Engineering Asset Management

“Engineering Asset Lifecycle Management” Proceedings of the 4thWorld Congress on Engineering Asset Management (WCEAM 2009) 28-30 September 2009

Editors Dimitris Kiritsis, Christos Emmanouilidis, Andy Koronios, and Joseph Mathew

Published by Springer-Verlag London Ltd

ISBN 978-1-84996-002-1

Proceedings of the 4th World Congress on Engineering Asset Management (WCEAM 2009) Ledra Marriott Hotel, Athens 28-30 September 2009

All Rights Reserved Copyright © Springer-Verlag 2010

No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the publisher.

PREFACE The 4th World Congress on Engineering Asset Management, WCEAM 2009, held at the Ledra Marriott Hotel in Athens, Greece from 28 to 30 September 2009, represents a milestone in the history of WCEAM. It is the first to be organised under the auspices of the newly formed International Society of Engineering Asset Management (ISEAM) who will host WCEAM on an annual basis as its forum for exchange of information on recent advances in this rapidly growing field. WCEAM 2009 was organised with the invaluable support of the newly formed Hellenic Maintenance Society (HMS) in Greece, who acted as the local host. After the inaugural WCEAM in July 2006 in the Gold Coast, Australia, visioned and initiated by the CRC for Integrated Engineering Asset Management (CIEAM) and organised in conjunction with the International Conference on Maintenance Societies (ICOMS) hosted by the Maintenance Engineering Society of Australia (MESA) and the International Maintenance Systems (IMS) Conference hosted by the Intelligent Maintenance Centre (IMS Centre), USA, the 2nd WCEAM was co-organised with the Condition Monitoring Conference hosted by the British Institute for Non-Destructive Testing (BINDT) in April 2007 in Harrogate, UK . The third congress in the series was organised in Beijing in October 2008 by a consortium comprising the Division of Mechanical and Vehicle Engineering, the Chinese Academy of Engineering, the China Association of Plant Engineering, the National Science Foundation Industry/ University Cooperative Research Center on the Intelligent Maintenance Systems Center (IMS), USA, the Diagnosis and Self Recovery Engineering Research Centre at Beijing Chemical Technology University and CIEAM. The theme of WCEAM 2009 was “Engineering Asset Lifecycle Management – A 2020 Vision for a Sustainable Future” and fits well with the lessons learnt from the recent financial and economical crises that impacted severely on the global economy and showed that a sustainable future requires consideration of the various lifecycle aspects of business, industrial and public good activities in their technical, organisational, economic, environmental and societal dimensions. The management of engineering assets over its lifecycle, which includes maintenance at its core, is a crucial element for global business sustainability and its importance is gradually being recognised by corporate senior management. Our seven distinguished keynotes at WCEAM 2009 presented developments in a number of these areas. ISEAM envisions WCEAM as a global annual forum that promotes the interdisciplinary aspects of Engineering Asset Management (EAM). In view of this vision, WCEAM seeks to promote collaboration between organisations who share similar objectives and where particular matters of common interests are discussed. The program for this year included a number of special sessions organised by the EFNMS, EURENSEAM and the Manufacturing Technology Platform (MTP) on Maintenance for Sustainable Manufacturing (M4SM) of the Intelligent Manufacturing Systems international program (IMS). In addition, a session on modern maintenance training along with a dedicated e-training workshop on Maintenance Management was organised by the EU project, “iLearn2Main”. WCEAM 2009 brought together over 170 leading academics, industry practitioners and research scientists from 29 countries to: • Advance the body of knowledge in Engineering Asset Management (EAM), • Strengthen the link between industry, academia and research, • Promote the development and application of research, and • Showcase state-of-the-art technologies. Over 120 scientific and technical presentations reported outputs of research and development activities as well as the application of knowledge in the practical aspects of: • Advanced Maintenance Strategies (RCM, CBM, RBI) • Condition monitoring, diagnostics, and prognostics

• • • • • • • • • • • • • • • • •

Decision support and optimization methods and tools Education and training in asset and maintenance management E-Maintenance Emerging technologies & embedded sensors and devices for EAM Human dimensions in integrated asset management Intelligent maintenance systems Lifecycle & sustainability considerations of physical assets Performance Monitoring and Management Planning and scheduling in asset and maintenance management Policy, Regulations, Practices and Standards for asset management Quality of information and knowledge management for EAM Risk management in EAM Safety, Health and Risk Management in EAM Self-maintenance and self-recovery engineering Strategic asset management for sustainable business Technologies for asset data management, warehousing and mining Wireless technologies in EAM

All full papers published in these proceedings have been refereed by specialist members of a peer review panel for technical merit. We would like to gratefully acknowledge the support of Bayer Technology GmbH as Gold Sponsor, the Intelligent Manufacturing Systems – IMS international organisation as Satchel and Dinner Sponsor, the CRC for Integrated Engineering Asset Management (CIEAM) as Bronze Sponsor of WCEAM 2009 and the ATHENA Research and Innovation Centre as the Welcome Reception Sponsor. Function sponsorship was undertaken by Gefyra SA and Atlantic Bulk Carriers Management Ltd while the SBC business channel, the Supply Chain & Logistics & Plant Management magazines and the supply-chain.gr portal were WCEAM 2009’s publicity sponsors. We would like to thank the Hellenic Maintenance Society (HMS) and the WCEAM Organising Committee for the enormous effort they have contributed to making this conference a success. Thanks are also due to members of the WCEAM International Scientific Committee for their efforts in both reviewing papers as well as promoting the congress within their networks. Athens, the “mythical” city of Athena, the ancient Greek goddess of Sophia, and the Parthenon offered WCEAM 2009 participants with an ideal environment for knowledge exchange and networking opportunities which will lead to the beginning of much new fruitful collaboration. Thank you for having been a part of WCEAM 2009. We look forward to meeting you at our next event, the 5th

WCEAM in Brisbane, Australia, from 25-27 October 2010. Dr. Dimitris Kiritsis Congress Chair

Dr. Christos Emmanouilidis Co-chair

Professor Joseph Mathew Co-chair

Congress Chairs Chair: Dr Dimitris Kiritsis, Ecole Polytechnique Fédérale de Lausanne, Switzerland. Co-Chairs: Dr Christos Emmanouilidis, C.E.T.I/R.C. Athena, Greece Professor Joseph Mathew, CRC for Integrated Engineering Asset Management (CIEAM), Australia. International Scientific Committee Adolfo Crespo Marquez, Spain Ajith Parlikad, UK Andrew Starr, UK Andy Koronios, Australia Andy Tan, Australia Anthony David Hope, UK Antonio J Marques Cardoso, Portugal Ashraf Labib, UK Basim Al-Najjar, Sweden Benoit Iung, France Birlasekaran Sivaswamy, Australia Bo-Suk Yang, Korea Brett Kirk, Australia Brigitte Chebel Morello, France Bruce Thomas, Australia Christos Emmanouilidis, Greece Dimitris Kiritsis, Switzerland Erkki Jantunen, Finland Gao Jinji, P R China George Tagaras, Greece Hong-Bae Jun, Korea Ioannis Antoniadis, Greece Ioannis Minis, Greece Jay Lee, USA Jayantha Liyanage, Norway

Jing Gao, Australia Joe Amadi-Echendu, South Africa Jong-Ho Shin, Switzerland Joseph Mathew, Australia Kari Komonen, Finland Kerry Brown, Australia Kondo Adjallah, France Len Gelman, UK Lin Ma, Australia Marco Garetti, Italy Margot Weijnen, The Netherlands Ming Zuo, Canada Mohd Salman Leong, Malaysia Mohsen Jafari, USA Nalinakash S Vyas, India Noureddine Zerhouni, France Peter Tse, Hong Kong, China Rhys Jones, Australia Roger Willett, New Zealand Seppo Virtanen, Finland Shozo Takata, Japan Stanislaw Radkowski, Poland Tadao Kawai, Japan Yiannis Bakouros, Greece

WCEAM 2009 Organising Committee Ian Curry, Hi Events, Australia Jane Davis, CIEAM, Australia Katerina Zerdeva, Zita Congress, Greece Professor Andy Koronios, CIEAM, University of South Australia, Australia Rhonda Hendicott, Hi Events, Australia Zacharias Kaplanidis, Zita Congress, Greece

Keynote Speakers Dr Claudio Boer - Intelligent Manufacturing Systems

Dr. Panayotis Papanikolas – GEFYRA S.A.

“IMS Global Network Support for Maintenance for Manufacturing”

“Construction, operation and lifecycle cost of ships - Realizing that before maintenance comes maintenability “

Claudio is Chairman of the International Steering Committee for Intelligent Manufacturing Systems (IMS). Over the past year, he has been involved with working to initiate an innovative program for researchers designed for easier global collaborations for new and ongoing research called the Manufacturing Technology Platform (MTP) program.

Heinz Cznotka - Bayer Technology Services GmbH “Proactive Asset Lifecycle Management” Dr Heinz Cznotka is the Director Competence Center Asset Management Consultancy and Sr. Risk and Reliability Consultant. He has 20 years experience in the oil, gas & petrochemical industry as: Project Manager, Reliability Manager, Production Manager, Managing Director Maintenance Services and Technical Risk Manager. Dr Cznotka’s core competences are Risk-based Maintenance (RBM), Reliability-centered Maintenance (RCM), Turnaround optimization; Lifecycle Length and Cost Optimization; Reliability & Maintenance Engineering and Optimization; Reliability & Maintenance related EHS Management; Reliability & Maintenance Production and In-Service Inspection.

Richard Edwards - EFNMS/IAM “Why do we need asset management trends and issues?” Richard is a member of the Council of the Institute of Asset Management and the current chairman of the Asset Management Technical and Professional Network, a joint venture between the IET and the IAM. He is also a Board member of the IAM. A Director of AMCL for ten years, Richard has extensive experience in the application and assessment of Asset Management. ‘

Eric Luyer - IBM Corporation “Leveraging Smart Asset Management in Engineering and Product Lifecycle Management” Eric Luyer has global responsibility for managing Industry Marketing activities for the Industrial Manufacturing sector. With more than 25 years of experience, Eric has developed in-depth expertise in the industry, working in Financial, ERP and Enterprise Asset Management software application environments, initially in Europe and the last seven years worldwide. Eric Luyer has held senior positions in Sales Management, Indirect Channels, Alliance Partner Management and Industry Marketing - working for software solutions providers, such as Comshare, Global SSA, Baan/Invensys, MRO Software. Currently Eric is taking the position of IBM’s worldwide Manager Product Marketing for Maximo Asset Management – positioning Asset Management in Industrial Manufacturing industries.

Panayotis is Vice-Chairman & Managing Director of GEFYRA S.A., the concessionnaire company of the Rion - Antirion Bridge. Holding the position of Technical Director during the construction of the longest multi-span cable stayed bridge in the world, built in aggressive environment in terms of durability, seismicity and wind, he was in charge of developing the inspection, monitoring and maintenance management plan of the bridge. Five years of operation have helped in fine-tuning what is considered today as one of the most complete structural asset management plans for bridges.

Jürgen Pothoff - Bayer Technology Services “Proactive Asset Lifecycle Management” Jürgen has 16 years experience with globally operating Engineering, Procurement & Construction (EPC) Companies working on projects in Europe, Americas, Asia, MENA and Australia servicing the petrochemical, refineries, chemical process industry. He has worked as a Lead Process Engineer and Feasibility Study Leader for world scale petrochemical plants; Sales & Project Manager for lump sum turnkey EPC projects including a 4 year local assignment for the development of the Australian market; Establishment / Management of Maintenance Management & Contracting services based on risk based maintenance concept and 3 years as a Reliability Engineering Manager with major chemical company, on the development & global implementation of risk based maintenance.

Kristian Steenstrup - Gartner Inc. “Operational Technology and the relationship to IT Management” Kristian is research vice president in Gartner’s Australian branch. He conducts and delivers research on Enterprise Business Systems including ERP, EAM, SCM, CRM and c-commerce. Kristian is responsible for vendor analysis in the Asia/Pacific market and is global research leader for Asset Intensive ERP II. Before joining Gartner Kristian worked in ERP and EAM system design and delivery for over 15 years in a number of global markets. During this period he was directly involved in emerging technologies in Asset Intensive industries such as Utilities, Mining, Rail and Defence.

Panos Zachariadis - Atlantic Bulk Carriers “Construction, operation and lifecycle cost of ships – Realising that before maintenance comes maintainability” Panos is Technical Director of Atlantic Bulk Carriers Management Ltd. From 1984 to 1997 he was Marine Superintendent for a New York bulk carrier and oil tanker shipping company. His shipping experience spans diverse areas such as sea service in bulk carriers and oil tankers, supervision of dry dock repairs, new building specifications and supervision, ship operations andchartering. Mr. Zachariadis holds a BSc degree in Mechanical Engineering from Iowa State University and a MSE degree in Naval Architecture and Marine Engineeringfrom the University of Michigan. He is a founding member of Marine Technical Managers Association (MARTECMA) of Greece.

SPONSORS Gold Sponsor

Satchel & Congress Dinner Sponsor

Welcome Reception Sponsor

Bronze Sponsor

Wednesday Lunch Sponsor

Publicity Sponsors

Exhibitors

Bayer Technology Services GmbH 51368 Leverkusen, Germany E-mail: [email protected] www.bayertechnology.com

Monday Morning Tea Sponsor

Ledra Marriott Hotel, Athens, Greece

Program The organisers reserve the right to make changes to the program.

Sunday 27 September 2009 1930

Dinner at Horizons Restaurant - Optional dinner at additional cost

Sponsored by Bayer Technology Services

Monday 28 September 2009 0700

Exhibitors and poster presenters mount displays

0830

Conference Registration

0900-0930

Welcome – Opening session

Plenary session sponsored by Bayer Technology Services

Chair: Dimitris Kiritsis 0930-1015

Keynote address 1: Heinz Cznotka & Jürgen Potthoff, Bayer Technology Services GmbH Proactive Asset Lifecycle Management

1015-1100

Keynote address 2: Kristian Steenstrup, Research VP, Gartner Inc Operational technology and its relationship to IT Management

1100-1120

Coffee break

1120-1300

Sessions

1300-1420

Session1: EURENSEAM - 1 - Strategic Engineering Asset Management

Session 2: Transport, Building and Structural Session 3: Lifecycle & Sustainability Asset Management Considerations Of Physical Assets

Chair: J P Liyanage

Chair: Tony Hope

Chair: Andy Koronios

Ahonen, T., Collaborative development of Maintenance investment management: A case study in pulp and paper industry

Kaphle, M., Tan, ACC., Kim, E. and Thambiratnam, D., Application of acoustic emission technology in monitoring structural integrity of bridges

Frolov, V, Mengel, D, Bandara, W, Sun, Y and Ma, L., Building an ontology and process architercture for engineering asset management

Al-Najjar, B. and Ciganovic, R., A model for more accurate maintenance decisions

Nayak, R., Piyatrapoomi, N. and Weligamage, J., Application of text mining in analysing road crashes for road asset management

Karray, MH, Morello, BC and Zerhouni, N, Towards A Maintenance Semantic Architecture

Gudergan, G., The House of Maintenance - Identifying the potential for improvement in internal maintenance organizations by means of a capability maturity model

Pérez, AA, Vieira, ACV, Marques Cardoso, AJ, School Buildings Assets - Maintenance Management and Organization for Vertical Transportation Equipment

Koronios, A., Steenstrup, C. and Haider, A., Information and Operational Technologies Nexus for Asset Lifecycle Management

Parida, A. and Kumar, U., Integrated strategic asset performance assessment

Phillips, P., Diston, D., Starr, A., Payne, J., and Pandya, S., A review on the optimisation of aircraft maintenance with application to landing gears

Matsokis, A., Kiritsis, D., An advanced method for time treatment in product lifecycle management models

Rosqvist, T., Assessing the subjective added value of value nets: which network strategies are really win-win ?

Piyatrapoomi, N. and Weligamage, J. Risk-based approach for managing road surface friction of road assets

Shin, J-H, Kiritsis, D., Xirouchakis, P., Function performance evaluation and its application for design modification based on product usage data

Lunch

World Congress of Engineering Asset Management 2009

Program Monday 28 September 2009 ... continued 1420-1540

Sessions Session4: EURENSEAM - 2 - Strategic Engineering Asset Management

Session 5: Transport, Building and Structural Asset Management

Session 6: Technologies For Asset Data Management, Warehousing & Mining

Chair: Bassim Al-Najjar

Chair: Joe Amadi-Echendu

Chair: Michael Purser

Al-Najjar, B., A Computerized model for assessing the return on investment in maintenance

Amadi-Echendu, J.E., Belmonteb, H., von Holdtc, C., and Bhagwand, J. A case study of condition assessment of water and sanitation infrastructure

Haider, A., Open Source Software Development for Asset Lifecycle Management

González Díaz, V., Fernández, JFG, Crespo Márquez, A., Case study: Warranty costs estimation according to a defined lifetime distribution of deliverables

Nastasie, D., Koronios, A., The role of standard information models in road asset management

Kans, M., Assessing maintenance management IT on the basis of IT maturity

Schuh, G. And Podratz, K., Remote service concepts for intelligent tool-machine systems

Nastasie, D., Koronios, A., The diffusion of standard information models in road asset management: - A study based on the human - technology environment

Mathew, A., Purser, M., Ma, L. and Barlow, M., Open standards-based system integration for asset management decision support

Wijnia, YC and Herder, PM, The state of Asset Management in the Netherlands

Ninikas, G., Athanasopoulos, Th., Marentakis, H., Zeimpekisa, V., and Minis, I., Design and implementation of a real-time fleet management system for a courier operator

Peppard, J., Koronios, A. Gao, J., The data quality implications of the servitization - theory building

1540-1600

Coffee break

1600-1645

Keynote address 3: Richard Edwards, European Federation of National Maintenance Societies & Institute of Asset Management Why do we need asset management: trends and issues Chair: Christos Emmanouilidis

1650- 1810 Sessions

1930-2100

Session7: EFNMS - Engineering Asset Management in Europe

Session 8: Decision Support & Optimisation Methods & Tools

17.00 - 19.00 ISEAM AGM

Chair: Kari Komonen

Chair: Marco Macchi

Chair: Joe Mathew

Benetrix, L., Garnero, MA and Verrier, V., Asset management for fossil-fired power plants: methodology and an example of application

Cenna, AA, Pang, K, Williams, KC, Jones, MG, Micromechanics of wear and its application to predict the service life of pneumatic conveying pipelines

Olsson, C., Labib, A. and Vamvalis, C. CMMS – Investment or Disaster ? Avoid the Pitfalls

Kim, JG, Jang, YS, Jeong, HE, Lim, J and Choi, BK, Flexible coupling numerical analysis method

Ulaga, S., Jakovcic, M. And Frkovic, D. Condition monitoring supported decision processes in maintenance

Godichaud M. , Pérès F. and Tchangani A., Disassembly process planning using Bayesian network

Welcome Cocktail Function - Sponsored by ATHENA Research & Innovation Centre

Ledra Marriott Hotel, Athens, Greece

Program Tuesday 29 September 2009 830

Delegate Arrival & Registration

0900-0945

Keynote address 3: Claudio Boer, Chairman, Intelligent Manufacturing System (IMS) IMS Global Network Support for Maintenance for Manufacturing Chair: Marco Garetti

1000-1140

Sessions Session 9: Maintenance for Sustainable Manufacturing - 1

Session 10 - Strategic asset management for sustainable business

S11 - Advanced Maintenance Strategies (RCM,CBM,RBI)

Chair: Marco Garetti

Chair: Ajit Parlikad

Chair: Lin Ma

Garetti, M. Welcome & Introduction to MTP M4SM Special Session

Furneaux, C.W., Brown, K.A., Tywoniak, S., and Gudmundsson, A., Performance of public private partnerships: an evolutionary perspective

Gorjian, N., Ma, L., Mittinty, M., Yarlagadda, P. and Sun, Y., A review on degradation models in reliability analysis

Pantelopoulos, S. (Industrial talk), Product maintenance in the ‘Internet of things’ world

Labib, A., Maintenance strategies: a systematic approach for selection of the right strategies

Gorjian, N., Ma, L., Mittinty, M., Yarlagadda, P. and Sun, Y., A Rreview on reliability models with covariates

Gómez Fernández J F*, Álvarez Padilla F J, Fumagalli L, González Díaz V, Macchi M, Crespo Márquez A, Condition monitoring for the improvement of data center management orientated to the Green ICT

Rezvani, A., Srinivasan, R., Farhan, F., Parlikad, AK., and Jafari, M., Towards Value Based Asset Maintenance

Muhammad, M., Majida, A.A. and Ibrahima, N.A., A Case study of reliability assessment for centrifugal pumps in a petrochemical plant

Liyanage J P, Badurdeen F, Strategies for integrating maintenance for sustainable manufacturing: developing integrated platforms

Yeoh, W., Koronios, A. And Gao, J., Ensuring Successful Business Intelligence Systems Implementation: Multiple Case Studies in Engineering Asset Management Organisations

Sun, Y., Ma, L., Purser, M. and Fidgeb, C., Optimisation of the reliability based preventive maintenance strategy

Tsutsui M, Takata S, Life Cycle Maintenance Planning System in consideration of operation and maintenance integration

Gao, J., Koronios, A., Kennett, S., Scott, H., Data quality enhanced asset management metadata model

Lee, WB, Moh, L-S, Choi, H-J, Lifecycle Engineering Asset Management

Session 12: Maintenance for Sustainable Manufacturing - 2

Session 13 - Planning, Scheduling & Performance Monitoring

Session 14 - Decision Support & Optimisation Methods & Tools

Chair: Shozo Takata

Chair: Seppo Virtanen

Chair: Colin Hoschke

Cannata A, Karnouskos S, Taisch M, Dynamic e-Maintenance in the era of SOA-ready device dominated industrial environments

Haider, A., Driving innovation through performance evaluation

Chebel-Morello, B., Haouchine, K., Zerhouni, N., Methodology to Conceive A Case Based System Of Industrial Diagnosis

Emmanouilidis, C. and Pistofidis, P. Design requirements for wireless sensor-based novelty detection in machinery condition monitoring

Kim, D., Lim, J-H., Zuo, MJ., Optimal schedules of two periodic preventive maintenance policies and their comparison

Lim, JI, Choi, BG, Kim, HJ, Kim, JG, and Park, CH., Optimim design of vertical pump for avoiding the reed frequency

Kans M, Ingwald A, Analysing IT functionality gaps for maintenance management

Lipia, TF, Zuo, MJ, and Lim, J-H., Optimal Replacement Decision Using Stochastic Filtering Process to Maximize Long Run Average Availability

Mokhtar, AA, Muhammad, M and Majida, MAA, Development of spreadsheet based decision support system for product distributions

Garetti, M - Session Wrap-Up

Seraoui R., Chevalier R. and Provost D., EDF’S plants monitoring through empirical modelling: performance assessment and optimization

Smalko Z, Woropay M, Ja, Z, The diagnostic decision in uncertainly circumstances

1140-1200

Coffee break

1200-1310

Sessions

World Congress of Engineering Asset Management 2009

Program Tuesday 29 September 2009 ... continued 1310-1430

Lunch

1430-1610

Sessions Session 15: Maintenance for Sustainable Manufacturing - 3 - Education & Training

Session 16: Advanced Maintenance Strategies

Session 17: Condition Monitoring, Diagnostics & Prognostics

Chair: Jan Franlund

Chair: Stanislow Radkowski

Chair: Tony Rosqvist

Bakouros, Y., Panagiotidou, S, and Vamvalis, C., Education and Training Needs in Maintenance: How you conduct a selfaudit in Maintenance Management

Bey-Temsamani, A., Engels, M., Motten, A., Vandenplas, S. and Ompusunggu, AP., Condition-based maintenance for OEM’s by application of data mining and prediction techniques

Chen T., Xu XL, Wang, SH, Deng SP. , The Construction and Application of Remote Monitoring and Diagnosis Platform for Large Flue Gas Turbine Unit

Emmanouilidis, C., Labib, A, Franlund, J., Dontsiou, M, Elina, L., Borcos, M., iLearn2Main: an e-learning system for maintenance management training

Gontarz, S. and Radkowski, S., Shape of specimen impact on interaction between earth and eigenmagnetic fields during the tension test

Gu, DS, Kim, BS, Lim, JI, Bae, YC, Lee, WR, Kim, HS, Comparison of vibration analysis with different modeling method of a rotating shaft system

Franlund, J, Training and Certification of Maintenance and Asset Management Professionals

Kiassat, C and Safaei, N., Integrating human reliability analysis into a comprehensive maintenance optimization strategy

Kim, BS, Gu, DS, Kim, JG, Kim, YC and Cho, BK, Rolling element bearing fault detection using acoustic emission signal analyzed by envelope analysis with discrete wavelet transform

Macchi, M., Ierace, S., Education in Industrial Maintenance Management: Feedback From Italian Experience

Mazhar, MI, Salman, M and Howard, I, Assessing the reliability of system modules used in multiple life cycles

Kim, H-E, Tan, ACC, Mathew, J, Kim, EYH, Cho, BK, Prognosis of Bearing Failure Based on Health State Estimation

Starr, A. Bevis, K., The role of education in industrial maintenance: the pathway to a sustainable future

Radkowski S, Gumiński R, Impact of vibroacoustic diagnostics on certainty of reliability assessment

Xu XL, Chen T, Wang SH, Research on Data-Based Nonlinear Fault Prediction Methods in Multi-Transform Domains for Electromechanical Equipment

Maintenance for Sustainable Manufacturing - 4 - M4SM Project kick-off 1° M4SM Meeting

Session 18: Emerging Technologies in EAM

Session 19: Condition Monitoring, Diagnostics & Prognostics

Chair: Marco Garetti

Chair: Bo-Suk Yang

Chair: Andrew Starr

Espíndola, D., Pereira, CE, Pinho, M., IM:MR - A tool for integration of data from different formats

Jasiński M., Radkowski S., Use of bispectral-based fault detection method in the vibroacoustic diagnosis of the gearbox

Mikail F. Lumentut , Ian M. Howard, Theoretical study of piezoelectric bimorph beams with two input base-motion for power harvesting

Maszak, J., Local meshing plane as a source of diagnostic information for monitoring the evolution of gear faults

Shim, M-C, Yang, B-S, Kong, Y-M, Kim, WC, Wireless condition monitoring system for large vessels

Yang, SW, Widodo, A, Caesarendra, W, Oh, JS, Shim, MC, Kim, SJ, Yang, BS and Lee, WH, Support vector machine and discrete wavelet transform for strip rupture detection based on transient current signal

Smit, JJ, Djairam, D., Zuang, Q., Emerging Technologies and Embedded Intelligence in Future Power Systems

Ierace, S., Garetti, M. and Cristaldi, L., Electric Signature Analysis as a cheap diagnostic and prognostic tool

1610-1630

Coffee break

1630-1750

Sessions

1915

Departure for Gala Dinner

1945

Conference Dinner - Athens Yacht Club Sponsored by IMS Keynote address 4 - Panos Zachariadis, Atlantic Bulk Carriers Construction, operation and lifecycle cost of ships - Realizing that before maintenance comes maintenability Chair: N. Nassiopoulos

Ledra Marriott Hotel, Athens, Greece

Program Wednesday 30 September 2009 0845

Delegate Arrival

0900-1300

WORKSHOP: INTEGRATION AND INTEROPERABILITY IN ENGINEERING ASSET MANAGEMENT (EAM)

0900-1020

Sessions Session 20: e-Maintenanance

Session 21: Policy, Regulations, Practices & Standards For Asset Management

Session 22: Condition Monitoring, Diagnostics & Prognostics

Chair: Errki Jantunen

Chair: Ashraf Labib

Chair: Andy Tan

Baglee, D., The Development of a Mobile e-maintenance system utilizing RFID and PDA Technologies

Haider, A., A Roadmap for information technology governance

Kim, EY., Tan, ACC., Mathew, J. and Yang, B-S., Development of an Online Condition Monitoring System for Slow Speed Machinery

Jantunen, E. Gilabert, E., Emmanoulidis, C. and Adgar, A., e-Maintenance: a means to high overall efficiency

Mathew, A., Purser, M. Ma, L. and Mengel, D., Creating an asset registry for railway electrical traction equipment with open standards

Rgeai, M., Gu, F., Ball, A., Elhaj, M., Ghretli, M. Gearbox Fault Detection Using Spectrum Analysis of the Drive Motor Current Signal

