On the maintenance concept for a technical system : a framework for design

On the maintenance concept for a technical system : a

framework for design

Citation for published version (APA):

Gits, C. W. (1984). On the maintenance concept for a technical system : a framework for design. Technische

Hogeschool Eindhoven. https://doi.org/10.6100/IR108981

DOI:

10.6100/IR108981

Document status and date:

Published: 01/01/1984

Document Version:

Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can be

important differences between the submitted version and the official published version of record. People

interested in the research are advised to contact the author for the final version of the publication, or visit the

DOI to the publisher's website.

• The final author version and the galley proof are versions of the publication after peer review.

• The final published version features the final layout of the paper including the volume, issue and page

numbers.

Link to publication

General rights

Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain

• You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:

www.tue.nl/taverne

Take down policy

If you believe that this document breaches copyright please contact us at:

openaccess@tue.nl

providing details and we will investigate your claim.

A FRAMEWORK FOR DESIGN

bVER

HET ONDERHOUDSCONCEPT VOOR EEN TECHNISCH SYSTEEM.

EEN ONTWERPKADER

PROEFSCHRIFT

TER VERKRIJGING VAN DE GRAAD VAN DOCTOR IN DE TECHNISCHE WETENSCHAPPEN AAN DE TECHNISCHE HOGESCHOOL EINDHOVEN, OP GEZAG VAN DE RECTOR MAGNIFICUS, PROF. DR. S. T. M. ACKERMANS, VOOR EEN COMMISSIE AANGEWEZEN DOOR HET COLLEGE VAN DEKANEN IN HET OPENBAAR TE VERDEDIGEN OP

VRIJDAG 19 OKTOBER 1984 TE 16.00 UUR DOOR

CHRISl"IAAN WAL THERUS GITS

GEBOREN TE DORDRECHT

PROF. IR. W. M. J. GERAERDS

Preface

At the Department of Industrial Engineering and Management Science of the Eindhoven University of Technology, research projects

concerning different aspects of maintenance are carried out since the mid 60's.

One of the primary projects is directed at the design of the maintenance concept for a technical system. The maintenance concept, essentially, is the set of ordered maintenance rules prescribing what maintenance should be carried out when. At the start of this project, attention was primarily directed at gaining insight in the relationship between failure behaviour of technical systems and possible maintenance rules influencing that behaviour. Later on, also other aspects of maintenance concept design were investigated. The technical systems studied are very diverse: conveyor belts in the cement industry, civil aircraft, plain paper copiers, medical equipment, buildings, diesel engines in seagoing vessels etc.

,This study is concerned with setting forth a framework for the systematic design of the maintenance concept for a technical system in an using organization, incorporating up to a high extent the results of the forementioned investigations.

It is of interest for the professional workers in the f~eld of maintenance management, who are responsible for the design and upkeep of the maintenance concepts in an organization, as well as for scientific workers, as lacunae in existing knowledge, requiring further research, are identified.

Acknowledgements

The project reported in this book was carried out under the supervision of Prof. W.M.J. Geraerds. Without his inspiring views on the field of maintenance and his continued support, I would never have finished this study.

J.H.J. Geurts provided many suggestions which, although often rather cryptically at first sight, after longer consideration always proved to be to the point.

Prof. W. Monhemius gave many thoughtful comments on earlier drafts and, at least as important, always encouraged me to continue. Last but certainly not least, thanks are due to A.C.J. Kirkels for typing the numerous revisions of the manuscript.

Abstract

This study aims at setting forth a framework that is intended to be useful in the systematization of the design of the maintenance concept for a technical system in an using organization. The maintenance concept, essentially, is the set of ordered maintenance rules pre-scribing what maintenance should be carried out when.

On the one hand, such a framework permits to fit in already available but fragmentary research results, thus identifying the lacunae in knowledge requiring further research. On the other hand, it can be used as an instrument, in actual design of a maintenance concept for a technical system to be introduced in an organization, or to evaluate existing maintenance concepts.

The introductio.n in chapter I shows the relevance of the maintenance concept with respect to maintenance control.

Chapter 2 presents a discussion of the aspects of the technical , system and of the using organization that have to be accounted for

in a framework for maintenance concept design.

In chapter 3, the literature dealing with the question "what maintenance should be carried out when" is reviewed. The contributions are classi-fied into a mathematical approach and an engineering approach. In the mathematical approach, the problem is considerably simplified in order to be able to obtain optimal maintenance policies. In the engineering approach, the real problem is considered, aiming at a satisficing solution.

In chapter 4, a framework for the fundamental design process of the maintenance concept is presented. A fundamental approach is concerned with the way in which ideally the maintenance concept should be designed. The approach is decomposed into the following four phases:

- technical system analysis;

- elementary maintenance generation; - combinatorial maintenance generatio.n; - maintenance concept determination.

Teahniaat system analysis

supplies that information about the technical system which is required in the subsequent phases in maintenance concept design.

Failure behaviour analysis identifies the process-failure combinations, i.e. the failures of the technical system unambiguously related to their underlying physical processes.

Failure consequence analysis concentrates on determining the con-sequences of the process-failure combinations identified.

Hardware structure analysis deals with the replaceability and acces-sibility of the parts of the technical system.

Elementary maintenanae generation

aims at relating a set of alternative elementary maintenance rules to each process-failure combination

distinguished, on the basis of its behaviour and taking into account its consequences and the constraints imposed by the using organization.

An elementary maintenance rule is a directive prescribing a maintenance operation and when it should be carried out.

With respect to a process-failure combination, it concentrates on the effectiveness of the categories of maintenance activation distinguished i.e.:

failure based maintenance; - use based maintenance; - condition based maintenance.

Further, it deals with the effectiveness of maintenance operations. This is considered to be a technical problem outside the scope of this study, as it requires knowledge and understanding of the physical processes underlying failure, of expected technical consequences of failure and of maintenance technology.

Finally, the nominal maintenance intervals are determined on the basis of trade off of the effort of maintenance and the resulting reduction in failure consequences.

