Applied Reliability For Industry 3 1st Edition David Delaux Editor Abdelkhalak El Hami Editor Henri Grzeskowiak Editor

Applied Reliability for Industry 3 1st
Edition David Delaux Editor

Abdelkhalak El Hami Editor Henri
Grzeskowiak Editor
Visit to download the full and correct content document:
https://ebookmeta.com/product/applied-reliability-for-industry-3-1st-edition-david-dela
ux-editor-abdelkhalak-el-hami-editor-henri-grzeskowiak-editor/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Applied anatomy for the FRCA 1st Edition Abdul Syed

Editor Harriet Scott Editor Bobby Krishnachetty Editor
https://ebookmeta.com/product/applied-anatomy-for-the-frca-1st-
edition-abdul-syed-editor-harriet-scott-editor-bobby-
krishnachetty-editor/
Intelligent Analytics for Industry 4 0 Applications 1st

Edition Avinash Chandra Pandey Editor Abhishek Verma
Editor Vijaypal Singh Rathor Editor Munesh Singh Editor
Ashutosh Kumar Singh Editor
https://ebookmeta.com/product/intelligent-analytics-for-
industry-4-0-applications-1st-edition-avinash-chandra-pandey-
editor-abhishek-verma-editor-vijaypal-singh-rathor-editor-munesh-
singh-editor-ashutosh-kumar-singh-editor/
Oxford handbook for medical school First Edition.

Edition Kapil Sugand Editor Miriam Berry Editor David
Metcalfe Editor
https://ebookmeta.com/product/oxford-handbook-for-medical-school-
first-edition-edition-kapil-sugand-editor-miriam-berry-editor-
david-metcalfe-editor/
Artificial Intelligence for a Sustainable Industry 4 0

Shashank Awasthi Editor Carlos M Travieso González
Editor Goutam Sanyal Editor Dinesh Kumar Singh Editor
https://ebookmeta.com/product/artificial-intelligence-for-a-
sustainable-industry-4-0-shashank-awasthi-editor-carlos-m-
travieso-gonzalez-editor-goutam-sanyal-editor-dinesh-kumar-singh-
Routledge Handbook of Applied Sport Psychology: A
Comprehensive Guide for Students and Practitioners 2nd
Edition David Tod (Editor)
https://ebookmeta.com/product/routledge-handbook-of-applied-
sport-psychology-a-comprehensive-guide-for-students-and-
practitioners-2nd-edition-david-tod-editor/
Applied Learning Algorithms for Intelligent IoT 1st

Edition Pethuru Raj Chelliah (Editor)
https://ebookmeta.com/product/applied-learning-algorithms-for-
intelligent-iot-1st-edition-pethuru-raj-chelliah-editor/
Lignans David Barker Editor
https://ebookmeta.com/product/lignans-david-barker-editor/
Walking Landscape and Environment 1st Edition David

Borthwick Editor Pippa Marland Editor Anna Stenning
Editor
https://ebookmeta.com/product/walking-landscape-and-
environment-1st-edition-david-borthwick-editor-pippa-marland-
editor-anna-stenning-editor/
Applied Veterinary Clinical Nutrition Team IRA 2nd

Edition Andrea J Fascetti Editor Sean J Delaney Editor
Jennifer A Larsen Editor Cecilia Villaverde Editor
https://ebookmeta.com/product/applied-veterinary-clinical-
nutrition-team-ira-2nd-edition-andrea-j-fascetti-editor-sean-j-
delaney-editor-jennifer-a-larsen-editor-cecilia-villaverde-
Applied Reliability for Industry 3
Reliability of Multiphysical Systems Set
coordinated by
Abdelkhalak El Hami
Volume 18
Applied Reliability
for Industry 3
Operational Reliability for the Automobile,

Aeronautics, Defense, Medical, Marine
and Space Industries
Edited by
Abdelkhalak El Hami
David Delaux
Henri Grzeskowiak
First published 2023 in Great Britain and the United States by ISTE Ltd and John Wiley & Sons, Inc.
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 and licenses issued by the
CLA. Enquiries concerning reproduction outside these terms should be sent to the publishers at the
undermentioned address:
ISTE Ltd John Wiley & Sons, Inc.

27-37 St George’s Road 111 River Street
London SW19 4EU Hoboken, NJ 07030
UK USA
www.iste.co.uk www.wiley.com
© ISTE Ltd 2023

The rights of Abdelkhalak El Hami, David Delaux and Henri Grzeskowiak to be identified as the authors
of this work have been asserted by them in accordance with the Copyright, Designs and Patents Act 1988.
Any opinions, findings, and conclusions or recommendations expressed in this material are those of the
author(s), contributor(s) or editor(s) and do not necessarily reflect the views of ISTE Group.
Library of Congress Control Number: 2022946026
British Library Cataloguing-in-Publication Data

A CIP record for this book is available from the British Library
ISBN 978-1-78630-693-7
Contents
Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
Phillipe EUDELINE
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
Abdelkhalak EL HAMI, David DELAUX, Henri GRZESKOWIAK
Chapter 1. Durability Approach: Applied to a Vehicle

Lighting Control System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Medoune NDIAYE and Caroline RAMUS-SERMENT
1.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2. Example of a vehicle lighting control system . . . . . . . . . . . . . . . . 2
1.2.1. Risks and reliability requirements . . . . . . . . . . . . . . . . . . . . 3
1.2.2. From failure modes to failure mechanisms . . . . . . . . . . . . . . . 3
1.2.3. From failure mechanisms to physical damage factors . . . . . . . . 5
1.2.4. From physical damage factors to mission profiles
or customer usage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.2.5. From failure mechanisms to component part strength
distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.2.6. Resistance distribution chart . . . . . . . . . . . . . . . . . . . . . . . 11
1.2.7. Proposal and study of a validation plan using the
stress–strength method: various real-world examples . . . . . . . . . . . . 14
1.3. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.4. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
vi Applied Reliability for Industry 3
Chapter 2. Structural Diagrams to Validate the Reliability

of Mechanical Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Paul SCHIMMERLING
2.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.2. Choice of methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.2.1. Criteria selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.2.2. Four basic methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.2.3. An applied example: the validation of disc brake pads . . . . . . . . 24
2.3. Feasibility study on the four methods . . . . . . . . . . . . . . . . . . . . 25
2.3.1. Animation principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.3.2. Comparison of Weibull laws under testing and in service . . . . . . 25
2.3.3. Comparing degradation under testing and in service . . . . . . . . . 28
2.3.4. Stress–strength method . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.4. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.5. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Chapter 3. How to Put an Efficient Methodology to

Design Innovative Products in Place . . . . . . . . . . . . . . . . . . . . . . 39
Claire SCHAYES, Ludovic NGAVOUKA and Eric MANOUVRIER
3.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.1.1. Reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.1.2. Variability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.1.3. “Lean Six Sigma” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.1.4. Quality according to the “Lean Six Sigma” approach
“is conforming to requirements” . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.2. DFSS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.3. DMAIC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.3.1. Introduction to DMAIC . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.3.2. Why launch DMAIC projects? . . . . . . . . . . . . . . . . . . . . . . 46
3.4. Feedback . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.4.1. Feedback on the define phase . . . . . . . . . . . . . . . . . . . . . . . 50
3.4.2. Feedback on the measure phase . . . . . . . . . . . . . . . . . . . . . 50
3.4.3. Feedback on the analyze phase . . . . . . . . . . . . . . . . . . . . . . 51
3.4.4. Feedback on the innovation phase . . . . . . . . . . . . . . . . . . . . 52
3.4.5. Feedback on the control phase . . . . . . . . . . . . . . . . . . . . . . 53
3.4.6. Can DMAIC be customized? . . . . . . . . . . . . . . . . . . . . . . . 54
3.5. How to design a reliable welding process with control
over the design of experience? . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.6. Definition of the objectives . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3.6.1. Determining the study space . . . . . . . . . . . . . . . . . . . . . . . 59
3.6.2. Building the DOE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Contents vii
3.6.3. Conducting the tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

3.6.4. Analyzing the results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.6.5. Process optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.6.6. Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.7. Big Data and process? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.8. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
3.9. Appendix 1: example of an ANOVA study . . . . . . . . . . . . . . . . . 74
3.10. Appendix 2: studying the variability of cycle times . . . . . . . . . . . 79
3.11. Appendix 3: example for the use of traditional statistics in Big Data . 87
3.12. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
Chapter 4. Reliability Study of the High Electron Mobility

Transistor (HEMT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Abdelhamid AMAR, Bouchaïb RADI and Abdelkhalak EL HAMI
4.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
4.2. HEMT technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
4.3. HEMT thermal modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
4.4. Reliability methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
4.4.1. Reliability study. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
4.4.2. Calculating the probability of failure . . . . . . . . . . . . . . . . . . 97
4.5. Thermo-reliability coupling . . . . . . . . . . . . . . . . . . . . . . . . . . 101
4.6. Calculating HEMT reliability . . . . . . . . . . . . . . . . . . . . . . . . . 102
4.7. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
4.8. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Chapter 5. Warranty Cost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

David DELAUX
5.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
5.1.1. The evolution of the warranty . . . . . . . . . . . . . . . . . . . . . . 107
5.1.2. The warranty cost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
5.2. Warranty and reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
5.2.1. Qualitative analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
5.2.2. Quantitative analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
5.3. Reliability estimation models . . . . . . . . . . . . . . . . . . . . . . . . . 113
5.3.1. Parametric, non-parametric and other models . . . . . . . . . . . . . 113
5.3.2. Mixed models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
5.3.3. Advantages and disadvantages . . . . . . . . . . . . . . . . . . . . . . 116
5.4. New models for estimating reliability from warranty costs . . . . . . . . 117
5.4.1. Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
viii Applied Reliability for Industry 3
5.4.2. Definition of the transition between “random” and

“wear-and-tear” phases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
5.4.3. New operational reliability model for the “random” phase . . . . . 125
5.4.4. New operational reliability model for the
“wear-and-tear” phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
5.5. Applied automotive case studies . . . . . . . . . . . . . . . . . . . . . . . 126
5.6. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
5.7. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
Chapter 6. Reliability Evaluation of a Luxury Watch Product:

Application of the Stress–Strength Method to a
Mechanical Component . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
Matthieu SALLIN and Anthony PONCET
6.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
6.2. Presentation of the watch and its case study. . . . . . . . . . . . . . . . . 136
6.2.1. The mechanical watch . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
6.2.2. Case study of the barrel spring . . . . . . . . . . . . . . . . . . . . . . 137
6.2.3. Identification of failure modes and damaging factors. . . . . . . . . 137
6.3. Evaluation of the customer usage profile . . . . . . . . . . . . . . . . . . 138
6.3.1. Classifying usage typologies . . . . . . . . . . . . . . . . . . . . . . . 138
6.3.2. Statistical quantification of usage . . . . . . . . . . . . . . . . . . . . 139
6.4. Characterizing experimental reliability . . . . . . . . . . . . . . . . . . . 140
6.4.1. Performance of failure tests . . . . . . . . . . . . . . . . . . . . . . . . 140
6.4.2. Evaluation of the accelerated lifetime law . . . . . . . . . . . . . . . 141
6.4.3. Constructing the law of resistance . . . . . . . . . . . . . . . . . . . . 142
6.5. Reliability evaluation of customers . . . . . . . . . . . . . . . . . . . . . . 143
6.5.1. Reliability calculation using the stress–strength method . . . . . . . 143
6.5.2. Transformation of the stress profile . . . . . . . . . . . . . . . . . . . 144
6.5.3. Numerical application to the barrel case study . . . . . . . . . . . . . 146
6.6. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
6.7. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
Chapter 7. RBDO of the High Electron Mobility Transistor. . . . . . . 149

