Ebook Product Maturity Volume 2 Principles and Illustrations 1St Edition Franck Bayle Online PDF All Chapter

Product Maturity, Volume 2 : Principles
and Illustrations 1st Edition Franck

Bayle
Visit to download the full and correct content document:
https://ebookmeta.com/product/product-maturity-volume-2-principles-and-illustrations-
1st-edition-franck-bayle/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Product Maturity 1 : Theoretical Principles and

Industrial Applications 1st Edition Franck Bayle
https://ebookmeta.com/product/product-maturity-1-theoretical-
principles-and-industrial-applications-1st-edition-franck-bayle/
Primary Mathematics 3A Hoerst
https://ebookmeta.com/product/primary-mathematics-3a-hoerst/
Principles of Accounting, Volume 2: Managerial

Accounting 1st Edition Mitchell Franklin
https://ebookmeta.com/product/principles-of-accounting-
volume-2-managerial-accounting-1st-edition-mitchell-franklin/
Accounting Principles, 8th Canadian Edition, Volume 2

Jerry J. Weygandt
https://ebookmeta.com/product/accounting-principles-8th-canadian-
edition-volume-2-jerry-j-weygandt/
Hayes' Principles and Methods of Toxicology [2 Volume
Set] A. Wallace Hayes (Editor)
https://ebookmeta.com/product/hayes-principles-and-methods-of-
toxicology-2-volume-set-a-wallace-hayes-editor/
Product Lifecycle Management (Volume 1): 21st Century

Paradigm for Product Realisation 5th Edition John Stark
https://ebookmeta.com/product/product-lifecycle-management-
volume-1-21st-century-paradigm-for-product-realisation-5th-
edition-john-stark/
Principles and Practice of Sleep Medicine - 2 Volume

Set 7th Edition Meir H. Kryger Md. Frcpc
https://ebookmeta.com/product/principles-and-practice-of-sleep-
medicine-2-volume-set-7th-edition-meir-h-kryger-md-frcpc/
Successful Product Management: Tool Box for

Professional Product Management and Product Marketing
1st Edition Klaus J. Aumayr
https://ebookmeta.com/product/successful-product-management-tool-
box-for-professional-product-management-and-product-
marketing-1st-edition-klaus-j-aumayr/
Function and Evolution of Repeated DNA Sequences 1st

Edition Guy-Franck Richard
https://ebookmeta.com/product/function-and-evolution-of-repeated-
dna-sequences-1st-edition-guy-franck-richard/
Product Maturity 2
Reliability of Multiphysical Systems Set
coordinated by
Abdelkhalak El Hami
Volume 13
Product Maturity 2
Principles and Illustrations
Franck Bayle
First published 2022 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 2022

The rights of Franck Bayle to be identified as the author of this work have been asserted by him 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: 2021952701
British Library Cataloguing-in-Publication Data

A CIP record for this book is available from the British Library
ISBN 978-1-78630-740-8
Contents
Foreword by Laurent Denis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
Foreword by Serge Zaninotti . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii
Chapter 1. Sampling in Manufacturing. . . . . . . . . . . . . . . . . . . . . . . . 1

1.1. Cost aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2. Considering the distribution of defects. . . . . . . . . . . . . . . . . . . . . . . 7
1.3. Considering the test coverage. . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Chapter 2. Compliance Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Chapter 3. Non-Regression Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

3.1. Non-regression on a physical quantity . . . . . . . . . . . . . . . . . . . . . . . 17
3.2. Non-regression depending on time . . . . . . . . . . . . . . . . . . . . . . . . . 20
Chapter 4. Zero-Failure Reliability Demonstration . . . . . . . . . . . . . . . . 23

4.1. Purpose of zero-failure tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.2. Theoretical principle. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.2.1. Non-maintained products. . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4.2.2. Maintained products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.2.3. Estimation of parameter β . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.2.4. Physical laws of failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.3. Optimization of test costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
vi Product Maturity 2
4.4. Specific cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

4.4.1. Imposed number of parts . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.4.2. Imposed testing time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.4.3. Imposed testing time and number of parts . . . . . . . . . . . . . . . . . . 49
4.4.4. A test was already conducted and the demonstrated reliability
should be estimated . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.4.5. One test was already conducted and failure to demonstrate reliability
must be known . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.4.6. Two tests were conducted . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.4.7. A second test is conducted . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
4.4.8. Reliability objective is a failure rate . . . . . . . . . . . . . . . . . . . . . . 69
4.4.9. Reliability data are available from the manufacturer . . . . . . . . . . . . 71
4.4.10. Demonstration of reliability at the product level . . . . . . . . . . . . . . 74
4.4.11. Taking into account a complex life profile . . . . . . . . . . . . . . . . . 76
Chapter 5. Reliability Management . . . . . . . . . . . . . . . . . . . . . . . . . . 79

5.1. Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
5.2. Physical architecture division . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
5.3. Classification of subsets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
5.4. Allocation of initial reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
5.5. Estimation of the reliability of subsets . . . . . . . . . . . . . . . . . . . . . . . 82
5.5.1. Consistency with the experience feedback . . . . . . . . . . . . . . . . . . 85
5.5.2. Estimation of the power of the test . . . . . . . . . . . . . . . . . . . . . . 85
5.5.3. Simulation algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
5.6. Optimal allocation of the reliability of subsets . . . . . . . . . . . . . . . . . . 90
5.7. Illustration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
5.8. Definition of design rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
5.9. Construction of a global predicted reliability model with
several manufacturers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
Chapter 6. Confirmation of Maturity . . . . . . . . . . . . . . . . . . . . . . . . . 115

6.1. Internal data from equipment manufacturer . . . . . . . . . . . . . . . . . . . . 115
6.2. System manufacturer data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
6.2.1. Original fit removal rate or “zero hour returns” . . . . . . . . . . . . . . . 117
6.3. End-customer data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
6.3.1. Burn-in effectiveness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
6.3.2. First failure analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
6.3.3. Method based on failure analysis . . . . . . . . . . . . . . . . . . . . . . . 124
6.3.4. Observed reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
6.3.5. Estimation of the forecasting number of catastrophic failures . . . . . . . 128
Contents vii
6.4. Burn-in optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134

6.4.1. Distribution of failures observed during HASS cycles . . . . . . . . . . . 134
6.4.2. Verification of the degradation of the manufacturing process . . . . . . . 136
List of Notations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
List of Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
List of Acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Foreword by Laurent Denis
Human beings are plagued by major worries, such as fear of death and fear of
illness. “How long will I live?” is a question that arises even in childhood. “Will I
one day have to deal with a condition similar to my neighbor’s?”. We live in an age
where disease, death, old age and disability are subjects to be avoided in polite
conversation. “How are you?” is a standard greeting to which a different and darker
reply than the traditional, “I’m very well, thank you, and you?” risks embarrassing
or even annoying the other party. Avoiding the problems of others, for fear they may
be contagious, gives us a sense of immortality on a daily basis.
This is a rather recent phenomenon, as many previous generations did not hide
the elderly or sick, although the risk of accidents in everyday life was higher and so
death was a more common occurrence. It was certainly a source of anxiety, but the
Church was there to alleviate it. Today we hide this subject by paying attention to a
society made up of young, healthy people whom we must emulate at all costs so as
to be part of it. Since our days are more or less the same, we succumb to
procrastination at the first opportunity and Seneca’s carpe diem loses its wonderful
charm to give way to flat Platonic reflection.
Surprisingly, a similar problem exists in industry: there is a willingness to forget