Corrbinator>iaZ maintenanae genemtion aims at transforming the sets of alternative elementary maintenance rules into alternative sets of maintenance rules, taking into account the interdependence of

the elementary maintenance rules and the constraints imposed by the using organizatio.n.

The transformation process requires combining the sets of alternative elementary maintenance rules, each related to a process-failure combination, into alternative sets of elementary maintenance rules, each set connected to the technical system as a whole, eliminating the redundant rules. Further, the individual nominal maintenance intervals, as prescribed by a set of elementary maintenance rules, are combined into common normative maintenance intervals, c.q. instants.

In maintenanae aonaept determination, the alternative sets of maintenance . rules are evaluated with respect to the design objective, and the set

of maintenance rules that yields the best results is selected to constitute the maintenance con;ept.

Chapter 5 is devoted to setting forth a framework for a satisficing design of the maintenance concept. A satisficing approach is concerned with the way in which the maintenance concept is designed in practice, taking into account the effort going into the design. process itself and the paucity of data about failure behaviour of the technical system.

A framework for satisficing design requires adaptation of technical system analysis and maintenance concept determination as presented in the fundamental approach.

Finally, in chapter 6, the main conclusions are summarized, and some proposals for further study are presented.

Samenvatting

Het onderzoek beoogt het opstellen van een kader voor het ontwerpen van het onderhoudsconcept voor een technisch systeem in een gebruikende organisatie.

Onder het onderhoudsconcept wordt het geordende stelsel onderhouds-regels verstaan,dat aangeeft welk onderhoud aan het technisch systeem moet warden uitgevoerd en wanneer dat dient te gebeuren.

In de inleiding (hoofdstuk I) wordt de relatie tussen onderhoudsconcept en de onderhoudsbeheersing van de technische systemen in gebruik in een organisatie belicht.

In hoofdstuk 2 warden de aspecten van het technisch systeem en van de gebruikende organisatie geidentificeerd die verdisconteerd dienen te warden in het onderhoudsconcept,en dientengevolge in het ontwerpkader.

In hoofdstuk 3 wordt·een literatuuroverzicht gepresenteerd aan de hand waarvan de lacunae in de aanwezige kennis met betrekking tot het ontwerpen van een cinderhoudsconcept bepaald warden,

Hoofdstuk 4 is gewijd aan een fundamenteel ontwerpkader dat bestaat uit een viertal fasen:

- het analyseren van het gedrag en.,de structuur van het technische systeem;

- het bepalen van de elementaire onderhoudsregels; - het combineren van deze onderhoudsregels;

- het selecteren van het uiteindelijke stelsel onderhoudsregels.

Indachtig de beperkte kennis van het storingsgedrag, het gebrek aan storingsdata en de ontwerpkosten komt in hoofdstuk 5 het opstellen van een ontwerpkader voor een "satisficing" onderhoudsconcept aan de orde.

Tenslotte warden in hoofdstuk 6 de conclusies s~engevat en enige voorstellen voor verder onderzoek gedaan,

CONTENTS

PREFACE ABSTRACT

SAMENVATTING (Abstract,in Dutch) CONTENTS

I. INTRODUCTION

1.1. Preliminary considerations 1.2. Maintenance control

1.3. Forecasting maintenance demand 1.4. Maintenance concept design

2. MAINTENANCE CONCEPT DESIGN CONSIDERATIONS 2.1. Introduction

2.2. Cost optimization 2.2.1. Definitions

2.2.2. The effectivity of maintenance 2.2.3. The efficiency of maintenance 2.3. The organizational requirements

2.3.1. Maintenance regulations

2.3.2. Safety of the production process 2.3.3. Continuity of the production process 2.4. The maintenance resources

2.5. The regularity of maintenance demand 2.6. Conclusion

3. LITERATURE REVIEW 3.1. Introduction

3.2. The evaluation criteria 3.3. The mathematical approach

3.3.1. Simple, preventive, two-state models 3.3.2. Complex, preventive, two-state models 3.3.3. Simple, preparedness, two-state models 3.3.4. Complex, preparedness, two-state models 3.3.5. Multi-state models vii ix xii xiii

7

10 12 12 12 13 17 20 24 24 25 25 28 31 36 39 39 39 41 42 44 45 46 46

3.4. The engineering approach 3.4.1. Geraerds

3.4.2. Kelly

3.4.3. Nowlan and Heap 3.5. Evaluation and conclusions

3.5.1. Evaluation of the mathematical approach 3.5.2. Evaluation of the engineering approach 3.5.3. Conclusions

4. A FRAMEWORK FOR FUNDAMENTAL DESIGN 4.1. Introduction

4.2. Technical system analysis 4.2.1. Introduction

4.2.2. Failure behaviour analysis 4.2.3. Failure consequence analysis 4.2.4. Hardware structure analysis 4.3. Elementary maintenance qualification

4.3.1. Introduction

4.3.2. Maintenance activator qualification 4.3.3. Maintenance operation qualification 4.4. Elementary maintenance quantification 4.5. Combinatorial maintenance qualification

4.5.1. Introduction

4.5.2. Maintenance rule classification 4.5.3. Maintenance rule elimination 4.6. Maintenance interval clustering

4.6.1. Introduction

4.6.2. Maintenance rule classification 4.6.3. Nominal interval combining 4.7. Maintenance demand structuring

4.7.1. Introduction

4.7.2. Periodic maintenance demand pattern 4.7.3. Cyclic maintenance demand pattern 4.8. Maintenance concept determination

47 47 50 52 54 54 55 57

59

59

65

65 66 67 68 70 70 7J 73 75 77 77 78 80 84 84 85 85 88 88 89 90 93

5. A FRAMEWORK FOR A SATISFICING DESIGN

5.1. Introduction

5.2. Failure behaviour information 5.3. Technical system analysis

5.4. Maintenance concept determination

6. CONCLUSIONS AND PROPOSALS FOR FURTHER STUDY

6. I . Conclusions

6.2. PrQposals for further study

Appendix A: List of abbreviations Appendix B: Glossary of terms References 95 95 96 98 104 108 108 1 1 1 113 114 118

1 • INTRODUCTION

1.1. Preliminary considerations

For a long time, maintenance in industrial organizations was not considered to be a problem requiring a systematic approach. The low degree of utilization and specialization of the technical systems in use in an organization made failure consequences negligible. Maintenance consisted of corrective maintenance only; attention concentrated on maintenance technology. However, changes inside and outside the organization, such as the growing attention for environmental

problems, the introduction of large and complex technical systems in the production process, increasing labour cost and increasing competition, have made improvement of maintenance control an economic and societal necessity.