Oyadiji, SO, Qi, S. and Shuttleworth, R., Development of Multiple Cantilevered Piezo Fibre Composite Beams Vibration Energy Harvester for Wireless Sensors

Stapelberg, RF, Corporate Executive Development for Integrated Assets Management in a New Global Economy

Zhu, Z., Oyadiji, SO, and Mekid, S., Design and Implementation of a Dynamic Power Management System for Wireless Sensor Nodes

1020-1040

Coffee break

1040-1200

Sessions

1200-1320

Session 23: Technologies For Asset Data Management, Warehousing & Mining

Session 24: Advanced Maintenance Strategies (RCM,CBM,RBI)

Session 25: Condition Monitoring, Diagnostics & Prognostics

Chair: Matt Barlow

Chair: Ioannis Bakouros

Chair: Ioannis Antoniadis

Grossmann, G, Stumptner, M, Mayer, W, and Barlow, M, A Service oriented architecture for data integration in asset management

Bohoris, G.A. and Kostagiolas, P.A., Inferences on nonparametric methods for the estimation of the reliability function with multiply censored data

Mpinos, CA and Karakatsanis, T., Development of a dynamic maintenance system for electric motor’s failure prognosis

Natarajan, K., Li, J. and Koronios, A., Data mining techniques for data cleaning

Gilabert, E., Gorritxategi, E., Conde, E., Garcia, A., Areitioaurtena, O. and Amaya Igartuaa, An advanced maintenance system for poligeneration applications

Gryllias, K.C., Yiakopoulos, C. and Antoniadis, I., Automated diagnostic approaches for deffective rolling element bearing using minimal training pattern classificaiton methods

Natarajan, K., Li, J. and Koronios, A., Detecting mis-entered values in large data sets

Kostagiolas, P.A., and Bohoris, G.A., Finite sample behaviour of the Hollander-Proschan goodness of fit with reliability and maintenance data

Pang, K., Cenna, AA, Williams, KC, and Jones, MD, Experimental determination of cutting and deformation energy factors for wear prediction of pneumatic conveying pipeline

Apostolids, H. (industrial talk), Crisis in Maintenance and Maintenance in Crisis: Opportunities for Maintenance Re-engineering

Yachiku, H., Inoue, R., and Kawai, T., Diagnostic Support Technology by Fusion of Model and Semantic Network

Lunch - sponsored by GEFYRA SA

World Congress of Engineering Asset Management 2009

Program Wednesday 30 September 2009 ... continued 1320-1520

Sessions Session 26: Workshop for e-training in Maintenance Management (HMS, M4SM, iLearn2Main)

Session 27: Safety, health and risk management in EAM

Session 28: Condition Monitoring, Diagnostics & Prognostics

Chair: Christos Emmanouilidis

Chair: Pantelis Botsaris

Chair: Tadao Kawai

Papathanassiou, N., Emmanouilidis, C., e-Learning in Maintenance Management Training and Competence Assessment: Development and Demonstration

Botsaris, P.N., Naris, A.D., and Gaidajis, G., A Risk Based Inspection (RBI) preventive maintanance programme: a case study

Fumagalli, L., Jantunen, E., Garetti, M. and Macchi, M., Diagnosis for improved maintenance services: Analysis of standards related to Condition Based Maintenance

Maintenance Management: Live & Interactive e-training and e-assessment workshop

Papazoglou, IA, Anezirisa, ON, Konstandinidou, M, Bellamy, LJ, Damen, M, Assesing occupational risk for contact with moving parts of machines during maintenance

Widodo, A., and Bo-Suk Yang, Machine prognostics based on survival analysis and support vector machine

Skroubelos, G., Accident causes during repair and maintenance activities and managerial measures effectiveness

Younus, AM, Widodo, A and Yang, B-S., Image Histogram Features Based Thermal Image Retrieval to Pattern Recognition of Machine Condition

Training in Maintenance Management Panel Discussion & Evaluation (Jan Franlund, Ashraf Labib, Andrew Starr, Yiannis Bakouros, Cosmas Vamvalis)

Elforjani, M. Mba, D. Acoustic emissions observed from a naturally degrading slow speed bearing and shaft

A Addali, S Al-lababidi, H Yeung, D Mba, Measurement of gas content in two-phase flow with acoustic emission

1520-1540

Coffee break

1540-1625

Congress Closing Keynote Address 6: Panayiotis Papanikolas, GEFYRA SA Maintenance Management of the Rion-Antirion Bridge

1625-1710

Congress Closing Keynote Address 7: Eric Luyer, IBM Corporation Leveraging Smart Asset Management in Engineering and Product Lifecycle Management Chair: Joe Mathew

1710-1730

Closing Remarks - End of WCEAM 2009

1830

Visit to new Acropolis Museum and dinner at Dionysos Restaurant - optional at additional cost

Proceedings of the 4th World Congress on Engineering Asset Management Athens, Greece 28 - 30 September 2009

COLLABORATIVE DEVELOPMENT OF MAINTENANCE INVESTMENT MANAGEMENT – A CASE STUDY IN PULP AND PAPER INDUSTRY Jouko Heikkiläa, Toni Ahonena, Jari Neejärvib, Jari Ala-Nikkolac a

VTT Technical Research Centre of Finland, P.O. Box 1300, FI-33101 Tampere, Finland; [email protected] b

Myllykoski Paper, Myllykoskentie 1, FI-46800 Anjalankoski, Finland; [email protected] c

ABB Service, Myllykoskentie 1, FI-46800 Anjalankoski, Finland; [email protected]

The interest in purchasing external maintenance services increases as a strategic choice, as manufacturing companies increasingly focus on their core business. In the case presented in this paper, an external service provider has taken the responsibility of the maintenance functions as a strategic partner. A development project was started to support the creation of profound and functional partnership. Focus of the project was to support the practical collaboration between the partners. The development focused on two activities, in which the both organisations clearly have important roles. These two activities are: management of maintenance investments and operator involvement in maintenance activities. The idea was to improve the collaboration practices by practical collaborative development of these two key activities. Maintenance investment decision making requires functional collaboration between the organisations. The maintenance organisation is responsible for making investment proposals and production organisation makes the final decisions. Originally not much collaboration was planned in this process. Furthermore there was a demand for improving and clarifying the reasoning for proposals. To meet this demand, a working group comprising representatives of both organisations and researchers developed proper commonly agreed criteria, a tool and a process for preparation of maintenance investment proposals. In the developed investment process, the collaboration aspect was taken into account. In this paper, we present the collaborative development process, resulting maintenance investment management development and key findings regarding the importance of practical development efforts done together in building trust and collaboration in a business partnership. Key Words: Maintenance investment, Pulp and paper industry, Decision making, Management

1

INTRODUCTION

Demand for paper products is not expected to grow, especially in Europe. Therefore investments in new production lines will be very rare and the role of maintenance investment management will increase in engineering asset management of ageing production lines. A maintenance investment here means typically a replacement of worn equipment, but excluding scheduled maintenance and reparation of failures. While scheduled maintenance and reparation activities both include replacements of worn equipment, too, maintenance investments are typically more expensive and require specific planning and decisionmaking which is substantially based on cost-effectiveness (which makes it as an investment). Unlike other investments, the main focus in maintenance investments is on restoring the original performance. The payback from a maintenance investment comes mainly from decreasing (corrective) maintenance and unavailability costs. Outsourcing of services in industry is a common trend that has been started in 1960s and which has been significantly increased in the last 20 years (as summarised by, for example, Hendry (1995), Bailey et al. (2002), and Huai and Cui (2005). Maintenance is one of the most commonly outsourced services (Bailey et al. 2002, Tarakci et al. 2009). The outsourcing trend is still going strong, even though numerous and severe drawbacks has been reported (Hendry 1995 and anon. 2007). In Finnish paper industry traditionally nearly all activities at plant has been accomplished by the in-house maintenance organisation. In a competition for market shares, paper companies have searched for effectiveness by focusing on core business. Thus, the companies are looking for and establishing outsourcing of functions, which are not seen to be their core business. The outsourcing trend was started from such activities as canteen and cleaning services, and now the maintenance services are

1

being outsourced. After such a long tradition of in-house operations, outsourcing is likely to raise fears and opposition in organisation, as well as practical challenges in partners’ roles, distribution of work and cooperation. Improving the performance of service, focusing management activities on core business and reducing costs are the most common motives for outsourcing. However, in many cases, there have been difficulties to reach these goals, because of unexpected management needs and “hidden costs” related to cultural and cooperation aspects of organisations (Hendry 1995, Bailey et al. 2002, anon. 2005, Huai & Cui 2005). In the case presented in this paper, the maintenance services were outsourced in the beginning of 2007. This started the build-up of a new form cooperation between a paper company and a maintenance company. To support managerial and practical cooperation, a development project was started. The development project focused on two practical issues in cooperation: 1) maintenance investment management process and 2) development of the role of operators in maintenance and development. In this paper the first task and its results are presented. 2 HOW THE MAINTENANCE INVESTMENT MANAGEMENT WAS DEVELOPED The initial state in the beginning of the development project was that the agreement between the paper company and the maintenance company had been signed and the agreed maintenance operations had been started. Most of the maintenance staff remained, but some changes were made in maintenance management. Maintenance tasks and responsibilities were defined in the agreement. One of the agreed tasks allocated to the maintenance organisation was to prepare maintenance investment proposals. The proposals are then evaluated and investment decisions will be made by the paper company. A need to improve and clarify the reasoning for investments had already been identified and the two organisations proposal-decision-model emphasized this need. Too often the proposed reason for a maintenance investment was only that “it has to be done”, which makes the evaluation of the necessity and profitability of the proposal quite difficult. Initially it was supposed – especially on the paper company side – that the maintenance company would prepare the maintenance investment proposals independently. Quite soon it appeared to be evident, that at least some cooperation would be useful or even required. There were two reasons for the need of cooperation: 1) the novelty of maintenance organisation at the plant and 2) the need to coordinate maintenance investments and development investments. Even though most of the employees in the new maintenance organisation had been working already in the old organisation, the organisation itself and many of its managers were new at the plant. In addition to that, the existing maintenance and equipment history documentation was not so complete that it could cover all the history knowledge of the experience personnel. Additional challenge was the change of the maintenance information system. Several activities have been made to improve the maintenance documentation in order to make it a more reliable and complete basis for maintenance investment planning. Meanwhile, experienced production and maintenance personnel has important additional role in producing and preparing information. Hopefully in near future, the maintenance history data will be more ready to be used for maintenance investment planning. This will reduce the need for cooperation in information collection for this specific purpose in future. In some cases there are connections and overlapping between the maintenance investment and development investment: A target of a planned maintenance investment may be also a part of a planned development investment. Or, a planned development investment may set some new requirements, which are worthwhile to be taken into account in maintenance investments. Since, in this case, the preparation of maintenance investments and development investments has been separated into different organisations, a frequent enough communication between these organisations should be especially planned and agreed. Coordination between development investments and maintenance investments would have taken place later in the decision-making process anyway, but coordination already while planning will save some preparation work and enable better results. Since improvement of collaboration between two organisations was the basic objective of the project, collaborative process was chosen to be used in developing maintenance investment preparation. Collaborative development process meant that representatives of both organisations together developed and agreed the method to be used starting from the criteria to be used in investment selection and including all important aspects, such as information to be collected, means for information collection, cooperation related to information handling, responsibilities and annual schedules for the phases of preparation. External consults from VTT guided the development process and supported it with their knowledge. The development consisted of 11 workgroup meetings during one year and additional development work between the meetings. The development could have been done as an expert work by the consults or by either of the organisations by themselves. Possibly, in that case, the developed method would have been more sophisticated and the development might have been more efficient (up to the method document). However, the equality based collaborative development process aimed to ensure an undisturbed start-up and sustainable use of the new practice, by dispelling suspicions and practically constructing collaboration between the two organisations. The developed method is at the moment practically implemented at the plant, which means still some learning and finalizing of practices. The final success of the work will still be seen in future.

2

3

THE DEVELOPED MAINTENANCE INVESTMENT IDENTIFICATION AND PROPOSAL PROCESS

The development work in this case focused on the first phase of a maintenance investment management process – the phase that includes the tasks from the collection of information until preparation of investment proposals. The second phase – handling of the proposals and the final decision making – was outside of the scope of this case. The criteria to be used in investment candidate identification, candidate selection and final decision making were carefully selected. From the beginning it was clear to all parties that the economic criteria – namely the profitability of an investment – would be dominant in most cases. Especially production loss and unavailability related costs would be important criteria. Maintenance costs could be triggering criteria in some cases. Opportunity to improve the system performance (above the original) is never a main criterion for a maintenance investment, but it may be an additional factor supporting the investment. The improvement may be related to production capacity, occupational or environmental safety, or quality. Plans for development investments should always be checked and taken into account, when maintenance investments are prepared. Capability to invest may be a restricting factor in some situations. Capability may be restricted by financial, resource or production reasons. In practice, concerning a single maintenance investment, capability to invest is not an important issue, since the maintenance investments are carried out within an annual budget. When proposals for maintenance investments are prepared, it should also be checked, if other options were possible and more profitable. The other options – instead of maintenance investment – may be an improvement of maintenance procedures or a development investment. One of the main objectives in the development was to improve the information collection for the basis of potential maintenance investment identification. It is known that collaboration and related information exchange is relevant from the integrated perspective of production and maintenance planning. E.g. Sloan et al. (2000) have studied the combined production and maintenance models and concluded that combined models result in significantly greater (25 %) reward compared to traditional method. An important source of information is the maintenance management information system. However, after a strict examination it appeared that the failure and maintenance history data collected in the system was not complete and detailed enough to be used as an only information source for the identification of potential maintenance investments. One specific drawback was that, even though very extensive data on downtime existed in production management system, it could not be automatically linked to the failure and maintenance data in maintenance information system. Thus, detailed enough information on how much each failure cause loss of production was not easily available. Possibilities to link the two information systems are being examined and management activities have been carried out to improve the maintenance and failure documentation. The quality of maintenance data has improved already during the project. In addition to above mentioned general improvements in maintenance data collection, an annual process was developed (figure 1) focusing to collect and analyse maintenance investment related information especially. At the end of the process, the maintenance investment proposals are prepared. The process consists of three main phases: 1) Identification and selection of potential management investments 2) Profitability assessment based on a cost-benefit analysis 3) Decision making and proposal finalisation. By making the preparation of maintenance investment proposals a year-long – or actually a continuous – process, better quality of proposals was sought.

3

Continuous tasks

Responsible

Identification of potential for maintenance investments on the basis of incidents causing downtime and maintenance costs. Rough estimate of investment benefits and costs. Supervisor Local service manager Responsible

Local service manager

•

Morning meetings

• •

Thursday meetings Root cause analyses Periodical tasks

Responsible

Examination of downtime January Decemb February information er Novemb er

March

October Examination of downtime information

Service manager

Preparation of investment proposals

August

Local service manager

April

Septemb er

Local service manager

Examination of downtime information

May

July

Examination of downtime information

Profitability assessment Proposals for major investments Selection of potential minor investments Examination of non-indicated risks

Service and production managers

June

Service and production managers

Figure 1. Annual schedule of preparation of maintenance investment proposals

The basis of the process is in daily and weekly maintenance and production meetings. In these meetings it is frequently checked, whether there has happened something, which might indicate a need for a maintenance investment. These potential targets for maintenance investments are documented in a formal common inventory (data base). The documentation in inventory should include a note on why the investment should be done (what should be improved) and a coarse cost-benefit evaluation. An additional information source for potential maintenance investments is the root cause analyses which are carried out for any failure or disturbance which have caused at least two hour downtime in production. An increased number of failures and increased need for maintenance is one kind of indication of a required maintenance investment. However, all aging equipment does not so clearly show such symptoms of replacement need. Such equipment requires regular risk assessment as means for identification of maintenance investment need. The risk assessment is based on the criticality classification of equipment: the criticality in relation to production and safety has been defined for each device in the plant. In risk assessment the main factors to be examined are: -

equipment age compared to the typical age of such equipment

-

changes in environment, use or functional requirements for equipment

-

changes in availability of spare parts and services.

Based on these factors, it will be evaluated, on which year the replacement would most probably be required. The evaluated year of failure is related to the likelihood part of risk assessment. The consequence aspect of risk assessment is taken into account by comparing planned and unplanned replacement cases. The difference between the costs of planned and unplanned cases affects to the preferable timing of replacement: if the unplanned replacement is much more expensive, the investment in replacement is likely to be profitable even long before the evaluated replacement date. On the other hand, if the difference in costs is small, a risk can be taken by postponing the

4

replacement even after the evaluated date. On the basis of risk assessment, potential targets for maintenance investment are added to the potential maintenance investment inventory. Risk assessment is carried out annually. The inventory of potential maintenance investments is examined four times a year. In this examination the information related to these potential investments is updated and complemented. This quarterly examination serves the follow-up. In the second examination of the year the cost-benefit analysis is carried out. On the basis of the analysis, budget proposals for major investments are prepared and forwarded to decision making, and other (minor) maintenance investments are selected for proposal finalisation. In the third annual examination, maintenance investment proposals are finalised. 4

PROFITABILITY ASSESSMENT BASED ON A COST-BENEFIT ANALYSIS

The practical approach and tool developed to support the profitability assessment in this context was based on an LCP calculation model. The developed model is a combination of the practices utilised at the plant earlier, cost-based failure analyses (e.g. Rhee & Ishii, 2003) and LCP principles for a dynamic environment (Ahlmann, 2002). So far the profitability assessments have been made mostly based on information on the investment payback time which was, however, found too one-dimensional a measure if applied alone. In practice, utilising payback time as the only criterion will not take into account the profits during the lifecycle but favours investments with short payback time. Thus, it was targeted that the developed approach and related tool should produce information on both the investment decision’s effects on the time periods for which capitals are invested (payback time) and the anticipated lifecycle profits of individual investment targets. However, economic fluctuation and other key features of the dynamic business environment can have effects on which criteria are emphasised. In capital-intensive industries, production downtime typically generate most of the total costs related to equipment failures. For this reason, the main driver for the implementation of a maintenance investment often comes from the resulted system downtime and related costs, as in this case. However, the more comprehensive list of cost items used in our profitability assessment model is as follows: -

unavailability costs (production downtime)

-

maintenance costs

-

o

failure based maintenance costs

o

preventive maintenance costs

energy consumption

By evaluating the assumed change in the fore-mentioned cost items due to the considered investment, one can calculate the cost effects of the investment. Thus, our practical LCP based calculation model is a comparative analysis tool for making assessments of future costs for two different scenarios where a) no investment is done or b) investment will be done during the next year. In addition, the costs of the implementation of the investment are taken into consideration in the investment calculations. Implementation costs are categorised as follows: -

investment purchase costs

-

installation related unavailability costs

-

other installation related costs and costs generated by the need to modify surrounding assets

The main results our practical investment evaluation tool produces are the payback time of the investment and investment’s effects on the lifecycle costs of the considered target. Qualitative information on the target is also given to support the decision-making: the description of the target, the identified need for the investment, and additional benefits related to the investment are to be described qualitatively. This complements the quantitative analysis which is purely focused on the main drivers of a typical maintenance investment: unavailability, maintenance and energy costs. Thus, the following aspects regarding the potential investment target are analysed: -

environmental and occupational safety

-

target’s capability to answer the future demands (e.g. increase of capacity/speed)

-

improvement of product quality

-

synergy potential in maintenance.

5

Out of the list of potential investment targets, most auspicious candidates are chosen for profitability assessment - based on commonly agreed preliminary criteria with qualitative and quantitative aspects. The number of candidates should at this phase be larger than the number of investment targets typically funded within the considered maintenance investment budget. Profitability assessment phase results in a list of maintenance investment proposals for the next year, with both quantitative and qualitative depictions of profitability and investment benefits as well as other key features having effects on decisionmaking. 5

CONCLUSIONS

In this case study maintenance investment management was developed as a collaborative development of operations models and practices, rather than a pure technical method and tool development task. The objective was to develop the best practical means for investment management rather than theoretically best means. The development was based on wide knowledge on maintenance investment management methodologies. The (theoretically) optimal solutions were modified for local circumstances and requirements of which a newly started cooperation of two companies was an important one. A practically useful and successful process and tool was strived for. How it was succeeded, is still being seen.

6

REFERENCES

1

Ahlmann, HR (2002) From Traditional Practice to the New Understanding: The Significance of Life Cycle Profit Concept in the Management of Industrial Enterprises. IFRIMmmm Conference, Växjö, Sweden, 67 May 2002. 16 s.

2

anon. (2005) Calling a Change in the Outsourcing Market, The Realities for the Worlds Largest Organisations, Deloitte Development LCC.

3

Bailey, W., Masson, R. & Raeside, R. (2002) Outsourcing in Edinburgh and Lothians. European Journal of Purchasing & Supply Management, 8, 83-95.

4

Hendry, J. (1995) Culture, Community and Networks: Hidden Cost of Outsourcing. European Management Journal. 13(2), 193-200.

5

Huai, J. & Cui, N. (2005) Maintenance Outsourcing in Electric Power Industry in China. In: Proceedings of International Conference on Services Systems and Services Management. 13-15 June 2005. IEEE. 2, 1340-1345.

6

Rhee, S. J., Ishii, K. (2003) Using Cost Based FMEA to Enhance Reliability and Serviceability. Advanced Engineering Informatics. 17(34). July October 2003. s. 179188.

7

Sloan, T.W. Shanthikumar, G. (2000) Combined production and maintenance scheduling for a multiple-product, singlemachine production system. Production and operations management. Vol. 9(4), Winter 2000.

8

Tarakci, H., Tang, K. & Teyarachakul, S. (2009) Learning effects on maintenance outsourcing. European Journal of Operational Research. 192, 138-150.

6

Proceedings of the 4th World Congress on Engineering Asset Management Athens, Greece 28 - 30 September 2009

A MODEL FOR MORE ACCURATE MAINTENANCE DECISIONS (MMAMDEC) Basim Al-Najjar a and Renato Ciganovic b a b

Terotechnology, School for Technology and Design, Växjö University, Sweden, [email protected]

Terotechnology, School for Technology and Design, Växjö University, Sweden, [email protected]

It is usual when using CM technology for assessing the state of a component and planning maintenance actions using predetermined levels for warnings and replacements. The replacement of a damaged component is usually done at lower or higher than the predetermined level, which both means losses. This is because the probability of doing replacements just at the predetermined level is negligibly small. The accuracy in the assessment of the condition of a component has big technical and economic impact on the output of the machine, production process and consequently company profitability and competitiveness. The higher the accuracy in assessing the condition of a component yields higher probability of avoiding failures and planning maintenance actions at low costs. In this paper, techniques for assessing the state of a component using both mechanistic and other statistical approaches are considered. This paper also applies Cumulative Sum (CUSUM) Chart for identifying the time of damage initiation and reducing false alarms. Techniques for assessing the probability of failure of a component and its residual life, and predicting the vibration level at the next planned measuring opportunity or planned stoppage are introduced, discussed, computerised and tested. The problem addressed is: How is it possible to increase the accuracy of assessing the condition of a component? The major result achieved is; development of a model for more accurate assessment of the condition of a component/equipment through combining different approaches. The main conclusion that can be drawn is; applying the model, it is possible to enhance the accuracy of assessment of the condition of a component/equipment and consequently maintenance decision since the integrated model provides comprehensive and relevant information in one platform. Key Words: Integrated Approach for Maintenance Decisions, Effective Maintenance Decision, Prediction of Vibration Level, Probability of Failure, Residual Lifetime, Vibration Monitoring, Total Time on Test. 1

INTRODUCTION

Today the companies try to reduce their production costs for gaining a competitive advantage in the market. Maintenance plays a key role in reducing production cost through enhancing availability and extending the life length of producing assets, Wu et al. (2007). Using condition monitoring (CM) technologies, production security and operating safety will increase because the probability of detecting and treating problems will increase Wang (2002), Al-Najjar (2004) and Wu et al (2007). Furthermore it will result in much lower operational costs, Xiaodong et al. (2005). According to White (1994) it also improves the safety for the labour because the probability for a sudden breakdown of a machine will decrease. In other words, lack of maintenance can also have a huge negative impact on the surrounding environment through, for example radioactive radiation, oil leakage, explosions etc. This is why maintenance can reduce safety hazard for the labour and make them feel safer when working with the machines. Furthermore, White (1994) states in many industries the production has increased with around two to ten percent when applying condition-based maintenance (CBM). It is important to keep the machines in good condition, plan maintenance at need and try to avoid failures and disturbances. Without these enhancements, the company may face difficulties in keeping-up and improving the customer service level, production quality, personal safety and company profitability and competitiveness. When using CM technology, it is quite common to use vibration signals analysis for detection of machine faults, Samanta and Al-Balushi (2003). Detection is made by comparing the vibration signals of a machine operating under normal conditions with a machine running under faulty conditions. Noise, randomness and deterioration may cause variation in the vibration level, Al-Najjar (2001) and Samanta and Al-Balushi (2003). Random fluctuating of the vibration measurements may occur according to some uncontrolled factors independent of component’s condition. These can obstruct the detection of component condition, Samanta and Al-Balushi (2003).

7

When using CBM system, there are analytical tools for interpreting the signals and assessing the condition of the component, Al-Najjar (2001 and 2003). If the prediction of the vibration level in the close future utilises previous and current measurements, current and future operating conditions, the predicted value may support and enhance the accuracy of maintenance decisions. But even if it provides reliable information about the condition of a component, we do not certainly know when the component will break, Al-Najjar (1997). Replacing components at unreliable assessment of its condition leads to two types of losses; either losing a big part of its residual life or at failures. 2

CUSUM CHART

Variation in the vibration level can occur due to noise, randomness in the vibration and a wide spectrum of deterioration causes, Al-Najjar (2001) and Samanta and Al-Balushi (2003). The variation in the vibration level may lead to over- (or under-) estimation of the component/equipment condition. The economic losses that arise due to the false signals are significant especially when the downtime cost is high, see Al-Najjar (2003). According to Al-Najjar (1997) the cumulative sum (CUSUM) chart of the condition measurements is a better indicator of a potential failure than the vibration measurement itself when the variation in the vibration level is appreciably big. The graph of CUSUM values shows the behaviour of the cumulative sum of the vibration level with reduced probability of false alarms. The CUSUM chart is obtained by plotting the cumulative sum of deviations from a target value along the monitoring period with respect to a predetermined level, see Al-Najjar (1997). 3

PREDICTION OF THE VIBRATION LEVEL

Effective usage of Vibration-based Maintenance (VBM) within TQMain concept for rotating machines provides the user with a long lead time for start planning cost-effective maintenance actions to avoid failures. It can be achieved through acting at an early stage in order to keep the probability of failure of a components low until a condition-based replacement is done, Al-Najjar (2000). In Al-Najjar (2001) and Al-Najjar and Alsyouf (2003) a model for predicting the vibration level during the close future is developed and tested, respectively. The tool for predicting vibration level utilises previous and current measurements and operating conditions to predict future vibration level. Components' mean effective life can be prolonged if the operating conditions have been assessed properly, Al-Najjar (2000). When assessing the condition of a component, deterministic and probabilistic or combined approaches can used. Deterministic method includes for example issues related to machine function, failures analysis and diagnostics, while probabilistic method includes assessment based on statistical tools, Al-Najjar (2007B). More specifically, it means prediction of the vibration level, assessment of the time to maintenance action and the probability of failure of the component. According to Al-Najjar (2000) damage initiation and development phase represents more than half the total usable life of a component/equipment. The past history of a component, its current status and its operating and environmental conditions during the near future, will affect the probability of failure, Ibid. During normal component state, i.e. when no damage is initiated, CM parameter value, e.g. vibration level, is fluctuating around its mean level xo, Al-Najjar (1997). Denote the phase of mechanical component life when its vibration level approaches the potential failure level (initiation and development of damage) by xp. Replacement level is denoted by the variable xr. When the damage under development has been confirmed, the function of CM parameter value, i.e. x(t) is assumed to be a non-decreasing function. While CM parameter level is usually considered stationary during the interval prior to the initiation of damage, see Herraty (1993) and Al-Najjar and Alsyouf (2003). According to Al-Najjar (1997), it is often difficult to tell for certain that the damage development has begun, especially when the number of CM measurements, e.g. vibration, is very small. The model shown in Eq. (1) is used for predicting the vibration level during the period until the time point for next measurement of a planned stoppage see Al-Najjar (2001). c

Yi +1 = X i + a * Exp(bi * Ti +1 * Zi i ) + Ei

(1)

Yi+1 is a dependent variable representing the predicted value of the CM level at the next planned measuring time. The model expressed in Eq. (1) consists of three independent variables (X, Z and T) and three parameters (a, b and c). Ti+1 is the elapsed time since the damage initiation has been confirmed. The current CM level is labelled by Xi. Zi is the deterioration factor, which in its turn is a function of the current and anticipated load and previous deterioration rate. (Z = dx` * Lf/Lc). The parameter 'a' is the gradient (slope) by which the value of the CM parameter increased since it started to deviate from its normal state xo due to damage initiation until detecting it at potential failure level, xp. Parameters bi & ci are model's non-linear constants. The model error, Ei, is assumed to be identical, independent and normally distributed with zero mean and constant variance, N (0, s). Finally i, for i = 1, 2…n, is number of measuring opportunities after damage initiation.

