the component that consumes a majority of the time needed to assemble a project. The impending scene is formed out of solid shapes and uses Boolean operations to add and subtract material as necessary for object creation. The addition of NURBS (non-uniform rational b-splines) allows a designer to assemble objects that include or are based on compound curves. The process of digital rendering places the solid model in an environment in which it is given realistic properties. The solid model is given material properties and the surrounding environment is defined. In this phase of the work, animation paths and cycles are cast and activated. A 4th dimension can be attained in this phase of the project by sequencing the solid model or a combination of solid models on the same animation path. Table 4.

Simplified process chart outlining tools and techniques.

Process

Cost rank

Solid modeling

2

Digital rendering

1

Post-manipulation

4

Production assembly

3

Description

Properties

Construction of the model and alternate scenes in four-dimensions The constructed scenes are given realistic properties and an environment in which to be viewed and manipulated

CAD Program with Boolean operations

Image touch-up and modification on a frame-by-frame basis. Also used to produce headliners and credits Construction assembly of the project for output to selected media

Scene and animation rendering based on individual floating camera and target points Image touch-up and manipulation

Off-line video editing suite with NTSC and PAL export

Note: The column entitled cost rank is a comparison of the components price relative to the other included tools. The lowest number is the most expensive.

208 R.J. Coble et al.

Post manipulation is the process of image editing to “touch-up” frames of the sequence. It can also be used to create slides from images for fade and dissolve transitions in the off-line editing phase. Time oriented material cannot be controlled with this process. Production assembly is an off-line audio and video construction process. In this phase of the project, the stills and/or animations created in the digital rendering process and the slides created in the digital rendering process are combined to create a final product with video transitions and audio accompaniment. This final product can be output to video or digital media.

IMPLEMENTATION Getting the foremen to buy into the use of this technology is something that will be dependent on the extent of their training and understanding they have as to the importance of this methodology. The implementation process is something that can be broken down into increments, which would start with clearly documenting the change with either job photos, which ideally would be digital. This again would be dependent on the foreman, in some cases it might be better to let the foreman use a camera that they are more familiar with and then scan in the photographs. On a more advanced foreman level, the ones that are able to use 3D CAD can actually be taught how to better document the changes to the 3D CAD drawings. In some cases this process can be extended to even include the time element, which would be ideal to have the foreman create the 4D CAD change documentation in the field. In addition, the National Center for Construction Education and Research (NCCER), located in Gainesville, Florida, is the main source in educating and training the future foremen and supplying this force to the construction industry. If initial steps towards educating the foremen in the 4D CAD applications in the construction field are taken while these future field supervisors are still in the learning mode in the NCCER classrooms, the effects will be tremendous for the construction industry. This class of 4D CAD-educated foremen will be the base for a computer literate generation of foremen. The role of NCCER and the volume of educated construction force that it is producing will make 4D CAD-educated foremen a reality in the near future.

RECOMMENDATIONS In order to implement 4D CAD on the construction site, foreman must understand and recognize the potential benefits of this new technology. Foreman are the

4D CAD in the construction workplace 209

closest form of supervision that management has on the construction site that have the ability to understand the intent and end result of a construction segment before it can be implemented and also have the authority to document and implement these changes. The new generation of foreman would be open to the introduction and daily use of this technology. However, if current foreman can be helped through the initial stages of this process with visualization tools, acceptance and use would increase. The second step of implementation of 4D CAD on the construction site is that the tools and training necessary to produce 4D CAD simulations would need to be made available. With the implementation of 4D CAD tools on the construction site, a changed product can be developed and understood by all parties before it is physically installed.

CONCLUSION Construction managers, owners, and designers have to visualize the relationship of the schedule with the actual physical product and rely on experience to make appropriate planning decisions. Visualizing the transition from a bar chart to a finished product in the field is the hardest for the owner of the facility. The existing need for a comprehensive tool that allows architects, engineers, and contractors to simulate and visualize construction sequences led to 3D and eventually 4D CAD. Since a CAD model provides the basis for a common language between all parties, adding time to a 3D model creates a visual simulation of the construction process. With 4D CAD design and construction planning alternatives can be assessed realistically within the context of space and time. It would be expected that the 4D CAD methods would be put to use by the large construction companies first. The projects these companies are involved with are of that size and complexity that can absorb the initial added design costs that come with 4D CAD. Furthermore, it is these companies that can afford further training and CAD literacy lessons for their field managers and foremen. 4D CAD applications in the construction workplace will become a reality for mid-size companies when the experiences of large companies account for better project delivery method. The major advantage of 4D CAD is the rate at which communication interference diminishes. The enormous coordination needed between the general contractor and the subcontractor will be the reason that will initiate a widespread use of 4D CAD. Change documentation is ideal for 4D CAD usage. Presently project engineers have to document these changed conditions with the use of photography, which will eventually be replaced with digital photography and further facilitate the importing of the pictures into the 4D CAD application. By documenting and producing potential changes in four dimensions, conflict resolution is taken from the

210 R.J. Coble et al.

initial cause to a negotiated and accepted analysis of cost and process. If the foremen could be taught to use this technology, then the coordination effort would be very straightforward. Foremen are the key to making 4D CAD a viable solution to documenting field changes and hopefully eliminating or greatly reducing changed conditions disputes. They are one of the last frontiers to using computer technology in construction. It is our belief that 4D CAD is a visual and communicative tool that will greatly help bringing the foremen into the new computer age.

REFERENCES Fischer, M., Spradlin, M. & Staub, S. 1999. Into the fourth dimension. Civil Engineering 69(5): 44–47. http://gaudi.stanford.edu/4D-CAD/INTRO-4DCAD.HTML

FURTHER READINGS Akinci, B. & Fischer, M. 1998. Time–space conflict analysis based on 4D production models. In K.C.P. Wang (ed.), International computing congress, Boston, 18–21 October: 342–353. ASCE. Fischer, M.A. & Aalami, F. 1996. Scheduling with computer-interpretable construction methods Models. Construction Engineering and Management 122(4): 338–345. ASCE. Collier, E. & Fischer, M. 1996. Visual-based scheduling: 4D modeling on the San Mateo County health center. In J. Vanegas & P. Chinowsky (eds), Proceedings of the 3rd computational civil engineering congress: 801–805. ASCE. Griffis, F., O’Brien, W. & Bronner, P. 1990. Columbia construction research: the applications of three-dimensional computer models in construction. Architectural and Engineering Systems. http://www.tier2.census.gov/cgi-win/cbp/cbp.exe http://www.census.gov/pub/const/C30/tab198.txt http://www.centex.com/about-centex/companyprofile.htm http://turnerconstruction.com/about.html http://www.fwc.com/industries/ http://www.gilbaneco.com/Inside-Gilbane/gbco/Frame_history.htm http://bovis.com/Services/frServices.asp http://www.enr.com/dbase/99db.asp http://www.enr.com/dbase/99cmrisk.asp Novitski, B.J. 1999. Two architects demonstrate how object-oriented CAD will change the way design is done. Architectural Record October 1999. Sahai, R. 1996 Inside Microstation 95. Onward Press, 4th edition. Thabet, W.Y. & Beliveau, Y.J. 1994. Modeling work space to schedule repetitive floors in multi-story buildings. Journal of Construction Engineering and Management 120(1): 98–115. ASCE.

VIRTUALLY REAL CONSTRUCTION COMPONENTS AND PROCESSES FOR DESIGN-FOR-SAFETY-PROCESS (DFSP) Steve Rowlinson, Bonaventura H.W. Hadikusumo Department of Real Estate and Construction, The University of Hong Kong, Hong Kong

Abstract Bulky 2D design drawings from consultants create difficulties for users to interpret into a 3D mental picture for construction purposes. If only visualizing a 3D mental picture of the project creates burdens to user, then he will find more problems in integrating this information with other plans such as the construction process and safety regulations. In other words, it is difficult to add more contents in the 2D-design representation. Virtual reality aims to allow the end-user to view a 3D model of a project. So, contents of the design can be added, such as construction processes and a design-for-safety-process (DFSP), before real construction is undertaken. This paper discusses some key elements required to build virtually real construction components and processes focusing on the concept of a reusable object library and variables for construction processes simulation. A brief discussion of DFSP is presented. Keywords: visualization, design-for-safety-process, construction site safety, high rise residential construction simulation, reusable objects

INTRODUCTION Construction design results in bulky drawing sets from architecture, structural, mechanical and electrical (M/E) engineering. This system causes difficulties in transferring the information of the bulky 2D drawings to develop a 3D mental picture of the project for construction purposes. Since the user is given a bulky set of drawings, he spends a lot of time and effort to create the 3D mental picture before the construction. Then, he prepares further plans such as construction methods and a safety plan which are documented as 211

212 S. Rowlinson & B.H.W. Hadikusumo

shop drawings and safety procedures. However, these additional plans also create another separate information set which must be interpreted and integrated with the 2D design drawing prior to construction. Safety planning has been implemented in Hong Kong projects but more than 300 accidents per 1,000 workers occurred every year from 1991 to 1997. Hong Kong’s accident rate is twice that of the United States, 25 times worse than Japan, and nearly 30 times worse than Singapore (Lingard & Rowlinson, 1991). One of the underlying reasons for accidents was the failure of workers to recognize hazards (Lingard & Rowlinson, 1997), which is an important element in accident occurrence (Ramsey, 1976 in Furnham, 1998). In this research, therefore, a virtually real construction process focusing on safety is proposed to represent the construction products/components and processes/activities of a typical Hong Kong Housing Authority building project. These virtually real construction products and processes will be used to develop a design-for-safetyprocess (DFSP) methodology which aims to view the design and point out the potential safety hazards inherited in the virtual project as well as the necessary actions to avoid the realization of those hazards. In this paper, the approach adopted to build a universe of construction projects using World Up™ virtual reality (VR) software and important elements to generate the construction process simulation are discussed. Also, a brief review of how the DFSP will be developed is included.

IMPROVING CONSTRUCTION DESIGN USING VIRTUAL REALITY With today’s technology, developing VR application is not a serious problem in terms of technology and cost. VR has been seen as a new way to visualize an object in many domains such as medical, military, and engineering as well as construction. Research into VR in construction has included safety training (Hadipriono & Larew, 1996), generating construction plan (Faraj & Alshawi, 1996) and visualizing construction components and processes (Adjei-Kumi & Retik, 1997). In the construction domain, VR can give advantages as follows:

• Reducing end-user’s effort and task to transfer by imagination the 2D data from the bulky construction drawings to a 3D mental picture of a project. Integrating information from different project participants (architect, structural, • and M/E engineer) into one universe. This can detect incomplete and ambiguous design prior to the construction stage. • Facilitating a user with a virtual walkthrough in which he can check any corner and location of the virtual project at a controllable distance and speed. Since VR has benefits as described, a research question is posed (Fig. 1), can VR be used to improve the design contents in terms of a DFSP?

Design-for-safety-process 213

M/E Eng.

Architect

Structural Eng.

Figure 1.

Complexity in transforming bulky 2D drawings to 3D mental picture.

APPROACHES TO BUILD A VIRTUAL REALITY UNIVERSE IN THE CONSTRUCTION DOMAIN A universe in VR is a virtual world composed of pre-defined objects such as the block, sphere, cylinder, 3D text and imported objects as well as light. These predefined objects can be created by two methods:

• Direct conversion from 2D graphic software to VR software. Faraj & Alshawi (1996) used Autocad™ by Autodesk™ to draw the construction plan. By developing an interface, this data can be directly converted into objects in the VR universe. • Developing and utilizing an object library to create objects in a VR universe (Adjei-Kumi & Retik, 1997). Once the library of objects is created, these objects are reused to compose a universe. Of these approaches, the second is considered more practicable for this research. The reasons are the core of this research is on the development of DFSP instead of developing an interface between 3D modeler software and World UP™ software; and the Hong Kong Housing Authority standard block, the phase one, can be built quite fast for the next phases, the representation of construction processes in the virtually real universe and the development of DFSP.

214 S. Rowlinson & B.H.W. Hadikusumo

REUSABLE OBJECT LIBRARY Reusability is the process of creating software/application systems from existing software rather than building them from scratch (Krueger, 1992 in Sametinger, 1997). The reusability concept facilitates a big benefit for application development in terms of time, quality, and cost. New application development time can be shortened since some part of the codes or components have already been created and tested. The tested source code or component also ensures the quality of the new application developed because error free codes are composed together to create a new software/application. Furthermore, the cost, such as human resources, also declines since unnecessary work is avoided.

PROBLEM WITH UNSTRUCTURED REUSE Although the reusability concept provides a lot of benefits, unstructured reusability, however, could also create a problem. The easiest way to reuse a source code is to copy and paste it to the new application being developed. This reusability method will cause a problem if the programmer has to modify the original source code, which means he has to change all of the copied and pasted source code piece by piece (Ng, 1992; Sametinger, 1997). OO for Reuse Coad & Yourdon (1991) noted that three important characteristics of objectorientation are class & object, inheritance, and communication through message. Two of these three characteristics influence the success of reusability of object. Class and object: An object is an abstraction of something in the domain of a problem or its implementation, reflecting the capabilities of a system to keep information about it, interact with it, or both; an encapsulation of attribute values and their exclusive services. Class is a description of one or more objects, describable with a uniform set of attributes and services; in addition, it may describe how to create new objects in the class. Class & object are relatively stable over time, and provide a basis for moving over time towards reusable analysis results (ibid.), for example, the mathematical function used for calculating the salary tax may change over time if the tax formula is changed, but the data represented as an object such as employee data will not change. Inheritance: The class & object can be related with other class & object by mapping. Mapping in object-orientation can be in whole–part relationship structure, generalization–specialization structure or message connection. Generalization– specialization structure can be described as “is–a” relationship, for example, chicken (sub-class) is a kind of bird (super-class), therefore chicken inherits the attributes and behavior of bird such as two wings.

Design-for-safety-process 215

The inheritance supports the reusability in which inheritance allows the sub-class to use the attribute and service of the super-class. The benefit of using inheritance for reusability is that modification at the super-class, source code, will automatically influence the sub-class. Reuse can be achieved not by modifying existing code, but rather by extending or specializing the classes found in the library, through the inheritance mechanism. Implemented attribute and service can be made to work with objects not foreseen at the time the attribute and service are created.

OBJECT-ORIENTATION CONCEPT IN World Up™ The object-orientated design is identified by class & object, attribute, and communication through a message (Coad & Yourdon, 1991). The class defines the behavior and properties. Objects are instances of a class and the way they behave is determined by the class they belong to. It is very important to understand the difference between classes and objects. In World Up™, the block, sphere, cylinder, text3d and imported items are considered as a class. An instance created—an object—from this class, for example, block-1 as a default name of block instance, inherits the attributes or properties of the block class. Therefore, block-1 has construction-related properties such as height, depth, width, and material, while sphere-1, as an instance of sphere class, has construction-related properties, such as initial radius, material, the number of latitudinal (north to south) subdivisions, and the number of longitudinal subdivisions, inherited from sphere class. An object can be created to have a more specific behavior from the parent class by creating user-defined properties. The user-defined properties can only be created by defining a sub-class called sub-type. For example, a sub-type, a sub-class, of block class must be created and an “ObjCtr” (object counter) user defined can be assigned for simulation purposes. In Figure 2, a building column class is created from the block class. The column class is assigned with more specific behavior than the parent class by assigning the ObjCtr as a user-defined property for simulation purposes.

BUILDING A UNIVERSE IN World Up™ A universe is a virtual world composed of pre-defined objects. The pre-defined objects could be geometry objects; such as spheres, cylinders, blocks, 3D texts, and imported objects; as well as lights. Primitive pre-defined objects, such as sphere, cylinder and block, can be created using World Up™. While a complex pre-defined object can be created from the modeling software provided by World

216 S. Rowlinson & B.H.W. Hadikusumo

Figure 2. A generalization–specialization concept in World Up™.

Design-for-safety-process 217

Up™ called World Up™ modeler or third party modeling software such as 3D Studio Max™. Among these three methods, modeling an object using primitive class provided inside the World Up™ is the most effective in terms of file size. Using third party modeling software would lead to a large file size and consequent degradation of performance, such as texture. An object created using third party modeling software can be used in the World Up™ universe to the extent that the file format of the object created are supported by World Up™. Some file formats which are accepted by World Up™ are Autodesk DXF format, Wavefront OBJ format, Autodesk 3D Studio mesh file (3ds), Pro/Engineer RENDER SLP file format, MultiGen/ModelGen Flight FLT file format, VideoScape GEO files, World Up NFF file format and binary NFF file format, and Virtual Reality Modeling Language (VRML) 1.0 WRL file format.

