Professional Documents
Culture Documents
Forensic Materials Engineering PDF
Forensic Materials Engineering PDF
2003048886
This book contains information obtained from authentic and highly regarded sources. Reprinted material
is quoted with permission, and sources are indicated. A wide variety of references are listed. Reasonable
efforts have been made to publish reliable data and information, but the author and the publisher cannot
assume responsibility for the validity of all materials or for the consequences of their use.
Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic
or mechanical, including photocopying, microfilming, and recording, or by any information storage or
retrieval system, without prior permission in writing from the publisher.
The consent of CRC Press LLC does not extend to copying for general distribution, for promotion, for
creating new works, or for resale. Specific permission must be obtained in writing from CRC Press LLC
for such copying.
Direct all inquiries to CRC Press LLC, 2000 N.W. Corporate Blvd., Boca Raton, Florida 33431.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are
used only for identification and explanation, without intent to infringe.
Dedication
Preface
Case studies of product failure are the most valuable source of information
for manufacturers to improve their products. There is no better way in which
product designs can be changed so as to minimize if not eliminate premature
failure. Case studies are the mainspring of forensic engineering, which is why
we are opening our own case books to a wider audience than the insurers,
lawyers and private clients who originally commissioned them. Some of them
have already been presented to our post-graduate students at The Open
University (T839 Forensic Engineering), and elsewhere such as the annual
ANTEC conferences of SPE (Society of Plastics Engineers). We make
acknowledgment at the appropriate point in the text.
There are now several compilations of case studies, in the U.S., 1,2 Japan3
and in Europe.4 A start has also been made on the World Wide Web, where
ease of access makes this medium an ideal source for users. 5 All such modern
compilations include cases from around the world, simply because of product
globalization. Indeed, many of the cases published by the journal Engineering
Failure Analysis, for example, emanate from rapidly developing economies
such as those of mainland and offshore China, Malaysia and other parts of
South East Asia as well as Eastern Europe.6 One serious problem with many
such compilations, however, is that they tend to ignore many vital aspects,
such as the litigation involved in a case, or the turning points in investigations.
We believe that these aspects are neglected at peril because if they focus on
just the technical details, they do not tell the story of what actually happened.
Product failures are more realistic when related as stories in chronological
sequence, rather than a logical way that makes one wonder why the products
failed in the first place. We have adopted the chronological approach because
it makes the technical matter more interesting, and so encourages the reader
to learn why products failed and what design changes are needed in the future
to prevent similar accidents. Traditional compilations are also limited by the
nature of the materials studied, nonmetallic materials being severely
neglected, despite their increasing importance in safety-critical and structural
roles. However, some case studies of plastic products have been published
independently in the last decade.7,8
Chapter 1 reviews some of the conceptual tools that that have become
widespread in industry. Although these tools can help improve product
vii
design to ameliorate if not eliminate failures, they stand or fall on the abilities
of product engineers to recognize failure modes in a very wide variety of
different situations. This is where the specific skills of forensic investigation
become paramount. One analogy that is appropriate to the work of a forensic
investigator is that of the forensic pathologist, and the parallels between the
two activities are explored in Chapter 2, a chapter that also includes an
introduction to the properties of different materials. A very high proportion
of failures are caused by unexpected loading, or repeated loading, or by
loading in the presence of aggressive chemicals, or by unexpected levels of
stress at specific points in a product. Such mechanical failures demand some
background information about the nature of loads, their magnitude and how
they relate to the properties of the material of which a product is made. The
nature of loading patterns is discussed in Chapter 3, which deals with the
difference between load and stress for an elementary situation that involves
only tensile forces. The concept of loading path is developed. The approach
is vital to understanding more complex loading patterns, such as the type
that occurs in automobile accidents. A variety of different techniques (the
so-called toolbox of Chapter 4) is available to study the physical remains
after an accident. Some tools are very simple: a hand magnifier is often the
only tool that is needed. Even so, the high magnifications from a microscope
are often needed to reveal the failure mode of a specific product. What is
seen is not always perceived, however, which is the justification for the small
amount of background theory supplied in the earlier chapters. All these parts
of the book are provided with illustrative examples taken from our case notes,
many of which have some intrinsic interest for their relevance to the changing
pattern of failure. Thus Chapter 4 includes mention of the failure of a soldered
joint in a circuit board. Thermo-mechanical fatigue is an important generic
failure mode for all electronic equipment, use of which has grown exponentially in the last decade. It is therefore a failure mode that will grow in
importance as equipment ages. Hopefully, designers will recognize the significance of the failure mode and rectify their products accordingly.
The following chapters are grouped according to the kind of defects that
lead to failure. Chapter 5 deals with manufacturing faults in products. It ends
with another area of interest: the application of new technology to improve
manufacturing performance. Thus gas molding seems to be a process that
offers everything without any drawbacks, apart from a license fee for the use
of patented technology. The reality is somewhat different, as the case study
about heavy-duty chair arms shows. Chapters 6 and 7 discuss the frequent
failures encountered with materials used for fluid containment (Chapter 6)
and storage vessels (Chapter 7). While pipe and cylindrical storage vessels
viii
could not be simpler in shape, they can suffer from poor design or materials
selection. Failure of such systems is very unforgiving because escape of fluid
can cause very serious property damage, especially if undetected in the early
stages of failure. The expansion in use of plastics in this area has been
remarkable, and they find widespread application for transportation of many
utilities such as gas, liquid fuels and water (quite apart from speciality chemicals).
The workplace is often the scene of severe accidents that may put employees at risk of serious injury. This area of investigation, which has grown from
legislative pressure to protect employees, especially in hazardous occupation
sites such as mines, steel mills and foundries, is the subject of Chapter 8. One
of the cases concerns failure of a scaffolding barrier, a case that illustrates the
importance of appreciating the wider picture. The immediate cause of failure
appeared to be at one end of a safety rail, although the actual cause of the
problem lay at the other end. One of the most common workplace accidents
involves slips and falls, often from a ladder. While a deceptively simple tool,
a ladder can be very dangerous if simple rules are not rigorously obeyed, as
the final case studies of the chapter show, making a link to earlier problems
with stepladders (Chapter 3).
One of the most rapidly growing markets is for medical devices, such as
implants or tools for surgeons and doctors. Legislation imposes onerous
duties on designers to protect the patient from unwanted side effects of
implants. Product quality and integrity must be of the highest standard
possible, and there are relatively few materials that meet these needs. The
human body is also a very unpredictable environment, which is why product
failures must be investigated thoroughly. Implants and prostheses have often
failed in the past, and it is only by learning from past mistakes that progress
can be made (Chapter 9).
Of all accidents that are familiar to the public, those on our roads and
highways can be the most serious both to drivers and other users. Vehicular
accident is the most common cause of accidental death, yet until Ralph
Nader9 began his quest to improve auto safety, that area had been largely
neglected by designers. Not so any longer. And it is one of the most challenging areas of investigation, often because of the destruction of vital evidence of causation. This is why analysis of trace evidence (such as headlamp
filaments) becomes a key task in reconstruction of the events leading to car
accidents (Chapter 10). Fraudulent insurance claims is the topic discussed
in Chapter 11, where forensic information allows insurers to check the validity of claims made upon them. And although the large majority of cases
discussed in this book are civil law cases, those of a criminal nature are equally
accessible to engineering analysis (Chapter 12).
ix
The final chapter examines patent cases where a patentee sues an alleged
infringer. Unlike most of the failed product cases, a high proportion of such
cases proceed to trial, so they are of particular interest for the detailed level
of forensic scrutiny. Failure to honor the intellectual property rights of others
can be just as serious a design disaster because of the financial consequences
to the loser. If a patent is found valid and infringed, backdated royalties must
be paid to the successful litigant, and the right to manufacture the product
suspended. The largest settlements (e.g., Polaroid v. Kodak) have involved
hundreds of millions of dollars in compensation, and such actions can seriously weaken companies. The cases describe common consumer products
like lawnmowers, wheeled garbage containers and residual current devices
(RCDs). This is a specialized area of forensic engineering that rarely receives
attention from engineers, but which may affect working practice more
severely than often supposed.
The terminology used throughout is American English (e.g., plaintiff
for claimant, coal mine for colliery, etc.).
REFERENCES
1. Shipley, R.J. and Becker, W.T. (Eds.), ASM Metals Handbook, 10th ed., Vol.
11, Failure Analysis and Prevention, ASM, Materials Park, OH, 2002. An
introduction to the subject is provided by Wulpi, D., Understanding How
Components Fail, ASM International (2002).
2. Esaklul, K.A. (Ed.), ASM Handbook of Case Histories in Failure Analysis, ASM,
Materials Park, OH, Vol. 1, 1992; Vol. 2, 1993.
3. Jones, D.R.H. (Ed.), Failure Analysis Case Studies, Elsevier, New York, Vol. 1
(1998); Vol. 2 (2001).
4. Seiko, R., Failure Case Studies in Engineering, Butterworth-Heinemann,
Oxford (1993).
5. For example, The Open University Department of Materials Engineering
maintains a website of case studies at http://materials.open.ac.uk.
6. The journal Engineering Failure Analysis is published bi-monthly by Elsevier
and edited by D.R.H. Jones.
7. Ezrin, M., Plastics Failure Guide: Causes and Prevention, Hanser-SPE, Munich,
1996.
8. Wright, D.C., Failure of Plastics and Rubber Products, RAPRA, Shropshire,
U.K. (2001).
9. Nader, R., Unsafe at Any Speed, Grossman, New York, 1965.
Acknowledgments
We thank The Open University for use of case study material first presented
in two post-graduate courses (T838 Design and Manufacture with Polymers,
1998 and T839 Forensic Engineering, 2000). The two courses, run jointly
with London Metropolitan University, were developed with financial assistance from EPSRC (one of the U.K. Government Research Funding Councils). One of the case studies (container failure of Chapter 5) has also appeared
in T353 Failure of Stressed Materials. The external examiners for T838 and
T839 (Professors Bob Young of UMITS. Roy Crawford FRAEng of Queeens
University, Belfast, and Dr. Colin Goodchild of Glasgow University) gave
helpful advice in preparation of case study material. Drs. Bob Dyson, Mark
Alger and Mike Fitzpatrick of London Met University also assisted in course
development. Dr. Dai (D.R.H.) Jones of Cambridge Universtiy was consulted
during preparation of T839, and encouraged publication of many of the cases
in the journal Engineering Failure Analysis.
We also express our appreciation to the experts and investigators with
whom we have worked on many of the case studies presented in this book,
and the many solicitors, loss adjusters and insurers for their instructions on
most of the cases.
P.R.L. thanks The Royal Academy of Engineering and The Open University for their financial support of several visits to the Failure Analysis and
Prevention Special Interest Group (FAPSIG) of the Society of Plastics Engineers ANTEC conferences in the U.S. Several of the cases mentioned in
Chapters 6, 7 and 9 were first presented there, and they have been updated
as necessary for this book. Drs. Meyer Ezrin (University of Connecticut),
Donald Duvall, and Professors Jan Spoormaker (Delft University), Alan Lessor (University of Massachusetts) and Alex Chudnovsky (University of Illinois at Chicago) provided helpful feedbadk for cases presented at several
FAPSIG sessions at recent ANTEC conferences.
P.R.L. would also like to acknowledge the inspiration of Sir Geoffrey
ALlen FRS in sponsoring the course PT614 (Polymer Engineering) and others
through SERC and Open University funding. He also thanks those of his
students who helped investigate some of the cases in this book, including
Drs. Geoff Attenborough, Bob Ward, Paul Hawkins and Kamal Weerapperuma (Haycarb, Sri Lanka).
We acknowledge the assistance of our able colleagues in the Department
of Materials Engineering at The Open University, especially Dr. George
Weidmann (co-investigator of the plastic tank failures of Chapter 7); Gordon
Imlach (micrography, DSC, FTIR); Dr. Jim Moffatt (mechanical testing);
Mike Levers (photography); Stan Hiller and Richard Black (micrography);
Naomi Williams (SEM) and Dr. Sarah Hainsworth (ESEM); Ian Norman
(computing and scanning); Jennifer Seabrook (secretarial); Peter Ledgard
and Charles Snelling (sample preparation); Clive Fetter and Robert Wood
(editors on T838 and T839); and Roger Dobson (course manager for T838
and T839).
P.R.L. also thanks Dr. John Bellerby and Dr. Mike Edwards of the Royal
Military College of Science (Cranfield University) for interesting discussions
on various case studies while he was an external examiner for their Forensic
Engineering and Science Masters Course. He acknowledges helpful discussion of cases with his fellow examiner on the course, Dr. Maurice Marshall,
lately Head of the Forensic Explosives Labs of the U.K. Government at Fort
Halstead. He also had fruitful discussions with Dr. Sarah Hainsworth of
Leicester University and Dr. Mike Fitzpatrick of The Open University.
Last, but by no means least, we thank our families for their patience and
understanding, which helped us bring this work to fruition.
Peter R. Lewis
Ken Reynolds
Colin R. Gagg
xii
The Authors
Nashville). He has produced more than 300 technical reports and several
books dealing mainly with failure of polymeric and composite products.
Ken Reynolds, M.Sc., the son of a village blacksmith, has had a lifelong
interest in the working and heat treatment of metals. He left school at 16
and took a job in the quality control laboratory of a nonferrous foundry and
tube works where he soon became the works troubleshooter and dealt with
customer complaints. Later he was transferred to the research department of
the same industrial group where he worked for 5 years on the development
of titanium alloys for aerospace. During this time he studied for external
degrees from London University and gained an honours degree in engineering metallurgy and later a masters degree with a thesis on the hot workability
of titanium tin aluminium alloys.
Mr. Reynolds entered the academic world as a lecturer in physical and
industrial metallurgy at Birmingham College of Advanced Technology, later
to become The University of Aston, which pioneered 4-year thin sandwich
degree courses with integrated industrial training. During this time he was
called upon to give evidence at several coroners inquests as to the cause of
failure of machinery and components involved in fatal industrial and road
traffic accidents. The forensic science service at that time was staffed mainly
by scientists, so Mr. Reynolds organized a series of short professional development courses in failure investigation at Aston specifically oriented toward
forensic engineering. These courses were well supported by industry as well
as independent engineers working for insurance companies.
As a UNESCO consultant in 19701971 Mr. Reynolds helped establish
the National Institute of Foundry and Forge Technology in India and immediately afterward undertook a project in Iraq advising on the setting up of
colleges for engineering technicians. Upon his return to the U.K. The Open
University had just been established. Mr. Reynolds was appointed a senior
lecturer in the technology faculty and contributed to all their courses in
materials over the next 25 years.
Mr. Reynolds retired from the academic world in 1996 but continues to
practice as a forensic metallurgist. To date, he has dealt with over 1300 cases.
While at The Open University, he met Peter Rhys Lewis. They have collaborated on several legal cases and compiled a masters course in forensic engineering, which led to Forensic Materials Engineering: Case Studies.
