Sổ tay bảo trì hệ thống phức tạp

x Contents Chapter 8: Preventive Maintenance Models for Complex Systems David F.. Part A: An Overview Chapter 1: An Overview Khairy Kobbacy and Pra Murthy Part B: Evolution of Concepts a

Springer Series in Reliability Engineering

Professor Hoang Pham

Department of Industrial Engineering

Other titles in this series

The Universal Generating Function in Reliability Analysis and Optimization

Gregory Levitin

Warranty Management and Product Manufacture

D.N.P Murthy and Wallace R Blischke

Maintenance Theory of Reliability

Toshio Nakagawa

System Software Reliability

Hoang Pham

Reliability and Optimal Maintenance

Hongzhou Wang and Hoang Pham

Applied Reliability and Quality

B.S Dhillon

Shock and Damage Models in Reliability Theory

Toshio Nakagawa

Risk Management

Terje Aven and Jan Erik Vinnem

Satisfying Safety Goals by Probabilistic Risk Assessment

Hiromitsu Kumamoto

Offshore Risk Assessment (2nd Edition)

Jan Erik Vinnem

The Maintenance Management Framework

Adolfo Crespo Márquez

Human Reliability and Error in Transportation Systems

B.S Dhillon

Khairy A.H Kobbacy • D.N Prabhakar Murthy

Editors

Complex System

Maintenance Handbook

123

Springer Series in Reliability Engineering series ISSN 1614-7839

British Library Cataloguing in Publication Data

A Complex system maintenance handbook - (Springer series in

Library of Congress Control Number: 2008923781

© 2008 Springer-Verlag London Limited

Watchdog Agent™ is a trademark of the Intelligent Maintenance Systems (IMS) Center, University of Cincinnati, PO Box 210072, Cincinnati, OH 45221, USA www.imscenter.net

Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act 1988, this publication may only be reproduced, stored or transmitted, in any form or by any means, with the prior permission in writing of the publishers, or in the case

of reprographic reproduction in accordance with the terms of licences issued by the Copy-right Licensing Agency Enquiries concerning reproduction outside those terms should be sent to the publishers

The use of registered names, trademarks, etc in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant laws and regulations and therefore free for general use

The publisher makes no representation, express or implied, with regard to the accuracy of the information contained in this book and cannot accept any legal responsibility or liability for any errors or omissions that may be made

Cover design: deblik, Berlin, Germany

Printed on acid-free paper

9 8 7 6 5 4 3 2 1

springer.com

To our wives Iman and Jayashree for their patience, understanding and support

Preface

Modern societies depend on the smooth operation of many complex systems (designed and built by humans) that provide a variety of outputs (products and services) These include transport systems (trains, buses, ferries, ships and aero-planes), communication systems (television, telephone and computer networks), utilities (water, gas and electricity networks), manufacturing plants (to produce in-dustrial products and consumer durables), processing plants (to extract and process minerals and oil), hospitals (to provide services) and banks (for financial trans-actions) to name a few

Every system built by humans is unreliable in the sense that it degrades with age and/or usage A system is said to fail when it is no longer capable of delivering the designed outputs Some failures can be catastrophic in the sense that they can result in serious economic losses, affect humans and do serious damage to the environment Typical examples include the crash of an aircraft in flight, failure of a sewerage processing plant and collapse of a bridge The degradation can be con-trolled, and the likelihood of catastrophic failures reduced, through maintenance actions, including preventive maintenance, inspection, condition monitoring and design-out maintenance Corrective maintenance actions are needed to restore a failed system to operational state through repair or replacement of the components that caused the failure

Maintenance has moved from being an engineering activity after a system has been put into operation into an important issue that needs to be addressed during the design and manufacturing or building of the system Maintenance impacts on reliability (a technical issue) with serious economic and commercial implications This implies that operators of complex systems need to look at maintenance from

an overall business perspective that integrates the technical and commercial issues

in an effective manner

The literature on maintenance is vast Over the last 50 years, there have been dramatic changes due to advances in the understanding of the physics of failure, in technologies to monitor and assess the state of the system, in computers to store

viii Preface

and process large amounts of relevant data and in the tools and techniques needed

to build model to determine the optimal maintenance strategies

The aim of this book is to integrate this vast literature with different chapters focusing on different aspects of maintenance and written by active researchers and/or experienced practitioners with international reputations Each chapter re-views the literature dealing with a particular aspect of maintenance (for example, methodology, approaches, technology, management, modelling analysis and opti-misation), reports on the developments and trends in a particular industry sector or, deals with a case study It is hoped that the book will lead to narrowing the gap between theory and practice and to trigger new research in maintenance

The book is written for a wide audience This includes practitioners from try (maintenance engineers and managers) and researchers investigating various aspects of maintenance Also, it is suitable for use as a textbook for postgraduate programs in maintenance, industrial engineering and applied mathematics

indus-We would like to thank the authors of the chapters for their collaboration and prompt responses to our enquiries which enabled completion of this handbook on time We also wish to acknowledge the support of the University of Salford and the award of CAMPUS Fellowship in 2006 to one of us (PM) We gratefully acknowl-edge the help and encouragement of the editors of Springer, Anthony Doyle and Simon Rees Also, our thanks to Sorina Moosdorf and the staff involved with the production of the book

Chapter 2: Maintenance: An Evolutionary Perspective

L Pintelon and A Parodi-Herz 21Chapter 3: New Technologies for Maintenance

Jay Lee and Haixia Wang 49

Chapter 4: Reliability Centred Maintenance

Marvin Rausand and Jørn Vatn 79 Part C Methods and Techniques

Chapter 5: Condition-based Maintenance Modelling

Wenbin Wang 111Chapter 6: Maintenance Based on Limited Data

David F Percy 133Chapter 7: Reliability Prediction and Accelerated Testing

E A Elsayed 155

x Contents

Chapter 8: Preventive Maintenance Models for Complex Systems

David F Percy 179Chapter 9: Artificial Intelligence in Maintenance

Khairy A H Kobbacy 209 Part D Problem Specific Models

Chapter 10: Maintenance of Repairable Systems

Bo Henry Lindqvist 235Chapter 11: Optimal Maintenance of Multi-component Systems: A Review

Robin P Nicolai and Rommert Dekker 263

Chapter 12: Replacement of Capital Equipment

P.A Scarf and J.C Hartman 287Chapter 13: Maintenance and Production: A Review of Planning Models

Gabriella Budai, Rommert Dekker and Robin P Nicolai 321

Chapter 14: Delay Time Modelling

Wenbin Wang 345 Part E Management

Chapter 15: Maintenance Outsourcing

D.N.P Murthy and N Jack 373

Chapter 16: Maintenance of Leased Equipment

D.N.P Murthy and J Pongpech 395Chapter 17: Computerised Maintenance Management Systems

Ashraf Labib 417

Chapter 18: Risk Analysis in Maintenance

Terje Aven 437Chapter 19: Maintenance Performance Measurement (MPM) System

Uday Kumar and Aditya Parida 459Chapter 20: Forecasting for Inventory Management of Service Parts

John E Boylan and Aris A Syntetos 479

Part F Applications (Case Studies)

Chapter 21: Maintenance in the Rail Industry

Jørn Vatn 509

Chapter 22: Condition Monitoring of Diesel Engines

Renyan Jiang, Xinping Yan 533

Chapter 23: Benchmarking of the Maintenance Process at Banverket

(The Swedish National Rail Administration)