REUSING OF OBJECT IN A UNIVERSE DEVELOPMENT The nature of typical building construction projects, having similar construction components, provides benefits in the development of a universe in terms of reusing one object to create other similar objects. Reusing of building construction objects in World Up™ exists at many levels:

• Reusing of objects for different locations. In construction, no two similar components occupy the same location, i.e. two walls with the same size are installed at different positions. The reusability of an object depends on the type of object source itself. • If the object source is a simple object created using World Up™, such as sphere, block or cylinder, a sub-class representing the object can be created and used to create many similar objects with similar properties. The value of translation property can be modified to place the new object in the location desired. • If the object source is created by World Up™ modeler or a third party modeler, the object source can be reused by adding it to the universe through the resource browser (see Fig. 3). Reusing of objects with different dimensions. Construction components can be created by using primitive geometry, i.e. columns and beams can be created using the block, primitive geometry. The block geometry object source can be stored as one unit of length, width and depth. This one-unit object source is used in the universe as a column by stretching the length, width and depth. Several columns can be created by using this object source; however, modifying the value of the stretch property of any column, for example, also changes the other columns from the same object source, the siblings. In order to create a column, as an example, from the same object source and also have an independent length, width, and depth control, the column must be assigned as a ForkImported object, a function of World Up™.

218 S. Rowlinson & B.H.W. Hadikusumo

Figure 3. Object reuse using resource browser of World Up™.

SIMULATING THE CONSTRUCTION PROCESS Adjei-Kumi & Retik (1997) noted that present planning systems, which use CAD technology to generate and simulate the process of constructing a facility graphically, only go as far as the visualization of project schedules at the component level. They proposed to visualize the construction components and processes in Provysis. That visualization of construction processes, however, does not mimic the real processes of a construction project. Provysis only fades in the construction component assumed to be constructed and fades out the successor components, instead of showing how the object is transported and installed. Bick et al. (1998) developed a visualization system in the manufacturing domain. In this system, the simulation supports object translation from one location to another location in a specified time. The translation of the object is determined by using origin and destination of object location while time control is achieved by specifying when the process takes place, tstart and tend. In the real world, location and time are the most important variables to represent a dynamic condition. Changing the value of these variables determines where the object will move in terms of location and how long the movement is in terms of time. These two variables are also important for the purpose of simulating the virtually real construction process. The most general construction process is

Design-for-safety-process 219

Distance Phase 1 (D1) = ((X1-X0)2 + (Y1-Y0)2 + (Z1-Z0)2)0.5 Distance Phase 2 (D2) = ((X2-X1)2 + (Y2-Y1)2 + (Z2-Z1)2)0.5 Total Distance (DTotal) = D1 + D2 Duration of Transporting = Total = ObjCtr Phase 1 Duration (TD1) = (D1/DTotal) * ObjCtr Phase 2 Duration (TD2) = (D2/DTotal) * ObjCtr

Figure 4.

Linear distribution of translation duration.

translation and rotation of an object. Other processes such as installing and removing an object can be derived from the concept of object translation and rotation. For example, installing a pre-fabricated wall can be simulated as translation from the stock location to the destination location. In order to transport an object from the origin location to the destination location in a specified time, four user variables—origin of location, destination of location, starting time to move, and duration of translating—must be defined. Origin of location and destination of location are user-defined properties that can be represented as x0, y0, z0, xt, yt, and zt. The object will move by changing the value of its translation properties from x0, y0, z0 to xt, yt, and zt. The duration of transporting is used to display a smooth movement of the object instead of leaping from one location to another location. A more complex translation can be achieved by increasing the number of location user-defined properties, such as origin of location (x0, y0, z0), middle of location (x1, y1, z1), and destination of location (x2, y2, z2). This method creates two phases of movement, from the origin to the middle (phase one) and the middle to the destination (phase two). Since this method creates two phases of movement, the duration of object translation must be distributed linearly in order to make a constant speed of translation (Fig. 4).

COUNTERING SYSTEM AS A SIMULATION ENGINE In World Up™, the location of an object in the virtual world can be determined by setting the value of the translation property. Adjei-Kumi & Retik (1997) noted that the main purpose of Provysis’ process data, such as start times, finish times, duration, and graphical images, is to facilitate the visualization of the simulation of the generated construction schedule. They connected the VR application with projectmanagement tool, Primavera Project Planner. This system supports the usual

220 S. Rowlinson & B.H.W. Hadikusumo

practice of scheduling in the industry. But it does not explain how to control the time scale. Time scale factor is important to adjust a speed level so that user can analyze the VR simulation at a desired and controllable speed. If the simulation is too fast, it may cause dizziness to the user and it will be difficult to observe the simulation, while a slow simulation may cause impatience. In order to represent time in the World Up™ universe, a counter system is used. When the simulation is run, World Up™ goes through the entire simulation loop for each frame that the simulation is run (Fig. 5). The counter works every time World Up™ simulates the universe (Table 1). The counter is increased every time that a new cycle of simulation is started. Since the counter will be increased every frame, an object can be designed to translate whenever the object start

Sensors Update

Paths Play/Records

Tasks Executed Universe Tasks

Node Tasks (bottom to top, right to left)

All Other Object Tasks

Scene Rendered and Displayed

Figure 5. Simulation cycle of World Up™. Source: World Up™ user’s guide (1997).

Table 1. Scheduling comparison between real practice of construction and VR simulation. Activity

Real practice

VR simulation

Time control Task started

Calendar As specified date, i.e. 1 January 2000

Task completion

Number of days, i.e. 5 days

Counter As specified counter, i.e. when the universe counter reach 200 Number of counter, i.e. 100

Design-for-safety-process 221

counter (OSC), a user-defined property, is equal with the counter for a certain duration of ObjCtr, a user-defined property (Fig. 6). For example, an object will be triggered to translate or move, when the counter is the same as OSC, 300, for duration of 100 (ObjCtr). So, the object will stop at its last position when the counter is 400. The increment of the counter determines how fast the object moves. This increment can be determined by creating an acceleration factor (Acc F) variable, which is a time scale (Fig. 6). The greater the factor, the faster is the translation of the object. This time scale factor influences only the speed of counter calculation, not the number of frames per second at which the universe is rendered, which depends on the computation capability of the computer, number of polygons rendered, number of pixel filled, and the display card. The faster the frame rate is rendered, the better quality of graphic animation is performed.

Figure 6.

Countering system for simulation engine.

222 S. Rowlinson & B.H.W. Hadikusumo

DESIGN-FOR-SAFETY-PROCESS General Electric developed guidelines for product design in the 1960s, but a significant benefit was not realized until systematic Design for Assembly (DFA) was introduced in the 1970s. Other design for “somethings”; such as manufacturability, inspectability, and quality; were developed in the 1980s. Then, in the 1990s, Design for X-Ability (DFX) has been used as an umbrella for all of these terms. The potential of DFX is enhanced by the availability of a powerful representation tool such as VR, where design can be represented in 3D graphical data and a walkthrough function enabling the user to discuss any aspect of the object in a virtually real location. VR of components and processes of a construction project will be used to develop a DFSP methodology, which aims to point out the safety hazards inherited by the construction components and activities. The main idea of DFSP is derived from DFX. The DFX aims to design a product from many viewpoints or characteristics (Gutwald in Prasad, 1996) with the following benefits: achieving a product exhibiting better qualities of X, i.e. Design for Manufacturability (DFM), DFA, Design for Disassembly, Design for Quality and Design for Environment.

EARLY FAILURE DETECTION Since DFX is an umbrella, the methodology development depends on the specific domain of x-ability developed. For example, DFM is designed by utilizing theories, such as the Taguchi method, and design axioms. However, the most important consideration is that DFX must work as a guideline to evaluate the design. This can be achieved by developing the DFX as an online or offline purpose (Huang, 1996). An online DFX checks its data/knowledge base to ensure that the design decision being considered will not violate the DFX rules. On the other hand, an offline DFX evaluates design decisions after they are made. In this research, an online DFSP is chosen. The safety process will be designed by utilizing a data/knowledge base compiled from theory and axioms. An axiom is a proposition, which is assumed to be true without proof for the sake of studying the consequences that follow from it and the axiom cannot be proven, rather it must be assumed to be true until a violation or counter-example can be found (Sush et al., 1978). In this research, the axiom will be compiled from safety regulations and safety best practice. Several theories of accident causation exist such as accident proneness, goalsfreedom alertness, adjustment-stress, unconscious motivation, situational theories, domino theory, and epidemiological theories (Hale & Hale, 1972; Petersen, 1984). Adams’ domino theory of accident causation developed from Heinrich is considered the most relevant accident theory to enable safety hazard recognition. The main

Design-for-safety-process 223

reason is that Adams’ theory of accident causation discusses the tactical error, a tangible approach, which contains the unsafe conditions as follows (Adams, 1976):

• • • • • • • • • • •

improper design, construction or layout; decayed, aged, worn, frayed or cracked; unnecessarily slippery, rough, sharp-edged or sharp-cornered; unsafely stored or piled tools or materials; poor housekeeping or congestion; unsafe established procedure, inadequate job planning, improper equipment provided; inadequate aisle space; improperly guarded; improper illumination; improper ventilation; personal protective equipment not adequate or not available.

CONCLUSION The traditional approach using 2D drawings results in bulky and separated design drawings and philosophies. This problem leads to difficulties for the end-user in order to create a 3D mental picture for construction purposes. VR can integrate this design information into one universe and also add in design content such as the construction process, method, and safety planning. The reusability concept is important in developing a VR universe since it can minimize development time. Reusability can be achieved by using the concept of objectorientation; class & object, and inheritance. Inheritance can be created by applying the concept of generalization–specialization which is also supported by World Up™. The two important variables, location and time, are identified for the purpose of construction process simulation. In order to translate the object, the location variable must be defined as the origin of location and the destination of location. The time for object translation can be distributed linearly if the location consists of three points or more, such as (x0, y0, z0), (x1, y1, z1), and (x2, y2, z2) in order to obtain a constant object translation. Since time must also be represented for the simulation purpose, a counter system, counter ⫽ counter ⫹ Acc_F, is proposed which can simulate a time standard, such as a calendar, in real life. The universe counter will be increased by the factor of “Acc_F” (acceleration factor), which also functions as a time scale to adjust the speed of the construction process as desired by the user every time a frame is calculated by World Up™. Adjustable speed of simulation is important for user friendliness. If the simulation is too fast, the user might get dizzy and might not be able to analyze the simulation properly, but a slow simulation might cause user’s impatience.

224 S. Rowlinson & B.H.W. Hadikusumo

An example of the use of the model is shown in Appendix A. The model shows a typical 40 story Hong Kong Housing Authority residential block and the model has been used to conduct a safety risk assessment for such units. The walkthrough was found to be realistic and the hazards identified were automatically logged in the risk assessment.

REFERENCES Adams, E.E. 1976. Accident causation within the management system. Professional Safety, October. Adjei-Kumi, T. & Retik, A. 1997. A library-based 4D visualisation of construction processes. IEEE conference on information visualization, 27–29 August: 315–321. ISBN: 0-8186-8076-8. Bick, B., Kampker, M., Starke, G. & Weyrich, M. 1998. Realistic 3D-visualisation of manufacturing systems based on data of a discrete event simulation. Proceedings of the 24th annual conference of the IEEE, Industrial Electronics Society: 2543–2548. Coad, P. & Yourdon, E. 1991. Object-oriented analysis, 2nd edition. New Jersey: Yourdon Press. ISBN 0-13-629981-4. Faraj, I. & Alshawi, M. 1996. Integrating virtual reality functionality with traditional design tools. DOE Research Contract 39/3/193, Department of Surveying, University of Salford, Manchester, U.K. Furnham, A. 1998. Personality and social behaviour. New York: Arnold. Hadipriono, F.C. & Larew, R.E. 1996. Safety training in virtual construction environment. Proceedings of the 1st international conference of CIB working commission W99/Lisbon/Portugal, 4–7 September. Hale, A.R. & Hale, M. 1972. A review of the industrial accident research literature. London: Her Majesty’s Stationery Office. ISBN: 0113608950. Huang, G.Q. 1996. Developing design for X tools. Design for X: concurrent engineering imperatives. London: Chapman & Hall. ISBN: 0-412-78750-4. Lingard, H. & Rowlinson, S. 1991. Safety in Hong Kong’s construction industry. The Hong Kong Engineer 19: 38–44. Lingard, H. & Rowlinson, S. 1997. Behavior-based safety management in Hong Kong’s construction industry. Journal of Safety Research 28(4): 243–256. Ng, K.G. 1992. Reusable components for business information systems. Dissertation for the Master Degree of Science in Data Processing, University of Ulster. Prasad, B. 1996. Concurrent engineering fundamentals Volume 1. New Jersey: Prentice Hall PTR. Petersen, D. 1984. Human-error reduction and safety management. New York: Aloray Inc. ISBN: 0-913690-09-0. Sametinger, J. 1997. Software engineering with reusable components. Berlin: Springer. ISBN 3-540-62695-6. Sush, N.P., Bell, A.C. & Gossard, D.C. 1978. On an axiomatic approach to manufacturing and manufacturing systems. ASME Journal of Engineering for Industry 100(2).

Design-for-safety-process 225

APPENDIX A

1. Use of the model to conduct a risk assessment on a Hong Kong Housing Authority residential block.

2. The user is observing the project from the ground level.

226 S. Rowlinson & B.H.W. Hadikusumo

3. The user is on the lift area at core.

4. At the installation of wall reinforcement bars.

THE POTENTIAL OF 4D CAD AS A TOOL FOR CONSTRUCTION MANAGEMENT Robert M. Webb1, Theo C. Haupt2 1

Bovis Lend Lease, Charlotte, NC, USA M.E. Rinker, Sr. School of Building Construction, University of Florida, Gainesville, FL, USA

2

Abstract 4D CAD presents several opportunities for use as a tool for construction management with respect to the way that it links the temporal and physical spatial aspects of a construction project. Bovis Lend Lease has used 4D CAD to graphically represent the relationship between space and project schedule, through the actual transformation of that space over time during the construction of a building or facility. When it is considered that construction managers are constantly on the look out for effective ways to gain competitive edge in a highly competitive industry, 4D CAD has the potential to provide such an edge. 4D CAD provides the vehicle by means of which it is possible to integrate the functions, roles, responsibilities and relationships of, and between, all the participants in the construction process. This process is examined in this paper. Some of the problems, which need to be overcome, to make 4D CAD more attractive for construction management are also explored. Keywords: 4D CAD, construction management, visualization, computer simulation

INTRODUCTION Construction managers are constantly being bombarded by the need to make rapid and informed decisions in order to satisfy the traditional project parameters of time, cost and quality. Decisions need to facilitate completion within the project schedule and within the project budget while satisfying the desired quality requirements. As the number of technological options increases, so does the complexity and the cost of choosing which combination of available options is the most appropriate for a given application. Informed decisions involve the management of vast amounts of information about the combinations of available options and the simulation of their performance (Papamichael, 1999). To further exacerbate 227

228 R.M. Webb & T.C. Haupt

matters, the industry has become more complex due to several factors that include the greater use of specialist contractors, more off-site manufacture and assembly and the increased use of bespoke systems (Marsh & Finch, 1999). Additionally, manual methods are becoming increasingly difficult to implement at comprehensive levels. Consequently, decisions are made that are partially informed, resulting in missed opportunities, and unaccountable, undesired effects (Papamichael, 1999). These consequences are undesirable in the context of the increasing competitive environment of the construction industry. The rapid advances in information technologies and the continuously decreasing cost of computing power present promising opportunities for the development of computer-based tools that may significantly improve decision-making. The combination of the graphic potential of 3D CAD with the construction project schedule is commonly known as 4D CAD. The 4D CAD technology presents several opportunities for use as a tool for construction management with respect to the way that it links the temporal and physical aspects of a construction project. It graphically represents the relationship between space and project schedule through the actual transformation of that space over time during the construction of a building or facility. Techniques that are presently being used to manage the design, planning and construction processes of a building facility, abstract the processes to produce a Gantt chart or CPM schedule (McKinney et al., 1996). A more comprehensive tool that will simulate and visualize construction activity sequences as part of an interactive experience is preferable. The interactive 4D CAD model provides just such a tool, in terms of which design and construction planning alternatives and decisions are evaluated, optimized and justified within the context of space and time. Bovis Lend Lease has used 4D CAD prominently in the marketing, procurement and preconstruction phases of their construction operations, where it is primarily used as a visualization tool to help best plan a construction project (Fig. 1). Bovis Lend Lease has only recently begun using it in the actual construction phase. In this paper, these 4D CAD applications by Bovis Lend Lease are referred to, as well as a few others, with reference to their impact on construction management.