Colin R. Gagg, M.Sc., holds an honours degree in engineering technology
and a masters degree in management and technology of manufacturing. He
is a chartered engineer and professional member of the Institute of Mechanical Engineering. His practical experience includes 2 years at the Structures
xiv
xv
Contents
Introduction
1
3
4
7
10
11
12
12
13
14
15
17
19
20
21
22
Materials in Distress
25
2.1 Introduction
2.1.1 Cause of Demise Engineering example
2.2 Manufacturing Defects
2.2.1 Defects by Omission
2.2.2 Welding Defects
2.3 Failures under Mechanical Overload
2.3.1 Polymeric Materials
2.3.2 Fracture Surfaces
2.4 Fatigue Failure
2.5 Corrosion
2.5.1 Stress Corrosion Cracking of Metals
2.5.2 Environmental Cracking of Plastics
2.6 Creep
25
27
30
31
33
35
39
41
43
52
55
58
58
61
61
62
66
73
85
4.1 Introduction
4.2 An Initial Approach to a Failure
4.2.1 Significance of Initial Observations
4.3 Assessing the Situation
4.4 Inside the Toolbox
4.4.1 Photography and Sample Handling
4.4.2 Product and Material Standards
4.4.3 Macroscopic Examination
4.4.4 Examination under Magnification: Optical
Microscopic Inspection
4.4.5 Metallographic Sectioning
4.4.6 Scanning Electron Microscopy (SEM)
4.5 Mechanical Testing
4.5.1 Tensile Testing
4.5.2 Hardness Testing
4.5.3 Mohs Hardness Test
4.5.4 Vickers Hardness Test
4.5.5 Brinell Hardness Test
4.5.6 Rockwell Hardness Test
4.5.7 Tensile Strength and Hv
4.5.8 Bend Tests
4.5.9 Torsion Testing
4.6 Indirect Strain Measurement
4.6.1 Brittle Lacquer
4.6.2 Crack Detection Methods
4.6.3 Other Crack Detection Techniques
4.7 Chemical and Thermal Analysis
4.7.1 Differential Scanning Calorimetry (DSC)
4.7.2 Infrared Spectroscopy
4.7.3 Chromatography
85
86
87
89
90
91
92
93
xviii
76
94
94
100
105
105
107
107
107
108
109
109
111
113
115
115
117
118
121
121
124
126
Fluid Transport
127
127
128
130
131
131
133
144
146
151
152
155
158
159
165
171
6.1 Introduction
6.2 Materials of Choice
6.2.1 Design Criteria
6.2.2 Corrosion and Environmental Degradation
6.2.3 Case Studies
6.3 Corrosion Failures
6.3.1 Failure of an Oil Delivery Pipe
6.3.2 Vehicle Brake Pipe Failure
6.3.3 Stainless Steel Pipe Clamp Corrosion
6.3.4 Other SCC Failures
6.4 A Plumbing Problem
6.4.1 Reasons for Failure
6.4.2 The Fracture Surface
6.4.3 Water Pressure
6.4.4 Molding Features
6.4.5 Stress Corrosion Cracking
6.4.6 Acetal Polymer
6.4.7 A U.S. Problem
6.4.8 Aftermath
6.5 Fuel Pipes
6.5.1 Cracking and Abrasion
xix
171
171
172
172
173
174
175
177
181
184
184
186
187
189
191
192
192
193
194
194
196
xx
197
198
199
199
200
202
204
204
205
208
208
210
212
212
215
216
217
218
218
219
221
223
226
226
227
228
228
231
233
236
239
241
243
245
247
248
248
250
251
252
253
254
255
xxi
255
255
257
258
260
261
261
263
264
265
267
270
273
273
275
275
276
277
278
281
282
284
286
287
287
288
288
289
289
290
290
10
xxii
292
293
295
297
297
298
303
305
305
306
307
307
310
313
315
316
317
319
320
320
323
325
326
328
331
332
335
335
336
338
340
340
344
11
12
Criminal Cases
350
352
354
358
365
13
349
xxiii
365
368
370
372
373
375
379
379
380
380
381
382
385
386
387
389
389
390
392
392
393
393
394
395
395
397
397
398
Index
399
405
405
406
406
407
408
409
411
413
414
414
414
415
418
420
423
425
425
426
429
xxiv
Introduction
But not all such failures lead to litigation. Product failure occurs frequently
with consumer products, for example, but it is rare to resort to the law to
alleviate the problem. A new car may develop an unexpected defect, such as
a leaking radiator. If covered by a warranty (as most new cars are today),
then the seller will repair the damage at no cost to the user. One typical
condition of warranties is that the car is maintained according to the manufacturers instructions; if not, the warranty may be void. Product or component failure is relatively common with a wide range of products, and there
are well-established routes to replacing failed products without resorting to
litigation. Components with a limited life are thus routinely replaced as a
precautionary measure, and others are replaced when wear becomes visible
and dangerous (such as car tires). However, some failures are more critical
than others, for example, because they may involve safety-critical parts. Such
parts are essential to the correct functioning of a product, and their failure
can lead to safety risk to the user.
A primary thesis of this book is that all unexpected failures, however
trivial they might at first glance appear, should be investigated to determine
the cause or causes. The more serious the failure, such as with safety-critical
components, the greater the need to investigate. Thus, in the case of the faulty
car radiator, if the failure led to seizure of the engine, then the consequence
of the original failure is much greater expense, both to the user and the
manufacturer. If seizure of the engine in turn led to a car accident, then the
safety issue becomes paramount.
So what are the consequences for the manufacturers? They will be concerned about their own quality methods used to make the part that failed,
whether or not there was a design defect, or the problem could have been
caused by the user (e.g., by not maintaining the water level in the radiator).
Introduction
heading. It is up to the user to investigate each possible fault mode systematically in turn to find the faulty component. By eliminating parts that are
functioning correctly, the faulty part should be located quickly and remedial
action taken. Since internal combustion engines work on the same general
principles, the tabulated data are generic to many different makes of car. Such
checklists are certainly a useful starting point for troubleshooting, but they
must be used with common sense so that unsuspected failure modes are not
discounted. Checklists often fail to mention a rather common failure mode:
using the device incorrectly or in a way likely to cause failure. An example
that springs to mind is spilling coffee on a computer keyboard; because water
is conducting and likely to penetrate onto the printed circuit below, failure is
almost inevitable. Even with simple mechanical tools such as ladders, unintended abuse by the user sometimes leads to very serious accidents.
1.2.1
Introduction
Overheating Engine
(Other Possible
Cause)
Coolant Level
Too Low
Leak to Outside
Faulty
Pressure Cap
(Other Possible
Cause)
Internal Leak
Leaking
Pump Seal
Broken Hose
or Clip
Cracked
Cylinder Head
Damaged Cylinder
Head Gasket
Figure 1.2 Fault tree diagram showing causal connections for an overheating
automobile engine.
Pressure Cap
Coolant
Reservoir
Engine Block
Fan
Radiator
a faulty pressure cap (Figure 1.3). There are several dotted lines on the fault
tree that indicate other possible causes (which might be peculiar to a particular model of a certain make of car). Fault tree diagrams are useful because
they place analysis on a systematic footing, and give a mechanic or fitter a
logical path to follow during examination. Each failure mode will exhibit
well-defined features, so if inspection of the engine near the cooling hoses
shows one point where there are dried traces of impurities from the cooling
system, then the adjacent hose will need further, much closer examination.
With machines in which the working principle is totally known or if the
principle of operation is known in outline form only, it is often possible to
Core Phases
Information
Identification
of Need
Patents
Legislation
Competitors
Techniques
Market Analysis
Specification
Prototyping
Standards
Evaluation
Materials Selection
Analysis
Detail
Design
Materials Properties
Costing
Manufacture
Codes of Practice
Market
Manufacturing Methods
Sales and
Usage
Use and
Failure
Wear-Out,
Failure
Testing
Failure Investigation
Figure 1.4 Product design flow diagram with core activities in center, information sources at left and methods at right.
evaluate failure using common sense ideas. The basis of analysis proceeds by
a logical analysis of the way failure has occurred. If the original failure is
buried under consequential damage, then the task is more difficult than for
simple failures. Alternatively, or in addition, the failed component may lie
deep within the machine, and the failure only inferred from the lack of a
particular function. However, clues to the failure may be exhibited by contact
traces of wear or abrasion, for example, where none would normally be
expected. Analysis of such traces is known as trace or contact analysis and
can be extremely useful in locating the source of a particular problem. Forensic investigation often starts after simple inspection by a mechanic or fitter
reveals a component that should not have failed, or shows defective design
Introduction
8
250
A
Number of Defects
200
150
100
50
the
ntr
ize
ce
rs
ve
7O
6O
on
5C
icit
ism
ize
llel
sity
oro
4P
ers
ara
3P
ick
2N
nd
1U
Vital Few
Trivial Many
Number of Defects
100
50
Defect
Figure 1.5 (A) Pareto diagram for several types of defects. (B) Interpretation of
Pareto diagram.
added-cards), using diagrams such as that shown in Figure 1.6, and flow
charts similar to that shown in Figure 1.7 to find root causes.4
Another technique of some value in tackling product defects is known
as failure modes and effects analysis, or FMEA.* This technique is in widespread use in many industries, and required by many standards and codes
of practice. It is based on assessment of the real or suspected defect in terms
of three criteria occurrence, severity and detection and is discussed
further below. All such techniques are, however, simply management tools
* In addition to Chapter 17 in Dales book, many large companies (such as Ford) have
produced their own FMEA manuals, and there is at least one standard, Analysis Techniques
for System Reliability: Procedure for FMEA, 1st ed., 1985, International Electrotechnical
Commission IEC Standard No. 812.5
Introduction
Procedures
Equipment
and Plant
Materials
Effect
Information
People
Hypothesis
Generation
Design
Studies
Statistical
Tools
Determine which
data would
identify potential
cause as a root cause
Collect
appropriate
data
Analyze
data
Prioritization
List
potential
causes
Select most
likely potential
causes; evaluate
each against
problem
specification
Potential
cause
determined
to be a
root cause
No
Yes
No
Problem
eliminated
Correct problems
and monitor results
Yes
Stop
10
Introduction
11
Design: geometry
Design: materials
Manufacture
Assembly
Finishing
1.4.1
Primary Defects
Secondary Defects
Degradation on surface
Primary Defects
Primary features that are often defects in components would typically include
the following:
1. Sharp corners in stressed areas
2. Internal voids or inclusions that raise the stress to unacceptable levels
when the product is loaded
3. Deviations from acceptable dimensions so that the component does
not fit with others during assembly (underfit or oversized, such as
sinking of the surfaces of castings or moldings)
4. Poor quality material of construction (e.g., contaminated material)
5. Cracks in components
Any such features can occur in combination with one another, and may
also interact, such as a void near a sharp corner, a serious defect in a product
where the stress is imposed at the corner. A more complete list of defects is
given in Table 1.1. Note, however, that not all such features may be defects
by themselves. Whether or not a feature can be regarded as a defect depends
on the product specification, and the role of that feature in product failure.
It may, for example, be acceptable to tolerate a certain level of cracks in a
product if they are well below criticality. Indeed, the subject of fracture
mechanics is devoted to the study of such subcritical cracks and how to
recognize under what conditions they may become critical, so as to avoid
those conditions and prevent catastrophic failure.
12
1.4.2
Secondary Defects
Features of the second kind that occur generally over a period of time typically
include:
1.
2.
3.
4.
5.
among many other potential features either within the body or at surfaces
in products (Table 1.1). As with primary features, they will become defects
if they could potentially lead to fracture or failure or rejection for aesthetic
reasons. The distinction between product features and defects is closely
related to the product function, its environment and usage (the specification).
Some of the most severe features that are often primary defects in product
parts are those caused by stress concentration, such as by sharp corners or
by cracks or from impurities or voids within samples.6,7 When products are
loaded as part of their normal function, the stress at such defects is raised
above acceptable limits. Such stress raisers cause the local stress to rise above
the intrinsic strength of the material concerned. The net result is that cracks
are initiated at such defects (to become secondary effects) and can either
grow very rapidly, producing a sudden fracture, or can grow stepwise over
time a number of load cycles in a process known as fatigue.
Introduction
13
FMEA
14
1.6 Litigation
Although the majority of product or process failures are resolved by on-thespot inspection, by compensation from the seller or from insurance cover,
there nevertheless remains a core of failures that cannot be easily resolved
without litigation. It may come from allegedly faulty components supplied
by an outside contractor, or as a result of an accident in which a component
is identified as defective. One party either does not believe, or does not accept
the claims of the other side or sides (for multiparty disputes). This may occur
for several reasons.
* A case study of the implementation of FMEA at Allied Signal in the U.K. is described
in Dales book on page 360 (Chapter 17). He discusses the tendency of many automotive
company suppliers to regard FMEA as a mere paper exercise.
Introduction
15
Trials of Disputes
Disputes are finally resolved by trials, when all other avenues have been
exhausted. So what is so important about a trial? A trial is a public forum
where the opposing parties in a dispute confront one another, and attempt
to convince a court of the rectitude of their own case. It is often the first time
for confrontation, or at least the first time when others are present to order
and marshal all the evidence in a systematic and fair way. Much research will
have been prepared well before the trial, with the collection of evidence to
be presented to the court, and the preparation of arguments to be used for
one case or the other. The arguments are prepared in the form of pleadings,
16
Introduction
1.6.2
17
The raw evidence available for a court to consider comes in several different
forms. It usually includes the following for a typical technical case:
1. Real evidence (the failed product or products, tool parts under question, models and prototypes, etc.)
2. Testimony from witnesses of fact relating to the incident(s)
3. Testimony from expert witnesses
4. Pleadings in the case (drawn up by the lawyers for both sides)
5. Documents discovered from both sides related to the facts at issue
The detailed sequence of events at trial varies from country to country, but
so-called common law countries (mainly English-speaking jurisdictions,
such as the U.S., Canada, Australia and the U.K. plus its many former
colonies) have broadly similar processes. The essential goal is to present
the evidence in a logical and coherent way to the judge or jury who, at the
end of the process, will need to decide which party wins the argument and
the damages to be awarded. Jury trial is more common in the U.S. for all
kinds of civil disputes, while single judges are the norm in Britain. Juries
are always used in criminal actions in both countries. The claimant or
plaintiff usually opens the trial with a speech from his or her lawyer, which
outlines the case against the defendants. It is usually short, and is followed
immediately by examination of the first witnesses of fact for the plaintiff
(possibly including the plaintiff). This is always started by the plaintiff s
lawyer, so the questioning is friendly, being designed to put the witness at
ease, and elicit favorable information about the plaintiff s case.
Cross-examination by the defendants lawyer is often aggressive in tone
and searching in nature. It is when the work of the court really begins, because
the testimony is scrutinized for inconsistency, ambiguity and deliberate falsehood. The latter is rare in civil actions, but unfortunately much more common in criminal trials. Sometimes cases collapse at this point if the plaintiff
(for example) contradicts his or her own previous evidence (from a witness
statement), and cannot explain the discrepancy or discrepancies. In a recent
case where the writer was to appear as an expert for the defense, this is exactly
what happened when the plaintiff contradicted himself on the witness stand,
even when the questions from the cross-examiner were put fairly and calmly.
The cross-examiner had intended only a few questions, because the case
actually hinged on expert evidence, but the questions came thick and fast as
the plaintiff s evidence began to unwind. The case ended the next day, because
the plaintiff faced yet more searching cross-examination, and his credibility
before the court was, by this time, rather low. In addition, the cross-examiner
18
had been using the plaintiff s numerous documents to frame his questions.
The judge had scrutinized many of the documents and found them lacking
key details (some pages appeared, on the face of it, to have been deliberately
obscured during photocopying). The judge asked for the originals, but they
were not produced by the time the case collapsed, with the defendant dropping his counter-claim, and winning all his (substantial) costs.
The end result could not have been predicted before the trial started but
this example demonstrates very effectively the power of cross-examination
to reach the truth of a matter in dispute. The experts were thus never called,
and the technical merits of the case never tested. Halted trials are not uncommon, and it is even more common for civil trials to stop at the doors of the
court when final offers are made between the opposing lawyers. It is a crunch
point for both parties, because they know that all the evidence will be tested
in minute detail. Any weaknesses will be exposed in the light of open court
(normally held in public), so both sides face stark reality before the trial
commences and often reach an acceptable compromise without needing the
added expense of a trial. In trials involving several defendants, the lawyers
must decide whether or not to proceed to trial irrespective of the merits of
a case. A settlement may be reached before trial because the costs of going
to trial are out of all proportion to the costs of the dispute itself. For example,
no fewer than ten different subcontractors were sued by a university due to
corrosion failures that necessitated extensive repairs. Because each party
agreed to a modest settlement, their costs were much lower than if the case
had been heard in a trial that was estimated to last 10 weeks. The defendants
had strong technical evidence that supported their case that the college was
actually at fault for not treating the water supply correctly with anticorrosion
chemicals. The costs of all ten companies would have included appearances
of their senior managers and operatives in the trial together with ten experts
instructed by each company, as well as the much greater costs of the ten teams
of lawyers. So they agreed to a commercial settlement without testing the
merits of the case, a sensible end to what could have been a long trial (with
no necessary clear conclusion), adverse publicity and substantial loss of goodwill.