that a product may be subject to failure during its lifetime, given it has been
optimally designed for the required functions. Some simple principles of upstream
reliability analysis, from the design phase onwards, are now well-established, but
they thwart the deep-seated notion that proper design outweighs everything else.
Two essential points are overlooked: when a technology naturally reaches maturity,
only a technological breakthrough can mark a distinction between two products
performing the same function, unless it can be demonstrated that product A will last
longer and be safer than product B. Moreover, the uses of the same product can
multiply according to its ability to adapt to multiple environments. A good
understanding of these uses in the field makes it possible to improve robustness
x Product Maturity 2
properly at the design stage, in order that it can withstand any mission profile
assigned to it during operation; this is one way to increase competitiveness.
Many companies still see the reliability study of a system before it becomes
operational as a mandatory step to be overcome, bypassing or minimizing it as soon
as possible. In the design phase, a signed product FMECA will end up in a folder, its
purpose merely to certify that the rules have been followed correctly. The objective
of the test phase is to confirm that the device being tested meets the requirements of
a standard, without taking the opportunity to validate that the mission profiles on the
ground will not unpredictably damage the product. During production, process
control cards are used to verify that tolerance limits are not exceeded, without
establishing forecasting instances that could lead to accidental stops. Hence, only
data in the form of returned products, found to be defective by the end user, are
subjected to a posteriori analyses by customer support. This can incur various costs
and may lead to product recall if a serious defect is found.
Fortunately, however, the reality tends to be a little less bleak than the situation
described above, with the emergence and dissemination of best practices that are
based on theories validated by various industry sectors. These are now adapting to
the challenges that companies face: making increasingly complex products that are
more adaptable and ever-faster, while maintaining quality standards and reducing
costs. This no longer involves applying deterministic models in which a single value
is assigned to an objective to be reached. Instead, it is about drawing up a range of
possible solutions that allow the supplier or integrator to make sure that the worst
case a product might be subjected to on the ground can still be controlled by
statistical modeling. The best way to achieve this is through the combined use of
theoretical and technical resources: an in-depth understanding of the possible
technological problems and solutions given by the manufacturer allows the qualified
reliability engineer to build the most suitable predictive models. Ideally, a single
person would have these two complementary sets of skills.
Franck Bayle is a perfect example of this. Throughout the second part of his
career as an electronics engineer, he relentlessly addressed challenges that no one
had previously openly solved, and he developed algorithmic solutions based on
cutting edge theories. He was nevertheless confronted with the ills that plague most
large groups: habit and fear of change. When he proposed significant advances
across the whole company, only his more informed colleagues considered these to
be opportunities for improvement. Sometimes his work was considered useless by
those whose feeling was: “Why consider risks when there are no problems on the
ground?”. This is reminiscent of: “Why would I get sick when I am fit and
healthy?”. We have to be forward thinkers to be able to act before any problem
arises, and Franck Bayle is such a person. His book presents all the best practices he
Foreword by Laurent Denis xi
has managed to implement within his department, as well as all the advances that
I have had the chance to see implemented, which he continues to improve.
This book is essential reading for any passionate reliability engineer, and it is a
real pleasure and an honor to write this foreword to accompany it.
Laurent DENIS
STATXPERT
November 2021
Foreword by Serge Zaninotti
When Franck invited me to work with him on his second book on system
maturity, I immediately accepted. My interest in the subject has grown largely as a
result of the rich technical exchanges we have had over the last 15 years, and
strengthened after reading his first book, published in 2019, on the reliability of
maintained systems under aging mechanisms.
Franck would tell me of his progress in the field of reliability, his field of
expertise, and I – having always wanted to maintain the link between quality and
reliability – would try to establish a connection with the standards.
Indeed, thanks to those who trained me as a quality engineer, I have always

known that quality assurance should never be dissociated from dependability.
I therefore felt instantly motivated by the opportunity to contribute to disseminating
the acquired knowledge by means of a book. The subject system maturity can be
mastered both through experience and through training.
It is often the failures or non-quality observed during the development or

operation of a system that indicate to us that our patterns of thinking lack dimension.
However, in order to find an appropriate response to prevent these unexpected

and feared events, and to be able to control them in the best way possible when they
do occur, it is important to master quality risk management techniques. Risk
management begins with risk prevention, the focus of this book.
In order to understand the problem of system maturity as a whole, before

addressing the actual techniques used, it is necessary to put it in context. This
context is provided by the quality standards for the systems.
xiv Product Maturity 2
Having trained as a general engineer within the Department of Energy and

Environmental Engineering (GEn) at INSA Lyon, I then gained experience as a
quality specialist, and have been a dependability supervisor since 1989. Franck
therefore asked me to present the standards environment and the links that tie it to
maturity, which the reader will find in Chapter 2 of his previous book, Product
Maturity 1.
Serge ZANINOTTI
Thales
Quality Expert
November 2021
Acknowledgements
This book would certainly not have been possible without the contribution of
certain persons. I therefore want to thank, first, my main supervisors throughout my
career with Thales: Jean Riaillon, Laurent Portrait and Claude Sarno, who gave me
the means to gain this experience.
For everything related to maturity, a special thank you goes to Serge Zaninotti,
quality expert with Thales, and also the author of Chapter 2 of my previous book,
Product Maturity 1, on the notion of maturity and the “quality” aspects, and to
Serge Parbaud of Thales for his advice and always appropriate corrections. I would
also want to extend my warmest thanks to Patrick Carton from Thales Global
Service for the passionate technical exchanges we have had in recent years, his
always apt remarks, his support and his listening.
Furthermore, I wish to thank Franck Davenel from DGA for our exchanges
during PISTIS upstream study related to accelerated tests and burn-in, and to give
my warmest thanks to Léo Gerville Réache for his valuable help.
Finally, I wish to thank my entire family, and particularly my wife, not only for
bearing with me, but also for encouraging me while writing this book.
Introduction
Reliability, availability, safety and so on are now major qualities that a product
must have, irrespective of the industrial application field (automobile, avionics, rail,
etc.) of its use. A significant literature related to these fields can be readily accessed,
and is generally grouped under the umbrella concept of “dependability”.
During the whole lifecycle of a product, from specification to operation by the

end user, a large number of actions are implemented in order for it to meet the
specified requirements. Reliability is the quantitative basis for dependability
activities, as poor reliability can lead to insufficient availability, for example,
although it should be reached as soon as the products are in service.
The maturity of a product is therefore its capacity to reach the desired reliability
level, from its launch into service until the end of its operation. Due to technical and
economic challenges, it is very difficult to reach product maturity. Indeed, defects
are very often generated during various phases of the lifecycle, reflected by failures
that occur very early on in product operation (a manufacturing defect, for example),
or during its operation (design flaw, integration flaw, etc.). This is particularly true
for products whose service life is becoming longer (e.g. 30 years for components in
the rail industry). It is important to note that this activity makes sense for maintained
products, which are predominantly in industrial applications.
There is abundant information on maturity, but this applies mostly to process

implementation within a company, and it is therefore often at the project
management level. Detailed literature describing the main theories (worst-case
analysis, derating analysis, etc.) and practical techniques (accelerated tests, burn-in,
etc.) for building product maturity is actually scarce, and many manufacturers often
use obsolete standards, which, at best, they modify according to their experience.
xviii Product Maturity 2
The main objective of this book is to fill this knowledge gap, which is often
detrimental to many manufacturers.
1
Sampling in Manufacturing
Chapters 1 to 6 of Volume 1 described various methods for building maturity.

However, from a manufacturing perspective, these methods must be cost-effective.
One of the solutions that can be considered to reduce costs is to test less than 100%
of the products before delivery to the system manufacturer. This is called sampling.
As expected, there are various standards dealing with this subject, such as
ISO 28590. These standards clarify the sampling rules to be applied, and the
interested reader is invited to read them for further details.
However, the standards do not cover several aspects that are very important for
the manufacturer:
– The cost aspect, which leads to the following questions:
- What is the benefit of applying a sampling rule?
- Is a sampling rule adapted for my application?
- What rule should I use to minimize costs?
– What is the impact of test coverage rates if they are not 100%?
– What is the impact of considering a distribution of potential defects?
This chapter aims to suggest a solution for each of these cases in order to
formulate optimum sampling in terms of quality and cost.
For a color version of all the figures in this chapter, see www.iste.co.uk/bayle/maturity2.zip.
Product Maturity 2: Principles and Illustrations,

First Edition. Franck Bayle.
© ISTE Ltd 2022. Published by ISTE Ltd and John Wiley & Sons, Inc.
2 Product Maturity 2
Theoretically speaking, sampling techniques rely on discrete probability

distributions (the random variable can only take certain values), unlike the
probability distributions for estimating the reliability of a failure mechanism, which
are continuous (e.g. exponential, Weibull, etc.). The Bernoulli distribution is used
for the result of a test (failure or success). When this test is repeated several times,
two cases are possible.
– The “draw” is unrestricted, and in this case the binomial distribution is
applicable.
– The “draw” is restricted, and in this case the hypergeometric distribution is
applicable.
Sampling is obviously a “restricted” draw. Therefore, the theoretical basis of the

sampling norms is the hypergeometric distribution.
1.1. Cost aspects
The function of “cost associated with a size n sampling” is a random variable

(the batch is accepted or rejected). Therefore, its average value (mathematical
expectation) is considered here.
Given:
– C is the cost per unit of the non-compliance test;
– K is the cost per unit of accepting a non-compliant product in a batch;
– N is the number of products to be tested;
– X is the number of “non-compliant” products in the batch;
– n is the size of the sample.
The average cost is equal to the cost of compliant products plus the cost of
non-compliant products. The cost of compliant products (if the batch was accepted)
is given by:
1 , , , , = C. n + X. K . P [1.1]
P1 is the probability of having “no non-compliant parts” in a size “n” sample,

randomly drawn from a batch of size “N”. Therefore it is:
1= [1.2]
where C is the number of combinations of x out of y elements.