ANTICIPATED GAINS The most obvious anticipated gains from the use of 4D CAD are that it will give the entire construction project team of clients, design consultants and contractors, a more effective tool to:

• improve communication between them while facilitating informed decisionmaking; • facilitate the evaluation, implementation and monitoring of design changes;

4D CAD as a tool for construction management 229

Figure 1.

Bovis Lend Lease use of 4D CAD.

• evaluate alternative materials or other processes to be used in the facility being planned or being built;

• evaluate and develop the most effective material staging and handling proce• •

• • •

dures for the project; identify and develop alternatives when disruption to the original plan on the construction project occurs; effectively train, and communicate with, construction crews (as well as other government regulatory organizations, community groups, etc.) specially before engaging in an intricate, challenging, or hazardous activity or a new construction method or technique; monitor progress on the project by comparing as-planned with as-built; improve the use of just-in-time material deliveries which are particularly important on construction sites where space is at a minimum or premium; and help overcome language barriers for members of the construction team, especially in the context of international construction activities.

However, to be able to meet the challenges that these applications present, 4D CAD as a tool must be capable of producing interactive 4D models for 4D animation (McKinney et al., 1996).

CONSTRUCTION SCHEDULES 4D CAD enhances the communication, approval and improvement of construction schedules by various parties, such as construction managers, clients, designers,

230 R.M. Webb & T.C. Haupt

sub-contractors and community members. For 4D CAD to be implemented effectively, however, a considerable amount of detail work is necessary that is project specific, and in many cases unique. Too little detail may result in critical elements of a work sequence being overlooked, or in a lack of allowances for uncertainty (Riley, 1998). Each construction project schedule is normally unique to a particular project. Very rarely it is possible that a set of schedule standards can be applied universally to multiple projects. By implication, the effort to associate the element of time or schedule with every single component in the 3D CAD system for a single project has to be repeated for each successive or new project. This process is further exacerbated by the many factors that impact a project construction schedule. Consequently, its universal incorporation becomes increasingly more arduous and difficult. Some of the factors identified during a survey that was conducted among numerous schedulers, superintendents, project managers and other leaders within Bovis Lend Lease, are listed below in random order:

• changing weather and site conditions; • availability of labor and materials, and use of specialty components in • • • • • • • • • • • • •

facility; regulatory requirements such as agency approvals and permitting; experience of project team, changes in project team during project, and project team preferences; owner involvement, approval and sign-off procedures (how much and fast), critical decisions and activities; technical complexity of facility components, site logistics and staging capabilities; trade contractor or sub-contractor performance; amount of design changes, speed with which changes to design are handled, schematic design, estimate and approval duration; involvement and requirements of community groups and trade unions with respect, for example, to minority requirements; legal contract requirements, bidding phase procedures, and documentation (volume and complexity), project budget and funding; environmental factors and compliance requirements; availability, condition and dependence on local utilities or infrastructure such as roads and rail; amount of value engineering process; site security; and safety requirements/accident impact.

Considering these factors, it is not surprising that contractors have not harnessed the potential that is presented by 4D CAD. For this scenario to change, the 4D CAD tool must empower construction designers, schedulers, superintendents and project managers to develop a project schedule directly related to a 3D model of the building facility (McKinney et al., 1996). At the same time, it is also necessary for the spatial and temporal relationships involved in the project, to be understood.

4D CAD as a tool for construction management 231

DEPLOYMENT OF 4D CAD STRATEGY The simple “visualization” use of 4D CAD is becoming relatively commonplace today. To date, Bovis Lend Lease and many other companies have deployed it across multiple market sectors and project sizes around the world. Some of the more common market sectors where it has been more routinely used include:

• • • • • •

Industrial, Semi-conductor, Pharmaceutical, Healthcare, Office Mixed Use, and Telecommunications.

Deployment of the schedule integration for actual construction is proving to be challenging and hard to achieve. Consequently, deploying, with schedule integration, remains very much an inexact science. Significantly more upfront commitment of project planning, resources and finances is needed. Very few projects globally have been completed using full 4D CAD schedule integration. While this is true, 4D CAD with schedule integration is feasible on any project, and makes the most long-term sense on the larger and more complex ones.

4D CAD AS A VISUALIZATION AND SIMULATION TOOL Several factors influence the selection of appropriate construction methods and related resources that impact the traditional project parameters of time, cost and quality (Liu, 1996). These factors include:

• • • • •

job specification, job size, site conditions, materials, and availability of equipment and suitably skilled labor.

Computer simulation provides a practical visual means of modeling construction activities and operations and identifying the characteristics of these before they actually begin (Liu, 1996). This simulation enables the creation of models that provide manipulative opportunities to construct situations that would not have been otherwise accessible (Horne et al., 1999). Experimenting with “what-if” scenarios on the computer model to arrive at the optimal operational plan can test decisions, and alternatives.

232 R.M. Webb & T.C. Haupt

Figure 2.

Month 4 of the project schedule.

Unfortunately, most of the available simulation programs make this objective difficult to achieve. They were originally developed by researchers, for research purposes, and are not easy to use (Papamichael, 1999). They require significant amounts of detailed information about the building and its context, are very expensive to use due to the time required for the preparation of input and interpretation of the consequent output. Bovis Lend Lease used an early generation of a 4D CAD visualization tool on its Lynchburg General Hospital project in the early 1990s. This hospital was preparing to undergo a renovation and addition to its existing facility. The visualization tool proved to be very effective in helping the client and the entire project team plan and understand the sequencing of the work in such a way as to cause the least amount of disruption, and yet complete the project at the lowest possible cost. More recently, Bovis Lend Lease has used 4D CAD visualization to facilitate the communication and planning for several projects in New York City. In an area like New York City where staging and logistics are particularly challenging, this application has been particularly helpful since nothing was left to the imagination in the visual presentation. The following graphics illustrate how it was used on one of the engagements. By clicking on the date in the construction schedule, the construction activities that are scheduled or planned to be in operation will be graphically represented in 3D. In actual fact, Figure 2 illustrates the erection of cranes in month 4 of the project schedule. At this time the foundations were nearing completion. Figure 3 shows month 15 of the project schedule indicating the expected project progress at that stage. Operations evident in this particular graphic, were the steel top out, setting boilers and cooling towers as well as con-ed permanent power. Month 22 of the project schedule is highlighted in Figure 4, where the operations included the removal of sidewalk bridges and the removal of the hoists.

4D CAD as a tool for construction management 233

Figure 3.

Month 15 of the project schedule.

Figure 4.

Month 22 of the project schedule.

It is clearly evident from this example how effectively 4D CAD can capture, and dynamically manage, the interaction between project components and resources over time, visualize these interactions, and support the real-time interaction of users with the 4D model. With this tool, it is possible to enhance the communication, approval and improvement of construction schedules by various participants in the construction process, such as construction managers, clients, designers, sub-contractors and community members. It is also possible to demonstrate the environmental impacts of the proposed project in order to allay fears, and gather support for it. Another one of the popular software packages on the market that is used for 4D CAD visualization allows the project team members of a project to streamline

234 R.M. Webb & T.C. Haupt

Figure 5.

Parallel workflow.

parallel workflow (Fig. 5). In a typical project environment, detailed engineering and scheduling run in two simultaneous yet independent work processes. This application integrates the design and construction planning disciplines, improving constructability and shortening the time from project concept to project completion. Overall, this particular 4D CAD application can potentially provide the following benefits:

• definition of the scope of projects at project proposal and conceptual design • • • • • • •

stages; early development of a construction and commissioning or start-up method statement; involvement of construction and commissioning personnel in conceptual design stage; full investigation of design, constructability and commissioning issues prior to commitment of costly site resources; improved design and procurement strategies; better focus on pre-fabrication, pre-assembly and just-in-time procurement; smooth materials management and handling; and exploration of alternative dispute resolutions.

Bovis Lend Lease used a fairly specialized 4D CAD visualization tool on a project in Sydney, Australia. On this 50-story tower, there was a need to keep the reinforced concrete frame as light and open as possible while still being adequate for wind and other general structural requirements. The Strand 7 software tool was

4D CAD as a tool for construction management 235

Figure 6.

Bovis Lend Lease project in Australia.

Figure 7.

Bovis Lend Lease project in Australia.

used to model the core structure. Wind was then introduced in time intervals to test for flaws in the design. Figures 6 and 7 illustrate some of the individual sequences from this visualization tool. This process added significant value resulting in a structurally sound building being constructed without the traditional amount of reinforcement that might have been required. The open, airy appearance of the building added to its aesthetic. The addition of open space throughout the structure was an added benefit.

ILLUSTRATION OF 4D CAD WITH SCHEDULE INTEGRATION Bovis Lend Lease is actively using 4D CAD with schedule integration on a mixed use office park in London. This engagement has the essential ingredients for allowing 4D CAD to make a meaningful contribution; a very supportive team

236 R.M. Webb & T.C. Haupt

comprising of owner, design team, construction manager and key trade contractors. The initial objectives of the project team for this effort were as follows:

• to graphically represent planned progress element by element; • to integrate graphical representation in time terms with current planning systems; • to involve trade contractors to a greater degree in the planning process by giving them the opportunity to test their individual programs on the model;

• to have greater discipline in the process with both planners and trade contractors; • to provide a tool to graphically test on-site program scenarios and recovery programs; and • to provide historical records with respect to progress achieved and the reasons for that progress. The following elements were built into the 3D models for all three buildings making up the project:

• • • • • • • • • • •

all foundations for the concrete columns, steel core and external steel columns; all concrete columns for the undercroft to the third floor level; all floor slabs; steel core; external glazing; external staircases, steel bracing and external metal sun shade louvres; roof mounted mechanical plant; all external glazing; raised floors; wall construction around the steel core; and buildings to be positioned according to the site plan.

Figures 8 to 10 illustrate several specific times during the sequencing of one of the office buildings in the project. Figures 11 to 13 are another illustration of project sequence on the office project in London.

Figure 8.

Bovis Lend Lease project in the United Kingdom.

4D CAD as a tool for construction management 237

Figure 9.

Bovis Lend Lease project in United Kingdom.

Figure 10.

Bovis Lend Lease project in United Kingdom.

Figure 11.

Bovis Lend Lease project in the United Kingdom.

238 R.M. Webb & T.C. Haupt

Figure 12.

Bovis Lend Lease project in the United Kingdom.

Figure 13.

Bovis Lend Lease project in the United Kingdom.

LESSONS LEARNED FROM 4D CAD WITH SCHEDULE INTEGRATION The time required to build a model The problem of design complexity versus programming time is one that poses the greatest risk to the successful development of the 4D CAD system being ready in time for use on a project. The current models are in a CAD platform and then animated, allowing them to operate a significant level of detail. This level of detail, however, carries with it a lengthy programming period. The fact that computers can handle complexity does not mean that there is no need to design for simplicity. The model does not necessarily have to be as powerful as the CAD program. A simple stick model which can be built and operated

4D CAD as a tool for construction management 239

by a contractor operative may be more usable and adaptable to change than a cumbersome all singing, all dancing model. Intricate details are not necessary on a 4D CAD tool. These details can be viewed adequately on detailed design drawings from either the trades or consultants. Simplification of the model With this point in mind, a simplification of the system currently being developed by the architect, sacrificing detail for usability, may offer a product that is more functional than the current envisaged format. The adaptation of the model currently being developed by the architect is a result of a hybrid of ideas from multiple sources, i.e. contractor, owner, consultants and architect. Indeed, simplification arises from knowing the job intimately; from a stand-back perspective or even from an innocent look at the project from the eyes of an outsider. Alternatively, the current system of modeling might be used on projects where the design is more complete and perhaps more basic. The design and build form of contract where the contractor heavily influences the design, may have more merit than a Contract Management contract where the design is still evolving when construction has often started. Project team summary observation Given the complexity involved in building a 4D CAD model with the level of detail and chronological animation used on this project, impracticalities might occur in the sheer amount of time required to produce such a model. As design is very rarely 100% complete prior to construction, it is unrealistic to expect the design to be sufficiently evolved early enough before construction to allow a detailed model to be built. Perhaps this product should be viewed initially as a problem-solving tool for complex interface areas on projects. Then, by following the natural evolution of a product in regular use, the ease and speed of operation should increase and the role could be expanded.

OPPORTUNITY FOR CONSTRUCTION PROCESS IMPROVEMENT The construction industry reputedly still suffers from one of the highest waste factors of all industries. It has been estimated that 25% of building costs in the United States are due to waste (CIOB, 1994). The introduction of 4D CAD technology into construction management could help reduce construction waste significantly. If the use of 4D CAD tools could reduce the cost of waste by 10%, the savings to the US $380 billion construction industry would be approximately US $38 billion, enough incentive to actively pursue 4D CAD or other efforts to reduce this waste factor.

240 R.M. Webb & T.C. Haupt

The Construction Industry Institute (CII) found, in a specific study of industrial projects, that the average cost of rework on industrial projects exceeded 12% (CIOB, 1994). For the projects studied, deviation costs averaged 12.4% of the total installed project cost. However, the same study revealed that not all of the deviations on a project were recorded. For example, construction changes made at the site were often not included in format reports. It was not unusual for errors to be made good and/or accepted immediately rather than expending the time and effort to file formal requests. The deviation data gathered included only the direct cost of the rework for the item in question and included no indication of impact on the rest of the project. It is therefore concluded that both the number and costs of deviations reported for the projects in this study are conservative estimates of the actual values. The two major categories resulting in deviations were design and construction. By managing and tracking design and construction changes effectively using 4D CAD, a fair proportion of the costs of construction waste can be reduced and even eliminated.

CAD SOFTWARE SELECTION CONSIDERATIONS According to a study conducted by Horne et al. (1999) in the United Kingdom, the following key independent variables for CAD software selection criteria were identified:

• modeling capabilities with respect to accuracy, ease of use and surface characteristics; • visualization capabilities with respect to high quality, animated images. In the same study, the dependent variables identified included:

• credibility of the data; • comparability with other representational methods; • appropriateness to differing needs of interested parties such as, for example, • • • • •

clients, designers, constructors, and sub-contractors; reliability; communicability; practicality in terms of time and other resources; reproducibility; and generic applicability.

The selection of the most appropriate visualization and modeling software is problematic because of the many features in different combinations, and incomparable user interfaces. Different simulation programs use different representations of buildings and their context, depending on the performance aspects that they address.

4D CAD as a tool for construction management 241

CONCLUSION While 4D CAD has potential for improving construction management, providing competitive edge, reducing construction waste and costs, contractors have not yet accepted this potential on a large scale, especially in the area of construction management. The 4D CAD tool empowers construction designers, schedulers, superintendents, and project managers to develop a project schedule directly related to a 3D model of the building facility while at the same time, facilitating the understanding of the spatial and temporal relationships involved. Furthermore, 4D CAD technology will need to be available at a more affordable cost to enable it to be applied to jobs other than large, complex projects, both with respect to physical enormity as well as dollars. Additionally, consideration has to be given to how the geometry of architectural form and structural design produced by architects and engineers can more easily form the basis for 3D representation while being linked at the same time to a construction schedule on a fully integrated basis. This aspect is essential if 4D CAD is to become the effective tool that it has potential to be with respect to construction project management. Interoperability of software is essential for the continued development and deployment of 4D CAD. Central to this effort is the work being done by the International Alliance for Interoperability (IAI). This group is supporting standards that allow objects to transfer seamlessly from one application to the next. Additionally, the aecXML effort is quite important to the continued development of 4D CAD. This technology provides an effective, cross-platform, cross-application transfer of defined information objects. Other areas of concern include the process of performance evaluation, the complexity of design information with respect to matching design and context parameters that are in conflict, and information overload caused by each decision being dependent on a large number of other decisions. Essential to the continued evolution and benefits of 4D CAD is the improvement of the environment in which it is to be used. 4D CAD needs to be promoted at all levels of the industry, including owners, contractors, designers, consultants, trade contractors. By having more projects online doing electronic collaboration will likely also increase the potential use of 4D CAD. While the potential benefits of 4D CAD for construction management are not in dispute, the challenge faces software designers to produce an integrated system which can effectively address the concerns raised on a cost- and time-effective basis. Collaborative efforts across the various building construction related disciplines are necessary to realize the overall vision of a computerized building industry. The ideal with respect to construction management would be multiple simulation tools and multiple databases that are all interoperable in a distributed, networked environment between all participants in the construction process and beyond, to the eventual end-user.