Ulla Espling and Uday Kumar 559

Chapter 24: Integrated e-Operations–e-Maintenance: Applications in North Sea

Offshore Assets

Jayantha P Liyanage 585

Chapter 25: Fault Detection and Identification for Longwall Machinery

Using SCADA Data

Daniel R Bongers and Hal Gurgenci 611

Contributor Biographies 643

Index 653

Part A

An Overview

1

An Overview

K.A.H Kobbacy and D.N.P Murthy

K Kobbacy and D Murthy

1.1 Introduction

The efficient functioning of modern society depends on the smooth operation of many complex systems comprised of several pieces of equipment that provide a variety of products and services These include transport systems (trains, buses, ferries, ships and aeroplanes), communication systems (television, telephone and computer networks), utilities (water, gas and electricity networks), manufacturing plants (to produce industrial products and consumer durables), processing plants (to extract and process minerals and oil), hospitals (to provide services) and banks (for financial transactions) to name a few All equipment is unreliable in the sense that it degrades with age and/or usage and fails when it is no longer capable of delivering the products and services When a complex system fails, the conse-quences can be dramatic It can result in serious economic losses, affect humans and do serious damage to the environment as, for example, the crash of an aircraft

in flight, the failure of a sewage processing plant or the collapse of a bridge

Through proper corrective maintenance, one can restore a failed system to an operational state by actions such as repair or replacement of the components that failed and in turn caused the failure of the system The occurrence of failures can

be controlled through maintenance actions, including preventive maintenance, inspection, condition monitoring and design-out maintenance With good design and effective preventive maintenance actions, the likelihood of failures and their consequences can be reduced but failures can never be totally eliminated

The approach to maintenance has changed significantly over the last one hundred years Over a hundred years ago, the focus was primarily on corrective maintenance delegated to the maintenance section of the business to restore failed systems to an operational state Maintenance was carried out by trained technicians and was viewed as an operational issue and did not play a role in the design and operation of the system The importance of preventive maintenance was fully appreciated during the Second World War Preventive maintenance involves additional costs and is worthwhile only if the benefits exceed the costs Deciding

4 K Kobbacy and D Murthy

the optimum level of maintenance requires building appropriate models and use of sophisticated optimisation techniques Also, around this time, maintenance issues started getting addressed at the design stage and this led to the concept of main-tainability Reliability and maintainability (R&M) became major issues in the design and operation of systems

Degradation and failure depend on the stresses on the various components of the system These depend on the operating conditions that are dictated by commercial considerations As a result, maintenance moved from a purely technical issue to a strategic management issue with options such as outsourcing of maintenance, leasing equipment as opposed to buying, etc Also, advances in technologies (new materials, new sensors for monitoring, data collection and analysis) added new dimensions (science, technology) to maintenance These advances will continue at an ever-increasing pace in the twenty-first century

This handbook tries to address the various issues associated with the tenance of complex systems The aim is to give a snapshot of the current status and highlight future trends Each chapter deals with a particular aspect of maintenance (for example, methodology, approaches, technology, management, modelling analysis and optimisation) and reports on developments and trends in a particular industry sector or deals with a case study In this chapter we give an overview of the handbook The outline of the chapter is as follows Section 1.2 deals with the framework that is needed to study the maintenance of complex systems and we discuss some of the salient issues Section 1.3 presents the structure of the book and gives a brief outline of the different chapters in the handbook We conclude with a discussion of the target audience for the handbook

main-1.2 Framework for Study of Maintenance

A proper study of maintenance requires a comprehensive framework that rates all the key elements However, not all the elements would be relevant for a particular maintenance problem under consideration

incorpo-The systems approach is an effective approach to solving maintenance lems In this approach, the real world relevant to the problem is described through

prob-a chprob-arprob-acterisprob-ation where one identifies the relevprob-ant vprob-ariprob-ables prob-and the interprob-action between the variables This characterisation can be done using language or a schematic network representation where the nodes represent the variables and the connected arcs denote the relationships This is good for qualitative analysis For quantitative analysis, one needs to build mathematical models to describe the relationships Often this requires stochastic and dynamical formulations as system degradation and failures occur in an uncertain manner In this section, we discuss the various key elements and some related issues

We use the term “asset” to denote a complex system or individual equipment It can include infrastructures such as buildings, bridges etc in addition to those listed

in Section 1.1

1.2.1 Stakeholders

For an asset there can be several stakeholders as indicated in Figure 1.1

Figure 1.1 Stakeholders for maintenance of an asset

The number of parties involved would depend on the asset under consideration

For example, in case of a rail network (used to provide a service to transport people

and goods) the customers can include the rail operators (operating the rolling

stock) and the public The owner can be a business entity, a financial institution or

a government agency The operator is the agency that operates the track and is

responsible for the flow of traffic The service provider refers to the agency

carrying out the maintenance (preventive and corrective) It can be the operator (in

which case maintenance is done in-house) or some external agent (if maintenance

is outsourced) or both (when only some of the maintenance activities are

out-sourced) The regulator is the independent agency which deals with safety and risk

issues They define the minimum standards for safety and can impose fines on the

owner, operator and possibly the service provider should the safety levels be

compromised Government plays a critical role in providing the subsidy and

assuming certain risks In this case all the parties involved are affected by the

maintenance carried out on the asset If the line is shut either frequently and/or for

long duration, it can affect customer satisfaction and patronage, the returns to the

operators and owners and the costs to the government

1.2.2 Different Perspectives

We focus our attention on the case where the asset is owned by the owner and

maintenance is outsourced In this case, we have two parties – (i) owner (of the

asset) and (ii) service agent (providing the maintenance) Figure 1.2 is a very

simplified system characterisation of the maintenance process where the

main-6 K Kobbacy and D Murthy

tenance activities are defined through a maintenance service contract The problem

is to determine the terms of the service contract

Figure 1.2 System characterisation for maintenance out-sourcing

Each of the elements of Figure 1.2 involves several variables For example, the maintenance service contract involves the following: (i) duration of contract, (ii) price of contract, (iii) maintenance performance requirements, (iv) incentives and penalties, (v) dispute resolution, etc The maintenance performance requirements can include measures such as availability, mean time between failures and so on The characterisation of the owner’s decision-making process can involve costs, asset state at the end of the contract, risks (service agent not providing the level and quality of service) and so on The interests and goals of the owner are different from that of the service agent

The study of maintenance is complicated by the unknown and uncontrollable factors It could be rate of degradation (which depends on several factors such as material properties, operating environment etc) and other commercial factors (high demand for power in the case of a power plant due to very hot weather)

1.2.3 Key Issues and the Need for Multi-disciplinary Approach

The key issues in the maintenance of an asset are shown in Figure 1.3 The asset acquisition is influenced by business considerations and its inherent reliability is determined by the decisions made during design The field reliability and degrada-tion is affected by operations (usage intensity, operating environment, operating load etc.) Through use of technologies, one can assess the state of the asset The analysis of the data and models allow for optimizing the maintenance decisions (either for a given operating condition or jointly optimizing the maintenance and operations) Once the maintenance actions have been formulated it needs to be implemented

Figure 1.3 Key Issues in maintenance of an asset

To execute effective maintenance one needs to have a good understanding of a

variety of concepts and techniques for each of the issues Another issue is the

computer packages that allow one to collect and analyze data and build models and

derive the optimal solutions

The linking of the technical and commercial issues is indicated in Figure 1.4

and this requires an inter-disciplinary approach

Figure 1.4 Linking technical and commercial issues

8 K Kobbacy and D Murthy

The disciplines involved are as follows

1.2.3.1 Engineering

The degradation of an asset depends to some extent on the design and building (or production) of the asset Poor design leads to poor reliability that in turn results in high level of corrective maintenance On the other hand, a well-designed system is more reliable and hence less prone to failures Maintainability deals with main-tenance issues at the design and development stage of the asset

1.2.3.2 Science

This is very important in the understanding of the physical mechanisms that are at play and have a significant influence on the degradation and failure Choosing the wrong material can have a serious consequence and impact on the subsequent maintenance actions needed

1.2.3.3 Economic

Maintenance costs can be a significant fraction of the total operating budget for a business depending on the industry sector There are two types of costs – annual cost and cost over the life cycle of the asset The costs can be divided into direct (labour, material etc.) and indirect (consequence of failure)

1.2.3.4 Legal

This is important in the context of maintenance out-sourcing and maintenance of leased equipment In both cases, the central issue is the contract between the parties involved Of particular importance is dispute resolution when there is a disagreement between the parties in terms of the violation of some terms of the contract