Improvement of maintenance control presupposes the introduction of preventive ~aintenance, which has to be based on the behaviour of the technical systems. Study of this behaviour results in a growing awareness of the imperfections of the technical systems in use with respect to an attractive behaviour from the point of view of mainte-nance. As a consequence, maintenance interests over the last decade extended as far as the specification of design requirements, which is known as terotechnology.

From the foregoing considerations, i t will be clear that in an industrial organization maintenance control should be a central mssue in mainte-nance management.

1.2. Maintenance control

In accordance with a systems approach to organizations (e.g. Bertalanfy, 1956; Boulding, 1956), an organization will not be considered as a static combination of components but as an ensemble of processes.

From the point of view of maintenance, the primary process in an organization is production, which transforms the production input

product. All types of equipment needed in connection with this transformation function will be called technical systems. A

teahniaal system (TS) is defined as a set of interrelated parts fulfilling a specifiable production function.

External causes, ageing and the use of the TS's inflict the physical TS properties considered relevant for fulfilment of their function. This results in an inevitable secondary production output: maintenanee demand. Execution of maintenance leads to a secondary production input:

potential produation aapaaity.

In this view, depicted in Fig. 1.1., maintenanae of a TS is the total of activities aiming at retaining each part of the TS in, or restoring it to, the physical state considered necessary for fulfilment of its function. primary inpu t seeondary produation input d pro uat~on produation maintenance duation ~t primary pro out seaondary produation out~t

Fig. 1.1. The relationship between produation and maintenanae

It should be noted that this definition of maintenance explicitly excludes:

- replacing the TS as a whole;

- changing the design of the TS (e.g. modification, design out maintenance);

- providing the primary input of the production process.

Maintenance defined as forementioned is a process, and, as an uncontrolled process, depicted in Fig. l.2.a., due to disturbances

distUPba.nae maintenance ._ demand ~~--~ maintenance exeaution __ potential 1----..,.p:t'oduation aapaaity

Fig. 1.2.a. The maintenance process, uncontrolled

maintenan demand forecast ce

--. "

t.

'l-n 0:1:'1'/1(; -z.on

r

maintenance cont:t'ol ~ 'ntewentions info:t'mation maintenanc demand e ~ maintenance execution

Fig. 1.2.b. The maintenance p:t'ocess, cont:t'olZ.ed

--

potential production capacity

will not always perform satisfactorily, the realization of its eventual result within desired norms requires control of bhat process as illustrated in Fig. 1.2.b.

Maintenance control has to be based on a forecast of maintenance demand in order to be able to decide on the necessity of intervention in the maintenance process. In principle, it should at each point in time evaluate all alternatives with respect to these decisions, and choose the best set of decisions. However, the influence of each decision on the objective of control is not necessarily clear and straightforward. Furthermore, the effect of a specific decision may depend on other decisions. This makes maintenance control a complex problem. Much of this difficulty can be dealt with by means of structuring the control problem into a number of separate less

complex control problems (Bertrand and Wortmann, 1981).

Structuring consists of decomposition and coordination.

Decomposition

aims at breaking down a complex control problem into a number of separate, less complex, subproblems. It will be clear that the process of determining viable ways of decomposing a problem is not straightforward. It depends to a large extent on existing knowledge about the problem at hand, on the aspects to be included, and, last but not least, on the preferences of the designer or the design team.

Coordination

refers to the problem of translating the overall objective of control into local objectives for each decision function distinguished, in such a way, that striving for attainment of the local objectives eventually results in realizing the overall objective.

Structuring the control problem in an organization has received much attention in the literature.(see e.g. Parsons, 1960; Ansoff, 1965; Anthony, 1965) •

A detailed evaluation of these structures with respect to their applicability in maintenance is considered to be outside the scope of this study. Within the context of maintenance concept design it is considered sufficient to group the maintenance control decisions according to their

futurity

(Drucker, 1974), which indicates the period of time over which a decision commits organizational resources. This results in distinguishing the following three levels in

maintenance control:

- long-term maintenance control; - mid-term maintenance control; - short-term maintenance control.

Long-term maintenance control

refers to the process of deciding upon the objectives of maintenance, on changes in these objectives, on the resources used to attain these objectives, and on the policies that are to govern the acquisition, the use and the disposition of .these resources.

Long-term maintenance control concerns decisions which are required when substantial changes are going to take place in the organization

with respec.t to the TS' s to be maintained. Such drastic changes are the introduction of a new type of TS, which incorporates new

technologies, or a large increase c.q. decrease in production volume.

Mid-term maintenanee

eont~ol refers to the process which ensures that the existing resources of maintenance are adjusted within the boundaries of the mid-term period, and which further ensures that these resources are used effectively and efficiently in the accomplishment of the long-term objective of maintenance.

Mid-term maintenance control involves both planning and controlling activities although these activities cannot always b4i! separated clearly. Maintenance planning is the decision process which results in the maintenance plan, specifying, over a plan period, '.what maintenance has

to be carried out, at what times and the norms to be used.

Sho~t-term

maintenanee

aont~ol refers to the process of assuring that individual TS-maintenance demand is carried out effectively and efficiently.

The mainiaoniJ:vities are p~oaess

planni'YI{l

(Buffa, 1965), which aims at preparation and presentation of the work to be carried out in a format appropriate for the subsequent activities,

saheduling,

which aims at the ordering for timely execution of the tasks to be carried out,

loading,

which aims at taking care of fluctuations in the

workload in respect of available capacity, and spare p~t

provisioning,

which aims at the timely availability of the spare parts required.

