Table Of ContentRELIABILITY FOR ENGINEERS
Other Macmillan titles of related interest
Mechanical Reliability, 2nd edition, A. D. S. Carter
Reliability and Maintainability in Perspective, 2nd edition, D. J. Smith
Reliability for Engineers
An Introduction
Eur Ing Michael Beasley
BSc (Hans), CEng, MIEE, FSaRS
M
MACMILLAN
© M. Beasley 1991
All rights reserved. No reproduction, copy or transmission
of this publication may be made without written permission.
No paragraph of this publication may be reproduced, copied
or transmitted save with written permission or in accordance
with the provisions of the Copyright, Designs and Patents Act
1988, or under the terms of any licence permitting limited
copying issued by the Copyright Licensing Agency, 33-4
Alfred Place, London WC1E 7DP.
Any person who does any unauthorised act in relation to
this publication may be liable to criminal prosecution and
civil claims for damages.
First published 1991
Published by
MACMILLAN EDUCATION LTD
Houndmills, Basingstoke, Hampshire RG21 2XS
and London
Companies and representatives
throughout the world
British Library Cataloguing in Publication Data
Beasley, M.
Reliability for engineers.
1. Reliability engineering
I. Title
620.00452
ISBN 978-0-333-54238-5 ISBN 978-1-349-21369-6 (eBook)
DOI 10.1007/978-1-349-21369-6
To my wife Frances and our children Ian and Susan
Contents
Preface viii
1 Introduction 1
2 Product Development 7
3 The Statistics of Failure-I 16
4 The Statistics of Failure -II 37
5 Component Reliability 61
6 System Reliability 75
7 Methods of Increasing System Reliability 91
8 Fault-tolerant Systems -I 95
9 Fault-tolerant Systems-II 102
10 The Behaviour of Maintained Systems 110
11 Elements of Markov Analysis 118
12 Spare Parts Provisioning 135
13 Software Reliability 143
14 Maintainability 153
15 System Reliability Prediction 161
16 Reliable Systems: Specification, Design, Manufacture and Installation 169
17 Reliability Growth 183
18 Reliability Demonstration 185
19 Some Analytical Methods and Computer Software 197
20 The Final Product 202
21 References 204
22 Bibliography 207
23 Answers to Exercises 209
x
Appendix 1: 2 Distribution Tables for Even-numbered Degrees of
Freedom 224
Appendix 2: 5%, 95%, and Median (50%) Rank Tables 230
Appendix 3: The Elements of Probability Theory 242
Appendix 4: Glossary of Terms 253
Index 256
Preface
This book is the outcome of lectures on reliability which I have given, firstly
in-house in STC plc and later at Hatfield Polytechnic. It is intended to help both
students who are new to the subject and practising engineers who are looking
for help in reliability engineering.
Until my retirement I spent my working life in the telecommunications
industry, and thus I had electrical engineers mostly in mind when writing this
text. However, the theory and principles of reliability engineering are of wide
applicability so that I hope this book will also be useful to workers in other
engineering disciplines.
It is difficult, if not impossible, to learn reliability engineering in complete
isolation. So my warm thanks go to all my former colleagues in STC plc who
by discussions and instruction helped my understanding of the subject; in
particular I should like to mention Dr G. G. Pullum, Mr G. H. P. Breuer and
Mr I. CampbelL
I have also been greatly helped as a part-time lecturer at Hatfield Polytechnic
by my colleagues Mr E. A. De Maria, Mr P. J. Bunn and Mr P. F. Jackson, to all
of whom I owe a debt of gratitude.
There are already a number of good textbooks on reliability; my excuse for
writing yet another is that I believe mine to be of practical help - to the student
as well as to the qualified engineer.
I am grateful to Mrs D. J. Ross and Mrs S.M. Pilkington who typed early
versions of the lecture notes from which this book has been developed; and also
to my wife Frances who had the arduous task of typing most of the final
manuscript.