1.2.3.5 Statistics

The degradation and failures occur in an uncertain manner As such, the analysis of such data requires the use of statistical techniques Statistics provide the concepts and tools to extract information from data and for the planning of efficient collec-tion systems

1.2.3.6 Operational Research

Operation research provides the tools and techniques for model building, analysis and optimization Often, analytical approaches fail and one needs to use simulation approach to evaluate the outcomes of different decisions and to choose the optimal (or near optimal) strategies

1.2.3.7 Reliability Theory

Reliability theory deals with the interdisciplinary use of probability, statistics and stochastic modelling, combined with engineering insights into the design and the scientific understanding of the failure mechanisms, to study the various aspects of reliability As such, it encompasses issues such as (i) reliability modelling, (ii) reliability analysis and optimization, (iii) reliability engineering, (iv) reliability science, (v) reliability technology and (vi) reliability management

1.2.3.8 Information Technology and Computer Science

The operation and maintenance of complex assets generates a lot of data One

needs efficient ways to store and manipulate the data and to extract relevant

information from data Computer science provides a range of artificial intelligence

techniques such as data mining, expert systems, neural networks etc., which are

very important in the context of maintenance

1.2.4 Maintenance Management

Maintenance management deals with the overall management of the maintenance

of an asset The management needs to be done at three different levels (strategic,

tactical and operational) as indicated in Figure 1.5

STRATEGIC LEVEL

TACTICAL LEVEL

OPERATIONAL LEVEL

MAINTENANCE STRATEGY

MAINTENANCE PLANNING AND SCHEDULING

MAINTENANCE WORK EXECUTION

- BUSINESS PERSPECTIVE

- TECHNICAL & COMMERCIAL

- IN-HOUSE vs OUT-SOURCING

- REPLACEMENT / DESIGN CHANGES

- DEGRADATION (RELIABILITY SCIENCE)

Figure 1.5 Maintenance management

The strategic level deals with maintenance strategy This needs to be

formu-lated so that it is consistent and coherent with other (production, marketing,

finance, etc.) business strategies The tactical level deals with the planning and

scheduling of maintenance The operational level deals with the execution of the

maintenance tasks and collection of relevant data

1.3 Structure of the Handbook

The handbook integrates the vast literature on maintenance with each chapters

focussing on a different aspect of maintenance and written by active researchers

with international reputation and/or experienced practitioners from industry Each

chapter either reviews the literature dealing with a particular aspect of maintenance

(for example, methodology, approaches, technology, management, modelling

ana-10 K Kobbacy and D Murthy

lysis and optimisation), reports on developments and trends in a particular industry sector, or deals with a case study

The book is structured into five parts and each of the last four parts contains several chapters The topic of the different chapters is as indicated below

Part A: An Overview

Chapter 1: An Overview (Khairy Kobbacy and Pra Murthy)

Part B: Evolution of Concepts and Approaches

Chapter 2: Maintenance: An Evolutionary Perspective (Liliane Pintelon and

Alejandro Parodi Herz)

Chapter 3: New Technologies for Maintenance (Jay Lee and Haixia Wang) Chapter 4: Reliability Centred Maintenance (Marvin Rausand and Jorn Vatn) Part C: Methods and Techniques

Chapter 5: Condition-based Maintenance Modelling (Wenbin Wang)

Chapter 6: Maintenance Based on Limited Data (David F Percy)

Chapter 7: Reliability Prediction and Accelerated Testing (Elsayed A Elsayed) Chapter 8: Preventive Maintenance Models for Complex Systems

(David F Percy)

Chapter 9: Artificial Intelligence in Maintenance (Khairy A.H Kobbacy) Part D: Problem Specific Models

Chapter10: Maintenance of Repairable Systems (Bo Henry Lindqvist)

Chapter 11: Optimal Maintenance of Multi-component Systems: A Review

(Robin P Nicolai and Rommert Dekker)

Chapter 12: Replacement of Capital Equipment (Philip A Scarf and Joseph

C Hartman)

Chapter 13: Maintenance and Production: A Review of Planning Models

(Gabriella Budai, Rommert Dekker and Robin P Nicolai)

Chapter 14: Delay Time Modelling (Wenbin Wang)

Part E: Management

Chapter 15: Maintenance Outsourcing (Pra Murthy and Nat Jack)

Chapter 16: Maintenance of Leased Equipment (Pra Murthy and Jarumon

Pongpech)

Chapter 17: Computerised Maintenance Management Systems (Ashraf Labib) Chapter 18: Risk Analysis in Maintenance (Terje Aven)

Chapter 19: Maintenance Performance Measurement (MPM) System

(Uday Kumar and Aditya Parida)

Chapter 20: Forecasting for Inventory Management of Service Parts

(John E Boylan and Aris A Syntetos)

Part F: Applications (Case Studies)

Chapter 21: Maintenance in the Rail Industry (Jorn Vatn)

Chapter 22: Condition Monitoring of Diesel Engines

(Renyan Jiang and Xinping Yan)

Chapter 23: Benchmarking of the Maintenance Process at Banverket

(The Swedish National Rail Administration)

(Ulla Espling and Uday Kumar)

Chapter 24: Integrated e-Operations–e-Maintenance: Applications in North Sea

Offshore Assets (Jayanta P Liyanage)

Chapter 25: Fault Detection and Identification for Longwall Machinery Using

SCADA Data (Daniel Bongers and Hal Gurgenci)

A brief outline of each chapter is as follows

Chapter 2: Maintenance: An Evolutionary Perspective

In the past few decades industrial maintenance has evolved from a non-issue into a

strategic concern During this period the role of maintenance has drastically been

transformed This chapter, while considering the fundamental elements of

main-tenance and its environment, describes the evolution path of mainmain-tenance

manage-ment and the driving forces of such changes It basically explains how and why

maintenance practice has evolved in time It includes basic notions of maintenance

and clearly classifies and distinguishes between different types of maintenance

actions, policies and concepts currently available The chapter concludes by

en-lightening the reader with some new challenges in maintenance

Chapter 3: New Technologies for Maintenance

Predictive maintenance is critical to any engineering system, especially complex

systems, in order to avoid system breakdown With the recent advances in pervasive

computing, prognostics can be easily embedded in any devices and systems When

smart machines are networked and remotely monitored, and when their data is

modelled and continually analyzed with sophisticated embedded systems, it is

possible to go beyond mere “predictive maintenance” to intelligent “prognostics”, the

process of pinpointing exactly which components of a machine are likely to fail and

then autonomously trigger service and order spare parts This chapter addresses the

paradigm shift in modern maintenance systems from the traditional “fail and fix”

practices to a “predict and prevent” methodology Recent advances in prognostic

technologies and tools are presented, and future work directions are discussed

Chapter 4: Reliability Centred Maintenance

This chapter gives an introduction to reliability centred maintenance (RCM) The

RCM analysis process is divided into 12 distinct steps Each step is thoroughly

described and discussed The main RCM process is similar to the processes

outlined in RCM standards and guidelines, but has more focus on the optimization

of maintenance intervals A new approach is proposed based on generic RCM

analyses related to specified classes of consequences The new approach will

significantly reduce the workload of the RCM analysis A computer tool OptiRCM

12 K Kobbacy and D Murthy

that has been developed by the authors, is used to illustrate the new approach Several examples from railway applications are provided

Chapter 5: Condition-based Maintenance Modelling

This chapter presents a model for supporting condition based maintenance decision making The chapter discusses various issues related to the subject, such as the definition of the state of an asset, direct or indirect monitoring, relationship be-tween observed measurements and the state of the asset, and current modelling developments In particular, the chapter focuses on a modelling technique used recently in predicting the residual life via stochastic filtering This is a key element

in modelling the decision making aspect of condition based maintenance A few key condition monitoring techniques are also introduced and discussed Methods of estimating model parameters are outlined and a numerical example based on real data is presented