In this study, it will be assumed that the long-term maintenance objectives and the maintenance resources are given. Consequently, no attention will be paid to long-term maintenance control and

to the adjustment of the existing maintenance resources.

Taking the foregoing simplifications into account, maintenance control, as shown in Fig. 1.3, concerns the allocation of maintenance resources to maintenance operations requiring these resources.

maintenance objectives

maintenance resources

maintenance

demand

forecast

mid-term

maintenanoe

p~anning p~anned

maintenance

demand

feedback

unp~anned mainte~n~~--~~

demand

short-term

maintenance

contra~

feedback

execution

potentia~

1---....,

production

capacity

Fig. 1.3. Maintenance

p~nning

and

contro~

in the using organization

Production control, which has been studied more comprehensively than maintenance control, essentially deals with a similar allocation problem.

Geraerds (1972) presents a comparision of the production and maintenance processes to reveal theoretical elements designed for production which are also applicable in maintenance and, further, to reveal the lacunae which require a separate approach. His conclusions can be summarized as follows:

- Analogies between production and maintenance exist in

mid-term

maintenance

p~anning,

process

p~anning, sahedu~ing and ~oading.

There is no need for an independent theoretical approach to these aspects in maintenance. Research would have to be directed at adaptation of the methods designed in production to include typical aspects of maintenance.

- FoPecasting maintenance demand

requires a different approach, primarily to be based on the study of the failure behaviour of the TS's to be maintained, as the methods used in market research to forecast demand for physical products are not applicable.

- Relatively more simple than in production control is the

reconciZiation

of

maintenance objectives

and

production wishes,

as the TS's to be maintained in an organization are a closed group and are known beforehand. Reconciliation requires a separate theore-tical approach if production a~ maintenance can be controlled within the same organization.

Inventory controZ for maintenance

requires an independent theoretical approach as conventional statistical inventory control theory covers a small and trivial fraction of the problem area only.

The decision which topic has to be studied first cannot be made on the basis of scientific considerations alone. In my opinion, within the context of maintenance control, forecasting maintenance demand is of primary importance as the two other lacunae, maintenance and production reconciliation and spare part inventory control, both require information about future maintenance operations.

1.3. Forecasting maintenance demand

Forecasting maintenance demand is the process supplying the information which forms the basis for mid-term maintenance planning. This information

specifies what maintenance should be carried out at what instants of time in the life of the TS's.in use in the organization concerned. The fact that the TS's form a closed group and are known beforehand makes it possible to consider the forecast of maintenance demand as

the superposition of the forecasts of TS-maintenance demand, i.e. maintenance demand resulting frQm one TS.

Forecasting TS-maintenance demand has to be based on the

maintenance

aonaept

for the TS, which is the ordered set of maintenance rules connected to the TS.

A

maintenance PUZe

(MR) is a directive prescribing a collection of maintenance operations and "when" it should be carried out. "When"

in this definition denotes the event activating ·maintenance execution. The event may be the occurrence of failure, the TS reaching a specific age in units of use, the occurrence of a specific opportunity, etc. The fact that the maintenance concept consists of a number of maintenance rules makes it possible to consider the forecast of TS-maintenance demand as the superposition of the forecasts of MR-maintenance demand, i.e. maintenance demand resulting from one MR.

Forecasting MR-maintenance demand, as depicted in Fig. 1.4, consists of the following three steps:

- maintenance requirement determination; - use-time conversion; - due-date assignment. produat1-on requirem·~e~n~tr---.! maintenance

rule

maintenance requirement determinatiOn MR-maintenance requirement use-time conversion deterministic MR-maintenance alaim due-date assignment

non-,.plcmnable

MR-maintenance de.l'f'I(11U]

·i? lannab 'le MR- maintenance demand

Maintenance requirement deter-mination

refers to establishing the MR-maintenance requirement specifying use-instants in the life of the TS at which the collection of maintenance operations prescribed by a MR should be carried out.

With respect to the information about the use-instant, two types of MR-maintenance requirements can be distinguished:

- probabilisitic MR-maintenance requirement: maintenance requirement resulting from a non-periodic MR, in which the use-instants can be expressed in probabilistic terms only;

- deterministic MR-maintenance requirement: maintenance requirement resulting from periodic MR1_{s, in which the use-instants are known}

with certainty.

Use-time conversion

refers to establishing the MR-maintenance claim, specifying the time-instants in the life of the TS at which the prescribed collection of maintenance operations should be carried out. It is primarily determined by the MR-maintenance requirement.

Furthermore, it is influenced by the

operation intensity

of the TS, which is the ratio of the relevant measure of use, measured over a certain period of time, and that period of time.

The operation intensity is determined by the

TS-production requirement,

specifying future use of the TS in the production process.

With respect to the information about the time-instant, it is evident that a probabilistic MR-maintenance requirement results in a listic MR-maintenance claim, specifying the time-instant in probabi-listic terms only.

A determinis.tic MR-maintenance requirement results in a deterministic MR-maintenance claim, if the operation intensity is predictable. If the operation intensity is not predictable, then a deterministic MR-maintenance requirement results in a probabilistic MR-maintenance claim.

Due-date assignment

refers to generating the dates before which the prescribed collection of maintenance operations should be carried out. A probabilistic MR-mainteannce claim does not allow due-date

A deterministic MR-maintenance claim in combination with the date of introduction of the TS in the organization, does yield a due-date and results in

ptannable

MR~aintenanae

demand.

From the foregoing discussion, it can be concluded that, as the date of introduction of the TS is given, and the TS-production requirement has to be regarded as given, from the point of view of maintenance, forecasting maintenance demand requires further investigation of the maintenance concept.

1.4. Maintenance concept design

Maintenance concept design in this study concerns devising the main-.tenance concept for a TS in an using organization.