Often the evidence is tested by without prejudice (i.e., nothing can
be used later in a court) meetings between the experts. It may end up in a
total disagreement (in the worst examples) with an agreement to disagree,
or with partial agreement on the technical issues. This is very valuable to
the court, which usually asks for a statement of the areas of agreement and
disagreement, because it saves time by focusing on the critical points in a
case.
Introduction
1.6.3
19
Expert witnesses such as forensic engineers and scientists have a special place
in any trial because they can give opinions about events, a role denied to
witnesses of fact.10,11 But most of their work is done well before trial, in the
form of site visits, laboratory tests and experiments, literature searches and
presentation of detailed reports. The expert will be given the failed product,
as well as background information on the accident (such as witness statements), and be instructed to prepare a report that should include
1.
2.
3.
4.
5.
6.
Criminal cases of course focus on the crime rather than a product failure,
mutatis mutandis. The expert may feel some responsibility to the solicitor or
company paying for the service or, if an expert within a company, some
loyalty to that company. Even if the case or failure never reaches court action,
the report should always be completely objective, and all attempts to bias the
conclusions one way or another from outside pressure (whether real or
imaginary) must be resisted. A report must be, and be seen to be, a fair
attempt to explain a failure in a perfectly neutral way. There are several ways
of achieving this central objective, but one of the most important is to let
the real evidence speak for itself. In a way, a piece of real evidence is a silent
witness that will reveal all kinds of detail about the failure that may not be
immediately obvious to the untutored eye but which an expert will locate
and show the court.
Photographic evidence (still images given directly, and reference to videos
or film footage) of such details is normally presented in the report, together
with analysis of the material of construction, an account of any mechanical
tests on the material, evidence from new, intact products and so on. A picture
of the sequence of events in the accident will emerge as the evidence from
the failed product builds up in the report. Finally, a reconstruction of events
allows the investigator to revisit the facts which have been established, and
in turn draw conclusions about the cause of the accident. Computer simulations or graphic animations are also useful in explaining mechanisms to a
judge with no technical training. The expert should be able to say how the
product has been made, and whether the material and process route has been
20
Expert Reports
Introduction
21
expert, or get a second opinion. Experts may differ as to the cause of the
accident, especially where critical information is missing. Experts can also
be wrong! It is a sad fact that many investigators fail to probe deeply when
presented with instructions either from within their own company (for internal reporting) or from lawyers. The authors have numerous examples of
expert reports where the investigators have made a superficial inspection of
the failed product and drawn the wrong conclusions. Such reports are brief,
and often include no plans or photographs of the product to support the
conclusions. The danger from the superficial approach to problems is that
the whole legal machine is put into motion on an unsupported base. The
machine often continues on unabated (sometimes for years) until another
expert produces a more firmly based report, which contradicts or qualifies
the original opinion. This situation creates an unwelcome build-up of costs,
money that will have been wasted. The moral is that failure reports must
from the outset examine the evidence in sufficient depth to draw reasonable
conclusions. It perhaps goes without saying that all reports should be written
in clear and concise language, without ambiguity or contradiction. After all,
it is a nontechnical judge or jury who will have to reach their own conclusions
about the facts of a case. It is easy for scientists and some engineers to make
a report too complex, and so lose the lay reader. If complex arguments or
higher level mathematics are needed, they should always be put in an appendix, so as not to break up the logic of a presentation.
22
fracture mechanism. It is now thought, for example, that many early failures
of bridges were caused in this way, although it was not recognized at the
time. Cast iron structures, in particular, fell with some regularity, culminating
in the Tay bridge disaster of 1879. Recent reanalysis of the witness evidence
and the photographic archive of the remains indicate that low-cycle fatigue
of the tie bar lugs contributed to its final demise.14 A similar problem occurred
in the early 1950s with the British Comet aircraft, one of the first to be
designed to fly at high altitude with a pressurized fuselage. Several disappeared over the Mediterranean, but careful and painstaking detective work
on the remains recovered at the scene established without doubt low-cycle
fatigue cracks from window cut-outs in the fuselage. When fatigue cracks
reached a critical size, they grew catastrophically and the planes disintegrated.
Many more recent disasters have also been caused by fatigue of structural
components in an airframe.
As a relatively new class of materials, polymers have found widespread
application in many consumer and industrial products, often in a safetycritical role. New polymeric or composite materials offer many advantages
in a variety of ways, be it in greater design freedom or lighter weight and so
on. But there are some disadvantages, especially as experience of long-term
usage is often absent. Many new polymers have been invented and used only
within the last 50 years or so (polysulfones, for example, are of even more
recent application), so longer-term experience of their behavior is either
absent or difficult to access. Knowledge of the failure modes of such materials
is a vital part of their safe application, and case studies are an important way
of gaining more precise knowledge of their properties. Textbooks tend to
ignore many of the realities of product failure, which is rather unfortunate
given the importance of the subject to designers. Textbooks also tend to
emphasize theoretical aspects of particular materials, without mention of the
practical problems of using those materials in often demanding and safetycritical roles. Reference works are available that provide basic definitions and
background information.15
References
1. Noon, R., Introduction to Forensic Engineering, CRC Press, Boca Raton, FL,
1998.
2. Brown, S., Ed., Forensic Engineering, Part I, ISI Publications, Bermuda, 1995.
3. Carper, K.L., Ed., Forensic Engineering, 2nd ed., CRC Press, Boca Raton, FL,
2001.
4. Dale, B., Managing Quality, 3rd ed., Blackwell Business, Oxford, 1999.
Introduction
23
5. Analysis Techniques for System Reliability: Procedure for FMEA, 1st ed., 1985,
International Electrotechnical Commission IEC Standard No. 812.
6. Young, W.C., Roarks Formulae for Stress and Strain, 6th ed., McGraw-Hill,
New York, 1989.
7. Pilkey, W.D., Petersons Stress Concentration Factors, 2nd ed., John Wiley &
Sons, New York, 1997.
8. Finlay, A. and Lewis, P.R., Total Quality Manufacture of Stationery Products
with FMEA Methods, Teaching Company Quality Seminar, Institute of
Marine Engineers, April 22, 1992.
9. Acland, A.F., Resolving Disputes Without Going to Court, Century Business
Books, Evergreen, CO, 1995.
10. Reynolds, M.P. and King, P.S.D., The Expert Witness and His Evidence, 2nd
ed., Blackwell Scientific, Oxford, 1992; Smith, D., Being an Effective Expert
Witness, Thames Publishing, Norwich, U.K., 1993.
11. Bronstein, D.A., Law for the Expert Witness, CRC Press, Boca Raton, FL, 1999.
12. Huber, P.W., Galileos Revenge: Junk Science in the Courtroom, Basic Books,
New York, 1991.
13. Ratay, R.T., Ed., Forensic Structural Engineering Handbook, McGraw-Hill, New
York, 2000.
14. Lewis, P.R. and Reynolds, K., Forensic engineering: a reappraisal of the Tay
bridge disaster, Interdisciplinary Sci. Rev., 27(4), 287298, 2002.
15. Walker, P., Lewis, P.R., Reynolds, K., Weidmann, G., and Braithwaite, N., Eds.,
Dictionary of Materials Science and Technology, Chambers-Harrap, Edinburgh, 1993.
Materials in Distress
2.1 Introduction
All materials have limits as to what they are able to withstand in terms of
forces that can be applied and environmental conditions under which they
can safely operate. Designers and manufacturers strive to ensure their products meet the appropriate materials and engineering specifications and
should survive an acceptably long, trouble-free service lifetime in a particular
application. However, no material is universally suited for every type of
engineered product and no product is unbreakable or will not fail in some
way if the service conditions are hostile. When a product does not achieve
the expected level of performance it is said to have failed. That does not
necessarily mean that it has broken but simply that it may have worn at an
unusually high rate, suffered debilitating environmental attack or degradation or anything that would prejudice its future performance if it had been
allowed to remain in service. The statistical incidence of failures for a particular product, whether it be a single item or a complex assembly such as a
motor vehicle, exhibits a characteristic bathtub shape as illustrated in Figure 2.1.
Initially the high incidence is associated with not getting the product
quite right in terms of satisfying every aspect demanded of the application,
or minor difficulties with quality of materials or manufacturing. When these
initial difficulties are resolved the incidence of failures flattens off to a very
low level and remains so for the intended lifetime of the product and beyond.
However, the incidence later rises steeply due to products literally wearing
out. A characteristic wear and tear failure mode of individual components
soon becomes recognized and, for this reason, preventive maintenance to
replace such parts before they fail catastrophically both greatly extends the
25
Number of Failures
26
Failures identifying
minor design fault or
need to adjust material
specification
Approaching end of
useful lifetime; failures
due to wear, fatigue,
creep, corrosion, etc.
Time in Service
useful lifetime of the machine and promotes safety. For an aircraft there are
no layovers at 40,000 ft.
The metallurgist and the materials engineer may aptly be regarded as the
pathologists of the engineering failure. The medical pathologist is called upon
to ascertain the cause of death by applying knowledge of the way the human
body works and how disease and trauma affect cells, tissues and organs. Hence,
if the driver of a crashed vehicle is found to be dead after an accident it is vitally
important to establish whether death was due to some natural cause such as a
stroke or heart attack that led to loss of control or whether it was the result of
violent impact in the collision. A number of parties will be keen to know the
findings, not only those investigating the cause of the accident but also insurers
of the driver and owners of property damaged in the accident. A whole chain
of litigation may be initiated on the basis of the pathologists findings.
But what if the driver was not killed but claims the vehicle suddenly went
out of control as it was entering a curve and, when the wreckage is examined,
it is discovered that some vital part of the steering or braking system is
broken? It is just as important to establish whether this was broken in the
collision or was the cause of the drivers loss of control. This calls for investigation by someone able to relate the way the part failed to the circumstances
of the accident. The necessary skills include the ability to interpret fractures
and identify failure modes in terms of the mechanical properties and manufacturing methods of metals, plastics and other materials used in engineered
products. Armed with such knowledge and experience, it becomes a fairly
straightforward matter to determine whether a particular component recovered after an accident was damaged or broken by sudden, massive overstress
in an impact or whether it failed spontaneously as the result of gradual wear
and tear or some progressive weakening (disease) such as corrosion or
fatigue.
Materials in Distress
27
28
Materials in Distress
29
(A) Unloaded
(B) Loaded
(C) Thermoplastic
(F) Woodworm
30
stresses acting along the bottom surface of the shelf. It would also reduce the
effective cross-sectional area at that point, which would further increase the
stress level. As books were placed on the shelf the bending action would
generate increasing tensile stress in the bottom fibers. If the defect were severe
the shelf might break at this position before it was fully loaded. However, if
the shelf could be fully loaded without causing instantaneous fracture, then
it is possible the defect might propagate slowly over a period during cycles
of stress associated with adding and removing books a few at a time. Eventually the defect could reach a size where the shelf was no longer capable of
withstanding one final stress cycle, so it would suddenly give way under a
load that had been satisfactorily withstood on many previous occasions
before the crack had grown so far into the section. This is what happens in
a fatigue failure.
It is a fairly straightforward matter to ascertain from the appearance of
the fracture surfaces whether the failure was of an instantaneous form or
whether it was a progressive fatigue failure. Such an examination would also
reveal the size and nature of the original defect, from which it should be
possible to establish at what stage in the manufacturing process it was introduced. In the above circumstances responsibility would rest with the manufacturer, not only on the grounds of introducing the defect in the first place
but also for failing to detect it in the finished product by appropriate quality
control procedures.
The failure in Figure 2.2F illustrates the kind of progressive weakening
in service that may be caused by environmental conditions. The illustration
represents a gradual eating away (literally) of cross-sectional area by woodboring insects, but a similar weakening could result from fungal attack under
moist conditions, as well as other forms of environmental degradation for
other materials. In either case, visual and microscopic study of the fracture
would reveal the cause. Tests might be conducted on samples cut from near
the break in order to demonstrate the reduction in strength that had resulted
from the attack. Clearly neither the designer nor the manufacturer could be
held responsible for the consequences of such failures.
Materials in Distress
31
32
Materials in Distress
33
34
Materials in Distress
35
3000
Steel
Wires
Carbon
Fiber
Aramid
Fiber
2500
Oriented
PE Fiber
2000
1500
Carbon-Fiber
Reinforced
Titanium Plastic
Alloys
1000
Bulk
Steels
500
Brasses
Iron
Copper
Alloys
Magnesium
Alloys
Glass
Glass-Fiber
Reinforced
Plastic
Aluminum
Alloys
Bulk Plastics
36
Materials in Distress
37
Necking
Tensile
Strength
Fracture
(Ductile)
Force
Force
Fracture
(Brittle)
Yield Point
(Onset of Plastic
Deformation)
Strain (Elongation)
Strain
Force
Fracture
Extension
Figure 2.5 Forms of tensile test curves for three different types of material.
(A) Brittle (e.g., ceramic). (B) Ductile (most metals). (C) Thermoplastic polymers.
38
Figure 2.6 Cup-and-cone tensile fractures of high tensile steel wires in bead
of burst tire (scanning electron micrograph).
more, that the direction of the external force that caused it must be consistent
with the nature of the deformation.
If the tensile test is continued until the test piece fractures, the curve
passes through a maximum, which is the value quoted as the tensile strength,
but then drops down until the test piece breaks at a lower value of applied
force. This is because the cross section of the test piece begins to neck down
in one position and all deformation thereafter is confined to the region where
the necking is taking place. If the applied force is divided by the actual area
sustaining it, then the true stress continues to increase, albeit at a much
lower rate than the earlier part of the deformation curve. The extension
axis is also no longer an accurate figure for the amount of deformation,
because this is now confined to the necking region, not over the whole gauge
length of the test piece. When this, too, is corrected and substituted on the
strain axis, the true stress-true strain curve continues to rise and the fracture
appears in a very different position from the engineering stress strain curve.
Usually this difference between the tensile strength and fracture stress
may be ignored but it does become significant when examining a component
that has broken in tension across a screw thread or a locally narrowed section
like a keyway or circlip groove. In estimating the applied force that caused
this damage account must be taken of the actual cross-sectional area over
which it acted, not the nominal cross sectional area.
Because of the way necking takes place in a tensile test the fracture of a
circular cross section exhibits a waisted appearance over the neck terminating
as a cup and cone fracture at the narrowest part, as illustrated in Figure 2.6.
Materials in Distress
39
Such fractures are commonly found in the high tensile bead wires from burst
tires and in wire ropes that have been subjected to gross overload.
All common engineering metals are ductile, although some alloys (i.e.,
mixtures of metal atoms on an atomic scale) may become brittle under
certain circumstances, for example, steel fully hardened to form the cutting
edge of a tool. Ceramic materials (e.g., glass, sintered oxides, etc.) are intrinsically brittle because their constituent atoms are chemically bonded and
unable to be displaced like metals. Nevertheless, they still behave initially as
if the atoms are held together by springs, so their force-extension curve is a
straight line, as in Figure 2.5B, terminating in sudden fracture. They do not
display any significant plastic deformation prior to fracture; if the broken
pieces are offered up to each other the specimen looks just as it did before
the test. When the applied force reaches a critical level, microcracks suddenly
run together, join up, and the product fractures. On account of this behavior
brittle materials tend to be restricted to applications where the loads are
predominantly compressive.
Elastic energy is stored as the applied force is increased and when fracture
occurs this is suddenly released, causing the product to break up into several
pieces as cracks propagate unrestricted from one free surface to another. By
careful replacement of the broken pieces it is possible to identify the nature
of the external force and the position where it acted. The ability of ceramic
products to withstand high levels of applied tensile force may be substantially
increased by building in compressive stress at the surfaces when the product
is manufactured as, for example, in toughened glass. External force must first
overcome this compressive force before it can place any tensile strain on the
glass. However, when such products do fail the stored elastic energy is greater
so the product breaks up explosively into myriad small pieces as this energy
is released.
2.3.1 Polymeric Materials
Thermoset plastics materials are also brittle because their long polymer
chains are linked by chemical bridges (cross-links) that anchor them together
to retain their as-molded shape. By introducing high tensile strength fiber
reinforcement (e.g., very fine glass, carbon or aramid fibers) numerous interfaces are formed within the brittle polymer matrix that prevent the microcracks running all through the material when the critical tensile stress is
reached. The microcracks do in fact begin to run through the plastic matrix
but they soon meet a fiber that blunts the tip and stops them from going any
further by reorienting the crack parallel with the tension axis at the interface.