8

4

ASSESSMENT OF PROBABILITY OF FAILURE AND RESIDUAL LIFETIME

The second technique uses failure and condition-based data for same or similar components using graphical Generalised Total Time on Test (GTTT-plot), Al-Najjar (2003). This technique is developed to asses the probability of failure of a component and its residual lifetime at demand. Assessment of probability of failure and residual lifetime is necessary to effectively enhance the information given to describe the condition of a component, e.g. bearing, at any time or after each vibration measurement. This is especially important for situations when the value of CM parameter, e.g. vibration level, is increasing rapidly or when it is relatively high and there is a risk of faster deterioration during the time until the next measurement, see Al-Najjar (2003). GTTT-plott can be obtained by plotting Ui = (Ti/Tn) = (Ti/n)/ (Tn/n) on y-axis versus ni/n on the x-axis. Here Ui indicates the proportion of the average exhausted life length of the bearings until their vibration levels approached or exceeded x(i) divided by the average time generated by n components until their levels equalled or exceeded x(n), for x(i)<x(n) and i=1,…,n. Consequently the ni/n gives an estimation of probability of failure occurrence, where ni is the number of replacements that have been done until x(i) is exceeded, while n represents the total number of the bearings under consideration and xr is the replacement vibration level, see Al-Najjar (2003). 5

MODEL DEVELOPMENT

It is better that predicting the vibration level in near future should not be done before the damage initiation has been confirmed for avoiding unnecessary work due to false alarms that may arise because of randomness. This is where CUSUM chart is useful due to its ability to identify the moment when damage has initiated, see Al-Najjar (1997). Assessment of the probability of failure of a component and its residual lifetime do not by itself consider the current and future state of the component in operation, which is the reason behind combining it with the tool for prediction of vibration level and CUSUM chart. The probability of failure and residual life help to enhance the information about the component by considering its age in comparison with historical data about same or similar components. CUSUM chart is computerized and the user-interface of CUSUM chart is shown by Fig.2. The CUSUM chart has been developed in Microsoft Excel since it is practical software for creating graphs. Thus, Fig. 2 displays the plot area intended for accumulated sum of the deviations between each vibration measurement and the reference value (k). The x-axis represents the measurement number and y-axis represents the CUSUM value. In this paper we apply CUSUM chart technique in the same way as done in Al-Najjar (1997). Therefore the reference level (k) is chosen in middle between xo and xp. Furthermore, the reference level is controlled against the mean vibration level of measurements and plotting of CUSUM chart is only obtained when the accumulated value of the vibration measurements is equal or less than zero. On the other hand the plotting of CUSUM is stopped as the measurements are declines under the k level because the process is assumed to be in control. Development of PreVib, ProFail and ResLife software prototypes: For easy use of the model and in order to achieve reliable and faster prediction of the vibration level, new software based on Eq. (1) and also called PreVib was developed. Also, ProFail & ResLife was developed for assessing the probability of failure of a component and its residual lifetime. The userinterface of PreVib is divided into two halves, Fig.3: left half representing input data for prediction and right half representing the result in form of a graph. Input data show both database and non-database data in the grey and white boxes, respectively. When a prediction is to be made, the segment (machine), asset (component, e.g. a rolling element bearing) and the location (direction of the vibration measurements) should be specified. Also, we should specify how many measurements in “Limit measurement” that we consider if we focus only on part of the measurements. Otherwise we should use “None” if we consider all the measurements in the prediction. Then, the measurements can be downloaded by pressing “Load” which uploads the data from a MIMOSA database. The “Mean vibration level (xo)”, “Prediction time period” and “Near future load/present load” should all be specified before clicking “Predict”. The desired future “Prediction time period” can be specified in, e.g. seconds, hours, days, weeks, months, etc. When “Predict” is clicked, the value of the predict vibration level and the associated date and plot it on the diagram on the right hand side of the user interface will be determined. By analogy, the same thing can be repeated any time the user need to predict the vibration level, for example after the next measurement of the vibration level. The right half of the user-interface presents a graph will contain two plots, the blue representing the actual vibration level values and the red showing the predicted vibration level values. Y-axis is the vibration level in, e.g. mm/sec and x-axis is the calendar time for measurements and predictions. The graph contains also two dashed lines representing the two different levels of life of a mechanical component i.e. mean vibration level (xo) and potential failure level (xp). These levels are determined and set based on data from identical components and is automatically retrieved when component data is uploaded. Some of the initial measurements belong to the first phase will not be shown in the graph, even if they have been considered when estimating model's constants b and c. The first prediction, as evident, should be done after confirming damage initiation, i.e. the vibration level exceeds or close to xp, because as long it fluctuates around xo it means there is no damage is initiated and thereby there is no need for predicting future level. To be able to predict future vibration level one need at least three measurements. This is why we need some of the measurements that have been considered part of the first phase of the component life. Observe that sometimes, the first two predictions are special cases due to the use of the measurements below the potential failure level to assess model's parameters b and c. The reason for that is; we want to predict future state as soon as we have confirmed the initiation of damage, instead of waiting for at least three measurements exceeded (xp). Waiting could also cause uncertainty in the maintenance decisions. Also, because deterioration process is a stochastic process, vibration level

9

can be changed randomly in time. Therefore, some of the level values may acquire higher level than it should, i.e. exceeds xp, which it may be irrelevant to the severity of the deterioration. For each prediction, all previous measurements are used for estimating model's parameters (b and c) and thus predicted vibration level itself. ProFail & ResLife software module is supplementary to PreVib module for enhancing the accuracy in maintenance decisions, which can be used jointly and independently. As for PreVib, ProFail & ResLife user-interface is also divided into two halves, Fig.4. The left half represents input data required for assessment and in the right half the result is shown in form of a plot (graph). When assessing the probability of failure and residual lifetime of a component, it demands to specify the “Segment” (machine) and “Asset” (component, e.g. a rolling element bearing in this case) and “Assessment time point”, i.e. the time at which the assessment should be done followed by clicking assess-button. Subsequently the program uploads the table of lifetime data from MIMOSA database and presents the graph on the right half of the user-interface. The y-axis in the graph shows the proportion of the average exhausted lifetime and the x-axis shows the probability of failure for the analysed component. “Probability of Failure” of the component in question and its “Residual lifetime” are then assessed and displayed in the left hand side of the user-interface. The values can be shown by using the cross-air pointer in the graph. If a new assessment is performed after the current assessment time point, then consequently the cross-air pointer will move forward on the curve. Each point on the graph represents each and one of the component lifetimes in relation to each other. The basic idea of developing a model for more accurate maintenance decisions (MMAMDec) through integrating CUSUM chart, PreVib, ProFail and ResLife based on; applying CUSUM chart reduces the probability of false alarms and confirms the initiation of damage, which makes the prediction of the vibration in the near future more effective. While, assessment of the probability of failure of a component and its residual lifetime increases the information underlying maintenance decision through improving the description of the condition of the component from additional dimensions. Fig.1. shows how the mentioned modules are integrated and the model working steps. Vibration measurements should firstly be analysed by using CUSUM chart. The vibration measurements can be exported to Microsoft Excel file from the original vibration measurement database. When damage initiation is confirmed by using CUSUM chart, i.e. the vibration level has approached xp, and then PreVib is applied. The software modules PreVib and ProFail&ResLife are MIMOSA compatible. MMAMDec working steps, see also Fig.1: 1. 2. 3. 4.

Use CUSUM chart to identify when xp is approached. Then, Predict the vibration level in the near future using PreVib module. From this step you can either go to step number 3 in order to enhance the information underlying maintenance decision through ProFail & ResLife module or you can jump to step 4. Assess the probability of failure of the component and its residual lifetime through using ProFail & ResLife module for the same dates as the predicted vibration levels or other date (after demand). It helps to enhance the information required to confirm the results in step 2, reduce its significance or reject it. Plan the required maintenance action by means of the information in previous steps

Fig.1. MMAMDec for integrating CUSUM, PreVib and ProFail&ResLife

10

6

MODEL TEST

The model is tested using real industrial data provided by CNC machine manufacturer Goratu from a specific machine/motor. The company also provided the reference levels for the machine type, i.e. mean vibration, potential failure and replacement levels. All the data needed for testing the models were retrieved from a MIMOSA database. Furthermore Goratu provided lifetime data, installation times and replacement/removal times for previous motors and the one under analysis. For the test of the model the same type of data that the three integrated modules use individually are needed. For the CUSUM chart we need vibration measurements and corresponding date of measurement. But we also need the different levels i.e. xo and xp in order to decide the reference level k. PreVib uses same type of data and level which are retrieved from MIMOSA database. ProFail & ResLife uses life data. First step in the model test is to identify when potential failure level xp is approached. CUSUM chart is appropriate tool for detecting systematic change from a prescribed level. The changes in the mean of the vibration measurements from the reference value can be plotted. If the mean of vibration measurements is equal to the reference value then cumulative sum fluctuates about zero. Variations in vibration level will make the CUSUM chart slope either to increase or decline. An upward slope in the CUSUM chart will indicate that accumulated sum of the deviations, between the vibration measurement and the reference value, has increased and vice versa for a downward slope. With help of CUSUM chart it is also possible to trace back the time point when the change occurred, see Fig. 2.

Fig.2. User-interface of CUSUM chart in Microsoft Excel using real data

For example by eye-balling the CUSUM chart in Fig. 2 one can see when the CUSUM chart slope changes its direction by observing the corresponding value on x-axis (Xi), which corresponds to the different measurements. Then it is easy to go back to the CUSUM chart file and trace the date for Xth measurement and start investigating the reason behind the change. CUSUM can also help us to find out if it is a false alarm or to confirm damage initiation. The characteristic of the CUSUM chart is that it cannot discover sudden deviation easily, this because usually at least two to three measurements are needed to indicate a deviation. From Fig. 2 we can see that decision interval is estimated to approximately 0,080 mm/sec, which means that passing this interval can be considered as an initiation of a potential failure. Furthermore we can see that the vibration measurement that passed the decision interval (potential failure initiation) was around measurement X50. But, prior X50 we have registered around 20 measurements have been near to or passed the potential failure level. It is by using the CUSUM chart that these false alarms have been discovered and thus reduced. We can also see that damage development is constantly increasing after the damage initiation point (X50), see Fig. 2. Once the damage initiation has been confirmed through CUSUM chart, we can start applying the PreVib module for predicting future vibration level and ProFail Module for assessing probability of failure and residual lifetime see Fig. 3 and 4.

11

Fig.3. User inter-face of the software program PreVib using real data The software module PreVib was used to predict vibration level for variable calendar time. The application of PreVib is shown in Fig. 3. One can see from Fig. 3 that predicted values are rather accurate since they are close to the real measurements. By predicting the vibration level the maintenance engineer will be able to model the optimum time for the replacement of the component.

Fig.4. User-interface of the software program ProFail and ResLife using real data Notice that the software module ProFail & ResLife is showing that probability of failure is 100% for the same dates as the predicted vibration values, see Fig. 4. This is because ProFail & ResLife makes an assessment based historical data for similar

12

or same components, which in this particular case have been shorter than the current, see life distributions on the left half of Fig.4. Also more historical data would be sufficient in order to consider the information provided by ProFail & ResLife. In the last step of MMAMDec, Fig. 1, the operator or maintenance engineer can plan the required maintenance action by means of the information in given from CUSUM chart, PreVib and possibly ProFail & ResLife. 7

RESULTS, DISCUSSIONS AND CONCLUSIONS

In this article we have presented an approach of how to integrate the three different tools with the purpose to increase the accuracy of assessing the condition of a component. By using CUSUM chart, the probability of false alarms can be reduced. After the damage initiation we started to apply the tool for predicting the future vibration levels so that an optimum time for replacement can be identified. At the same time we applied the tool for assessing the probability of failure and residual lifetime in order to enhance the information available. The main conclusion from this paper is that the developed model has the ability to integrate different type of data in one platform. The model has also clear and systemized way of application. The model creates a link between these modules so that the decision maker can for example for same date predict the vibration level and assess the failure probability and residual lifetime of the component. In order to make more accurate maintenance decisions; we should also sort out false alarms due to randomness in the vibration signals. In the final step the maintenance engineer can then plan the maintenance actions required with better accuracy. 8

REFERENCES

1

Al-Najjar, B. (1997) Condition-based maintenance: Selection and improvement of a cost-effective vibration-based policy in rolling element bearings. Doctoral thesis, ISSN 0280-722X, ISRN LUTMDN/TMIO—1006—SE, ISBN 91-628-2545X, Lund

2

University, Inst. of Industrial Engineering, Sweden.

3

Al-Najjar, B. (2000) Accuracy, effectiveness and improvement of Vibration-based Maintenance in Paper Mills; Case Studies. Journal of Sound and Vibration, 229(2), 389-410.

4

Al-Najjar, B. (2001) Prediction of the vibration level when monitoring rolling element bearings in paper mill machines. International Journal of COMADEM 4(2), 19-27.

5

Al-Najjar, B. (2003) Total Time on Test, TTT-plots for condition monitoring of rolling element bearings in paper mills. International Journal of COMADEM 6(2), 27-32.

6

Al-Najjar, Basim (2007A) The Lack of Maintenance and not Maintenance which Costs: A Model to Describe and Quantify the Impact of Vibration-based Maintenance on Company's Business. International Journal of Production Economics IJPPM 55(8).

7

Al-Najjar, Basim (2007B) Establishing and running a condition-based maintenance policy; Applied example of vibrationbased maintenance. WCEAM2007, 106-115, 12-14 June Harrogate, UK

8

Bergman, B. (1977) Some graphical methods for maintenance planing. Annual Reliability and Symposium, 467-471.

9

Bergman, B.and Klefsjö, B. (1995) Quality from customer needs to customer satisfaction. Studentl., Lund, Sweden.

10

Herraty, A.G. (1993) Bearing vibration-Failures and diagnosis. Mining Technoloy, 51-53.

11

Jardine, A.K.S. and Joseph, T. and Banjevic, D. (1999) Optimizing condition-based maintenance decisions for equipment subject to vibration monitoring, Journal of Quality in Maintenance Engineering, 5(3), 192-202.

12

Lin, C. and Tseng, H. (2005) A neural network application for reliability modelling and condition-based predictive maintenance. International Journal of Advanced Manufacturing Technology, 25(1), 174-179.

13

SAMANTA B. AND AL-BALUSHI, K.R. (2003) ARTIFICIAL NEURAL NETWORK BASED FAULT DIAGNOSTICS OF ROLLING ELEMENT BEARINGS USING TIME-DOMAIN FEATURES, Mechanical Systems and Signal Processing, 17(2), 317-328.

14

Xiaodong, Z. and Xu, R. and Chiman, K. and Liang, S.Y. and Qiulin, X. and Haynes, L. (2005) An integrated approach to bearing fault diagnostics and prognostics, American Control Conference, 2005. Proceedings of the 2005, 2750-2755.

15

Wang, W. (2002) A model to predict the residual life of rolling element bearings given monitored condition information to date. IMA Journal of Management Mathematics, 13(1), 3-16.

16

Wang, W. and Zhang, W. (2007) An asset residual life prediction model based on expert judgments, European Journal of Operational Research, 2, 496-505.

13

Maintainability

17

White, Glenn (1996) Maskinvibration, Vibrationsteori och principer för tillståndskontroll, Landskrona: Diatek vibrationsteknik

18

Wu, S. and Gebraeel, N. and Lawley, M. A. and Yih, Y. (2007) A Neural Network Integrated Decision Support System for Condition-Based Optimal Predictive Maintenance Policy, IEEE Transactions on Systems, Man, and Cybernetics - Part A: Systems and Humans, 37(2), 226-236.

Acknowledgement We would like to thank EU for the support that we had in EU-IP DYNAMITE, where this paper is part of the work done within DYNAMITE.

14

Proceedings of the 4th World Congress on Engineering Asset Management Athens, Greece 28 - 30 September 2009

THE HOUSE OF MAINTENANCE - IDENTIFYING THE POTENTIAL FOR IMPROVEMENT IN INTERNAL MAINTENANCE ORGANISATIONS BY MEANS OF A CAPABILITY MATURITY MODEL Prof. Dr. Guenther Schuh, Bert Lorenz, Cord-Philipp Winter, Dr. Gerhard Gudergan Research Institute for Operations Management at RWTH Aachen University (FIR), Pontdriesch 14/16, 52062 Aachen, Germany

In order to guarantee an efficient and effective employment of production equipment, it is essential to identify any possible potential for improving performance, not only in the production process, but also in supporting areas such as maintenance. One of the major tasks in increasing maintenance performance consists of systematically identifying the company’s most significant weaknesses in maintenance organisation and thus being able to implement improvements there where they are most needed. But how is a company to tackle this important task? To answer this question, this paper describes an assessment and improvement approach, based on a capability maturity model (CMM). By means of this approach, the status-quo of a maintenance organisation can be analysed and its individual improvement opportunities identified. Key Words: Maintenance Management, Capability Maturity Model, Maintenance Assessment, Improvement Program, House of Maintenance Framework, Maintenance Performance 1

INTRODUCTION

Ever changing market conditions, shorter product life cycles and an increase of competitive pressure create the need for a more effective and efficient deployment of production facilities. Maintenance is a key factor for both efficiency and effectiveness in production [1]. Being an internal service provider, maintenance is essential in ensuring a high degree of equipment availability, as well as meeting quality requirements and thus being able to meet customer’s needs without wasting resources. Due to its increasing significance concerning production, maintenance has become a major cause for cost increases. Nevertheless, it is essential to identify maintenance as a strategic factor for success and not a mere cost centre. Also, maintenance departments must continuously be improved, in order to enhance their performance. Improvement, however, requires the identification and understanding of unutilized potentials within maintenance. Target-oriented improvements are only possible, if the object and organisation respectively, and their current situation are measurable, and hence, rateable. It is, however, particularly difficult to develop a suitable method for objectively measuring and accessing a complex structure like a maintenance organisation. A further requisite for an improvement process is the fact that both starting point (in this case, the status quo of maintenance) and the target strived at, are known beforehand. The maturity-level-system theory offers a promising approach for solving such problems. With this in mind, the Research Institute for Operations Management at RWTH Aachen University (FIR) has attempted a research effort, aimed at projecting the maturity-level-system onto the management of the internal maintenance. For this purpose, the CMM was first assigned to the internal maintenance as well as to all other relevant supporting departments. Finally, the assessment framework (the House of Maintenance of FIR) was validated in practice, using business cases. The results of the research project are shown below for a practical case. As part of its research, the FIR has developed a diagnostic instrument, the “IH-Check” (“Maintenance-Check”), for recognizing strengths and weaknesses in the maintenance departments of mainly small and medium-sized enterprises (SME), and for recognizing potentials for improvement and methods for utilising the same.

15

Modern concepts of management, like Total Productive Maintenance (TPM), Reliability Centred Maintenance (RCM) or Risk Based Inspection (RBI), can help improve maintenance performance [2, 3, 4, and 5]. Although these and other similar concepts have been successfully used by large enterprises, their application to SMEs is, however, limited. This is validated by results of expert surveys, conducted among heads of maintenance departments [6].

„According to your oppinion, which concepts / methods are suitible for improving the maintenance performance of SMEs?“

Reliability oriented maintenance Rist Based Maintenance (RBM) Total Productive Maintenance (TPM) Outsourcing 0%

20%

40%

60%

suitable

suitable to a limited extent

not suitable at all

unknown

80%

100%

hardly suitable

N = 56 Source: FIR Expert Study „Trends and DevelopmentPerspectives in Maintenance“ 2004

Figure 1 Acceptance of existing maintenance concepts in operational service It is essentially the first step of such improvement processes, namely the realistic evaluation of one’s own strengths and weaknesses, which causes considerable problems within enterprises. Experts believe that an absence of an integrated approach is one of the causes of these problems (see fig. 1). Many existing management concepts in maintenance consist of single isolated solutions which are not part of an integrated improvement process. Additional challenges accompanying the application of modern concepts of maintenance management are described in figure 2. „According to your opinion, which are the main problems preventing the application of Maintenance concepts in SMEs?“ Systematic support in identifying potentials for improvement A systematic and integrated view of maintenance Internal analysis (Estimation of potentials in maintenance) Insufficient involvement of employees in the process of improvement Insufficient consideration of resources (human resources as well as financial resources) of SMEs Setting realistic goals concerning maintenance Consideration of qualitative assessment criteria 0%

N = 56

20%

40%

60%

80%

Source: FIR Expert Study „Trends and DevelopmentPerspectives in Maintenance“ 2004

Figure 2 Causes of the lack of acceptance of existing maintenance concepts in SMEs

16

In order to support enterprises in this situation, FIR has developed “IH-Check”, a diagnostic instrument which helps to systematically reveal organisational weaknesses in maintenance.

Customer

Partnerships

Materials management

Maintenance controlling

Maintenance organisation Maintenance object

Information and knowledge management

Maintenance policy and strategy

Maintenance staff Figure 3 House of Maintenance The assessment is based on a framework called the “House of Maintenance” (see fig. 3), which consists of nine fields of action, describing the elements of a typical maintenance organisation on a generic level. These fields are defined by an individual set of nine assessment criteria, each of which contains a set of specific levels of maturity. The levels of maturity are developed according to the CMM [7]. The CMM is a structure of elements that describes certain aspects of maturity in an organization. It aids in defining and understanding processes within an organization and is based on a five-level process maturity continuum.

Analysis of each individual field of action Evaluation of maturity level of

-

-

the specific field of action Field of action „Maintenance Controlling“

Calculation of individual maturity level for each criteria

Criteria


Tasks




Data collection




Identification of key figures

Do you collect maintenance specific data?

Answer-possibilities




Use of key figures







Key figure comparisons







Cost accounting




Budget




Calculation of profitability




Indirect costs

maturity level

No data is collected for maintenance Little data is collected. Figures are only sporadic and cannot be verified. No evaluation of figures.




Evaluations are carried out regularly.

consequences is possible. All the key figures can be determined without any uncertainty.

17

4

The collected data is complete without redundancy. It is evaluated in such a way that a forecast on the

Figure 4 Elements of the maintenance assessment

3

A large amount of data is collected, certain data often more than once. For collection a standard is specified.




2

Data is collected, but collection is not completely standardised. Evaluations are only carried out irregularly.




1

5

Following the assessment, a company’s individual maturity profile regarding maintenance management which determines the company’s potential for improvement in maintenance, is developed. In combination with a prioritisation for identifying the crucial fields of action – e.g. by means of a pair-wise comparison – specific measures can be developed to exploit the company’s full potential in maintenance. This paper describes the House of Maintenance’s structure on the one hand, enlarging on its different fields of action and referring to dedicated maturity levels, while discussing the procedure of maintenance assessment and the identification of specific measures for improvement in maintenance organization within a practical example on the other hand. Further details under discussion are the House of Maintenance’s different fields of action as well as companies’ specific levels of maturity.

2

FRAMEWORK: THE “HOUSE OF MAINTENANCE”

The assessment is based on a framework called the “House of Maintenance”, which consists of nine fields of action originating from our vast experience in maintenance organisation. The House of Maintenance is oriented towards current models concerning excellence and maintenance management [2, 8, 9, and 10]. The fields of action describe the elements of a typical maintenance organisation on a generic level. Their significance has been validated in the study among maintenance experts mentioned above. The nine fields of action represent all persons and sections/departments with a significant impact on the Overall Equipment Effectiveness (OEE) – the most important key figure for measuring maintenance performance. Maintenance staff provides the basis for all maintenance activities and is thus a key to a company’s performance whilst both production uptime and production quality depend on the maintenance department’s performance. Requirements for uptime and quality within production are determined by the production department, which therefore has a major impact on the configuration of all other fields of action. The latter have to be managed accordingly.

3

•

While developing the House of Maintenance, we considered state-of-the-art developments in maintenance as well as issues specific to SMEs:

•

The House of Maintenance has a practical and easy-to-communicate visualisation, readily understood by both management and shop floor workers.

•

While maintenance is recognised as an internal service provider, it is essential to evaluate maintenance performance from the customer’s viewpoint – the customer being the production department. It is, therefore, necessary to include aspects such as service quality and customer orientation in the House of Maintenance structure

•

A growing degree of linking between maintenance and other internal organisational units, as well as external service providers, has increased the importance of interface-management and IT-support.

•

Maintenance policy is supposed to aim at satisfying the needs of production departments in order to ensure efficient and effective production.

DEVELOPING STEPS FOR CAPABILITY MATURITY

As is the case in quality management and software engineering, the evaluation of individual criteria is based on a CMM. Such models can be employed to systemise and structure varying processes. They initiate a stable long term process of optimisation, by pointing out the course for future developments. The progress of development is quantified and can be checked at regular intervals. The IH-Check uses a total of five maturity levels, based on the typical CMM levels [7]. The adaption of the content to inhouse maintenance, however, represents a totally new development. Each benchmark is structured according to the House of Maintenance and is based on the five parameters (characteristics) of the CMM, namely, “improvisation”, “orientation”, “commitment”, “implementation” and “optimisation” (see fig. 5 top part). The first level of CMM characterises a chaotic condition, in which improvements in maintenance are introduced sporadically. No consistent understanding of a fully integrated maintenance management exists in the company. The second level is characterised by an awareness of the importance of internal maintenance as an internal service provider and its contribution to a company’s value and operational performance. Very frequently, this development is initiated and carried out by individual employees.

18

Steps Optimisation “There will always • Integrated maintenance management be room for • Continuous improvement improvement!“

CIP 5

Implementation • Usage • Measurement • Company-wide implementation

4

Commitment • Company-wide rules • Introduction • Organisation

3

Orientation • Target planning • Target formulation • Breakup period

2

1

Improvisation • To become conscious • To get real • Chaos

„Not bad at all!“

„We are Getting better!“

„Let´s get started“

„There is much to do!“

Time

Call for action:




Need for documentation of adopted measures Need to analyse adopted measures and initiation of further activities Identification of priority measures most essential to maintenance performance

Orientation:






Often, only a small number of employees realise the importance of Maintenance organisation Improvement is possible in several cases Employees begin to realise the importance of maintenance performance

Figure 5 CMM for an effective facility-oriented maintenance (modified version based on [7]) and development of the CMM (example: “Orientation”) In the third level, all measures undertaken to improve the performance of maintenance are documented, evaluated, standardised and laid down in the form of operating instructions. The fourth level may be considered to be the stage in which most important maintenance processes are understood. At this stage, a high level has already been attained. Further improvements can only occur in small steps, demanding large inputs of effort. The fifth and highest level denotes a stage of integrated maintenance management which performs with high efficiency and effectiveness. Up to this moment, the maintenance has been reorganised using effectively coordinated steps. All employees, including external service providers, have, meanwhile, adopted these aims and regulations and are prepared to accept and to improve them continuously. The five levels of the CMM may be considered to be a relatively stable condition of maintenance organisation, based on actual and durable activities and processes. This implies that a maintenance organisation cannot be changed overnight and that no steps or levels may be skipped. The individual steps are consecutive and support each other interactively, so that any single level can be attained only after the requirements of the previous one are fulfilled.

4

PROCEDURE

Assessments are handled in multiple interactively working steps. The first step is to select those fields of action which are relevant and are able to significantly improve the performance of internal maintenance. These are then compared and evaluated, thus depicting to the individual situation of the company. The above is achieved by using a pair wise comparison to give a weighting to the nine fields of action as described in the House of Maintenance [11]. In the following step, the decisive fields of action are evaluated using the consigned criteria. The criteria are depicted as open ended questions based on five standard statements. These statements are interactively related and represent the five levels of the applied CMM. The use of previously formulated statements within the assessment generally prevents any subjective evaluations (see fig. 2).