Sampling in Manufacturing 3
It is worth considering the details of this result. The probability of an event can
be estimated by the ratio of the number of possible cases to the total number of
cases. It is clear that the total number of cases is equal to the number of
combinations of “n” taken out of “N”, or C .
The number of possible cases is equal to the number of combinations of

compliant parts, or N–X, and is therefore equal to C , hence the result of
equation [1.2].
The cost of non-compliant products (if the batch was rejected) is given by:
2 , , , , = C. N. 1 − P1 [1.3]
Based on equations [1.1], [1.2] and [1.3], the total cost is:
, , , , = C. n + X. K . + C. N. 1 − [1.4]
As an illustration, assume that:

– there are 100 products to be delivered  N = 100;
– the cost of a compliant part is C = 30€;
– the cost of a non-compliant part is K = 1,500€.
According to this data, the cost of a non-compliant part is very high compared
to the cost of a compliant part. Let us now consider the evolution of the total
average cost depending on the size (n) and the number of defective products (X) (see
Figure 1.1).
It can be noted that starting with X > 1, the sampling rate of 100% is optimal. In
this example, the sampling is not interesting in terms of cost.
Now assume that:

– the cost of a compliant part is C = 1,000€;
Figure 1.1. Evolution of the total average cost depending on

the size of the sample and the number of defective products
Although low, the cost of a compliant part is of the order of the cost of
non-compliant parts. Let us now consider the evolution of the total average cost
depending on the sample size (n) and the number of defective products (X) (see
Figure 1.2).
Figure 1.2. Evolution of the total average cost depending on

the size of the sample and the number of defective products
In contrast to the previous example, in this case the sampling is better than the
test at 100%.
Now assume that:

Let us now consider the evolution of the total average cost depending on the
sample size (n) and the number of defective products (X) (see Figure 1.3).
Figure 1.3. Evolution of the total average cost depending on the

size of the sample and the number of defective products
This situation is an intermediate one between the two previously mentioned

examples, as the sampling is optimal when the number of defects is equal to or
greater than 7.
1.2. Considering the distribution of defects
Given p(X), the probability density of the defects observed during this phase of
the test, the mathematical expectation of the cost is:
, , , =∑ p x . C. n + X. K . + C. N. 1 − [1.5]
If this distribution is not known, the uniform probability distribution can be used:
Given that the number of non-compliant products ranges between 0 and N, the
probability density is then:
= [1.6]
Using equations [10.5] and [10.6], the mathematical expectation of the cost is
given by:
, , , =∑ . C. n + X. K . + C. N. 1 − [1.7]
Once again, as an illustration, let us resume the following example and assume
that:
Let us now consider the evolution of the total average cost depending on the
sample size (n) (see Figure 1.4).
It can be noted that for a small sample size, the “average” cost can be very high.
Figure 1.4. Evolution of the total average cost

depending on the sample size – Example 1
Assume that:
Let us now examine the evolution of the total average cost depending on the
It can be noted that for a small sample size, the “average” cost can be very high.
However, a slight optimum is obtained for n = 20, or a sampling rate of 20%, in this
example.
Assume that:
– the cost of a compliant part is C = 1,000€;
Let us now examine the evolution of the total average cost depending on the


It can be noted that for a small sample size, the “average” cost is the lowest.
1.3. Considering the test coverage
Given Pt, the probability of detecting a non-compliant product for a test,

knowing that it will be detected by the client, let us denote by pnMNXt the probability
of having M “non-compliant” parts in a sample of size n, drawn from a batch of size
N, in which X parts are non-compliant.
Figure 1.7. Illustration
Here, the number of possible cases is the number of possibilities of drawing M

defective parts from X and drawing “n–M” compliant parts from the “N–X”
compliant parts of the batch. The probability pnMNXt is then written as
(hypergeometric distribution):
.
= [1.8]
where C is the number of possibilities of having defective products among the X

products from the batch of size N, C is the number of possibilities of having
compliant products in the size n sample among the compliant parts of the batch, and
C is the number of possibilities of having n products among the N products.
The probability of detecting no non-compliant products in the M non-compliant

products drawn in the test sampling is:
= 1−p [1.9]
Finally, let us denote by pn0NXt the probability of having no non-compliant parts

detected in a sample of size n, drawn in a batch of size N, in which X parts are
non-compliant. Hence, according to the formula of total probabilities:
,
=∑ P .P [1.10]
or, based on equations [1.9] and [1.10]:
, .
=∑ . 1−p [1.11]
Let us denote by Xd the number of parts detected as non-compliant when all the
parts are tested following the rejection of the batch, as a result of the detection of at
least one non-compliant part in a sample of size n. The “cost” function associated
with a sample of size n, knowing that Xd non-compliant parts were detected, is
defined as follows:
C n, N, X, C, K, t = C. n + X. K . P + C. N + X − X . K . 1 − P [1.12]
Since Xd follows a binomial law and E[Xd] = X.pt, the average cost depending on
X is:
C n, N, X, C, K, t = C. n + X. K . P + C. N + X. 1 − p . K . 1 − P [1.13]
where 1 − p is the additional cost due to the lack of detection capacity of tests pt.
The cost depends on X number of non-compliant products in the batch, unknown by
definition. Given p(X), the probability law of X for X = 0 to N, the expectation of
the total cost of a test of n parts is therefore:
, .
C n, N, C, K, t = ∑ . C. n + X. K . ∑ . 1−p +
, .
C. N + X. 1 − p . K . 1 − ∑ . 1−p
2
Compliance Test
The purpose of this test is to verify whether the measurements conducted on a

batch of parts comply with a specification, standard, etc. For example, a pull test can
be conducted to verify whether the weld type complies with the bonding strength.
However, the specification can be seen as a normal distribution of the average of the
specified value and zero standard deviation.
The proposal is to use the “resistance/constraint” method, the resistance being

what the test data yields, while the constraint is the specification. The principle is
then to calculate the probability that the constraint is greater than the resistance and
to verify that this probability is below a maximal value Pmax, fixed in advance.
The diagram shown in Figure 2.1 illustrates this method.
The probability Po that the constraint is greater than the resistance is given by:
=ϕ [2.1]
where φ is the distribution function of the standard normal distribution.
In the specific cases when the constraint can be described by a normal

distribution of zero standard deviation, the above formula becomes:
=ϕ [2.2]

Figure 2.1. Overview diagram of the non-compliance test

Compliance Test 15
EXAMPLE.–
Pull tests were conducted on 30 parts using the data in Table 2.1.
3.692 1.731
3.835 1.904
5.706 2.532
5.062 2.015
5.095 0.099
4.157 5.944
3.321 1.342
4.005 5.175
6.193 3.874
5.621 1.91
2.046 5.52
2.346 6.639
1.047 7.828
5.773 3.33
6.828 2.645
Table 2.1. Example of data for a non-compliance test
The Anderson–Darling normality test was conducted to verify data normality

(see Figure 2.2).
Figure 2.2. Test of normality

Therefore, the normality hypothesis cannot be rejected since the p-value is

greater than the accepted risk level of 5%. The average resistance is therefore
μr = 3.663 and the standard deviation is σr = 2.138. The constraint is defined by the
MIL-STD-883G standard. Its average is μc = 1.5 and the standard deviation is
σc = 0. The maximal probability Pmax is fixed at 5%.
This data yields Po = 15.6%, which is greater than Pmax. Consequently, the
batch of parts is rejected.
3
Non-Regression Tests
This chapter aims to verify whether the new definitions of products, subsets or
components are at least as good as the previous ones. As already noted, two cases
are possible.
3.1. Non-regression on a physical quantity
The assumption made here is that the two definitions can be modeled by normal
distribution. This point must be reinforced by a test of normality. To compare two
samples of observations, the classical approach involves using a central tendency
measure, which is generally the mean. Therefore, a comparison of two means will
be made.
The two underlying hypotheses are:
Ho: µ_New ≥ µ_Old
H1: µ_New < µ_Old
To illustrate the proposed method, let us consider the data in Table 3.1.
First, a test of normality should be conducted on the two variables (see

Figures 3.2 and 3.3).