242 R.M. Webb & T.C. Haupt

REFERENCES Horne, M., Hill, R. & Giddings, R. 1999. Visualization of photovoltaic clad buildings. Building Research and Information 27(2): 96–108. Liu, L.Y. 1996. ACPSS—Animated construction process simulation system. Computing in civil engineering; Proceedings of the third congress, Anaheim, California, 17–19 June: 397–403. New York: American Society of Civil Engineers. Marsh, L.E. & Finch, E.F. 1999. Using portable data files in the construction supply chain. Building Research and Information 27(3): 127–139. McKinney, K., Kim, J., Fischer, M. & Howard, C. 1996. Interactive 4D CAD. Computing in civil engineering; Proceedings of the third congress, Anaheim, California, 17–19 June: 383–389. New York: American Society of Civil Engineers. Papamichael, K. 1999. Application of information technologies in building design decisions. Building Research and Information 27(1): 20–34. Riley, D.R. 1998. 4D space planning specification development for construction work spaces. Computing in civil engineering; Proceedings of international computing congress, Boston, Massachusetts, 18–21 October: 354–363. Virginia: American Society of Civil Engineers. The Chartered Institute of Building (CIOB) 1994. Constructing total quality handbook: 7. Berkshire: CIOB.

VIRTUAL REALITY: A SOLUTION TO SEAMLESS TECHNOLOGY INTEGRATION IN THE AEC INDUSTRY? Raja R.A. Issa M.E. Rinker, Sr. School of Building Construction, University of Florida, Gainesville, FL, USA

Abstract A construction project is often divided into work packages because of its complexity. Although the construction of a big project becomes easier in a specialized industry, it also brings difficulty to the communication and cooperation between the participants of the project. It is proposed that a computer-integrated system may reduce the downsides brought by the fragmentation of the construction industry and improve the productivity and efficiency of the construction project. Different models of integration have been suggested, however, a uniformly accepted integration system has not yet been defined. The introduction of virtual reality (VR) technology into integration research may provide a general solution to this dilemma. A VR platform supported by knowledge-based database systems can become the main interface to construction information for every specialty throughout the construction (life) cycle of the project. All major application packages would be developed under or integrated in the VR system. As a consequence, we can foresee a marked decrease in legal disputes among the owner, architect, and constructor because of misinterpretation of design drawings and specifications and unmet owner expectations. Keywords: construction industry, integrated construction environment (ICE), project modeling and integration, immersive, non-immersive, virtual reality

INTRODUCTION In order for virtual reality (VR) applications to be successfully implemented in a complex industry such as construction, they must be part of a vertically integrated construction environment (ICE). Whether immersive or non-immersive techniques are used in the VR applications, users must be able to visualize design and construction information in 3D, photo-realistic, and interactive images. The user 243

244 R.R.A. Issa

must also be able to interact with external applications at real-time, thus, allowing VR systems not only to be used as presentation tools, but also as a universal interface for all construction applications. Finally, construction professionals must be able to view, alter, test, etc. any function or part of the proposed design and at any stage of the project life cycle through the virtual space. Due to the magnitude and complexity of construction projects, the traditional way of doing business in the construction industry is to divide the whole project into work packages according to well-established specialization. The work packages are assigned to specialty designers and contractors respectively. Although a system like this brings significant benefit to the industry, it also results in difficulties in communication and it requires extensive collaboration among the participants of the project. The communication between the segments of the project relies mostly on drawings and specifications. Project participants acquire from these paper-based media information only relevant to their own specialty. Confusions and delays often occur due to the abstract nature of the said media and the process of constant reinterpretation by the project participants. Although computer applications in every specialty benefit the industry very much, most of these applications can only keep information integrity inside their specific areas. The communications between these independent systems are very limited and sometimes frustrating at best. An established concept, Computer Integrated Construction (CIC) may provide a solution to this dilemma. Teicholz & Fischer (1994) defined CIC as “a business process that links the project participants in a facility project into a collaborative team through all phases of a project”. The process included in this concept covers the whole duration of the project from design, construction to facility management. The main purpose of CIC is to facilitate information exchanges and collaborative efforts among the project participants. A summary of the objective of CIC was given by Teicholz & Fischer (1994) as: (a) rapid production of high-quality design, (b) fast and cost-effective construction of facility, (c) effective Facility Management. VTT (1998), the Technical Research Center of Finland, proposed the interesting analogy of the current integration research in the construction area as shown in Figure 1. The independent computer applications in specific areas like design, construction and project management, which shows the fragmentation of the construction project, was referred to as “Islands of Automation” or “Islands of Information”. The contour line is actually the time axle. The current coastline means the frontier of the research and applications at present, while the coastline of 2000 was the goals that the researchers may achieve before the next century. With the advances of the computer technology, breakthrough of some key concepts, and the effort of both researchers and industry practitioners, “the water level has dropped” (Froese, 1994), and bridges are built between the islands. This process will eventually lead to a “unified continent”, an integrated construction management (CM) system. Figure 1 is an imaginative description of the evolving process of integrated computer applications in construction industry.

A solution to seamless technology integration in the AEC industry? 245

Figure 1. The islands of Information (from VTT, the Technical Research Center of Finland). Table 1.

The development of integration.

Time

Architectural Engineering design design

1960s 1970s 1980s

– – 2D drawing

2000 and beyond

Construction

Structural design Accounting, data management Parametric component design – CAD in drawing, prefabricated Quantity calculation, component modeling production planning ISO STEP, VRML, Internet, EDI, DXF, Information Broker …

The VTT analog can be converted into a tabular format, as shown in Table 1. It provides a historical perspective for the general scenario. The first use of computer in engineering design was in the area of structural design. Although some sophisticated structural analysis theories were developed long before that time, they could not be implemented without powerful computers. The computerization of accounting systems that happened in the 1960s was the first full-scale acceptance of

246 R.R.A. Issa

computer application in business management. At the same time, computers also began to be used in construction process. The development of computer hardware advanced in the 1980s to enable higher graphic processing ability. CAD, which stands for “Computer Aided Design”, became popular. However, it is more like “Computer Aided Drawing” in most cases. The software that was used to generate drawings for the construction or manufacturing process had difficulty exchanging information with other software, such as structural analysis or other computation software. Estimating software, which is closely related to accounting system, came into use in the 1980s. During this time the CPM method was computerized as well and became the mainstream of construction planning software. The great advances in technology integration came in the 1990s. Our whole society was affected by the fast development of computer technology. CAD technology advanced from 2D drawings to 3D visualization and VR using VRML became possible. It was recognized that the lack of integration between computer applications in this area could become a major obstacle to the further development. Integration plans were proposed and tested in order to achieve full-fledged production automation under the control of a unified computer system. The rapid expansion of the Internet has resulted in numerous possibilities and opportunities for the construction industry to make improvements to many aspects of its business operations. Some new areas of applications started emerging, such as product databases, and facility management archives. The current active areas of standardization research include:

• ISO STEP (Standard for the Exchange of Product Model Data): BCCM Core • • • •

Model, Express, etc.; CORBA (Common Object Request Broker Architecture) from OMG (Object Management Group); IFC (Industry Foundation Classes) by IAI (International Alliance of Interoperability); PDMS (Product Data Management); Multimedia (Video and Audio), Internet, VR, and DXF.

These standards form the three important sub-areas of computer integration research:

• Integration of the existing applications: The major purpose is to establish information standard between computer applications developed independently. It is like the transportation facilities between the islands shown in Figure 1. • Investigation of new specialties: Some applications are specialized in new specialty areas, such as product databases and facility management. • Introduction of new concepts and utilities from computer science: The latest developments in computer science give rise to new possibilities for solutions to current problems in the construction industry. Examples of such developments include, VRML, Internet, XML, EDI, DXF, etc. These new developments appear as objects floating around the islands in Figure 1.

A solution to seamless technology integration in the AEC industry? 247

HISTORICAL PERSPECTIVE The history of Computer Aided Project Management (CAPM) can be traced back to the 1950s. The ultimate aim of such a system at the time was toward creating an integrated management entity for the construction project. The actual developmental process did not go as predicted, however. Constraints from both technical and managerial aspects hindered further investigation. The pursuance of integrated systems began in the past decade again because of the new possibilities brought up by the following changes in construction industry and advances of computer technology.

CONSTRUCTION INDUSTRY NEEDS The construction industry was long considered slow in adopting new technology. It “has viewed innovation with suspicions or attempted to protect new thinking by protectionism” (Brandon & Betts, 1997). While the manufacturing industries improved their productivity and quality by leaning production and applying worldwide manufacturing benchmarking studies of production standards, the construction industry remained low profile. Once the construction industry realized this, major companies in the industry along with research institutions began to investigate a solution that might bring profound innovation to the whole industry. This lead to the setup of an international network, Construct IT, which represents one of the most productive efforts ever made in the quest for integration in construction. A major purpose of this network is to promote the application of information technology, system integration and standardization in the construction industry. The development of these applications is a necessary step toward a fully integrated construction industry.

THE CHANGING STYLE OF PROJECT MANAGEMENT The evolution of project delivery systems brought changes to the general style of project management. The basic project delivery system, which has a long history, is called the “traditional method”. It can be described as a process of “design– bid–build”. The owner, architect, and contractor are three independent parties bonded by contractual or administrational relations. Some new delivery systems, which were referred to as “alternative methods”, emerged in the practice of the past decades and began to challenge the domination of the traditional method. These systems include: CM, design-build, and BOT.

248 R.R.A. Issa

A notable feature of the change is that the later systems tended to have more centralized management (Clough, 1986). The project management responsibilities were conveyed to another independent party in the CM method, the construction manager in construction stage. The project management team, consisting of the architects, contractor, and owner’s representative, is headed by the construction manager. In the Design-Build method, a single contract including both design and construction responsibilities, is awarded to a “design-builder”. The design-builder is responsible for controlling the project activities for the duration of the project. While in the BOT method, the financing, operation, and limited time ownership appear in the job tasks of the builder.

DEVELOPMENT OF COMPUTER TECHNOLOGY The information produced from a construction project can be enormous because of the complexity and large scale of the construction project. It can be extremely hard to manage construction activities in an integrated manner without the help of computer facilities. The major factors that influence the further progress of integration research include: The computational capability of computers: Graphic and database application need the support of higher process ability. Hardware cost: Sharp decreases in hardware prices make possible the expanded usage of computers in the construction industry. The concept of databases: Orderly organized information provides efficiency and increases productivity. Networking: Brought a revolution to the method of communication in construction, which is crucial to the cooperation and coordination in construction process. AI and neural networks: Added further strength to the integrated construction system. The emergence of new concepts and methodology like Object-Oriented languages and databases, and the Internet are providing even more possibilities for the construction industry to integrate its systems. Sometimes it is just a matter of using our imagination to discover new potentials of computer in construction domain.

CURRENT RESEARCH The integration issues in construction were investigated intensively in Nordic countries, such as Sweden and Finland. Some of their publications are leading studies in this area. Their research actually defined the structure and trends of the integration

A solution to seamless technology integration in the AEC industry? 249

research. The major institutions endeavoring in this area in North America include CIFE at Stanford University and the University of British Columbia. VTT, the Technical Research Center of Finland, Finland VTT, the Technical Research Center of Finland, is an expert organization that carries out technical and techno-economic research and development work. There are two research groups in VTT related to CIC, RATAS and Project Planning and Building Design group. The active researchers include Matti Hannus and Mika Lautanala. Center for Integrated Facility Engineering, Stanford University The Center for Integrated Facility Engineering was founded in 1988 as an industry affiliated program of the Departments of Civil Engineering and Computer Science within the School of Engineering at Stanford University. The center is working on applying information technologies to the construction industry to improve integration in the construction process from the design to the management of the constructed facility. The research involves a wide range of technical, social, economical and managerial issues. The major topics explored include: CAPM, the strategy issues of the CIC, and information exchange standard. CIFE has many publications and has made great contributions to the establishment of some basic concepts in this area. Key researchers include Paul M. Teicholz, Hans Bjornsson, Raymond Levitt, and Martin Fischer. Department of Civil Engineering, University of British Columbia, Canada The integration research is very impressive due to the efforts of T.M. Froese, who received his Ph.D. degree in Civil Engineering from Stanford University in 1992. One of their research interests is the design of integrated, computer-based decision tools to support project design and construction. Their major works include the development of improved tools for modeling projects, representing and selecting construction technologies, encoding construction expertise into systems, automating the interpretation of construction records, and capturing multi-media project information. Findings from this work have been put into practice on many construction projects. Other research within this area focuses on information sharing and the integration of project functions throughout the construction life cycle. A major methodology of their integration research is to introduce new concepts of computer science and technology, such as Object-Oriented database principles, into construction industry.

VIRTUAL REALITY: A SOLUTION TO INTEGRATION? The basic concept of VR is to model the shape of the objects in three dimensions. The idea of VR appeared decades ago, but the inferior ability of the primitive

250 R.R.A. Issa

computers at the time hindered data-intensive implementations. The price of equipment was so prohibitive that the application of VR had to stay in a virtual status. However, VR does have some advantages that put it among the most promising solutions to implement system integration. The “ideal” solution A VR Integrated Construction System can be expected to

• enable designers, developers, and contractors to use the VR system and virtually test a proposed project before construction actually begins; • offer “walk through” view of the project so that problems can be found and design improvements can be made earlier; • provide free flow of information between CAD systems and other applications work packages by professionals in industry, minimize the misinterpretation between participants in the project, especially between designers and clients; • facilitate the selection of alternative designs by allowing different plans to be tested in the same virtual world. In a VR Integrated Construction System, VR becomes the main interface for all application packages and construction information for every specialty throughout the construction (life) cycle of the project. Two ways of interacting with a VR world There are two approaches to implementing a VR World: immersive and nonimmersive. In an immersive approach, the user is surrounded by the virtual world through curved screens and body suits or headmounted devices (HMD). The audio and visual perception of the user will form a virtual world. The non-immersive approach, also known as desktop VR, enables users to interact with the virtual world with conventional devices such as a keyboard, mouse and a monitor. Although this does not give the same level of spatial awareness as the immersive approach, it does provide users with a low cost solution and does not require the use of the HMD. This solution seems to be an attractive compromise for many users who are uncomfortable about spending a long time in a helmet (Issa, 1999). VR can be interpreted as a bridge between subject and human perception. These two ways of implementing VR provide solutions from two ends of the bridge. The immersive approach makes human perception its focus, while the non-immersive approach started from the description of the subject. The distinctions between these two styles of VR may eventually diminish with technological advances. But for the current investigation of VR in construction, the non-immersive approach seems to be more applicable. Problems of current VR systems Currently the two major areas of functionality of VR in construction are interaction with objects in real-time and walk-through presentation. These features are

A solution to seamless technology integration in the AEC industry? 251

mainly about visualization and simulation, instead of providing a basic interface between users and the project (subject). Most of the time VR systems are just supplementary to CAD packages. They cannot perform standalone design let alone be the bases of 2D drawings and all engineering design. Lots of implementation problems come from the supplementary role of VR systems, and include difficulty in use, requirement of special skills, and expensive to implement. These problems, which mainly come from the lack of integration between application packages, constitute tremendous barriers to the implementation of VR systems in the real world. What is needed to make it happen? To make the dream of VR come true, a scheme similar to the following needs to be set up:

• VR must become the general interface among the different applications instead of their individual interface.

• 2D and 3D images must become not just a way of presentation, but more importantly they must become interface for interactivity. • A central core which is a database system (most likely a knowledge-based database system) will be the basis of the whole VR system, the application and the interface. • The VR integrated construction system must be able to reside on a communications network (the Internet or more precisely the WWW). A serious challenge to the actual deployment of a VR system is whether an Industry Standard is developed or not. Before a complete solution can be provided to the user, the industry must be persuaded to adapt and move to a totally new, standards-based system.