19

As an intensive evaluation process is very costly and consumes a vast amount of a company’s resources, which, especially in SMEs, are not always available in the desired quantities, it was necessary to curtail the number of evaluation criteria for each field of action to a maximum of nine, giving a total of 81 evaluation criteria for the diagnosis, requiring four to six hours for the accomplishment of the assessment. The assessment takes place as a workshop using a questionnaire based approach and including collective discussions to guarantee a highly objective evaluation process. The questionnaire, as seen in fig. 2, is used for the identification of individual maturity levels for each evaluation criteria. An evaluation of maturity levels for all relevant fields of action follows implicitly in accordance with the House of Maintenance. In order to consider all relevant views of maintenance as an internal maintenance provider, the employees and head of the maintenance department, together with employees of production, controlling and purchasing should be included in the analysis. Hence, an integrated view of maintenance as part of the enterprise as a whole, is guaranteed on the one hand. On the other hand, an evaluation including representatives from all divisions within a company ensures a high commitment of the employees involved. At the end of the questionnaire based survey, the whole information is consolidated and the individual maturity levels are calculated. For further analysis, results can be depicted for every field of action and its related criteria. This maturity evaluation is depicted in a radar chart (see fig. 6).

25%

1 Tasks 1

50%

100%

2 Data collection

80%

9

2

25%

40%

4 Use of key figures

20%

8

25%

3 Identification of key figures

60%

3

0%

5 Key figure comparison

0%

50%

6 Cost accounting 7

4

6

5

50%

7 Budget 8 Calculation of profitability

0%

0%

9 Indirect costs 0%

100%

Figure 6 Evaluation of results in the form of radar charts (example: field of action “Maintenance Controlling”) In addition, the maturity levels of all nine fields of action are condensed into a single diagram, giving a maturity profile for the company’s maintenance as a whole (see fig. 7). On the basis of this profile, it is possible to judge which fields of action should be developed primarily, and which level should be strived for. The aggregated result determined by the IH-Check also delivers a collective maturity score in the form of a percentage (0 to 100 %). It shows the stage that the maintenance organisation of the company has reached, regarding a maintenance organisation aiming at excellent equipment effectiveness. This key performance index (KPI) can be effectively applied for internal marketing purposes within the organisation. Using the information derived in the weighting process within the House of Maintenance and in the creation of the enterprise’s individual maturity profile, “IH-Check” then prioritises the fields of action, considering both the importance of each field of action to the individual company and its current level of maturity. The prioritisation’s results are depicted in a prioritisation matrix shown in figure 11. Within the matrix, fields of action with a high importance as well as a low level of maturity can clearly be identified as most crucial concerning improvement. Based on these insights, measures can be developed to improve maintenance performance and thereby increase the enterprise’s operational performance.

20

44%

100%

1 Customer 1

80%

9

2

100%

75%

3 Maintenance organisation

60%

72%

4 Information and knowledge management

40% 20%

8

33%

2 Maintenance policy and strategy

100%

3

0%

25%

5 Maintenance controlling

100%

100%

100%

59%

100%

6 Maintenance object 7

4

67%

7 Materials management 53%

6

5

100%

100%

8 Partnerships 50%

100%

9 Maintenance staff 0%

100%

AS- IS TO- BE scenario

Figure 7 Maturity profile (regarding a company’s maintenance as a whole) (example)

5

PRACTICAL APPLICATION

Practical experience has, in a variety of cases, shown that “IH-Check” supports organisations to conduct systematic, object oriented and substantiated diagnosis of their internal maintenance. This applies to a first self-assessment, which a project team can conduct in a short time and is sufficient to lay the basis for a detailed discussion of problems within individual areas. The evaluation of criteria within a team has proved to be beneficial. Firstly, this encourages an exchange among the different lobbies and ultimately promotes an improved mutual understanding regarding the different points of view existing within the organisation. On the other hand, it is just this team evaluation which is absolutely necessary to ensure the objectivity necessary for a realistic determination of the strengths and weaknesses existing at the different locations. It occurs often enough, that the highly distorted views which employees have, regarding their own maintenance performance have to be put into proper perspective. In addition, it is already in the course of these discussions, that initial valuable contributions for potential improvement are suggested. The presentation of results as radar charts has likewise proved its merit. The results of the assessment, independent of the depth and scope of the survey, can be clearly and easily communicated to and interpreted by both maintenance employees and management.

6

CASE STUDY EXAMPLE

With the objective of optimising the maintenance management of a company operating in the gas industry and the FIR have conducted a project for assessing the potential of the maintenance management and deriving the optimal maintenance strategy. An analysis of the actual situation of the company’s maintenance management was conducted using the “IH-Check”. Both the current relevance of the respective fields of action and the level of maturity within maintenance management were assessed. After defining the relevant fields of action as described in the House of Maintenance, a pair wise comparison was used for a weighting of these fields of action, specifically adapting the House of Maintenance to the company’s individual situation (see fig. 9). “Customer”, “Maintenance Object” and “Information and Knowledge-Management” were identified as fields of action with a high relevance for the company’s maintenance management, which is reflected in their weighting scores of 14.9%, 14.0% and 12.2%.

21

Materials management 2 Maintenance Partnerships staff 2 1 11 22 Materials management Maintenance staff 0 0 Partnerships 12 Maintenance staff Partnerships

Maintenance staff

0

201 2

222 1

002 1

11

21

2

20

12

0

1 Information 2 2 Maintenance organisation management and knowledge 0 Partnerships 10 1 2 2 Maintenance organisation management 01 0 Materials 1 management Customer Maintenance policy and 01

02

0

2

2

1

1

1

1

0

1

2

2

1

1

1

1

0

2

0

1

1

1 Customer strategy Maintenance 1

Maintenance policy and strategy9,6 % Maintenance organisation 14,0 %

10,2 %

9,6 %

7,2 %

Partnerships

Reset

Reset Maintenance organisation policy and 11 Maintenance 0 0 Reset Maintenance controlling strategy Maintenance policy and object/asset 02 12 Maintenance 0 0 1 Maintenance controlling strategy 22 12 Maintenance 21 Maintenance 1 2 policy Customer organisation and Maintenance object/asset 2 0 11 12 strategy 22 21 Maintenance 1 2 Customer organisation 0 Maintenance object/asset Information 12 02 Information 21 Maintenance 2 Maintenance object/asset and knowledge knowledge 0Maintenance 02 0and 1 0controlling 1 11 1 staff Information and knowledge management 12 management object/asset Maintenance 2111 0Maintenance 102 101staff 00 Maintenance 00 Maintenance 1 0controlling policy and management 10 00 20 Maintenance 0Materials 0 12 00 strategy 0 Information 0object/asset Materials management and knowledge 0 00 management Partnerships 10 000 010 020 Maintenance object/asset 12 Information 01 0 0 management andMaterials knowledge 1 Materials management management 22 12 11 Maintenance 0 11 01 staff 0 1staff Maintenance controlling Partnerships management 1 01 Maintenance 22 002 021 Maintenance staff 210 Partnerships 12 01 Maintenance 0 1staff Maintenance controlling 2 02 211 Materials Partnerships 2Customer 20 20 organisation 1management 1 Maintenance object/asset 0 21 Maintenance 02 201 Materials organisation 20 0 2Customer 20 22 Maintenance 1 management 1 Maintenance object/asset 2 2 2and 0 Partnerships 1 1 Maintenance staff Information knowledge 2 01 Maintenance 1 1 controlling 2 1 Maintenance staff 20 2Information 2 20 and Partnerships knowledge 2 2management 02 1 10 Materials 2 1 management Maintenance staff 10 management 0 0 Customer 0 1 Maintenance organisation 11 0 Materials 1 1 2 1 Partnerships management 2 controlling 100 120 00 10 Customer 11 Maintenance 2 1 Partnerships Information and knowledge 22 12 2 1 Maintenance controlling

201 1

2

12,2 % Sort

Materials management

Reset Maintenance staff

strategy Maintenance policy and 0 strategy Maintenance organisation 1 2 Customer

Maintenance object

staff

Sort

Reset

10,2 %

Maintenance controlling

Sort

Reset

Information and know ledge management

Sort

Reset

Maintenance policy and

Partnerships Maintenance

Maintenance staff

Sort

2

Maintenance staff

2

Customer

Sort

Materials management Maintenance staff Partnerships

Structural fields Main 0 Structural 0 1fields0 0 1 0 0 tena Main Customer 0 0 1 fields 0 0 1 0 Structural Maintenance policy and tena Main 2 1 2 1 Customer 1Structural 21 2fields 22 12 21 strategy Maintenance policy and tena Main 2 1 Customer 02 01 11 02 02 strategy Maintenance policy and Maintenance organisation 2 1 2 Main 1 2 2 1 1 tena 21 Customer 00 01 10 02 strategy Maintenance and Maintenance organisation policy 2 1 2 Main Information and knowledgeCustomer 2 tena 11 121 212 212 1 0 policy 0 0 0 strategy Maintenance and 0 tena Maintenance organisation 00 0 20 Main management Information and knowledgeCustomer 2 11 220 212 Maintenance 1 policy 0 and 0 0 tena 0 Main 0 0 strategy Maintenance organisation 2 1 1 2 2 1 0 1 management Information knowledge Customer Maintenance controllingand strategy 2 1 1 2 2 211 Maintenance 01 policy 22 and 21 Maintenance organisation 20 1 02 tena 21 Main management Customer Information andMaintenance Maintenance controlling strategy 2knowledge 1 1 2 and Maintenance organisation 11 02policy 00 0 02 tena 20 Maintenance object/asset 2 1 Maintenance 1 2 002 1 management Information and knowledge Maintenance controlling 01 12 strategy Maintenance organisation 11 policy 00 and 00 0 01 Information2and strategy knowledge Maintenance object/asset 1 1 2 1 2 02 management Maintenance controlling2 1 organisation 01 22 0 1 Maintenance 2 2 Materials management 1 0 0 2 0 0 2 0 management Information and knowledge Maintenance object/asset 1 2 2 Maintenance controlling2 11 112 12 212 Maintenance organisation Materials management 1 0 0 2 0 0 2 management Information Maintenance Maintenance object/asset controlling 2 and 1 0knowledge 1 12 111 20 Partnerships 2 0 Information 0 0 0 0 Materials management 00 02 management Maintenance Maintenance object/asset controlling 20and knowledge 11 111 120 121 Partnerships 2 0 0 0 0 management Materials Maintenance management object/asset 10 02controlling1 10 22 20 10 Maintenance Maintenance staff 2 1 1 1 2 2 2 Partnerships 11 02 Materials Maintenance management object/asset 10 controlling 01 002 221 1102 Maintenance Maintenance staff 2 1 1 1 2 2 Partnerships Materials management 21 02object/asset 00 02 00 10 Maintenance Maintenance staff 0 0 2 1 2 21 11 Partnerships Materials management 2 0 0 0 2 2 Maintenance object/asset 200 Maintenance Partnerships staff 2 1 11 02 01 11

Sort

Reset Maintenance object/asset Partnerships Maintenance staff Materials management

Structural fields Structural fields Customer

14,9 %

Sort

Reset

Maintenance controlling Materials management Partnerships Maintenance object/asset

Customer

Structural fields

Weighting of fields of action

Sort

Information and knowledge Maintenance organisation Maintenance controlling Maintenance object/asset Materials Maintenance management staff Partnerships management

Structural fields

Information and knowledge Maintenance object/asset Materials management Maintenance controllingstaff management Partnerships Maintenance

Structural fields

Maintenance policy and Customer strategy Maintenance policy and Customer Maintenance organisation strategy Maintenance policyand andknowledge Information Customer Maintenance organisation strategy management Informationpolicy and knowledge Maintenance and Maintenance organisation Customer Maintenance controlling strategy management Information and knowledge Maintenance policy and Customer Maintenance controlling Maintenance organisation Maintenance object/asset management strategy Maintenance policy and Customer Information and knowledge Maintenance Maintenance organisation controlling Maintenance object/asset strategy Materials management management Maintenance policy and Information and knowledge Maintenance organisation Customer Maintenance object/asset controlling Materials management strategy Maintenance Partnerships management Information and knowledge Maintenance policy and Customer Maintenance organisation Materials management Maintenance controlling Partnerships Maintenance object/asset Maintenance staff management strategy Maintenance policyand andknowledge Information Maintenance Maintenance organisation controlling strategy Partnerships Maintenance object/asset Maintenance staff management Materials management

Pair wise comparison

Maintenance staff12,0 %

policy and

Maintenance policy and strategy strategy

Figure 8 Weighting of individual fields of action (example) Using the CMM approach, an evaluation of specific maturity levels followed for each field of action, revealing the company’s status quo regarding maintenance. The evaluation took place in a questionnaire based workshop including collective discussions as described above (see fig. 9). The hereby identified status quo in maintenance is depicted in figures 6 and 7.

Field of action „maintenance controlling“ Criteria


Tasks




Data collection




Identification of key figures




Use of key figures




Key figure comparisons




Cost accounting




Budget




Calculation of profitability




Indirect costs

Do you collect maintenance specific data?

Answer-possibilities

maturity level




No data is collected for maintenance




Little data is collected. Figures are only sporadic and cannot be verified. No evaluation of figures.




Data is collected, but collection is not completely standardised. Evaluations are only carried out irregularly.




0%

2

25%

x 3

50%

4

75%

5

100%

A large amount of data is collected, certain data often more than once. For collection a standard is specified. Evaluations are carried out regularly.




1

The collected data is complete without redundancy. It is evaluated in such a way that a forecast on the consequences is possible. All the key figures can be determined without any uncertainty.

Figure 9 Evaluation of specific maturity levels using a CMM based approach Considering the weighting within the House of Maintenance as well as gap between the status quo and the desired level of maturity for each field of action, the company’s current state was transformed into a prioritisation matrix (see fig. 10). The Prioritisation matrix was used to identify those fields of action, which combine a high weighting and a high gap – representing the potential for improvement within the respective field of action – and thus have a high priority concerning the improvement of maintenance performance.

22

By assessing the weighting of relevance of each field of action against the gap between the actual and achievable level of maturity (potential for improvement) the following fields of action were identified as most crucial in the company’s maintenance management: 1. Maintenance Controlling and Performance Management (MC) 2. Maintenance Policy and Strategy (MPS) 3. Customer of Maintenance (C) Within these fields of action lay the biggest opportunities for improvement. These improvements were essential for building up a maintenance organisation that could provide the maximum value and profit to the whole company and its customers.

C:

Customer

MPS:

Maintenance policy and strategy

MO:

Maintenance organisation

IKM:

Information and knowledge management

MC:

Maintenance controlling

MOA: Maintenance objects and assets MM:

Material management

P:

Partnership

MS:

Maintenance staff

Figure 10 Prioritisation Matrix (example) Based on the results of the maintenance assessment, the FIR suggested an improvement of maintenance controlling (as well as other measures, which will not be discussed in this paper) by introducing a performance management system (PMS) based on balanced scorecards [12]. In detail, the FIR recommended a two-level PMS for maintenance consisting of strategic and execution level and five perspectives which provide a consistent monitoring of the company’s maintenance, focussing on Processes, Customer, Finance, Staff and External Services. After its implementation, the PMS would work as a tool for creating transparency and provide the basis for the continuous measurement of the actual maintenance performance in sense of effectiveness and efficiency. The PMS system was successfully implemented and increased the maintenance department’s performance within the company.

7

CONCLUSION

The “IH-Check” assessment is a powerful tool, clearly identifying shortcomings in maintenance performance as well as potentials for improvement and thus enabling the introduction of specific measures ensuring a continuous improvement. A periodic application of this instrument ensures a successful derivation and implementation measures aiming at an improvement of maintenance performance and thereby ensuring effectiveness and efficiency in production. Based on the analysis of the current condition of a company’s maintenance, “IH-Check” helps in efficiently introducing measures for improvement. The sequence of reorganisation projects in maintenance, till now more a result of reactions to

23

external causes, is converted to a process of systematic improvements. Employees are no longer confronted with the introduction of new management concepts, but are now directly involved in measures specifically derived for the company. The problem of measuring and assessing internal maintenance was solved with the introduction of the House of Maintenance framework and the associated assessment tool “IH-Check”. The actual status of a company’s maintenance can be measured easily, using the status quo analysis described. Applying the maturity-level model, it is also possible to identify salient fields of action, thus being able to derive appropriate measures. Using the “IH-Check”, it is possible to convert a mere reorganisation of internal maintenance to a continuous improvement process. The resulting progressive improvements in maintenance contribute in improving the operational performance of the whole enterprise.

8

REFERENCES

1

Schuh G., Berbner J., Lorenz B., Franzkoch B. & Winter C.-P., (2008) Reliability leads to a better performance – results of an international survey in continuous process industries. Proceedings of the 3rd World Congress on Engineering Asset Management and Intelligent Maintenance Systems (WCEAM-IMS 2008), 1366-1374, Springer-Verlag London Ltd. Beijing.

2

Shirose K. (2007) Total Productive Maintenance. New Implementation Program in Fabrication and Assembly Industries. Seventh edition. JIPM-Solutions. Tokyo.

3

Moubray J. (1997) Reliability-Centered Maintenance. Second edition. Industrial Press Inc. New York.

4

McKone K. E., Schroeder R. G. Cua K. O. (2001) The impact of total productive maintenance practices on Manufacturing performance. Journal of Operations Management, Vol. 19, 39-58.

5

Khan F. I., Haddara M., Krishnasamy L. (2008) A new Methodology for Risk-Based Availability Analysis. IEEE Transactions on Reliability, Vol. 57, No. 1, 103-112.

6

Schick E. (2004) unpublished Expert Study “Trends and Development-Perspectives in Maintenance” conducted by the Research Institute for Operations Management at RWTH Aachen University (FIR).

7

Mark C. Paulk, Charles V. Weber, Bill Curtis & Mary Beth Chrissis (1994). The Capability Model: Guidelines for improving the software process (SEI Series in Software Engineering). Addison-Wesley Professional. Boston.

8

Marquez A. C. (2007) The Maintenance Management Framework: Models and Methods for Complex Systems Maintenance (Springer Series in Reliability Engineering). Springer-Verlag. Berlin.

9

Moore R. (2004) Making Common Sence – common practice: Models for Manufacturing Excellence. Third edition. Butterworth-Heinemann. Oxford.

10

Al Najar B., Asyouf I. (2000) Improving effectiveness of manufacturing systems using total quality maintenance. Integrated Manufacturing Systems, Vol. 11, No. 4, 267-276.

11

Carrizosa E., Messine F. (2007) An exact global optimization method for deriving weights from pairwise comparison matrices. Journal of Global Optimization, Vol. 38, No. 2, 237-247.

12

Lorenz B., Winter C. (2008) Identification of Optimal Maintenance Strategy Mixes for Small and Medium Enterprises (SME). Euromaintenance Papers. Bemas. Brussels.

24

Proceedings of the 4th World Congress on Engineering Asset Management Athens, Greece 28 - 30 September 2009

INTEGRATED STRATEGIC ASSET PERFORMANCE ASSESSMENT Aditya Paridaa and Uday Kumara a

Division of Operation and Maintenance Engineering, Luleå University of Technology, Luleå, Sweden

Asset performance assessment forms an integral part of a business process for heavy and capital intensive industry to ensure performance assurance. Managing the asset performance is critical for long term economic and business viability. Assessing the asset performance is a complex issue as it involves multiple inputs and outputs; and various stakeholders’ dynamic requirements. Lack of integration among various stakeholders and their changing requirements in strategic asset performance assessment is still a problem for the companies. It is a challenge to integrate a whole organization, where free flow and transparency of information is possible; and each process is linked to integrate to achieve the company’s business goals. In this paper, various issues associated with an integrated strategic asset performance assessment are discussed. Key Words: Asset performance assessment (APA), asset Performance Indicators (API), employee involvement, Maintenance process. 1

INTRODUCTION

Global dynamic and competitive business scenario with technological development and changes during last couple of decades, besides the prevailing economic slowdown have dominated the global industrial scenario demanding an effective, safe and reliable integrated strategic engineering asset management. Further, outsourcing, separation of asset owners and asset managers, and complex accountability for the asset management, make the assessment of asset performance and its continuous control and evaluation more critical. Organizations operating today are facing several kinds of challenges, like; highly dynamic business environments, complicated intellectual work at all levels of the company, efficient use of information and communication technologies (ICT), and a fast pace of information and knowledge renewal [1]. Thus, under this technological advancement and global competition scenario, asset owners and managers are striving for monitoring, assessing and following up the asset performance. Health monitoring of the strategic engineering asset is an important issue and challenge for the management as it provides information on plant and system health status to achieve higher productivity with minimum cost and safety with high reliability. The advancement in computer and information technology has a significant impact on the process of asset management information system for finding the asset health status facilitating timely decision making. Advancement of sensor technologies, automated controls and data telemetry have made possible new and innovative methods in asset health monitoring. Rapid growth in the networking systems, especially through the internet has overcome the barriers of distance, allowing a real time data transfer to occur easily from different locations [2]. Corporate strategy of an organization describes how it intends achieve its mission and objectives, and to create value for its stakeholders like; the shareholders, customers, employees, society, regulating authorities, suppliers and alliance partners. Without a comprehensive description of strategy, executives cannot easily communicate the strategy among themselves or to their employees [3]. Therefore, it is essential that the corporate strategy and objectives of an organization is converted to the specific objectives integrating different hierarchical levels of the organization. Under challenges of increasingly technological changing environment, implementing an appropriate performance assessment (PA) system in an organization becomes a necessity. This is because, without integrated assessment of performance, it is difficult to manage and verify the results of the desired objectives of an organization. Maintenance of an asset is considered as an important support function for the management and perceived as “It can be planned and controlled” and “It creates additional value” [4]. To know the amount of additional value created, the assessment of asset performance needs to be integrated into the business process. Assessing the asset performance of an organisation is a complex issue due to multiple inputs and out put which are influenced by stakeholders and other sub-processes. Many times, contribution of maintenance on asset performance can only be assessed in terms of the losses incurred due to lack of

25

maintenance activities. Coupled with lack of maintenance activities have resulted in disaster and accident of extensive losses and changes in legal environment, the asset managers are likely to be charged with “corporate killing” for the future actions or omissions of maintenance efforts [5]. These societal responsibilities to prevent loss of life and property, besides high asset maintenance cost are compelling the management to undertake asset PA as part of the business measurement system. An asset PA system ensures that all operational activities are aligned to the organization’s corporate strategies and objectives in a balanced manner. The organization has to satisfy and meet the requirements of both the external and internal stakeholders need and identify the performance indicators (PIs) from a integrated and balanced point of view. The purpose of this paper is to discuss various issues associated with integrated strategic asset performance assessment. The structure of the paper is as follows: after providing an introduction to the topic at section 1, strategic issues of engineering asset are discussed at section 2. Section 3 deals with the integrated issues in engineering asset performance assessment. A discussion and conclusion is provided at section 4.

2

STRATEGIC ISSUES IN ENGINEERING ASSET

Strategy; is concerned with the long-term directions of the firm, deals with the overall plan for deploying resources that the firm possesses, entails the willingness to make trade-offs to choose between different directions and between different ways of deploying resources, achieving unique positioning vis-à-vis competitors, and sustainable competitive advantage over rivals and ensuring lasting profitability [6]. An organization’s strategy indicates how it intends to create value for its stakeholders, like the shareholders, customers, employees, the society, etc though effective use of its asset. For maximum impact, the measurement system should focus on organization’s strategy, how it expects to create future and sustainable value [4]. No two organizations develop and follow the strategy in the same way, because some follow strategy from financial perspectives for revenue and growth; some by the services or products focusing on their customers, others from marketing or quality perspectives; and some others from a human resource perspectives. Observing different organizations and critical analysis of authors, strategic policy exist around shareholders value, customer satisfaction, process management, quality, innovation, human resources and information technology amongst others. Engineering asset strategy is formulated from the corporate strategy considering the integrated and whole life cycle of the asset. An integrated approach is essential as an asset performance management is associated with various stakeholders with their conflicting needs with multiple inputs and outputs. From asset performance objectives, two set of activities are undertaken. One set of activity develops the key performance indicators for bench marking performance with similar industry and the other set formulate the activity plan, implementation, measurement and review as given at Figure 1. As shown in the figure, asset performance objectives are formulated as per stakeholders’ requirements and organizations integrated capability and capacity. In order to achieve the asset performance objectives, critical success factors are identified from which key result areas of activities are identified. From the key result areas, key performance indicators (KPIs) are developed for measuring and assessing the asset performance. On the other set of activity, activity plans are made based on which the implementations are carried out. After implementation, measurement and assessment of the asset performance is undertaken, so that feedback and reviewing action can be undertaken to validate the asset performance objectives.

Corporate strategy Feedback & review Asset performance objectives

PM

Asset strategy

Implementation Activity plan

Critical success factors Key result area Key performance indicators

Figure 1 Strategic asset performance measurement process (Adapted from [7])

26

Companies are using the scorecards as a strategic management system to manage their strategy over their long run and using the measurement focus to accomplish critical management processes [8] like; 1. Clarify and translate vision and strategy 2. Communicate and link strategic objectives and measures 3. Plan, set targets, and align strategic initiatives 4. Enhance strategic feedback and learning In an asset management strategy, various industry forces play important roles and are required to be considered for analysis. For new entrants, it is the entry barriers of experience and culture; for suppliers, there may be many service providers; for alternate products, it is better system and processes; for customer, it is trust and good relationship; and for the industry, it is the competitors. Importance of strategic aspects for engineering asset cannot be overlooked, especially under the present business scenario context. Some of the examples of asset performance objectives could be to achieve higher OEE level, zero defects (zero quality complaints), and nil accidents. The KPIs translate aggregate measures from shop floor levels to the strategic level. The real challenge lies in measuring all the KPIs, as some of the KPIs are difficult to measure being intangible in nature and cannot be quantified. Organizations need a framework to align their performance measurement system with the corporate strategic goals of a company by setting objectives and defining key performance at each level [9]. The performance measurement PM which forms part of the asset performance measurement system needs to be aligned with the organizational strategy [10]. The PIs are required to be considered from the perspective of the multi-hierarchical levels of the organization. As per [11], maintenance management needs to be carried out in both strategic and operational contexts and the organizational structure is generally structured into three levels. The three hierarchical levels considered by most of the firms are; the strategic or top management level, the tactical or middle management level, and the functional/operational level [12]. Two major requirements of a successful corporate strategy relevant for the performance measurement are namely: 1. 2.

Cascading down the objectives from strategic to shop floor level Aggregation of performance measurements from shop floor to strategic level.

2.1 Cascading down the objectives from strategic to shop floor level. The strategic objectives are formulated based on the requirements of the stakeholders, both internal and external. The plant capacity and resources are considered from long-term objectives and matched. These corporate objectives are cascaded down the hierarchical level of the organization though the tactical level which considers the tactical issues such as financial and nonfinancial aspects both from the effectiveness and the efficiency point of view. The bottom level is represented by the functional personnel and includes the shop floor engineers and operators. The corporate or business objective at the strategic level is communicated down through the levels of the organization and translated into the objective measures in a language and meaning appropriate for the tactical or functional level. This cascading down of strategy forms part of the goal deployment of the organization.

2.2 Aggregation of performance measurements from shop floor to strategic level. The performance at the shop floor level is measured and aggregated through the hierarchical levels of the organization to evaluate the achievement of the corporate objectives. The adoption of fair processes is the key to successful alignment of these goals. It helps to harness the energy and creativity of committed managers and employees to drive the desired organizational transformations [13]. This aggregation leads to empowerment of employees in the organization.

3

INTEGRATED ISSUES IN ENGINEERING ASSET PERFORMANCE ASSESSMENT

Observing different organizations and critical analysis of authors, strategic policy exist around shareholders value, customer satisfaction, process management, quality, innovation, human resources and information technology amongst others. [14] Thompson (1997) listed eight clusters of organizational competencies which are linked to the integrated and strategy contents competencies, needs to stay aware and strategic change competencies. The eight clusters are: 1. 2. 3. 4. 5. 6.

strategic awareness abilities stakeholders satisfaction abilities competitive strategic abilities strategic implementation and change abilities competency in quality and customer care functional competencies

27

7. 8.

ability to avoid failure and crises Ability to manage ethically and with social responsibility.