Chapter 6: Maintenance-based on Limited Data

Reliability applications often suffer from paucity of data for making informed maintenance decisions This is particularly noticeable for high reliability systems and when new production lines or new warranty schemes are planned Such issues are of great importance when selecting and fitting mathematical models to improve the accuracy and utility of these decisions This chapter investigates why reliability data are so limited and proposes statistical methods for dealing with these difficulties It considers graphical and numerical summaries, appropriate methods for model development and validation, and the powerful approach of subjective Bayesian analysis for including expert knowledge about the application area Chapter 7: Reliability Prediction and Accelerated Testing

This chapter presents an overview of accelerated life testing (ALT) methods and their use in reliability prediction at normal operating conditions It describes the most commonly used models and introduces new ones which are “distribution free” Design of optimum test plans in order to improve the accuracy of reliability prediction is also presented and discussed The chapter provides, for the first time, the link between accelerated life testing and maintenance actions It develops procedures for using the ALT results for estimating the optimum preventive maintenance schedule and the optimum degradation threshold level for degrading systems The procedures are demonstrated using two numerical examples

Chapter 8: Preventive Maintenance Models for Complex Systems

Preventive maintenance (PM) of repairable systems can be very beneficial in reducing repair and replacement costs, and in improving system availability Strategies for scheduling PM are often based on intuition and experience, though considerable improvements in performance can be achieved by fitting mathemati-cal models to observed data For simple repairable systems comprising few compo-nents or many identical components, compound renewal processes are appropriate This chapter reviews basic and advanced models for complex repairable systems and demonstrates their use for determining optimal PM intervals Computational

difficulties are addressed and practical illustrations are presented, based on

sub-systems of oil platforms and

Chapter 9: Artificial Intelligence in Maintenance

AI techniques have been used successfully in the past two decades to model and

optimise maintenance problems This chapter reviews the application of Artificial

Intelligence (AI) in maintenance management and introduces the concept of

developing intelligent maintenance optimisation system The chapter starts with an

introduction to maintence management, planning and scheduling and a brief

definition of AI and some of its techniques that have applications in maintenance

management A review of literatures is then presented covering the applications of

AI in maintenance We have focused on five AI techniques namely Knowledge

Based Systems, Case Based Reasoning, Genetic Algorithms, Neural Networks and

Fuzzy Logic This review also covers “hybrid” systems where two or more AI

techniques are used in an application A discussion of the development of the

prototype hybrid intelligent maintenance optimisation system (HIMOS) which was

developed to evaluate and enhance PM maintenance routines of complex

en-gineering systems then follows The chapter ends with a discussion of future

research and concluding remarks

Chapter 10: Maintenance of Repairable Systems

A repairable system is traditionally defined as a system which, after failing to

per-form one or more of its functions satisfactorily, can be restored to fully satisfactory

performance by any method other than replacement of the entire system An

extended definition used in this chapter includes the possibility of additional

main-tenance actions which aim at servicing the system for better performance, referred to

as preventive maintenance (PM) The common models for the failure process of a

repairable system are renewal processes (RP) and non-homogeneous Poisson

pro-cesses (NHPP) The chapter considers several generalizations and extensions of the

basic models, for example the trend renewal process (TRP) which includes NHPP

and RP as special cases, and having the property of allowing a trend in processes of

non-Poisson type When several systems of the same kind are considered, there may

be an unobserved heterogeneity between the systems which, if overlooked, may lead

to wrong decisions This phenomenon is considered in the framework of the TRP

process We then consider the extension of the basic models obtained by introducing

the possibility of PM using a competing risks approach Finally, models for

peri-odically inspected systems are studied, using a combination of time-continuous and

time-discrete Markov chains

Chapter 11: Optimal Maintenance of Multi-component Systems: A Review

This chapter gives an overview of the literature on multi-component maintenance

optimization focusing on work appearing since the 1991 survey by Cho and Parlar

A classification scheme primarily based on the dependence between components

(stochastic, structural or economic) is introduced Next, the papers are also

classi-fied on the basis of the planning aspect (short-term vs long-term), the grouping of

maintenance activities (either grouping preventive or corrective maintenance, or

opportunistic grouping) and the optimization approach used (heuristic, policy

14 K Kobbacy and D Murthy

classes or exact algorithms) Finally, attention is paid to the applications of the models

Chapter 12: Replacement of Capital Equipment

This chapter deals with models of replacement of capital equipment Capital ment models may be classified as economic life models or dynamic programming models The former are concerned with determining the optimal lifetime of an item

replace-of equipment taking account replace-of costs over some planning horizon The latter siders replacement decisions dynamically, determining whether plant should be retained or replaced after each period We begin by looking at simple economic life models These are applied in a case study on escalator replacement Economic life models are then extended to consider first an inhomogeneous fleet and then second a network system viewed as an inhomogeneous fleet with interacting items A number

con-of different dynamic programming models are introduced for singular systems and then expanded to homogeneous and inhomogeneous fleets and networks of assets Chapter 13: Maintenance and Production: A Review of Planning Models

This chapter gives an overview of the relation between planning of maintenance and production Production planning and scheduling models where failures and maintenance aspects are taken into account are considered first The planning of maintenance activities are considered next, where both preventive as well as corrective maintenance are discussed Third, the planning of maintenance activities

at such moments in time where the items to be maintained are not or less needed for production, also called opportunity maintenance is considered Apart from describing the main ideas, approaches, and results a number of applications are provided

Chapter 14: Delay Time Modelling

This chapter presented a modelling tool that was created to model the problems of inspection maintenance and planned maintenance interventions, namely Delay Time Modelling (DTM) This concept provides a modelling framework readily applicable to a wide class of actual industrial maintenance problems of assets in general and inspection problems in particular The delay time defines the failure process of an asset as a two-stage process The first stage is the normal operating stage from new to the point that a hidden defect has been identified The second stage is defined as the failure delay time from the point of defect identification to failure It is the existence of such a failure delay time which provides the oppor-tunity for preventive maintenance to be carried out to remove or rectify the identified defects before failures With appropriate modelling of the durations of these two stages, optimal inspection intervals can be identified to optimise a criterion function of interest This chapter first gives an outline of the delay time concept then introduces two delay time inspection models of a single component and a complex system respectively The parameters estimation techniques used in DTM are discussed Extensions to the basic delay time model are highlighted and future research in DTM concludes the chapter

Chapter 15: Maintenance Outsourcing

It is often uneconomical for businesses to carry out their own maintenance on

complex equipment The alternative is to ‘out-source’ the maintenance function

and use an external agent, under a service contract, to carry out some or all of the

maintenance actions (preventive and corrective) This chapter develops the

frame-work needed to study decision-making for maintenance outsourcing from both the

customer (equipment owner) and service agent perspectives The relevant literature

is reviewed and a game theoretic approach to maintenance outsourcing and the use

of agency theory is discussed The link between maintenance outsourcing and

extended warranties is highlighted and the scope for future research in both areas is

examined

Chapter 16: Maintenance of Leased Equipment

For leased equipment, the lessor has to carry out the maintenance of the equipment

over the lease period To ensure satisfactory performance and maintenance, the

lease contract has penalty terms which result in the lessor having to compensate the

lessee if the number of failures exceeds some specified number and/or the time to

rectify each failure exceeds some specified value This implies that the lessor needs

to take into account these penalties in determining the optimal maintenance

strategy The chapter starts with a conceptual framework to discuss the different

issues involved and then looks at models to help the lessor in developing the

optimal maintenance strategy

Chapter 17: Computerised Maintenance Management Systems

Computerised maintenance management systems (CMMSs) are vital for the

co-ordination of all activities related to the availability, productivity and maintainability

of complex systems Modern computational facilities have offered a dramatic scope

for improved effectiveness and efficiency in, for example, maintenance CMMSs

have existed, in one form or another, for several decades In this chapter, the

characteristics of CMMSs have been investigated and have highlighted the need for

them in industry and identified their current deficiencies

A proposed model is then presented to provide a decision analysis capability

that is often missing in existing CMMSs The effect of such model is to contribute

towards the optimisation of the functionality and scope of CMMSs for enhanced

decision analysis support The use of AI techniques in CMMSs is illustrated The

features of next generation maintenance systems are finally highlighted

Chapter 18: Risk Analysis in Maintenance

Risk analysis can be used for selection and prioritisation of maintenance activities,

and this application of risk analysis has been given increased attention in recent

years This chapter presents and discusses the use of risk analysis for this purpose