VR APPLICATION PROTOTYPES Construction material specification integration The integration of construction drawings, design and material specifications within a VR environment allows the AEC professionals and the owner/procurer of construction services to preview the final product of their effort. This preview allows the participants in the project to more realistically determine the soundness of the design; the appropriateness of the construction techniques and the adequacy of the facility and material finishes in meeting the owners needs, prior to the execution of the project. Consequently, the expectations of the parties will be more realistic and the risk of costly disputes will be reduced considerably.

252 R.R.A. Issa

Collaborative virtual prototyping Even though VR-based tools can be useful at every stage of the construction process (to convince clients, to design the project, to organize and follow the construction site, etc.) important applications are related to the “design phase”. Decisions taken during the early design phase are of paramount importance due to their possibly dramatic effects on the final project, timing and costs. Virtual prototyping allows architects, engineers, contractors, and clients to create a design and evaluate it simultaneously for function, cost and aesthetics very early in the design process. The visual capabilities and the interactive inspection features offered by VRbased tools are much more extensive than those offered by standard CAD tools. Furthermore, coupled with distributed technologies such as STEP and CORBA, VR tools offer cooperative capabilities very useful in the design, by geographically distant teams, of large engineering projects. In that case, the virtual prototype can be considered as the starting point of the design process. After the first stage where the design teams test and validate the virtual prototype, relevant data is extracted from this prototype and is fed into CAD/CAM tools in order to be completed with more technical and detailed data (Fig. 2). Link with CAD tools The reverse process (i.e. extract data from CAD/CAM tools in order to visualize objects in a VR tool) is also possible. Nevertheless, it requires a fair amount of simplification (for evident reasons of performance optimization, detailed data cannot be

Figure 2.

Collaborative virtual prototyping (CIB, 1999).

A solution to seamless technology integration in the AEC industry? 253

fed into VR tools as a whole). Furthermore, existing techniques of simplification (polygonal reduction, re-meshing, etc.) still have some limitations particularly for granularity management (a small component that highly effects the virtual scene, e.g. a key hole when simulating lighting effects in a dark room, might be suppressed in an automatic re-meshing procedure). CAD models aim to represent the geometry of components for their manufacturing or for executing physical simulations (deformations, thermal analysis, etc.) by using methods such as the Finite Elements Method. On the other hand, VR models aim to represent objects visually, in order to interact with them. CAD models, therefore, can only be used within VR platforms after being processed by optimization procedures such as tessellation. Tessellation can be described as the processing of a 3D model in order to reduce the number of triangles of the model while maintaining an acceptable visual aspect. This procedure has some limitations:

• it could change the frontiers on the components of the initial model (which might be a problem when, for instance, two components should keep a perfect fit); • it is rather limited in handling gaps and intersections in the model. In both cases, manual corrections are usually needed to rectify the simplified model before using it in a VR application. Furthermore importing CAD models within VR tools usually yields a model where some of the facets are missing. This is due to the fact that, in CAD tools, a common way of constructing 3D models is based on symmetry (i.e. only half of the model is described and the other half is deduced by using symmetry axis). The “symmetrical copy” of the 3D model would be identical to the original one but would have inverted normals. When imported into a VR platform that uses backface culling for optimization issues, the “symmetrical copy” of the model will not be visible. A manual action from the user is then needed in order to invert the normals of the model. An interesting optimization tool for CAD/VR coupling is CAD-Real-Time Link from Prosolvia Clarus (http://www.clarus.se). VR applications for detailed design During the detailed design phase, virtual prototyping tools will allow the design office to refine the design proposed by the architect by adding constraints and modifications induced by the technical calculations (structural, thermal, lighting, etc.):

• Acoustics: The results of acoustic calculations can be related to the sound going through a window or a wall or the sound inside a room (e.g. a meeting room). These results are usually 3D sound WAV files associated to the related building components. • Lighting: Different lighting calculation methods can be used. The most effective ones are based on radiosity computation and raytrace rendering.

254 R.R.A. Issa

These methods combined give a high realistic visual feedback on the architectural options taken. • Thermal analysis: At this stage, thermal analysis is done in order to estimate the performance of HVAC systems and/or the comfort in the built environment. This should give a quick feedback on the architectural options taken (orientation, glazed surface, etc.). • Documentation/annotations: During the design, users should be able to access, in line, to relevant documentation and standards about the building components. This can be done by supporting hypertext links between building components and related URLs. Furthermore, users can attach annotations to a given component or the overall project so they can leave a message or explain a choice to other users (that are not in the same work session). Construction projects can very easily become complex. Therefore, performance optimization procedures are of paramount importance. Two optimization procedures are particularly efficient in the AEC sector: scene graph culling (when the walkthrough takes place in the first floor, there is no point in loading the geometry of the other floors) and Levels of Details (LOD) (each of building components, that can be very complex if represented with all these details, have several representations that will be displayed depending on the LOD required based on the distant of the component from the camera). These methods, combined with more generic optimization methods (such as visibility culling and backface culling) should allow complete scalability of the system regardless of the complexity of the construction project.

THE INTEGRATED CONSTRUCTION ENVIRONMENT In order for VR applications to be successfully implemented in a complex industry such as construction, they must be part of an ICE (Fig. 3). In such an environment, construction applications packages are integrated through a central intelligent core whereby project information is controlled, maintained, and manipulated. The user interface for this environment should have the ability to convey project information in a humanly acceptable level, i.e. elements, spaces, resources, etc. At this end, VR can play a major role in the development of a human computer interface for the ICE. Whether immersive or non-immersive techniques are used, users can visualize design and construction information in 3D, photo-realistic, and interactive images. The latter facility allows users to interact with external applications at real-time, thus, allowing VR systems not only to be used as presentation tools but as a universal interface for all construction applications. Construction professionals can view, alter, test, etc. any function or part of the proposed design and at any stage of the project life cycle through the virtual space.

A solution to seamless technology integration in the AEC industry? 255

Figure 3.

Conceptual presentation of the ICE.

In the short term, VR (non-immersive) can be used, as a modeling tool, to complement current design tools such as CAD systems. This implies that VR can be considered as an application package within the ICE, which aims at providing flexible, realistic, and interactive presentations. Once VR models are generated in virtual space, users can navigate through the product, at its current stage of development, and interact with any design elements or spaces to access further information or run external simulation programs. Users’ movements and queries are monitored and controlled by the intelligent central core of the ICE. VR, as a universal interface, can be enhanced by video conferencing. Communica-tions between different members of the design team or between design team, builder and owners can be significantly improved by using a combination of VR and video conferencing techniques. If VR models are generated automatically from the traditional design tools at the local design office, such models can be transmitted to the client’s remote site. Clients can navigate through the product and/or request alterations to the design or part of the design by simply pointing or moving the concerned elements. Alternative solutions can then be suggested by the designers and represented to the clients for final approval. The same scenario can be applied to improve communications between various members of the design team. A prototype of such product has already been developed by the Automation and Integration in Construction (AIC) research group at the TIME Research Institute,

256 R.R.A. Issa

University of Salford. At its current stage of development, the prototype “SPACE” (Simultaneous Prototyping for An integrated Construction Environment) integrates six construction applications with the central data models. The applications are: design, specifications, estimating, construction planning, site layout planning, and VR (Alshawi & Budeiri, 1993). In the long term, VR (fully immersive) will offer the average user the potential to enhance the final presentation by combining 3D images, headmounted displays, sounds, and self-movements. The ability to support the illusion of the individual’s movement through the virtual space will make the implementation of VR much more acceptable to humans. Users will be able to feel/see their movements in space, thus, improving the performance and well-being of the ultimate human user. Users’ movements and requests, in virtual space, will be monitored and controlled by an intelligent and integrated knowledge-based system and other external construction applications where all communications with external applications’ are carried out in virtual space in either a textual or graphical format. The flexibility offered by virtual environments to visualize and interact with the virtual world, provided that these technologies are available at a reasonable cost, will enable designers, clients, and contractors to use VR to rapidly construct and test their prototypes before constructing the actual project. But this only happens if the strengths of the technology are emphasized and the hype is significantly played down. VR should be treated not as a technology in its own right, but in terms of a suite of technologies, which when carefully implemented, are capable of matching the capabilities of humans to the requirements of the application or task he or she is required to work with. The potential of VR can only be realized if it is integrated with construction applications packages. An ICE should be developed where all construction applications are integrated through a central intelligent core. VR can play a major role in the development of a human computer interface for such an environment. Whether immersive or non-immersive techniques are used, users can visualize design and construction information in 3D, photo-realistic, and interactive images. Moreover, VR displays and interactive devices should only be selected on the basis of (a) human factors issues, i.e. what is expected of the performance and well-being of the ultimate human user, and (b) customer requirements.

ROBOTICS INTEGRATION IN THE CONSTRUCTION WORKFORCE THROUGH VR Mobility on the job site Robots in construction are part of a system made up, as shown in Figure 4, of four basic, interacting components: operator, computer, robot, and the construction environment. The design of new robots to supplement the construction workforce

A solution to seamless technology integration in the AEC industry? 257

Figure 4.

Robotic system model.

can only be achieved with the help of VR. VR can be of valuable assistance in both geometric aspects, such as link dimensions, work envelope and dexterity, as well as in control aspects, such as, visualizing sensor data and virtual navigation controllers. By combining and integrating reflex control and virtual environments, great progress can be made toward completely autonomous robots. Reflex control allows us to establish a direct link between information and action, thus bypassing the high resource overhead requirements associated with the decision making stage. This inclusion of decision in information is only possible in well-identified environments (Burdea & Coiffet, 1994). Virtual environments and fixtures Applying VR to unstructured environments involves two categories of virtual objects. The first category would involve the modeling of the minimal information known about the unstructured work environment. The result of the process will be the replacement of an unknown characteristic of the real environment by a known virtual environment. This principle could be extended to most characteristics of the unstructured construction environment. The second category of virtual objects is “virtual” fixtures or guides, which help during the task execution. Lines, curves surfaces, or volumes of known geometry, along which the robot is restricted to move (Coiffet, 1993). Figure 5 shows a conceptual representation of a system involving virtual environments and fixtures. Teleoperation Another approach to dealing with the unstructured construction environment is by keeping an operator involved in the control loop. Using VR to integrate the

258 R.R.A. Issa

Figure 5.

Virtual environments and fixtures concept.

operating environment with the operator’s environment, facilitates maneuvering the robot on the unstructured environment of the construction job site (Rosenberg, 1992). One of the difficulties involved in dealing with teleoperation is that the operator is remote from the robot and the feedback data may be time delayed or insufficiently detailed for correct control decisions. Stereo viewing Human vision is the most powerful sensorial channel and has extremely large processing bandwidth. Our depth perception is associated with stereopsis, in which both eyes register an image and the brain uses the horizontal shift in image registered by the two eyes to measure depth (Julesz, 1971). Depth perception is what allows us to maneuver in our environments, because it gives us our ability to see scenes in 3D. Integrating robots in the construction workforce and in the work environment will involve designing a display and vision system that can adequately provide this type of stereo vision and allow for its integration and interpretation in terms of maneuverability. Dexterity in manual functions Once the autonomous robot has reached its designated work area on the construction job site, the focus shifts from mobility to dexterity in performing construction tasks. Dexterity training for robots can be achieved by fitting a human construction worker with an instrumented glove and then asking that worker to perform tasks identical to those expected from a robot in the future. The system, as shown in Figure 6, uses a pair of electrodes placed on the forearm, which are connected to a neural network computer. Once the neural network has been trained, the robotic arm can be used to perform the function it has been trained on with relative dexterity.

A solution to seamless technology integration in the AEC industry? 259

Figure 6.

Enhancing dexterity in manual functions using a neural network.

CONCLUSION Research into CIC has just begun to draw the attention of both industry and academic institutions. Non-immersive applications using VR as a modeling tool to verify the integrity and constructibility of designs have been gaining in popularity. At the same time collaborative design-build efforts are making the use of VR as a universal interface among construction team members ever more popular. With the ever-increasing computational power available to users, it is expected that VR technologies and peripherals will develop rapidly and their application will have the potential to change dramatically the way of doing business in construction industry. What is not certain is what path the advances will take and what kind of impact these advances will bring to the industry. Using the VTT analogy presented in Figure 1, we can reasonably conclude that the ground under the water is still not clear, but is getting clearer.

260 R.R.A. Issa

REFERENCES Alshawi, M. & Budeiri, M. 1993. Graphical simulation of construction sequence by integrating CAD and planning packages. The International Journal of Construction Information Technology 1(2): 35–46. Brandon, P. & Betts, M. 1997. Creating a framework for IT in construction. The Armathwaite Initiative, the Formation of a Global Construction IT Network, Construct IT Centre of Excellence. Burdea, G. & Coiffet, P. 1994. Virtual reality technology. New York: John Wiley & Sons. Clough, R.H. 1986. Construction contracting, 5th edition. New York: John Wiley & Sons. Coiffet, P. 1993. Robot Habilas and Robot Sapiens. Paris: Editions Hermes. Froese, T. 1994. Information standards in the AEC industry. Canadian Civil Engineer 11(6). Issa, R.R.A. (ed.) 1999. State of the art report: virtual reality in construction. International Council for Building Research Studies Documentation (unpublished). Julesz, B. 1971. Foundations of cyclopean perception. Chicago: University of Chicago Press. Rosenberg, L. 1992. The use of virtual fixtures as perceptual overlays to enhance operator performance in remote environments. Technical Report, Center for Design Research, Stanford University, September. Teicholz, P. & Fischer, M. 1994. Strategy for computer integrated construction technology. Journal of Construction and Management Engineering 120(1): 117–131. ASCE. VTT, the Technical Research Center of Finland 1999. http://www.vtt.fi/cic/ratas/ islands.html

CONSTRUCTION MANAGEMENT PULL FOR nD CAD Peter Barrett University of Salford, Salford, UK

Abstract 4D CAD work at present could be typified as “techno-construction-centric”. This paper endeavors to provide a wider construction management perspective that will open up high value alternative areas for consideration. Construction is a dynamic, fragmented and combative industry. There is just not a stable platform for the adoption of sophisticated tools. In addition it is usual to speak of managing for time, cost and quality which is really quite misleading. The following performance dimensions are suggested: location (planning), function, aesthetics, cost, time, health and safety and environmental performance. This implies a broader, longer-term perspective beyond immediate project needs. Given the tacit–tacit emphasis of the industry, the mismatch with the explicit–explicit character of 4D CAD systems is stark. Instead of accuracy and detail, coarse robustness and connectedness are needed in systems that cover the important hard and soft dimensions. The implication is that 4D CAD systems need to shift emphasis towards the tacit–explicit mode by accommodating the above wide range of hard and soft, long- and short-term performance dimensions (nD CAD). In parallel with this a push towards supporting explicit–tacit knowledge conversion is needed with an emphasis on richer communications. The developments suggested will create a closer fit between the characteristics of the systems and the reality experienced by those in the industry. As such it will simply make more sense for such systems to be taken up through industry pull. Keywords: construction management, knowledge transfer, nD CAD, tacit knowledge, industry pull

INTRODUCTION: CURRENT 4D CAD FOCUS To date 4D CAD appears to be focused on integrating the technical design information respectively within the design and construction phases (e.g. Aalami & 261

262 P. Barrett

Fischer, 1998). This has great potential to unlock the synergies between the knowledge and experience of the designers and that of the constructors. These islands of know-how are typically isolated by education, tradition, orientation, contracts and processes. The thrust of the work, however, seems to be limited to the technological issues, with a heavy emphasis on the construction phase. In short, 4D CAD work at present could be typified as “techno-construction-centric”. This paper endeavors to provide a wider construction management perspective that will open up high value alternative areas for consideration. In this way, it is hoped, the full benefits of the emerging technology can be developed. Unless otherwise stated a UK perspective is being taken. Implicit in this paper is the view that very seldom is the technology itself the area where fundamental problems in practice arise, there is usually some better or worse technical solution. Limiting underlying assumptions, however, the innovation processes necessary for take-up and the management of people are much more difficult to handle. The usual “best practice” solution is to advise that people should behave more rationally (Barrett & Stanley, 1999), but this belies the reality of human nature. In practice what has to happen is that the nature of construction players is accepted as a given (at least in the short to medium term) and that initiatives, such as 4D CAD are developed taking this into account. Direct experience of using shared CAD systems to integrate project planning on a large trial project in Norway (studied in Barrett & Stave, 1993) reinforces this view. One manager closely involved stated: “The most important basis of success in the introduction and development of new technology is not the technology itself, but the people who are to use it.” It was how the technology supported changing perceptions, relationships, information flows and working systems that mattered, not so much the particular form of the technology chosen. Thus the remainder of this paper will seek to elaborate on some of the wider factors and opportunities that any 4D CAD system should ideally attempt to address and support.