Research with respect to a diversity of maintenance concept design aspects has been carried out at the Eindhoven University of Technology in the context of M.Sc. projects (e.g. Van Geijn and Ramaekers, 1975; Molenaar, 1976; Van den Brink, 1976; Bergmans, 1978; Ruisendaal, 1979; Delnoy, 1980; Rensen, 1981; and Smetsers, 1982). These studies showed that the available techniques and knowledge from existing maintenance theory can rarely be applied in a straightforward manner, due to the wide gap existing between the relatively simple problems subjected to analytic treatment, and the complex nature of the maintenance concept design problem in practice.

This study, carried out as a Ph.D. project, aims at setting forth a framework for design of the maintenance concept, incorporating the insights gained from the forementioned research projects. Such a framework introduces a generalized ordering of design steps, which on the one hand assists in logically fitting in already

available, but fragmentary, research results, and which on the other hand assists in identifying the lacunae in knowledge requiring further research.

Chapter 2 concentrates on identification of relevant aspects of the TS and of the using organization, that have to be taken into account in a framework for the design of the maintenance concept.

In chapter 3, the literature dealing with the question what maintenance should be carried out when is reviewed and evaluated. Chapter 4 is devoted to setting forth a framework for fundamental design of the maintenance concept, with the objective to show the way in which the maintenance concept

should

be designed. Establishing a framework for satisficing design of the maintenance concept, showing the way in which the maintenance concept can be designed in practice, taking into account the design effort and incompleteness of failure behaviour information, is the subject of chapter 5.

Finally, in chapter 6, the main conclusions are summarized, followed by some proposals for further study.

2. MAINTENANCE CONCEPT DESIGN CONSIDERATIONS

2.1. Introduction

This chapter aims at identifying the relevant aspects of the,TS and of the using organization which have to be taken into account in a framework for design of the maintenance concept.

From the discussion about the relationship between maintenance concept and maintenance control, presented in chapter I, it follows that the maintenance concept should contribute to attainment of the long-term objective of maintenance and should take into account the existing maintenance resources.

It will be assumed that, in general terms, the long-term objective of maintenance can be expressed as striving for cost optimization, while meeting the organizational requirements. These elements will be explored in sections 2.2 and 2.3 respectively.

Section 2.4 deals with the impact of the available maintenance resources on the concept. The quality of short-term maintenance control that can be achieved in an organization up to a high degree depends on the regularity of maintenance demand. The short-term maintenance control wishes with respect to regularity will be discussed in section 2.5. Section 2.6. presents the eventual propositions with respect to maintenance concept design.

2.2. Cost-optimization

The cost of maintenance execution have to be borne by the products manufactured by means of the TS. Therefore, optimization of maintenance has to be found in criteria which directly or indirectly minimize these cost. This necessitates to concentrate on the effectiveness as well as on the efficiency of maintenance as prescribed to be

carried out according to the maintenance concept. But before discussing these subjects, a number of key words will be defined as understood in this study.

2.2.1. Definitions

Production essentially is interested in the

produation fUnction

of the TS which is the set of performance characteristic activities production requires to be realized by the TS. Such a function may have a composite structure consisting of a number of sub-functions, which are of different importance and which, in addition, may be

combined into various

opePational modes.

In this study, it will be assumed that the TS functions in one operational mode only.

With respect to the production function of TS, the following two functional states of a TS will be distinguished:

- available:

the functional state in which the TS is fulfilling its production function, or can fulfil its production function when called upon; and

- unavailable:

the functional state in which the TS cannot fulfil its production function due to maintenance to be carried out.

Maintenance, as defined in section 1.2.1., essentially is a technical process which aims at keeping the TS in, or returning it to, the available state by means of altering the physical properties of its parts.

The

aondition

of a part of the TS is the value of one of these, clearly defined, physical properties.

With respect to the physical ability of a part, to fulfil its function, the following two states will be distinguished:

- operable:

the physical state of the part that is considered sufficient for fulfilment of its function; and

inoperable:

the physical state of the part that is considered insufficient for fulfilment of its function.

Failure

of a part is defined as the transistion of the part from the operable into the inoperable state.

Failure consequences are damage inflicted upon the TS and its environment as well as functional consequences concerning the effect of failure on the function of the part and, eventually, on the production function.

With respect to the effects of failure on the function of a part, the following two types of failures can be distinguished:

- absolute failure:

a failure which results in the physical impossibility of a part to fulfil its function, and

- normative failure:

a failure due to the passing of a specified ,threshold by a failure relevant physical property (e.g. thread depth

of a tyre).

The rationale of normative failure in particular lies in the expected occurrence of a

multiple failure,

defined as a sequentially generated series of failures with consequences which are incomparably more serious than the originating failure in isolation.

With respect to the observability of failure, the following two types of failure can be distinguished:

- evident failure:

a failure, the occurrence of which signals itself directly during.normal operation of the TS; and

hidden failure:

a failure, the occurrence of which is not signalled during normal operation of the TS.

Detection of a hidden failure requires a separate inspective maintenance effort.

With respect to the transistion of the TS from the available into the unavailable state, a distinction has to be made between:

- breakdown:

the transistion of the TS from the available into the unavailable state, due to absolute failure; and

- shutdown:

the transistion of the TS from the available into the unavailable state, due to a maintenance control decision.

The rationale of maintenance lies in reduction of failure conse-quences. With respect to the objective to be reached by maintenance execution, maintenance operations can be divided into the following two categories:

aorreative maintenanae

(CM): maintenance with the objective to restore a part of a TS to the operable state; and

-preventive maintenanae

(PM): maintenance with the objective to retain a part of a TS in the operable state.

Maintenance execution is primarily activated by the maintenance concept for the TS, as it essentially prescribes what and "when" maintenance should be carried out.

With respect to a well defined failure, the maintenance concept can prescribe maintenance to be activated by:

- the

event

of failure, or by

- a

maintenanae control decision,

before the event of failure.

Failure based maintenance

(FBM) prescribes maintenance to be generated by the event of failure only.