This effect is readily recognizable when a semitransparent fiber-reinforced
40
product has been subjected to gross overstress by the milky white translucent
area that develops around the point where the damaging force was applied
due to the scattering of light by the myriad interfaces.
In amorphous (noncrystalline) elastomeric materials the long molecular
chains are not cross linked by strong chemical cross-link bridges but are
simply tangled together, held by very weak bonds. Such materials include
natural rubber and the large range of synthetic elastomers such as NBR (used
in fuel pipes) and fluoroelastomers such as Viton (which became more widely
known after the Challenger space disaster). Their inherent flexibility and
reversible elasticity make them key engineering materials because they can
absorb vibration, and so reduce the debilitating effects of fatigue, for example.
The same property makes them ideal seals against fluid movement in engines
and motors.
Thermoplastic materials such as polyethylene or polypropylene are usually partly crystalline, which makes them much more rigid and thus suitable
for use in engineered products. However, they are still much less stiff than the
same geometry in a typical metal, so their design to withstand imposed loads
must be changed quite drastically. They can also be transformed into very
strong flexible products by straightening out the molecular chains. This shows
up well in their tensile force-extension curves, as in Figure 2.5C.
When the force is first applied there is only slight resistance so the
extension increases rapidly while the applied force is doing nothing more
than straightening out the chain. Extensions of several hundred percent may
occur without any sign of fracture, though the cross-sectional area of course
becomes correspondingly smaller as the strain increases. Eventually, all the
chains become straightened out so the tensile force is now acting along the
line of the molecular chains. Because these are strong chemical bonds the
force rises steeply as very little strain is possible. Eventually the test piece
breaks where the chains have the least overlap. Narrow strips may be split
away easily to give very strong fibers (e.g., polypropylene), which can subsequently be woven into products like string, binding tape and rope, which
have high tensile strength and are flexible. Even stronger materials such as
carbon fiber can be made by simultaneously heating and stretching, while
other high performance polymers can be made by control of chain orientation during manufacture (e.g., UHMPE fibers such as Spectra and Dyneema)
or synthesis (e.g., aramid fibers like Nomex, Twaron and Kevlar).
One disadvantage of some common thermoplastic materials for loadbearing applications is that they soften and deform slowly under quite modest
loads at high temperatures. There are two thermal measures that are important: Tg (the glass transition temperature) and Tm (the melting temperature)
if the polymer is partly crystalline. Tm is always above Tg and many engineering thermoplastics have a Tm above 200C. The ultimate use temperature lies
Materials in Distress
41
42
Figure 2.7 Characteristic fracture paths in ductile (left) and brittle (right) materials subjected to different overstress modes.
skin becomes brittle. While the bulk of such a component will be heat treated
to achieve the desired high strength and toughness and satisfy all the design
parameters, it may fail in a brittle manner, if subjected to bending impact or
some condition where the tensile stresses at the surface are not uniform. A
crack initiated in the brittle case or surface layer may run right across the
section in a brittle manner. More often though, the case cracks and the
underlying material deforms before it eventually fractures, as happened to
the ball stem shown in Figure 2.8.
Steels are also anomalous inasmuch as under impact bending a section
that is ductile when tested at room temperature may fracture in a brittle
manner at subzero temperatures. This ductile-brittle transition temperature
as it is called is sensitive to a number of metallurgical factors, including
carbon and impurity levels, grain structure, the presence of any nonmetallic
Materials in Distress
43
Figure 2.8 Cracks in the brittle surface hardened layer on the stem of the
steering ball joint damaged in the accident.
44
Figure 2.9 Single bending fatigue fracture in a pneumatic tool. (A) Fracture
surface on the shank. (B) Blunted cutting edge that set up abnormally high cyclic
bending stresses in the shank.
approximately 40% of their tensile strength but for most materials there is
no safe limit below which it ceases to propagate. Eventually the effective
cross-sectional area becomes so reduced that the component ruptures under
one final cycle of a load near the top of its normal service spectrum, but
one at a level that had been satisfactorily withstood on many previous
occasions before the crack propagated. The final fracture occurs suddenly
in a ductile or a brittle mode, depending on the characteristics of the
material.
Fatigue fractures exhibit a characteristic appearance, which reflects the
initiation site and the progressive development of the crack front, culminating in an area of final overload fracture. Figure 2.9A shows a typical single
initiation fatigue fracture, in this instance a percussion tool used for road
breaking. This particular tool was a tarmac cutter having a wide, chiselshaped blade that had been allowed to become blunt (Figure 2.9B). As quite
usual for such a tool, the shank was subjected to a levering action that applied
bending cycles just above where the blade flared out, predominantly in the
Materials in Distress
45
Figure 2.10 (A) Double bending fatigue failure of a scaffold spanner handle.
(B) Double bending fatigue failure of a rotating shaft.
same direction. A tiny indentation near the shoulder had served as the initiation site for the fatigue crack. The tool had been hardened and tempered
to give a high hardness but in this condition the steel had fairly low ductility
and fracture toughness. Consequently, the crack did not have to penetrate
very far into the section before it suddenly ran to completion during one
particularly severe bending cycle.
A characteristic of all fatigue fractures is that the initiation site is almost
invariably at the surface where there exists some change in the section (like
a slot for a keyway or circlip, or the root of a screw thread) that acts as a
localized stress concentrator, although very exceptionally it may start just
below the surface at the site of some internal flaw or inclusion. Usually the
initiation site is easily recognized with the naked eye or with the aid of an
eyeglass. From this site the crack front spreads into the underlying section,
forming a series of concentric wave fronts resembling the ripples on a still
pool disturbed by a pebble. These are actually positions where the developing
crack arrested while it awaited another load cycle sufficiently high to drive
it further. In fact, the surface marks are commonly referred to as beach
markings because they resemble the striation patterns left in the sand when
an incoming wave pauses before retreating. With an incoming tide the next
beach mark is a little further up the beach, although broadly of the same
shape.
Figure 2.10 shows a double bending fatigue fracture that initiated at two
sites diametrically opposite at the root of a slot. This was a reversible box
46
Figure 2.11 Multisite initiation fatigue fractures. (A) Crop sprayer bearing arm
that failed in reverse bending. (B) Rotavating machine drive shaft that failed under
cyclic torsional forces.
spanner used for tightening and releasing clamp bolts on scaffolding, where
the box end could be fitted over the nut and the hand bar swung side to side
through 180 when tightening or releasing, to avoid lifting the box off the
bolt head and repositioning it for the next half turn. Fatigue cracking initiated
on opposite sides of the groove cut-away in the handle. Cracking had reached
the stage where only the narrow band of metal at right angles to the pivot
was still holding the handle to the box section. This suddenly broke in a
ductile overload manner as a workman was striving to loosen a tight nut
while dismantling the scaffold, causing him to fall to his death. (This accident
happened in the days before the present safety at work legislation was introduced.)
A fatigue fracture may exhibit several initiation sites, all of which begin
to grow into the cross section under the externally applied loadings until
eventually there is insufficient cross section remaining to withstand one final
cycle. Thus a rotating shaft failed suddenly by propagation of the top crack
in Figure 2.10B, which was initiated at a small defect near the arrow. There
are two initiation sites, however, showing that the shaft was subject to bending
along the axis defined by the arrows. Figure 2.11A is an example of a multiinitiation site fatigue failure of the pivot bearing arm for a crop sprayer. This
predominantly follows the path of a double bending stress system, but the
Materials in Distress
47
Figure 2.12 Fatigue crack surface in an NBR lip seal of an automotive brake
hydraulic system.
sprayer was tilted at different angles, all subjected to up and down forces as
the tractor bounced over uneven ground. Hence one crack might initiate at
one side but before it had grown very far another initiated further along on
the same side, followed by another at a different place, and the same thing
was happening on the opposite face. The net result is a fracture such as
illustrated here, with several distinct initiation sites, all eventually linking up
before running together in the final ductile overload fracture along the centerline of the cross section. In the drive shaft (Figure 2.11B), separate fractures
initiated at the roots of individual splines and began to propagate in helical
paths, but then linked up as they entered the main shaft. The final failure
was ductile overstress of the last remaining area at the central axis.
Fatigue is a disease that affects all materials, but without a doubt the
majority of such failures, certainly those that are publicized, are metals, often
steels, simply because these are the most widely used structural materials.
Recent rail disasters in the U.K. (Hatfield) and Germany (Eschede) were
caused directly by fatigue of steel rail and a wheel, respectively. The single
worst aircraft disaster ever (JA8119 on a Boeing 747) in Japan in 1985 was
caused by fatigue of a rear pressure bulkhead from a faulty repair. One of
the worst crashes in the U.S. was the Chicago crash of 1979, when an engine
fell off a plane in flight due to a fatigue crack in the support pylon. Composite
materials have a much better record in safety-critical components, and have
displaced many metal parts such as helicopter rotor blades and propellors.
Elastomer seals can fail by fatigue just as any other material, such as the lip
seal from an automotive hydraulic brake system shown in Figure 2.12. The
sudden failure of the brakes caused a road accident, which is one very good
48
Figure 2.13 Classical fatigue failure of an automobile engine crankshaft, initiating at the shoulder of the journal and propagating through the web.
reason why small leaks on brake systems should always be investigated and
the safety-critical parts replaced.
If single incident mechanical overstress failures are discounted then
something on the order of 75% of all service failures of engineered products
are attributable to fatigue, a figure certainly borne out by engineers involved
in accident investigation. The insidious nature of fatigue is that the crack
does not appear in the initial stages although damage is occurring on an
atomic scale at the site where it does eventually appear. Only cycles of stress
above a particular threshold value will actually cause this damage and propagate the crack. But all these stress cycles are in the elastic regime, so there
is no external sign of the distress in the form of deformation. Plastic deformation only appears during the final fracture and even then only if the
material is ductile. In high strength/low ductility materials such as the heattreated steel in Figure 2.9 there is often no external sign of deformation,
although examination of the last area to fracture may well exhibit a ductile
mode; just as often, it may be brittle. After the fatigue crack has initiated and
while it is still propagating into the cross section, it may be detected using
one or more of several different nondestructive inspection techniques, like
dye penetrant, magnetic particle inspection and ultrasonics (which are
described in Chapter 4), but the difficulty is that many initiate in positions
where they are inaccessible to routine inspection, for example, the crankshaft
deep inside an internal combustion engine as illustrated in Figure 2.13. This
particular failure initiated where the original radius of the bearing journal
had been partially removed when the crankshaft was reground, leaving a step
that served as the local stress raiser.
Extensive use of fatigue data is made when load bearing components are
designed, particularly if they are subject to cyclic loadings. The reader is
advised to refer to the many specialist books on this subject, as a detailed
discussion is beyond the scope of this text. After all, the forensic engineers
Materials in Distress
49
500
Stress Amplitude/MPa
Alloy steel
Low carbon steel
Aluminium alloy
A
B
C
400
300
200
100
104
105
106
107
108
Figure 2.14
S-N for two steels and heat-treated aluminum alloy. Notice the
leveling off at 40% of tensile strength for the steels.
task is usually to identify the mode and cause of a particular failure and this,
more often than not, is because a product has been exposed to unusual
conditions or was abused in some way that the designer never intended or
foresaw. However, it is useful to refer to the form of S-N curve, as depicted
in Figure 2.14. These diagrams summarize data from a series of tests in which
standard specimens are subjected to cyclic loads of a particular stress amplitude, S, until fracture occurs. The number of cycles to failure, N, is represented on a logarithmic scale. Tests are carried out over a range of stress
levels, all below the elastic limit of the material. S-N curves supply useful
data for design, provided that the frequency and maximum amplitude of
cyclic stresses likely to be encountered in service can be estimated. There are
various ways of taking into account the variations in stress amplitude and
frequency likely to be experienced by a specific component in a particular
application.
For steels the S-N curves level off at approximately 40% of their tensile
strength, and this is defined as the fatigue limit. In principle steels should
survive indefinitely if a component is used in situations where the maximum
level of cyclic stress is unlikely to exceed this fatigue limit, but there are
complications in relying on this, not the least of which is because the level
of stress at all positions within a shaped component under different applied
loads is difficult to predict, even using finite element analysis methods. However, for nonferrous metals and most other solids used in engineering there
50
Stress Uniform
throughout Section
Stress Concentration at
Root of Notch
Materials in Distress
51
Figure 2.16 Fatigue initiating at the corners of a keyway slot in an 80-mmdiameter drive shaft and following the root of the screw thread.
and the nominal level was below the fatigue limit, then clearly the introduction of a stress raiser could raise them locally to well over the limit and a
fatigue failure is to be expected sooner or later. Figure 2.16 is a typical example
of a fatigue fracture initiating at the corner of a keyway in a circular shaft
and subsequently following the stress concentration around the root of a
screw thread. The minimal area of cross section at the moment of the final
break is a clear indication that in the absence of the keyway the uniform
stress level would have been well below the fatigue limit of the metal. The
local stress concentrating effect could have been reduced considerably if the
bottom of the keyway had been radiused instead of being machined to a
sharp 90 corner.
The above example should make it abundantly clear that in conducting
an examination to ascertain the cause of a failure it is vitally important to
identify the initiation site(s) of a fatigue fracture and hence ascertain the
nature, magnitude and direction of the forces that caused it. From there
onward it is a fairly straightforward matter to establish what or who was
responsible.
The fatigue path follows a route that is at 90 to the maximum principal
tensile stress within the component, whereas ductile failures are at 45 to this,
following the direction of maximum shearing stress. For example, a torsional
fatigue fracture of a shaft follows a helical path (Figure 2.17B). In this instance
the torsional forces were caused by the shaft being driven in both directions
so fatigue initiated in two places and both cracks grew independently toward
each other following the characteristic helical paths until they met.
52
Figure 2.17 (A) Overstress torsional failure of a 30-mm steel shaft. (B) Reverse
bending torsional fatigue failure of a 20-mm-diameter steel drive shaft.
2.5 Corrosion
Metals and alloys corrode because the electrons associated with each atom
are not tied up in chemical bonding to other atoms but, instead, exist like
a cloud or gas surrounding the regular array of atoms in the crystal. If
placed in a chemical solution, the outermost metal atoms can lose electrons
and become positively charged ions, in which form they are able to enter
Materials in Distress
53
solutions. In this state they are able to take part in chemical reactions as
positively charged ions (written Fe2+ or Fe3+ for iron, depending on whether
two or three electrons have been given up) and the freed electrons are able
to form an electric circuit. This process is called corrosion, and results in
a wasting away of the surface of the metal and the formation of a chemical
corrosion product. The consequential loss of cross-sectional area thus
reduces the capacity of the component to support load. The propensity to
corrode depends on the potential of the atoms at the surface to enter
solution is called the electrode potential of the particular metal. The rate
of corrosion is determined by the current that flows through whatever
electrical circuit is established. Table 2.1 lists the electrode potentials of the
common metals.
Table 2.1 Standard Electrode Potentials of
Common Metals in Conducting Solutions of
Their Ions in Relation to a Hydrogen Electrode
Metal
Gold
1.42
Silver
0.80
Mercury
0.79
Copper
0.34
Lead
-0.13
Tin
-0.14
Nickel
-0.25
Cobalt
-0.27
Cadmium
-0.40
Iron
-0.44
Chromium
-0.70
Zinc
-0.76
Titanium
-1.63
Aluminum
-1.66
Magnesium
-2.37
54
If any two of these metals are in contact in a solution that can accept
ions (a solution that need be nothing more than slightly acidic moisture),
they may set up an electrical circuit in which the driving voltage is the
difference between their respective electrode potentials. The metal having the
more negative potential will give up electrons and corrode, while the less
negative or positive metal does not. The corroding metal becomes the anode
and the receiving metal becomes the cathode in the corrosion cell. Thus if
aluminum is connected to iron in a corrosive medium their potential difference is (0.44 (1.66) = 1.22 V) and the aluminum, being the more electronegative, becomes the anode and corrodes at a rate proportional to the
electrical resistance of the circuit. If copper were in contact with the iron
then their potential difference would be 0.34 (0.44) = 0.78 V and the iron,
being the more electronegative, would be anodic to the copper and would
corrode.