The chapter reviews some critical aspects of risk analysis important for the

successful implementation of such analyses in maintenance This relates to risk

descriptions and categorisations, uncertainty assessments, risk acceptance and risk

informed decision making, as well as selection of appropriate methods and tools

Both qualitative and quantitative approaches are covered A detailed risk analysis

is outlined showing the effect of maintenance on risk

16 K Kobbacy and D Murthy

Chapter 19: Maintenance Performance Measurement (MPM) System

It is important that factors influencing the performance of maintenance process should be identified, and measured, so that they can be monitored and controlled for improvement In this chapter, besides an overview of performance measurement, maintenance performance indicators, associated issues and challenges for developing

a maintenance performance measurement framework, and indicators as in use by different industries are discussed The framework considers stakeholders, business environment, multi-criteria and hierarchical needs amongst other

Chapter 20: Forecasting for Inventory Management of Service Parts

This chapter addresses issues pertinent to forecasting for the inventory management

of service parts In some sectors, such as the aerospace and automotive industries, a very wide range of service parts are held in stock, with significant implications for availability and inventory holding Their management is therefore an important task First, a number of possible approaches to classifying service parts for forecasting and inventory management related purposes are reviewed Second, parametric and non-parametric approaches to forecasting service parts requirements are discussed followed by the presentation of appropriate metrics for measuring the performance

of the inventory management system The existing empirical evidence on various forecasting methods is then summarised Finally, the conclusions of this work are presented along with the identification of some natural avenues for further research Chapter 21: Maintenance in the Rail Industry

The chapter presents two case studies in railway maintenance The first case study presents an optimisation model preventive maintenance of a train bogie In the model

a dynamic approach to grouping of maintenance activities is used enabling, e.g., opportunity maintenance Data from the Norwegian State Railways have been used

in the calculation example The second case study present a life cycle cost approach

to prioritization of larger maintenance and renewal projects under budget constraints Chapter 22: Condition Monitoring of Diesel Engines

Various techniques have been widely used to monitor the condition of diesel engines Analysis of engine lubricant is a most widely used condition monitoring technique In this chapter, a case study applying oil analysis technique to monitor the condition of marine diesel engines is presented The case study focuses on analysis and modelling of oil monitoring data The study first introduces the con-cept of state discriminant capability of condition variables and uses it to identify the significant condition variables, and then develops a state discriminant model to determine the state of the monitored system based on the current observation The model parameters are obtained by directly minimizing the misjudgment probabil-ity We believe that the proposed model has a great potential to be used due to its plausible mathematical basis and simplicity though it needs further testing with new data

Chapter 23: Benchmarking of the Maintenance Process at Banverket (The Swedish

National Rail Administration)

For sustaining a competitive edge in the business, railway companies all over the

world are looking for ways and means to improve their maintenance performance

Benchmarking is a very effective tool that can assist the management in their

pursuit of continuous improvement of their operation Three different benchmarks

have been studied based on a project benchmarking of the maintenance process

across borders, another project dealing with benchmarking of maintenance

out-sourcing by different track regions in Sweden, and a third project studying the level

on transparency among the European railway administrations The chapter discuss

the pro and cons, the areas for improvement and the need for improvement of

benchmarking metrics and framework

Chapter 24: Integrated e-Operations–e-Maintenance: Application in North Sea

Offshore Assets

Ongoing developments in Norway brings a good example of how an industry-wide

re-engineering process has triggered major changes in operations and maintenance

practice of complex and high-risk assets leading towards what is termed integrated

e-operations e-maintenance It aims towards a step-change to the conventional

operations and maintenance practices of offshore assets Initiatives have already been

taken to exploit new methods, smart techniques, and digital technologies to enable

remote monitoring of offshore equipment condition and asset performance in

land-based onshore support facilities using large ICT networks This has already proved to

have direct positive implications on the technical and safety integrity of assets, and

subsequently on the plant economics This chapter shares current experience and

knowledge with reference to ongoing developments in the Norwegian oil and gas

industry It highlights current offshore asset maintenance practice, changing technical

and economic environment that lead towards an e-approach, development and

im-plementation of integrated e-operations and e-maintenance solutions in the North sea,

key features of the e-approach in North sea assets, and future challenges to be

fully-integrated and fail-safe

Chapter 25: Fault Detection and Identification for Longwall Machinery Using

SCADA Data

In an attempt to improve equipment availability and facilitate informed,

preventa-tive maintenance, engineers may choose to implement one or more fault detection

and identification (FDI) technologies For complex systems (systems for which

component interactions are not understood and model uncertainties are significant),

data-driven methods of FDI are often the only practicable solution The

develop-ment of a data-driven FDI system for longwall mining equipdevelop-ment using SCADA

data is described here

Significant data preprocessing was required to generate a quality example set

Missing value estimation (MVE) techniques were required to complete the

high-dimensional stream of condition monitoring data from existing sensors A cost

function, in combination with a linear discriminant analysis, was used to ‘align’ the

inaccurate, categorical delay records with those delays inferred by the SCADA

data A neural network was developed to determine the state of the system as a

18 K Kobbacy and D Murthy

function of the real-time SCADA data input Validation of this algorithm with unseen condition monitoring data showed misclassification rates of machine faults

as low as 14.3%

1.4 Target Audience

The unique features of the book are as follows:

1 A coverage of the different approaches to maintenance

2 Deals with many different aspects (scientific, technical, commercial,

management, quantitative modelling) etc

3 Blends theory with practice

As such it should appeal to both researchers and practitioners For researchers (from different disciplines) it should provide a starting point for new research into different aspects of maintenance For practitioners it should provide the concepts and tools so that these can be used for improvements in the overall business performance Also we hope that it will serve as a reference book for use in postgraduate programs

in maintenance

Evolution of Concepts and Approaches

2

Maintenance: An Evolutionary Perspective

Liliane Pintelon and Alejandro Parodi-Herz

L Pintelon and A Parodi-Herz

2.1 Introduction

Over the last decennia industrial maintenance has evolved from a non-issue into a strategic concern Perhaps there are few other management disciplines that under-went so many changes over the last half-century During this period, the role of maintenance within the organization has drastically been transformed At first maintenance was nothing more than a mere inevitable part of production, now it is

an essential strategic element to accomplish business objectives Without a doubt, the maintenance function is better perceived and valued in organizations One could considered that maintenance management is no longer viewed as an under-dog function; now it is considered as an internal or external partner for success

In view of the unwieldy competition many organizations seek to survive by producing more, with fewer resources, in shorter periods of time.To enable these serious needs, physical assets take a central role However, installations have become highly automated and technologically very complex and, consequently, maintenance management had to become more complex having to cope with higher technical and business expectations Now the maintenance manager is confronted with very complicated and diverse technical installations operating in

an extremely demanding business context

This chapter, while considering the fundamental elements of maintenance and its environment, describes the evolution path of maintenance management and the driving forces of such changes In Section 2.2 the maintenance context is described and its dynamic elements are briefly discussed Section 2.3 explains how main-tenance practice have evolved in time and different epochs are distinguished Further, this sections devotes special attention to describe a common lexicon for maintenance actions and policies to further focuss on the evolution of maintenance concepts Section 2.4 underlines how the role of the maintenance manager has been reshaped as a consequence of the changes of the maintenance function Finally, the chapter concludes with Section 2.5 identifying the new challenges for maintenance

2.2 Maintenance in Context

To discuss the context in which maintenance management is embedded, one may raise the question what is maintenance as such? Most authors in maintenance management literature, one way or another, agree on defining maintenance as the