THE NATURE OF CONSTRUCTION Construction is a dynamic, fragmented and combative industry, certainly in the UK, and it would seem worldwide (e.g. Latham, 1994; DETR, 1998). The ability to absorb technologies is hindered by the industry having its own unique “recipe” of assumptions, knowledge-bases, technologies and practices (Huff, 1982; Spender, 1989). This “recipe” considerably erodes the ease with which technologies can be transferred into an industry by creating “incompatibility barriers.” These barriers generally can only be surmounted by the technologies being carefully interpreted and transformed to blend comfortably with, and enhance, the recipient industry’s “recipe”. It has to be said that the construction industry is highly reactive and action orientated. This can suit coping with short-term emergencies, but is problematic when more reflective initiatives are needed.

Construction management pull for nD CAD 263

Figure 1.

Impervious construction!

From the evidence of the industry’s reaction to, say, quality, health and safety and environmental imperatives (Barrett & Sexton, 1999), there is a tendency to do the minimum, as late as possible. Drawing from Leavitt et al. (1973: 306–310), Figure 1 provides a continuum, from companies having impervious boundaries with their business environments to companies with open boundaries. Very often construction companies are at the impervious extreme. Sometimes they are “selectively impervious”, taking on some proposals where they seem to fit (or are unavoidable!), but grafting them on so that over time the company develops an array of incompatible systems that do not deliver synergies. In fact “initiative fatigue” is more likely. A few firms will organically adapt to their environment reactively, but hardly any will actively manage their business environment (“actionadaptation”) for symbiotic benefit. The economic turbulence of the industry is one undoubted cause of this inability to deal with major change in a positive way. Work by Sarshar et al. (1999) has articulated the problem in a way that will be familiar to IT specialists. Taking the Capability Maturity Model for the software industry, developed at Carnegie Mellon University for the US Department of Defense (SEI, 1994), she has worked with colleagues and industry to develop a version for construction companies. Apart from finding that the supply chain aspect is under-represented in the original model, they have found that the systems of even very good companies in construction appear to only be at Stage 1 or maybe 2 (Fig. 2), i.e. they are “chaotic” or moving towards “repeatable”. In fact quite a lot of time on this industry-collaborative project was spent debating whether to create a Stage 0! Thus, the “organizational readiness” (Hersey et al., 1996) for taking up 4D CAD is likely to be rather low. There is just not a stable platform for the adoption of sophisticated tools, particularly if they are prescriptive in the way work is done.

IMPLICATIONS OF TURBULENCE AND IMMATURITY “The construction industry is very old, but not very mature” (Barrett & Sexton, 1999). It does not, and arguably should not, have a high level of standardization or

264 P. Barrett

Figure 2.

CMM/SPICE maturity levels.

a large body of explicit knowledge. However, there is a massive accumulation of fractured, formal, façade systems dating from as far back as the Middle Ages. Construction needs to be recognized as a “new” industry in which an emphasis on innovation, customization and the use of tacit knowledge is celebrated and supported (Hansen et al., 1999). Cairncross (1998) has graphically illustrated by extrapolation “how the communications revolution will change our lives” in general and the OECD (1999) has highlighted the effect electronic commerce will have on the time dimension in particular. It is reasonable to assume that rapid change will occur within construction in the coming decade. More powerful, cheaper computing power will be available to more computer-literate workers. This must be used to bind back together the industry by supporting strong informal horizontal linkages as well as formal vertical integration. Galbraith’s (1977) model given in Figure 3 further illustrates this suggested emphasis. His model is proposed as a complete set of alternatives to absorb “exceptions” in a company, i.e. gaps between work demands and worker capabilities. Traditionally the construction industry has relied on formal (contract) mechanisms (1, 2 and 3) together with self-containment (5), witness the prevalent division between designers and constructors, and, of course alternative 4—“slack resources”. This last is a euphemism for “sub-optimal performance”. However, new technology will undoubtedly support an increased capacity to process information. This may be aimed at creating vertical information systems (6), but as a complementary approach, or indeed, an alternative, supporting the creation of

Construction management pull for nD CAD 265 1. Rules and programmes 2. Hierarchical referral 3. Goal setting

4. Creation of slack resources

5. Creation of selfcontained tasks

Reduce the need for information processing

Increase the capacity to process information

_

+

+

+

Growing commitment

+

Mini experiment

_

Explicit understanding

_

_

Tacit action

Figure 4.

7. Creation of lateral relations

An information processing view of the firm.

Performance

Figure 3.

6. Investment in vertical information systems

/ n tio ion p t o ta Ad ap ad

Time

Incremental change in construction.

stronger “lateral relations” (7) has much to commend it. This approach concerns horizontal communications. The importance of this aspect will be further drawn out in this paper. These changes will need a “fusion” of technological and organizational innovations (OECD, 1998). Work on the implementation of change in construction has highlighted the need for an incremental approach that emphasizes the adaptation of the technology to the company for success (Barrett & Stanley, 1999). Figure 4 shows the process revealed when attempts were made by the author to implement consensus improvements to the briefing process. The collaborating companies started from a position of doing what they always had because it had worked so far, that is their actions were based on a tacit knowledge base. At this stage there were only minor irritations such as “driving forces” and quite significant “restraining forces” (Lewin, 1947).

266 P. Barrett

First of all, as we tried to implement the agreed changes together, nothing happened! The firms decided their clients “didn’t want to be guinea pigs” and carried on as before. Even though the researchers were disappointed we had to try to understand this resistance. It became clear that it was based, not so much on antagonism towards the proposals, which they had all been involved in developing, but rather it was grounded in a highly rational (from their perspective) aversion to risk. Each firm was not acting in a vacuum, they had many relationships with other parts of the industry. If they changed in isolation it could cause real problems of disjuncture in these relationships. Given the current nature of the industry, any resulting problems would be blamed on our partners. In these conditions a company would have to be either foolish or brave and highly motivated to move first. Initially our companies were not sufficiently motivated, but they did see more clearly what was going wrong and why. This led to the explicit understanding stage shown on the model, i.e. no change in their actions, but a significant change in the companies’ appreciation of the impact of their actions, even as they rushed from incident to incident. As a consequence of this heightened awareness the companies tended to gain motivation to change “now that we have seen it go wrong again!” This increased motivation led to the design of low risk “mini-experiments” with consequently reduced restraining forces. The forms were beginning to move. As these experiments delivered some benefits the companies commitment to the implementation of the ideas grew. The perception of the risks dwindled and the motivation to carry on grew. Further experiments were built in and the ideas progressively adapted and adopted by the companies. This rather extended description serves to highlight the rocky road to implementation and so the need for a sustained incremental approach and a tolerance (indeed expectation) of a mixture of success and failure on the way. This rather uneven progress is reflected very well in the juggling analogy (Gelb & Buzan, 1994), where it is stressed that to make progress at all mistakes must be accommodated. In litigious construction this means trying things out behind the scenes or on a limited basis. Trying to pick up three balls in front of an audience and just juggle without any practice is likely to have comical results. However, not many construction practitioners want reputations as clowns! So take-up of new systems must allow adoption in stages through non-threatening, low risk, incremental access. There is very strong resonance between the above view of managing change within construction companies and the findings of MIT’s recent study of the processes of architectural design. This is summed up in the following view: “No process worth replicating is replicable. Put in less jarring terms: A worthwhile process must be reinvented rather than mechanically reproduced.” (Horgen et al., 1999: 269). The point is that there is no short cut in the adoption of new technologies and integrating them with a company’s processes. Each firm has to engage in its own tailored learning process. If technology is at the center of this it is still only a part of a much more complex whole. And beyond the firm itself and all of its dimensions there is the industry context to take into account. A firm may choose to move, but on its own it cannot move very far.

Construction management pull for nD CAD 267

Level 5: Innovation network

Level of innovation through the supply chain

Level 4: Innovation chain Level 3: Knowledge collaboration Level 2: Knowledge exchange

Level 1: Information transfer

Figure 5.

Innovation and the supply chain.

The briefing study quoted above also stressed the high-leverage potential for improvements that create and maintain a shared vision amongst all involved in a construction project. This has been reinforced and extended in a study of construction innovation which stressed that to make significant gains a strategic approach, utilizing a carefully selected portfolio of company-to-company relationships, is needed (Barrett & Sexton, 1998). Various possible levels of interaction are set out in Figure 5. In these collaborative endeavors, soft factors such as trust and longerterm plans were found to be central, reflecting Doz’s (1996) formulation of a developing cycle in strategic alliances against the three criteria of efficiency, equity and flexibility. Doz makes the telling point that “the impact of initial conditions quickly fades away” in successful alliances. Again this argues against top-down imposition of the ready-made, complete, “right” solution. For 4D CAD this seems to argue for flexible shell systems that support a good deal of integrating features to reflect the reality that diverse players will be using them. Trying to make everyone play the game by a single set of rules is not likely to work. Providing an environment that a company can incrementally take-up could. Providing better means of flexible communication could. For an example of this latter aspect, during case studies of the operation of supply networks for hybrid concrete systems (Barrett, 1998), it became very apparent that the formal system of controls could not cope with the volume and rapid changes of information. However, the communications technologies supporting the “informal system of controls” (Tavistock, 1966) that took over, such as radios, facsimile machines and mobile telephones, were very widely used with no resistance, no hesitation. They met the needs of the workers and fitted with the culture of the industry (if 4D CAD systems can do the same then there will be no problem in achieving take-up!).

268 P. Barrett

The above observation links well with the analysis by Hansen et al. (1999) of knowledge management in management consultancies. They make the distinction between two principal strategies, namely “codification” or “personalization”. The aim with the codification approach is to “Provide high quality, reliable, and fast information-systems implementation by reusing codified knowledge”. Personalization aims to “Provide creative, analytically rigorous advice on high-level strategic problems by channeling individual expertise”. (p. 109). Codification assumes a volume of work with a lot of reuse of knowledge by large teams with a lot of juniors, using people-to-document systems, highly supported by IT. Personalization assumes high margin work carried out by well-qualified small teams using personto-person knowledge management, supported by moderate IT systems. … companies that use knowledge effectively pursue one strategy predominantly and use the second strategy to support the first. We think of this as an 80-20 split … Executives who try to excel at both strategies risk failing at both. (p. 112). So, they are arguing you have to make a broad choice, but which alternative relates best to construction? To choose a predominant strategy (codification or personalization respectively) depends on whether: the service is standardized or customized; the organization/sector mature or innovative; the knowledge used to solve problems explicit or tacit knowledge. Much of construction is customized and we have already seen that company systems are of low maturity and the knowledge used is predominantly tacit. This all points towards a strategy that emphasizes teams, person-to-person knowledge management and only moderate IT support with an emphasis on communications. It is not possible to generalize, but it seems that construction faces the uncertainty and, doubtless as a consequence, has many of the characteristics of a new, dynamic, thrusting industry. These characteristics should be celebrated and supported. The work of Hansen et al. connects with more substantial work specifically focused on construction professionals. For example, Coxe et al. (1987) set out a parallel choice between “practice-centered businesses” and “business-centered practices”. The great majority of design firms are primarily practices and this again fits with the customization strategy. This emphasis on communications rather than codification underpins a speculative view based on work with the UK concrete industry. This is drawn from case study work specific to concrete, but also from an innovation study investigating supply chains in which the prominent role of materials and components suppliers became very evident. Figure 6 sets out the idea of creating a web-based environment to link the various players. In particular, design support is made accessible to designers and tendering and ordering made easy for contractors and suppliers. Within the shared environment a set of broad generic systems are provided with associated interactive cost, etc. models. The starting position is shown in the diagram, but over time it is anticipated that consortia and market mechanisms

Construction management pull for nD CAD 269

5 Supply chain models 5 Models of elemental costs

Standardised design details

5 generic Hcc structural systems

Clients / designers

5 Models of programmes

Contractors / suppliers

Suppliers cost / delivery details

X 5 standard building layouts

Figure 6.

A speculative, neutral web-based environment for concrete.

Table 1.

Employment and number of construction enterprises in the EU.

Size of firm (employees)

Number of firms

Percentage of firms

Number of employees

Percentage of employees

Percentage of total turnover

0–9 10–19 20–99 100–199 200–499 500⫹ All firms

1,700,797 76,618 48,695 3,543 1,585 585 1,831,822

92.80 4.20 2.70 0.20 0.10 0.03 100.00

3,512,969 1,025,263 1,820,354 492,320 483,257 761,345 8,095,509

43.3 12.7 22.5 6.1 6.0 9.4 100.0

36.1 12.4 24.7 7.2 7.5 12.1 100.0

would evolve the content of the shared space as experience and opportunities became evident. The above examples generally relate to studies of large organizations with relatively robust in-house management capabilities. So, it is worth mentioning in passing that a series of benchmarking studies by ConstructIT (1998) of large construction companies has found that there is great variability in the beneficial use of IT systems even amongst these leading companies. However, the message that non-prescriptive, flexible communications tools are needed is heavily reinforced when the size structure of the industry is considered. Construction is populated mainly by very small firms. Although drawing from slightly old data, in the EU for example, about 97% of firms, doing 49% of the work are less than 20 employees strong (Atkins, 1994; see Table 1 for more detail). This means that sophisticated bespoke technologies are unlikely to be adopted by enough players to create sufficient momentum for an industry-wide, top-down change. This has led construction researchers in the UK towards using standard software platforms where possible, or at least making the interface appear familiar. For example, a database tool used to encourage cross-organizational learning in

270 P. Barrett

construction (COLA) is available in Microsoft Access, as “smaller organizations are more likely to have access to the … products and, more importantly, access to the skills required to adapt and manipulate the database for their own purposes” (Boam, 1999).