This explanation is oversimplified in order to illustrate the general principles. A number of scientific factors have not been taken into account and
in practice real corrosion cells are more complex. For example, the relative
areas of the two metals forming the anode and the cathode are important.
A small aluminum rivet in a large sheet of steel (iron) would corrode very
rapidly because it constitutes a small anode in a large cathode, whereas the
rate of the corrosion of aluminum sheet if the rivet were steel would be less
because the anode current would be spread over a much greater area. The
insulating effect of films and corrosion product set up in a corrosion cell also
exerts a pronounced effect on the rate of attack.
The cause of failure resulting from such general wastage type of corrosion
is usually obvious, for example, where a steel member of a vehicle has rusted
through or an aluminum alloy structural member flaked away over a period
of several years due to exposure to road salt in winter weather. Occasionally,
however, the corrosion may not be visible from the outside and could remain
undetected up to the moment when sudden catastrophic failure takes place
in an area inaccessible to painting or protection such as between two closely
fitting plates or inside a boxed structure (Figure 2.18).
Two distinct types of corrosion are recognized, namely, wet corrosion
and dry corrosion. Both result in the wasting away of a component, reducing
the thickness of section and thereby the load carrying capacity of the affected
structure until failure takes place by mechanical overstress or fatigue. Dry
corrosion is essentially high temperature oxidation and is controlled by the
rate at which oxygen can diffuse through the existing oxide skin and react
with the metal beneath to form additional oxide. This may be an adhering
tenacious film that largely prevents further oxygen from reaching the metal
or it may be porous and nonadherent, allowing a continuing reaction progressively eating away the cross section. Dry corrosion is seldom a problem
Materials in Distress
55
Figure 2.18 Galvanic corrosion and perforation of steel pipe caused by contact
with copper heating spiral inside wet lagging.
for structures and components operating in normal environmental conditions, although it clearly has serious consequences for parts operating at
elevated temperatures.
It is not generally realized that stainless steels rely on a thin surface film of
chromium oxide which forms virtually instantaneously in an oxidizing atmosphere. If the film is broken by scratching or abrasion it reforms and effectively
becomes self healing. However, under chemically reducing conditions stainless
steel is no better than plain unalloyed varieties. A family went away on vacation
and inadvertently left a pork chop on the draining board of their new stainless
steel sink unit. When they returned 6 weeks later and scraped off the offending
mess they discovered the shape of the chop outlined in pinholes through a
disintegrating draining board. Several other metals, for example, aluminum
and titanium, derive their corrosion resistance from a tightly adherent surface
oxide film that is very stable despite the underlying metal itself being chemically
reactive, particularly in the case of titanium. In contrast, if the surface oxide
film is porous or cracks up because it has a larger specific volume than the
metal from which it formed, like hydrated iron oxide (rust), then it will have
practically no protective effect whatsoever.
2.5.1 Stress Corrosion Cracking of Metals
For certain metallic alloys the effects of brittle fracture, stress and corrosion
combine to produce the troublesome phenomenon of stress corrosion cracking. This requires all of the following acting together:
1. A susceptible composition and microstructure.
56
Figure 2.19 Stress corrosion cracking in a deep drawn brass cup (left) beside a
cup from same material (right) but stress-relief annealed before exposure to the
same environment.
2. External or residual internal tensile stress from a manufacturing operation. Residual stresses exist as balanced regions of tension and
compression and cancel each other out across the body of the component, but the susceptible state is when the tensile component exists
at an outside surface.
3. A mildly aggressive chemical environment that on its own would
cause no significant corrosive attack.
If these are all present then crack-like defects that start as corrosion
crevices can grow rapidly and proliferate throughout the microstructure. In
brasses the phenomenon is known as season cracking because it was first
observed when cartridge cases showed spontaneous cracking in the monsoon
season as the British Empire spread across India. Deformed grains of singlephase brass provided the susceptible microstructure, residual stress from
crimping the deep drawn cartridge cases gave the residual stress pattern and,
because they were usually stored near stables, traces of ammonia from horse
urine in the humid atmosphere provided the mildly aggressive chemical
agent. Figure 2.19 shows stress corrosion cracking in a deep drawn brass cup
exposed to ammonium (NH4) ions, such as often present in proprietary
metal polishes. As the cracks appeared the residual hoop stress was relaxed,
so the strips of broken metal bent outward under the stimulus of the residual
component of longitudinal stress. The cup placed beside this was made in
exactly the same way and from the same sheet of metal, but before exposing
to the corrosive atmosphere was given a stress relief anneal. This did not
soften the metal or alter its grain structure in any way, but it did eliminate
the residual stresses; hence, no cracks. Figure 2.20 shows a silver-plated
goblet, one of an expensive set purchased overseas, which failed for the same
reason.
Materials in Distress
57
Most of the higher strength engineering metals may suffer stress corrosion failure if the conditions are right. Stainless steels and aluminum
alloys are susceptible in chloride (Cl) environments. One obvious remedy
for all is to eliminate or significantly reduce the level of residual stress
resulting from forming or welding processes. Hydrogen is also a troublesome element in this respect as it is believed it may form molecular hydrogen within the crack region and extend the crack by causing an internal
gas pressure, or the formation of solid hydrides within the microstructure
that can also subject the crack to additional mechanical stress. These
stresses may extend the cracks themselves or enhance the corrosion mechanism as before. Examples of both these types of failure are discussed in
later chapters.
58
2.6 Creep
Creep is a time-dependent phenomenon in which a component gradually
deforms under a steady tensile load that is well below the elastic limit of the
material. It is due to the atoms or molecules within the microstructure
continually adjusting to accommodate the externally applied forces. In this
process damage to the crystal structure may occur, internal voids appear and
cracks may spread which eventually culminates in a creep rupture failure.
This may not take place until the component has been in service for several
hundreds or thousands of hours. In metals and alloys the temperature at
which creep becomes significant is approximately 30% of their melting temperature in Kelvin and strain occurs at an increasingly faster rate up to 60%,
at which temperature the metal softens and begins to recrystallize. Above
60% of the Kelvin melting temperature the component will rupture in a very
Materials in Distress
59
62
63
Figure 3.1 Stepstool ladder placed in the doorway where an accident occurred.
have been caused by the forces generated if these safety instructions were
being properly observed.
Figure 3.1 shows a stepstool ladder alleged to have failed within an hour
or so of purchase while being used to paint the gutters of a house. The user
was a tall, heavy man but within the weight specified on the labeling of the
steps for light domestic use. The foot of the left stile is bent inward and the
steps can no longer be placed on the floor without the front frame tilting so
far that anyone attempting to climb them would topple sideways. The stile
is a rectangular box section aluminum alloy that has buckled inward opposite
the bottom tread, and the rivet holes at that side are torn, as shown in
Figure 3.2. The plaintiff alleges the steps were of defective construction
and/or that the box section was not strong enough or that there was a fault
in the metal.
As in all designs of step stool ladders the front and back legs splay outward
so that, whether viewed from the front, back or the sides, when they are
opened out and placed on a firm, level floor the four feet are farther apart
than any of the steps or side braces. Thus any load acting downward from
someone mounting or standing on the upper treads generates compressive
forces in the stiles that are reacted to by the floor and hence set up force
components acting to spread the feet outward. How then could the bottom
64
Figure 3.2 Closer view from the underside of the collapsed stile showing
buckle.
of the left stile have been bent inward? Yet the deformation opposite the
bottom tread is on one side only and has been caused by an inward acting
horizontal force acting with the tread serving as a fulcrum. Loosening of the
rivets and slight tearing of the rivet holes were consistent with this action
and no sign of any similar distress could be found anywhere else in the
structure. Moreover, the rivets had left witness marks showing that they had
been properly set when the front frame was assembled. There is no evidence
of faulty construction or material weakness and, even if there were, the force
that caused this failure was directed from the outside of the left foot and was
reacted to by the bottom tread but not producing any damage whatsoever
to the tread itself or the right-hand stile.
The most likely way a force system of sufficient magnitude could have
been set up is if the steps had been lying sideways with the right stile flat on
a firm surface while a downward force was directed against the bottom of
the left, as depicted in Figure 3.3. This would bend the left foot downward
about the fulcrum of the bottom tread and in doing so would stretch the
rivets and produce the buckle. How could such a situation arise? Simply
by the user overreaching sideways while near the top of the steps, causing
them to topple sideways onto their right side. On the way down some part
65
Figure 3.3 Position of the steps when the leg was bent, after they had become
unstable.
of their anatomy struck the projecting foot of the left stile as the right stile
lay nearly flat on the ground. The full weight of their falling body was applied
to the foot of the left stile, producing exactly the form of damage shown in
Figure 3.1.
Sometimes the falling body may land on the back frame and bend the
leg in a similar manner and the back cross brace may be buckled. In one
particular accident where the users body landed directly above the cross
brace the back leg was pushed down over the brace and penetrated the users
buttock, causing severe injury. He was a large man weighing some 130 kg
(280 lb), and stated he had stood tiptoe on the very top of a set of steps to
remove a suitcase from his attic and, as he pulled it out, the steps suddenly
collapsed.
Such damage as described following an accident is invariably the result
of some kind of instability while the person injured was climbing or working
from the steps near the top. Sometimes the instability is caused by overreaching, possibly to lift something heavy off a shelf, and other times by the steps
not being opened fully and locked in position to prevent them folding forward, or if the feet were not set on a firm, level surface. With tall ladders it
is essential to have a second person standing on the bottom rung (footing
the ladder) so that no rocking movement is possible while the user is working
near the top. In fact, the stiles of ladders are usually so strong that the
maximum recommended static weight of a user acting at any angle will not
bend a stile. Dynamic forces are necessary that are generated only when the
66
67
Figure 3.4 Fracture at the top of the steering tube of a motorcycle. Sections
have been cut away for metallurgical examination.
tube had sheared off near the top, just below the upper taper roller bearing.
He commented that if this tube had failed as the vehicle was entering the
bend the rider would have been unable to steer the machine, although all
cable controls (in particular to the front disc brake) would be unaffected.
The engineering expert instructed by the young mans lawyer concentrated his attention on this fracture. The engineer observed that the top of
the steering tube had both an internal thread and an external thread and that
the fracture had initiated where their roots coincided, i.e., across the thinnest
cross section of the steering tube (Figure 3.4). He sent the broken pieces to
a metallurgical laboratory, which cut sections and reported it was a virtually
instantaneous ductile tearing fracture displaying no evidence of metal fatigue.
In addition, the laboratory carried out hardness tests and a microstructural
examination (hence the pieces cut away in Figure 3.4) and found these to be
typical of the quality of steel used for such steering tubes. The laboratory
concluded that the tearing fracture had started at the front of the tube where
the roots of the internal and external thread coincided, i.e., where the wall
cross section was least.
The engineer then sent the remains of the front wheel (Figure 3.5) that
had shattered into several pieces to another laboratory that reported it was
a magnesium alloy casting to a standard specification. This second laboratory
found a small casting defect in one of the fractures but concluded this defect
would have exerted no significant effect on the strength of the wheel and had
been exposed as the wheel fractured.
In the light of these findings the engineer concluded that the cause of
the accident was loss of steering control due to sudden failure of the
steering tube as the front wheel went over a pot hole in the ridged road
surface. He advised the injured mans lawyer that they had grounds to
pursue a claim against the motorcycle manufacturers on the basis of
68
69
Figure 3.6 Bead of a front tire showing the kink where the steel bead has been
deformed by the tire being impacted against the wheel rim.
of abrasion or wear. Assembling the pieces revealed that the fractures spread
out from a single point on the left rim, consistent with a forceful impact
concentrated at one spot. The impact had been cushioned by the tire, as there
was no deep indentation in the metal of the rim itself. Figure 3.5 shows the
major pieces juxtaposed, and Figure 3.6 shows the inside of the tire with a
pronounced kink in the bead consistent with being flattened against the outer
flange of the rim.
It was a tubeless tire and no evidence of puncture could be found, so the
air must have either escaped through cracks in the metal wheel or by displacement of the tire bead from the wheel rim by an external force. The
bundle of high tensile steel wires inside the rubber forming the bead (which
holds the tire onto the wheel rim) had been deformed to the profile of the
outer rim, as evident in Figure 3.6. This forceful impact had clearly occurred
when the wheel hit the tree, as there were several small splinters of bark
embedded in the fracture of the major part of the wheel still attached to the
hub, which could only have occurred after the tire had been displaced from
the disintegrating wheel.
This big force against the front tire had been reacted to by the magnesium
alloy wheel and must, in turn, have been reacted to by the axle in order for
70
Figure 3.8 Brake disc, bent and bolts missing where trapped in the calliper.
the wheel to break up in this manner. Figure 3.7 shows the axle placed on 1mm2 paper, clearly revealing that it is bent toward one end. A steel axle does
not deform in this way unless one side is restrained while the other is acted
upon by a force sufficient to deform it. This necessary displacement is
revealed by the way the fork stanchions have been twisted out of line, as
illustrated in Figure 3.8. The same force had twisted the brake disc, breaking
two bolts at one side, while it had been restrained by the calliper (Figure 3.9).
Thus the big force against the tire must have been transmitted through the
wheel into the axle and then to the forks, twisting them sideways, as illustrated
in Figure 3.10. This force must have then been reacted to by the head frame
tube of the chassis. All external forces on the forks are transmitted to the
headframe of the machine via the bearings at the top and bottom of the
steering tube.
The steering tube exhibited a ductile shear failure just below the top
bearing, as already shown in Figure 3.4, and had bent immediately above the
bottom bearing as a result of the bottom of the fork being pushed backward,
as illustrated in Figure 3.9. (Surprisingly, the first engineer had made no
mention of this in his report and had not taken any photographs of this
region. From the dimensions of the tube in the vicinity of the bend and
71
Figure 3.9 Front forks, twisted out of line. Notice the position of the brake
calliper and the bottom of the steering tube bent as the forks were pushed
rearward.
estimating the tensile strength of the steel from its hardness value, it is
possible to make a reasonable estimate of the bending force, but this was
unnecessary in this instance.)
No sign of damage or deformation could be found beyond the bearings
in the frame of the machine. Hence the steering tube is where the energy of
external impact had been finally expended and is where the chain of damage
terminates. Although the force that broke the steering tube had been reacted
to by the headframe of the machine and transferred into the chassis, it must
clearly have been below the level necessary to cause any plastic deformation
of the metal structure. Sometimes a bearing may show small indentations
on the race where rollers or balls have been forced in during collision impact,
but in this instance there was no sign of this effect (often referred to as
brinelling after the hardness testing method that forces a hardened steel
ball into the test specimen to produce a spherical indentation).
The above analysis demonstrates that the steering tube had been fractured by collision forces generated when the front wheel hit the tree. A
shearing fracture had occurred across the threads where the roots of the inner
72
Figure 3.10 Closer view of the lower section of a steering tube, sawn to remove
the mating fracture of the threaded portion still remaining in the upper bearing.
Notice the absence of bending deformation on the upper portion of the steering
tube still remaining in the top bearing.
and outer threads coincided, which was the minimum cross-sectional area
of the entire length of tube. As this section fractured in shear the bottom of
the forks continued to be pushed backward, which consequently bent the
bottom of the steering tube forward just above the lower bearing, as evident
in Figure 3.10. Hence the steering tube had not failed spontaneously and
caused the rider to lose steering control while the machine was traveling
along the road. A steering tube is never subjected to such forces as would
cause it to fail in this manner during normal riding, even on the most bumpy
of road surfaces. If the steering tube had contained any significant manufacturing fault at this position, the fracture would have exhibited a progressive
fatigue mode, not ductile tearing. The defect would have been exposed by
the fracture. Furthermore, if this section had failed before the machine hit
the tree the steering tube could not have reacted to forces of sufficient magnitude to disintegrate the front wheel, bend the axle and cause so much
damage to the forks and brake disc.