From the point of view of effectivity and control, it is essential that the decision to activate maintenance is based on

failure behaviour

characteristics,

which are properties meeting the following criteria:

-relevance,

with respect to the failure concerned;

- quantifiability,

in order to measure to what extent the objectives are met;

- forecastability.

With respect to forecastability, the following two types of models can be distinguished:

- probabilistic models; and - mechanistic models.

In the

probabilistia model,

forecasting of the event of failure is based on (the distribution of) the duration of failure free operative periods. Based on this failure probability density function, f(t), the failure behaviour characteristics reliability and failure rate can be established.

Barlow and Proschan (1965) defin~

reliability,

R(t) as the probability of an object performing its purpose adequately, for the period of time intended, under the operating circumstances encountered.

In mathematical terms:

""

R(t)

I

f(T)dT

t

They define failure rate, z(t) by:

z(t) = f(t)

R(t)

(2.1)

(2.2)

So that z(t)d(t) represents the probability that an object of age t will fail in the interval (t, t+dt).

Use based maintenanee

(UBM) is maintenance based on the probabilistic model. It prescribes maintenance to be carried out after a specific period of use, or after the event of failure, if failure occurs earlier.

In the

meohanistie model.,

prediction of individual failures is based on the understanding of the physical process underlying failure during failure free periods.

Geraerds (1972) presents a fundamental model required to allow individual prediction of those failures of a TS, which are the result of under-lying physical processes, which show properties that change from an initial value to a fatal limit, being the value at which failure occurs. Such properties can show typically different behaviour, as illustrated in Fig. 2.1. Curve C is, for practical purposes, rather unsuitable as the point at which the fatal limit will be crossed varies considerably if the value of the property varies a little, or if measurement is not very accurate.

Failure prediction pPopePty

(FPP) is a property showing behaviour

according to curves A and B which have the potential to predict failures.

Condition based maintenance

(CBM) is maintenance based on the mechanistic model. It prescribes maintenance primarily based on assessing the value of a FPP and, if a predetermined norm has been reached, the remedial activities to be executed.

CBM is usually, but not necessarily, carried out after fixed intervals of use.

physiaaJ

property

use

-Fig. 2.1. Enduranae ourves (Geraerds, 1972)

From the foregoing discussion about categories of maintenance operation and categories of maintenance activation, it must be concluded that FBM can result in CM only. UBM and CBM primarily result in PM, but in CM for the fraction of failures that will turn up in spite of maintenance execution at the times prescribed by PM.

With respect to the consequences of maintenance execution on the production function of the

rs.

a distinction has to be made between:

- running maintenanae:

maintenance that can be carried out while the TS is in the available state, or that requires the TS to be operative; and

shutdown maintenanae:

maintenance that requi.res the TS to be ino.perative, consequently in the unavailable state.

2.2.2. The effectivity of maintenance

Considerations about the effectivity of maintenance concern the possibility of achievement of the objective pursued. As a maintenance concept prescribes "when" and what maintenance should be carried out, a distinction has to be made between the effectivity of:

- maintenance activation in view of the correctness at the moment prescribed; and of

the maintenance operation to be carried out in view of its achievement of the physical state desired.

With respect to maintenance activation, the following three

aategoriea of maintenanae aativation

have been distinguished: -failure based maintenance (FBM);

-use based maintenance (UBM); -condition based maintenance (CBM).

From section 2.2 it follows, that FBM is effective in any case, and that CBM is effective if a FPP is known.

UBM is effective if it results in a reduction in the probability of failure after PM execution. This implicates to concentrate on the failure rate. With respect to the failure rate, ideally, three distinct types of failure behaviour can be distinguished, characterized by the failure rate decreasing, constant or increasing.

In case of a decreasing failure rate (Fig. 2.2.a), execution of PM results in an increased probability of failure after its execution. In case of a constant failure rate, (Fig. 2.2.b) execution of PM does not alter the probability of failure.

In case of an increasing failure rate (Fig. 2.2.c), execution of PM results in a reduction of the probability of failure after its execution. Thus, UBM is effective only if the failure rate increases.

It should be noticed that a constant failure rate of a TS, as Cox and Smith (1953, 1954) have shown, may be the result of a superposition of a number of failure processes (Fig. 2.3). The time between

superposed failures of a TS tends to become exponentially distributed, irrespective of the distributions of the underlying individual

failure~. Consequently, considerations about the effectivity of UBM should fundamentally be based on individual failures.

Conside~ations about the effectivity of maintenance operations to be carried out require knowledge and understanding of the physical

z(u)

1

z(u)

t

z(u)

f

use -

U'

use

-(a) Decreasing failure rate (DFR)

z(u)

f

u s e - use

-(b) Constant failure rate ( CFR)

z(u}

i

u s e - U'

use-(c) Increasing failure rate (IFR) Fig. 2.2. The effeat of UBM with interval U' on the failure rate

supe:r>position

'

f

'

t •

•

f

(A+B)

I I

_I

_I

I

I

I

I

_I

failure proaess

I

I

I

I

I

part A

•

_{J •}

_I

*

I

failure proaesa

I

I

I

J

part B

•

~

u s e _ l(: _failure

processes underlying failure and of maintenance technology, and therefore belong to the fields of. the technical disciplines involved.

The foregoing discussion concentrated on the effectiveness of

maintenance with respect to a single failure. However, considerations about the effectivity of maintenance with respect to a part subject to a diversity of failures should further take into account that replacement of the part in respect of maintenance required as a result of one of its failures, not only effects the condition concerned, but also includes a return to their initial values of the physical processes underlying the other failures of the part.

Consequently, replacement of a part makes those maintenance operations, which would have to be carried out at a later time than the replacement of the part, superfluous.

2.2.3. The efficiency of maintenance

Considerations about the efficiency of maintenance have to be based on a trade off of the effort of maintenance execution and the resulting reduction in failure consequences.

FBM, to be carried out as prescribed in the maintenance concept, results in CM.

With respect to absolute failure, the ao~eative maintenanae aost

(CMC) can be expressed as:

CMC BDC + CMEXC + EDC

in which:

BDC breakdown cost; CMEXC: CM execution cost;

EDC environmental damage cost.