Abdelhamid AMAR, Bouchaïb RADI and Abdelkhalak EL HAMI
7.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
7.2. Description of the HEMT technology . . . . . . . . . . . . . . . . . . . . 151
7.3. Electrothermomechanical modeling of HEMT . . . . . . . . . . . . . . . 152
7.3.1. Electrothermal modeling of HEMT . . . . . . . . . . . . . . . . . . . 152
7.3.2. Thermomechanical modeling of HEMT . . . . . . . . . . . . . . . . 154
7.4. Reliability methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
Contents ix
7.5. Reliability analysis of HEMT . . . . . . . . . . . . . . . . . . . . . . . . . 156

7.6. Reliability optimization of systems . . . . . . . . . . . . . . . . . . . . . . 158
7.6.1. The classic RBDO approach . . . . . . . . . . . . . . . . . . . . . . . 158
7.6.2. The hybrid RBDO approach . . . . . . . . . . . . . . . . . . . . . . . 159
7.7. HEMT reliability optimization using the hybrid RBDO approach . . . 160
7.7.1. Description of the optimization problem . . . . . . . . . . . . . . . . 160
7.7.2. Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
7.8. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
7.9. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
List of Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
Summaries of other volumes . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

Foreword
Predicting and then guaranteeing the reliability of an industrial system is a major

challenge for manufacturers in the automotive, aeronautics and defense industries, as
well as for those in the rail, telecommunications, nuclear and medical sectors,
among many others. But above all, it is a major challenge for us, who use this
equipment on a day to day basis, who must have absolute confidence in the
information transmitted and the decisions made in real time. The increasing
development of connected objects (autonomous cars, home automation, etc.) will
lead to a drastic reduction of human intervention in favor of the intervention of
mechatronic systems. All these systems can only be deployed if users have absolute
confidence in the reliability of the equipment.
These forms of equipment are divided into two main functionalities: a hardware
part mainly composed of electronic boards (coupled with mechanical systems) and a
real time software part enabling the implementation of the equipment and the
realization of the expected tasks.
Predicting and guaranteeing the reliability of electronic equipment is a huge and

endless task. On the one hand, the number of component types used to build these
boards is very high and on the other hand, the new functionalities in innovative
equipment require numerous reliability and robustness tests.
Thanks to the support of the Astech and MOV’EO competitiveness clusters, the
NAE aeronautics industry, the Normandy and Ile de France regions, and the Rouen
and Versailles Chambers of Commerce, programs have been implemented and have
produced exceptional results (AUDACE, FIRST-MFP, PISTIS, CRIOS,
SYMSTECK, etc.).
Given the richness of the results of these three books that have been published, I
would like to thank very warmly Abdelkhalak El Hami, David Delaux and Henri
xii Applied Reliability for Industry 3
Grzeskowiak for their remarkable work in the implementation of six volumes in

total as well as all the authors who have devoted many hours to shaping all their
results. This has meant that this essential information does not remain the property
of few but shared by the largest number of engineers, technicians, researchers and
students.
The first volume is devoted to Predictive Reliability, for estimating or predicting

the reliability performance of systems.
The second volume is dedicated to Experimental Reliability using tools and test
methods to demonstrate the reliability of systems.
The third volume presents Operational Reliability; it verifies the reliability

performance of systems in real life by way of an analysis of the field data.
I would like to thank all those involved in this program as well as the financiers
(State, Regions) without whom these projects could not have succeeded and I wish
that the members of the hard core set up in these programs (AUDACE, FIRST-MFP,
PISTIS, CRIOS, SYMSTECK) continue their association because their skills will be
essential for guaranteeing the reliability of the very many technologies which are in
the process of emerging and which will equip the mechatronics systems of
tomorrow.
Philippe EUDELINE
President of Normandie AeroEspace (NAE)
October 2022
Preface
March 2020 was a turning point in our lives, for industry worldwide, etc., when
Covid-19 struck the world! This pandemic has highlighted the fragility of our
industries while exacerbating the paradigms of development, innovation,
commercial competition and the acceleration of research phases.
More than ever, we need to be agile in our American, European, Asian and even
global business plans, while at the same time cementing the use of our products in
their operational use. It is clear that the challenges of combining the concepts of
variability and stability, standardization and acceleration of product development are
enormous. The “post-Covid” world is pushing all industries to the very limits of
reliability engineering.
It is true that the word “reliability” is often used as a marketing cosmetic without
any real care for its underlying scientific substantiation! Nonetheless, behind this
word lies an applied science, validated by analysis, research and talented people.
Which type of the reliability approach can a designer, an engineer and a manager
use in their professional environment? Whatever the industrial field (aeronautics,
space, defense, automotive, home automation, etc.), how is it possible to design a
reliable product more quickly, all the while upholding safety measures?
This book illustrates, in great detail, the state of the art for the multidisciplinary
science of operational reliability. The authors wish to clearly depict this discipline
through concrete examples from the industry, based on user experience feedback.
xiv Applied Reliability for Industry 3
Many experts are convinced that reliability is not limited to the statistical
sciences. Various disciplines interact to bring the product to its highest level of
reliability, achievable through current development and production technologies and
methods.
Chapters 1 and 2 elucidate the perspectives of two major automotive

manufacturers, Renault and Stellantis, to design reliable products through two
similar yet distinct approaches.
A top-tier automotive supplier adds an important concept to Chapter 3 by

answering this question: how do you implement an efficient method for designing
innovative products?
We cannot talk about operational reliability in the manufacturing industry

without talking about the “warranty” risk. To this end, the reader will find Chapter 4
very interesting as it provides a methodology for a warranty cost analysis to estimate
reliability performances between 1 and 8 years, and beyond.
However, even if the automotive world occupies a privileged place on the global
stage, the Covid-19 health crisis has brought the medical world to the forefront.
Chapters 5 and 6 address a study for the reliability optimization of high electron
mobility transistors (HEMTs).
Likewise, the world of luxury watchmaking is very rarely cited in reliability

studies, and how wrongly so!
The reader will have the opportunity to discover a hidden facet of reliability
research through the case study of the Richemont watchmaking system in Chapter 7.
It may be surprising to learn that methodologies from the automotive, aeronautical
and medical world are also applicable to the luxury consumer market.
To conclude, there is a before and an after Covid. The reliability challenges for
our industries to address during this new period are significant. The global context,
but also the green, sustainable and carbon-neutral goals are the next steps in a fierce
economic competition that will span the next decade. The authors hope that this
book on operational reliability as applied to industry will impart meaningful
reflections and help to guide readers toward the best practices to encourage efficient
design and effective decision-making.
Preface xv
This book is dedicated to all those who have lost a loved one during the
Covid-19 pandemic worldwide.
October 2022
Abdelkhalak EL HAMI
David DELAUX
Henri GRZESKOWIAK
1
Durability Approach: Applied to a

Vehicle Lighting Control System
Ensuring the durability of system components is a strategic activity for

companies in key industrial sectors, such as transportation. The stakes are twofold:
economic impact and brand image. Indeed, a lack of control over the reliability of a
product can lead to significant costs. This is becoming all the more of a reality given
the recent trend to increase the length of the warranty period [DEL 16]. Although it
is not the only factor, the reliable or durable design of components contributes to
these costs, and thus their improvement leads to a reduction. However, beyond the
direct economic impact, the brand image of the product is also affected. This is
difficult to quantify objectively. However, a customer, whether a private individual
or a professional, who encounters problems with their vehicle before its resale
period (which is on average 5 years of use worldwide [CCF 19]) will naturally be
disappointed and may not choose to make their next purchase from the same brand.
However, in a sector as competitive as the automotive industry of today, a bad or
good image can change our place in the market in a significant way.
1.1. Introduction
Designing a reliable car requires knowledge of information, such as the

customer’s expected usage requirements (region, country, type of use and changes in
mobility solutions), as well as the strength of the parts that make up the car. It is also
necessary to consider all the constraints and evolutions of the component, such as
For a color version of all the figures in this chapter, see http://www.iste.co.uk/elhami/applied3.zip.
Chapter written by Medoune NDIAYE and Caroline RAMUS-SERMENT.
2 Applied Reliability for Industry 3
the mass (which has an important impact on consumption) or the choice of materials
(associated with the recyclability of products).
In this chapter, a global approach to sustainability will be illustrated through the

example of a vehicle lighting control system. This approach is similar to the one
presented in the guide “Aide à la conception et validation de la fiabilité
automobile”1 [SIA 19]. It is based on the concepts of reliability and probabilities:
– quality requirements at mid-life and/or over lifetime;
– technical standards and methodological tools (stress–strength method);
– reliability management plan between manufacturers and suppliers.
Figure 1.1. Instantaneous rate of failure over

time and study of the durability perimeter
As shown in Figure 1.1, a component is subject to different causes of failure

which occur in the early-stage of lifetime (less than one year of usage). These
failures represent pathologies to which the component does not resist for a long time
before showing symptoms. By definition, they are caused by insufficient burn-in or
a faulty process.
– Random failures: Under these failures, several of the actual causes of failure are
hidden. These are generally due to a lack of robustness in terms of the design or the
process.
– Wear-and-tear failures: These failures are characteristic for the end of a
component’s life. They are the result of reaching the physical and chemical limits of
the materials used.
1.2. Example of a vehicle lighting control system
The lighting control of a vehicle allows for the activation of the lights: light
position, high beam or low beam. When a driver selects the desired position, a
1 Design assistance and validation of automotive reliability.