Not surprisingly, on the strength of this second opinion, the litigation
was abandoned.
73
kg
5 1000 1
4
2
3
Steel Chain
two limbs on each link
(total area 100 mm2)
20000
Stress =
= 200 N mm2
100
Wood Blocks
50 50 = 2500 mm2
20000
Stress =
= 8 N mm2
2500
Superstixit Adhesive
same area as blocks,
same stress
Steel Chain
two limbs on each link
(total area 100 mm2)
Stress =
20000
= 200 N mm2
100
20000
= 25 N mm2
2500
74
crane weigher. The weigher registers 2 t. The intention is to show how good
the glue is. But let us estimate the stresses in the various links comprising
this load transfer path, first to discover which is likely to fail first if the load
were increased up to the point where something would break and then to
investigate differences if a defect were present in one or another of the
component members of the linked assembly. For the purpose of this exercise
we shall ignore the weight of the straps, the wooden blocks and the chains,
as these would be relatively insignificant compared with the 2 t being lifted.
We take the order of mechanical strength of each of the three components
to be:
Stress (N mm2)
under 2000-kg
Load
Mechanical Strength of
Material (N mm2)
Nylon straps
Four, each 100 mm 2 mm
Total cross section = 800 mm2
25
Tensile failure 75
Steel chain
Two limbs each 50-mm2 area
Total = 100 mm2
200
Yields at 400
Tensile fracture 650
Fails at 15
Component
While the 2000-kg mass is off the ground it applies a force of 20 kN (near
enough) acting on all components in the load transfer path (including those
inside the weigher) which is reacted to by the crane or whatever is supporting
the weighing machine. On the 2500-mm2 area of adhesive this generates a
stress of 8 N mm2. The total cross-sectional area of the four 100-mm 2mm nylon straps is 800 mm2, so the mean stress when supporting the load
of 20 kN is 25 N mm2. Every chain link comprises two limbs totaling 100
mm2 cross-sectional area, so the mean stress in each limb is 200 N mm 2.
Joints made with this adhesive are known to fail at stresses in the region of
15 N mm2, so the joint is, in fact, by far the weakest link in this load transfer
system. It does not fail (in the short time, at least) because the stress acting
is only just over one half this level. Its weakness relative to the other members
in the system is not immediately apparent due to the much greater crosssectional area of the wooden blocks. However, if the applied load were
increased this is where the failure should occur because the chain links and
the nylon straps are working well within their limiting stresses and neither
would be expected to show any kind of mechanical distress, although they
would stretch elastically under the stress. Elastic deformation is recovered
when the force is removed.
75
Which member would fail first if the load were increased to, say, 4 t? This
would double the stresses given in the table, from which it is quite obvious
that the adhesive would have failed before the stress reached 16 N mm 2.
However, at a 4-t load the stress in the chain links would be 400 N mm 2,
which is just at its yield point, so some of the links might have begun to
elongate. The nylon straps would have reached only two thirds of their
breaking load.
Now let us remove the wooden blocks and, using only the chain and the
nylon straps, gradually increase the load until something broke. Which would
it be? When the load reached 6 t the nylon straps would fail as the stress in
them would be 75 N mm2. However, the chain would also have stretched
considerably because the stress of 600 N mm2 is well above the yield point
of the metal but still below its tensile strength. So we would expect to see an
appreciably elongated chain attached to the broken straps.
Now let us suppose that the weld in one of the chain links was defective.
By how much would its load-bearing cross section have to be reduced for
the link to fail before the adhesive joint? Clearly this must be a load below
that to break the adhesive joint, i.e., 15 N mm 2 2,500 mm2 cross-sectional
area = 37,500 N. (In reality this would be a far more complex problem
than it seems, because the fracture toughness of the steel and the form of
the defect would need to be taken into account to make an accurate calculation.) Nevertheless, if we take the simplistic view that the cross-sectional area of the chain link must be so reduced that its remaining area (A)
multiplied by the nominal tensile strength of the steel (650 N mm 2) has
to be equal to or less than the breaking load of the adhesive joint (37,500
N), then A works out at 57.7 mm2. As each limb of the chain should have
been a 50-mm2 cross section, then the one with the defective weld must
have been effectively reduced to only 7.5 mm 2. Hence, on inspection we
should expect to find a sizeable defect occupying something like 85% of
the original cross-sectional area of that limb. Even though this is not a
precise figure it shows we should be looking for a sizeable pre-existing
defect in the fracture surface of the broken limb.
There would also be another feature to look for. Because steel is a ductile
material and yields before it fractures, the defective limb would break under
the applied load but the opposite limb would deform plastically, open up
and allow the connecting link on the loaded side to escape. None of the other
links in the chain would display any sign of distress because under a load of
37,500 N the mean stress in the limbs of nondefective links would be only
375 N mm2, which is below the steels yield stress of 400 N mm 2.
Finally, to bring this semiquantitative consideration of load transfer to a
conclusion, if such a chain were used on its own, for towing a broken down
vehicle, for example, and if the entire length of chain were overloaded beyond
76
the yield stress of the steel, then every link would deform before the weakest
one broke. The entire length of chain between the attachment points would
be found to have stretched when the pitches of the individual links were
measured. However, if there was no general stretching of the chain, this would
indicate a failure caused by some localized condition leading to overstress,
such as a defect in that particular link or some unusual mechanical condition
that led to overstress, for example, a kink, while the rest of the chain was not
loaded to its yield stress.
Practicing engineers may question the validity of assumptions and
approximations made in estimating forces as outlined in this section. Of
course it would be better to carry out laboratory tests and make simulations
to obtain accurate values, but the problem is that damaged components
produced are one-offs from a particular accident and the evidence must be
preserved with minimum disturbance for examination by other experts representing various parties who may be joined in litigation. Quite often there
is insufficient material to remove sections for standard test pieces without
destroying what some expert may later assert was vital evidence. The usual
material and mechanical specification for a part may be available but the
reason for a particular one failing may be that it did not meet that specification at the time of manufacture or that its properties may have been altered
by something that was done to it in service, like inadvertent heat treatment
from a nearby welding operation softening or cracking a quenched and
tempered steel. Hence the forensic investigator must rely on nondestructive
(or nondamaging) tests and observations to ascertain whether the material
was ductile or brittle or had deteriorated in some way in service. With metals
a Vickers diamond pyramid hardness test, which leaves only a tiny indentation on the surface, is very useful for estimating the yield and tensile strength
of a broken component in the vicinity of a fracture and, with a small item
like a rivet or fastener, is the only means of doing so.
77
Figure 3.12 View from the rear of the racing motorcycle sidecar as recovered.
The sidecar wheel bracket is broken and completely separated from the chassis.
78
was broken, though it was said this was badly cracked but still hanging
together when the machine was lifted off the track. The tire bead had separated all around the outer rim and the inner tube had a 3-in.-long split near
the valve. The sidecar chassis frame was unaffected apart from slight bruises
where the clamps holding the bracket had been attached, but both front and
rear clamps had opened out as a result of the bracket being pushed down
onto the chassis, and two of the four bolts on each had been sheared off at
their brazes. After considering the load transfer path, this experts conclusion
was that all the damage had been caused by the sidecar wheel violently
impacting the ground as the machine rolled over toward the end of the
somersault, after it had passed over the prostrate body of the rider but before
it came to rest further along the racetrack. A video recording had been made
of the incident by an amateur cameraman who had positioned himself with
a good view of the machines at the first curve. This experts conclusion was
entirely based on visual examination and nothing was disturbed or dismantled pending expert examination commissioned for the claimant, who had
instituted proceedings against the designers and the manufacturer of the
sidecar chassis and the eccentric hub.
The plaintiff s engineer agreed all damage occurred as described above,
but he reached an entirely different conclusion. He observed that all the braze
fractures had occurred at the interface between the steel and the braze alloy
and that nearly 30% of the fracture area of the brazed joint to the eccentric
hub was more heavily stained than the rest. He regarded this as evidence that
this part of the braze had been cracked for some time before the accident.
He also found different degrees of staining on the fractured interfaces of the
smaller brazes on the clamp bodies (Figure 3.13) and interpreted these stains
as signs of stress corrosion. There were several other features that we need
not discuss here, but which he regarded as further evidence that the brazed
joints had been progressively weakening for some time before the accident.
His conclusion therefore was that failure of these weakened joints had
occurred as the machine was rounding the curve and, as the eccentric hub
cracked, its movement caused the tire to rub against the chassis and exert a
strong braking effect. It was this, coupled with misalignment of the sidecar
wheel, that caused the machine to go tail light and begin to somersault.
This engineers detailed allegations may be broadly summarized as follows:
1. The design of the eccentric hub and the clamps was defective in
relying on butt joints in high tensile steel tubes to be hand made by
bronze welding.
2. The brazing had been carried out defectively. The large braze to the
eccentric hub had not wetted properly (based on the area of discoloration on the exposed fracture surfaces) and this had caused fatigue
79
Figure 3.13 Close view of the clamp to the sidecar chassis showing the sheared
lug. Notice the left end of the clamp has lifted off the chassis tube under the
same force system that sheared the lug weld.
80
Figure 3.14 Sidecar wheel offered up to the broken bracket. Notice the gap
(arrowed) suggesting the axle is bent downward.
Figure 3.15 Sidecar wheel turned around and placed under the chassis. Notice
the broken tie rod and how the rear leg of the bracket has broken at the hub
joint. Notice also that the bottom of the rear clamp has been broken away,
allowing the bracket to separate from the chassis tube.
the axle was carefully marked and then removed from the hub. It was found
to be bent downward by some 15. The hardness of the steel was measured
in order to estimate its tensile strength and confirm that it was in the specified
hardened and tempered condition. The defendants expert later applied engineering standard bending formulae to calculate the external force that would
have been required to produce this degree of bend. The figure obtained was
8.5 t force applied in the plane of the wheel at the midpoint of the wheel
bearings. The axle had bent downward so this force must have acted from
the top of the wheel or possibly, but much less likely, by the bottom of the
81
Figure 3.16 Unused bracket tested to destruction in the compression test. The
fracture is virtually identical to the accident bracket.
wheel being pushed inward. The defendants expert then obtained an unused
bracket from the manufacturer, fitted with an axle, and tested this in a
compression test machine. The clamps began to deform at 7.4 t, and the
brazed joint could be heard to crack at 8.2-t force and finally split open as
shown in Figure 3.16 at 9.1-t force, with the break virtually identical to that
of the accident bracket. There was no sign of any defect in the fracture surfaces
of the bronze weld. The cracking followed the interface between the bronze
and the steel.
Suspecting that the discoloration of the fracture surfaces of the plaintiff s
bracket was postaccident and most probably resulted from exposure to moist
atmospheres during the 8-month storage period between the accident and
the experts examinations, part of this laboratory-produced fracture was
masked off and the hub was left in an outdoor store. The exposed area of
fracture quickly started to become discolored. This was because the fracture
followed a brittle intermetallic layer formed at the interface of the bronze
and the steel, which corrodes rapidly in moist air. The masked region was
hardly affected after 2 weeks but 3 months after the masking was removed
it was indistinguishable from the rest of the fracture interface. Hence, as
neither expert had seen the hub fracture until several months after the accident and the machine had been stored in a farm barn for this period, there
was no way of deducing what the fracture would have looked like immediately
after the accident. The topography of the fracture displayed none of the
characteristics of metal fatigue, so the defendants expert continued to believe
it was an instantaneous fracture caused when the wheel hit the ground.
The plaintiff s expert had obtained a replacement clamp and put it
through tensile tests to determine the force required to break the allegedly
unsuitable bronze welded joints on the lug. Instead of the interface failing as
82
he had predicted, the steel lug stretched, opened out and eventually fractured,
and the bolt used to hold it in the grips of the testing machine was badly
deformed. The bronze weld was unaffected, mainly because it was of much
greater cross-sectional area than the steel tubes that it joined. The plaintiff s
expert also carried out stress corrosion tests but was only able to produce
cracks by exposing the replacement clamp to concentrated ammonia fumes
with the joint under so much tension it had deformed. In all normal moist
atmospheres, including that developed in contact with animal dung, the
defendants expert determined that no cracking could be produced in the
braze.
Still refusing to accept the simple interpretation of the load transfer path
from the wheel through to the sidecar chassis, the plaintiff s expert commissioned a thorough FE (computer modeling of finite elements within a structure) strain analysis of the bracket, but, for some unaccountable reason, not
including the axle or the wheel. The analysts were unable to offer any figure
for the forces necessary to cause the kind of fracture observed because the
data necessary for bronze welded joints were unavailable.
After spending 7 years and a great deal of money conducting technical
investigations, the plaintiff s engineer had nevertheless not considered the
load transfer path associated with the deformation damage to all the components from the sidecar wheel to the chassis. For example, he devoted
considerable attention to the fractured lugs of the clamp but made no attempt
to explain why the clamps themselves were heavily deformed, as illustrated
in Figure 3.17 and Figure 3.18. Eventually, the two experts were ordered to
produce a joint report summarizing and commenting upon each others
evidence and their opinions as to the cause of the accident. As a consequence
of this, the plaintiff was advised by his legal representatives that he had no
chance of succeeding at trial, so the claim was abandoned.
The lawyers accepted the defendant engineers explanation for the cause
of the accident, based on simple consideration of the load transfer path from
the wheel to the sidecar chassis. The crucial evidence was the bending of the
axle. Stub axles do not deform until they are subjected to loads that exceed
the yield stress of the steel from which they are made. Thus a force on the
order of 8.6 t must have been directed at the rim of the wheel but, more
importantly, this must have been reacted to by whatever was holding the
opposite end of the axle and, in turn, by all the components transferring this
force to the sidecar chassis. If any of these linkages had not been capable of
transferring forces of this magnitude they would have failed in some way
before the axle bent. If we start with the wheel, we see that the axle was
stressed above its yield point and bent downward. This force was reacted to
by the hub at the top of the bracket. The bracket was subjected to compression, which cracked the bronze weld at the hub exactly as reproduced in the
83
Figure 3.18 End view of rear clamp, illustrated in Figure 3.17, showing how
compression force has deformed semicircular shape.
laboratory test (Figure 3.16). This compression force was transferred to the
clamps holding the bracket to the chassis, which opened out as illustrated in
Figure 3.19 and confirmed in the photographs in Figure 3.13, Figure 3.17
and Figure 3.18. The damage terminated at the chassis frame, which had
suffered only slight bruising by the clamps. All the energy of the impact
against the wheel was dissipated in fracturing and deforming the other components.
If the plaintiff s expert had considered this possibility at the outset
before starting his detailed scientific tests, he would have realized that if
the weld at the hub had broken first, there would have been no path by
which the load to deform the clamps and bend the axle could have been
84
FIGURE 3.19 Upper portion of the rear clamp, as in Figure 3.18, but drawn to
show extensive deformation caused by compressive loads acting through brackets.
transmitted. Whenever and however the axle became bent the hub mounting must have been strong enough to react to and transmit the applied
force to the chassis.
The final capitulation by the plaintiff s expert was when analysis of the
video, frame by frame, revealed the wheel to be in its normal position and
alignment on the sidecar chassis during most of the somersault, after the
machine had struck the riders prostrate body. One of the last frames showed
the sidecar wheel violently striking the ground at the end of the somersaulting, after which the wheel appeared to be thrown outward almost horizontally, clearly establishing that this was when the damage was done.
A Toolbox for
Forensic Engineers
4.1 Introduction
A range of scientific instruments and techniques is necessary to solve the
variety of cases presented in the remaining chapters. Such procedures are
needed to obtain evidence about the products involved and the materials of
which they are made. In other words, tools must be selected that are
appropriate to report on the issues put to the investigator. The interpretation
of these data requires some expert knowledge. Furthermore, expert findings
must be presented in a form that lawyers can understand and incorporate
into their legal proceedings, as well as one that allows the courts to reach a
judgment.
The objective of any forensic engineering investigation is to determine
the cause or causes of accidents or failures. Forensic engineering is, therefore,
primarily a problem-solving exercise. With any analytical assignment, there
is a need to be aware of available tools, the information they provide, their
limitations and any trade-offs that may be inherent in a particular technique. However, there is no pressing need to understand any deep underlying
theory, just as when driving a car it is not necessary to know the way a car
engine works.