“set of activities required to keep physical assets in the desired operating condition

or to restore them to this condition” While this defines what maintenance is about,

it may suggest that maintenance is simple, which it is not, as will be confirmed by any maintenance practitioner Hence “maintenance management” is needed to ingrain maintenance practice in a complex and dynamic context From a pragmatic view, the key objective of maintenance management is “total asset life cycle optimization” In other words, maximizing the availability and reliability of the assets and equipment to produce the desired quantity of products, with the required quality specifications, in a timely manner Obviously, this objective must be attained in a cost-effective way and in accordance with environmental and safety regulations Figure 2.1 clearly shows that maintenance is embedded in a given business context to which it has to contribute What is more, it shows that the maintenance function needs to cope with multiple forces and requirements within and outside the walls of the organization Beyond any doubt, the tasks of main-tenance are complex, enclosing a blend of management, technology, operations and logistics support elements

Figure 2.1 Maintenance in context

To cope with and to coordinate the complex and changing characteristics that constitute maintenance in the first place, a management layer is imperative Management is about “what to decide” and “how to decide” In the maintenance arena, a manager juggles with technology, operations and logistics elements that mainly need to harmonize with production Technology refers to the physical assets which maintenance has to support with adequate equipment and tools Operations indicate the combination of service maintenance interventions with

life cycle optimization

Information Technology Legislation

Competition

Outsourcing Market

People

e-business

life cycle optimization

Information Technology Legislation

Competition

Outsourcing Market

People

e-business

Maintenance: An Evolutionary Perspective 23

core production activities Finally, the logistics element supports the maintenance activities in planning, coordinating and ultimately delivering, resources like spare parts, personnel, tools and so forth In one way or another, all these elements are always present, but their intensity and interrelationships will vary from one situation to another For example, the elevator maintenance in a hospital vs the plant maintenance in chemical process industries stipulates a different maintenance recipe tailored to the specific needs Clearly, the choice of the structural elements

of maintenance is not independent from the environment Besides, other factors like the business context, society, legislation, technological evolution, outsourcing market, will be important Furthermore, relative new trends, such as the e-business context, will influence the current and future maintenance management enor-mously A whole new era for maintenance is expected as communication barriers are bridged and coordination opportunities of maintenance service become more intense

2.2.1 Changes in the Playing Field of Maintenance

One should expect that neither maintenance management nor its environment are stationary The constant changes in the field of maintenance are acknowledged to have enabled new and innovative developments in the field of maintenance science

The technological evolution in production equipment, an ongoing evolution that started in the twentieth century, has been tremendous At the start of the twentieth century, installations were barely or not mechanized, had simple design, worked in stand-alone configurations and often had a considerable overcapacity Not surprisingly, nowadays installations are highly automated and technologically very complex Often these installations are integrated with production lines that are right-sized in capacity

Installations not only became more complex, they also became more critical in terms of reliability and availability Redundancy is only considered for very critical components For example, a pump in a chemical process installation can be con-sidered very critical in terms of safety hazards Furthermore, equipment built-in characteristics such as modular design and standardization are considered in order

to reduce downtime during corrective or preventive maintenance However, dominantly only for some newer, very expensive installations, such as flexible manufacturing systems (FMS), these principles are commonly applied Fortunate-

pre-ly, a move towards higher levels of standardization and modularization begins to

be witnessed at all level of the installations As life cycle optimization concepts are commendable, it becomes mandatory that at the early design stages supportability and maintainability requirements are well thought-out

Parallel to the technological evolution, the ever-increasing customer focus causes even higher pressure, especially on critical installations As customers’ service in terms of time, quality and choice becomes central to production deci-sions, the more flexibility is required to cope with these varying needs This calls for well-maintained and reliable installations capable to fulfil shorter and more reliable lead-times estimation Physical assets are ever more important for business success

Maintenance does not escape from the (r)evolution in information tion technology (ICT), which has tremendously changed business practices How-ever, we comment further on this topic in Section 2.3, by illustrating the impact on the role of the maintenance manager as such

communica-Furthermore, new production and management principles such as Just-in-time (JIT) philosophy, Lean principles, total quality management (TQM) and so forth, have emerged These production trends intend, by all means, to reduce waste and remove non-value added transactions It is not surprising that work-in-process (WIP) inventories are one of the key issues for improvement Clearly, WIP inven-tories incur high costs as a consequence of the capital immobilization, expensive floor space, etc As processes happen to be streamlined, WIP inventories are no longer a buffer for problems; accordingly, asset availability and reliability are ever more imperative Albeit, these principles were initially inspired for production and manufacturing environments are currently also applied and translated in service context

Above all, the business environment has also changed Competition has become fierce and worldwide due to the globalization The latter not only implies that competitors are located all over the world, but also that decisions to move production or service activities from a non-efficient site (e.g due to high opera-tions and maintenance costs) to another site are quickly taken, even if the other location belongs to another continent Obviously, with the advent of globalization and intense competitive pressures, organizations are looking for every possible source of competitive advantage This implies that the nature of business environ-ment has become more complex and dynamic requiring different competitive strategies Many companies are critically evaluating their value chain and often decide to drastically reorganize it This results in focusing on the core business Consequently outsourcing of some non-core business activities and the creation of new partnerships and alliances are being considered by many organizations Not surprisingly, maintenance as a support function is no exception for out-sourcing Yet, it may not be so simple Outsourcing maintenance of technical systems can become a sensitive issue if it is not handled with diligence Technical systems are unique and situation specific For example, outsourcing maintenance

of utilities or elevators can be relatively straightforward, but when it comes to production floor equipment it can be a strategic issue that has to be handled with extreme care These circumstances suggest that outsourcing needs to be considered

at operational, tactical and strategic level; see Figure 2.2

The simplest, and also the most common, form of outsourcing is “operational outsourcing” At this level, a specific task is outsourced and the relationship between supplier and customer is strictly limited to a sell-buy situation The impact

on the internal organization of the customer is also limited As outsourcing moves

up in the organizational pyramid the relationship between supplier and customer changes and “tactical outsourcing” maybe required At this level of outsourcing the customer shares management responsibility with the supplier and a simple kind of partnership is established The impact on the internal organization is also greater Finally, moving towards the organization’s top and for more critical maintenance services, a new form of outsourcing is created, the so-called “strategic out-sourcing” This type of outsourcing is also labelled as “transformational out-

Maintenance: An Evolutionary Perspective 25

sourcing” because of its impact on the customer’s internal organization Here a complete outsourcing is carried out, the maintenance department is cut away from the customer and moved to the supplier The relationship between customer and supplier is a strong partnership: the customer has fully entrusted the supplier with one of its strategic maintenance activities This level of outsourcing is yet less common than the former ones The rationales of whether or not to outsource main-tenance activities are complex and require a well-thought and structured outsourcing process As mentioned maintenance outsourcing can cover a lot of alternatives Fortunately, besides, traditional outsourcing of maintenance activities to equipment suppliers or the use of some small local firms, there is nowadays a growing market

of medium sized and large outsourcing firms These firms offer a range of consulting support, specialized services and even full service to allow strategic outsourcing to work

Figure 2.2 Outsourcing decision levels

Societal expectations concerning technology is also creating boundary tions for maintenance management The attention paid to sustainability (3P: people, profit, planet) is a clear sign of this Legislation is getting more and more stringent This is especially important here because of its impact on occupational safety and environmental standards

condi-Note that most of the above-mentioned trends for industrial installations can be easily translated to the service sector Think, for example, of automated warehouses

in distribution centre, hospital equipment or building utilities

Generic services Specialised services Projects Service package

Full service

e.g renovation, shutdown,

e.g MRO, utilities, facilities,

e.g outsourcing of all maintenance, BOT,

To think with…

To manage…

To carry out…

To organise… Generic services

Specialised services Projects Service package

Full service

e.g renovation, shutdown,

e.g MRO, utilities, facilities,

e.g outsourcing of all maintenance, BOT,

To think with…

To manage…

To carry out…

To organise…

2.3 Maintenance Practices Over Time

Consequent to the transformation the maintenance context, the maintenance function has also drastically evolved from a non-issue into a strategic concern (see Figure 2.3) At first maintenance was nothing more than an inevitable part of production; it simply was a necessary evil Repairs and replacements were tackled when needed and no optimization questions were raised Later on, it was conceived that maintenance was a technical matter This not only included optimizing technical maintenance solutions, but it also involved attention of the organization

on the maintenance work Further on, maintenance became a full-blown function, instead of production sub-function Clearly, now maintenance management has become a complex function, encompassing technical and management skills, while still requiring flexibility to cope with the dynamic business environment Top management recognizes that having a well thought out maintenance strategy together with a careful implementation of that strategy could actually have a significant financial impact Nowadays, this has led to treating maintenance as a mature partner in business strategy development and possibly at the same level as production In turn, these strategies formally consider establishing external partnerships and outsourcing of the maintenance function