Durability Approach: Applied to a Vehicle Lighting Control System 3
mechanical contact is established between a moving part and a printed circuit track
(see Figure 1.2). This mechanical contact allows for an electric current to flow to the
lamps (headlights).
1.2.1. Risks and reliability requirements
When contact is lost between the moving part and the printed circuit board
(PCB) track, power is interrupted and the lights go out. This can lead to a significant
safety issue.
Indeed, the untimely shutdown of the lights is a safety event classified as

ASIL B, according to several Original Equipment Manufacturers (OEMs) following
the rating rules of the ISO 26262 Standard [ISO 11].
For any event classified as ASIL B, the ISO 26262 Standard recommends a
maximum probability such that this event is equivalent to 1E-7 per hour [ISO 11].
Assuming a constant in-service failure rate, this is equivalent to having a

maximum probability of 7.5E-4 of the lights going out unexpectedly (regardless of
the failure mechanism) over a vehicle service life of approximately 7,500 h.
In addition, a clause in ISO 26262 requires that the estimated failure for
hardware (raw materials, hardware, as opposed to software), using statistical
calculations based on user experience or testing, must have a confidence level of at
least 70% [ISO 11]. For this example, we set a confidence level of 75%.
Figure 1.2. Representation of the electrical contact made via a gold plate
1.2.2. From failure modes to failure mechanisms
We will now focus on the technical points. If we carry out a risk analysis such as
a Failure Mode Effects and Analysis (FMEA) or a Fault Tree Analysis (FTA) for the
event “unexpected loss of lights”, we can see, according to the technical definition
of the light control, that several failure mechanisms [SIA 17] can lead to the event
occurring. In the example being studied, we will focus mainly on the increase in
electrical resistance of the contact between the moving track and the PCB.
Physical–
Failure Causes of
chemical failure Comments
mode failure
mechanisms
Oxidation
Gold does not oxidize, but copper alloy
does. As long as the gold layer is
present, it protects against oxidation.
Fretting – The presence of the gold layer prevents

corrosion the phenomenon of frictional corrosion.
If this gold layer tends to disappear, the
vibrations transmitted to the component
by the chassis/road will lead to corrosion
of the part.
Mechanical Inadequate component design can
breakage compound problems of aging.
Mechanical Track: due to bending movement of the
fatigue PCB.
Moving contact: due to repetitive
Headlight loss: bending stresses.
Headlight Increased Differential Track circuit: due to temperature cycling
malfunction contact thermal during operation and/or ambient
resistance expansion temperature variation (negligible).
Track – moving contact: the contact time
is not long enough to be important.
Thermal ageing Copper diffusion in gold tinning
(diffusion) [PIN 79]. Once the entire thickness
of the gold film is penetrated, copper
accumulates on the surface and a layer
of high copper concentration exists
immediately below the gold–air
interface.
Pollution Not relevant because IP5x component is
protected against dust.
Wear-and-tear The surfaces of the moving contact
and the track are rough.
The thin gold layer is subject to
frictional wear-and-tear.
Table 1.1. Potential versus probable failure mechanisms

First, we can identify all the potential failure mechanisms that can occur, next we
will focus on the most likely ones.
The most likely major failure mechanisms are listed in bold in Table 1.1.
On our PCB, we have a copper alloy track covered by a gold layer (see
Figure 1.2). The gold layer prevents two main types of mechanism failures:
– oxidation of the copper alloy layer;
– corrosion of the copper alloy through friction.
However, it introduces two other failure mechanisms:

– the diffusion of copper through gold;
– the wear-and-tear of the gold layer.
In Table 1.1, the following is presented:

– in italics: the potential failure mechanisms (by no means an exhaustive list).
Here, we have two main competitive failure mechanisms:

– competitor A: diffusion of copper through the gold layer, due to thermal aging;
– competitor B: disappearance of the gold layer, due to contact wear-and-tear.
The first one that occurs defines the life of the component.
1.2.3. From failure mechanisms to physical damage factors
Now we need to look for the physical factors that will activate these failure
mechanisms.
Failure mechanisms Physical damage factors

Thermal aging (diffusion,
Track temperature
Kirkendall effect, etc.)
Contact force (N) and relative displacement (mm)
Wear-and-tear
between the PCB track and the moving contact
Table 1.2. Failure mechanisms and physical damage factors

1.2.4. From physical damage factors to mission profiles or customer

usage
Once we have identified the physical factors that activate the failure
mechanisms, we investigate:
1) their links with the driver’s actions, the vehicle environment or the system
design;
2) sources of amplitude or level deviation.
Physical damage factors Linked to

Vehicle environment: vehicle storage temperature
Track temperature System design: current through the track
System design: location of components in the car
Contact force (N) System design: contact design
Relative displacement (mm) Driver: number of light activations by the driver
Table 1.3. Relationship of damage factors between design and use
The system design is the same for all vehicles. Therefore, any variance of the
physical factors will come from the driver’s behavioral constraints or the vehicle
environment.
Table 1.4 summarizes the variance in the magnitude of the physical factors.
Physical damage factors Source of amplitude or level deviation

Track temperature Vehicle environment: different climate
Contact force (N) System design: negligible
Relative displacement (mm) Driver: different usage patterns of customer cars
Table 1.4. Sources of variation among damage factors
Now, for simplicity of illustration, let us consider that the thermal environment is
not severe enough to activate the copper diffusion failure mode. We could also
assume that even if we have enough copper diffusion, we can be quite sure that the
gold layer will disappear before the first phenomenon becomes significant.
In any case, this assumption does not fundamentally change the approach.
Otherwise, we would have to treat the two ruin mechanisms one by one using the
same methodology.
For the wear-and-tear failure mechanism, we need to define the mission profile
of the associated stress. This is a statistical distribution since the number of control
activations is a random variable.
Below, we present the characterization of the stress mission profile:

– lognormal distribution with parameters µ = 8.5 and σ = 0.8;
– environmental conditions:
- dust (negligible since introduction of IP5x),
- thermal cycling (negligible for gold layer wear-and-tear).
Figure 1.3. Probability density of the constraint
1.2.5. From failure mechanisms to component part strength distribution
Obviously, the thickness of the gold layer will not be the same for all tracks.
Therefore, the number of activations before the copper alloy is exposed and will be
different for each lighting control.
The ability to resist activations is defined as the “strength” of the contact, if

we focus on the wear-and-tear of the gold layer. The resistance distribution for
wear-and-tear is often represented by a Weibull distribution. With failure test

procedures, it is possible to identify the appropriate statistical distribution in order to
have the values of the shape (β) and scale parameters (η).
In practice, failure tests often involve a limited number of components (<15).

The fitting of the strength distribution must therefore be done with robust methods
for small samples such as the median rank method [ABE 06].
The fitting procedure that can be easily performed with a spreadsheet is as

follows:
– Step 1: Order the number of activations to failure found in the test in ascending
order.
– Step 2: Estimate the values of the distribution function of the Weibull
distribution (see illustration in Table 1.5) by:
.
𝐹 = with an accuracy to the order of 5E-2 [1.1]
.
Equation [1.1] is the median rank by the Benard’s approximation or by:
𝐹 = 𝐵𝑒𝑡𝑎 (0.5, 𝑖, 𝑁 − 𝑖 + 1) for an exact value. [1.2]
Equation [1.1] is the median rank by the inverse beta law.
Here:
– i: the order of the failure;
– N: the size of the tested sample;
– 𝐵𝑒𝑡𝑎 : the inverse of the incomplete beta distribution.
In Table 1.5:
– the i values are in cells F17 to F22;
– the sample size N is in cell F15;
– the values of the number of activations to fail here come from a
Weibull distribution with a shape parameter beta of 5.7 and a scale parameter eta
of 68,000.
Number of Fi (Benard
i Fi Exact Beta Law
activations Approximation)
1 51082 =(F17-0.3)/($F$15+0.4) =BETA.INVERSE(0.5;F17;$F$15-F17+1)
Table 1.5. Formulas for calculating

median ranks in a spreadsheet
This gives the results presented in Table 1.6.
N
6
Number of
i Fi (Benard Approximation) Fi Exact Beta Law
Activations
1 51082 10.94% 10.91%
2 57828 26.56% 26.44%
3 61777 42.19% 42.14%
4 67850 57.81% 57.86%
5 75502 73.44% 73.56%
6 83229 89.06% 89.09%
Table 1.6. Results of the calculation of median ranks

with Benard’s formula and the incomplete Beta distribution
– Step 3: Identify the parameters of the linear regression 𝑦 = ln ln and

𝑥 = ln(𝑛𝑖) with 𝑛𝑖 as the numbers of activations at failure.
In Table 1.7, the Fi values calculated previously go from cell I17 to cell I22 and
the numbers of activations to failures are filled in from cell G17 to cell G22.
Least squares regression on rows
yi = ln [ln (1/(1 – Fi))] xi = ln (ni)
=LN(LN(1/(1 – I17))) =LN(G17)
=LN(LN(1/(1 – I18))) =LN(G18)
=LN(LN(1/(1 – I19))) =LN(G19)
=LN(LN(1/(1 – I20))) =LN(G20)
=LN(LN(1/(1 – I21))) =LN(G21)
=LN(LN(1/(1 – I22))) =LN(G22)
Table 1.7. Formulas for computing yi and xi

for ordinary least squares linear regression
This gives the following results in Table 1.8.
Least squares regression on rows
yi = ln [ln (1/(1 – Fi))] xi = ln (ni)
–2.158 10.841
–1.180 10.965
–0.603 11.031
–0.146 11.125
0.285 11.232
0.795 11.329
Table 1.8. Results of yi and xi computation

for ordinary least squares linear regression
In accordance with the linear form of the Weibull distribution by double

logarithmic transformation [ABE 06] and by identification:
– the shape parameter of the Weibull distribution corresponds to the slope of the
regression line;
– the scale parameter of the Weibull distribution corresponds to
𝜂=𝑒 .
Table 1.9 shows the application in a spreadsheet.
Beta =SLOPE(J17:J22;K17:K22)
Eta =EXP(-ORIGINAL.ORDER(J17:J22;K17:K22)/K24)
Table 1.9. Formulas for calculating Beta and

Eta by the median rank method in a spreadsheet
The range J17:J22 contains the values of yi, and the range K17:K22 contains the
values of xi.
The results obtained are presented in Table 1.10.
Parameter Estimated value True value

Beta 5.8 5.8
Eta 71,156 68,000
Table 1.10. Results estimation of beta and eta

by the median ranks method in a spreadsheet
Remember that it is extremely important to design the failure test correctly in

order to avoid the activation of failure mechanisms other than the one we are trying
to characterize. For example, during the wear-and-tear test, high-frequency
activations should be avoided as this would lead to abnormal heating of the contact.
1.2.6. Resistance distribution chart
In the absence of failure tests, which can be costly in the design phase, it is
however possible to define the properties that the strength distribution must respect
in order for the component to be compatible with the reliability requirements and the
mission profile.
This is done by constructing a chart (see Figure 1.4) that defines a level line, that
is, an acceptance boundary corresponding to the maximum allowable failure
probability and defined by unique pairs of the shape parameter β and the scale
parameter η.
Therefore, for each value of β (image of the strength dispersion), the minimum
value of the scale parameter η (image of the average strength) ensuring a failure
probability equal to the target is calculated.
This is done by solving the following equation for fixed 𝑃 , 𝜇 and 𝜎 fixed:
𝑃 =∫ 𝑓 (𝑥; 𝜇; 𝜎). 𝐹 (𝑥; 𝛽; 𝜂) ∙ 𝑑𝑥 [1.3]
Equation [1.3] is the probability of failure calculation by strength–stress method.
Figure 1.4. Resistance distribution chart
Here,
( )
𝑓 (𝑥; 𝜇; 𝜎) = 𝑒
√
𝐹 (𝑥; 𝛽; 𝜂) = 1 − 𝑒
This calculation of the integral of equation [1.3] can be easily implemented in a

spreadsheet using a quadrature method of the Simpson, trapezoid, rectangle or
midpoint type, as well as the statistical functions LAW.LOGNORMAL and
LAW.WEIBULL.
We then simply use the spreadsheet solver to find the value of η which, for a
given value of β, satisfies the equality of equation [1.3].
This results in the construction of an acceptance/rejection boundary that allows

both the customer and the supplier to estimate the impact of the design (average gold
layer thickness) and the process (gold layer thickness dispersion) for the
achievement of reliability objectives.
Indeed, the relative strength dispersion, which is the ratio between the strength
dispersion and the average strength, is expressed as:
𝑞= [1.4]
For a Weibull law, we get:
∙ ∙
𝑞= = [1.5]
∙
Equation [1.5] is the relative dispersion as a function of beta, with the Gamma
function Γ.
The application of the formula from equation [1.5] in a spreadsheet (with the
value of 𝛽 in cell A1) is given as:
Q = ROOT((GAMMA(1 + (2/A1)) – GAMMA(1 + 1/A1)2))/GAMMA(1 + 1/A1)
The relationship between 𝑞 and 𝛽 can be correctly approximated by:

.
𝑓𝑜𝑟 𝛽 ∈ [1 − 10[ ∶ 𝑞 = 0.989375 ∙ 𝛽
.
𝑓𝑜𝑟 𝛽 ∈ [10 − 100[ ∶ 𝑞 = 1.167036 ∙ 𝛽
Table 1.11 shows the typical coefficients of variation for different levels of
process control.
Level of process control Q β

Low mastery 19.99% 5.8
Good control 10.04% 12.1
High proficiency 5.01% 24.9
Excellent command 2.50% 50.6
Table 1.11. Correspondence between beta and coefficient

of variation according to the level of process control
Therefore, the a priori knowledge of the process control level (thus of 𝛽) allows
from the design phase to identify the value of 𝜂 that the component must have a
minimum and consequently the average strength level because:
𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝑠𝑡𝑟𝑒𝑛𝑔𝑡ℎ = 𝜂 ∙ Γ 1 +
where:
– β is estimated from the process variation coefficient;

– η is determined from the chart.
1.2.7. Proposal and study of a validation plan using the stress–strength

method: various real-world examples
Two main tests are used to validate reliability.
The zero-failure test (Figure 1.5): Practice sheet 7 is applicable only if feedback
for the strength distribution is available (see [SIA 19]). Confidence in the test result
is related to the confidence in the feedback. These tests usually have a shorter
duration.
The failure test (Figure 1.6): Practice sheet 7 usually requires at least seven
components, generates knowledge (REX: Return on Experience) but it is a more
expensive and longer duration test (see [SIA 19]).
Figure 1.5. Zero-failure test
Figure 1.6. Failure test
A) Illustration of the zero-failure test

Question 1: What is the number of contacts to test?
The resistance shape parameter 𝛽 = 3.3 obtained from the supplier. A

resistance distribution chart (as shown in Table 1.12) constructed by the reliability
referent is as follows:
– Test acceptance criteria: zero failures observed during the test.
– Damage z: number of activations during the test before complete wear-and-tear
of the gold layer = 115,550.
– c = confidence level for the test = 75%.
1 2
1 2 9.318·105
2 2.1 7.726·105
3 2.2 6.534·105
4 2.3 5.623·105
5 2.4 4.913·105
6 2.5 4.35·105
7 2.6 3.897·105
8 2.7 3.527·105
9 2.8 3.223·105
10 2.9 2.969·105
11 3 2.755·105
12 3.1 2.574·105
13 3.2 2.419·105
14 3.3 2.285·105
15 3.4 2.169·105
16 3.5 ...
Table 1.12. Column 1 = beta value; Column 2 = eta value
Weibull’s law: cases where the failure mechanism was the wear-and-tear of the
gold layer.
𝐷𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑊𝑒𝑖𝑏𝑢𝑙𝑙 𝑙𝑎𝑤, 𝐹(𝑧) = 1 − 𝑒

( )
𝑁= ( )
Excel ®: Nb_test = LN (1 – 0.75)/LN (1 – WEIBULL.LAW.N (115,550; 3.3;

228,500; TRUE)) – 1 = 12.2
Therefore, the number of parts to be tested is 13.
Questions 2 and 3: The test has been performed and there is one failure among
13 parts. Calculate the confidence level [SIA 17] obtained to prove the reliability
target and the number of new parts to be tested, without failure, to prove the
reliability target?
𝑂𝑏𝑗𝑒𝑐𝑡𝑖𝑣𝑒 = 𝐵𝑒𝑡𝑎(𝐹(𝑥), 𝑥 + 1, 𝑁 − 𝑘 + 1)
Excel ®: Confidence Level = BETA.LAW.N (WEIBULL.LAW.N (115,550; 3.3;

228,500; TRUE); 115,550 + 1; 13 – 1 + 1; TRUE) = 41.5%
𝛿 = 𝐹(𝑥) ≥ 0.1 for a well-defined test
Therefore, the confidence is 41.5% instead of 75%; if there is one failing part out
of the 13 tested. The test is therefore inconclusive.
What is the number of new parts to be tested without failure in order to prove the
reliability target?
Thirteen new parts are needed to meet the 75% confidence level. There is no
other choice than to redo the complete test.
B) Illustration of zero-failure test with different results

Let us assume that you have received three different proposals:
– Supplier A: 150,000 activations, 5 pieces
– Supplier B: 120,000 activations, 10 pieces
– Supplier C: 100,000 activations, 22 pieces
Questions 4 and 5: What is the confidence level of the test to prove the reliability
objective: for each of the suppliers? Which supplier provides the most appropriate
reliability test?
𝑐 = 𝐵𝑒𝑡𝑎(𝐹(𝑥); 𝑥 + 1; 𝑁 − 𝑘 + 1)
Excel ®: Confidence = BETA.LAW.N (WEIBULL.LAW.N (150,000; 3.3;

228,500; TRUE); 150,000 + 1; 5 + 1; TRUE)
Response to Q4: The confidence is:

– Supplier A = 77.60;
– Supplier B = 73.11%;
– Supplier C = 77.79%.
Which supplier provides the most suitable durability test?
Response Q5: Supplier A provides the best reliability test. The confidence in
supplier B’s test is too low. The confidence in supplier C is sufficient, but the test is
too short: delta = 6.3% (delta = 10% is the minimum value).
Recall, 𝛿 = 𝐹(𝑥) must be ≥ 0.
Recall, δ = F(x) must be ≥ 0.1 for a valid test, 1 for a valid test.
Illustration of a failed test

The test was done with seven parts. We have received the results (see
Table 1.13): each piece failed after N activations.
N activations to failure
187000
221000
235000
260000
320000
357000
437000
Table 1.13. Failure test results
Question 6: What are the values for beta and eta – Weibull distribution
parameters – for a 75% level of confidence?
We analyze these failure test results using Johnson’s median rank method.
First, arrange the results of the N activations in ascending order.
Calculate the table below:

Excel ®:
Fi = SI ($ ri <> ""; INVERSE.BETA (confidence; ri; nb_parts-ri + 1); "")
Ln (x-d) = SI ($ ri <> ""; LN (zi); "")
Ln Ln (1/1 (1 – F)) = = SI (Fi <> ""; LN (LN (1/(1 – Fi))); "")
Slope = SLOPE (Ln (x – d); lnln (1/(1 – F)) = 3.13 = beta
Y-intercept = ORIGINAL.ORDER (Ln (x – d); lnln (1/(1 – F)) = –39.28
Quality of Model R² = (CORRELATION.COEFFICIENT (Ln (x – d); Lnln (1/
(1 – F))2
Eta = EXP(–Slope/order) = 287,140

See the good quality of the regression with the R² value.
Question 7: What is the reliability obtained? Are we within the target?
P = ∫ F (z, θ , θ ) × f z, x 𝑘𝑚 or 𝑦 years, θ , θ × dz ≤ P ∗
Pf = 6.7 × 10–5 < Pf* = 10–4
The reliability target is met with a minimum confidence level of 75%.
1.3. Conclusion
This chapter describes a common method and illustrates this through a simple
example. Reliability engineers size the parts that make up a vehicle with the view
that keeps in mind product quality and the cost impact for the customer and the
brand image. After the launch, they usually carry out a follow-up during production
life through the analysis of the quality department’s warranty returns. This
verification ensures that the customer is genuinely satisfied.
1.4. References
[ABE 06] ABERNETHY R., The New Weibull Handbook: Reliability and Statistical Analysis for
Predicting Life, Safety, Supportability, Risk, Cost and Warranty Claims, 2006.
[CCF 19] CCFA, “Durée de possession moyenne d’une voiture”, L’industrie automobile
française, available at: https://ccfa.fr/wp-content/uploads/2018/01/Analyse_Statistiques_
2017_FR.pdf, 2019.
[DEL 16] DELAUX D., Warranty and reliability: new methodology for a good design, ESA
AWARE, 2016.
[ISO 11a] ISO 26262 Véhicules routiers – Sécurité fonctionnelle Part 3, ISO26262-3: 2011,
2011.
[ISO 11b] ISO 26262 Véhicules routiers – Sécurité fonctionnelle Part 5, ISO 26262-5: 2011,
2011.
[PIN 79] PINNEL R., “Diffusion-related behaviour of gold in thin film systems”, Gold
Bulletin, vol. 12, pp. 62–71, 1979.
[SIA 17] SIA, Glossaire de langage commun fiabilité, available at: https://www.sia.fr/
publications/432-glossaire-langage-commun-fiabilite, 2017.
[SIA 19] SIA, Guide d’aide à l’estimation et à la validation de la fiabilité automobile,
Référence: DC-01-03, available at: Société des ingénieurs de l’automobile: https://
www.sia.fr/publications/431-guide-aide-estimation-validation-fiabilite-automobile, 2019.
2
Structural Diagrams to Validate the