DIMENSIONS OF SYSTEMS It is usual to speak in project management circles of managing for time, cost and quality. This is really quite a misleading formulation. For example, time and cost can easily be seen as quality dimensions themselves, and, what else is to be included under the quality heading anyway? It seems more fruitful to think in terms of generic performance criteria, or at least to use quality as an overall descriptor for a comprehensive set of such criteria. The former “quality-free” approach is advocated convincingly by Sjoholt (1989), however, the latter approach was taken in a project involving the then newly independent states of Estonia and Lithuania, together with Danish and UK partners (CONQUEST, 1995). In this challenging cross-cultural project we strived for an objective assessment of quality in construction. The following list of eight performance dimensions was created: location (planning), function (fitness for user’s purpose), aesthetics, cost, time, technical performance, health and safety and environmental performance. These criteria were found to be “owned” variously by different selections of stakeholders, ranging from the “paying client” to “society at large”, with any control achieved by a range of mechanisms, ranging from socialization through custom, to legislation. The resulting mapping achieved is summarized in Table 2. Solid squares are the primary mechanisms and unfilled squares the secondary mechanisms. This table was developed for a particular comparative analysis, however, the blank sheet stimulus provided by the new states studied helped the team take an objective view of construction in general. The notion of a range of performance dimensions controlled (managed) using a range of mechanisms in a variety of combinations seems robust. This incidentally fits with work on the interactions of markets, hierarchies and networks, which Bradach & Eccles (1991) typify as underpinned by price, authority and trust respectively. They claim that these mechanisms are: “overlapping, embedded, intertwined, juxtaposed and nested … typically control mechanisms are grafted on to and leveraged off existing social structures”. This seems a realistic view from our studies of construction. The implication is again that top-down “designed” solutions are unlikely to be capable of reflecting the complexity of reality. Taking this wider range of, say eight, criteria should help to overcome “overmeasurement” of the easily measurable (Etzioni, 1964). It can be seen to lead naturally to the idea of a broader perspective beyond the immediate project needs. Construction can in fact be seen to be merely a change agent for the built environment, which itself supports society’s needs (see Fig. 7). That is, although construction

1. Market 2. Custom 3. Legislation 4. Institution 5. Experts 6. Clients

1. Market 2. Custom 3. Legislation 4. Institution 5. Experts 6. Clients

1. Market 2. Custom 3. Legislation 4. Institution 5. Experts 6. Clients

1. Market 2. Custom 3. Legislation 4. Institution 5. Experts 6. Clients

Controlled by (Who)

䊏 ⵧ

䊏 ⵧ

䊏

ⵧ ⵧ

ⵧ

䊏

ⵧ 䊏

䊏

ⵧ

ⵧ

ⵧ

ⵧ

ⵧ ⵧ ⵧ 䊏

ⵧ

䊏

ⵧ

Estonia

ⵧ

Denmark

䊏

Lithuania

ⵧ

UK

ⵧ

Estonia

ⵧ

ⵧ

䊏

ⵧ ⵧ

䊏

䊏

Denmark

ⵧ ⵧ

ⵧ ⵧ ⵧ ⵧ

ⵧ ⵧ ⵧ 䊏

ⵧ

䊏

䊏

ⵧ ⵧ ⵧ ⵧ ⵧ 䊏

䊏 ⵧ

ⵧ ⵧ 䊏

䊏 ⵧ

ⵧ

ⵧ

䊏 ⵧ

Lithuania

ⵧ

UK

䊏

䊏

ⵧ

䊏 ⵧ

ⵧ 䊏

Estonia ⵧ

ⵧ

䊏

ⵧ ⵧ

䊏

䊏

Denmark ⵧ ⵧ ⵧ ⵧ ⵧ 䊏

䊏 ⵧ

ⵧ

ⵧ

䊏 ⵧ

ⵧ ⵧ

䊏 ⵧ

Lithuania ⵧ ⵧ

䊏

ⵧ 䊏 ⵧ

ⵧ

ⵧ 䊏 ⵧ

䊏

ⵧ

ⵧ 䊏

ⵧ

Lithuania

Denmark 䊏 ⵧ

䊏

ⵧ

ⵧ

䊏 ⵧ

ⵧ ⵧ 䊏 ⵧ

ⵧ 䊏

䊏

䊏

ⵧ 䊏 ⵧ

ⵧ

Construction stage

ⵧ

䊏

Design stage

䊏

ⵧ

ⵧ

䊏

ⵧ

䊏

䊏

ⵧ

Estonia

䊏

ⵧ

ⵧ ⵧ ⵧ ⵧ ⵧ 䊏

ⵧ ⵧ ⵧ ⵧ ⵧ 䊏

䊏

Post construction stage

UK ⵧ 䊏

Estonia ⵧ 䊏

UK

Brief stage

Denmark ⵧ ⵧ ⵧ ⵧ ⵧ 䊏

ⵧ ⵧ ⵧ ⵧ 䊏

䊏 ⵧ

ⵧ ⵧ

ⵧ 䊏

ⵧ ⵧ

Lithuania ⵧ 䊏

ⵧ

䊏 ⵧ

ⵧ

䊏 ⵧ

ⵧ ⵧ 䊏 ⵧ

ⵧ

UK ⵧ 䊏

䊏

ⵧ ⵧ

䊏

䊏

ⵧ

Estonia 䊏

䊏

䊏

ⵧ 䊏

ⵧ ⵧ ⵧ ⵧ ⵧ 䊏

ⵧ ⵧ ⵧ 䊏

䊏 ⵧ

ⵧ ⵧ

ⵧ 䊏

ⵧ

Denmark

6 Fitness for

䊏 ⵧ ⵧ ⵧ ⵧ ⵧ

䊏 ⵧ

ⵧ

ⵧ ⵧ ⵧ ⵧ 䊏 ⵧ

䊏 ⵧ ⵧ ⵧ ⵧ ⵧ

Lithuania

5 Technical

UK ⵧ 䊏

䊏 ⵧ

䊏

7 Environment

ⵧ

ⵧ

ⵧ ⵧ

䊏

ⵧ

䊏

ⵧ

Estonia

4 Aesthetics

ⵧ 䊏

ⵧ

䊏

ⵧ

䊏 ⵧ

ⵧ

ⵧ ⵧ

䊏

Denmark

3 Time

ⵧ

䊏 ⵧ

䊏 ⵧ ⵧ ⵧ

䊏 ⵧ ⵧ

䊏 ⵧ ⵧ ⵧ

Lithuania

2 Cost

䊏

䊏

ⵧ

䊏

ⵧ

䊏

UK

1 Location

8 Health and safety

䊏

ⵧ

ⵧ

䊏

ⵧ

䊏

ⵧ

Estonia

Quality criteria (What) by country

ⵧ 䊏

ⵧ

䊏

ⵧ

䊏 ⵧ

ⵧ

䊏

Denmark

Performance criteria and their stakeholders.

ⵧ ⵧ ⵧ 䊏

䊏 ⵧ ⵧ ⵧ

ⵧ ⵧ 䊏

ⵧ ⵧ 䊏 ⵧ

Lithuania

Table 2.

ⵧ ⵧ

䊏

䊏

ⵧ

䊏

ⵧ

䊏

UK

272 P. Barrett

Figure 7.

Construction as a means to a means to an end.

is a very big and important industry, it is a service industry and as such is a means (to a means) to an end, and not an end in itself. The built environment is there to serve society, which in the UK’s Government’s terms equates to improving competitiveness and quality of life. Construction is the change agent through which this change is achieved. Thus, at a minimum the whole project life cycle needs to be addressed and here the “4” in 4D CAD comes into focus as time is a key linkage to many of these issues. This takes into account a long-term perspective as well as the project duration highlights issues, such as the whole life cycle of facilities, including user views, but also societal factors such as planning and environmental impacts.

IMPLICATIONS OF MULTI-DIMENSIONALITY Creating systems that can model, in detail, physical project attributes over time, perhaps even including the variation of construction project costs, is clever, but does it address the critical problem areas? Systems like these will probably support improved efficiency in design and logistics on site, however, these are not the problems highlighted in strategic reports on the industry (e.g. Latham, 1994; DETR, 1998). In these reports there is a heavy emphasis on trust and relationships, as well as some fairly mechanistic thinking. This lag is evident in construction supply chain theories to emphasise almost exclusively logistics (O’Brien, 1997), whereas, in the general field of logistics there is an interesting development towards studying the softer and longer-term aspects of supply networks, such as perceptions of requirements and performance and customer satisfaction (Harland, 1996). This appears to be following the burgeoning services literature, but particularly Grönroos’ (1984) work, which highlighted the “expectation—perception gap”.

Construction management pull for nD CAD 273

Supplier’s perception of requirements

Mismatch 3

Mismatch 1

Mismatch 4

SUPPLIER

Mismatch 2

Supplier’s perception of performance

Figure 8.

Customer’s perception of requirements

CUSTOMER

Customer’s perception of performance

Harland’s supply chain model.

Specialist subcontractor 1

R5 P5 R4 P5

R5 P5

Specialist subcontractor 2

Package Contractor R4

R5

P5

P5 R3

P5

R3 P3 R3 P5

Construction manager

Specialist subcontractor 3

R5 P5 R3 P5

R4 P5

Client

R4 P5

R5

Designer

P5 R4 P4

Figure 9.

Example analysis of supply network.

The complex “gaps” picture given in Figure 8 is, of course, a great simplification because each “customer” is someone else’s “supplier”. But, more than this, supply chains are really supply networks as shown in Figure 9. This was based on a study of a specific project. The difficulty of handling complex, “soft”, interacting data like this is a big challenge for a 4D CAD system. However, it is a necessary effort if relevant, useful systems are to be created. Interestingly, this move in emphasis is paralleled in general UK research funding which is shifting away from technical “single loop” research (Argyris & Schon, 1978), on how to do better what we already do, towards supporting organizational and sociological research on how effectiveness can be enhanced through customer-orientated, knowledge-based innovations (e.g. HMSO, 1998). The UK Government is pushing forward on open electronic communications in its regulatory role (Cabinet Office, 1999), and this could facilitate and stimulate linkages right the way through from land survey data, to design, to construction, to

Figure 10.

Example soft data analysis (of innovation in small construction firms).

274 P. Barrett

Construction management pull for nD CAD 275

use and post-occupancy evaluation data. This will then allow a flow of key items of information throughout the life cycle of the built artifact, and beyond, through feedforward from use, to design. These are sound ideas, but there is a long way to go to achieve tangible results in practice. This sort of development will call for 4D CAD systems that can represent a wider range of performance criteria, as set out in the previous section, including soft factors, such as aesthetics, but also with a strong longer-term flavor. This has significant implications for the type of data the systems need to be able to handle and leads again towards an emphasis on soft or tacit forms of data. Accommodating factors like these in strategy is evident in Eden & Ackerman’s (1998) cognitive mapping approach (see, e.g. Fig. 10). Beer’s (1985) recursive, multidimensional model of organizational forms stresses both the distinction and the linkage between the long- and the short-term and the place of technology in mediating between high and low variety situations as amplifiers or attenuators. Of course, the present is simply the consumption of the future, but for all that, data on the future is inevitably very different from that from the present or the past. It seems that, instead of accuracy and detail, coarse robustness and connectedness are needed (e.g. Argenti, 1980). This links in an interesting way with Tenner’s (1996) diagnosis that pushing systems further and further in terms of sophistication and detail will inevitably lead to problems and that the solution is “finesse”. For me, this means aiming for systems that cover the important hard and soft data dimensions and use technology and ingenuity to make the resulting system accessible, easy to use, robust, integrative, dynamic and flexible. There is a real danger that in addressing and integrating some aspects of the overall picture disintegration can be the result. This can be seen to have happened with quality, health and safety and environmental systems in the way illustrated in Figure 11. Increasing formalization has led to fragmentation. Such systems should aim to remain holistically integrated, in the top zone, whether they are informal or formal.

Ideal zone Integrated

e ut ro al

tre nd

Systems integrated by design!

Some systems More systems formalised formalised

Theoretical isolation introduced

Cu rre nt

ic Typ

Systems informal and integrated by real world focus

Divisions between systems harden

Disintegrated Informal

Figure 11.

Maintaining an integrated view.

Formal

276 P. Barrett

In terms of traditional logic the argument is for extension over connotation. Assuming a finite capacity for a system it is argued that it is better to include all of the important dimensions that need to be integrated, but as a consequence the volume of detail will have to be sacrificed. This approach will, in addition, reduce the denseness of the system to new users and so make it more accessible and transparent. Concentrating on effectiveness should also lead to tools that emphasize the communications aspect and exploit widely available, increasingly familiar technologies, such as the web. This will leave the industry to deal with the minutiae of the process, but within a informing context in which the major short- and long-term performance criteria of the full range of stakeholders is given center stage. This can be seen as a holographic approach (Morgan & Ramirez, 1983), with the major performance criteria being reflected at various levels and time-frames for different stakeholders with, in each case, an appropriate level of detail.

CONCLUSION Underlying a lot of this discussion has been the distinction between hard and soft data, which parallels to a degree the distinction between explicit and tacit information. Figure 12 (based on Nonaka & Takeuchi, 1995) uses this classification and endeavors to summarize the suggested shift in 4D CAD. If the tacit–tacit emphasis of the industry is taken as a given, then the mismatch with the explicit– explicit character of 4D CAD systems is stark. For the industry to operate to its optimum the synergies of all four modes of knowledge conversion are needed

Figure 12.

Modes of knowledge conversion.

Construction management pull for nD CAD 277

(Barrett & Sexton, 1999). The implication is that 4D CAD systems need to shift emphasis towards the tacit–explicit mode by accommodating a wider than normal range of key hard and soft, long- and short-term performance criteria (nD CAD) at a coarse, but robust level of resolution. In parallel with this a push towards supporting explicit–tacit knowledge conversion is needed with an emphasis on richer communications. The developments suggested will create a closer fit between systems’ characteristics and the reality experienced by those in the industry. As such it will simply make more sense for such systems to be taken up through industry pull, especially if an incremental, but progressive trajectory is supported. Thus, building from simple, but useful implementations, based on “loose couplings” (Baumard, 1999), evolution can take place as the general systems and processes in industry improve. More sophisticated modular applications can then be built into the robust performanceorientated framework and communications infrastructure created. Progressively, greater and more widely available computer power will doubtless be available, but initially any capability should be used to make systems as easy and flexible as possible to use, with particular attention on ease of data exchange between companies. CAD is an important tool for the journey from idea to artifact implicit in every construction project. Based on the above analysis the following keywords for developments in 4D CAD are suggested from a “construction management pull” perspective, namely: outwards-looking, multi-perspective, hard and soft data processing, coarse, robust, strategically-directed, open, informing, integrative, communications orientated, evolutionary and intuitive. nD CAD systems that succeed in the future will provide a rich, broad and integrated knowledge context to the tacit– tacit essence of the construction industry. As such they could be termed wisdombased systems!

REFERENCES Aalami, F. & Fischer, M. 1998. Joint product and process model elaboration based on construction method models. The life cycle of construction innovations. CIB Publication 226. Argenti, J. 1980. Practical corporate planning. Allen and Unwin. Argyris, C. & Schon, D. 1978. Organisational learning: a theory of action perspective. Addison-Wesley. Atkins, W.S. 1994. Strategies for the European construction sector. Luxembourg: EC. Barrett, P.S. 1998. Hybrid concrete structures for the UK market: business processes and desirable process improvements. Reading: Reinforced Concrete Council. Barrett, P.S. & Sexton, M. 1998. Integrating to innovate. London: Construction Industry Council. Barrett, P.S. & Sexton, M. 1999. The transformation of ‘out-of-industry’ knowledge into construction industry wisdom. Report to the CRISP Motivation Theme Group, London. Barrett, P.S. & Stanley, C. 1999. Better construction briefing. Oxford: Blackwell Science.

278 P. Barrett Barrett, P.S. & Stave, O. 1993. Integrated project planning using CAD: a case study. Proceedings of the 7th CIB W-65 international symposium, University of the West Indies. Baumard, P. 1999. Tacit knowledge in organisations. London: Sage. Beer, S. 1985. Diagnosing the system for organisations. Chichester: John Wiley. Boam, J. 1999. COLA user manual. Leeds Metropolitan University, UK. For details of the COLA system see (24 September 1999), http://is.lse.ac-uk/b-hive Bradach, J.L. & Eccles, R.G. 1991. Price, authority and trust. In Thompson et al. (eds), Markets, hierarchies and networks: 277–292. London: Sage. Cabinet Office. 1999. Modernising government. London: Cm 4310. Cairncross, F. 1998. The death of distance. London: Orien Business Books. CONQUEST. 1995. Achieving quality in construction. EC PHARE-ACE Final Report, Salford. ConstructIT. 1998. Benchmarking best practice reports: supplier management and project programming and control. Construct IT, University of Salford. Coxe, W. et al. 1987. Success strategies for the design professional. New York: McGraw-Hill. DETR. 1998. Rethinking construction. The Egan Report, DETR, London, (16 July 1998). www.construction.detr.gov.uk Doz, Y.L. 1996. The evolution of cooperation in strategic alliances: initial conditions or learning processes? Strategic Management Journal 17: 55–83. Eden, C. & Ackerman, F. 1998. Making strategy. London: Sage. Etzioni, A. 1964. Modern organizations. Englewood Cliffs, NJ: Prentice-Hall. Galbraith, J.R. 1977. Organisation design. Reading, Mass: Addison-Wesley. Gelb, M.J. & Buzan, T. 1994. Lessons from the art of juggling. London: Aurum Press. Grönroos, C. 1984. Strategic management and marketing in the service sector. Bromley: Chartwell-Bratt. HMSO. 1998. Our competitive future: building the knowledge driven economy. Cm 4176, London. Hansen, M.T., Nohria, N. & Tierney, T. 1999. What’s your strategy for managing knowledge. Harvard Business Review March–April: 106–116. Harland, C.M. 1996. Supply chain management: relationships, chains and networks. British Journal of Management 7(special issue): S63–S80. Hersey, P., Blanchard, K.H. & Johnson, D.E. 1996. Managing organisational behaviour: utilizing human resources, 7th edition. New Jersey: Prentice-Hall. Horgen, T.H., Joroff, M.L. & Schon, D.A. 1999. Excellence by design: transforming workplace and work practice. New York: John Wiley and Sons Ltd. Huff, A.S. 1982. Industry influences on strategy reformation. Strategic Management Journal 3: 119–131. Latham, M. 1994. Constructing the team. The Latham Report, London: HMSO. Leavitt, Dill & Eyring. 1973. The organizational world. New York: Harcourt Brace Javanovich. Lewin, K. 1947. Frontiers in group dynamics. Human Relations I(1): 5–41. Morgan, G. & Ramirez, R. 1983. Action learning: a holographic metaphor for guiding social change. Human Relations 37(1): 1–28. Nonaka, I. & Takeuchi, H. 1995. The knowledge creating company: how Japanese companies create the dynamics of innovation. New York: Oxford University Press. OECD. 1998. 21st century technologies: promises and perils of a dynamic future. OECD. OECD. 1999. The economic and social impact of electronic commerce. OECD. O’Brien, W.J. 1997. Construction supply-chains: case study, integrated cost and performance analysis. In L. Alarcon (ed.), Lean construction: 187–222. Rotterdam: A.A. Balkema.