Many failure problems will lead to litigation, perhaps because the damage
is capable of several different interpretations. Thus the approach taken to any
material evidence, such as a fractured component, is fundamentally constrained by the need to preserve and conserve it for possible examination by
others. Items submitted for examination may be vital legal evidence that
cannot be chopped up or machined into test specimens. The exception is
where the item is very large, so selecting samples for test specimens will not
85
86
destroy anything critical. Forensic investigations are, therefore, fundamentally different from routine quality control, where thousands of components
will be available for examination. The different approach of forensics is biased
to nondestructive methods of inspection, simply so that the evidence is
preserved. That is not to say that destructive or partly destructive methods
cannot be used at all, but rather that they should be used only as the last
resort. Indeed, all the parties in a dispute will need to agree on a testing
program, at which all can be present when any element of destruction is
proposed.
The purpose of this chapter is to outline a typical range of techniques
and associated instruments that can be called upon by the forensic engineer.
These range from photography, crack detection, metallography, and highand low-power microscopy through to chemical and mechanical analysis.
It is essential to ensure that the right objects are selected for examination,
and the correct circumstances surrounding an event become familiar.
87
88
Hoist motor
Frame (in Roof)
Cable
Carrier Bar
Fracture
Hooks to Carry
Rolls of Foam
Chain
?
Fractured Surface
Figure 4.1 Graphic sketches of hoist, lifting bar, foam roll and fracture surface.
89
90
Figure 4.2 Undisturbed road dirt layer on a wheel rim (A), compared with
impressed markings transferred from the wheel hub to the wheel (B).
sense approach along with a refusal to jump to conclusions. This can mean
a long and often tedious examination of the evidence left by an accident.
Other debris may hide critical trace evidence, for example. Observations
made at the scene must be chronicled, usually with the aid of sketches, note
taking and photography of the debris. The more complex the debris field,
the greater the need for multiple images of the remains, so that the final
resting position of key components is recorded. Such activity is essential
before laboratory analysis.
91
can help glean information from a failed artifact. These range from a simple
eyeglass and macro-photography on the one hand, to hardness testing,
mechanical testing and microanalysis on the other.
4.4.1
92
Among the first on the scene of an accident would be the police photographer, or safety officer (or factory inspector) if the accident is an industrial
accident. When recovering evidence for further observation, preservation of
any dirt or grease marks on samples can be vital if critical evidence of its
past history is not to be lost. Such trace evidence (witness marks) can help
in reconstructing past movements of the sample, such as contact with other,
perhaps mating parts. The need to maintain fracture surfaces in their state
as found immediately after an accident cannot be emphasized enough. With
certain materials, touching such surfaces can initiate corrosion, so it is important to minimize handling. When being removed from the accident scene,
failed samples should be protected by wrapping carefully in bubble-wrap
film, and then sealed into labeled plastic bags. Blu-Tack or Plasticine can also
prove useful in protecting surfaces, as well as for mounting samples for
photography. It finds an additional application in removing grease and dirt
at a later stage in an inspection, after original pictures have been taken. By
using a dabbing action, Blu-Tack will lift and clean most dirt and loose
corrosion product from fracture surfaces.
4.4.2
Compliance with product standards and regulations is now a major issue for
many manufacturers. Unless products comply with specific standards, the
ability to sell can be compromised in many markets. The American Society
for Testing and Materials (ASTM) and British Standards Institution (BSI)
have produced standards covering safety, performance and reliability of most
products, including the influence of mechanical and environmental factors.
In addition, each standard is reviewed and updated periodically, thus ensuring continued relevance.
93
When considering any failure, an appropriate standard written to encompass the component or its application (e.g., wire ropes, chains, seatings, etc.)
may be applicable. Standards are an invaluable comparator against which
failed components can be matched. Furthermore, it is not just ASTM and
BSI that can be searched; equivalent standards exist in Europe (ISO), Germany (DIN), Japan (JIS), and elsewhere. All are excellent sources of reference.
Because the majority of accidents occur within the home, domestic appliances are particularly well covered by appropriate standards. In the U.K.,
there more than 50 deaths and some 12,000 accidents requiring hospital
treatment each year, involving consumers using ladders around the home
(Source: DTI, U.K., July 2001). A person falls and suffers injury, and damage
is found on the stiles or rung hooks, feet, etc. The user seeks to sue the
supplier, alleging manufacturing defect or flaw, poor construction, insufficient strength of alloy sections, etc. Such allegations are usually unfounded
as there are stringent standards (i.e., BSI, 2037:1994) covering the design,
materials, construction and performance for such ladders and steps. Analyses
of stepladder failures are presented in Chapter 3, and associated workplace
accidents are presented in Chapter 8.
Within any standard, much of the required background work will be
included and will be of particular value in product liability disputes. All that
may be required of the expert is a comparative analysis between the relevant
standard, and the product/component in dispute. It should become apparent
if a product did not conform to the required standard.
4.4.3
Macroscopic Examination
Macroscopic examination involves observation of a failed component, fracture surface or section under low magnification (10 to 50), preferably using
a stereomicroscope. For field observations, a simple eyeglass is all that will
be required. The macroscopic study will reveal information representative of
the entire piece.4 In addition, the amount of information that can be obtained
from examination of a fracture surface under low magnification is surprisingly extensive. Fracture characteristics such as direction of crack growth and
origin of failure can be determined, as in the fatigue failure shown in Figure
4.4. Furthermore, size and distribution of inclusions, uniformity of structure
and grain size, segregation and presence of fabricating defects will be readily
discernible. Macroscopic examination of fracture surfaces can also reveal
information relevant to failure mode (fatigue, shear, brittle or ductile), indentations and surface damage suffered in service (some of which may have
induced a fatigue failure) and environmentally enhanced failure (corrosion,
stress corrosion cracking, hydrogen embrittlement, etc.).
94
Figure 4.4 Single bending fatigue failure of the drive shaft from an agricultural
baling machine. Shaft diameter: 25.4 mm (1 in.).
4.4.4
Metallographic Sectioning
95
96
Figure 4.5 Microstructures of four different types of cast iron. Linear magnification: 75. (A) Flake graphite (gray) cast iron. (B) Nodular (spheroidal
graphite) cast iron. (C) Chill (white) cast iron where carbon is present as Fe3C.
(D) Malleable (blackheart) iron.
plane section, the lettuces look like separate flakes, but in three dimensions,
each group of flakes has grown from the same nucleus.
By making additions to the ladle of fractional percentages of either magnesium or cerium (a rare-earth metal), the graphite separates in the form of
spheroids, quite separate from each other, in matrices that again can be
anything from the equivalent of pure, soft iron to a high strength steel. These
are called nodular or spheroidal graphite (S.G.) cast irons. 4
Figure 4.5A shows a typical microsection of a gray iron casting and Figure
4.5B shows a typical microsection of a ferritic SG iron, both as polished and
photographed at a linear magnification of 75. By making a casting solidify
in the mold very quickly, the carbon does not appear as graphite but instead
separates as iron carbide (Fe3C), rendering the casting extremely hard and
brittle. Figure 4.5C shows such a structure, the section polished and etched
but again as a linear magnification of 75. If a casting with such a microstructure is broken it exposes a silvery white, sparkling fracture, which is why
these types of casting are called white cast irons. For certain applications
where high hardness and maximum abrasion resistance are required, for
example, the nose of a plowshare or the surface of a roll for cold working
metals, castings are often made with chills inserted in the molds so that the
metal that solidifies in contact with them develops the iron carbide microstructure.
97
Finally, there is a fourth group, the malleable cast irons, which have to
be cast as wholly white iron and subsequently heated for several hours at
temperatures where the iron carbide breaks down to form clusters of graphite.
Strength is determined by whether they are produced with a ferritic or a
pearlitic matrix after the heat treatment. Figure 4.5D shows such a structure,
where the graphite appears as the ragged black areas and the matrix is equivalent to a medium carbon steel. A casting with this type of microstructure
will be strong, tough and reasonably ductile, hence the name malleable cast
irons. However, the application of such malleable castings has been largely
ousted by the SG irons which are cast directly into the molds and require
virtually no subsequent treatment.
As an analytical technique, metallography is relatively simple to undertake; it consists of four steps:
Selection of section
Surface preparation
Etching
Examination under magnification
98
99
metal eye splinter will illustrate just how valuable metallography can be as a
tool. A workman in a machine shop was removing a large drill, with a Morse
taper shank, from a radial arm drilling machine. He said that he had used a
soft (steel) drift to remove the drill from the machine, striking the drift with
an ordinary engineers hammer. This method was a normal everyday (and
correct) procedure for drill removal. However, while undertaking the routine,
the workman was hit in the eye by a flying metal splinter. The eye splinter
was submitted to ascertain which of the tools being used was the likely source.
The components and their respective compositions are shown in Figure 4.8.
Because all four components in this case were made from different materials,
it was decided to examine their respective microstructures and compare them
with those of the splinter, in order to establish the source of the splinter. The
microstructures are shown in Figure 4.9.
Direct microstructural comparison clearly reveals the eye splinter is a small
piece of HSS, broken from one of the outer flutes of the drill, not the hardened
steel hammer head, the medium carbon steel drift or the soft iron Morse taper
shank. It was therefore a straightforward task to arrive at a sequence of events
that caused the metal splinter to pierce the workmans eye. The simple answer
was that the workman must have struck the drill flutes with the hammer when
extracting the drill from the machine, and NOT used the drift as claimed. A
trained workshop engineer would know full well that drill flutes are exceptionally hard, with little or no temper therefore prone to splinter or shatter
under impact loading and the hardened steel face of a hammer is perhaps
the worst tool he could use to knock it free. The resultant eye injury was
therefore a direct consequence of the workmans irresponsible mistake or poor
working practice, as well as lack of eye protection.
100
Figure 4.9
4.4.6
(SEM). The fracture surface shown is that of a steel fracture toughness specimen. The power of the SEM to differentiate between
different fracture mechanisms is clearly demonstrated.
Figure 4.10 Montage of fractographs illustrating the depth of field and magnification capabilities of scanning electron microscopy
101
102
Large depth of field, so that features of a rough surface are all in focus
Very high degree of resolution (down to 0.5 106 m)
Nondestructive for conducting materials
Elemental analysis using EDAX (energy dispersive x-ray analyzer)
4.4.6.1
103
Figure 4.11 Micrographs of birch (A), and pine (B) (magnification: 45).
and-run accident. A witness saw the car drive off but was unable to get the
vehicle license plate number. Nevertheless, the witness was able to provide a
detailed description of make, color, etc. of the suspect vehicle. A vehicle
matching the description was located sometime later, and a sliver of wood
was found behind one of its headlights. The vehicle owner recalled that he
had run into his own wooden fence, which had damaged the front wing. The
wooden sliver must have become trapped behind the headlight as a direct
result of his bump.
Comparative observation on a SEM revealed a different story. A sample
taken from the suspects fence proved the wood to be birch (Figure 4.11A
micrograph), whereas the sliver recovered from behind the headlight was
pine (Figure 4.11B micrograph). Samples from the post that had been struck
at the accident scene showed that it was also pine. Thus a simple comparison
of respective microstructures revealed that the car in question was the vehicle
involved in the hit-and-run incident.
Both the SEM and ESEM may be coupled to an energy dispersive x-ray
analyzer (EDAX) for quantitative, or semiquantitative analysis. EDAX analysis can be performed in a SEM for understanding the chemistry of materials.
When the electron beam in the SEM strikes the sample, it excites the atoms
of an element, producing x-rays characteristic of the element. The scan is
typically displayed in graphical format, with the X-axis showing photon
energy and the Y-axis indicating the number of times energy of that photon
keV has been generated. Chapter 5, Section 5.2 shows an elemental scan for
an aluminum alloy rivet that had been recovered from a failed container. The
EDAX scan showed that the rivet was of the correct composition. No sample
was destroyed, or altered microstructurally to make this analysis. It is not as
104
precise as a chemical analysis but quite adequate for checking material specifications.
The SEM will provide fractographic and microstructural information,
where the EDAX will provide accurate quantitative information on local variations in material composition. In addition to resolving chemical composition,
EDAX can be utilized to ascertain unusual microconstituents or contaminants
within the material, even down to a single inclusion particle. The advantages
of EDAX are that it is fast and easy, but sometimes requires followup work; for
example, it will reveal Fe and O, i.e., the elements, but will not readily distinguish Fe2O3 or FeO phases. In addition, EDAX can be utilized for quantitative
analysis of thick and thin films, semiconductors and some polymers. In difficult
cases involving chemical attack, the data provided by microanalysis can be
crucial to successfully completing the failure investigation.
An investigation that readily demonstrates the use of the SEM coupled
with the power of EDAX involved the theft of a large quantity of copper
alloy. A factory, manufacturing extruded copper alloy tubing for steam
condensers, suffered a break-in and a large quantity of metal was stolen.
The metal in question was a rather unusual brass of 76Cu-22Zn-2Al, and
the factory was the only one in that area using aluminum bearing brasses
in its manufacturing process. At that time, commercial brasses used for
steam condensers and general manufacture of tubing were 70Cu-30Zn and
70Cu-29Zn-1Sn.
A suspect who worked in a nearby iron foundry was apprehended in
connection with the robbery. He claimed that he had spent the evening in
question drinking with his four buddies, and had never been near the brass
tube factory. However, swarf was dug out of the suspects boot, and compared
with a piece of swarf from the floor around a large lathe used to skim billets
prior to extrusion. The SEM comparison is shown as a montage in
Figure 4.12, where the top half was swarf dug out of the suspects boot and
the bottom half was swarf recovered from the crime scene. As a lathe tool
wears, it develops a serrated cutting edge that will produce characteristic
score marks rather like the marks on a bullet fired down the barrel of a
particular rifle. It can be seen that there is a perfect wear pattern match
between the two pieces of swarf that can be considered comparable to a
fingerprint in conventional detective work. Elemental x-ray analysis on the
SEM confirmed that the sample removed from the suspects boot was an
aluminum bearing brass. Therefore the combination of wear fingerprint and
alloy composition was more than sufficient evidence to place the suspect at
the scene of the crime. It later transpired that the suspect and his drinking
buddies were perpetrators of the robbery.
Examples of ESEM are given in Sections 6.6 and 9.5.
105
Figure 4.12 SEM montage of two separate pieces of swarf. The top half was
recovered from the boot of a suspect and the bottom half was recovered from the
crime scene. Note the perfect tool wear pattern match between the two pieces.
Tensile Testing
106
Figure 4.13
Specimen geometry for tensile test bars, taken from BS18: 1987.
Specimens for tensile tests are usually machined to a circular cross section
or are wires. However, specimens from plates or sheets may be of rectangular
section. Each must have a portion of uniform cross-section on which extension measurements are made. To give an indication of size, a typical tensile
test bar is shown in Figure 4.13.
In addition to verifying the mechanical response of materials, tensile
testing is used as a quality control tool, at the point of product manufacture.
Testing at this point will ensure that quality and design specifications are met
and maintained before the component or product enters service.
A few of the limitations to the use of tensile testing as a forensic tool are
listed below:
4.5.2
107
Hardness Testing
The Vickers hardness test involves measuring the size of an impression left
by a 136o diamond pyramid indenter (see Figure 4.14), under a specified
load. A hardness machine is used to hold the sample being examined and
for applying the diamond indent. The indent diagonals are measured with a
calibrated, low power optical microscope and the measured value is known
as the Vickers hardness number Hv (see Table 4.1).
Microhardness measurement is possible using the Vickers test, but does
require a polished surface. With indentation loads of less than 500 g the
results are not quite the same as those obtained with the standard Vickers
hardness test.
108
Figure 4.14 (A) Vickers pyramidal diamond indenter. (B) Impression on the
surface of a test specimen.