Reliability of Mechanical Components
2.1. Introduction
When an accelerated test breaks a prototype, the design office sometimes

challenges the result by saying that, it is normal to break our prototype, given the
excessive severity of the test. This does not make the component any less reliable for
in service conditions. The same conclusion would have been reached if the
prototype had not broken during testing. Taking this reasoning a step further,
accelerated testing could simply be eliminated, given that whatever the results, the
conclusions are the same! This anecdote, which many reliability test managers could
tell, highlights the major challenge in developing accelerated testing: knowing how
to credibly link its results to in-service reliability. If we can demonstrate that a given
result leads to a given level of in-service reliability for the component that will be
manufactured at a large scale, we will have succeeded in developing a validation
system that fully plays its role: filtering out unreliability and giving a timely warning
when future reliability in series production is insufficient.
The credibility of an accelerated test can be quantified by its confidence level γ,

defined by its probability of detecting a level of reliability deemed unacceptable in
service.
Let us assume that the unacceptable level is set to F < 1%. F is the probability
that a failure will occur before a given service life t. Not generating more than 1%
failure is a very unambitious goal for many components. Under normal use, a
For a color version of all the figures in this chapter, see http://www.iste.co.uk/elhami/applied3.zip.
Chapter written by Paul SCHIMMERLING.
probability calculation shows that 230 tests and no failures must be performed in
order to demonstrate that this objective has been reached with a confidence level of
γ = 90%.
In fact, γ is the probability that the test will trigger a warning, that is the
probability of occurrence of at least 1 failure among n tests. When the probability of
failure is F, the confidence level can be evaluated with the binomial distribution:
γ = 1 – (1 – F)n [2.1]
By performing n = 230 tests, we have γ = 0.9, when F = 1%. This relationship

shows that accelerated testing is essential to demonstrate high levels of reliability.
By increasing the applied stresses, as compared to their in-service level, we increase
the failure probability F, which allows us to reduce the number of tests n. However,
it is important to know how to relate the results of accelerated testing to in-service
reliability.
Experience in the validation of automotive mechanical parts has allowed us to

identify some methodological diagram tools through which to establish this link.
These diagram tools allow us to specify the hypotheses and the data to be collected
in order to design and size credible validation tests. Designed for the validation of
automotive components, these diagram tools can be applied to the validation of any
mechanical part intended for mass production, subject to a well-known and
identified critical failure mode. The sizing consists of determining the test profile
(number of cycles to be applied, amplitude of the cycles, test duration, etc.), the
number of parts to be tested and the acceptance criterion of the test so as to
demonstrate the achievement of the specified reliability objective. The approach is
also applicable to the definition of loading data for a numerical simulation that
faithfully reproduces the identified critical mode. We will illustrate the approach
through the development of a reliability validation test for disc brake pads.
2.2. Choice of methods
2.2.1. Criteria selection
Two criteria appear to be decisive. Firstly, the existence of a technical reference

in terms of customers presenting the same failure modes as the prototype, and
secondly, the nature of the failure: is it sudden or is the degradation progressive?
When a technical reference has already been put into service, it will be possible
to exploit feedback. Feedback includes all the information reported by after-sales
services. In the automotive industry, this means that the repairs carried out by the
Structural Diagrams to Validate the Reliability of Mechanical Components 23
dealer and garage network. Feedback can also refer to failures and incidents, as well
as examinations and measurements of used, non-faulty parts. The absence of
incidents observed on a volume of parts in service also constitutes statistically
exploitable feedback. The previous formula, equation [2.1], shows that the
probability of failure is less than F, with a probability of γ, if the n parts put into
service do not fail.
In the absence of a technical reference – which is the case for innovations – the
mission profiles will serve as the only information source that can be used to
characterize the operating phase in service.
When the failure manifests itself as a progressive, not brutal, degradation, it will
be very interesting to set up a measurement of it. We thus go from observations of
the “broken, not broken” type, to a quantitative measurement. It will hence be
possible to estimate an average degradation, instead of a proportion of failures,
thereby reducing the number of tests necessary to restore a given level of reliability
[SIA 17b].
2.2.2. Four basic methods
The combination of the above two criteria allowed us to define four basic
methods for establishing the link [SCH 03, SCH 14, SIA 17a, SIA 17b]: the
comparison of Weibull laws in service and under testing, the comparison of
degradations in service and under testing, the stress–strength method and the
stress–strength method with degradation analysis (see Figure 2.1).
Figure 2.1. Four basic methods linking

reliability under testing to reliability in service
2.2.3. An applied example: the validation of disc brake pads
The feasibility study of the four methods will be illustrated through the example
of brake pads, presented in the guide [SIA 17b], which details the reliability
calculations.
A hydraulic mechanism rubs the pads against a disc (see Figure 2.2). The critical
failure mode, which is unavoidable, is abrasive wear-and-tear, which manifests itself
as a loss of pad thickness, can be measured easily.
Figure 2.2. Disc brake diagram
A bench test simulates braking cycles by applying a force onto the piston of the
hydraulic circuit’s master cylinder in order to reproduce the effect of the brake pedal
force.
A new brake will be fitted to a new vehicle. The reliability target set for the pads
is F < 1% at 20,000 km. How can the test be sized to demonstrate that this objective
is met?
The feasibility study will specify the data to be collected and the hypotheses to
be consolidated in order to answer this question.
2.3. Feasibility study on the four methods
2.3.1. Animation principle
The reliability specialist will bring together at least one brake disc designer, one
test manager and also a representative of the after-sales quality department. The
principle is to review each of the four methods in order to specify with the working
group, the data and the hypotheses necessary for their application. Each method will
be explained through the aid of a diagram, so that all participants understand the
principle and can give their opinion on the data and assumptions.
The facilitation principle is to ask questions in order to make the group clarify
the meaning of all the elements of the methodological diagrams represented in
Figures 2.3, 2.4, 2.6 and 2.7, in the context of the problem being studied.
Ideally, it is best to review all four methods, without focusing on one method
a priori. Indeed, it is not always easy to discard a method a priori. The review of the
four methods allows us to identify data that we may not have considered.
Before starting this analysis, we will first make sure that the following three
prerequisites are met:
– B1: the test faithfully reproduces the failure mode observed in service. This
leads to verify that the test activates the same damaging factors. The identification of
the damaging factors is facilitated by the identification of the physical mechanism of
failure at the origin of the material failure;
– B2: failure in service is caused by a physical mechanism, affecting all the parts
in series and not by a poorly controlled manufacturing problem, affecting production
batches (defects, non-conforming parts, etc.). The slope of the Weibulll laws on the
after-sales data can help to identify this case [ABE 83, BLI 11];
– B3: a reliability objective to be reached in service is defined.
2.3.2. Comparison of Weibull laws under testing and in service
2.3.2.1. Principle
The method consists of comparing the Weibull law applied with the failures
observed under testing and in service (see Figure 2.3). The test objective is
readjusted if the serial reliability objective is not reached. The test target life (often
specified by a Bq corresponding to q% failure) is multiplied by a factor K equal to
the ratio of the series target Bq to the series Bq of the technical reference.
The severity of the test is evaluated by locating the test and service lives
associated with the same failure probability F.
Figure 2.3. Method for comparing Weibull

laws under testing and in service
Other probability laws than the Weibull Law can be applied; however, the
Weibull Law is particularly adapted to the modeling of mechanical failures. The
slope of the line, in transformed scale, reflects an aging mechanism if it is greater
than 1. A slope less than 1 indicates an early-stage failure, often related to a
manufacturing defect. Other Weibull models, which are more complex, can reveal
the presence of two failure modes and mixed populations [ABE 83].
2.3.2.2. Application conditions and assumptions

It is necessary to verify two applied conditions and three assumptions:
– C1: time-to-failure data is available in service;
– C2: time-to-failure data is available under testing, for the same component;
– H1: the mission profiles do not change for the new component when compared
to the technical reference;
– H2: the x-axis, under testing and in service, can be defined by a preponderant
damaging factor;
– H3: the improvement of service life through testing induces a proportional
improvement of the service life for customer usage. This assumption is sometimes
too pessimistic. In terms of fatigue, the service lifespan can be increased
considerably, if by strengthening the part, the stresses in service go below the
endurance limit. The lifespan will increase far less for an accelerated test that applies
stresses above the endurance limit.
2.3.2.3. Variant: comparing the reliability testing of 2 components

This comparative approach aims at verifying whether a second component is
more reliable than the first one in series by comparing the Weibull Laws under
testing. The approach is similar to the method of comparing Weibull laws under
testing and in service provided that the component in series is sufficiently reliable
and that the above-mentioned assumptions are verified.
2.3.2.4. An applicability to brake pad validation

Conditions B1 and B2 are met: abrasive wear-and-tear is indeed the aging
mechanism affecting the pads in service, which is reproduced on the bench.
A special case is the absence or rarity of incidents in service. Does this mean that
the method is not applicable? Provided that the slope of the Weibull distribution in
service and the monthly volumes in service are known, it is possible to apply an
“envelope” Weibull distribution with increasing F and a level of confidence γ. The
actual value of F, unknown, is below the threshold defined by the envelope, with
probability γ [REL 18].
Condition C2 is more problematic. It is rare to push a test to the point

of complete pad wear-and-tear. However, it is possible to extrapolate pad
wear-and-tear measurements in order to deduce times to failure, using a statistical
model of degradation (see Figure 2.4).
Hypothesis H1 leads to an analysis of the mission profile: is it comparable to the

profile defined for the technical reference? Is it the same market and the same use?
Does the brake sizing lead to the same pressure on the brake pads, with the same
pedal force?
Hypothesis H2 defines the predominant damaging factor. In service, the number

of kilometers covered by the vehicle will be used, if we do not know the number of
brake strokes suffered by the pads replaced by the network. The number of brake
hits is the main damaging factor of the bench test.
The method of comparing Weibull laws appears to be poorly adapted to the