Therefore all successful organizations have to be aware, formulate the winning integrated strategy, implement and manage it in a dynamic and competitive business environment. Companies are using the PM scorecards as a strategic management system to manage their strategy over their long run and using the measurement focus to accomplish critical management processes [8] (Kaplan and Norton, 1996): • • • •

Clarify and translate vision and strategy Communicate and link strategic objectives and measures Plan, set targets, and align strategic initiatives Enhance strategic feedback and learning

The integrated issues in engineering asset PA are discussed as under [15] (Parida, 2006): 1.

2.

3.

4.

5. 6.

Stake holder’s requirement. The stakeholders’ external needs are to be assessed and responded with matching asset and resource requirements with planning and that of internal stakeholders’ capability and capacity, which formulate the corporate objectives and strategy and translate into targets and goals at the operational level or converting a subjective vision into objective goals. While considering the external stakeholders needs, the prevailing and futuristic business scenarios are looked into besides the competitors. Internal stakeholders’ need from employees, management & organizational culture perspectives are also considered, besides the asset and other resources capacity and capabilities. Organizational issues. The asset PM system needs to be aligned and form integral part of the corporate strategy. This will require commitments from the top management and all employees to be aware of the asset PM system through effective communication and training, so that they all speak same language and are fully involved. The involvement of the employees in the asset PM system at every stage, like the planning, implementation, monitoring and control, and at each hierarchical level can ensure the success of achieving the asset performance and business strategies. Besides, all functional processes and area like, logistics, IT, human resources, marketing and finance need to be integrated with engineering assets. Engineering asset requirements. From the stakeholders’ need, the demand analysis of engineering asset is perceived and designed. After concept development, validation and engineering asset specifications are worked out. Besides, competitive product, cost of maintenance, risk management, correct product design, asset configuration and integration are considered from strategic and organizational perspective. From operation and maintenance, the engineering asset may be outsourced partially or entirely. How to measure? It is essential to select the right PIs for measuring asset PA from an integrated whole life cycle perspective for benchmarking besides, collecting the relevant data and analysis for appropriate decision making. The asset PM reports developed after the data analysis are used for subsequent preventive and/or predictive decisions. The asset PM needs to be holistic, integrated and balanced [12] Sustainability. Sustainability development is the development that is consistent while contributing for a better quality of life for the stakeholders. This concept integrates and balances the social, economic, and environmental factors amongst others. Linking Strategy with integrated asset performance assessing criteria. The linkage between integrated Enterprise Asset Management (EAM) measuring criteria with condition monitoring, IT and hierarchical level for decision making at different hierarchical level is given at Figure 2. This figure describes the linkage between the external and internal stakeholders needs and considers the concept of integrated enterprise asset management from the different hierarchical needs, while linking the performance measurement and assessment from engineering asset’s operational level to decision making at strategic level. The external effectiveness is high lighted by stakeholders need like return on investment and customer satisfaction. The internal effectiveness is high lighted through the desired organizational performance reflected by optimized and integrated resources utilization for EAM. For example; availability and performance speed of the equipment and the machineries forms part of the internal effectiveness or back end process. Quality is the most important aspect, which is not only related to the products quality of the back end process, but also with customer satisfaction of external effectiveness. From external stakeholders, the quantity of annual production level is decided, considering the customer’s requirements, return on investment and internal plant capacity and plant availability etc. From internal stakeholder’s, the organization considers department’s integration, employee requirements, organizational climate and skill enhancement. After formulation of the asset PA system, the multicriteria PIs are placed under the multi hierarchical levels of the organization.

28

External Stakeholders’s Needs

Organization’s Vision & Objectives

Company’s Internal Needs

Decision

making STRATE GIC LEVEL

Checked with PIs

Organization’s

Production/Op erational Strategy & Policy

Business & Marketing Strategies

Production Planning Scheduling & Control

Maintenance Planning, Scheduling & Control

TACTICAL/ MANAGERIAL LEVEL

OPERATIONAL LEVEL

& KPIs

Information

Data

Performance Indicators For: Productivity Process, Cost, Environment,

Employee Satisfaction Growth & Innovation, Etc.

IT system

Embedded sensors

Enterprise asset management (EAM)

Figure 2. Linkage between integrated Enterprise Asset Management’s (EAM) measuring criteria with condition monitoring, IT and hierarchical levels for decision making

4

DISCUSSION AND CONCLUSION

An asset cannot be managed with out considering the integrated strategic issues for an appropriate PA system. This is because of the various stakeholders conflicting interest needs, associated multiple inputs and outputs including the tangible and intangible gains from the asset. For engineering asset PA, strategic issues are essential to be considered. Under prevailing dynamic business scenario, asset PA is extensively used by the business units and industries to assess the progress against the set goals and objectives in a quantifiable way for its effectiveness and efficiency. An integrated asset PA provides the required information to the management for effective decision making. Research results demonstrate that companies using integrated balanced performance systems perform better than those who do not manage measurements [16]. In this paper the complexities of asset PM are considered and discussed from corporate strategy perspective of the organization. Since no two organizations are exactly similar, the asset PM framework and PIs formulation from corporate strategy need to be specific to that organization. The concepts and strategic issues are discussed for a successful engineering asset PM.

29

5

REFERENCES

1

Antti Lönnqvist. (2004) Business Performance Measurement http://www.pmteam.tut.fi/julkaisut/HK.pdf, visited on 2004 Aug 22.

for

Knowledge-Intensive

Organizations,

2 Toran, F., Ramirez, D., Casan, S., E. Navarro, A and Pelegri, J., (2000) Instrumentation and Measurement Technology, Vol. 2 (Ed. IEEE) IMTC, pp. 652-656 3 Kaplan, R. S and Norton, D. P, (2004) Strategy maps, converting intangible assets intangible outcomes, Harvard Business School Press, USA 4

Liyanage, J. P. and Kumar, U. (2003) Towards a value-based view on operations and maintenance performance management, Journal of Quality in Maintenance Engineering,

5

Mather, D (2005), An introduction to maintenance scorecard, The plant maintenance News letter edition 52, dated 13 April 2005.

6

Jelassi, T. and Enders, A. (2005) Strategies for e-business, Prentice Hall, Essex, London

7

Parida, A, Ahren, T and Kumar, U. (2003) Integrating Maintenance Performance with Corporate Balanced Scorecard, COMADEM 2003, Proceedings of the 16th International Congress, 27-29 August 2003, Växjö, Sweden, pp. 53-59. 8

Kaplan, R. S and Norton, D. P (1996) The Balanced Scorecard: Translating Strategy into Action, Harvard Business School Press, pp. 322.

9 Kutucuoglu, K. Y; Hamali, J; Irani, J; Sharp, J. M (2001) A framework for managing maintenance using performance measurement systems, International Journal of Operation and Production Management, Vol. 21, No. ½, pp. 173-194. 10 Eccles, R. G. (1991) The performance measurement manifesto, Harvard Business Review, January-February, pp. 131-137. 11 Murthy, D. N. P, Atrens, A. and Eccleston, J. A. (2002) Strategic maintenance management, Journal of Quality in Maintenance Engineering, Vol. 8, No. 4, pp. 287-305. 12

Parida, A. and Chattopadhyay, G. (2007) Development of Multi-Criteria Hierarchical framework for Maintenance Performance Measurement (MPM). Journal of Quality in Maintenance Engineering, Vol. 13, No. 3, pp 241-258

13 Tsang, A. H. C. (1998) A strategic approach to managing maintenance performance, Journal of Quality in Maintenance Engineering, Vol.4, No. 2, pp.87-94 Vol. 9, pp. 333-350. 14

Thompson, J. L. (1997) Lead with Vision: manage the strategic challenge, International Thompson Business Press, London

15 Parida, A (2006) Development of Multi-criteria Hierarchical Framework for Maintenance Performance Measurement: Concepts, Issues and Challenges, Doctoral thesis, Luleå University of Technology, Sweden, http://epubl.ltu.se/14021544/2006/37/index.html 16 Lingle, J. H. and Schiemann, W. A. (1996) From balanced scorecard to strategy gauge: Is measurement worth it?” Management Review, March, pp. 56-62.

30

Proceedings of the 4th World Congress on Engineering Asset Management Athens, Greece 28 - 30 September 2009

ASSESSING THE SUBJECTIVE ADDED VALUE OF VALUE NETS: WHICH NETWORK STRATEGIES ARE REALLY WIN-WIN? Tony Rosqvista, Toni Ahonenb, Ville Ojanenc, Arto Marttinend a

VTT Technical Research Centre of Finland, PO Box 1000, FI-02044 VTT, Espoo

b

VTT Technical Research Centre of Finland, PO Box 1300, FI-33101 Tampere, Finland

c

Lappeenranta University of Technology, PO Box 20, FI-53851 Lappeenranta, Finland d

Metso Automation Inc, PO Box 310, FI-00811 Helsinki, Finland

As manufacturing companies increasingly focus on their core business, the interest in the utilisation of external services provided by system suppliers and service companies increases. Currently an increasing number of services are purchased from service supply networks. Furthermore, globalisation, complexity of technological innovations and demand for integrated solutions also create need for networking and collaboration. Establishing or improving the performance of the networked service providers, the value net, is a long-term effort, requiring the build up of trust between the partners. The necessary condition of moving from a subcontractor relationship to a strategic network or partnership is the sharing of the view of joint gains in a prospective value net. How do we then evaluate the added value of moving to a new partnership? What network strategies provide the win-win network solution? This paper is a tentative effort in answering these questions based on Decision Analysis. Keywords: value net, strategic network, network orchestrator, win-win strategy, collaborative maintenance network 1

INTRODUCTION

A current trend is to outsource operations that do not belong to the core competence of a company. The main rationale for this development is the assumption that by subcontracting, the company can buy certain services cheaper and receive them better managed and implemented, thus providing more added value compared to keeping the same functions in-house. Furthermore, embedded technology in the assets, ever-increasing demand for production efficiency, and dynamic end customer requirements increase the need for a variety of services. The main assumption here is that there exists, or will rapidly emerge, a competitive market of services that is able to provide collaborative services cost-effectively. Of course, there is always a tendency in the area of manufacturing and servicing, that some services may be based on know-how that is restricted to only some service providers, resulting in the creation of an oligopolistic service market which is not cost-effective (market effect of scarcity power). The strive to get competitive edge on a special area will naturally lead to this type of oligopolistic service market, but on the other hand companies must avoid to seize all the customer value of their related service. This is characteristic in networked environments. where the market players aim at strategic partnering, forming a value net, which can be characterised as in Fig. 1, showing the ‘strategic’ differences between the network partners and the role of a network orchestrator.

31

Figure 1. Illustration of the requirements for trust, information transparency and the roles of the value net partners. As the figure shows service providers have different strategic roles in the value net: the principal service providers belonging to the core of the network whereas other service providers lie at the fringe of the value net with decreasing strategic significance. Such a governance structure is typically worked out by the network orchestrator or network leader who has the closest relationship with the customer. From customer’s perspective, there has been a clear need to actually decrease the number of closest partners in order to ease the management. This favours the one stop shops that can integrate solutions and services and act as the main partner in the customer’s direction. The network leader establishes the values and culture of the network, developing its guiding principles (e.g. centralisation vs. decentralisation, incentive systems, accountability rules - all these issues entangled with their own tensions and trade-offs) while utilising the best practices from the network itself. In contrast to rigid control systems used to manage production units, the network orchestrator relies not just on rewards, but also upon a combination of empowerment and trust, as well as training and certification, to manage a network that it does not own. Finally, orchestrators have a different way of creating value: value in the traditional firm comes from specialisation, refining skills in specific areas, protecting trade secrets, and keeping out rivals and even partners, whereas value nets create value by integration, bridging borders, leveraging intellectual property across the network. In other words, the social rationale of a value net is simply: i) doing ‘more’ than the organisation knows, and ii) knowing ‘more’ than the organisation does. Of course, this has implications in managerial areas such as contracting / strategic partnering, relationship management (culture, trust), knowledge management (social networks, IT systems), and change management (strategy, leadership, ‘integrator’) – all looking at organisational boundaries from their own perspectives: legal, physical, work/activity and knowledge. The managerial challenges of strategic networks or partnership are best illustrated by the results from a survey made in Finland between 2007-08 on challenges that Finnish industrial service providers meet in offering a client extended services in managing the technical and economic life time of fleet assets [1]. In the survey covering 5 industrial partners (4 vendors, 1 client) and 11 interviewees, success factors for a service network where identified. The following key elements are crucial for building trust between the network partners:

32

-

Pricing in relation to the added value created - How transparent should the price and added value determinants of the services be?

-

Service offering (coupling between service bundling and pricing) - What kind of bundling of product/services will satisfy the client’s need of getting everything from ‘one hatch’?

-

Information management (open access to reliability performance of equipment and maintenance performance and plans) - To what extent should the information systems be open, or even shared, based on a common platform?

-

Mixed cultures (balance between service oriented and product manufacturing cultures) – How to reconcile effectiveness of production to flexibility of services?

-

Intermediate operator / system integrator (ability to connect product and services from separate vendors) - ‘Is an intermediate operator or system integrator’ needed – in the beginning of the new service development process only, or over the life time of the physical assets serviced?

-

Knowledge management (scope of access to each others data, IPR, business models, etc.) - What kind of knowledgerelated asymmetries [2] do we need to worry about?

Good management of the key elements above can be viewed as a necessary condition of moving towards a true value net in the particular industrial area considered in the survey. It is obvious that this transition takes a lot of time and commitment of the managers of the companies in question. The remainder of the paper is organised as follows: In section 1 the different roles in a value net are defined and some demarcations are made for presenting a framework supporting value judgments on alternative service networks, presented in section 2. Section 3 presents a value measurement framework based on value-tree analysis that is a key feature in the generic assessment process presented in section 4. The paper is concluded by a discussion of directions for future research in section 5.

2

ROLES IN A VALUE NET

REGULATOR

OWNER

SERVICE PROVIDERS PHYSICAL ASSET

END-USER

SYSTEM SUPPLIER

OTHER STAKEHOLDERS

OPERATOR

In principle, the network leader has several governance choices for a value net servicing a physical asset. The choices are reflected in the management of the different roles, accountabilities and functions conducted in collaboration. The roles and functions should be defined with particular attention to customer needs which typically derive from the asset owner’s needs together with end-users needs. This is to ensure that the established network processes are truly value adding for the customer. In practice, network and partnership choices have to be made based on current relationships and roles adopted. Basic roles tied together by the physical asset are illustrated in Fig. 2.

Figure 2. The basic question of the network orchestrator is: Which network strategies provide the highest joint gains in the value net, and at the same time meet possible constraints (e.g. regulatory)?

To narrow down the decision context of developing network strategies we will make following demarcations:

33

-

the owner of the physical asset is also the user and customer in the value net

-

one of the suppliers or service providers is also the network orchestrator (see Fig. 1)

-

the ‘value’ of running and maintaining the assets is also assessed by ‘other stakeholders’ (e.g. environmentalists) and the regulator, but these are usually not considered as the principal partners in the value net

-

the added value provided by the value net is basically determined by end-user, but the ultimate rationale of the value net is to provide ‘value’ for all partners i.e. a win-win or a joint gains outcome

-

the service agents are not all ‘equal’ but have different strategic positions in the value net (Fig. 1), and accordingly different network strategies

Formulation of the strategic objectives of the value net aims at the maximisation of customer received value and thus the analysis of customer objectives lays the foundation for the derivation of the network objectives, together with the strategic objectives set by the individual service providers. In other words, the network strategy should be based on customer’s strategic objectives. For instance in a situation where customer’s objectives are highly influenced by the dynamics in business environment, network strategy must include aspects of these dynamics as well. Or, if the customer wants the (physical) assets to have a certain performance (availability, reliability, maintainability, etc), the value net has to manage its processes in a way that the corresponding performance levels are met. Direct business related value of networking for the partners can be e.g. increasing business opportunities and profit, coping with the challenges resulted from a dynamic market, benefiting from network level reputation and reaching economy of scale in the business. In the following, we discuss the value measurement further.

3

VALUE MEASUREMENT

From the point of view of the individual partner, his/her perceived ‘value’ can be assessed using the standard Balance Scorecard approach [3-5]. The BSC perspectives are fourfold and for each of them, strategic objectives, goals and indicators can be defined according to the managerial plans. For instance, plant and maintenance objectives, and related Key Performance Indicators, are discussed in Rosqvist et al. [6]. The Learning & Growth Perspective This perspective includes employee training and corporate cultural attitudes related to both individual and corporate self-improvement. In a value net, learning from each other and sharing knowledge can offer a competitive edge for the partners. The emphasis is therefore in increased understanding of openness, transparency requirements and construction of mutual trust, and promotion of systematised feedback and interaction mechanisms. Indicators can be developed to support managing these issues. Such indicators are usually interpreted as leading indicators, i.e. indicators that signal future outcome of customer satisfaction and financial performance. The Business Process or Internal Perspective This perspective refers to internal business processes. The managerial areas are the operations management, customer management, innovation management and the regulatory&social issues management. Proper indicators allow the managers to know how well their business is running, and whether its products and services conform to customer requirements. Again such indicators are usually interpreted as leading indicators. The above managerial areas may to large extent be jointly managed by partnering. It is thus important for a value net to identify key managerial areas where joint gains can be achieved by reallocating managerial tasks and responsibilities within the network. In particular, information sharing principles need to be addressed: what is openly accessible to all partners, what is restricted to certain partners, what information is entered by whom, what ICT is used and who maintains it, what happens with IPR, etc. The Customer Perspective Recent management philosophy has shown an increasing realisation of the importance of customer focus and customer satisfaction in any business: if customers are not satisfied, they will eventually find other suppliers that will meet their needs. The mutual trust between the customer and the network is crucial. As customer value is something perceived by customers rather than objectively determined by the supplier, the significance of understanding and management of customer knowledge has to be emphasised. The main responsibility of the network orchestrator is to read the customer’s ‘signals’ and ‘transmit’ them properly up to the furthermost partner in the network (see Fig.1). The signals can relate to many attributes such as price, quality, service availability and selection, but also trust and branding and overall functionality of the network. In essence, the value propositions given by the value net need to be achieved, maintained and monitored.

34

The Financial Perspective The financial perspective relates to immediate economic determinants and economic results. Currently, financial metrics have been criticised for their emphasis on short-term performance by moving into quarterly financial reporting. For each partner in a value net, the financial performance determines the success of the network in the eyes of the partners. How this affects the decision to continue as a partner or move somewhere else depends on the position and the role of the partner in the value net: the furthermost partners are expected to base their partnering decisions more on short term than long term financial performance. It is also expected that the financial performance is one key driver of changing the network structure: if revenues and risks are perceived to be distributed in an unfair way by some partners, then the cohesion of the network is clearly threatened. Kaplan and Norton, augmented their performance measurement BSC system with a strategy planning tool – the strategy map – which depicts the cause-effect relationships between the four standard perspectives or objectives in the BSC system [7]. A key insight is that factors in the ‘Learning&Growth’, ‘Business Process’ and the ‘Customer Perspectives’ can be interpreted as leading indicators for the financial performance and competitiveness with their respective lagging indicators [8,9]. It has to be remembered that the developments by Kaplan and Norton are connected to the strategy development of a single corporation. The strategy development of a value net, can, basically, utilise the strategy map idea. Any prospective network strategy can be formulated in terms of managerial objectives that the partners share and jointly try to achieve. The network partners may value a network strategy by assessing its impact on determinants of customer value, and even further on financial determinants. In principle, the added value related to a prospective network strategy may be subjectively assessed, by each partner, by comparing it to the existing network strategy. This is illustrated in Fig. 3 which is an adaptation of the strategy map by KaplanNorton [7]. The strategy map illustrates the structure of a shared network strategy in order for the value assessment to verify that a prospective network strategy is a win-win strategy. It is important to note that a network strategy that yields added value to the customer may for the other network partners produce negative added value due to complexities introduced in contracting and knowledge management, inducing extra transaction costs. In principle, the value assessment framework includes all the strategic elements relevant in the survey referred to in the Introduction, providing a direction for further application-specific refinement with respect to the concerned managerial areas. Some of these aspects are discussed next. The role of a network orchestrator in a service supply network has been emphasised in our survey as, for instance, capabilities for structured strategic planning are often lacking particularly in smaller service companies. In the case, where an network orchestrator will take full responsibility network strategy development, smaller companies may compensate this lack by adopting dynamic working methods of good quality in the implementation of the strategy, making them an important partner in the service provision. Also with respect to information management, our survey has identified strategically important differences between typical players in a service supply network, especially with respect to transparency requirements as indicated in Fig.1. Regardless of these differences, the internal managerial perspective to information management requires integrated solutions to align the combination of strategies. Based on the survey, the surrounding business environment and the related dynamics create the most fundamental needs for partnership development with effects on pricing, service offering, as well as, leadership (culture) and knowledge management. From the customer point-of-view, one of the most important, but at the same time most difficult aspects to control during the contract period, seems to be the network’s capability to continuously develop the services found strategically important. Openness and transparency are found important when finding common ground for assessing and capturing opportunities for the added value of improved services and/or business processes in the network. In the next section, a value assessment process is describe that operationalises the strategy map – based added value model in Fig. 3.

35

Partnership Added Value (dependent on network strategy)

Added Value of service providers

Network Orchestrator Added Value

Customer Added Value (user and/or owner of physical asset)

Financial determinants

improve cost structure

improve asset utilisation

expand revenue opportunities

enhance customer value

Customer value determinants

price

quality

availability

selection

functionality

Product/service attributes

flexibility

brand

Relationship

Internal processes of the partners, where some functions are the property of the network and defined in the network strategy.

causal relationship

defining relationship

Each partner valuates the performances in the various managerial areas differently.

Current network strategy (zero reference)

partnering

Value adding alternative network strategy

?

Figure 3. A strategy map linking the internal managerial perspective (objectives) with the customer and financial perspectives (objectives) for the assessment of added value of a prospective network, formulated in terms of shared managerial objectives and goals that are expected to lead to added value for the customer and the partners, i.e. a win-win situation. The added value assessment is subjective and comparative in nature. (The reader is referred to the work of Kaplan and Norton on strategy maps).

4

THE ASSESSMENT PROCESS OUTLINE

To be able to use the added value model in the previous section, following key issues need to be addressed, mainly by the network orchestrator: -

what managerial elements are incorporated in the current network strategy, and what changes could provide added value?

-

how are the impacts of the prospective network strategy assessed?

-

how are uncertainties (risks) incorporated in the assessment?

36

Figure 4 shows two distinct activities that are needed: the Network Strategy Formulation and the Added Value Assessment.

Figure 4. Activities to support network strategy development: the outcome is an action plan that outlines how the current partnership should be changed to create an improved win-win network strategy. Network Strategy Formulation is expected to be a sensitive process led by the network orchestrator with numerous mutual discussions with the existing and potential partners on wanted or possible changes in roles, accountabilities, etc. with respect to the current network strategy. The sense of trust in the leadership and orchestration is crucial. Basically, the outcome of the activity could be in the format of Table 1 which shows how the current and the prospective network strategies are formulated. The structuring is based on the BSC perspectives. In addition, risk management issues can be included in the form of real options that can be executed conditional on the occurrence of random events. For instance, if there is a sudden price increase for a certain raw material or component, or a radical breakdown of one partner’s production, there are options for the affected partner to switch the material/component to another, or an option for the network to be temporarily supplied by some other company outside the network, respectively. Such real options should be identified and agreed upon in the network strategy. Table 1 Network strategy formulation – generic template Management area

Current strategy

Prospective strategy

financial customer business process

What is in place to implement the current network strategy?

What should be in place to implement the prospective improved network strategy?

regulation&social

The prospective new network strategy is then valuated in order to assess the added value for each partner, as well as for the customer. The valuation techniques follow the methods and techniques presented in Decision Analysis, e.g. [10,11]. The customer may, or may not be included in the assessment. This depends on the character of the customer-network relationship. At the same time will be verified that the new network strategy produces added value for each (principal) partner, i.e. the prospective network strategy is a win-win strategy. The outcome of this activity is an action plan that indicates the next concrete measures to be performed in order to implement the new network strategy. In general, it is expected that the alternative network strategy entails only a small change in terms of strategy, but anyhow, reflects a major change in attitude and trust among the partners. Any formal change in strategy is always connected with changed expectations on the outcome and the way partners will act. If expectations are met, trust will build up, and the willingness to develop deeper partnerships will increase. Such a partnering process is incremental: evolving step-by-step. The formulation and valuation activities may be supported by a Group Decision Support System (GDSS), allowing effective ways of generating ideas, commenting, voting and arriving at an action plan. The use of GDSS is focusing on supporting the whole multi-phased process of group decision making with certain technologies, and it can be seen to fall under the larger umbrella of the concept of GSS (Group Support Systems), which may include any technologies used to make groups more productive [12,13].

37

Earlier studies have revealed several benefits of GDSS that are worth noticing in coping with the problem area of this paper. These are, for example, process structuring, goal oriented process, parallelism (many people able to communicate at the same time), allowance of larger group sizes, automatic documentation, anonymity of group members, access to external information, and automated data analysis [13-16]. The problem area in this paper is a complex one, it needs structuring, it influences on several people, and several people are needed to make decisions with regards to it. Therefore, it can be assumed that the use of GDSS would provide significant benefits. A large amount of results could be done during, e.g. two strictly phased GDSS workshop sessions, of which the first one would formulate the prospective network strategy, and the second one would have the principles of value assessment, metrics, and action plans as outputs. However, earlier studies have also stressed the significance of a detailed pre-planning: In pre-planning meetings, the purpose and goals for each GDSS sessions need to be carefully defined in order to make the process itself as efficient and effective as possible, and in order to improve the likelihood of making the concrete results of the sessions the best possible.

5

CONCLUSIONS

In the proposed framework for assessing the added value of a network strategy value is a measure of the subjective preferences of the network partners. Thus, a win-win strategy cannot be proved other than by monitoring the customer-network relationship and the relationships between the partners. The partners have different roles, skills, expectations, etc. that need to be aligned in a network strategy for adding value to the customer and to each other. The Balance Scorecard perspectives and strategy maps provide a good basis for developing a value theoretic assessment framework that supports the network orchestrator in the formulation and the valuation of a prospective network strategy. In the framework, the valuation is comparative with the current network strategy as the reference point. The framework is developed as one answer to the needs to improve partnering for maintenance and production services for a customer in the fertiliser business in Finland. The presented approach needs many test cases for refinement and validation and should be viewed as a reference for further research rather than a readily implementable method to develop value nets.

6

REFERENCES

1

Ojanen V, Lanne M, Reunanen M, Kortelainen H & Kässi T. (2008) New service development: success factors from the viewpoint of fleet asset management of industrial service providers. Fifteenth International Working Seminar on Production Economics, Pre-Prints Volume 1, 369-380.

2

Cimon Y. (2004) Knowledge-related asymmetries in strategic alliances. Journal of Knowledge Management, 8(3), 17-30.

3

Kaplan RS & Norton DP. (1992) The balanced scorecard - Measures that drive performance. Harvard Business Review, Jan-Feb, 71-79.

4

Kaplan RS & Norton DP. (2001a) Transforming the Balanced Scorecard from Performance measurement to Strategic Management: Part I. Accounting Horizons, 15, 87-104.

5

Kaplan RS & Norton DP. (2001b) Transforming the Balanced Scorecard from Performance measurement to Strategic Management: Part 2. Accounting Horizons, 15, 147-160.

6

Rosqvist T, Laakso K & Reunanen M. (2009) Value-driven maintenance planning for a production plant. Reliability Engineering and System Safety, 94, 97-110.

7

Kaplan RS & Norton DP. (2004) Strategy maps: Converting intangible assets into tangible outcomes. Boston: Harvard Business School Press.

8

Neely A, Bourne M & Kennerley M. (2000) Performance measurement system design: developing and testing a processbased approach. International Journal of Operations & Production Management; 20 (10): 1119-1145.

9

Fitzgerald L, Johnston R, Brignall S, Silvestro R & Voss C. (1991). Performance measurement in service business, Chartered Institute of Management Accountants (CIMA), London.

10

Keeney L & Raiffa H. (1993) Decisions with Multiple Objectives: Preferences and Value Trade-offs. Cambridge University Press.