Figure 3.1. Overview diagram of non-regression on a physical quantity
Variable 1 Variable 2
60 58
54 52,5
45 44
41,5 41
41 39
38,5 38,5
37,5 37,5
35 29
32 28,5
28,5
26
26
26
Table 3.1. Example of a non-regression test

Non-Regression Tests 19
Figure 3.2. Test of normality on variable 1
Figure 3.3. Test of normality on variable 2
The Anderson–Darling test yields a p-value of 47.9% and 36.4% for variables 1
and 2, respectively. Consequently, the hypothesis according to which the two
variables are normal is not rejected, at a risk of 5%.
NOTE.– Other tests can be used (Shapiro, Lilliefors, etc.), but the conclusion is
identical for all tests.
The equality of variances should be verified to compare the means. For this
purpose, a Fisher test (or F-test) is used, and a statistic can be built based on the ratio
of two variances according to a Fisher law.
For our example, this yields a p-value of ~ 73%. Therefore, the hypothesis of
equality of variances is not called into question. Finally, a Student test (or t-test) can
be used to compare the means of the two samples. This is a parametric test whose
statistic t is given by:
= [3.1]
. .
.
where:
– n1 is the size of the sample of population 1;

– n2 is the size of the sample of population 2;
– X is the mean of data of population 1;
– X is the mean of data of population 2;
–S is the standard deviation of data of population 1;
–S is the standard deviation of data of population 2.
It can be proven that this statistic follows Student’s law. Our example yields a
p-value of ~ 49% and therefore the hypothesis of the equality of means is not called
into question.
3.2. Non-regression depending on time
This section proposes the use of a method that compares the survival
distributions of two definitions. A non-parametric test is proposed to avoid the
estimation of the parameters of a model of accelerated reliability. The test that
compares the survival distribution, known as the “log-rank test”, is what we are
looking for (see Figure 3.4).
NOTE.– This test is non-parametric and therefore avoids the difficulty of estimating
the parameters of a model, particularly that it can be accelerated.
– Wrongly rejecting a new design is reflected by an improvement of the design,

therefore useless additional costs are avoided.
Non-Regression Tests 21
– Similarly, wrongly accepting non-regression is reflected in operations by

unplanned additional costs and a strong degradation of the brand image in relation
to the customer.
– This test is the optimum in terms of costs since it can be stopped as soon as a
significant difference between the two designs is noted.
Figure 3.4. Diagram of a non-regression test
This test relies on the hypothesis that the two survival distributions are identical.
As an illustration, let us consider the following example.
A test was conducted on a reference design, with 20 parts tested, with the
following results:
A non-regression test was then conducted on a new definition, with eight parts
tested, with the following results:
The graphical representation is presented in Figure 3.5.
Figure 3.5. Example of a non-regression test
The log-rank test shows that the survival distributions are different, and there is a
regression in terms of reliability for the new design.
4
Zero-Failure Reliability Demonstration
4.1. Purpose of zero-failure tests
The purpose of this type of test is to demonstrate a certain level of reliability

with a given confidence level. When a reliability objective must be demonstrated,
the common approach is to submit the components to testing and note the number of
failures. As a result, the tests may prove time-consuming and expensive and thus, in
some cases, a “zero-failure” test may be used.
While such a test has the advantage of being faster and less expensive, it also has
disadvantages. These are generally accelerated tests and since only one test is
conducted, assumptions must be made on the acceleration law. In addition, the
demonstration may fail as soon as a failure is detected. Finally, when no failure is
observed, reliability cannot be estimated; however, it is possible to estimate a lower
bound (for a probability of correct functioning, also known as the survival function)
or a lower bound (for an MTBF). If this bound is above the objective, this means
that the reliability is demonstrated.
4.2. Theoretical principle
As already noted, there are two industrial application categories for which the
reliability objective is different.

4.2.1. Non-maintained products
For this type of application, a Weibull distribution can be viewed as a reliability

model. If Robj is the probability that the mission will be completed successfully after
a duration Tm, the duration of the required test is given by:
T =T . [4.1]
.
where:
CL is the confidence level;

N is the number of parts submitted to the test;
β is the shape parameter of the Weibull distribution.
Demonstration
Given that a random variable X follows a Weibull distribution W(η,β) then the
variable X1/β follows an exponential distribution Exp(1/η). On the other hand, the
higher one-sided bound of the failure rate following an exponential distribution is
given by:
2. N. T
=
χ CL, 2
Given that χ CL, 2 = −2. ln 1 − CL , this yields:
N. T
=
−ln 1 − CL
The “p-quantile” of the Weibull distribution is given by the time after

which “p%” of defective parts were observed, hence:
= −η. ln 1 − p
The previous two equations lead to:
N. T
=−
−ln 1 − CL
ln 1 − p
Zero-Failure Reliability Demonstration 25
or still
ln 1 − CL
T =T .
N. ln 1 − p
or the proportion of functional parts “1- p” at the moment Tm is in fact the objective
of the survival function Robj, which finally yields:
ln 1 − CL
T =T .
N. ln R
End
This type of test is generally conducted under accelerated conditions in order to

reduce the duration of the tests, and therefore the costs. The quid pro quo is that this
requires knowledge of the acceleration law that makes it possible to estimate the
acceleration factor AF between the test conditions and the operational conditions. To
account for the effect of physical constraints on reliability, the most commonly used
model is the Acceleration Failure Time (AFT) model. This model relies on the
hypothesis that an increase in the level of physical contributions accelerates the
time.
Hence, equation [4.1] can be written as:
T = . [4.2]
.
Let us interpret this last theoretical result. It can be seen that the duration of the
test depends on six parameters:
– The duration of the mission Tm, whose value is imposed by the specifications.
It seems obvious that the test time is even longer as this duration is significant.
– The probability of successfully completing the mission Robj. Similarly, this
parameter is imposed by the specifications. The differentiation of Ttest with respect
to Robj yields:
ln 1 − CL −ln 1 − CL
=T . .
N. ln R N. β. R . ln R
This derivative is always positive so that the test time is even longer because the
probability of successfully completing the mission is high, which is a logical result.
– The confidence level CL. This parameter is fixed before the test dimensioning.
Hence:
ln 1 − CL
N. ln R
= −T .
β. 1 − CL . ln 1 − CL
Here, the derivative is always positive which means the test time will be longer
as the confidence level is significant. Obviously, the longer the test time, the lower
the risk involved by demonstrating a reliability above the objective.
– The parameter N, which can be controlled subject to financial considerations.
It can be noted that the greater the number of parts in the tests, the shorter the testing
time, but this parameter is present at a power 1/β.
– The parameter AF, which depends on the level of physical and operational
contributions of the test.
– The parameter β. This is not known, but in the spirit of reliability
demonstration, a conservative value of β is looked for. Equation [4.1] can be written
as:
=x
with
= and a =
.
As β is by definition positive, the parameter “a” is also positive. This type of test
is expected to evidence aging mechanisms, which implies β > 1. On the other hand,
the confidence level CL < 1 and therefore ln(1-CL) < 0. Similarly, Robj < 1 and
therefore ln(Robj) < 0. Consequently, it can be stated that x is positive. On the other
hand, there are three possible cases for the value of x depending on the properties of
the power function:
– either x < 1, and in this case it is a low value of “a” that minimizes the
function f(x). Therefore, a high value of β should be chosen;
– or x > 1, and in this case it is a high value of “a” that minimizes the function
f(x). Therefore, a low value of β should be chosen;
– or x = 1 and equation [4.1] can then be written as T = T , irrespective of
the value of β. Therefore, an arbitrary value of β can be chosen.
As mentioned previously, the parameters Robj, CL and Tm are fixed before the
demonstration of reliability. Let us find the values of parameters N and β, for which
the variable x takes values around 1.
If x > 1, a low value of β must be chosen, therefore β = 1. This leads to:
T ln 1 − CL
T = .
AF N. ln R
We therefore tend to consider the highest possible value of N, since the testing
time is proportional to N. In fact, these two parameters (N and Ttest) can be chosen to
optimize other constraints. If test duration is restrictive (either in terms of time or in
terms of costs), we tend to consider the highest possible value of N in order to
minimize the testing time. If, on the contrary, the tested parts are expensive, a
minimum number should be considered, even if this extends the duration of the test.
If x < 1, a high value of β should be considered. However, for physical reasons,

the parameter β cannot take an infinite value. Indeed, it depends on the aging
mechanism that may be activated during this test. For a given component, several
competing aging mechanisms are possible. It is known that the aging mechanisms
depend on physical contributions applied to the tested component. This is all the
more true in the case of a subset or an entire product.
As a result, when this type of test is conducted at the component level, knowing
the aging mechanism that is the most likely to be activated is essential. This
information is paramount for identifying the type of physical contribution to be used
for the test, and therefore the acceleration law to be employed for the estimation of
the acceleration factor.
When this test is conducted at the subset or product level, it is useful to know the
component with the lowest reliability. All of these considerations also depend on the
operational life profile. For example, if the product is watertight and the
environment has low moisture content, a moist heat test is not necessary.
EXAMPLE.– Assume the aim is to demonstrate a reliability Robj = 92% with a

confidence level CL = 80% after 10 years of operation. The reliability demonstration
test is conducted under conditions such that the acceleration factor with respect to
the operational conditions is AF = 100.
The first step is to observe the evolution of x = depending on the

.
number of parts. In this example, Figure 4.1 is obtained.
Figure 4.1. Example of demonstration of reliability of non-maintained products
It can be noted that:

– for N∈[1 ;19], x > 1. A low value of β should therefore be chosen;
– for N ≥ 20, x < 1. A high value of β should be chosen.
The graphical representation of the testing time as a function of N, with β as a

parameter, yields Figure 4.2.
Figure 4.2. Testing time depending on the

number of parts of non-maintained products
Figure 4.2 reflects the theoretical principles presented above. When the number
of parts being tested is below 20, a low value of β is conservative (which yields the
longest testing time). Starting with N = 20, the opposite is true. Therefore, in this
example, if for other reasons (costs, testing duration, overall external dimensions of
the parts being tested, etc.) we choose, for example, N = 10, a parameter β = 1
should be considered. In this specific case, the test duration is 1,691 hours.
4.2.2. Maintained products
The same approach can be used to find an MTTF, instead of an observed

downtime. It can then be written as:
= . [4.3]
Demonstration
.
As previously noted, η = . Furthermore, as noted in Chapter 1 of
the previous book, numerically MTBF ~ MTTF. Finally, as far as the Weibull
distribution is concerned, it is known that MTTF = η. Γ 1 + . Hence, the following
can be written as:
N. T
=
−ln 1 − CL 1
Γ 1+
β
or still
MTBF − 1−
= .
1
Γ 1+
β
End
Equation [4.3] can be written in the following form:
= b. x
with
= , = and a =
Since β is by definition positive, the parameter “a” is also positive. This type of
test is expected to evidence aging mechanisms, which implies β > 1.
Then, for β > 1, 0.8 < Γ 1 + < 1 and hence b > 1. On the other hand, the
confidence level CL < 1 and therefore –ln(1–CL) is positive. Consequently, it can be
said that x is positive. On the other hand, there are three possible cases for the value
of x depending on the properties of the power function:
– either x < 1, and in this case a low value of a minimizes the function f(x).
Therefore, a high value of β should be chosen;
– or x > 1, and in this case a high value of a minimizes the function f(x).
Therefore, a low value of β should be chosen;
– or x = 1 and this yields the following condition: CL = 1 − exp −N . This case

is numerically unlikely as N is an integer. On the other hand, equation [4.2] can be
written as T = , irrespective of the value of β since x = 1. An arbitrary
value of β can therefore be taken.
Similar to the non-maintained products, the acceleration factor between the test
and operational conditions should be taken into account. This leads to:
= . [4.4]
.
EXAMPLE.– Using the data from the previous example for a maintained product,
there is no need for the parameter Robj. It must be replaced by an MTTF that is
assumed to be MTTF = 100,000 hours.
The first step is to observe the evolution of x = depending on the

number of parts. In this example, Figure 4.3 is obtained.
Figure 4.3. Example of reliability

demonstration for maintained products
Another random document with
no related content on Scribd:
him that we had a cavalry fight a couple of weeks before at
Bardstown, where no doubt his boys were engaged, when on his
further enquiry about the engagement, I told him how we scattered
them all over the country, killing and wounding a great many. The
ladies burst into tears and went back into their room, and the old
man had nothing more to say about his boys.
I then again tried to persuade him to give us at least one
wagonload of bacon, promising him that if he would send his team
and a boy to drive it to Lebanon, he would surely have them
returned, when he again refused in a spiteful, insulting manner. I told
him that we had understood he had some six or eight yoke of work-
oxen and in Texas we knew all about handling oxen and we would
go into his pasture and drive them up and hitch them to the wagons
that were at the house, but this was only a threat. We gave him up
as a bad job and when we reached the pike about three miles from
there, we met a citizen who told us that Wheeler’s cavalry had
evacuated Lebanon and burnt all the meat stored there, which we
were induced to believe, and decided to ride back to the
Harrodsburg Pike and get with our command, which we did.
Our army then continued to retreat, the main part of the army
moving towards Crab Orchard, where we struck the Cumberland
Gap road, while the army, under Kirby Smith, was struggling over Big
Hill, and had still to join the main army at the junction of the roads at
Pitman’s. The army then moved into the mountains on the
Cumberland Gap road, which, owing to the character of the country,
was generally restricted to a single wagon track. This stretched out
our columns of retreat for perhaps twenty miles or more and cut up
the road very badly, frequently causing wagons to stall. Two
infantrymen consequently were detailed with every wagon, of which
we had thirty-eight hundred, laden with provisions and valuable
stores. This detail of two with each wagon was ordered to assist any
wagon that was stalled by taking hold of the wheels, thereby helping
the team to pull the wagon out of the rut.
A division commander was detailed every day to take charge of
the wagon train and artillery and keep it moving. When a wagon
stalled, the whole line of retreat, infantry, wagons and artillery behind
it, would have to wait until it would move again, thereby seriously
impeding our line of march and causing the cavalry in the rear
desperate fighting sometimes to hold off the enemy.
It was reported of General Cheatham, when he had charge of the
train, that one of his wagons was stalled, and he put spurs to his
horse and rode up the line and reached the wagon. The driver was
whipping his mules and the two infantrymen were standing by the
roadside, resting on their guns. At the sight of this, he jumped off his
horse, took hold of the spokes of the wagon wheel and tried to turn
it, but all to no purpose. The two guards still stood resting on their
guns. General Cheatham lost his patience and turned around and
slapped one of the guards in the face. This happened to be an
Irishman, who said, “Be God; if you were not Gineral Cheatham you
couldn’t do this.” General Cheatham pulled off his sword belt, coat
and hat and threw them down by the side of the road and said, “Now,
there lies General Cheatham and here is Frank Cheatham; now light
in.” They say that at this invitation the Irishman lit in and got the best
of the bargain, of which General Cheatham never made any
complaint. The two men then took hold of the wheels in conjunction
with General Cheatham, and started up the wagon, and with that the
whole line of retreat.
This incident was currently reported and generally believed by all
who knew General Cheatham, but I would not be willing to vouch for
the same, as it is almost past belief.
After leaving Crab Orchard, General Buell dispensed with his
cavalry, as they were unable to cope with ours and moved only with
his infantry and artillery in advance. To enter into the details of the
rest of this campaign, would require too much space and will only
say that the brigade of General Wharton, which always includes the
Terry Rangers, in conjunction occasionally with other cavalry, were
expected to and did succeed in retarding the pursuit of the enemy,
restricting his advance to from six to eight miles a day only, thereby
protecting our infantry column, as well as the artillery, ordnance and
thirty-eight hundred wagons loaded with valuable army stores. On
this retreat the infantry were called on only one time to fire a gun. We
met the enemy in a general engagement at Mount Vernon, Barren
Valley, Rocky Hill, Bushy Mound, Wild Cat, Pitman’s Road, Little
Rock, Castle River and many other points, inflicting on them
considerable loss. This mountain service on the part of the Rangers
proved a most severe tax on their endurance, on account of being
deprived of rations. At one time, for nearly two days, we depended
on picking up raw corn left in the camps of artillery and wagons,
where the horses and mules had been fed. A number of times, after
fighting all day long, we had to go out into the hills ten or twelve
miles to find forage for our horses before we could retire to get a little
rest. Our camping places were frequently by the light of the enemy’s
fires.
To give the reader a better idea of the valuable service we
rendered, I will quote an order issued by General Wheeler, read to
us at Cumberland Gap, October 23, 1862.
GENERAL ORDER NUMBER THREE:
“Soldiers of the Cavalry Corps, Army of Mississippi:

“The autumn campaign in Kentucky is over, your
arduous duties, as the advance and rear guard, for the
present, are finished. Your gallantry in action, your cheerful
endurance in suffering from hunger, fatigue and exposure,
render you worthy of all commendation. For nearly two
months you have scarcely been for a moment without the
range of the enemy’s musketry. In more than twenty pitched
fights, many of which lasted throughout the day, you have
successfully combated largely superior numbers of the
enemy’s troops of all arms. Hovering continually near the
enemy, you have engaged in no less than one hundred
skirmishes. Upon the memorable field of Perryville, alone
and unsupported, you engaged and held in check during
the entire action, at least two infantry divisions of the
opposing army. By your gallant charges on that day you
completely dispersed and routed a vastly superior force of
the enemy’s cavalry, driving them in confusion under their
artillery and infantry supports, capturing in hand-to-hand
conflicts many prisoners, forces and arms. Your continuous
contact with the enemy has taught you to repose without
fear under his guns, to fighting wherever found and to
quietly make your bivouac by the light of his camp fires. On
this continued series of combats and brilliant charges, many
great men have fallen. We mourn their loss. We commend
their valor. Let us emulate their soldierly virtues.
“JOSEPH WHEELER,
“Chief of Cavalry.”
CHAPTER XVII.
Omissions in Preceding Chapters