available data in the absence of customer crisis experience feedback and destructive
bench tests.
Another random document with
no related content on Scribd:
Sir Harold approached the young man, extending his hand.
“Behold ‘the gentleman from India,’” he said, faintly smiling. “My dear
boy, ask me as many questions as you like. Don’t you know me,
Arthur, that you stare at me so? I am no ghost, although our friend
Atkins took me for one.”
Another cry, but this time a cry of rapture, broke from the young
earl’s lips. He bounded forward and clasped Sir Harold’s hands in
his, and both were silent with an emotion too mighty for speech.
Atkins turned aside to add fresh fuel to the blazing fire, his own
features working.
“Sir Harold! O, Sir Harold!” cried Lord Towyn at last, in a very ecstasy
of gladness. “What a joy this will be to my poor little Neva! She has
mourned for you as dead, and I have thought that the shadow of
your supposed fate would darken all her life. How glad she will be,
my poor little girl!”
“Your little girl?” said Sir Harold.
Lord Towyn’s fair face flushed.
“I love Neva, and she loves me,” he said frankly. “She has promised
to marry me, and I hope, Sir Harold, that you retain your former good
opinion of me, and will sanction our union.”
“We will see,” said the baronet, pressing the young earl’s hand
warmly. “It has always been my desire, as it was that of your father,
to unite my family to yours. Your face tells me that you have fulfilled
the glorious promise of your boyhood. If Neva consents to marry
you, my dear Arthur, I shall not refuse my consent.”
Lord Towyn looked his delight, and then cast a quick, inquiring
glance at Atkins.
“Does Sir Harold know?” he asked significantly.
“I have told him,” answered the solicitor, “that Miss Wynde has
disappeared in the most mysterious manner and that she is in the
power of a couple of adventurers—”
Sir Harold interrupted Atkins by a passionate gesture.
“Arthur,” exclaimed the baronet, his proud face drawn with pain.
“Atkins tells me that I have been deceived in—in Lady Wynde, and
that he has discovered her to be an adventuress, unscrupulous and
unprincipled. Is this his prejudice? I cannot give utter credence to it.”
“It is God’s truth, Sir Harold,” said Lord Towyn solemnly, holding the
baronet’s hand in a strong, firm pressure. “It is better that you should
know the truth from us than to hear it from strangers, or be further
deceived by the woman you made your wife. Lady Wynde is an
adventuress, bold and false and wicked.”
“You forget that I knew her history even back to her childhood,” cried
Sir Harold eagerly. “I did not marry her with my eyes blindfolded. She
never attempted to impose herself upon me as other than she was.
She made known her whole life to me. She was the daughter of a
naval officer, and the niece of Mrs. Hyde, a lady of good family and
position, who lives a very retired life in Bloomsbury Square, London.
We ate our wedding breakfast in Mrs. Hyde’s house. Lady Wynde’s
first husband was the Honorable Charles Hathaway, the younger son
of a Viscount. Lady Wynde’s family connections both by birth and
marriage are excellent. I knew all this beyond a peradventure before
I married her. And yet you call her an adventuress!”
“And so she was, Sir Harold,” exclaimed Atkins. “Her past life, her
family and her connections were all you say. Her record was all fair.
Not a word had ever been whispered against her reputation, and she
went in the best society, and had admirers and suitors. All this I
grant. But she was none the less an adventuress at heart. She had
an income of three hundred pounds a year and spent a thousand,
sponged from relatives, or given her by Craven Black, from his
winnings at the gaming table or at the races. She was engaged to
marry Craven Black soon after Mr. Hathaway’s death, and before her
marriage with you. Mrs. Hyde is not overfond of her niece, and told
me this fact herself. This marriage, owing to the meagre fixed
income of the pair, was deferred, and finally they conceived the idea
that Mrs. Hathaway should contract a wealthy marriage, secure a
comfortable jointure, become a widow, and then marry Craven Black.
There can be no doubt that your marriage with Mrs. Hathaway was
the result of a conspiracy against you by these two villains, male and
female—that they had set a trap for you, Sir Harold, and that you fell
into it!”
Sir Harold turned his haggard eyes upon Lord Towyn.
“It is true,” said the young earl, full of the tenderest sympathy. “You
were imposed upon, Sir Harold. The woman you married, so fair and
spotless in seeming, was like some fair fruit with a worm at its core.
There are adventuresses in good society, of good birth and spotless
reputations, as there are well-born adventurers. Mr. Atkins is right.
Craven Black and Mrs. Hathaway have played a daring game, but
they have not yet won. This is a terrible stroke to you, dear Sir
Harold; but bear it bravely. You are not desolate because Lady
Wynde feigned a love for you, and has proved false and wicked. You
have the holy memories of your first wife to keep pure and steadfast
your faith in woman. You have Neva to love you. You have your
friends.”
But Sir Harold threw up his arms with a gesture of despair.
“I loved her!” he said brokenly. “I have thought of her in my Indian
dungeon, and on the lonely sea, and have planned how to break to
her the news of my return tenderly and gently, that her reason might
be spared a shock which I feared might destroy it. And, O God! all
the while she never loved me! While I thought of her upon the deck,
with longings for wings, that I might sooner reach her, she was the
wife of another, and exulted in the thought that she was rid of me
forever! Ah, this is a dreary coming home!”
“It is, Sir Harold,” said Lord Towyn sorrowfully; “but the wickedness
of one person whom you have loved need not darken your life, nor
paralyze your energies. Neva is in peril. Rouse yourself from this
great grief for her sake. Think what joy your return will be to her. We
must find her, and save her.”
The young earl had touched the right chord. Sir Harold aroused
himself from his despair, and said:
“Yes; we must find her, and save her. But where are we to look for
her? If the detectives have failed to find a clue to her whereabouts,
how are we to succeed?”
“I have been upon the Continent,” said Lord Towyn, “and have
traveled from one end of England to the other. I have been upon a
score of false trails, and failed to find a trace of those I sought. I have
now been three or four days in this town, consulting every day with
Atkins or Sir John Freise, while the detectives continued the search.
And to-night I have received news which for the first time gives me
hope that we are nearing the end. A messenger, sent by one of my
detectives, came to me by the last down train from London, with a
report of discoveries.”
“They have been found?” cried Sir Harold eagerly.
“Not yet. The object of Craven Black and his wife—I hardly know
how to call her, Sir Harold—was to marry Neva to Black’s son, and
so obtain control over the Hawkhurst property,” said Lord Towyn. “It
is to effect this marriage that Craven Black and his wife are engaged
in persecuting Neva. When they left Hawkhurst, they left Rufus Black
behind them. It occurred to me that when they should deem matters
in a fair state of progress, or when Neva showed signs of relenting,
they would send for Rufus to come and plead his cause, or to marry
her, wherever they might be. I therefore hired a detective to watch
Rufus, and it is from this detective, and not from those in search of
Neva, that I have to-night heard.”
“And what does he say?” demanded Atkins breathlessly.
“Young Black has remained at Hawkhurst ever since the marriage—
some five weeks. Two or three days ago he went up to London. The
detective, who had been stopping at Wyndham as a commercial
traveler in broken health, went up on the same train. It seemed at
first, my messenger says, as if young Black had had no object
beyond a day’s saunter in town. He visited picture shops and so on,
but that night he went to the Great Northern railway station, and
found the train gone. That movement of his, as the detective said,
began to look like business. Black went to his hotel, the detective still
on his track. The next morning young Black sold his watch and
chain, and the next evening he was off again to the Great Northern
railway station. He caught the night express, and went on it, the
detective on the same train. The detective sent a note from
Edinburgh to a fellow-officer, who brought it to me to-night. I am
convinced that Rufus Black has gone to rejoin his father, and that if
we follow him we shall find Neva.”
“To what place did he book himself?” asked Atkins.
“To Inverness. It is plain that while the Blacks tried to persuade us
that they were upon the Continent, they were safely hidden with
Neva in the Scottish Highlands. They may have gone there from
some idea of bringing about an informal Scottish marriage between
Neva and young Black. Neva can know nothing of the marriage laws
of Scotland, where a declaration from a woman that a certain man is
her husband, when he hears and does not contradict the assertion,
and vice versa, constitutes a legal and binding marriage. The Blacks
may calculate upon Neva’s ignorance, and hope to avail themselves
of the facilities of Scottish law in marrying her to Rufus.”
“It is very probable,” said Atkins, knitting his brows.
“Young Black has the start of us. He must have arrived at Inverness
to-day. I came here to propose, Atkins, that we start for the north by
the earliest morning train. We are on the right track now,” said Lord
Towyn. “Let us follow it up promptly.”
“We will go in the morning,” declared Atkins.
“I shall go also,” said Sir Harold. “Let the secret of my return be kept
a secret still. I do not wish to warn this Craven Black, or put him on
his guard. Call me Mr. Hunlow. It is the name I traveled home under.
And be careful not to betray my secret until I myself declare it.”
The three sat together by the office fire all the remainder of the night
and talked. In the morning Atkins wrote a note to his wife, and
another to his clerks, and leaving the notes upon his desk, went out
with his two guests before the family were astir. Sir Harold muffled
his face beyond recognition, and conducted Lord Towyn and Atkins
to his hotel. Here they were served with breakfast, and soon after
they proceeded to the station, and took the train for London.
Sir Harold breathed more freely when they had left the Cathedral
town behind them. He was well known in Canterbury, and with a
strange nervous shrinking, he dreaded recognition before he should
choose to make his return known.
On arriving in London, the three pursuers hastened to the Great
Northern Railway station, and an hour later they were on their way to
Scotland, upon the trail of Rufus Black.
CHAPTER XIX.
REUNITED.
Upon the day after the storm, a high wind still prevailed. No sailing
vessel dared put out to sea from Inverness. The sky was dun and
gray, with now and then a fitful gleam of sickly yellowish sunlight.
The black waters were all alive with “white caps,” and the sullen roar
of the waves, as they hurled themselves against the cliffs upon
whose summit stood the house of Heather Hills, filled all the house
with its monotonous tumult.
Lally Bird spent the morning in her own room, upon a sofa in a
recessed window. Mrs. Peters came and went softly, bestowing
pitying glances upon the round gipsy face lying so white and
sorrowful against the cushions, but the dusky eyes were looking
seaward with a strange, far off, steadfast gaze, and it was evident
that the young girl was not even conscious of the presence of her
attendant.
At noon Mrs. Peters brought up a tray on which was spread a
tempting luncheon of chocolate, hot rolls, delicate game birds, and
jellies. She placed the tray upon a low table, and wheeled it beside
the sofa. Still Lally did not stir.
“Miss Lally,” cried the good woman, her lips quivering. “Are you not
going to eat to-day? You had no breakfast. You will be ill. I know that
I have offended you beyond all forgiveness, and that my face must
be unpleasant in your sight, but I would undo what I have done if I
could. Better almost any kind of a marriage than to see you lying
here looking so wan and hopeless. Oh, Miss Lally, if you would only
speak to me!”
Lally turned her face slowly, with a look of surprise mingling with her
expression of pain.
“Why, Peters,” she said kindly, “I did not know you were so troubled
about me. I am not angry at you. You meant what you did for the
best. There, don’t cry, Peters. I am not angry; indeed, I am not. You
are as much my friend as ever. Sit down by me, and we will eat our
luncheon together.”
Peters complied as soon as she could command her emotion, and
Lally aroused herself to speak cheerfully, and to inquire concerning
the results of the storm.
After the luncheon the young mistress of Heather Hills announced
her intention of going out for a solitary walk. The wind was not so
high as it had been in the morning, and Mrs. Peters did not venture
any objections. Lally attired herself in a bombazine walking dress
and astrachan jacket, hat and muff, and about two o’clock she went
out alone for a walk along the cliffs.
For an hour or more she rambled on, stopping now and then to rest,
and keeping near the sea, over whose wide wild waters her gaze
strayed and fixed itself with singular steadfastness. At last she sat
down upon a great bowlder, and the slender black figure was
outlined against the gray sky with startling distinctness.
Before her lay the wild and restless sea, behind her the undulating
fields of her new domain. At one side of her, in the gray distance,
was the house of Heather Hills, and on the other hand, and nearer,
was the low range of heath-clad hills which gave the estate its name.
It was a lonely spot, that upon which she had paused to rest, with a
bold bluff surmounted by a very chaos of rocks, upon whose summit
she had perched herself. A few sea-gulls were screaming in the air,
but besides them and the wild birds on the heath there was no sign
of life far or near.
An hour passed. The wind still blew strong and fierce, tugging at her
hat and garments with strong, despoiling hands. Her vail was swept
over the bluff into the abyss of waters, and her hair was torn from its
confining braids, and tumbled over her shoulders in a dusky
cascade. But still Lally sat high up upon the rocky mass, paying no
heed to wind or murmur of wave, her soul being busy with the great
problem of her destiny.
And so, looking seaward with great longing eyes, she did not see a
human figure coming toward her over the fields. It came nearer and
nearer—the figure of Rufus Black!
The young man had gone back to Inverness upon the previous night,
but he had not been content to accept his dismissal at the hands of
Mrs. Peters. His old love for Lally was strong and fierce, and he was
determined to win back his lost young wife, if energy and patience
and love and sincere repentance could win her back. So, after a
sleepless night, and a morning spent in indecision and irresolution,
he had come out again to Heather Hills. Mrs. Peters was in her own
room, and the housemaid had answered his knock. Rufus had
inquired for Miss Bird, but the housemaid had never heard the name.
He then asked for Mrs. Black. That name was also unknown at
Heather Hills. In this dilemma, believing Lally to be at the Hills, as a
companion to Miss Wroat, and believing her to have taken a new
name as a disguise, he boldly asked for Miss Wroat, determined to
see Lally’s supposed employer, and to entreat her to intercede in his
behalf with Lally. The housemaid had told him that Miss Wroat had
gone out for a walk, indicating the direction, and calling up all his
courage, Rufus had started in pursuit.
He saw the dark and slender figure perched on the peaks while yet
afar off. Something in its droop reminded him of Lally, and he came
on at a swinging pace, his eager gaze never swerving from her; and
as he came nearer and yet nearer, the conviction stole upon him that
it was Lally at whom he looked.
“She must have come out with Miss Wroat,” he thought. “Rich ladies
never walk without an attendant. She has dropped behind, being
tired. It is Lally! it is—it is!”
He came up swiftly, the damp soil deadening the sound of his
footsteps. He gained the rocks, and began to climb them to Lally’s
side, but the girl did not stir, nor notice his approach.
A sudden sound at her side at last startled her. With a quick
exclamation, she turned her head—and beheld him!
She did not speak, but her great black eyes grew larger, and her
face grew suddenly so deathly white that he thought she must be
fainting.
“Lally! O Lally!” he cried to her, in an anguished, broken voice.
“Thank God! I have found you! Oh, my darling, my little wife, whom I
have mourned as dead!”
He knelt down before her, in the shadow of a projecting rock, the
tears streaming over his face, and his eyes regarding her in wild
imploring. So a devotee might have knelt to his patron saint, feeling
unworthy to approach her, but longing and praying with his whole
soul for forgiveness and mercy.
Lally felt her soul melt within her.
“Oh, Rufus!” she gasped, in a choking whisper.
He put up his arms to enfold her. She shrank back, not with loathing,
but with a sudden dignity, a sort of majesty, that awed him.
“You must not touch me, Rufus,” she commanded. “I am not your
wife—”
“You are! You are! Before God, I declare that you are my wife—”
“Hush, Rufus! You wrote to me that I was not your wife. Don’t you
remember? You said that our marriage was ‘null and void.’”
“I thought it was. My father told me so!” cried Rufus. “O Lally, I have
been a poor, weak-souled wretch. I am not worthy of your love. I
should have stood by you, instead of basely deserting you through
my own personal cowardice. My father threatened to have me
indicted for perjury, in swearing that we were of age at the time of
our marriage, and I—I was afraid. You can never respect me, Lally,
nor love me again, I know, but if you knew how I have suffered you
would pity me.”
“I have always pitied you,” she murmured.
“I thought you dead. I saw your poor mutilated drowned body in my