Figure 2.3 The maintenance function in a time perspective

The fact that maintenance has become more critical implies that a thorough insight into the impact of maintenance interventions, or the omission of these, is indispensable Per se, good maintenance stands for the right allocation of resources (personnel, spares and tools) to guarantee, by deciding on the suitable combination

of maintenance actions, a higher reliability and availability of the installations Furthermore, good maintenance foresees and avoids the consequences of the failures, which are far more important than the failures as such Bad or no main-tenance can appear to render some savings in the short run, but sooner or later it will be more costly due to additional unexpected failures, longer repair times, accelerated wear, etc Moreover, bad or no maintenance may well have a signi-ficant impact on customer service as delivery promises may become difficult to fulfil Hence, a well-conceived maintenance program is mandatory to attain busi-ness, environmental and safety requirements

Despite the particular circumstances, if one intends to compile or judge any maintenance programme, some elementary maintenance terms need to be unam-biguous and handled with consistency Yet, both in practice and in the literature a lot of confusion exists For example, what for some is a maintenance policy others refer to as a maintenance action; what some consider preventive maintenance others will refer to as predetermined or scheduled maintenance Furthermore, some argue that some concepts can almost be considered strategies or philosophies, and

Maintenance: An Evolutionary Perspective 27

so on Certainly there is a lot of confusion, which perhaps is one of the breathing characteristics of such a dynamic and young management science The terminol-ogy used to describe precisely some maintenance terms can almost be taken as philosophical arguments However, the adoption of a rather simplistic, but truly germane classification is essential Not intending to disregard preceding terminol-ogies, neither to impose nor dictate a norm, we draw attention, in particular, to three of those confusing terms: maintenance action, maintenance policy and maintenance concept In the remainder of this chapter the following terminology is adopted

Maintenance Action Basic maintenance intervention, elementary task carried out

by a technician (What to do?)

Maintenance Policy Rule or set of rules describing the triggering mechanism for the different maintenance actions (How is it triggered?)

Mainenance Concept Set of maintenance polices and actions of various types and the general decision structure in which these are planned and supported (The logic and maintenance recipe used?)

2.3.1 Maintenance Actions

Basically, as depicted in Figure 2.4, maintenance actions or interventions can be of two types They are either corrective maintenance (CM) or precautionary main-tenance (PM) actions

2.3.1.1 Corrective Maintenance Actions (CM)

CM actions are repair or restore actions following a breakdown or loss of function These actions are “reactive” in nature; this merely implies “wait until it breaks, then fit it!” Corrective actions are difficult to predict as equipment failure behavior

is stochastic and breakdowns are unforeseen Maintenance actions such as replacement of a failed light bulb, repair of a ruptured pipeline and the repair of a stalled motor are some examples of corrective actions

2.3.1.2 Precautionary Maintenance Actions (PM)

PM actions can either be “preventive, predictive, proactive or passive” in nature These types of actions are moderately more complex than the former To describe fully each one of them, a book can be written on its own Nonetheless, the fundamental ideas aim at diminishing the failure probability of the physical asset and/or to anticipate, or avoid if possible, the consequences if a failure occurs Some

PM actions (preventive and predictive) are somewhat easier to plan, because they can rely on fixed time schedules or on prediction of stochastic behaviours How-ever, other types of PM actions become ongoing tasks, originating from the attitude concerning maintenance Somehow they became part of the tacit knowledge of the organization Some precise examples of precautionary actions which can be mentioned are lubrication, bi-monthly bearing replacements, inspection rounds, vibration monitoring, oil analysis, design adjustments, etc All these tasks are considered to be precautionary maintenance actions; however, the underlying prin-ciples may be different

Figure 2.4 Actions, policies and concepts in maintenance 1

Although it seems a very clear-cut way of defining elementary maintenance interventions, it still may be difficult in practice to assign some interventions to either class An example here is routine maintenance on medical equipment such as

a breathing device Cleaning and sterilizing this equipment can be called cautionary maintenance since the equipment is not defective at the moment of the intervention On the other hand, it is very difficult to predict when an intervention will be needed, and this is a typical characteristic of a corrective intervention Furthermore, even within precautionary maintenance, it is not always simple to classify certain actions into simple types This is due to the changing perception on maintenance and the fast evolution of its techniques

pre-2.3.1.3 Acuity of Maintenance Actions

As maintenance knowledge is enhanced and more advance enabling technologies are available, the perception on which maintenance action is “right” has changed a lot during the last decennia In the 1950s almost all maintenance actions were corrective Per se maintenance was considered as an annoying and unavoidable cost, which could not be managed Later on, in the 1960s many companies switched to precautionary (preventive) maintenance programs as they could recognize that some failures on mechanical component had a direct relation with the time or number of cycles in use This belief was mainly based on physical wear

of components or age-related fatigue characteristics At that time, it was accepted

existing concept

Customized concept LCC

existing concept

Customized concept LCC

RCM

TPM

Ad hoc Optimizing

existing concept

Customized concept LCC

Precautionary

Predictive, preventive, proactive and passive reactive

Maintenance: An Evolutionary Perspective 29

that preventive actions could avoid some of the breakdowns and would lead to cost savings in the long run The main concern was how to determine, based on historical data, the adequate period to perform preventive maintenance Certainly, not enough was known about failure patterns, which, among other reasons, have led to a whole separate branch of engineering and statistics: reliability engineering

In the late 1970s and early 1980s, equipment became in general more complex

As result, the super-positioning effect of the failure pattern of individual ponents starts to alter the failure characteristics of simpler equipment Hence, if there is no dominant age-related failure mode, preventive maintenance actions are

com-of limited use in improving the reliability com-of complex items At this point, the effectiveness of applying preventive maintenance actions started to be questioned and was considered more carefully A common concern about “over-maintaining” grew rapidly Moreover, as the insidious belief on preventive maintenance benefits was put at risk, new precautionary (predictive) maintenance techniques emerged This meant a gradual, though not complete, switch to predictive (inspection and condition-based) maintenance actions Naturally, predictive maintenance was, and still is, limited to those applications where it was both technically feasible and economically interesting Supportive to this trend was the fact that condition-monitoring equipment became more accessible and cheaper Prior to that time, these techniques were only reserved to high-risk applications such as airplanes or nuclear power plants

In the late 1980s and early 1990s a different footprint on maintenance history occurred with the emergence of concurrent engineering or life cycle engineering Here maintenance requirements were already under consideration at earlier product stages such as design or commission As a result, instead of having to deal with built in characteristics, maintenance turned out to be active in setting design requirements for installations and became partly involved in equipment selection and development All this led to a different type of precautionary (proactive) main-tenance, the underlying principle of which was to be proactive at earlier product stages in order to avoid later consequences Furthermore, as the maintenance function was better appreciated within the organization, more attention was paid to additional proactive maintenance actions For example, as operators are in straight and regular contact with the installations they could intuitively identify and “feel” right or wrong working conditions of the equipment Conditions such as noise, smell, rattle vibration, etc., that at a given point are not really measured, represent tacit knowledge of the organization to foresee, prevent or avoid failures and its consequences in a proactive manner Yet these actions are indeed typically not performed by maintenance people themselves, but are certainly part of the structural evolution of maintenance as a formal or informal partner within the organization