After leaving Cumberland Gap our army again moved into Middle
Tennessee, with headquarters at Murfreesboro. Our cavalry in the
advance camped near La Vergne, at Nolandsville and Triune. The
enemy concentrated at Nashville, from whence they sent out
foraging parties, supported by large infantry forces with which we
had daily engagements, restricting their foraging within a small area
of country. At Nolandsville, where General Wharton made his
headquarters, we camped nearly a month, when Lieutenant Decherd
was instructed to select about fifteen men and cross the Cumberland
Mountain, for the purpose of buying fresh horses, which were very
much needed. I was ordered to go with this party.
While camped near Winchester, Tennessee, intending to cross
the mountain the next day, we heard the distant roaring of the guns
of the battle of Murfreesboro, which was not expected so soon when
we left the command, and which proved a great disappointment to
our party, as we felt that every man was needed for such an event.
We, therefore, hastened back to the army, which we found
evacuating Murfreesboro, and reported. Of the Rangers’ part in that
great battle I will not mention in this, as that is of record in the
general reports of General Bragg and others, and will only say that
they fully sustained their character as one of the leading regiments in
this army, capturing prisoners, artillery, wagon trains, etc., and finally
covering the retreat of the army off the field.
Our army then continued its retreat through Shelbyville to
Tullahoma, our cavalry still operating on the north side of Elk River.
Before crossing Elk River a courier reached General Wheeler from
General Forrest, after Wheeler had crossed the bridge, requesting
him to hold the bridge until he (Forrest) could cross with his
command. Promptly on receipt of this information, General Wheeler,
with a portion of his command, notably the Fourth Alabama Cavalry,
recrossed the bridge to the north side, determined to hold the same
until General Forrest had crossed with his command. Before Forrest
reached Shelbyville, however, General Stanley, with a heavy force of
cavalry, outnumbering Wheeler’s little force ten to one, charged and
forced them back across the river, cutting General Wheeler off from
the bridge. General Wheeler spurred his horse to the bank and over
it, into the dangerous river, which had been swollen by excessive
rains, making a leap of not less than twenty feet, with Stanley’s
cavalry shooting after him and continuously firing on him until he
reached the opposite bank. This was, perhaps, the most miraculous
escape he had during the war.
Before reaching Tullahoma, a Captain Gordon, who had
distinguished himself near Bardstown, where he held in check a
whole brigade of the enemy’s infantry on the Bloomfield Road for a
whole day with only twenty men, was ordered to select twenty men
from the Rangers and enter Kentucky, for the purpose of gaining
information of the disposition of the enemy’s forces, preparatory to a
general raid by our cavalry. The history of this trip, which resulted in
my being wounded and captured and held a prisoner just one year,
lacking a day, I have already recorded, and by an oversight, it crept
into this history ahead of the proper time.
Recurring to the hard service sustained by us in the mountains
between Crab Orchard and Cumberland Gap: The last night we were
on picket duty our company had dwindled down to seven men and I
happened to be on vidette with a messmate, John Cochran. Just at
daylight, when the enemy usually made its appearance, we were
relieved by two others of the command and when we reached the
reserve picket, discovering a grassy spot in the middle of the road, I
told Cochran I must try to steal a little nap, and laid down on this
grassy spot, holding my horse by the bridle, when I was awakened,
only about ten minutes after, by Cochran stooping down from his
horse and jabbing me with his pistol. The reserve picket had formed
a line across the road, just a little back of where I was sleeping and
were firing on the enemy’s advancing skirmish line, the noise of
which failed to awake me and it was only his prodding me with the
end of his six-shooter that got me awake. I had just time enough to
swing on to my horse and get out of there. Here Cochran’s
prediction, frequently made, that he would bet Graber would wake up
some fine morning with a Yankee bayonet sticking in him, came very
near being verified. I merely mention this to give the reader a fair
idea of our complete exhaustion for the want of sleep, continuous
hunger and arduous duties.
CHAPTER XVIII
General Johnston’s Failure to