11

Keeney R. (1992) Value-focused thinking – a path to creative decision making. Harvard University Press.

12

Nunamaker J, Briggs R & Mittleman D. (1996) Lessons from a decade of Group Support Systems Research. Proceedings of the 29th Annual Hawaii International Conference on Systems Sciences, Jan 3-6, Maui, Hawaii.

38

13

Elfvengren K. (2006) Group Support System for Managing the Front End of Innovation: case applications in business-tobusiness enterprises. Lappeenranta University of Technology, Acta Universitatis Lappeenrantaensis 239, doctoral dissertation Lappeenranta, Finland.

14

Jessup L & Valacich J. (1993) Group Support Systems: New Perspectives. Macmillan Publishing Company.

15

Weatherall A & Nunamaker J. (1995). Introduction to electronic meetings. Technicalgraphics.

16

Turban E, Aronson J. & Liang TP. (2004) Decision support systems and intelligent systems, 7th ed. Prentice-Hall.

39

Proceedings of the 4th World Congress on Engineering Asset Management Athens, Greece 28 - 30 September 2009

APPLICATION OF ACOUSTIC EMISSION TECHNOLOGY IN MONITORING STRUCTURAL INTEGRITY OF BRIDGES Manindra Kaphle a, Andy CC Tan a, Eric Kim a and David Thambiratnam a a

CRC for Integrated Engineering Asset Management, Faculty of Built Environment and Engineering, Queensland University of Technology, Brisbane, Australia.

Bridges are an important part of a nation’s infrastructure and reliable monitoring methods are necessary to ensure their safety and efficiency. Most bridges in use today were built decades ago and are now subjected to changes in load patterns that can cause localized distress, which can result in bridge failure if not corrected. Early detection of damage helps in prolonging lives of bridges and preventing catastrophic failures. This paper briefly reviews the various technologies currently used in health monitoring of bridge structures and in particular discusses the application and challenges of acoustic emission (AE) technology. Some of the results from laboratory experiments on a bridge model are also presented. The main objectives of these experiments are source localisation and assessment. The findings of the study can be expected to enhance the knowledge of acoustic emission process and thereby aid in the development of an effective bridge structure diagnostics system. Key Words: Structural health monitoring, acoustic emission, damage, bridge structures 1

INTRODUCTION

Bridges are an important part of a nation’s infrastructure and reliable monitoring methods are necessary to ensure their safety and structural well-being. Many bridges in use today were built decades ago and are now subjected to changes in load patterns that cause localized distress, which may result in bridge failure if not corrected. Early detection of damage and appropriate retrofitting can help in prolonging the lives of bridges and preventing failures. Bridge failures can cause huge financial losses as well as loss of lives, an example being the I-35W highway bridge collapse in Minnesota, USA in August 2007 which killed 13 people and injured 145 people. There are altogether 33500 bridges in Australia with replacement value of about 16.4 billion dollars and about 100 million dollars annual maintenance expenditure [1]. In USA, out of a total 593,416 bridges, 158,182, that is, around 26.7 percent were identified as being either structurally deficient or functionally obsolete [2]. These statistics point to the need of cost effective technology capable of monitoring structural health of bridges, ensuring they remain operational during their intended lives. The aim of the paper is to compare various methods currently used in health monitoring of bridge structures and in particular discuss the application and challenges of acoustic emission (AE) technology. Some of the results from laboratory experiments, which are aimed at finding source location and assessment, are also presented. The findings of the study can be expected to enhance the knowledge of acoustic emission wave propagation and signal analysis techniques and their applications in structural health monitoring.

2

LITERATURE REVIEW

2.1 Structural health monitoring techniques Visual inspection has been the traditional tool for monitoring bridges. Bridges are inspected at regular interval for visible defects by trained inspectors. Though simple, visual inspection results depend solely on inspectors’ judgements and small or hidden defects may go unnoticed. A range of newer techniques are available today that provide more reliable information than visual inspection. In commonly used vibration monitoring techniques, damage to a bridge is assessed by measuring changes in

40

the global properties (such as mass, stiffness and damping) of the whole structure and identifying the shifts in natural frequencies and changes in structural mode shapes [3-5]. But some damage may cause only negligible change in dynamic properties and therefore may go unnoticed. Additionally, these methods generally give the global picture, indicating the presence of damage in the whole structure; and local methods are often necessary to find the exact location of the damage. Several non-destructive techniques are available for local health monitoring of bridge structures. Most commonly used techniques are based on the use of mechanical waves (ultrasonic and acoustic), electromagnetic waves (magnetic testing, eddy current testing, radiographic testing) and fibre optics [4, 6]. Ultrasonic technique detects the geometric shape of a defect in a specimen using an artificially generated source signal and a receiver [7]. Magnetic particle testing uses powder to detect leaks of magnetic flux [8]. It can be an economic alternative to other methods but cannot be used for nonferrous materials. Eddy current testing is based on the principle that there is a change in the eddy- current pattern due to the presence of a flaw in a structure [4]. It can detect cracks through paint and is effective for detecting cracks in welded joints, but testing can be expensive. In radiographic methods, a suitable energy source is used to generate radiation and flaw is detected when the radiation is recorded in the other side of the specimen. Laboratory results of radiographic testing are promising but large size of the equipment hampers its use in field investigation. Fibre optics can detect various parameters; displacement and temperature are the common ones. Sensing is based on intensity, wavelength and interference of the light waves [9]. Advantages include their geometric conformity, capability for sensing a variety of perturbations and no electric interference [9]. But they can be costly and placement within the structure during the construction may be needed, excluding their use for already built bridges.

2.2 Acoustic emission technique Acoustic emission (AE) waves are the stress waves that arise from the rapid release of strain energy that follows microstructural changes in a material [10]. Common sources of AE are initiation/growth of cracks, yielding, impacts, failure of bonds and fibre failure. AE waves generated within a material can be recorded by means of sensors placed on the surface. AE technique involves the analysis of these recorded signals to obtain information about the source of the emission. Physically, AE waves consist of P waves (primary/longitudinal waves) and S waves (shear/transverse waves) and might further include Rayleigh (surface) waves, reflected waves, diffracted waves and others [11]. In plate like structures, as signals travel away from the source, Lamb waves become dominant mode of propagation [12]. Lamb waves primarily travel in two basic modes namely symmetric (S0) or extensional and asymmetric (A0) or flexural, though higher modes such as S1 and A1 can exist [13]. These modes travel with different velocities depending on the frequencies and thickness of the plate. Dispersion curves based on solutions to Lamb’s equation are used to relate velocity with the product of frequency and plate thickness [13]. Some of the advantages of AE technique over other non-destructive techniques are its high sensitivity, source localization capability and ability to provide monitoring in real time, that is, damage detection as it occurs. Study of AE started in 1950s and AE technique found its initial application in monitoring pressure vessels and aerospace structures. AE was first applied for bridge monitoring in early 1970s but the use and study of AE for monitoring bridge structures rose with the rapid increase in computing resources and development of sensor technology.

2.3 Applications of AE technology for bridge monitoring and challenges faced AE is well suited for the study of the integrity of bridge structures as it is able to provide continuous in-situ monitoring and is also capable of detecting a wide range of damage mechanisms in real time [12]. A general overview of applications of AE for monitoring bridges has been given in [13]. Application of AE for steel bridges has been discussed by [14] and application for concrete has been covered by [11] . A number of previous studies have explored the use of AE technology for monitoring bridge structures made of different materials, such as steel [12, 15, 16], concrete [17-19] and composites [20, 21] as well as masonry bridges [22]. Most of the studies have combined field testing with experiments performed in laboratory. The traditional way in AE monitoring involves the use of parameters of the recorded AE signals such as amplitude, energy content, rise time and duration in characterising damage [23]. This parameter based approach is simple but may be insufficient as all waveform information is not used during analysis. Waveform based approach involves recording the whole AE signal waveform and studying its features; and is often regarded as better than the parameter based method. Though AE technique has been successfully applied for bridge monitoring, several challenges still exist. Large size of bridges creates practical problems, for example with accessing desired areas and the need for a large number of sensors to monitor the whole structure. The solution is to identify critical areas and then to monitor these targeted areas. As large volume of data is generated during monitoring owing to high sampling rate, effective data management becomes important. AE signals can arise from a number of sources, so distinguishing the sources of origin is critical in understanding the nature of damage. For instance, in steel bridges likely sources of AE include crack growth, sudden joint failures, rubbing, fretting and traffic noises. Presence of noise sources that can mask the AE signals from real crack has been identified as one of the biggest limitation of the AE monitoring system. Different suggestions have been made for noise suppression [24] but scope exists for further research. Proper analysis of recorded data to obtain reliable information about the source is another major challenge in

41

AE technique. Frequency based analysis provides important information about the nature of the source. Along with traditional Fourier based analysis, other signal analysis techniques such as short time Fourier transform (STFT) and wavelet analysis (WA) are gaining popularity [7].

3

EXPERIMENTS

Since three important aspects of AE monitoring have been identified as source location, source identification and severity assessment [13], laboratory experiments were carried out to address them. Two sets of experiments were carried out to determine the source location in small and large plates using popular time of arrival (TOA) method. In TOA method, differences in arrival times of the signals at different sensors and velocity of the waves are used to find the locations of the source using triangulation techniques [25]. Influence of AE wave travel modes in source location process was also studied. Next set of experiments was aimed at finding a way to determine similarity between two different sources of AE signals. Knowing how to differentiate and classify signals from different sources can be expected to help in source identification and assessment. The AE analysing system used was micro-disp PAC (Physical Acoustics Corporation) system, along with AEwin software provided by the same company. Sensors used were R15a sensors (PAC) resonant at 150 KHz and preamplifiers used had a gain of 20 dB. Data was acquired at a sampling rate of 1 MHz for duration of 15 ms. The system recorded a hit when signals reached a certain threshold, set at 60 dB.

3.1 Source location experiments 3.1.1 Source location in small plate A 300 mm by 300 mm aluminium plate of thickness of 3 mm was used as a test specimen. Three sensors were placed in three different locations to record AE signals. The source of AE signals were pencil lead breaks which involved breaking 0.5 mm pencil leads at selected locations within the plate. Signals from pencil lead break tests have been found to closely resemble crack signals. MATLAB (R2008a, The MathWorks) codes were used for iterations purposes. Calculated locations were compared with the exact locations to verify the accuracy of the TOA method. 3.1.2 Source location in larger plate Similar experiments were conducted on a 1.8 m by 1.2 m steel plate of thickness of 3 mm that acted as a part of a deck for slab-on-girder bridge model. Again three sensors were used to record data.

3.2 Signal similarity experiments In a long steel beam of dimensions (3m long, 0.15m wide, 75mm thick), two sources of AE signals were generated: pencil lead breaks and steel ball drops (6 mm diameter balls dropped from a vertical height of 15 cm). 10 sets of each test were carried out. The signals were recorded by a sensor placed at a distance of 1.5 m from the source. They were then analysed to determine similarity, with an aim of finding a way to classify signals from various sources. A parameter called magnitude squared coherence (MSC) was used to measure similarity. The magnitude squared coherence estimate is a function of frequency with values between 0 and 1 that indicates how well two signals correspond to each other at each frequency, with the value of 1 indicating exact match (MATLAB help guide R2008a, The MathWorks). MSC is calculated using the power spectral densities and the cross power spectral density of the signals [7].

4

RESULTS

4.1 Source location experiments For small plate source location experiment, longitudinal wave velocity in aluminium ( c L = E / r = 5128 m/s, E being Young’s modulus and ρ density) was used to calculate the source location. Fig.1 shows the calculated positions and the exact locations of pencil lead breaks, along with the sensor locations. A good correlation between the calculated and exact values is seen.

42

The result of the larger plate source location experiment using the longitudinal wave velocity in steel (cL=5188 m/s) is shown in Figure 2a. The exact and calculated values do not show good match. Results using c = 3000m/s, a value close to the transverse velocity (cT = E / 2 r (1 + u ) , where ν is the Poisson’s ratio) is shown in Figure 2b, where a much better correlation is obtained. The results indicate that the waves recorded by the sensors are not longitudinal waves, as in the first experiment.

Figure 1 Source location in small plate

(a) Using c = 5188 m/s

(b) Using c = 3000 m/s

Figure 2 Source location in larger plate with two different velocities Initial parts of sample signals from the two source location experiments are shown in Figure 3, along with the threshold value (dotted line); demonstrating how the threshold is crossed by the recorded signals.

43

(a)

(b)

Figure 3 Record of hits in small plate (a) and larger plate (b) In Fig. 3a, the first arriving wave crosses the threshold and records a hit. On the other hand, in Fig. 3b, the initial portion consists of low amplitude signals that do not cross the threshold; and therefore do not trigger a hit. Also seen is that the initial component arrives about 90 µs before the triggering wave component. Using the velocity of the triggering wave (c=3000m/s) and the distance between the source and signal (in this case signals were recorded by Sensor S3 for source at position of (0.3, 1.2), so the distance was calculated to be 0.67 m); the velocity of the initial arriving wave can be calculated to be around 5000 m/s. This value is close to the longitudinal velocity of waves in steel. Initial conclusions can be drawn that though longitudinal waves are present they have attenuated to a level below the threshold; and the waves that record a hit by crossing the threshold are the transverse waves. More investigation is needed to check whether the waves seen are Lamb wave modes, as Lamb waves are common in large plate like structures. Detailed frequency analysis of the signals by means of Fourier analysis and short time Fourier analysis (STFT) are expected to be useful in identifying the modes. Therefore, Fourier analysis was carried out for two parts of the sample signal from the large plate source experiment: the initial 90 µs portion and the next 230 µs portion. The frequency response diagrams are shown in Figure 4.

(a) Initial 90 µs

(b) Next 230 µs

Figure 4 FFT of two portions of signal from large plate source location experiment The major difference observed between Figures 4a and 4b is that frequency peaks around 47, 70 and 90 kHz appear in Fig. 4b, indicating that these lower frequency wave modes arrive late and trigger a hit. To obtain more information, Short time Fourier transform (STFT) analysis was carried out using time-frequency toolbox [26] and the results are shown in Fig. 5.

44

Figure 5 STFT analysis of the signal From the STFT plot in Fig. 5, it is clear that waves with frequencies around 100 to 180 kHz arrive at the beginning. Starting at around 90 µs, waves with a large variation in frequencies arrive. The frequencies gradually decrease from around 350 kHz till 30 kHz, with a peak value at around 45 kHz and 250 µs. Another similar wave pattern with decreasing frequency values emerges after 200 µs; these could be the reflected waves. For further insight into Lamb wave phenomenon, study of the dispersion curve for steel is useful. It is given in Figure 6 and shows the variation of the velocities of modes S0, A0, S1 and A1 with frequencies and thickness of the plate. Using frequency f of 100 kHz and 180 kHz and 3 mm thickness t of the plate (f.t = 0.3 or 0.54 MHz.mm), group velocity of around 5000 m/s is seen for S0 mode in Fig. 6. This value matches calculations made before for initial fast arriving component. But for the triggering slow arriving wave (of velocity 3000 m/s) to be flexural mode (A0), frequency of around 333 kHz or more is required (f.t = 1 MHz.mm). This high frequency component is not conspicuous in FFT analysis in Fig. 4b, though it can be seen in STFT analysis in Fig. 5. But since signals with a wide range of frequencies are present, it is hard to ascertain that 333 kHz component crosses the threshold first. It has to be added that due to sensor sensitivity, frequencies in higher range are not recorded properly, as sensors tend to be sensitive to the signals near their resonant frequency, in this case 150 kHz.

Figure 6 Dispersion curves in steel [13]

45

4.2 Experiment 3 To judge the similarity of signals, magnitude squared coherence (MSC) is used. MSC values between signals from pencil lead break experiments have a mean of 0.78 while the MSC values between pencil lead break and ball drop signals have a mean of 0.38. These differences are significant, showing the possibility for further exploration of magnitude squared coherence for signal classification purposes. A sample MSC values between two pencil lead break signals (a) and between pencil lead break and ball drop signals (b), using MATLAB code mscohere (R2008a, The MathWorks), are shown in Fig. 7. Fig. 7a indicates close match between the signals, especially up to 400 kHz, whereas Fig. 7b indicates less coherence.

(a)

(b)

Figure 7 MSC values versus frequencies

5

DISCUSSIONS

Experiments in this study have explored the use of frequency analysis for better interpretation of recorded AE data. Ability to determine the location of the source is an important advantage of acoustic emission technique. AE propagation in solids is a complex phenomenon as signals travel in various modes and mode conversions occur. For accurate source location, identification of the modes is necessary. It has been seen that progressive attenuation of longitudinal waves precludes their use in locating the source of AE events in larger scale [15]. The results from source location experiments confirm this case. For signal analysis, STFT is found to be more informative than Fourier transform, as both frequencies and times of occurrence of different wave modes are seen in STFT analysis. More rigorous analysis is required to ascertain whether the wave modes recorded are the Lamb wave modes, transverse waves or other waves. Use of finite element analysis and other techniques such as wavelet analysis are expected to be beneficial and will be carried out in the next stage. Similar sources have been found to give similar waveforms. Magnitude squared coherence (MSC) based on the power spectral frequency analysis of the signals provides a simple way of judging signal similarity, as verified by the experimental results. Signal similarity can be an effective tool for signal classification and thus for source identification and assessment, which are some of the important aspects of AE monitoring method. A crack waveform obtained from a laboratory experiment can act as a template for distinguishing a similar signal obtained in field testing from other noise sources. Waveform recorded by a sensor is influenced by the path travelled by the signals and by the sensor characteristics; the influence of these parameters needs further consideration.

6

CONCLUSIONS

The study of acoustic emission technique for monitoring bridge structural integrity is growing continually. Though AE technique has several distinct advantages over other non-destructive methods of monitoring, it still has not become the preferred choice in bridge monitoring. Existence of sources of noises that can mask AE signals from real damage has been identified as a major hindrance for the use of AE technique in monitoring bridges. Signal processing techniques including the frequency analysis tools used in this study can be effective tools to distinguish and remove noises from real signals. Proper analysis of recorded AE signals to deduce useful information about the nature of the source is another challenge. This study has aimed to address some of these issues by analysing the signals for source location and source assessment purposes. Though only laboratory tests have been carried out so far, knowledge from these tests is valuable in interpreting the results from actual

46

field tests. Plate like structures and beams are common in bridges, and hence make ideal experimental specimen. As AE is generally used as local monitoring technique, critical areas where damage is likely are identified and specifically monitored. In experiments in this study, signals were recorded for a distance up to slightly more than 1m for plates and 1.5 m for beams. Studies on attenuation of AE signals (not shown in this paper) have shown possibility of recording AE signals for larger distances (5-7 m) effectively; hence fairly wide area could be monitored using strategically placed AE sensors. The approach in this study was to record experimental data first and then transfer and analyse data later on. Real time analysis is an attractive option and, especially with advanced computing resources available today, should be possible for implementation. To conclude, it can be said that monitoring structural integrity of bridges by AE provides insight into the current state and helps determine if further steps are necessary to extend the lives of bridges and to ensure they perform safely and reliably.

7

REFERENCES

1

Austroads, (2004) Guidelines for Bridge management - Structure Information. Austroads Inc: Sydney, Australia.

2

USDoT, (2006) 2006 Status of the Nation's Highways, Bridges, and Transit: Condition and Performance. U.S. Department of Transportation Federal Highway Administration Federal Transit Administration.

3

Shih, H.W., D.P. Thambiratnam, and T.H.T. Chan, (2009) Vibration based structural damage detection in flexural members usign multi-criteria approach. Journal of sound and vibration. 323, 645-661.

4

Chang, P.C. and S.C. Liu, (2003) Recent research in nondestructive evaluation of civil infrastructures. Journal of materials in civil engineering. p. 298-304.

5

Chang, P.C., A. Flatau, and S.C. Liu, (2003) Review paper: Health monitoring of civil infrastructure. Structural health monitoring. 2, 257-267.

6

Chong, K.P., N.J. Carino, and G. Washer, (2003) Health monitoring of civil infrastructures. Smart Materials and structures. 12, 483-493.

7

Grosse, C.U., et al., (2004) Improvements of AE technique using wavelet algorithms, coherence functions and automatic data analysis. Construction and building Materials. 18, 203-213.

8

Rens, K.L., T.J. Wipf, and F.W. Klaiber, (1997) Review of non-destructive evaluation techniques of civil infrastructure. Journal of performance of constructed facilities. 11(2), 152-160.

9

Ansari, F., (2007) Practical implementation of optical fiber sensors in civil structural health monitoring. Journal of intelligent material systems and structures. 18, 879-889.

10 Vahaviolos, S.J., (1996) Acoustic emission: A new but sound NDE technique and not a panacea, in Non destructive testing, D. Van Hemelrijck and A. Anastassopoulos, Editors. Balkema: Rotterdam. 11 Ohtsu, M., (1996) The history and development of acoustic emission in concrete engineering. Magazine of concrete research. 48(177), 321-330. 12 Holford, K.M., et al., (2001) Damage location in steel bridges by acoustic emission. Journal of intelligent material systems and structures. 12, 567-576. 13 Holford, K.M. and R.J. Lark, (2005) Acoustic emission testing of bridges, in Inspection and monitoring techniques for bridges and civil structures, G. Fu, Editor. Woodhead Publishing Limited and CRC. p. 183-215. 14 Lozev, M.G., et al., (1997) Acoustic emission monitoring of steel bridge members. Virginia transportation research council. 15 Maji, A.K., D. Satpathi, and T. Kratochvil, (1997) Acoustic emission source location using lamb wave modes. Journal of engineering mechanics. p. 154-161. 16 Sison, M., et al., (1998) Analysis of acoustic emissions from a steel bridge hanger. Research in Nondestructive Analysis. 10, 123-145. 17 Colombo, S., et al., (2005) AE energy analysis on concrete bridge beams. Materials and structures. 38, 851-856. 18 Shigeshi, M., et al., (2001) Acoustic emission to assess and monitor the integrity of bridges. Construction and building materials. 15, 35-49. 19 Yuyama, S., et al., (2007) Detection and evaluation of failures in high-strength tendon of prestressed concrete bridges by acoustic emission. Construction and building materials. 21, 491-500. 20 Rizzo, P. and F.L. di Scalea, (2001) Acoustic emission monitoring of carbon-fiber-reinforced-polymer bridge stay cables in large-scale testing. Experimental mechanics. 41(3), 282-290.

47

21 Gostautas, R.S., et al., (2005) Acoustic emission monitoring and analysis of glass fiber-reinforced composites bridge decks. Journal of bridge engineering. 10(6), 713-721. 22 Melbourne, C. and A.K. Tomor, (2006) Application of acoustic emission for masonry arch bridges. Strain - International Journal for strain measurement. 42, 165-172. 23 Vallen, H., (2002) AE testing fundamentals, equipment, applications. NDT.net. 7(09). 24 Daniel, I.M., et al., (1998) Acoustic emission monitoring of fatigue damage in metals. Nondestructive testing evaluation. 14, 71-87. 25 Nivesrangsan, P., J.A. Steel, and R.L. Reuben, (2007) Source location of acoustic emission in diesel engines. Mechanical systems and signal processing. 21, 1103-1114. 26 Auger, F., et al., (1996) Time-Frequency Toolbox - For use with MATLAB. CNRS (France) and Rice University (USA).

Acknowledgments The authors gratefully acknowledge the financial support from the QUT Faculty of Built Environment & Engineering and the Cooperative Research Centre for Integrated Engineering Asset Management (CIEAM).

48

Proceedings of the 4th World Congress on Engineering Asset Management Athens, Greece 28 - 30 September 2009

APPLICATION OF TEXT MINING IN ANALYSING ROAD CRASHES FOR ROAD ASSET MANAGEMENT Richi Nayak1, Noppadol Piyatrapoomi2 and Justin Weligamage2 1

Faculty of Science and Technology, Queensland University of Technology, Brisbane, QLD 4001, Australia

2

Road Asset Management Branch, Queensland Government Department of Main Roads Brisbane, Queensland, Australia

Traffic safety is a major concern world-wide. It is in both the sociological and economic interests of society that attempts should be made to identify the major and multiple contributory factors to those road crashes. This paper presents a text mining based method to better understand the contextual relationships inherent in road crashes. By examining and analyzing the crash report data in Queensland from year 2004 and year 2005, this paper identifies and reports the major and multiple contributory factors to those crashes. The outcome of this study will support road asset management in reducing road crashes. Key Words: Text Mining, Road Crashes, Data Analysis 1

INTRODUCTION

Traffic safety is a major concern in many states around Australia and including the state of Queensland. Since year 2000, there had been 296 fatalities on average per year in Queensland as recorded by the Office of Economic and Statistical Research [1]. Data obtained for analysis in this paper shows that during the years 2004 and 2005, there were over 20,000 traffic crash investigation reports recorded involving Queensland motorists. The annual economic cost of road crashes in Australia is enormous - conservatively estimated at $18 billion per annum - and the social impacts are devastating [2]. The cost to the community through these crashes is very high. They also have a devastating impact on the emergency services and a range of other groups. In addition, it is inevitable that the insurance companies will have to increase the cost of premium to cover their ongoing cost of insuring those motorists and their vehicles. It is therefore in both the sociological and economic interests of society that attempts are made to identify the major and multiple contributory factors to those crashes. Statistical analysis of road crashes is not a new realm of research by any means. For many years, road safety engineers and researchers have attempted to deal with large volumes of information in order to gain an understanding of the economic and social impacts of car crashes. The hope is, that with this understanding, more efficient safety measures can be put into place to decrease the number of future road crashes [3]. Various data mining and statistical techniques have been used in the past in the domain. Researchers have attempted to investigate crash analysis through ordinary statistical tables and charting techniques [4, 5].The issue with these techniques is that they limit human involvement in the exploration and knowledge discovery tasks. Researchers have also attempted advanced data analysis methods of data mining that include clustering, neural networks and decision trees to reveal relationships between distractions and motor vehicle crashes. Major focuses of research on road crashes are the use of data mining to analyse freeway or highway accident frequency, the development of models to predict highway incident durations and the use of data mining in the classification of accident reports [6, 7, 8]. Other studies include the use of data mining and situation-awareness for improving road safety; a comparison of driving performance and behaviour in 4WDs versus sedans through data mining crash databases [5] and a study in the safety performance of intersections [9]. These studies revealed some interesting results, however, they are unable to properly analyse the cognitive aspects of the causes of the crashes. They often opt to leave out significant qualitative and textual information from data sets as it is difficult to create meaningful observations. The consequence of textual ignorance results in a limited analysis whereby less substantial conclusions are made. Text mining methods attempt to bridge this gap. Text Mining is discovery of new, previously unknown information, by automatically extracting it from different written (text) resources. Text mining methods are able to extract important concepts and emerging themes from the collection of text sources. Used in a practical situation, the possibilities for

49

knowledge discovery through the use of text mining is immense. To our knowledge, there is limited or no reputable studies that have utilised text mining in this data domain, however, earlier studies in the field indicate a real need for textual mining in order to better understand the contextual relationships of road crash data. This paper presents a text mining based method to better understand the contextual relationships inherent in road crashes. By examining and analyzing the crash report data in Queensland from year 2004 and year 2005, this paper identifies and reports the major and multiple contributory factors to those crashes. Analysis is performed to identify links between common factors recorded in crash reports. Of key concern are the causes of crashes, rather than the consequences. The outcome of this study will support road asset management in reducing road crashes. With those findings on hand, we hope it can be useful for reviewing the limitations of existing road facilities as well as planning better public safety measurements. Most importantly, implementing and continuing a long term public education on road safety issues especially amongst the young generations and male gender which historically are involved in a high proportion of road crashes each year.

2

TEXT MINING METHOD

Text mining is discovery of new and previously unknown information automatically from different text resources using natural language and computation linguistics, machine learning and information science methods [10]. The key element is linking of the discovered information together to form new facts or new hypotheses to be explored further by more conventional means of experimentation [10]. Text mining methods include the steps of processing the input text data, deriving rules and patterns within the newly processed data and finally the evaluation and interpretation of the output rules and patterns.