dreams. Day and night it haunted me. I was nearly beside myself. I
thought I should go mad. My father’s mind was set upon my
marriage with a great Kentish heiress who loved another than me. I
appealed to her to save me—to save me from my anguish, torture
and remorse, produced by continual thoughts of you! I had no heart
to give her. I was base and unmanly in offering her the dregs of the
cup that had been filled for you; but oh, Lally, I was half mad and
wholly despairing! I wanted the love of some good woman to
interpose and save me from going to perdition.”
“I heard your offer of marriage to her,” said Lally. “And you are
engaged to marry her?”
“No; she refused me. I am free, Lally, and I thank God for it. What
should I have done if I had married her and then discovered that you
still live? I love you and you alone in the whole world. I am of age
and my own master. I have thrown off the shackles my father has
kept upon me. I mean to be brave and honest and true
henceforward, so help me God! I mean to be a man, Lally, in the
best and noblest sense of the word. It shall never be said again of
me that I am ‘unstable as water,’ or that I am a coward. Lally, I offer
you a second marriage, which no one can contest. Will you forgive
me, and take me back?”
His words found echo in Lally’s heart, but she did not speak. Her
pallor gave place to a sudden rose stain, and she began to tremble.
“I came to-day to entreat Miss Wroat to intercede with you for me,”
said Rufus, becoming alarmed at her silence. “I have not a fine home
to offer you, such as Miss Wroat gives you, but I will work for you,
Lally. I will make myself a great painter for your sake. Those
worthless daubs I painted at New Brompton belong to the past life.
Henceforward I will paint better pictures, and show that there is
something in me. We will have two cosy rooms somewhere in the
London suburbs, and you shall have a sunny window for flowers,
and I will work for you, and you shall never know want or misery
again. I can do anything with and for you, Lally, but if left to stand by
myself I shall surely fall. Lally, little wife, take me back!”
He crept up nearer to her and held out his arms.
She crept in to them like a weary child.
She might justly have reproached him for his weakness and
cowardice, and have taunted him with having courted the heiress of
Hawkhurst, but she did neither. She nestled in his arms, and looked
up at him with great tender eyes full of a sweet compassion and love
and offered him her lips to kiss!
And so they were reunited.
For a while they sat in silence, their hearts too full for words. Then
Rufus Black reverently touched her black garments, and asked
simply:
“Are these worn for me?”
Lally shook her head.
“For the lost love and vanished trust?” he asked. “Yes, I see. But, my
wife, if you will love and trust me again, I will try to make your life all
rose color. Poor little wife! How you have suffered! I know the whole
story from Miss Wroat. When I called at the house yonder last
evening and asked for you as Mrs. Peters, a tall bony woman who
stood in the hall came forward and said she was Mrs. Peters. I was
completely mystified, for I had decided in my own mind that you were
known here as Mrs. Peters, but I now see how it is. The old lady
knows your story and was angry at me, and called herself Mrs.
Peters to throw me off your track. She told me all your adventures
since we parted. And now, little wife, let us seek your employer and
tell her that you have taken me back, and that we are to be married
to-morrow morning at Inverness.”
“So soon, Rufus?”
“Yes. I mean to make you mine in a new bond that no one can
contest. I have never taken steps to have our first marriage set
aside, and I think it still stands. But we will be married quietly to-
morrow morning in a Presbyterian church, and we can be so married
without a license or publication of bans. May I take you to church to-
morrow, little wife?”
“Yes,” said Lally, softly. “Oh, Rufus, I do think you are going to be
strong and brave and true henceforward, and if so I shall not regret
what I have suffered. It has been very bitter,” and she shuddered;
“but God is good to us at the last. I will try and be a good wife, and to
strengthen and uphold you.”
“You were always a good wife to me,” sobbed Rufus, with a sudden
remembrance of her gentleness, her tenderness, her strong faith in
him, and her resolute faith that he would some day achieve honors
and wealth. “Oh, Lally, I am not worthy to touch the hem of your
garments, but for your sake I will be a man.”
Lally stroked his cheek softly, as she had been wont to do in the
long-ago, at the dingy lodgings at New Brompton.
“My poor boy!” she whispered yearningly. “My poor dear boy!”
“Shall we go now in search of Miss Wroat?” asked Rufus, drying his
eyes. “I do not see her on the shore. I own I am afraid to meet her,
Lally. It’s a remnant of the old cowardice, you see. But last night,
when she told me your pitiful story, I quailed before her. She must
despise me, and she will surely try to persuade you to cast me off.”
“My poor Rufus!” said Lally, with a gay, sweet smile, such as had not
visited her roguish mouth since the blight had fallen on her life. “Mrs.
Peters is harsh in seeming, but her heart is true and tender. She
loves me dearly, and I love her more as a friend than as a mistress.
One thing we must understand, Rufus,” and Lally’s gayety increased,
“I can’t part with dear old Peters.”
Rufus looked aghast.
“You—you won’t marry me then?” he gasped.
“Yes, Rufus; but I must keep Peters. She won’t leave me; and
besides, it was only yesterday I thought her the only friend I had in
the world.”
“Her name is Peters then?” said Rufus, bewildered. “I traced you two
up from London under the names of Miss Wroat and Mrs. Peters. I
didn’t notice a third name as belonging to the party. By what name
are you known here then, Lally?”
“As Miss Wroat, dear.”
Rufus looked his amazement.
“I—I don’t understand,” he said helplessly. “They said that Miss
Wroat was an eccentric old lady, who was rich, and odd as Dick’s
hatband. Has she adopted you?”
“Do you remember, Rufus, that last morning we spent together at
New Brompton?” said Lally gravely. “I told you then that I had no
relative living except a great-aunt, an old lady who lived in London,
and who was rich, but whose name I did not know. That aunt I
afterward discovered. Her name was Mrs. Wroat. She was an
eccentric old lady, but good and sweet at heart, and I loved her. She
is dead, and it is for her I wear mourning.”
Rufus looked open-eyed astonishment.
“That is not all,” said Lally. “I took my aunt’s name at her death, at
her request. She made me her heiress. I am the owner of the town
house in Mount street, and of the estate of Heather Hills, and have
besides fifty thousand pounds safely invested, so that I have an
income of about three thousand pounds a year.”
Rufus’ arms dropped from his wife’s waist.
“An heiress!” he muttered. “And I have dared to dream that you
would take me back? An heiress! A trifle of money will set you free,
Lally, from any marriage claims, and you can marry according to
your new position. I do not wonder that Mrs. Peters turned me out of
your house, a poor, good-for-nothing coward unfit even to address
you. An heiress! O Heaven! The word is like a two-edged sword
between us!”
He moved backward, white and trembling.
A mischievous gleam shot from Lally’s gipsy eyes.
“I have known so much of poverty,” she said, “that I should like to
keep this wealth. It would make a good basis to build upon. But if it is
‘like a two-edged sword between us,’ I suppose I can endow some
already rich hospital with it, or give it to Peters, or send it to the
heathen.”
“You don’t mean, Lally,” cried Rufus, all agitation, “that you, a rich
lady, will stoop from your high estate and marry me, and try to make
something of me?”
“I do mean just that!” cried Lally, with spirit. “For you know, Rufus, I—
I love you.”
Rufus was at her side again in an instant.
The hour wore on, and the early dusk of the gathering evening fell
around the reunited lovers. Lally started at last, crying out:
“How dark it grows! It must be five o’clock, and Mrs. Peters will be
distracted about me. I don’t know as it is just etiquette, Rufus, but the
circumstances are peculiar, and I don’t believe that Mrs. Grundy has
laid down any rule to fit the precise case, and the situation is so
remote, and I don’t believe anybody will know or care; and so—and
so I’ll invite you to remain to dine with me. But at an early hour—by
ten o’clock, mind!—you must start for Inverness.”
“And you will meet me there at eleven o’clock in the morning, at the
kirk, little wife?”
“Yes,” said Lally solemnly, and with a holy joy in her black eyes, “I
will be there! Who shall part us now, Rufus?”
CHAPTER XX.
AN UNTIMELY ARRIVAL.
In the early dawn of a gray wild morning, Sir Harold Wynde, Lord
Towyn, and Mr. Atkins arrived at Inverness. They proceeded directly
to the Railway Hotel, and secured bedrooms and sitting-room, and
ordered breakfast. Some attention to their toilet was necessary, and
after baths and vigorous brushing, they met in their sitting-room, and
here a very tempting Scottish breakfast was served to them.
They were still lingering at the table, discussing their future
movements, when a knock was heard at the door, and the detective
who had been set to watch Rufus Black entered the room.
He was a thin, small man, with a sharp business face, and looked
the very ideal of a keen commercial traveler, and Sir Harold for the
moment supposed that such was his vocation, and that the man had
strayed into the wrong room. This impression was speedily
corrected.
“Good-morning, my lord,” said the officer, addressing himself to the
young earl. “Expected your lordship up yesterday. Good-morning,
Mr. Atkins. Fine morning, sir—for Scotland. My lord, can I see you
alone for a few minutes?”
“Say what you have to say here, Ryan,” said the young earl. “Mr.
Atkins is a fellow-guardian of the young lady of whom we are in
search, and this gentleman, Mr. Hunlow, is also a friend of Miss
Wynde. Speak out, therefore. Have you any news?”
The detective glanced half uneasily at the baronet, whose striking
face announced that he was no common personage. The gray hair
and gray military beard had not greatly changed Sir Harold’s looks,
but Ryan had never seen the baronet before, and of course
conceived no suspicion of his identity.
The baronet arose and went to the hearth, sitting down before the
fire, his face half turned away from the detective, who again
addressed himself to Lord Towyn.
“There is news, my lord,” he announced. “I succeeded in tracing
Rufus Black up to this place. He stopped at the Caledonian. In fact,
he is stopping there now.”
The young earl’s face kindled with excitement.
“Then we cannot be far from Miss Wynde!” he exclaimed. “He has
stopped two or three days at Inverness, and that proves that he has
not much further to go. Has he been out of Inverness since he
came?”
The detective’s face clouded a little.
“Rufus Black arrived at Inverness the day before yesterday,” he said.
“Upon the afternoon of the very day on which he arrived, while I was
at dinner, he went off in a cab, and did not return till late in the
evening. I was lounging about the door when he came back, and he
looked the very picture of despair, and came in recklessly and went
to his room.”
“That proves that Miss Wynde is not many miles from here,” said
Lord Towyn. “His despair may be readily accounted for, if he had just
come from an interview with her.”
“Yesterday,” continued the officer, “he strolled about the town all the
forenoon, and went down to the river, and visited the wharves on the
Canal, and seemed to be making up his mind to something that
required courage. After luncheon at the Caledonian, he took a cab
and went off again, not returning till midnight last night.”
“And you followed him?” cried Atkins.
“What, and put that sharp old ferret Craven Black on his guard!”
ejaculated Ryan, in astonishment. “No, sir. We’ve got an uncommon
pair to deal with. Mr. Black and his lady are as shrewd and keen as
any old stagers I ever knew. It wouldn’t do to let them suspect that
we are on their track, or they would outwit us yet, and perhaps put
this young lady in peril. People do a many things when they get
desperate that will do them no good, and is sure to harm them if
found out.”
A stifled groan came from Sir Harold. Ryan shot a quick, suspicious
glance at him.
“Then you are at a stand-still, Ryan,” said Mr. Atkins impatiently. “You
have treed the game and sat down to wait for us?”
“By no means, sir,” answered Ryan deliberately. “I saw the cabby
after Rufus Black had gone to bed, and arf-a-crown drew out of the
fellow all that he knew. Mind you my lord, that gold and silver make
the best cork-screws in existence. The cabby owned up all he knew,
as I said, and a pity it wasn’t more. He drove out with his fare to an
estate called Heather Hills, between this and Nairn, on the coast,
and a wild, bleak spot it is, according to cabby. They went up a long
drive and stopped in a carriage porch, and Rufus Black he knocked
and rung, and a house-maid came to the door, and he asked her
something, and she pointed down the coast. And, telling cabby to
wait, young Black went down the bluffs and struck across the fields.
Cabby put an oil-cloth on his horse, for the wind was blowing free
and strong from the sea, and sat there on his box, and sat there till it
began to grow dark and he began to swear a blue streak; and then,
at last, young Black came back with a young lady dressed in black
upon his arm, a hanging on to him so very fond, and a looking up at
him so very tender, that cabby saw that they were lovers.”
“Impossible!” cried Lord Towyn, turning pale. “I will stake my very
soul on Miss Wynde’s courage, and her fidelity to me. No personal
fears, no cruelty even, could drive her into accepting Rufus Black. I
know her brave and glorious nature; I know that she could never
know a moment of weakness or yielding. The cabman has deceived
you, Ryan.”
“No, my lord,” said the detective doggedly. “I’ll stake anything your
lordship likes on his good faith. Rufus Black hung over the lady as if
the ground wasn’t good enough for her to walk on, and she smiled
up at him as loving as—as a basket of chips,” said Ryan, at fault for
a simile, and concluding his comparison rather ignominiously. “The
lady saw cabby, and says she, blushing and smiling, ‘The gentleman
will stay to dinner, and you can put up your horses in the stable,’
says she, ‘and go into the servants’ hall and get a glass of ale and
your dinner.’ And cabby put up his horses, and went into the
kitchen.”
“A queer story,” muttered Atkins. “Perhaps Miss Wynde was playing
a part—pretending to love Rufus Black in order to throw her jailers
off their guard, and so obtain a chance of escape?”
The young earl’s face now flushed.
“I can’t understand it,” he said. “It is not like Miss Wynde to play such
a part, even to effect her escape from her enemies. She is truth
incarnate. She could never summon to her lips those false smiles;
she could never for one moment allow Rufus Black to consider
himself her favored lover.”
“The earl is right,” said Sir Harold. “Neva could never play such a
part. She is too truthful and straightforward.”
The detective bent another quick glance at the baronet.
“Did the cabman make any further discoveries?” inquired Atkins.
“One or two, of some importance,” said Ryan. “In the first place,
there were too few servants for so grand a house. In the second
place, the young lady, with an older woman, had come up here
within a week. In the third place, the house-maid said that her young
mistress was called Miss Wroat, but that that was not her real name,
for the young gentleman had asked for her by another name. And
altogether, an air of mystery seems to hang about the young lady.
But the fact of the most importance of all is, that on the way home
from Heather Hills last night, young Black got up on the box with the
cabby, and asked him no end of questions about the Scotch laws
concerning marriage—if licenses were necessary, if publication of
banns was usual, and so on. And the young man asked him which
was the best church to step into for a quiet, informal marriage,
without license or publication of banns, but the marriage to be
perfectly legal and binding.”
“Ah!” said Atkins. “That begins to look as if he meant business.”
“Young Black seemed to be in gay humor all the way home,” said
Ryan. “He sung to himself, and talked and laughed, and acted as if
he had had a fortune left to him. And as they drove into Inverness,
he told the cabby that he wanted him to take him to church this
morning at a quarter to ten o’clock, and he told him that he was
going to be married to a great heiress whom he adored.”
“Is there not some mistake?” asked Lord Towyn excitedly. “Can he
be in love with some other lady?”
“I should say not,” said Atkins dryly. “Heiresses are not as plenty as
oat cakes in Scotland. He’s been courting Miss Wynde since last
July, and was dead in love with her, as any one could see. He could
not shift his affections so soon, and fix them upon another heiress.
The young lady is Miss Wynde, fast enough. And she is either
deluding him, meaning to denounce him to the minister at the altar,
or to escape from him in Inverness, or else her courage is
weakened, and she believes herself helpless, and has yielded to her
enemies in a fit of despair.”
“If she were alone upon the cliffs, she might then have attempted an
escape,” said Lord Towyn, thoroughly puzzled. “I cannot feel that this
smiling, loving bride is Neva. I know she is not. But we will present
ourselves at the marriage, and if the bride be Neva, we will save
her!”
“I cannot think that she is Neva,” said Sir Harold thoughtfully. “And
yet, as Atkins says, where could he have found another heiress so
soon? And how, if he loved Neva so devotedly, could he be so
deeply in love with this young lady who has just come up to
Inverness?”
“She comes from Kent,” said Ryan. “The housemaid has heard her
speak of being at Canterbury within the month.”
“That settles it!” cried Atkins. “It is Miss Wynde!”
“Ryan,” exclaimed Lord Towyn, “you must go now and discover to
what church Rufus Black is going. We will wait here for you to guide
us.”
Ryan bowed and departed.