Finally, it is a pleasure to express my thanks to Mr Malcolm Stewart and his
colleagues of Macmillan Education Ltd for their help in bringing this book to
press.
viii
1 Introduction
1.1 The Study of Reliability
Reliability (which is defined more precisely later) is the study of how, why and
when failures occur. This book does not cover the physics of failure of com
ponents since this is a substantial field in its own right. Although the statistics
of component failure are described, the main purpose of this book is to introduce
the reader to the methods of specification, design, analysis, manufacture and
testing of reliable systems. The systems envisaged are chiefly in the telecom
munications field although the principles introduced here can be used much
more widely. From our viewpoint the study of reliability needs a combination
of electrical engineering (for system analysis), statistics (for the analysis of
failure data), and probability theory (for the prediction of failure patterns).
We shall be concerned chiefly with system failure caused by component
failure (either catastrophic or by deterioration) although software reliability is
also briefly considered. An example of a component's catastrophic failure is a
capacitor which goes open-or short-circuit. Deterioration of a component is
exemplified by a pair of contacts in a relay which gradually develop a high
contact resistance. When the contact resistance becomes excessive then the
relay is considered to have failed.
It is important that the distinction between relevant and non-relevant
failures should be clearly understood. As an example, it would be possible for
a badly-constructed amplifier to oscillate because of the proximity of the input
and output connections. An examination of the amplifier would not reveal a
faulty component yet nevertheless the performance would be unsatisfactory
because of poor design. The amplifier may loosely be called 'unreliable' but its
unsatisfactory performance is due to design inadequacy rather than unreliability.
Similarly a TV receiver which might be dubbed 'unreliable' may be giving an
unsatisfactory picture because the signal strength is too low. The problem then
lies with the siting of the aerial rather than improving the reliability of the
receiver In a similar fashion a user may misunderstand the controls on a piece
of equipment and declare it to be 'unreliable' because he cannot make it work
properly. It is the role of Human Factors to ensure that the controls on a piece
of equipment are easily understood.
1
2 Reliability for Engineers
Because so many issues are involved in the design and manufacture of 'good'
systems, the study of reliability should be regarded as a part of System Effective
ness, the study of which developed historically from that of reliability.
1.2 Historical Survey
Ever since man first made objects, he has been faced with the problem of un
reliability; shoes wore out, spears broke, bridges collapsed. Improvements were
made partly by trial-and-error and partly by a simple process of analysis and
re-design: if a part was of insufficient strength then it was made stronger. Even
in ancient times this process led to some spectacular successes· in the reliability
field. If the builder of the Great Pyramid at Giza had been asked 'How reliable
is your pyramid?', he would probably not have understood the question. But
if he had been asked 'How long do you expect your building to last?', he would
probably have replied 'For ever' and 4500 years can be considered a good
approximation to 'for ever' in terms of the human life span.
Trial-and-error continued for many years to be the way in which manufactured
objects were made more reliable. In the 19th century the steel plates of steam
boilers were at first made thicker than was actually necessary (although the
builders did not know this at the time). Over the years the thickness was
reduced until catastrophic failures showed that the process had gone too far.
Attention to detail in design and emphasis on quality in manufacture were
shown by Rolls Royce in the early years of this century to produce highly
reliable cars. But no coherent theory on how to make reliable objects was
available.
Although a certain amount of work on aircraft reliability in the 1930s was
of a theoretical nature, it was not until the German development of the Vl
weapon in the 1939-1945 war that the basis of reliability theory was really
laid. The Vl design team had great difficulty (fortunately for the British!) in
making the weapon reliable. Only when a reliability assessment was made by
Lusser was it realised that a system relying on many component parts was like
a chain with many links. Since failure of any link causes failure of the whole
chain it is thus necessary for each individual link to be highly reliable if the
complete chain is to be reliable.
Immediately after the 1939-1945 war, little development of reliability
theory took place. However, during the Korean war of 1950-1953 the USA
found that the availability of its fighting aircraft was disastrously low, largely
owing to the unreliability of the avionics, more and more of which was being
carried by the aircraft. Thereafter the US military authorities made a deter
mined and largely successful attempt to build a thorough understanding of
reliability theory and practice.
The pace of understanding of reliability principles increased during the
1950s and 1960s. The next important milestone was the development, again by