2.1

Objectives

The focus of this paper is to determine the most common causes of road crashes so that appropriate measures can be taken by road asset management in the future to prevent these accidents from occurring. The objective is to investigate the nature of crashes and roads that result in the crashes reported by Queensland traffic accident investigators within the period 2004 to 2005. Performing text mining on the crash description gives information about the causes of a crash that cannot necessarily be categorized into any particular field within a database. The crash reports, when pre-processed and grouped into clusters, can reveal insights that may have formally been unrecognisable. The identified unusual and hidden relationships may be useful to government, businesses (insurance organisations, motoring associations) and individuals in better road asset management. As compared with simply having a quantitative data set, textual information can enable conclusions to be drawn from the circumstances that caused the accident as opposed to simply looking at what the accidents were.

2.2 Dataset The data used for this analysis, is collected from files containing information related to road crashes from the state of Queensland. The two files supplied are actual reports produced by traffic accident investigators within the period 2004 to 2005. They contain data for each reported road accident in this period according to 29 attributes including date, time, location and road conditions of the crash. More specifically they are: Atmospheric, Carriageway, Crash_Description, Crash_Nature, Crash_Area, Crash_Date, Crash_Day_of_Week, Crash_Distance, Crash_Divided_Road, Crash_Landmark, Crash_Number, Crash_Speed_Limit, Crash_Time, District, Horizontal_Allignment, Lighting, Owner_ID, Roadway_Feature, Road_Section, Road_Surface, Traffic_Control and Vertical_Allignment. Preliminary review of the dataset helped us in determining the most interesting and important attributes to be used in our text mining analysis. However, for the purposes of text mining the crash description was of highest significance. This is a character attribute containing data with values up to 403 characters. The attribute “Street 2” with data of for example: “Warrego highway” had over 241 reported crashes which is also an interesting piece of information to look into. It is noted that there were as many as over 80% crashes occurred with clear conditions as shown in the attribute “Atmospheric”, and over 65% cases occurred during “daylight hours”. Also, a large number of reported crashes occurred on sealed and dry road surface as shown in the attribute “Road_Surface”, only small number of crashes occurred on wet surface. Another interesting attribute that came out of our observations was “Owner_ID”. It represents the gender of the person who was involved in each of these described crashes and it appears that almost all of them are MALE. The assumption is that this is due to the limitation of this particular crash report data. It may or may not be a true reflection of the correct distribution of gender involved in road crashes in a broader sense.

50

2.3 Process: Step1 - Pre-processing Pre-processing of textual information is a time-consuming task but is essential in order to achieve results that are of value to the users of the information. An initial scan through the data set identified a number of potential problems that will need to be addressed before any text mining could take place. The main cause of this is mainly due to noise and various inconsistencies between the different records, which could possibly be due to different forensic experts writing the notes. Punctuation: Punctuation was often omitted or used extraneously. No consistent information was apparent in the use of punctuation. Therefore, to simplify the text mining all punctuation was removed and replaced with spaces. Specifically, the following characters were replaced: ~`!@#$%^&*()_+-={}[]|\;':'<>?/,. Broken Words: The previous step resulted in some words with gaps. Also there were many gaps (spaces) between words that will need to be removed in order to obtain any value from the words. This is because gaps between words are not actually “new” words, and during the text mining process they will have a low frequency, meaning that they are unlikely to be used. These gaps were consequently removed in order to obtain a more accurate result. Some examples are “trave lling”, “ro ad”. Inconstancy due to the user of abbreviation and different cases: Another problem encountered with the data set is that there are many inconsistencies in different records. An example of this would be “unit 1”, where variants including “u1” and “unit one” were used throughout the data set. This presents a problem in the context of road crash text mining because names of roads and highways could be abbreviated in a multitude of ways. In order to provide any meaningful recommendations, abbreviations also had to be standardised. It has been agreed that most value can be provided to the end user if the full word was used. Another example of inconsistency was using lowercase and uppercase to representing the same word. Consequently data was transferred to lowercase to prevent text mining tools from separating the same words that started with either upper or lower case. If this action was not taken, same words would not be grouped, therefore misleading the results and the integrity of the analysis. Converting all the text to lower case meant there were less combinations to code for transforming abbreviations to their full descriptions. Spelling mistakes: Spelling mistakes was another common problem encountered during the pre-processing phase. They were removed by filtering through each record and correcting mistakes through the data set. If spelling mistakes are not corrected, text mining tools do not recognise the word, nor would it group same words. Some examples are “uint”, “unti”, “utni” user frequently to write the word “unit”. Common phrases: Finally as part of the formatting functions, common phrases that comprise of more than a word were combined into a single word to assist the text analysis. The data was processed for words which were in close proximity to each other to create common combinations. This was important as combinations such as green light and police station do not have the same meaning if they were not combined. For instance the car could have been green and it crashed into a light pole, instead of the car went through a green light and was hit in the middle of an intersection. Table 1 presents examples of the phrases replaced with concatenated words.

Table 1: Example replacements for common phrases

Original Text Traffic light red light turning lane stop sign green light give way lost control police station parking bay failed to stop road side bruce highway round about towing a trailer

Replace With red-light stop-sign turning-lane stop-sign green-light give-way lost-control police-station parking-bay failed-to-stop road-side bruce-highway roundabout towing-a-trailer

51

2.4 Process: Step 2 – Text Mining The process of text mining includes converting unstructured text data into structured data, clustering the crash reports to identify links between common factors reported in crash reports, and viewing the concept links. We employ the Leximancer tool [13] based on the bayesian theory [14] to assess each word in the dataset to predict the concepts being discussed. It learns which word predicts which concept (or cluster) and forms concepts (or clusters) based on associated terms. It thus positions clusters based on the terms that they share with other clusters. It constructs a conceptual graphical map by measuring the frequency of occurrence of the main concepts and how often they occur close together within the text. A concept is treated as a cluster. Each term appearing in the text data is analyzed to form a concept to allow blackbox discovery of patterns that may not otherwise be known. Concepts that are similar are merged and edited. For example, the concept list included concepts such as turn, turning and turned; direction, north, south, east and west; light and lights; approached and approaching; road and street; lane and lanes. Each of these combinations of concepts relates to the same thing and one word is merely a stem of the other. As a result these similar concepts are merged into one and are renamed to reflect the true meaning of the concept, for example: Day, Time, Years, and Week are merged into the single concept ‘Time’. Some concepts are removed which may not be pertinent to the crash senario being analysed, for example, preceded and occurred. Many concepts are then put together to form a theme.

Table 2: Example stop words excluded from the standard stop-word list Word Excluded from the stop list Bald Hit Look, looking Fast Indicate, Indicated Right Following Two 3

Rational “Bald” tyres may be a cause of an accident. Hit may imply a collision. Look and Looking may be referring to where a driver was looking when the accident occurred. Speed which may be the cause of an accident. Whether a driver indicated left or right. Turning and merging right as opposed to left may have more of an impact in collisions. (i.e. turning across traffic). Car could be following too closely to another vehicle. Could refer to Unit two or number of vehicles involved in the accident.

ANALYSIS AND RESULTS

3.1 Dataset examination Data distribution of the data set is displayed in figures 1 to 5 showing some significant correlations as well as disassociations between various attributes. The atmosphere is usually clear when the crash occurred indicating weather condition was not a big factor in this dataset. The distribution of crash time is mostly in the afternoon especially between 3pm to 5pm. This is during afternoon peak times when drivers are tired from working all day. The area with speed limit of 60km/h was where most crashes occurred, following by the area with speed limit of 100km/h. This is expected as the majority of roads have a speed limit of either 60km/h or 100km/h. No traffic control showing as most important contributory factor for the crash in this dataset. The three most significant crash natures were angle, hit fixed obstruction/ temporary object and rear-end. Figure 6 shows that over a quarter of the accidents (28%) are classified as “rear-end” which is a high proportion of the data given there are 14 categories for this attribute. The data in Figure 7 displays the count of accidents grouped by the characteristic of the road where the accident occurred according to the “Roadway Feature” attribute. The proportion of accidents that occur at some form of intersections (i.e. Cross, interchange, multiple road, roundabout, T junction and Y junction) is 95% (excluding Not Applicable). This would indicate that there might not be enough controls in place around intersections to avoid an accident.

52

Figure 1 - ATMOSPHERIC attribute

Figure 2 - CRAS_TIME attribute

Figure 3 - CRAS_SPEED_LIMIT attribute

Figure 4 - TRAFFIC_CONTROL

Figure 5 - CRASH_NATURE attribute

Figure 7: Roadway Feature Summary

Figure 6 - Crash Nature Summary

53

Figure 8 – The cluster map 3.2 Cluster Analysis This Cluster Map in Figure 8 shows the different Clusters that were produced by the Leximancer text mining tool after the pre-processing had been preformed. Several clusters are immediately obvious by communicating possible causes for road accidents including intersections, rear-ending and loss of control. The list of concept terms generated by the Leximancer tool includes roundabout, intersection, traffic, bend, lane, injuries, left and right, give-way, rear and speed, among others. These terms alone give a good indication of possible causes of road accidents as they are the most frequently appearing terms in the sample text once stop words have been removed. The travelling, Unit 1, Unit 2, road, right, vehicle and intersection have the highest relative counts. The two highest frequency concept terms are Unit 1 and Unit 2. Unit 1 occurs 12,774 times whilst Unit 2 only occurs 7,286 times. Similarly, the concept terms ‘left’ and ‘right’ appear 3111 and 6446 times respectively. An analysis into why ‘right’ might appear more than twice as many times as ‘left’ revealed that perhaps more accidents occur in right-hand lanes or whilst performing right-hand turns. Indeed, the relationship between ‘right’, ‘left’ and ‘intersection’ showed that the concept term ‘intersection’ will be accompanied by the concept term ‘right’ 73.5% of the time whilst it was only accompanied by ‘left’ 18% of the time. An immediate assumption that could be made regarding the reason for Unit 1 appearing nearly twice as many times as Unit 2 would be that single vehicle accidents occur more frequently than multi-vehicle accidents. However, this assumption may not necessarily be true. For example, Unit 1 may just be repeated more times than Unit 2 within the same passage. Figure 9 indicates the strength of the relationship between Unit 1 and all other concept words, whilst Figure 10 indicates the strength of

54

the relationship between Unit 2 and all other concept words. The relationship from Unit 1 to Unit 2 is moderately strong whilst the relationship from Unit 2 to Unit 1 is significantly stronger. The relative count of the first relationship shows that the term ‘Unit 2’ (7286) is closely related to the term ‘Unit 1’ (7286) 100% of the time. However, the second relationship, shows that the term ‘Unit 1’ (12774) is only closely related to the term ‘Unit 2’ (7268) 57% of the time. This information indicates that 43% of the time a second vehicle is not involved. It is therefore possible to conclude that nearly half of all road crashes in this case study are single vehicle accidents.

Figure 9: Unit 1 Concept

Figure 10: Unit 2 Concept

With the above discoveries in mind, the clusters can now be analysed to identify meaning in the grouping of concept words and their relative locations. The first meaningful clusters are the ‘vehicle’ and ‘lost-control’ clusters (as shown in Figure 11). These clusters appear in close proximity to each other, in fact overlapping, indicating a strong relationship between the two clusters. These two clusters include key words such as ‘towing’, ‘trailer’, ‘speed’ and ‘lost-control’. One possible conclusion that could be drawn from these concept words is that drivers can often lose control of their vehicles when speeding. Another is that drivers can easily lose control of their vehicles when towing a trailer. The ‘vehicle’ cluster may also indicate a relationship between these conclusions and ‘single driver’ accidents or ‘single vehicle’ accidents.

Figure 11: Driver Control Cluster

Figure 12: Rear-ending Cluster

The second meaningful cluster (as shown in Figure 12) is the ‘rear’ cluster which is also overlapping with the ‘Unit 2’ cluster. The concept words of the cluster include ‘slowed’, ‘stop’, ‘time’, ‘collided’ and ‘rear’. These combination of words could indicate a scenario of rear-ending, a common form of car crash in suburban areas. The concept terms ‘collided’ and ‘rear’ alone would suggest this to be the case. However, this is also supported by the terms ‘stop’ and ‘time’ indicating that perhaps a vehicle could not ‘stop in time’ and as a result collided with the vehicle in front.

Figure 13: Intersection Cluster

Figure 14: Speed Concept

55

The ‘intersection’ cluster (as shown in Figure 13) also seems to indicate quite meaningful information. An immediate observation that can be made is that this cluster overlaps with both the ‘Unit 1’ and ‘Unit 2’ clusters, indicating that perhaps accidents at intersections often involve two vehicles. The ‘intersection’ cluster includes interesting concept words such as ‘red light’, ‘green light’, ‘intersection’, ‘intending’ and ‘give way’. These key words might indicate that ‘giving way’ (or lack thereof) at intersections in perhaps a common cause of road accidents. This cluster may also suggest that traffic lights are often involved in crashes at intersections. Although this data alone does not tell us exactly how the traffic lights might be related to the accident one conclusion is that perhaps people are not stopping for red lights or simply do not see them. However, the relationship between traffic lights and causes of crashes is an area for further investigation. Although these three clusters were perhaps the most meaningful, further conclusions could be drawn from the remaining clusters with further analysis. These three clusters in particular were chosen for analysis as they indicated possible causes for road accidents. The implications of these findings are discussed later. One last area of interest is how road accidents are influenced by speeding. An analysis of the relationship between the concept word ‘speed’ and all other concept words indicates that perhaps speeding is a cause of more accidents in ‘low speed’ areas such as roads or streets rather than ‘high speed’ areas such as highways. As can be seen in Figure 14, the “speed” concept can be correlated with “road” or “street” (these two words were grouped in pre-processing) 63.6%. Whereas, highway and motorway only appear with speed 17.6% and 13.2% of the time respectively. Whilst the word ‘speed’ alone does not necessarily indicate speeding, it can be assumed that if the word were to appear in a crash report the speed must have been an influence factor.

4

DISCUSSION AND CONCLUSION

Several conclusions were drawn from the analysis conducted above. These conclusions involved: (1) the likelihood of a second vehicle being involved in an accident; (2) the likelihood of an accident when turning right as opposed to turning left; (3) the influence of towing a trailer in losing control of a vehicle; (4) the influence of speed in losing control of a vehicle; (5) a person’s inability to stop resulting in a rear-ending accident; (6) the likelihood of more than one vehicle being involved in an intersection accident; and (7) the influence of speed zone category in speeding accidents From these conclusions, various recommendations can be made. Our proposed recommendations are as follows: (a) greater awareness be raised regarding following another vehicle too closely or better known as tail-gating. Such awareness may help reduce the number of incidents related to rear-ending. (b) determine new as well as improving existing controls to prevent these sort of rear ending accidents through signalling by the immobile vehicle. This would involve developing and improving methods of displaying that a vehicle is immobile to vehicles approaching it on either side. This includes for both the vehicle as well as for trailers that are attached to it to prevent the rear ending happening. (c) further enhance future analysis of accident data, the improvement of information capture can be achieved by recording the presence/absence of right hand turning lanes at the intersection (for those accidents occurring at an intersection). (d) determine new as well as improving existing roadway features. As mentioned in point (c), turning lanes are used to improve safety at intersections, however, if these are not able to be installed at certain intersections there may be a requirement to develop alternate controls. Another consideration is if turning lanes do not reduce accidents at intersections, which could be the subject of additional research. (e) drivers purchasing trailers should be made aware of the difficult in controlling such vehicles and the implications associated with this. (f) Speeding campaigns should target low speed zone areas rather than high speed zone areas and speed cameras should be utilized in low speed areas more often to discourage speeding in these problem areas. (g) Furthermore, drivers should be reminded of ‘give way’ rules and there should perhaps be a greater focus on these rules during driving exams, particularly focusing on right-hand turns. Table 4 lists recommendations for improvement in road assets according to the features that are highlighted during text mining analysis of the crash data set. Finally, this paper has focused on the causes of road accidents and has not considered the consequences of such accidents but recognises that this is an equally significant area of concern. Whilst there is some information regarding injuries and damage to vehicles it is recommended that further research be conducted into the consequences of road accidents.

56

Table 4: Recommendation according to the features from the crash data set of interest Other features Losing control of vehicle or rolling on embankment Motorcyclist accidents Speeding through intersections Failure to obey signs. Collision with inanimate objects Blood samples taken Accidents at traffic lights Police did not attend scene/minor accidents Collisions due to right-hand turns Towing trailers Serious accidents requiring hospitalisation Rear-end collisions

Recommendations for improvement (if possible) Increased signage of accident-prone areas. Install road barriers if feasible. Increased regulations for gaining a motorcycle licence. Encourage motorcyclists to be more careful on roads at all times. This figure is concerning. More driver education is needed. Fixed speed cameras could be considered. Public awareness campaign might be required. Decrease speed limits if warranted. Stricter consequences for violation of traffic regulations. Reflector strips on guardrails, other inanimate objects. Install parking lane for stopped vehicles. Increase visibility of traffic lights or install signage on approach to lights. Improve traffic lights at Southport and Nerang. Consider installing a right-hand turn lane. Create signalised intersections. Educate drivers with long/ heavy loads how to manoeuvre vehicle properly. Investigate each of these accidents separately. Remind drivers to keep a safe distance behind other vehicles at all times.

5

REFERENCES

1

Australian Government, Department of Infrastructure, Transport, Regional Government and Road Safety. (2008) Road Safety, http://www.infrastructure.gov.au/roads/safety/, Retrieved October/2008.

2

Queensland Fire and Rescue. (18/09/2002) Firefighters called to http://www.fire.qld.gov.au/news/view.asp?id=207 , Retrieved October/2008

3

Abugessaisa, I. (2008). Knowledge discovery in road accidents database – Integration of visual and automatic data mining methods. International Journal of Public Information Systems, 2008 (1), 59-85. Retrieved October 20, 2008, from Emerald Insight database.

4

Gitelman, V. and Hakkert, A. S. (1997) The evaluation of road-rail crossing safety with limited accident statistics. International journal of Accident Analysis Prevention, 29 (2), 171-179. Retrieved October 20, 2008 from Emerald Insight database.

5

Gurubhagavatula, I., Nkwuo, J. E., Maislin, G., and Pack, A. I. (2008). Estimated cost of crashes in commercial drivers supports screening and treatment of obstructive sleep apnea. International Journal of Accident Analysis & Prevention, 40 (1), 104-115. Retrieved October 20, 2008 from Emerald Insight database.

6

Chatterjee, S. (1998). A connectionist approach for classifying accident narratives. Purdue University.

7

Li-Yen, C., & Wen-Chieh, C. (2005) Data mining of tree-based models to analyze freeway accident frequency. Journal of Safety Research.

8

Tseng, W., Nguyen, H., Liebowitz, J., & Agresti, W. (2005) Distractions and motor vehicle accidents: Data mining application on fatality analysis reporting system (FARS) data files. Industrial management and data systems, 109 (9), 1188-1205. Retrieved October 20, 2008, from Emerald Insight database. Journal of Safety Research.

9

Queensland University of Technology. (2008) Retrieved October 22, 2008, from QUT Centre for Accident Research and Road Safety: http://www.carrsq.qut.edu.au

record

number

of

road

crashes,

,

10 Hearst, M. A. (1999) Untangling Text Data Mining. The 37th Annual Meeting of the Association for Computational Linguistics, Maryland, June 20-26, (invited paper). http://www.ischool.berkeley.edu/~hearst/text-mining.html. Accessed 20 April 2007. 11 Jain, A.K., M.N. Murty, and P.J. Flynn, (1999) Data Clustering: A Review. ACM Computing Surveys (CSUR), 31(3): p. 264-323.

57

12 Grossman, D. & Frieder, O. (2004) Information Retrieval: Algorithms and Heuristics. 2nd edn., Springer. 13 Smith A. E. Humphreys, M. S. (2006) Evaluation of unsupervised semantic mapping of natural language with Leximancer concept mapping. http://www.leximancer.com/documents/B144.pdf Accessed 28 May 2007, 14 Han, J., & Kamber, M. (2001) Data Mining: Concepts and Techniques. San Diego, USA: Morgan Kaufmann

Acknowledgments We will like to thank CRC for Integrated Engineering Asset Management (CIEAM) to provide us the opportunity to conduct this case study. We will also like to thank students of ITB239 and ITN239: Enterprise Data Mining to conduct some of the experiments and Dan Emerson to assist us in reformatting the figures.

58

Proceedings of the 4th World Congress on Engineering Asset Management Athens, Greece 28 - 30 September 2009

SCHOOL BUILDINGS ASSETS – MAINTENANCE MANAGEMENT AND ORGANIZATION FOR VERTICAL TRANSPORTATION EQUIPMENTS Andrea Alonso Pérez a,b, Ana C. V. Vieira b,c and A. J. Marques Cardoso b a

b

Universidade de Vigo, ETSII, Campus Universitario Lagoas-Marcosende, C.P. 36310 Vigo, España.

Universidade de Coimbra, FCTUC/IT, Departamento de Engenharia Electrotécnica e de Computadores, Pólo II – Pinhal de Marrocos, P – 3030-290 Coimbra, Portugal. c

Instituto Politécnico de Tomar, Escola Superior de Tecnologia de Tomar, Departamento de Engenharia Electrotécnica, Estrada da Serra – Quinta do Contador, P – 2300-313 Tomar, Portugal.

Maintenance of educational building assets is an important tool not only for the wellness of students and other users, but also as an important economic instrument maximizing items life cycle and minimizing maintenance costs. The law in force regarding the Abolition of Architectural Barriers was conceived to facilitate the access to buildings of people with physical deficiencies, but even after its endorsement, there are still Portuguese Schools without an easy access or even none at all, for such individuals. The vertical transportation of persons helps to avoid this discrimination, as it provides access to all the building floors, with the highest possible comfort, for people whose mobility is reduced. This paper addresses the maintenance management and organization of Vertical Transportation Equipments. It will be proposed a standard example of a Maintenance Plan for the Portuguese Schools focusing on the elements and infrastructures related to vertical transportation of people with physical deficiencies. It will be also presented a cost analysis of these infrastructures maintenance and operation in order to reach the optimum point that allows the reduction of these costs without reducing the reliability of components, thus improving their useful life and safeguarding the wellness of users. Key Words: Costs evaluation, assets maintenance management, maintenance planning and scheduling 1

INTRODUCTION

For the particular case of school buildings, a great number of maintenance actions may be organized in a systematic way, with foreknowable costs and controlled funds. It is considered adequate to adopt preventive maintenance strategies, condition based or planned and scheduled in time. Although this might be an effective method for preserving schools, preventive maintenance is highly dependent on the availability of human resources and budgets. Preventive maintenance efforts range from visual inspection only, to performance testing and analysis; from minor adjustments, cleaning and/or lubrication, to complete overhauls; from reconditioning, to complete replacement. One must identify the adequate maintenance strategy to follow. Simultaneously, for each item a specific maintenance plan must be developed. The maintenance program may be divided into several lists of planned activities, allowing for unplanned reactive activities and deferred ones. The elevator life cycle is longer than other transport systems, reason why the design, operation, safety and accessibility can be delayed in relation with new technologies, making difficult the access of people with disability. For example, in the European Union there are more than four million of elevators, 50% of them having more than twenty five years, and they don’t have an appropriate level of safety and accessibility for nowadays [1].

59

The Portuguese Law 123/97 from the 22nd of May of 1997 enumerates the technical standards that regulate the buildings accessibility to people with disability [2]. A study developed in the framework of the Portuguese Secondary School Level of Education, the CARMAEE study [3], indicates that only 64.87% of the schools participating in the CARMAEE study, refer the amount of money expended in outsourcings, for example in maintenance of elevators and acclimatization systems. According to the same study, Portuguese schools have only adapted the building for the access of people with disability with the following elements and percentages [3]: - Elevators: 40.24% - Ramps: 71.60% - Wheelchair elevator platforms: 2.96% To accomplish the maintenance plan, it has to be considered a period of time that may, according to literature, allow to guarantee reliability of the elevator and safety of its users [2-8]. To settle on the maintenance strategy to apply, one must consider the procedures complexity, process time, and necessary technicians. According to literature, maintenance operations regarding vertical transportation of persons may be grouped in two different types. The first group includes maintenance activities that can be accomplished by an employee of the school due to the low incidence in the reliability of the equipment as well as users safety. This group of maintenance activities include ground cleaning and button illumination, for example. The second group of vertical transportation maintenance activities includes those activities that, according to current law, should be accomplished by certified companies, Elevators Maintenance Companies (EMC). These activities have a technical character and, since they have a high incidence in the element reliability and safety, they therefore should be part of the maintenance objectives. When developing the maintenance program, two different types of maintenance procedures where considered: the inspections and the maintenance actions procedures. By inspections, one means the procedures regarding maintenance activities aiming to verify the global state of operation of the elevator, and they can be partial inspections or tests. The inspections will be made in preset dates and, according to their technical complexity, they can be carried out by a school employee or a technician from the EMC. Maintenance actions procedures include operations released according to the inspections result, or due to a bad operation or damage of the equipment. As well as the inspections, they will be accomplished by a technician or an employee, depending on its technical complexity. The operations to be accomplished by the technicians of the EMC shall be performed according to a preventive maintenance plan, including procedures to be differentiated according to months, quarters, etc. On the other hand, the maintenance activities to be accomplished by school employees shall result from the following events: - At the beginning of the academic period: preventive maintenance. - During the academic period: preventive maintenance. - At the end of the academic period: preventive maintenance. - When there is a damage or bad operation: corrective maintenance. It is considered opportune to use as a general model the one of preventive maintenance, with the end of accomplishing the defined objectives, and to appeal to corrective maintenance when necessary. The maintenance plan is a very important tool since its periodicity will regulate the vertical transport function reverberating in the reliability and safety. There are several manuals, legislation, normative and technical documents, according to which it should be accomplished tests and inspections of the equipments following a stipulated periodicity, in order to increase equipments reliability and safety. The elevator is a total automatic equipment. The departure, acceleration and stop functions are automatically accomplished in agreement with the calls, without the necessity of a human command to work. Therefore, the accomplishment of regular inspections is indispensable to guarantee to users trips with safety and comfort. According to elevator manufacturers, elevators must be check-up every month to guarantee the users safety, the reliability of the equipment and its performance [4]. All maintenance reports must be kept near the equipment documents. These registers should be maintained actualized whenever there are modifications in the equipment characteristics and they should be available in the School Unit for those people in charge of the maintenance as well as for the organisms responsible for the accomplishment of tests and inspections.

60

2

MAINTENANCE PLAN OF VERTICAL TRANSPORT ELEMENTS 2.1 Maintenance activities to be accomplish by school employees

Schools that have between their staff, employees prepared to accomplish maintenance work, may organize these persons work to accomplish the proposed activities. This will allow them not only to optimize their productivity but also it will reduce costs, so important nowadays, since schools usually struggle with reduced budgets. Table 1 shows the activities to accomplish in each period of the year. Moreover, employees can also accomplish other daily recommendable inspections such as [6]: - Test if there are abnormal noises when the elevator is in movement. - Test that the platform moves without any interference. - Test that the bridge plate moves without any interference when it is folding or unfolding. - Test that the front barrier closes and blocks immediately after the platform ascends from the ground. Table 1 Maintenance activities to be accomplish by school employees [5] When to accomplish

Operations to accomplish

Before the academic period

-

During the academic period

- Coordinate the cabin and corridors cleaning. - Verify that there are no objects obstructing the accesses and -

After the academic period

circulations. Travel inside the elevator stopping in every floor. Test alarm button. Test the opening door button. Verify the telephone operation. Verify the equipment integrity.

- Send the elevator to the first floor. -

In case of damages or bad operation

Turn on the equipment according to manufacturer orientations. Travel inside the elevator stopping in every floor. Test alarm button. Test the opening door button. Verify the telephone operation. Turn on the lights of the floors and platforms of the elevators, also verifying if the access doors to the corridors are locked. Verify the equipment integrity.

-

Turn off the elevator in the command board, according to the manufacturer's orientation, and lock the board. Verify that the corridors elevator doors are locked, and turn off the corridor lights. Contact the EMC, requiring assistance. Place a poster indicating “Elevator in Maintenance”.

Reprinted from: School Elevators Usage and Maintenance (in Portuguese) 2005. Elevator damages and/or bad operation may usually be associated with the following scenarios [5]: - The minimum time of permanence of cabin door opening is less than 3 seconds. - The unevenness between the cabin floor and the pavement floor does not guarantee a safe and perfect accessibility. - There are buttons without light, do not obey the command or the elevator doesn't assist the call in the pavement. - There is an alarm information in the cabin or pavement panels. - There is cabin swinging, vibrations or noises in excess during the transport.