The last type of precautionary (passive) maintenance actions are driven by the opportunity of other maintenance actions being planned These maintenance actions are precautionary since they occur prior to a failure, but are passive as they

“wait” to be scheduled depending on others probably more critical actions Passive actions are in principle low priority for the maintenance staff as, at a given moment

in time, they may not really be a menace for functional or safety failures However, these actions can save significant maintenance resources as they may reduce the

number of maintenance interventions, especially when the set up cost of tenance is high For example, when maintenance actions are planned or need to be carried out on offshore oil platforms or on windmills in remote locations, getting to the equipment equipment can be costly Therefore, optimizing the best combina-tion of maintenance actions, at that point in time, is mandatory This may invoke replacing components with significant residual life that in different circumstances would not be replaced

main-2.3.2 Maintenance Policies

As new maintenance techniques happen to be available and the economic tions of maintenance action are comprehended, a direct impact on the maintenance policies is expected Several types of maintenance policies can be considered to trigger, in one way or another, either precautionary or corrective maintenance interventions As described in Table 2.1, those policies are mainly failure-based maintenance (FBM), time/used-based maintenance (TBM/UBM), condition-based maintenance (CBM), opportunity-based maintenance (OBM) design-out main-tenance (DOM), and e-maintenance

implica-Table 2.1 Generic maintenance policies Policy Description

FBM Maintenance (CM) is carried out only after a breakdown In case of CFR

behaviour and/or low breakdown costs this may be a good policy

TBM / UBM PM is carried out after a specified amount of time (e.g 1 month, 1000 working

hours, etc.) CM is applied when necessary UBM assumes that the failure behaviour is predictable and of the IFR type PM is assumed to be cheaper than

CM

CBM PM is carried out each time the value of a given system parameter (condition)

exceeds a predetermined value PM is assumed to be cheaper than CM CBM is gaining popularity due to the fact that the underlying techniques (e.g vibration analysis, oil spectrometry, ) become more widely available and at better prices The traditional plant inspection rounds with a checklist are in fact a primitive type of CBM

OBM For some components one often waits to maintain them until the “opportunity”

arises when repairing some other more critical components The decision whether or not OBM is suited for a given component depends on the expectation

of its residual life, which in turn depends on utilization

DOM The focus of DOM is to improve the design in order to make maintenance easier

(or even eliminate it) Ergonomic and technical (reliability) aspects are important here

CFR = Constant failure rate, IFR=Increasing failure rate

For the more common maintenance policies many models have been developed

to support tuning and optimization of the policy setting It is not our intention to explain the fundamental differences between these models, but rather to provide an overview of types of policies available and why these have been developed Much

Maintenance: An Evolutionary Perspective 31

has to do with the discussion in the previous section regarding the acuity of tenance actions Therefore, it is clear that policy setting and the understanding of its efficiency and effectiveness continues to be fine-tuned as any other management science We advocate the reader, particularily interested in the underlying principles and type of models, to review McCall (1965), Geraerds (1972), Valdez-Flores and Feldman (1989), Cho and Parlar (1991), Pintelon and Gelders (1992), Dekker (1996), Dekker and Scarf (1998) and Wang (2002) for a full overview on the state-of-the-art literature

main-The whole evolution of maintenance was based not solely on technical but rather on techno-economic considerations FBM is still applied providing the cost

of PM is equal to or higher than the cost of CM Also, FBM is typically handy in case of random failure behaviour, with constant failure rate, as TBM or UBM are not able to reduce the failure probability In some cases, if there exists a measurable condition, which can signal the probability of a failure, CBM can be also feasible Finally, a FBM policy is also applied for installations where frequent

PM is impracticable and expensive, such as can be the maintenance of glass ovens Either TBM or UBM is applied if the CM cost is higher than PM cost, or if it is necessary because of criticality due to the existence of bottleneck installation or safety hazards issues Also in case of increasing failure behaviour, like for example wear-out phenomena, TBM and UBM policies are appropriate

Typically, CBM was mainly applied in those situations where the investment in condition monitoring equipment was justified because of high risks, like aviation

or nuclear power regeneration Currently, CBM is beginning to be generally accepted to maintain all type installations Increasingly this is becoming a common practice in process industries In some cases, however, technical feasibility is still a hurdle to overcome Another reason that catches the attention of practitioners in CBM is the potential savings in spare parts replacements thanks to the accurate and timely forecasts on demand In turn, this may enable better spare parts management through coordinated logistics support

Finding and applying a suitable CBM technique is not always easy For example, the analysis of the output of some measurement equipment, such as advanced vibration monitoring equipment, requires a lot of experience and is often work for experts But there are also simpler techniques such as infrared measuring and oil analysis suitable in other contexts At the other extreme, predictive techniques can be rather simple, as is the case of checklists Although fairly low-level activity, these checklists, together with human senses (visual inspections, detection of “strange” noises in rotating equipment, etc.) can detect a lot of potential problems and initiate

PM actions before the situation deteriorates to a breakdown

At present FBM, TBM, UBM and CBM accept and seize the physical assets which they intend to maintain as a given fact In contrast, there are more proactive maintenance actions and policies which, instead of considering the systems as “a given”, look at the possible changes or safety measures needed to avoid maintenance

in the first place This proactive policy is referred to as DOM This policy implies that maintenance is proactively involved at earlier stages of the product life cycle to solve potential related problems Ideally, DOM policies intend to completely avoid maintenance throughout the operating life of installations, though, this may not be realistic This leads one to consider a diverse set of maintenance requirements at the

early stages of equipment design As a consequence, equipment modifications are geared either at increasing reliability by raising the mean-time-between-failures (MTBF) or at increasing the maintainability by decreasing the mean-time-to-repair (MTTR) Per se DOM aims to improve the equipment availability and safety Some equipment modifications may merely request ergonomic considerations to reduce MTTR, others may need totally new designs Often DOM projects are combined with efforts to increase occupational safety or increase production capacity, such as set up reduction programs

A rather passive, but considerably important maintenance policy that needs to

be mentioned is OBM Typically OBM is applied for non-critical components with

a relatively long lifetime For these components no separate maintenance programs are scheduled; maintenance happens if an opportunity arises due to a maintenance intervention for another component of that machine

More recently in the mid-1990s, with the emergence of the Internet as an enabling technology and the growth of e-business as the standard on business communication, e-maintenance also appeared in the radar of maintenance policies E-maintenance rather than a policy can also be considered as a means or enabler to some, if not all, the previous policies However, it is more than just an acronym; it

is a step forward to full-integrated maintenance techniques without the boundaries

of place It is in fact a maintenance policy on its own that can support other policies In particular, academics and practitioners watch with anticipation the great impact it may have on CBM Conditions measured on site can be remotely monitored, opening entirely new dimensions and opportunities for maintenance services Therefore, e-maintenance has captured much attention of maintenance re-searchers given its great impact on business practice An example of this evolution

is telemaintenance, which allows the diagnosis of installation and to perform limited type of repairs from a remote location using ICT and sophisticated control and knowledge tools

2.3.3 Maintenance Concepts

The idea of an “optimized” maintenance program suggests that an adequate mix of maintenance actions and policies needs to be selected and fine-tuned in order to improve uptime, extend the total life cycle of physical asset and assure safe working conditions, while bearing in mind limiting maintenance budgets and environmental legislation This does not seem to be straightforward, and may require a holistic view Therefore, a “maintenance concept” for each installation is necessary to plan, control and improve the various maintenance actions and policies applied A maintenance concept may in the long term even become a philosophy, tenet or attitude to perform maintenance In some cases advance main-tenance concepts are almost considered strategies on their own What is certain is that maintenance concepts determine the business philosophy concerning main-tenance, and that they are needed to manage the complexity of maintenance per se

In practice, it is clear that more and more companies are spending time and effort determining the right maintenance concept

As a matter of fact, maintenance concepts need to be formulated considering the physical characteristics and the context within which installations operate Not