Strike—Sherman
Recurring to my service in Captain Britton’s company, acting as
escort to General Hood at Dalton, Georgia, where I described the
meeting of the several generals with General Hood at his
headquarters in the rear of Railroad Gap: On our return to camp that
night after supper, Captain Britton suggested he should go up to
headquarters and pump Major Sellars on the meaning of the meeting
that morning. He reported on his return from a visit to headquarters
that General Mower, commanding Hooker’s old corps, had moved
down to Snake Creek Gap during the day, which was located about
nineteen miles in our rear and about ten miles west of Dalton.
General Hood plead with General Johnston that morning for
permission to move out of his works through Railroad and Rocky
Face Gaps with his corps and defeat Sherman’s Army before Mower
could return to reinforce them. Captain Britton said that he would bet
our army would be in full retreat that night, falling back to Resaca,
which prediction was verified, as, by daylight next morning, our
infantry and artillery were engaged with the enemy at Resaca, where
we came very near losing a large part of our army by having their
retreat cut off.
Had General Johnston yielded to General Hood’s plan, there is
no question but what he could have destroyed Sherman’s Army;
here was a golden opportunity lost by General Johnston, and was
the beginning of the downfall of the Confederacy.
After about two weeks I succeeded in getting a horse with the
regiment and continued with the regiment during the whole of the
North Georgia campaign, the details of which I will not venture to
insert, as they will be recorded fully in a history now being written by
Colonel Ben F. Weems of Houston.
During the siege of Atlanta General Sherman started out two
cavalry expeditions, one under a General Stoneman to move around
the right wing of our army, and one under General McCook around
the left wing of our army, both to unite on the Macon line of railroad,
and to destroy and tear up the same, then move on to Andersonville
and release our prisoners. Had these expeditions proven a success,
with an army of probably twenty-five or thirty thousand released
prisoners turned loose in our rear, it would have wound up the
Confederacy. At Atlanta, General Hood took command of our army,
not exceeding thirty-six thousand muskets and, to use his words,
“This army through General Johnston’s retreating from Dalton, had
become an army of laborers by day and travelers by night,” while the
army at Dalton, including Polk’s corps at Rome, numbered eighty-six
thousand muskets, and was better equipped and organized than any
army the West had ever had. The North Georgians and
Tennesseans, largely constituting this army, with their families inside
of the enemy’s lines, were anxious and eager for an advance, and
there is no question of doubt had General Hood been permitted to
give battle at Dalton, our army would have recaptured Tennessee
and Kentucky.
Referring back to the enemy’s cavalry expedition out of Atlanta:
General Stoneman, with a large part of his force, and a lot of
convalescents in the town of Macon, Georgia, were captured near
Macon by General Iverson, commanding Georgia cavalry. General
Wheeler with our brigade, Ross’ and Roddy’s, forced McCook to a
general engagement on the evening of the second day between
Noonan and Philpott’s Ferry, where they finally surrendered, with the
exception of himself and staff, and Colonel Brownlow and some
other line officers, who swam the river that night and made their
escape.
General Wheeler issued an order that night for no man to cross
the river after these fellows, when I, with several of our regiment,
decided there must be some mistake about it and crossed the river
to try to catch these fellows, specially anxious to capture Colonel
Brownlow. Immediately after crossing the river we found a
quartermaster’s clerk, so he represented himself to be, left wounded
at a house. His wound, however, was not very serious we thought.
He had on a magnificent pair of boots, which just about fitted me and
I had been unable to secure boots, only wearing shoes, when I
proposed to him to exchange with me, which he readily did. While he
was pulling off his boots, the lady of the house came in and opened
a tirade of abuse on me for taking a poor, wounded man’s boots. I
told her I had but just come out of a Federal prison where they
treated us worse than that and I was satisfied that my shoes would
prove more comfortable to this man at Andersonville, than the boots,
to which our prisoner agreed. We then continued our pursuit on the
main road to Wedowee, the county seat of Randolph County,
Alabama, occasionally taking a prisoner, whom we would turn over
to reliable citizens, to be taken to West Point where we had a
garrison. We were unable to secure many prisoners, probably not
exceeding eight or ten, as those afoot would hear us coming in the
road and dart into the brush, while their officers impressed every
horse they could lay their hands on and soon outdistanced us with
their fresh horses.
At Wedowee we found a tanyard, where I purchased a lot of good
leather, sufficient to rig a Texas saddle. We had some men detailed
to make saddles, who were experts in such work and moved down
with the army as fast as it retreated. Our first shop was at Ackworth,
Georgia, where they did a good deal of work, but were prevented
from turning out anything extensive ever after, for the reason they
were unable to get leather. I paid one hundred and twenty-five
dollars for the leather I got at this tanyard. Colonel Harrison
promised me, after my return from prison, that if I would furnish the
leather he would have rigged for me one of the finest saddles that
could be made, which was the inducement for me to carry this roll of
leather on my horse’s back.
Going back into the town from the tanyard, we stopped at a hotel
to get some dinner. This was one of the ordinary country hotels with
a porch in front and large square columns under the porch. While
eating dinner, I had a seat at the end of the table where I could see
out on the street. The hotel was located somewhat under the hill,
away from the square, when I discovered Carter Walker, one of our
party, who had finished dinner, behind one of the posts with his pistol
out, talking to some one on the street towards the courthouse.
Having his pistol out suggested to me that there was trouble ahead,
so I jumped up and told the boys to come on. As we got out on the
porch we discovered about fifteen or twenty men on their horses
near the courthouse, with one of them talking to Carter Walker, about
fifty yards distant from us. As soon as we came out, he retired and
when he got back with his crowd, said something to them and
immediately they wheeled and left town. This proved to be a party of
bushwhackers, who were not anxious for a fight with us. We now
decided to return and when a few miles from town, we heard of an
old gentleman, whose name I have forgotten, the only Rebel citizen
in that section, whom we decided to go and see and get some
information from.
After reaching his house and getting acquainted, we decided, on
his urgent request, to stay with him that night, as we were very tired,
as were also our horses, and we did not suppose there was any
great need for our services immediately after the destruction of the
enemy’s cavalry. This old gentleman had had considerable trouble
with his Tory neighbors, who came to his house several nights and
opened fire on him, which he, his old lady and his daughter, a
barefooted girl of eighteen, returned with their squirrel rifles through
port holes cut in the logs of his house.
On the information of our old friend, we decided to visit the house
of a Tory neighbor of his, across the mountain, who belonged to the
Tory regiment in camp at Rome, which we did. Riding up to the
house in blue overcoats, we called for a drink of water, when a lady
invited us in, supposing that we were Federal soldiers. In our talk
with them, there being two other ladies in the house, we represented
that we were Federal spies on our way to Andersonville to make
arrangements about the escape of our prisoners there, which
created quite an interest with these women, who told us that a large
number of young men of the neighborhood belonged to the First
Federal Alabama Cavalry, stationed near Rome, and quite a number
of them were expected home pretty soon on a furlough. We then
arranged with them to tell their boys about our visit and tell them that
we expected to return there in about ten days, as we would probably
need their assistance and we wanted to confer with them. Our idea
was that we would return there at that time, with our company, and
capture the whole outfit.
After making complete arrangements, we started back towards
Philpott’s Ferry, where we again recrossed the Chattahoochie and,
on our arrival at Noonan, found that Wheeler had moved over to
Covington, on the Augusta road.
Riding all that day in a drizzling rain, we called at a house for the
purpose of getting some feed for our horses and something to eat for
ourselves. Night had already set in. We asked the gentleman if he
could take care of us that night, give us a place to sleep on the floor,
as we never slept in a bed, and get something to eat for our horses
and ourselves. His answer was, “Certainly, gentlemen; light and
come in.” I told him before we got off our horses that we were about
out of money and did not have enough, perhaps, to pay our fare,
when he stated that if his wife had anything left from supper we
could have it and he would give us some shattered corn for our
horses. We, of course, didn’t feel very comfortable under such
liberality, but decided to stay, nevertheless, and sleep down in his
barn, some distance from the house.
While we were waiting for his wife to gather what she had left
from supper, he asked us if we were that command the other day
that fired on the Federals when they were tearing up the railroad
near his house. I told him that we were, and he said, “They were in
my pasture trying to catch my horses, when they heard the guns fire
and you ought to have seen those devils run.” When we went in to
supper we found a little piece of cornbread and a little butter, all they
had left from supper, so the woman stated, not enough to satisfy one
man’s hunger. We did not sit down at the table, didn’t touch anything
they had to offer us, and went down to the crib to get the shattered
corn for our horses, which he consented for us to take, fed our
horses and laid down to rest for the balance of the night. Next
morning we got up early and without going to the house, proceeded
on the road towards Covington. Here now, was a fair illustration of
the want of appreciation of a Confederate soldier, with a selfish lot of
people, whom we occasionally met. Rest assured it was very
discouraging to us. The idea of coming all the way from Texas to
fight for and protect these people! He had told us that we saved his
horses from capture by engaging the enemy near his house; you can
imagine our disgust at such treatment. We now proceeded on the
Covington road. When about two miles from there we came to a
large, white house, a magnificent place, and rode up to the gate. A
man about twenty-five years old, well dressed, wearing a white
starched shirt, the first we had seen in a long time, came out to the
gate. When within twenty feet of us, espying the leather on my
horse’s back, tied to the rear of the saddle, he called out, “I want that
leather.” I said, “If you need it any worse than I do, you are welcome
to it.” He said he did, he wanted to make shoes out of it. I told him
that I wanted to make a saddle out if it, to ride to keep Federals off of
him, when he insisted that he needed it worse. I then told him that
we wanted some breakfast and some feed for our hoses. He said,
“All right, gentlemen; light and come in.” Before getting down I said,
“I had better tell you that we are nearly out of money, not enough to
pay for breakfast and feed, away from our command unexpectedly,
but as soon as we get with them and we have an opportunity, we will
send it to you.” He stated that he couldn’t afford to feed us without
pay, that the armies had been around him for some time and had
nearly eaten him out of house and home. I told him that he needn’t
say anything more, that we didn’t want anything he had, although our
horses were hungry, as well as ourselves. As we rode off he called
after us, “I’ll feed you for that leather,” thus adding insult, but we
decided not to notice him.
About three miles further down the road we came to another
house, a somewhat humble cottage, and stopped to make some
inquiry, when a lady came out to the gate and we asked how far
down the road we could find a house where we could get something
to eat for ourselves and feed for our horses. She asked us if we had
tried at the big, white house we had passed on the road. We told her
that we had and were refused because we had no money. She then
insisted that we come in and partake of such as she had, telling us
that she had very little left, as the commissary from Atlanta had
visited her and taken all the corn she had, except five barrels, which
in Georgia, means twenty-five bushels. This, she and her two
daughters had made with their own hands, her husband being in the
Virginia army. She then told us about this man at the big, white
house, who had never been in the army, but had an exemption on
pretense of working in a saltpetre cave and had never had any
forage taken by the commissary from Atlanta, as he had protection
papers, so she called them, from his general at Atlanta. I merely
mention these cases to show you the condition at that time, of the
State of Georgia, the worthy people submitting patriotically to all
manner of abuse by some of our army officials, while some of the
rich, through nefarious practices, escaped the weight of war.
Thanking this lady for her kind offer, which we could not afford to
accept, we continued on this road and two miles further on struck a
large cornfield with tempting roasting ears and decided to stop, build
a fire, dry our clothes and roast corn for our meal, feeding our horses
on the same, in moderation. We had to build our fire of rails taken off
the fence and very soon were enjoying our roasting ears and the
warm fire, being somewhat chilled by the rain. The proprietor of the
place came up the road and, judging from his manner and looks,
was pretty mad, when he said, “Gentlemen, if you had come to the
house I would have gladly given you a good meal and fed your
horses, rather than to see the destruction of my rails.” I told him that
we didn’t believe it, that we had tried several places up on the road
and were refused because we had no money and he, no doubt,
noting that we were in no mood for argument, decided that he had
better say no more. We then proceeded on our road to Covington.
When on our arrival there we found that Wheeler, with all the cavalry
having horses fit for service, had gone on a raid into Middle
Tennessee, by way of Dalton, tearing up the railroad in Sherman’s
rear for many miles, and finally entering Middle Tennessee, returning
by way of Mussels Shoals, rejoining the army below Atlanta.
After the battle of Jonesboro, Hood started on his fatal Middle
Tennessee campaign, his march to the Tennessee River being
covered by our cavalry, making a feint at Rome, Georgia, to which
point General Sherman had followed, confidently expecting to give
Hood battle at Gadsden and never suspecting his move towards the
Tennessee River. While concentrating his army at Rome, Harrison’s
Brigade, under Colonel Harrison, commanding our regiment, made a
feint on Rome by dismounting, hiding our horses in the rear in the
woods, out of sight, and advanced on the outer works of Rome,
preceded by a line of skirmishers. For this purpose, not having our
battle flag with us, we used a new flag, sent us from Nashville, made
by a couple of young ladies from their silk dresses, with the name of
Terry’s Texas Rangers worked in gold letters and some Latin words
on the other side. After skirmishing with Sherman’s infantry a short
time, we retired down the valley, which at this point was perhaps a
couple of miles wide, from the hills to the bottoms.
Falling back that night some six or eight miles, we struck a
wooded ridge, running from the hills to the bottom, perhaps nearly
three miles long. This ridge overlooked the country in front towards
Rome, several miles. General Sherman coming out in person with a
corps of his infantry, expecting to give Hood battle the next morning,
discovered there was only a handful of cavalry in his front, which
was Harrison’s Brigade, and which he was specially anxious to
capture. For this purpose he sent a heavy cavalry force, perhaps
three times our number, into our rear, flanking our position by moving
through the hills on our left, then occupying nearly every road in our
rear, for eight or ten miles. During the night we received
reinforcements of Pillow’s Brigade, a new command, which had been
in only one engagement, at La Fayette, Georgia, where they were
badly handled, causing the loss of a great many killed and wounded
and in consequence, they were a little demoralized. We also
received a section of artillery, two pieces, under a lieutenant, whose
name I do not remember.
This artillery was stationed on a hill to the left of our position,
under an old gin house.
Immediately after taking position the artillery opened on the
enemy, a heavy line of battle making its appearance in the edge of
the woods, about a mile distant. The Rangers were kept mounted,
drawn up near this old gin house, supporting the battery, when all the
rest of the two brigades had been dismounted with their horses
immediately in the rear, out of sight of the enemy.
Very soon a courier from the right of our line, dashed up to
Colonel Harrison and reported that the enemy were flanking us,
down in the bottom, with a heavy force. Harrison abused him, told
him to go back and tell his colonel if he sent him another such
message he would have him court martialed, but very soon a
lieutenant dashed up from the extreme right of our line, reporting the
enemy advancing in the bottom, and about to outflank us, when
Colonel Harrison decided to ride down in the rear of our line and
ascertain conditions for himself. Immediately the enemy raised a
shout and charged. The lieutenant of the battery, concluding that his
guns were in danger of being captured, limbered up and ran down to
the road, where he met Colonel Harrison returning and was by him
ordered to unlimber and open again on the enemy, when he
succeeded in firing one shot and was sabered right over his guns by
the enemy’s cavalry. In the meantime, through some
misapprehension of orders, the Alabama Brigade broke for their
horses, followed by the balance of our brigade, when our regiment
was ordered to charge their cavalry, which we did, striking them on
their flank, using our six-shooters, to which they paid no attention,
simply calling out, “Clear the road for the Fourth Regulars!” This
Fourth Regulars was commanded by a Captain McIntyre from
Brenham, Texas, who was in the United States Army, a lieutenant,
when the war broke out, having just graduated at West Point.
It is hardly necessary to say that finding the enemy’s cavalry in
our rear for a great many miles, resulted in a general stampede,
everybody trying to make their escape out of it. In recording this
engagement I regret to have to mention the loss of our beautiful flag
which, encased in a rubber cover, slipped off its staff and was found
by a Major Weiler, commanding a battalion of the Seventeenth
Mounted Indiana Infantry, and after many years, returned to us at
Dallas, Texas, by Governor Mount and staff, instructed to do so by a
joint resolution of the Indiana Legislature, in response to a memorial,
drawn up and sent by me.
In this engagement the Terry Rangers lost no prisoners, had only
a few wounded and none killed, while the Alabamians’ loss was quite
heavy in prisoners and the balance of Harrison’s Brigade had very
few men taken prisoners. I made my escape by crossing the big
road, being joined by about eight or ten Alabamians, one of whom
was shot in the fleshy part of the thigh, which somewhat demoralized
him, when he called on me, “Texas, can you take us out of here?” I
told him, “Yes, follow me; I’ll take you out.” I struck out straight for the
river bottom, the Federal cavalry not following us, and when out of
sight of the main road, in a little branch bottom, I called a halt and
told the men my plan of trying to swim the river, as the road ahead of
us seemed to be occupied for many miles, judging by the scattered
firing a great distance ahead of us. The wounded man straightened
up in the saddle and asked me if I was an officer. I told him, “No,”
and he said that he was a lieutenant and would take command of the
squad. I told him he could take command of his own men, but he
couldn’t command me, and told his men, “Now, all of you boys that
want to go out with me, come on,” when they all followed me,
including the lieutenant.
Reaching the high ground on the other side of the branch, I
discovered a house, with a lone cavalry-man at the front gate, and,
getting a little nearer, I recognized him as one Joe Harris, of our
company, who was well acquainted in that section, having married,
near Cedartown, the daughter of a Doctor Richardson, just on the
other side of the river. He suggested to me that he knew of a batteau
about seven miles this side of Rome; that we go up there, put our
saddles and equipments into the boat, swim our horses across, then
go to Doctor Richardson and get a good dinner; to which I, of course,
readily consented. On our way to this batteau, following the river in
the bottom, we struck hundreds of Alabamians trying to find a
crossing place. These men we took along with us and when we
reached the boat we were the first ones to cross, leaving the
Alabamians there to cross as fast as they were able. Joe and I then
rode to Doctor Richardson’s, about ten or fifteen miles, and by three
o’clock sat down to a sumptuous dinner. Here we stayed all night
and the next morning recrossed the river, finally striking the main
Gadsden road and finding our stampeded forces gathering at some
gap, the name of which I have forgotten. Here we met General
Wheeler, with the balance of his command. We then moved down to
the town of Gadsden, where we recrossed the river and spent
several days resting our horses and ourselves.
General Hood, in the meantime, with his army, crossed the
Tennessee River, and General Sherman returned to Atlanta, leaving
Thomas’ Corps to follow Hood into Middle Tennessee. Wheeler and
his cavalry returned to below Atlanta, where we struck Sherman’s
forces moving in the direction of Macon, Georgia, by way of Augusta
to Savannah. We then had daily engagements with Kilpatrick’s
cavalry, often driving them into their infantry. Sherman used his
cavalry to forage for the army, depending altogether on the country
for his commissary. To enter into detail of the many engagements
had on this trip would occupy too much time and space. Our service
was largely, as stated, to keep his cavalry from foraging, burning and
destroying the country. In connection with this I would mention an
incident at Macon:
I was at a blacksmith’s shop with a comrade by the name of
Freeman, who was about seven years my senior in age. While
waiting to get our horses shod we heard artillery, supported by small
arms, open at our works, about a mile across the river. We
immediately mounted our horses and dashed over there and just as
we got in sight of the roadway through the breastworks we witnessed
a lone trooper of Kilpatrick’s cavalry coming up the road through the
works, having his horse shot just as he reached inside. His horse fell
on his leg, from which position he was trying to extricate himself and
was about to be shot by an excited militia of young and old men, who
had never been under fire before, when Jim put spurs to his horse
and with his pistol raised, dashed up to where this man lay under his
horse, and drove off the excited militia, I, of course, following him. He
called up a lieutenant, asked his name, company and regiment; told
him to take charge of that prisoner and see that he was well treated,
that he would hold him personally responsible for his safety, and
immediately wheeled his horse, I following him, and returned to town
without giving the lieutenant a chance to ask questions. On our
return I asked Jim Freeman his reasons for doing as he did, risking