Applied Reliability For Industry 3 1st Edition David Delaux Editor Abdelkhalak El Hami Editor Henri Grzeskowiak Editor

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Applied Reliability For Industry 3 1st Edition David Delaux Editor Abdelkhalak El Hami Editor Henri Grzeskowiak Editor

Uploaded by

Copyright:

Available Formats

Applied Reliability for Industry 3 1st

Edition David Delaux Editor

Applied anatomy for the FRCA 1st Edition Abdul Syed

Intelligent Analytics for Industry 4 0 Applications 1st

Oxford handbook for medical school First Edition.

Artificial Intelligence for a Sustainable Industry 4 0

Applied Learning Algorithms for Intelligent IoT 1st

Lignans David Barker Editor

Walking Landscape and Environment 1st Edition David

Applied Veterinary Clinical Nutrition Team IRA 2nd

Operational Reliability for the Automobile,

ISTE Ltd John Wiley & Sons, Inc.

© ISTE Ltd 2023

Library of Congress Control Number: 2022946026

British Library Cataloguing-in-Publication Data

Chapter 1. Durability Approach: Applied to a Vehicle

Chapter 2. Structural Diagrams to Validate the Reliability

Chapter 3. How to Put an Efficient Methodology to

3.6.3. Conducting the tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

Chapter 4. Reliability Study of the High Electron Mobility

Chapter 5. Warranty Cost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

5.4.2. Definition of the transition between “random” and

Chapter 6. Reliability Evaluation of a Luxury Watch Product:

Chapter 7. RBDO of the High Electron Mobility Transistor. . . . . . . 149

7.5. Reliability analysis of HEMT . . . . . . . . . . . . . . . . . . . . . . . . . 156

List of Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167

Summaries of other volumes . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

Predicting and then guaranteeing the reliability of an industrial system is a major

Predicting and guaranteeing the reliability of electronic equipment is a huge and

Grzeskowiak for their remarkable work in the implementation of six volumes in

The first volume is devoted to Predictive Reliability, for estimating or predicting

The third volume presents Operational Reliability; it verifies the reliability

Chapters 1 and 2 elucidate the perspectives of two major automotive

A top-tier automotive supplier adds an important concept to Chapter 3 by

We cannot talk about operational reliability in the manufacturing industry

Likewise, the world of luxury watchmaking is very rarely cited in reliability

Durability Approach: Applied to a

Ensuring the durability of system components is a strategic activity for

Designing a reliable car requires knowledge of information, such as the

In this chapter, a global approach to sustainability will be illustrated through the

Figure 1.1. Instantaneous rate of failure over

As shown in Figure 1.1, a component is subject to different causes of failure

1.2. Example of a vehicle lighting control system

1 Design assistance and validation of automotive reliability.

1.2.1. Risks and reliability requirements

Indeed, the untimely shutdown of the lights is a safety event classified as

Assuming a constant in-service failure rate, this is equivalent to having a

1.2.2. From failure modes to failure mechanisms

Fretting – The presence of the gold layer prevents

Table 1.1. Potential versus probable failure mechanisms

However, it introduces two other failure mechanisms:

In Table 1.1, the following is presented:

Here, we have two main competitive failure mechanisms:

1.2.3. From failure mechanisms to physical damage factors

Failure mechanisms Physical damage factors

Table 1.2. Failure mechanisms and physical damage factors

1.2.4. From physical damage factors to mission profiles or customer

Physical damage factors Linked to

Table 1.3. Relationship of damage factors between design and use

Physical damage factors Source of amplitude or level deviation

Table 1.4. Sources of variation among damage factors

Below, we present the characterization of the stress mission profile:

Figure 1.3. Probability density of the constraint