61

Cleaning must have a special mention since it is usually outsourced, despite its importance to avoid accidents ensuring users safety. Water infiltration in the facilities is harmful to the equipment, therefore it must always be kept dry: cabin, ground and walls of the elevator. This procedure is also important in order to avoid slipping and possible wounds to users. Special careful should be taken in the cabin interior cleaning, following the orientation of the manual supplied by the elevator manufacturer. As much for the finish in stainless steel as for the laminated one it must be used soap or neutral detergent and water, applied with cloth or sponge. Aggressive chemical products or abrasive material, as straw-of-steel should never be used [6].

2.2 Maintenance Plan to be accomplished by certified Elevators Maintenance Companies (EMC) The legislation that regulates elevators maintenance in Europe is scarce. With regard to school facilities for example it only states the necessity for inspection every two years by competent entities. On the other hand, manufacturers and other entities offer several recommendations. Although the legislation on safety inspections in elevators may be different according to the locality, country, etc, technicians usually employ similar procedures [7]. In the hoist way, in the ditch, or on top of the cabin, technician must organize the inspection in a safe and ordinate way to avoid injuries not only for himself but also for people occupying the immediately adjacent areas. To accomplish this work, it is necessary to avoid wide clothes. Technician must be equipped with safety glasses or protector glasses in the hoist way, since the airborne particles swept there might cause eye injury. Minimum tools include a flashlight, a 183 cm ruler of wood, a magnifying glass, a small mirror, and a multimeter to measure voltages and grounds [7]. The minimum specific maintenance for all traction elevators may be accomplished according to the periodicity and procedures presented in Table 2:

Table2 Maintenance plan to accomplish by the EMC [7, 8] Periodicity

Procedures to accomplish

Actions according to inspection’s result

Weekly

Perform general inspection of machinery, sheaves, worm and gear motor, brake and selector of floor controllers.

Lubricate as necessary.

Empty drip pans, discard oil in an approved manner and check reservoir oil level. Observe brake operation.

Adjust or repair if required.

Inspect machinery, contacts, linkage and gearing.

Lubricate.

Inspect brushes and switches.

Clean and repair.

Inspect controllers, selectors, relays, connectors, contacts, etc.

Clean and repair.

Ride car and observe operation of doors, levelling, reopening devices, pushbuttons, lights, etc. If rails are lubricated, check conditions and lubrication service lubricators. Verify lamps in elevator cars, machine room, pit, hall lanterns, etc.

Replace all burned out.

Inspect all machine room equipment.

Remove garbage, dust, oil, etc.

Clean trash from pit and empty drip pans. Check condition of car switch handle.

Replace emergency release glass.

Check governor and tape tension sheave lubrication.

Lubricate.

Verify lamps in all lanterns, push buttons, car and corridor position indicators, direction indicators and in other signal fixtures.

Replace all burned out.

62

Table2 Maintenance plan to accomplish by the EMC [7, 8] Periodicity

Procedures to accomplish

Bi-monthly

Observe operation of elevator throughout its full range of all floors. This procedure serves to test controls, safety devices, levelling, relieving, and other devices.

Quarterly

Actions according to inspection’s result

Check door operation: brakes, checks, linkages, gears, wiring motors, check keys, set screw, contacts, chains, cams and door closer.

Clean, adjust and lubricate.

Check selector, brushes, dashpots, travelling cables, chain, pawl magnets, wiring, contacts, relays, tape drive and broken tape switch.

Clean, adjust and lubricate.

Check car: car door and gate tracks, pivots, hangers, car grill, side and top exits.

Clean, adjust and lubricate.

Inspect interior of cab: test intercommunication system, normal and emergency lights, fan, emergency call system or alarm, car station.

Repair.

Visually inspect controller, contacts and relays.

Adjust or replace.

Observe operation of signal and dispatching system.

Repair.

Inspect compensating hitches, buffers, rope clamps, slack cable switch, couplings, keyways and pulleys. Check load weighing device and dispatching time settings.

Clean, adjust, repair, lubricate and replace.

Check oil level in car and counterweight oil buffers.

Add oil as required.

Check brushes and commutators. Inspect switches for finish, grooving, eccentricity and mica level.

Clean, adjust, repair, replace or refinish to provide proper commutation.

Inspect brushes for tension seating and wear.

Replace or adjust.

Check car ventilation system, car position indicators, direction indicators, hall lanterns and car and hall buttons.

Replace or adjust.

Check levelling operation: levelling switches, hoist way vanes, magnets, and inductors.

Clean, adjust and repair.

Check hoist way doors: car door or gate tracks, hangers, and up thrust eccentrics, linkages jibs and interlocks.

Clean and lubricate.

Check car door or gate tracks, pivots, hangers. On hoist ways doors: tracks, hangers and eccentrics, linkages jibs and interlocks.

Clean, adjust and lubricate.

Inspect all fastening and ropes for wear and lubrication: governors and hoist ropes. Inspect all rope hitches and shackles and equalize rope tension.

Clean, lubricate, balance the tension rope, repair.

Inspect hoist reduction gear brake and brake drum, drive sheave and motor, and any bearing wear. In the car, test alarm bell system: fixtures, retiring cam devices, chain, dashpots, commutators, brushes, cam pivots, fastenings.

Clean and adjust.

Test emergency switch: Inspect safety parts, pivots, setscrew, switches, adjustment of car and counterweight jibs, shoe or roller guides.

Replace, lubricate and repair.

In the pit, verify compensating sheave and inspect hitches.

Lubricate.

Inspect governor and tension sheave fastenings.

Tape tension sheave fastenings

63

Table2 Maintenance plan to accomplish by the EMC [7, 8] Periodicity

Procedures to accomplish

Actions according to inspection’s result

Quarterly

Verify oil drip pans.

Clean and empty.

Verify all parts of safeties and moving parts.

Clean and lubricate.

Check clearance between safety jaws and guide rails.

Adjust.

Visually inspect all safety parts. SemiAnnually

Annually

Examine governor rope.

Clean and replace.

Check controller, alignment of switches, relays, timers, contacts, hinge pins, etc.

Clean with blower, adjust and lubricate.

Check all resistance tubes and grids, oil in overload relays, settings and operation of overloads.

Lubricate and adjust.

Inspect fuses and holders and all controller connections. In hoist way examine guide rails, cams and fastenings.

Clean.

Inspect and test limit and terminals switches.

Replace or repair.

Check car shoes, jibs or roller guides.

Adjust or replace.

Check all overhead cams, sheaves, sills, bottom of platform, car tops, counterweights and hoist way walls.

Clean.

Inspect sheaves to ensure they are tight on shafts. Sound spokes and rim with hammer for cracks.

Adjust or repair.

Examine all hoist ropes for wear, lubrication, length and tension.

Replace, lubricate and adjust.

On tape drives, check hitches and broken tape switch.

Replace or repair.

Check car stile channels for bends or cracks; also car frame, cams, supports and car steadying plates.

Replace or repair.

Examine moving parts of vertical rising or collapsible car gates. Check pivot points, sheaves, guides and track wear.

Lubricate and replace.

Inspect guide shoe stems.

Lubricate and replace.

Check governor and tape tension sheave fastenings.

Tape tension sheave fastenings

For bi-parting doors, check: chains, tracks and sheaves, door contacts.

Clean, lubricate, repair or replace.

Clean car and counterweight guide rails using a non-flammable or high flash point solvent to remove lint dust and excess lubricant. Examine brake cores on brakes, linings, and inspect for wear.

Remove, clean, and lubricate. Correct excess wear and adjust.

Examine reservoirs of each hoisting motor and motor generator. Drain, flush and refill. Check all brushes for neutral settings, proper quartering and spacing on commentators.

Restore.

Group supervisory control systems installed must be checked out. The systems, dispatching scheduling and emergency servicing must be tested and adjusted in accordance with manufacturer’s literature. Reprinted from: Procurement Services Group: Elevator Maintenance, 2000 and Maintenance Engineering Handbook, 2002.

64

Elevators maintenance must consider three different areas: hoist way, pit and machine-room [7]. - Hoist way: • Mechanical equipment: sheaves, buffers, door closers, floor selectors, limit switches, hoist way door hangers, closers, and door gibs, interlocks, to be sure they are in the right form. • Hoist and governor ropes and their fastenings, to avoid wear and rust. • Travelling cables to verify its state, and cables for vibration and wear or tear. • Rails for a correct alignment. • Steadying plates for excessive fluctuations. - Pit: oil level in the buffer, rope stretch, debris or water leaks and safety's shoes. - Machine-room: • Motors and generators should be clean, undercut and conditioned correctly, and brushes verified. • Bearings must be inspected for wear or tear. • Electric equipment must be grounded. • Brakes and brake belts, to avoid defective safety. • Shafts into the pulley, for a correct alignment. • Gears and bolts must be inspected to ensure that they are not loose or broken. • Controllers, for a correct operation. • Switches. • Safety equipment, for blocked or shorted contacts. • Governors, to certify that the rope is appropriately placed in the sheave, correctly lubricated and that the rust don’t affect its operation. • Landing equipment, to verify if broken buttons exist and if the illumination is correct. Maintenance operations to accomplish in case of hydraulic elevator would be the same ones that in the case of a traction elevator, but in addition it must be considered [7]: - Packing of the superior part of the cylinder and piston, to make sure that the oil is not draining excessively or it is returning to the tank through a deviation. - Hydraulic machines, to be sure that there is enough oil so that the car still reaches the top of the landing with oil left in the tank. - Load oil average, for cleaning. It is recommended that safety tests should be conducted by the elevator manufacturer. In addition to ensuring compliance with safety regulations, inspections contribute to equipment’s life extension by preventive care and adjustment. A comprehensive elevator maintenance program may also include a full-load safety test, to be carried out every 5 years, and annual no-load tests [7]. 3

MAINTENANCE CONTRACT

As previously stated, school organizations must decide how to organize the elevator maintenance, whether or not trying to accomplish internal maintenance or just to hire the maintenance services of an independent company as well as those offered by the company that manufactured and installed the elevators [5, 7]. EN 13269:2006 enunciates the instructions for maintenance contracts preparation. Maintenance contracts may be differentiated by the offered coverage, namely total maintenance or partial/conservative maintenance [9]. Partial or conservative maintenance is the simplest form of contractual service. This type of maintenance contracts covers lubrication and minor cleaning of elevator equipment and they usually do not include parts replaced during maintenance activities nor overtime emergency calls. The maintenance agreement limits the responsibility of the contractor and, while protecting the elevator contractor, they are potentially costly to the owner. Contract prices shall be adjusted every year according to the changes in the labour costs [7]. Total maintenance contracts usually cover both parts and work demanded to maintain the elevator system. Emergency calls service may be or not included, depending on individual needs. The exclusions in contracts of total maintenance are normally less than in the partial/conservative maintenance contracts and the contractor party's responsibility is larger since it is responsible for the preventive maintenance. Whenever a total maintenance contract excludes mandatory safety tests; the school administration must provide trained employees to shut down the equipment in case of unsafe conditions development. These items omission by the contractor increases the maintenance cost supported by the owner [5, 7]. Contract prices and services vary a lot from company to company, depending on the contract type and services required. Table 3 presents some considerations about procedures to be followed before making a maintenance contract or when revising it [5].

65

Table 3 Advices on how to choose a maintenance contractor and procedures to adopt during contract’s life [5] Terms to consider when choosing a maintenance contractor

Procedure to adopt during contract’s life

Demand the company’s certificate of register in the regulating agency, as well as the document corroborating that the responsible engineer is properly registered.

Demand the technician's functional identity.

Verify if the company has vehicle, telephone, workshop, and parts of the installed equipment.

Control frequency and visiting hours.

Verify if the contract includes 24 hours attendance.

Demand a copy of the record service filled out and signed by the responsible.

Verify the references and services given to other customers.

Demand receipt, guarantee of use of original parts and the executed services.

Reprinted from: Foundation for the Education Development, 2005.

Together with the maintenance contract the following documents and information have to be provided in order to be able, thus, to have an easy access when it is necessary: - The date when the equipment was put in service. - Basic characteristics. - Traction handles’ characteristics. - Characteristics of the parts for which it was asked for inspection certificate. - Designs of the installation. - Schematic diagrams of the electric circuits. - Licenses of installation and operation (or documentation that substitutes them, according to legislation). - Technical guarantee of the equipment. - Maintenance contract. - Technical Dossier of the equipment (supplied by the manufacturer and/or constructor). - Manuals of the equipment (of the manufacturer). - Maintenance manual of the equipment. - Inspections, rehearsals and verifications. 4

MAINTENANCE COSTS Nowadays the maintenance costs depend on several factors like: - Charge capacity - Velocity - Number of floors - Automatic/semi-automatic doors - Kind of work: universal, with memory, etc. - Type of maintenance - 24 hours service

In the following it is presented an example of an annual maintenance cost for a standard elevator, provided by an elevator company. - Total Maintenance: € 2 400.00 - Partial Maintenance: € 1 056.00 - 24 hours Service: € 180.00 There is a big difference between the total and the partial maintenance cost. The school management executive board has to consider both options and choose the one that better fits their maintenance objectives.

66

By comparing this information with the one from CARMAEE Project it can be concluded that the average cost of € 1 679.05 for the elevator maintenance of Portuguese schools is within the limits provided above. 5

CONCLUSIONS

There are more than four million of elevators in the E.U., 50% of them having more than twenty five years, and they don’t have an appropriate level of safety and accessibility for nowadays. The elevators life cycle is longer than other transport systems, reason why the design, operation, safety and accessibility can be delayed in relation with new technologies, making difficult the access of people with disability. This paper proposes a Maintenance Plan for the Portuguese schools vertical transport equipment with the aim to provide some general guidelines for school management executive boards.

6

REFERENCES

1

Technical Specification CEN/TS 81-82:2008 (2008) Safety rules for the construction and installation of lifts. Existing lifts. Part 82: Improvement of the accessibility of existing lifts for persons including persons with disability, AENOR, Madrid.

2

Ministry of Solidarity and Social Security (1997) Law nº123/97 from the 22nd of May of 1997 (in Portuguese), Diary of the Republic nº 118 the 22nd of May of 1997, Portugal.

3

Cação, C.; Silva, F. and Ferreira, H. (2004) CARMAEE: Characterization of the Maintenance in School Buildings (in Portuguese), Final Year Project Dissertation; Department of Electrical and Computer Engineering, University of Coimbra, Portugal.

4

ThyssenKrupp Elevators (2005) Your Elevator: Preventive Maintenance (in Portuguese), 12th edition.

5

Foundation for the Education Development (2005) School Elevators Usage and Maintenance (in Portuguese), São Paulo.

6

Phantom, Phantom Elevator Maintenance Guidelines (in Spanish).

7

Robertson, J. (2002). Maintenance of Elevators and Special Lifts. In: L. R. Higgins and R. K. Mobley eds. Maintenance Engineering Handbook, 6th edt, McGraw-Hill, United States of America.

8

New York State Office of General Services (2000) Procurement Services Group: Elevator Maintenance, New York.

9

European Committee of Standardization CEN/TC 319 (2006) EN 13269: Maintenance. Guidelines on Preparation of Maintenance Contracts, Brussels.

67

Proceedings of the 4th World Congress on Engineering Asset Management Athens, Greece 28 - 30 September 2009

A REVIEW ON THE OPTIMISATION OF AIRCRAFT MAINTENANCE WITH APPLICATION TO LANDING GEARS P. Phillips¹, D. Diston¹, A. Starr², J. Payne³ and S. Pandya³ ¹School of Mechanical, Aerospace and Civil Engineering, University of Manchester, Manchester, UK ²School of Aerospace, Automotive and Design Engineering, University of Hertfordshire, Hatfield, UK ³Messier Dowty Ltd, Gloucestershire, UK Current maintenance programmes for key aircraft systems such as the landing gears are made up of several activities based around preventive and corrective maintenance scheduling. Within today’s competitive aerospace market innovative maintenance solutions are required to optimise aircraft maintenance, for both single aircraft and the entire fleet, ensuring that operators obtain the maximum availability from their aircraft. This has led to a move away from traditional preventive maintenance measures to a more predictive maintenance approach, supported by new health monitoring technologies. Future aircraft life will be underpinned by health monitoring, with the ability to quantify the health of aerospace systems and structures offering competitive decision-making advantages that are now vital for retaining customers and attracting new business. One such aerospace system is the actuator mechanisms used for extension, retraction and locking of the landing gears. The future of which will see the introduction of electromechanical replacements for the hydraulic systems present on the majority of civil aircraft. These actuators can be regarded as mission critical systems that must be guaranteed to operate at both take-off and landing. The health monitoring of these actuation systems can guarantee reliability, reduce maintenance costs and increase their operational life span. Aerospace legislation dictates that any decisions regarding maintenance, safety and flight worthiness must be justified and strict procedures followed. This has inevitably led to difficulties in health monitoring solutions meeting the necessary requirements for aerospace integration. This paper provides the motivation for the research area through reviewing current aircraft maintenance practices and how health monitoring is likely to play a future strategic role in maintenance operations. This is achieved with reference to current research work into developing a health monitoring system to support novel electromechanical actuators for use in aircraft landing gears. The difficulties associated with integrating new health monitoring technology into an aircraft are also reviewed, with perspectives given on the reasons for the current slow integration of health monitoring systems into aerospace. Keywords: Maintenance management, health monitoring, actuators, landing gear 1

INTRODUCTION

The airline industry is considered as one of the most unique businesses in the world. The business suffers from a variety of complex operations. These include moving aircraft loaded with passengers and cargo over large distances and the scheduling of flights, crews and maintenance. These all lead up to substantial operating and maintenance costs measured in time and money. Aircraft maintenance forms an essential part of airworthiness, with its main objective being to ensure a fully serviced, operational and safe aircraft. If an aircraft is not maintained to a required level then this inevitably risks passenger and crew safety. Table 1 lists examples of incidents that have occurred with the probable cause being due to insufficient maintenance [1]. Also, the risk is that the aircraft may be unable to take-off leading to passenger dissatisfaction; likewise it is plausible that the aircraft may be forced to land in undesirable locations, where spare parts or maintenance expertise is unavailable. Maintenance actions therefore have to be carried out at regular scheduled intervals, but ideally be performed with minimum cost to the operator.

68

Table 1: Aircraft maintenance related accidents Airline Aloha Airlines 737

Location Hawaii

Year 1988

Incident Inspection failure led to fuselage failure

BM AirTours 737

Manchester

1989

Wrong bolts led to windshield blowout

United Airlines DC10

Iowa

1989

Continental Express

Texas

1991

Engine inspection failure led to loss of systems Tail failure as task not completed before flight

Northwest Airlines

Tokyo

1994

Incomplete separation

ValueJet

Florida

1996

Fire in hold due to incendiary cargo

assembly

led

to

engine

One of the key systems which have to be maintained and kept fully operational is the aircraft landing gears. Landing gear’s are an essential part of any aircraft; even though they remain redundant for most of the flight. Their main task is to absorb the horizontal and vertical energy of the aircraft as it touches down on the runway. During flight most modern aircraft have their landing gears retracted and stowed and only extend them during the approach to landing. Aircraft extend and retract their landing gears using a variety of methods, which include pneumatic, hydraulic or electrical motor-driven drives, with the majority of retraction mechanisms being hydraulically powered. Most landing gears contain three actuators; the largest of which is the retraction actuator that generates a force about a pivotal axis in order to raise the landing gear against weight and aerodynamic loads. The other two actuators are the lock-stay actuator, which locks the landing gear in place once extended and the door actuator that ensures that the bay doors are successfully opened and closed for landing gear deployment. Figure 1 shows a typical arrangement of the down lock and main retraction actuator positions.

Figure 1: Airbus A320 main gear Hydraulic actuation systems have found popular use in aerospace due to their reliability, relative simplicity and their wide spread use has generated engineering experience and familiarity. They are also ideally suited for landing gear operation as the hydraulic fluid provides a constant lubrication and natural damping. There are disadvantages also, when used in aircraft they are heavy, require large volumes of space, operate noisily and require the correct disposal of hydraulic fluids in accordance with environmental legislation There is currently a move within the aerospace industry towards replacing hydraulic drives with electrical counterparts as part of the ‘more electric aircraft’ concept [2]. The motivation for the use of Electro-Mechanical

69

Actuators (EMA) is driven by the desire to reduce aircraft weight arising from a combination of increasing fuel costs and environmental concerns. For example, environmental damage associated with air traffic has created the need to reduce aircraft fuel consumption and polluting emissions, a key factor in achieving this is the reduction of weight. Landing gears contribute a significant amount of mass to the aircraft, on average they contribute 4% on civil aircraft and 3% on military aircraft [3]. The European drive for the replacement of hydraulic systems on landing gears is driven in part by a large DTI funded project known as ELGEAR [4]. The aim of which is to demonstrate the potential to reduce operating noise, increase operating efficiency, reduce installation volumes and most significantly reduce the landing gear mass by up to 12%. A further motivation for utilising electrically powered actuators is the real possibility that with the move towards optimising engine efficiency, future aircraft engines will not produce hydraulic power. Innovative maintenance solutions, such as health monitoring systems are required to support the introduction of the new electrical actuator technology to provide reliability assurances. Health monitoring systems will enable decisions to be taken regarding aircraft flight worthiness. Importantly for aircraft operators it can aid in providing a unique optimised maintenance scheduling allowing maintenance decisions to play a key part in driving forward operations strategies and providing a business winning advantage This paper will provide the motivation and justification for the current research area of electromechanical actuator health monitoring for landing gears. This is done through reviewing the current maintenance practice for aircraft with reference to landing gears, and what path future advanced maintenance solutions are likely to follow. How health monitoring is likely to play a strategic operations role offering benefits to operators, manufactures and maintenance service providers is also highlighted. The difficulties associated with integrating new health monitoring technology into an aircraft are reviewed with perspectives given on the reasons for the current slow integration of health monitoring systems into aerospace. 2

ELECTROMECHANICAL RETRACTION ACTUATOR DESIGN

The design for the retraction actuator is based around that of a roller screw. A roller screw converts the rotary torque produced by the motor into a linear motion. Roller screws consist of a multiple arrangement of threaded helical rollers assembled in a planetary arrangement around a screw captured in place by a nut. Rotation of the nut with respect to the shaft enables axial translation of the nut; likewise rotation of the screw also can enable translation of the nut. A primary duplex motor connected to a gearbox linearly displaces the nut by rotation of the screw, which moves a lever arm about a pivot achieving retraction/extension of the landing gears. If the primary motor fails there is an emergency control that will ensure successful displacement of the actuator. Figure 2 provides a schematic of the actuator arrangement.

Brake

Emergency Gear Box

Primary Gear Box

Duplex Motor

Emergency Motor Primary Roller Screw

Emergency Roller Screw

Figure 2: The EMA retraction actuator Hydraulic actuation has been used in aerospace successfully for many decades, proving to be reliable and robust, gaining the confidence of aircraft operators. Any replacement drive will therefore need to provide assurances that they are of equal robustness and reliability to the preceding system [5]. The replacement of a key aircraft system, such as actuators, will inevitably require changes in the way in which maintenance actions are performed. As an example, it is easy too visually inspect a hydraulic actuator for faults such as fluid leaks or corrosion. An EMA, such as the schematic in figure 2, is a complex mechanical system and it is not so easy to inspect individual subsystems and components. Most of theses key components (i.e. gears, electrical wiring) are often sealed within housing units making access exceptionally difficult. The use of traditional visual inspections would therefore require a certain degree of landing gear dismantling. The move towards ‘more electric aircraft’ could therefore increase the time an aircraft spends in the maintenance hangar, increasing aircraft downtime costs. This aids in justifying the need for additional automated fault detection and diagnostic health monitoring incorporated into the design

70

3

CURRENT MAINTENANCE PRACTICE

Maintenance programmes for aircraft, which include key systems such as the engines and landing gears are made up of several activities based around preventive, corrective, on-condition and redesign maintenance. Preventive actions are taken at pre-determined intervals based upon the number of operating hours, or often in the case of landing gears, the number of landings. This is supported by regularly scheduled inspections and tests in which on-condition maintenance is performed based upon observations and test results. Each of these activities is finally supported by corrective maintenance conducted in response to discrepancies or failures within the aircraft during service. The final action type, redesign maintenance takes the form of engineering modifications that are made in order to address arising safety or reliability issues, which were unanticipated in the original design. It is essential that aircraft maintenance be performed at appropriate times and to the highest standard to ensure system reliability and guaranteeing passenger safety. So that any potential unsafe conditions can be identified and addressed, the country of aircraft registration and the civil aviation authority of the manufacturing country, generate a set of mandatory guidelines known as airworthiness directives. These directives notify the aircraft operators that their aircraft may not conform to the appropriate standards and if there are any actions (i.e. maintenance) that must be taken. It is a legal requirement that operators follow the airworthiness directives and country specific authorities closely regulate them. Such authorities include the Federal Aviation Administration (USA), The Civil Aviation Safety Authority (Australia) and The Joint Aviation Authorities (Europe). Much of the major maintenance and repair work performed on aircraft is provided through service providers who carry out Maintenance, Repair and Overhaul (MRO) operations for the aircraft operators. The landing gear is a critical assembly and a major key to maintaining the overall aircraft value. Operators cannot afford, or are willing to risk compromising their landing gear MRO activities and will look for the best combination of affordability, expertise, flexibility and the ability to offer customised solutions when faced with the choice of MRO provider. An example of how maintenance support of landing gears would be as follows. In the event of a series of incidents such as ‘hard landings’ reported by the operators, major repair operations, or complete gear overhauls will be conducted at a MRO provider’s maintenance site. The operators themselves can carry out, minor repairs and on-wing maintenance, also at predetermined intervals. Once landing gears have been received at the MRO maintenance facility, they will be dismantled and individual parts will be put through a serious of non-destructive tests. This testing will identify any developing failures, such as structural fatigues or internal corrosion. The results of which will determine if the parts are repaired, replaced, scrapped or recycled [6]. There are a vast number of parts on a typical landing gear which need to be maintained and inspected, An example of key inspection areas along with typical timescales would be: 1.

After 300 hours or after 1 year in service inspection -

2.

Shock absorber Nitrogen Pressure check

After 600 hour inspections -

Landing gear hinge points visual inspections

-

Leak inspection (oil, hydraulic fluid etc)

-

Inspection of torque link play

3. After 7 years or 5000 cycles : Landing gear overhaul To understand maintenance costs it is necessary to look at the elements of maintenance in terms of time. Figure 3 gives a breakdown of the time elements covering maintenance actions. A breakdown such as this can show designers the areas in which they can influence related activity times. In corrective maintenance much of the time is spent on locating a defect which often requires a sequence of disassembly and reassembly. Being able to predict fault location times is extremely difficult using traditional inspection techniques. The ability to automate this fault diagnosis, with advanced technologies and techniques, can help accurately predict the downtime required [7].

71

Time

Up Time

Flying Time

Down Time

Available for Flying Time

Flight Prep. Time

Maintenance Time

Turn Around Time

Pre-flight inspection Time

Preventive Maintenance

Access Time

Modification Time

Corrective Maintenance

Inspection Time

Preparation Time

Defect Location Time

Defect Rectification Time

Rectify by adjustment time In-situ repair time Remove repair & refit time Remove and replace time

BITE effectiveness Fault diagnostic aids Equipment test/Read out capability Technician skill, experience & training

Figure 3: Civil aircraft maintenance time relationships 4 CHANGING MAINTENANCE STRATEGIES Currently the European market holds a 26% share of the worldwide MRO business compared to 39% held by North America and is expected to experience dramatic world wide growth during the next 10 years [8]. There are however several hurdles which must be overcome by these MRO providers in order to continue their leading global market shares [9]. Examples of which include:


Growing competition from the Middle East




Greater competition from Original Equipment Manufacturers (OEM)




Continuing pressure from airlines to reduce costs

These hurdles coupled with increased demand for airline MRO are forcing changes in the global aviation maintenance industries, including:


MRO providers are expanding their geographical reach and capabilities in a bid to become regional and global full service providers.




Spending on MRO is expected to universally increase

72




Airlines are now seeking how to make the next level of savings, which has raise the demand for more predictive maintenance strategies, with more reliability and material solutions to compliment outsourced maintenance repair work.




To drive further cost reductions, airlines are seeking to incorporate sophisticated maintenance management solutions into their aircraft, reducing investments in inventory and to aid in improvements in airline operations and reliability.

Such factors have begun to dictate a change in maintenance strategy for operators and the solutions in the services