(2.3)

The breakdown cost consist of the cost associated with the incon-venience caused by the event of breakdown and the cost of production loss.

CM execution cost are the cost of the activities involved in restoration of the parts of the TS to their operable states. The cost of environmental damage are the cost of the activities required to restore the environment of the TS to the state considered necessary, and possibly, a penalty.

The reduction of failure consequences achieved by FBM, related to an absolute failure follows from the fact that the TS is restored to the available state without renewal of the TS as a whole.

With respect to normative failure, CMC can be expressed as:

CMC SDC + CMEXC

in which:

SDC shutdown cost CMEXC: CM execution cost.

The shutdown cost consist of the cost associated with the event of shutdown and the cost of production loss.

Reduction of failure consequences is found in reduction of the probability of occurrence of multiple failure and eventually in reduction of multiple failure consequences.

(2.4)

In case of hidden failure, FBM primarily prescribes an activity, which aims at detecting whether or not the failure has occurred, followed by a restoration activity, if failure has occurred.

Considerations about the efficiency of FBM with respect to a hidden failure concern the detection interval as this interval determines, on the one hand, the lapse of time during which the failure may go unnoticed, which influences the probability of multiple failures, and, on the other hand, the CMC.

UBM to be carried out as prescribed in the maintenance concept results in PM and CM for the failures which turn up in spite of PM.

The

preventive maintenance cost

(PMC) can be expressed as:

PMC SDC + PMEXC

in which:

SDC shutdown cost PMEXC: PM execution cost.

PM execution cost are the cost of the activities involved in retaining a part of the TS in the operable state.

The

reduction of failure consequences

(RFC) resulting from carrying out PM in accordance with UBM can be expressed as:

FRC

=

CMC - (CM/PM)C in which:

CMC cost of CM, if only CM is applied (CM/PM)C: cost of CM if PM is applied.

Optimal PMC will be achieved for:

max(RFC - PMC)

Equation 2.7, taking into consideration equations 2.5 and 2.6, is identical to:

max[CMC - {PMC + (CM/PM)C}]

As CMC is constant, equation 2.8 comes down to:

min{PMC + (CM/PM}C} (2.5) (2 .6) (2.7) (2.8) (2.9)

With respect to equation 2.9 it should be noted that, on the one hand, PMC decreases with increasing length of the PM interval, on the other hand, (CM/PM)C decreases with decreasing length of the PM interval. Hence, a trade off of the PM and (CM/PM)C will eventually result in the optimal PM interval.

Considerations about the efficiency of CBM concentrate on the interval with which the FPP is assessed and compared with a predetermined norm. With respect to the determination of the FPP-assessment interval, essentially the same argument holds for CBM as for UBM, eventually resulting in the optimal FPP-assessment interval.

The foregoing discussion concentrated on the efficiency of maintenance in view of a single failure. However, a TS will be subject to a diversity of failures and therefore considerations about the efficiency of

maintenance should also take into account the possible reduction of maintenance effort that can be achieved by means of execution of a diversity of maintenance operations at the same time.

Reduction of the•.llllaintenance effort may result from a reduction in: - maitenance execution cost, and/or

- shutdown cost.

Simultaneous execution of a diversity of maintenance operations results in a reduction of maintenance execution cost compared to the cost of execution of these maintenance operations individually, if the operations require the same set up activity to be carried out. A

set up activity

is defined as an activity which has to be carried out to enable execution of the prescribed maintenance operation. Set up activities may result from the accessibility of the parts of the TS for execution of the prescribed maintenance operations, and from the administrative activities which precede the actual execution of the maintenance operations.

The reduction of maintenance execution cost can be introduced in the maintenance concept by means of combining the individual maintenance intervals, based on considerations about the efficiency with respect to failure in isolation, into common maintenance intervals.

Considerations about the reduction of shutdown cost through simultaneous execution of a diversity of maintenance operations, have to distinguish between the cost associated with the event of shutdown and with

A reduction of the cost associated with shutdown can be introduced in the maintenance concept by means of combining the individual maintenance intervals of the maintenance operations requiring shutdown

into common maintenance intervals.

From the point of view of maintenance, a reduction of the cost asso-ciated with production loss can only be achieved by means of

a reduction in the time the TS will be required to be unavailable due to the execution of the maintenance operations concerned.

Simultaneous execution allows parallel execution of operations, which may result in a reduction of the time the TS is unavailable, as only

the operations on the critical path determine the duration of unavailability. However, this requires decisions about the allocation of

capacities-to maintenance operations, and, as these decisions lie in the domain of maintenance control, the resulting reduction in production loss cannot be accounted for in the maintenance concept.

2.3. The organizational requirements

From the total of organizational aspects, the maintenance concept for a TS has to take into account those organizational requirements which have a direct or indirect impact on what and 1_{'when" maintenance}

of the TS should be carried out. These requirements concern: - maintenance regulations,

- safety of the production process, and - continuity of the production process.

2.3.1. Maintenance regulations

A maintenance regulation is a directive prescribing a maintenance operation or a collection of maintenance operations and "when" it should be carried out.

Maintenance regulations may be

aompuZsory,

e.g. issued by external bodies, legislative or insurance, such as the Federal Aviation Authority and Lloyd's Register of Shipping, or may be

reaommended

by the manufacturer of the TS, e.g. the activities that should be carried out after an aircraft is struck by lightning, or after a

rough landing. In thii study it will be assumed that maintenance

regulations of this nature are known, clearly defined, and have to be accepted in the design of the maintenance concept.

2.3.2. Safety of the production process

The requirement of safety of the production process and its environment, results in prescribing PM, possibly in combination with a specific reliability to be achieved with respect to a specified failure. This requirement dominates in the determination of the maximum maintenance interval after which the maintenance operation should be carried out.

2.3.3. Continuity of the production process

The requirement of continuity of the production process concerns those TS's which are expected to function in correspondence with a specific production pattern, defined as the planned sequence of productive and non-productive periods.