Ebook Product Maturity Volume 2 Principles and Illustrations 1St Edition Franck Bayle Online PDF All Chapter

Uploaded by

Copyright:

Available Formats

You might also like

Ebook Product Maturity Volume 2 Principles and Illustrations 1St Edition Franck Bayle Online PDF All Chapter

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

Ebook Product Maturity Volume 2 Principles and Illustrations 1St Edition Franck Bayle Online PDF All Chapter

Uploaded by

Copyright:

Available Formats

Product Maturity, Volume 2 : Principles

and Illustrations 1st Edition Franck

Product Maturity 1 : Theoretical Principles and

Primary Mathematics 3A Hoerst

Principles of Accounting, Volume 2: Managerial

Accounting Principles, 8th Canadian Edition, Volume 2

Product Lifecycle Management (Volume 1): 21st Century

Principles and Practice of Sleep Medicine - 2 Volume

Successful Product Management: Tool Box for

Function and Evolution of Repeated DNA Sequences 1st

Principles and Illustrations

ISTE Ltd John Wiley & Sons, Inc.

© ISTE Ltd 2022

Library of Congress Control Number: 2021952701

British Library Cataloguing-in-Publication Data

Foreword by Laurent Denis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix

Foreword by Serge Zaninotti . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii

Chapter 1. Sampling in Manufacturing. . . . . . . . . . . . . . . . . . . . . . . . 1

Chapter 2. Compliance Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Chapter 3. Non-Regression Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Chapter 4. Zero-Failure Reliability Demonstration . . . . . . . . . . . . . . . . 23

4.4. Specific cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

Chapter 5. Reliability Management . . . . . . . . . . . . . . . . . . . . . . . . . . 79

Chapter 6. Confirmation of Maturity . . . . . . . . . . . . . . . . . . . . . . . . . 115

6.4. Burn-in optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134

List of Notations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

List of Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

List of Acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

Surprisingly, a similar problem exists in industry: there is a willingness to forget

Indeed, thanks to those who trained me as a quality engineer, I have always

It is often the failures or non-quality observed during the development or

However, in order to find an appropriate response to prevent these unexpected

In order to understand the problem of system maturity as a whole, before

Having trained as a general engineer within the Department of Energy and

During the whole lifecycle of a product, from specification to operation by the

There is abundant information on maturity, but this applies mostly to process

Chapters 1 to 6 of Volume 1 described various methods for building maturity.

Product Maturity 2: Principles and Illustrations,

Theoretically speaking, sampling techniques rely on discrete probability

Sampling is obviously a “restricted” draw. Therefore, the theoretical basis of the

1.1. Cost aspects

The function of “cost associated with a size n sampling” is a random variable

P1 is the probability of having “no non-compliant parts” in a size “n” sample,

where C is the number of combinations of x out of y elements.

The number of possible cases is equal to the number of combinations of

As an illustration, assume that:

Now assume that:

Figure 1.1. Evolution of the total average cost depending on

Figure 1.2. Evolution of the total average cost depending on

Now assume that:

Figure 1.3. Evolution of the total average cost depending on the

This situation is an intermediate one between the two previously mentioned

1.2. Considering the distribution of defects

Figure 1.4. Evolution of the total average cost

Figure 1.5. Evolution of the total average cost

Figure 1.6. Evolution of the total average cost

1.3. Considering the test coverage

Given Pt, the probability of detecting a non-compliant product for a test,