Construction management pull for nD CAD 279 SEI. 1994. The capability maturity model. SEI, Addison Wesley Longman Inc. Sarshar, M., Haigh, R., Finnemore, M., Barrett, P. & Aouad, G. 1999. Standardised process improvement for construction enterprises: an EU update. International conference on technology watch and innovation in the construction industry, Brussels, April 2000 (submitted). Sjoholt, O. 1989. From quality assurance to improvement management. Oslo: NBI. Spender, J.C. 1989. Industry recipes: the nature and sources of management judgement. Oxford: Basil Blackwell. Tavistock. 1966. Interdependence and uncertainty: a study of the building industry. London: Tavistock Institute. Tenner, E. 1996. Why things bite back: new technology and the revenge effect. London: Fourth Estate.

CLOSURE R.R.A. Issa, I. Flood, W.J. O’Brien M.E. Rinker, Sr. School of Building Construction, University of Florida, Gainesville, FL, USA

CURRENT STATE OF 4D CAD 3D/4D CAD tools are in their second decade of use in the design and construction industry. From scattered early efforts in the late 1980s, the 1990s saw a proliferation of such tools in practice. Certain industries such as industrial construction now routinely design projects in 3D. In other areas, such as general building construction, there are owners, architects, contractors, subcontractors and vendors that have taken the lead in applying 3D/4D tools across their operations. Owners require 3D for design communication and decision-making and use the models for facilities management. Architects use 3D models to communicate with clients, render dramatic sculptural forms, and share these models with contractors and subcontractors to enable construction of these sculptural forms. Contractors use 3D models for coordination of building systems and materials procurement. Contractors use 4D models to ensure schedules are buildable and for trade coordination. Subcontractors use 3D and 4D tools for much the same purposes as contractors, but often use more detailed models to plan field production. Many vendors have developed 3D models of their offerings to allow architects and engineers to place them directly into their designs. Vendors also use 3D models to direct their internal production processes. The 1990s also saw increased sophistication in software packages. At the high end, tools have grown more powerful in storing intelligence about building objects and their relationships (e.g. an object is a beam connected to a column; the beam has attributes of strength, material properties, a manufacturer and tracking number, etc.). An infrastructure of tools has been developed to support implementation of 3D/4D models. These tools include: libraries, global positioning systems for surveying linked to the 3D model objects, portable display devices, and data exchange standards. Collectively, these developments make possible the integrated application 281

282 R.R.A. Issa et al.

of 4D analysis across all stages of the project lifecycle (although an integrated offthe-shelf package does not yet exist, requiring investment on the part of project team members). Advanced software applications promising more integrated functionality are in the commercial product development pipeline. Mid- and lower-level software applications have also seen increasing sophistication in their ability to manipulate 3D models. Mainstream commercial CAD packages now allow users to rapidly develop 3D models at different levels of detail. It is also possible to perform 4D analysis by linking the 3D models to scheduling packages and/or by manipulating the 3D objects using layer functions. 3D technology has also enhanced lower-end CAD systems, perhaps best evinced by the sub-US$100 home design CAD packages sold in retail stores. These systems allow homeowners to rapidly build and render 3D models of their homes, producing drawings usable by contractors. Cost and capability are no longer considered a barrier for any firm that wishes to employ 3D/4D tools. Despite the power and availability of 3D/4D tools, their use is still not common in most areas of design and construction. Apart from modeling of piping and related systems in large-scale industrial construction, the most common use of 3D models is in marketing and conceptual design. Clients are sold on the building concept in 3D “walk-thrus.” Sometimes these 3D concept models have a 4D element to portray the impact of construction on existing sites or to portray stages of project development. These models have little detail, and are seldom further developed through detailed design or for construction planning. This unfortunate circumstance is not solely the fault of uncreative practitioners. Further development in at least four technical and business areas is needed to fully realize the potential of 4D CAD on practice.

FUTURE STEPS—FOUR AREAS OF DEVELOPMENT TO TAKE 4D TOOLS TO THE NEXT LEVEL 1. The visual interface: There are two related problems with the interface of current tools. First, the more sophisticated tools (and many of the lower-end tools) are difficult to use. The models cannot easily be manipulated by anyone other than experts, and simpler representations (e.g. static models including printouts) lose much of the power of full models. Difficult-to-use software also limits the ability of users to add information to the model, consequently restricting the model’s usefulness. Second, it is difficult to customize visualization of the 3D/4D models and related data. Every user will have a different set of needs for information and different preferences for visualization of that information. Even the most sophisticated of existing applications have limited abilities to display information in various forms. Hence, even if an integrated 4D model is available, the model may not support business decisions beyond its ability to display information. 4D tools

Closure

283

are by nature information rich. Improved ease of use in accessing and contributing information and greater fluidity in customizing the presentation of that information are necessary developments if 4D models are to be fully leveraged. 2. Data exchange between applications: Improvements in the visual interface require seamless exchange of data among the software applications behind the interface. Currently, code must be written for each link between applications. This is a lengthy and difficult set-up process that most projects are unwilling to support. While writing code on a per-project basis does allow customization, the approach is not scalable. Nor is the customization necessarily fluid; as project needs and project participants change, it is unclear that the code can be easily adapted to meet those changes. What is needed is theory and methodology about the sharing of information that supports implementation on projects without detailed coding by experts. Currently, standardized data models are under development to allow software applications to share data. These data models provide a basis for sharing data such as ⬍schedule activity⬎ . It is less clear that these data models will directly support higher level reasoning, particularly with regard to user level customization of information representation. Further developments in sharing and manipulating data are required for widespread use of 4D tools. 3. Job design to leverage the tools: While there does not currently exist a recipe for collaboration to make best use of 4D tools, it is clear from the existing implementations that the technology requires new ways of working together. These range from the simple changes of design review with a 3D model to the more sophisticated questions of who contributes what to the 3D/4D model. How job design and responsibilities should change is a fundamental issue with implications for firm and project organization, legal responsibilities, and, not least, contractual incentives. Many of the projects that have used 3D/4D models collaboratively across firms have done so in the spirit of experimentation. Thus, they have not addressed the issues of standard operating procedures using the new technologies. What these procedures should be is very much a subject for future research and development. 4. Benefits and contracts for 4D: Closely related to the idea of job and organization design is the appropriation of benefits using a 3D/4D environment. While many of the firms using 3D/4D tools report benefits in a wide variety of applications that they believe easily outweighs the cost of 3D/4D model development, only limited cost-benefit analysis has been performed. It is clear that early development of the model in the project lifecycle pays dividends later in the lifecycle. However, given the fragmented nature of the construction industry, many of the early developers of design are not responsible for later stages of the project and do not accrue the benefits from savings in those later stages. In any phase of the project, it is the rare firm that wishes to absorb the cost of model development without a clear understanding of the benefits and how they will be distributed. As it is difficult to predict where benefits will be realized (e.g. on one project in design coordination, on another project in productivity improvement), it is unclear how to

284 R.R.A. Issa et al.

assemble contracts that appropriately reward those firms that bore the cost of developing a 3D/4D model. The construction industry has a poor record of consistently supporting improved planning processes, and 3D/4D tools may not be different from other approaches. How to structure contracts to support the use of 3D/4D tools is unclear. Three approaches come to mind: first, the owner can simply pay for the models and assume that they will get their fair share of benefits. Some owners are taking this approach. Second, individuals in the industry can adopt the tools and create new forms of firms that simply do everything better than traditional firms. These “category killer” firms have redefined other industries and may do so in design and construction. Third, the tools can become so powerful and cost effective that they replace existing 2D tools and methods in the various firms involved in the project process. Thus, 3D/4D tools may organically replace 2D work practices. We have seen developments in each of these approaches. However, how these benefits are supported by contractual structures remains an open question. So where do we stand on the use of 4D in construction? The 1990s saw the premature proclamation of a coming revolution in practice based on the early benefits and tools seen in the late 1980s. Now early in the first decade of the millennium, we are hesitant to make sweeping claims. But we do have the experience of the 1990s to guide us, and concrete benefits have been seen in practice. The contents of this workshop have established a roadmap to move forward, and the editors tentatively suggest that at the end of the next decade we will be writing not about novel developments in 4D tools, but about incremental improvements in a technology that is well accepted.

Subject index

4D 1–7, 10–15, 18–19, 21–22, 24–30, 43–44, 48, 55–61, 74, 77–78, 80–81, 83, 86, 94–95, 97–98, 101–103, 113–115, 116–122, 125–130, 132–135, 137–143, 150, 152, 165–171, 173–174, 176, 186–187, 195–197, 199, 201–204, 207–210, 227–235, 238–241, 261–263, 267, 272–273, 275–277, 281–284 4D CAD 2–3, 6, 11, 14, 31, 43–44, 48, 55–61, 101–103, 113, 116–119, 121–122, 125, 150, 152, 195–197, 199, 201–204, 207–210, 227–235, 238–241, 261–263, 267, 272–273, 275–277, 281–282 4D modeling 1–2, 7, 11, 13–15, 18–19, 21, 24, 30, 94, 102, 125–130, 132, 135, 139, 141–143, 165, 167–168, 171. advantages 55, 95, 100, 133, 151, 212, 250 AEC 1, 15, 165–171, 173, 176, 187, 241, 251, 254 assessment of design and process 145 benefits 1, 3–7, 30, 55–57, 61, 65, 76–77, 87, 95, 98, 100, 121, 125–126, 128, 142, 144, 151, 156, 163, 166, 208, 212, 214, 217, 222, 234, 241, 262, 266, 283–284 CAD 2–3, 5–7, 10–11, 14–15, 30, 33, 35, 37, 38, 40, 43–45, 48, 51, 55–62, 72, 80, 86, 101–103, 113–114, 116–122, 121–123, 125–127, 130, 132,

150, 152, 162, 167, 169–170, 186–187, 195–199, 201–204, 207–210, 218, 227–235, 238–241, 245–247, 249–253, 255, 259, 262–264, 267, 272–273, 275–277, 281–282, 284 case studies 1, 11, 75, 86, 98, 100, 125–126, 128, 132, 141, 267 change order 4, 12, 46, 48, 51, 195–197, 204 communications 33, 45–51, 195, 201, 244, 251, 255–256 computer 2, 5, 21, 30, 33, 37, 40–41, 44, 48, 50, 55, 58, 60–62, 65–67, 72, 77, 80, 94, 122–123, 126, 145, 147, 150, 152, 155, 166–169, 176, 184, 195, 205, 208, 210, 221, 227–228, 231, 238, 241, 264, 277 computer applications 80, 195, 244, 246 computer simulation 176, 227, 231 concurrent engineering 165, 170 construction 1–16, 19, 24, 26, 27–30, 33, 35, 37–40, 43–53, 55–58, 60–63, 65–67, 69–70, 75–78, 80–84, 86–88, 92, 94–95, 96–98, 101–105, 110, 112, 114, 118–130, 132–134, 137–159, 168–174, 176, 178, 184, 186, 195–203, 205, 207–209, 211–213, 215, 217–220, 222–223, 227–234, 236, 239–241, 243–252, 254–259, 261–274, 277, 281–284 construction company 7, 75 285

286 Subject index construction costs 101, 118 construction foreman 195 construction industry 55–56, 72, 81, 122–123, 146, 151–152, 162, 165, 195–196, 199, 202–203, 207, 208, 210, 228, 239–240, 243–244, 246–249, 259, 262–264, 277, 281, 283–284 construction management 2, 55–57, 96, 103, 198, 201–202, 227–228, 239, 241, 244, 261–262, 277 construction planning 1, 3, 7, 11–14, 24, 75–76, 123, 170, 196, 209, 228, 234, 246, 256, 282 construction site safety 211 coordination 5, 11, 14, 46, 69, 105–106, 111, 114, 118, 125, 127–129, 141–143, 172–173, 197, 202, 204, 209–210, 248 cost planning 101, 118 design 1–5, 7, 11, 13–15, 25–26, 28–30, 33, 35, 37, 40, 43–44, 48, 51, 53, 55–58, 60, 62–65, 67, 69–72, 75–77, 79, 81, 95, 102, 104, 107, 114, 116, 118–119, 121, 126–129, 132, 141–142, 145–163, 167–178, 180–184, 186–187, 196–205, 207, 209, 215, 220, 222–223, 228–230, 233–236, 238–241, 248–256, 258–259, 261–262, 264, 266, 268–273, 275, 281–284 design for safety process 211, 222 design/technology innovation 145 dynamic process simulation 145, 151–152, 154–155 FIAPP 55, 57–58, 59, 61–62, 65 field applications 195 high rise residential construction simulation 211 immersive 44, 243, 250, 254–256 industry pull 261, 277 information technology 72, 75, 98 integrated construction environment (ICE) 243 integrated product and process development 165, 171, 174, 176, 187 interdependence 165, 170 knowledge 7, 24, 30, 76–77, 80, 95, 119–121, 128, 146, 154–155, 166, 170,

178, 187, 222, 243, 251, 256, 261–262, 264–265, 267–268, 273, 276–277 lean construction 165, 278 modeling 1–3, 7, 10–11, 13–15, 18–19, 21, 24, 30, 33, 40, 43–44, 46–48, 49–53, 61, 75, 77–78, 95, 102, 118, 125–130, 132–133, 135, 137–143, 146, 151–156, 162–165, 168, 170–172, 196, 198, 207, 215, 217, 231, 239–240, 243, 245, 249, 255, 257, 259, 282 non-immersive 243, 250, 254–256, 259 planning 1–3, 7, 11–14, 23–24, 27–28, 30, 43–44, 50, 57, 59–60, 62–63, 66–67, 72, 75–76, 78, 86, 93–94, 96, 98, 101, 104, 106–107, 109, 114, 125–130, 132–133, 136–142, 150, 163, 165, 167, 170–173, 186, 196–197, 209, 212, 218, 223, 228, 231–232, 234, 236, 245–246, 249, 256, 261–262, 270, 272, 282, 284 production planning 76, 94, 114, 116–117, 125–127, 129, 132–133, 136–137, 141–142, 165, 170, 186, 245 project modeling and integration 243 resource planning 101, 122 reusable objects 211 set-based design 165, 170 simulation 10–11, 24, 31–32, 53, 58, 84, 86–87, 90–92, 118, 125, 145, 150–162, 165, 172, 174, 176, 196, 204, 207, 209, 211–212, 215, 218–221, 223, 227, 231–232, 240–241, 251, 253, 255 subcontractors 1–6, 11–12, 14, 29, 47, 53, 65, 93, 98, 101–107, 109–114, 119–122, 196–198, 202, 281 supply chain management 165 tacit knowledge 261, 264–265, 268, 277 training 29, 34–35, 55, 68, 72, 110, 146, 154, 197, 207, 208–209, 212, 258 transfer, 4D CAD 261 uncertainty 43, 76, 117, 121–122, 137, 140–141, 147, 163, 165, 167, 169–171, 174–175, 187, 230, 268

Subject index 287 virtual reality 29, 31, 75–76, 80–81, 83, 166, 211–213, 217, 243, 249 visual 2, 14, 33, 41, 43, 48, 52–54, 150, 196, 209–210, 231, 232, 250, 252–254, 282–283

visualization 2, 12, 14, 38, 43–44, 47, 51, 56, 67, 76, 83, 87, 121, 125–126, 130–132, 151, 162, 168, 174, 195, 209, 211, 218–219, 227–228, 231–235, 240, 246, 251, 282