The impact of this requirement on the maintenance concept is twofold. First of all, it results in a preference for PM, not only because of the expected contribution of PM to the availability of the TS, but also because it allows some tolerance in the planning of the actual time of execution. This tolerance can be used to take advantage of the opportunity to execute maintenance during periods in which it will not require to interrupt production.

The impact of the continuity requirement on the maintenance concept further depends on the maintenand.e float, defined as the number of identical TS's allowed to be in the unavailable state simultaneously without interruption of the production process. Determination of the size of the maintenance float requires a trade off of the investment in additional TS's and the benefits accrueing from reduction in down time and from better plannability of maintenance demand. This decision lies in the domain of long-term maintenance control, and consequently, the maintenance float has to be regarded as given in the context of maintenance concept design.

With regard to the maintenance float, essentially the following two situations have to be distinguished:

- maintenance float > 0; - maintenance float = 0.

If a maintenance float exists, then the production process depends on the availability of a nominal fraction of all the identical TS's in the organization. The other TS's are allowed to be unavailable without interrupting the production process.

To illustrate the effect of a maintenance float on the maintenance concept, assume a planned periodic maintenance demand pattern, as depicted in Fig. 2.4.

unavailable

state

---PMDI

MTT

---available

state

_{t i m e}

-Fig. 2.4. A

pe~iodia

planned maintenance demand

patte~

No interruption of the production process for execution of planned maintenance will occur if the planned maintenance demand interval meets the following requirement (Geraerds, 1970):

NRP

*

MTT PMDI ~ TNTS - NRP - MFD

in which:

PMDI: planned maintenance demand interval in units of time MTT : maintenance throughput time

TNTS: total number of (identical) TS's NRP number of TS's required by production

(2 .1 O)

MFD maintenance float needed for CM and fluctuations in MTT and

In view of plannable maintenance requirements of the TS, the plannable maintenance requirement interval (PMRI) can be expressed as:

PMRI PMDI

*

OI (2. 11)

in which:

OI: operation intensity, expressed in units of use per unit of time

Given MTT, the continuity requirement can be accounted for in the maintenance concept by means of structuring the maintenance intervals

to be prescribed by the plannable MR's into normative maintenance intervals which meet:

NMI.

l.

in which:

NMI.: ith normative maintenance interval

l.

i 1, 2, 3, .•.••

If no maintenance float exists, then the continuity requirement implies maintenance of the TS preferably to be carried out during the non-productive periods in the production pattern.

With respect to these non-productive periods, the following three patterns can be distinguished:

- continuous;

- intermittent, non repetitive; - intermittent, repetitive.

(2 .12)

In case of a

eontinuous production

pattern~ strict continuity cannot be attained. Generally then, this requirement is relaxed to a minimum availability to be achieved.

A diversity of definitions of availability are in use, depending on the organizational s.etting they are meant for (Lie c. s., 1979). In general terms,

cwailabiUty

can be expressed as the ratio. of the sum of the times that the TS is available for production over a period of time and the period of time over which it is measured.

The minimum availability requirementdontains no information about maintenance intervals in the maintenance concept.

In case of an

inter.mittent, non repetitive produation pattern,

the non-productive periods occur at random. Consequently, making use of them is the sole responsibility of short-term maintenance control. The production pattern contains no information that can be accounted for in the maintenance concept.

In case of an

intermittent, repetitive produation pattern,

the repetitive nature of the occurrence of the non-productive periods makes it possible to account for these periods in the maintenance concept through structuring of the maintenance intervals to be prescribed in the maintenance concept to integer times the periodicity of the production pattern.

In this study, it will be assumed, that the non-productive periods in the pattern are of the same duration. Hierarchic production patterns, which imply non-productive periods of different duration, require a separate theoretical approach.

The"foregoing considerations about the continuity requirement are based on the production patterns of individual TS's. However, in case of a

fleet of TS's

(e.g. busses, aircraft), the production

process consists of fulfilling a number of missions, usually prescribed in a timetable. As the TS's in a fleet are interchangeable, i.e. each TS can perform each mission, the continuity requirement makes it necessary to take into account the production pattern of the fleet as a whole. The production pattern then determines the number of TS's required by production at any point in time. The interchangeability also makes it possible to control the operation intensities of the individual TS's up to a certain degree,by means of assigning the TS's to the missions to be performed.

2.4. The maintenance resources

The types of maintenance resources (e.g. machines, space, tools, manpower and skills) available with~n the organization determine the limitations in the type of maintenance operations which can be

carried out within the organization, thus restricting the maintenance concept for a TS used in that organization. Up to what extent

maintenance operations demanding outside resources are prescribed in the maintenance concept depends on the organizational policies with respect to contracting-out maintenance.

Some maintenance operations require special maintenance facilities such as docks for ships, hangars for aircraft and special bays for motorvehicles. If their utilization is high, then these resources will put a constraint on the planned maintenance demand pattern, i.e. on the planned maintenance demand interval and the maintenance throughput time. Furthermore, waiting for them to become available for use will result in an increase of downtime of the TS's demanding these resources.

The decision to increase the capacity of these resources has to be based on a trade off of the investment necessary against the benefits of the resulting reduction in downtime and increased plannability. This decision belongs to the domain of long-term maintenance control. Consequently, the capacity of these resources has to be regarded as given within the context of maintenance concept design.

Although coping with capacity restrictions essentially is the task of maintenance control, an exception can be made for these

critical

capacities,

as these capacities have a dominating effect on the decisions "when" maintenance should be carried out.

spare

part

is defined as a part of the TS which can be replaced in

maintenance execution.

The spare

part inventory poUey

in an organization has a direct impact on the maintenance concept in view of replacement of parts of ~he TS. Furthermore, organizational considerations may have led to an

eehelonized maintenanee structure,

as depicted in Fig. 2.5. In such a structure specific maintenance resources are centralized, to take advantage of specialization and of the increased rate of demand for certain maintenance operations.

The impact of an echelonized maintenance structure is twofold. Firstly, at a certain level, a maintenance operation is limited