4.5.5
Similar in principle and operation to the Vickers test, the Brinell hardness
test involves a hardened steel ball indenter. Hardness is expressed as the
Brinell hardness number, BHN or HB. The Brinell test utilizes either a 1- or
10-mm-diameter steel ball in place of a diamond pyramid. Because the ball
is steel, it can itself deform under load and will result in misleading readings
on hard surfaces (hardness > 600 Hv). Use of the 10-mm ball allows one
indentation to sample over a larger area of the component than can be
accomplished with a Vickers diamond indenter. For the failure investigator
it is, therefore, a useful method for testing items such as castings, where
coarse grains can give different readings in different places if a small indenter
(such as a 1-mm ball or Vickers diamond) is used.
4.5.6
109
The Rockwell C scale (Rc) is the most widely used hardness scale for workshop
measurements of steel tools and hardened artifacts since they are relatively
unaffected by shallow surface effects such as decarburization. The objects to
be tested are indented with a diamond cone and the depth of penetration
measured. The significant difference in principle is a prior application of an
initial light load, followed by a further load being applied, at which point
the additional depth of penetration is measured. Application of a preload
means no surface preparation is required. Readings are taken directly from
a dial and therefore minimal skill is required to perform the test. For testing
the hardness of nonferrous materials, the Rockwell B hardness test is used,
where a steel ball is used as an indenter.
4.5.7
It can be shown that the Vickers hardness test (Hv) bears a direct relationship
to the tensile strength (sTS) of a material. However, the way in which a
material responds will depend on the ratio of its yield stress (sy) to Youngs
modulus (E). For softer metals with a low sTS/E, it is found that
Hv = 3sTS
(4.1)
(4.2)
Typical values of Hv for a range of material are shown in Table 4.1. When
specimens are small or when standard test bars for tensile or shear testing
cannot be taken from material evidence, a hardness test may be the only
means of obtaining a useful estimate of the tensile strength of the material.
A relationship between the different hardness testing methods can be shown
in an approximate equivalent hardness table, shown in Table 4.2.
Any investigative tool can be used in isolation, or in combination with
others to reinforce findings. A particular case relating to production difficulties of surgical eye probes concisely illustrates the versatile use of both hardness testing and electron microscopy during the course of an investigation.
110
Table 4.1 Typical Values of Vickers Hardness Number, Hv, for a Range of Different
Materials
Material
Tin
Aluminium
Gold
Copper
Iron
Mild steel
Fully hardened
steel
Hv
N mm2
5
25
35
40
80
140
900
Material
Hv
N mm2
Limestone
250
MgO
Window glass
Fused silica
Granite
Quartz
Al2O3
500
550
720
850
1200
2500
Material
Hv
N mm2
Tungsten
carbide
Polycarbonate
PVC
Polyacetal
PMMA
Polystyrene
Epoxy
2500
14
16
18
20
21
45
Table 4.2 Relationship between Brinell and Rockwell Methods of Hardness Testing
Hv 30
700
600
400
200
100
Brinell
Brinell
Rockwell
(Rc)
615
550
379
190
95
665
564
379
190
95
60.1
55.2
40.8
(11.0)
Unsuitable
Equivalent tensile
strength
of steel (MPa)
2100
1850
1250
620
300
111
Figure 4.15 Optical micrographs obtained from a good eye probe (A) and
from a partially fabricated probe from the suspect batch (B). Magnification: 400.
x-ray analysis was undertaken and confirmed that two different titanium
systems had been used, i.e.,
The successful batch of eye probes had been manufactured from Ti6Al-4V alloy.
The unsuccessful batch had been manufactured from CP titanium.
Bend Tests
Tension is also encountered in bending, and there are several kinds of test
for bending simple shapes like beams of uniform section. Thus a beam can
be loaded at three points and strained to failure. As the beam is deformed,
112
one surface is deformed into a convex shape and the material here is in
tension, while the opposite surface is deformed into a concave shape where
it is under compression. It is often thought that forensic work, being scientific
by nature, must always be accompanied by long mathematical arguments.
However, this is not the case as simple logic matched against the evidence is
the only way in which causation is determined. Nevertheless, there are occasions where quantitative evaluation becomes valuable. A simple equation
used for evaluating stresses and strains on beams in bending is the engineers
bending equation:
M/I = s /y = E/R
where M is the moment on the beam, I the second moment of area of the
beam section, s the stress at a distance y across the beam section, E the tensile
modulus of the beam material and R the radius of curvature of the bent
beam (Figure 4.17). Some of these variables become constant when a single
type of beam is considered under two different loading conditions. Thus E
is constant (same material) and I is constant (same section shape).
A particular case that required calculation of strain on the outside of a
bend occurred when a manufacturer encountered a problem when forming
113
R
F
F
d
Figure 4.17 A state of bending caused by applying equal and opposite forces
at the ends of a beam.
5-mm-thick sheet aluminum in a vehicle body pressing plant. The manufacturer supplied samples of both intact and cracked press formed bends and
verified that all parameters (tooling, processing, etc.) were identical.
Visual observation of the two panels showed that forming processes had
not been identical (Figure 4.18). Measurement of the uncracked sample
revealed an inside radius of 7 mm, whereas the cracked sample had been
formed to an inside radius of 3 mm. A simple bending strain calculation
(Figure 4.19) indicated that the strain of the outer skin was of the order of
26% in the satisfactory batch, compared with 45% in the unsatisfactory.
It was therefore considered that the bend radius was the major cause of
the cracking, as the smaller radius was introducing 45% strain at the outer
surface of the bend well above the ductility of the aluminum sheet.
4.5.9
Torsion Testing
114
Figure 4.18 Sections across 90 bends from cracked and uncracked aluminum
vehicle body panels. Note the two different radii. For scale, leg length = 50 mm
(2 in.).
Bending strain calculation:
y
where y = distance from neutral axes and R = radius of curvature of neutral
R
axes. In this case, y = ymax = distance to outer fiber
x =
y = 2.5 mm
ymax: 5 mm
R = 3 mm + 2.5 mm = 5.5 mm
and
R = 7 mm + 2.5 mm = 9.5 mm
Neutral axis
115
pins that limit the amount or torque or bending moment that may be
transmitted. In this way these cheap devices protect the rest of the more
expensive mechanism from expensive overload. They function in a similar
way to a fuse in an electrical circuit. Indeed in some circumstances shear
pins are called fuse-pins or fuse links. Flywheels, storing energy in
rotating machines, may cause severe damage if the machine is suddenly
stopped for any reason. By incorporating a shear pin the damage is greatly
reduced. Their designers need to know the applied force that will cause
them to fail by shear overload. Many airliner engines are attached to their
mounting pylons in this way, and are designed to shear away if engine
vibrations risk damage to the airframe.
Brittle Lacquer
Brittle lacquers can be used to provide a simple and direct way of estimating
the degree of strain in deformed bodies. The product is coated with a thin
layer of brittle polymer, using a solution of the polymer in an organic solvent.
After drying it forms a thin uniform film that is well bonded to the underlying
metal. When strained to a high degree, the metal deforms plastically, and the
thin film also deforms. Because it is brittle, however, the film cracks at only
very small strains, so that it is covered with a pattern of cracks that reflect
the underlying plastic deformation of the metal product. The cracks are
oriented at right angles to the tensile strains in the metal, so the method
allows easy identification of the plastic deformation. In other words, the crack
pattern bears witness to the loading history of the product, as shown in Figure
4.20, for example.
This basic idea can be used for identifying loading patterns in failed
products, and the case where a cyclist was injured after riding into the back
116
A
Uncracked Coating
Cracked
Coating
Tensile
Loading
Tensile
Loading
Envelope
of Cracks
Set of Cracks
117
Figure 4.21 Head tube of a cycle frame showing cracking, resulting from plastic
strain prior to fracture, in its brittle lacquer paint coating.
On receipt of the head tube section, the first observation was the clear
presence of cracks in the yellow paint (Figure 4.21). It was determined that
other markings on the head tube were severe scratches, emanating from vice
jaws the frame had been held in a vice while hack-sawing the section and
had nothing to do with the actual failure. However, the weld at the bottom
of the head tube of the cycle frame revealed an interesting pattern of strain
cracks. The paint had acted as a brittle lacquer, and bore witness to the fact
that the steel tube had undergone extensive deformation under frontal impact
before the weld failed.
The scenario can be reconstructed: the rider, head down, half asleep after
a night shift, gritting his teeth as he rode into the rain, ran straight into the
back of the unlit parked van; the weld failed under the impact. There was
no evidence whatever of fatigue cracking in the fracture surface. If there had
been a preexisting crack, the paint film on either side of it would not have
exhibited strain cracks.
4.6.2
Cracks on the surface of a component can often be seen by eye. These cracks
are convincing pieces of evidence for structural failure. A crack is an opening
of the material that must have been formed by strain, no matter how small.
Several methods are available for detecting hairline cracks and are used to
discover whether there are any cracks in a worn component, such as a crankshaft, before regrinding or refurbishing. Such methods are routinely used for
inspection of safety-critical parts, such as bodies and wings of aircraft, pressure vessels and so on. For example, in July 2000 hairline cracks in parts of
118
119
Figure 4.22 Broken ice-skate blade (A), optical photograph of fracture surface
(B) and surface cracking revealed by dye-penetrant examination (C).
parallel with the x-ray beam, Figure 4.23. Radiographs will reveal whether
defects or discontinuities are present in castings, forgings and weldments, in
addition to polymeric and other nonmetallic parts. Information that can be
gathered from radiographs would include surface cracking, internal voids,
reinforcing flow patterns, density of reinforcing and fiber orientation (particularly at critically stressed points).
Soft x-rays are useful for showing the internal structure (including internal
cracks) of composite material products, and even thermoplastic components.
The technique was used in a dispute concerning the quality of gas injectionmolded office chair arms (Chapter 5, Section 5.10). The defendants
would not allow the samples to be sectioned physically, so the expert for the
claimants suggested a trip to the local hospital to radiograph the arms nondestructively. On inspection of the radiograph, it was seen that the gas pocket
120
Figure 4.23 X-ray orientation for crack detection (A), compared with threedimensional voids (B).
had penetrated the lower zone of attachment to the shell. The point being
made is that the defect was detected without physically altering the component.
The technique has further advantages for failure investigation in that
information obtained from a radiograph is often of value when choosing a
location for sectioning. For example, radiographs can be utilized to determine a site that will provide optimum information for metallographic sectioning and examination. More to the point, it will allow choice of sectioning
site that will not destroy useful information. However, one major limitation
of the technique is that, for all samples, cracks that are inclined or perpendicular to the beam are very difficult to detect as they have little effect on
beam absorbance (Figure 4.23).
This method is vital to personal injury cases, such as when a young man
suffered injury when using a circular saw. A tooth came off an almost new
blade and went right through his eye, lodging close to his brain. The final
position was so delicate that surgeons could not risk removing it. However,
x-rays taken at different angles allowed identification as a tungsten carbide
tooth of the same dimensions as those on the failed blade, two of which were
found to have broken off due to faulty brazing.
121
0.1mW
Solder standard
(Sn 0.5Cu)
Suspect
Sn 0.5Cu
solder
177
178
179
180
181
182
183
(C)
Figure 4.24 Comparative differential scanning calorimetry (DSC) scan of suspect eutectic Sn-0.5Cu solder against that of a known standard sample. Melting
temperature for the solder is 227C. The suspect solder displayed a low temperature (~186C) melting phase.
122
123
Figure 4.25 (A) Ceramic capacitors on printed circuit board (PCB). (B) Section through a failed ceramic capacitor. (C and D) Cracking within solder
joint. (E) EDAX analysis of solder.
aramid (Kevlar, Twaron) or UHMPE fiber (Dyneema, Spectra) are exceptional in that the polymer chains are fully aligned along the fiber axis. The
degree of crystallinity in common partly crystalline polymers such as
polypropylene is thus variable, depends on how the polymer has been processed, and can be measured using DSC. Values of T g and Tm are tabulated.6
A case example where DSC proved of great value in unraveling the truth
124
2
210.56C
-2
Intact Seal
30.72 mJ/mg
-4
205.95C
-6
218.89C
Failed Seal
-8
43.32 mJ/mg
-10
+
220.03C
-12
-14
100
120
140
160
180
200
220
240
260
280
Temperature (C)
Figure 4.26 Melting behavior of new intact and failed seals made from Hytrel
thermoplastic elastomer as shown by DSC spectra.
concerned the widespread failure of simple round washers for sealing radiator
bleed valves. A new polymer, a thermoplastic elastomer (Hytrel, a block
copolymer of a polyester and a polyether) was introduced in the U.K. to
overcome the problem of leaking taps from fiber washers. It worked well in
both hot and cold water taps, but failed in radiator valves. Comparison of
an intact new seal and a failed seal showed a significant increase in crystallinity (Figure 4.26). Exposure to continuous high temperatures of a central
heating system could increase the crystallinity with adverse effects on the
strength of the seal, ultimately leading to cracking. The DSC curves show the
increase in heat of fusion from about 30 mJ/mg to well over 40 mJ/mg as a
result of exposure.
4.7.2
Infrared Spectroscopy
Spectroscopy involves exposure of a thin sample to a beam of infrared radiation and recording the absorption spectrum of the material. The nature of
the polymer chains produces absorption at particular wavelengths which are
characteristic of the polymer, and so is a primary aid to identification of the
material. When fillers and other additives are used in the material, more
complex spectra are obtained simply because extra absorption occurs. There
are at least two ways of preparing the sample thin enough for analysis. Films
must generally be 20- to 60-m thick, so that the infrared beam can pass
through the sample. A thin film can be cast from a suitable solution of the
polymer (not always easy since many polymers are insoluble in common
125
61.24
1735.7
%T
C = 0 Absorption
14.4
4000 3500 3000 2500 2000
cm1
1500
1000
solvents) or a thin slice cut using a microtome. Even a slice cut using a sharp
scalpel can yield useful information if a microtome is unavailable. Solvent
casting is useful where the matrix polymer must be determined because most
inorganic fillers are insoluble and not carried over into the thin film. However, cut sections are vital if the material is to be examined in its original
state. FTIR (Fourier transform IR) is a sophisticated form of spectroscopy
where the thin sample can be scanned repeatedly so as to improve resolution.
The method is so important for polymers that it is routinely used for
failed materials, although care is needed since the method is partly destructive. For example, when a disabled woman was badly injured by fracture of
her crutch, the plastic part that cracked was analyzed as shown in Figure
4.27. The spectrum from a solvent-cast film showed that the material was a
polypropylene copolymer. However, it exhibited anomalous absorption at
1736 cm1, as found by comparison with reference spectra.7 Oxidation of the
material is usually found at this position, and can be caused by poor processing at too high a temperature. When material degrades, the chains are
broken and the strength falls, which explains why the crutch broke when the
woman put her weight onto the device. The plastic connector fractured and
she was injured as a direct result.
A second, quite different example reinforces the point. A rotational
molder of large products like mancabs and road cones suffered a number of
premature failures due to brittle cracking. The process route involves placing
preweighed polyethylene powder into a tool that is then closed and rotated
126
Wavelength (m)
6
15
10
100
50
% Transmission
70
20
C=0
0
1800
1600
1400
1200
1000
800
Figure 4.28 Infrared spectrum of a roto-molded polyethylene road cone showing intense absorption due to oxidation and UV attack leading to brittle cracking.
in a large oven. The plastic melts and sticks to the walls, where it solidifies
when the oven is cooled. High process temperatures are needed due to the
high viscosity of molten polymer. Oxidation tends to occur at the outer
surfaces of the product, and when exposed to sunlight, the chains degrade
preferentially at these points. Figure 4.28 shows the enlarged oxidation peak
in a thin solvent-cast film from a road cone that had been attacked by the
UV radiation present in sunlight (the same radiation produces sunburn).
The problem was solved by adding a UV absorber (which acts like sunburn
cream) and antioxidant to the raw power polymer.
Derived infrared techniques offer completely nondestructive analysis of
small samples. Thus FTIR microscopy involves using an infrared microbeam
to be shone through film, fibers or even dust samples in an optical microscope. The technique was used almost incidentally in a failed catheter case,
and showed the presence of degraded products that should not have been
present if the material had showed any integrity (Chapter 9). The surfaces
of samples can be analyzed using attenuated total reflection (ATR) spectroscopy, where the IR beam is channeled onto the surface and reflected back
for analysis in the usual way.
4.7.3
Chromatography
Another technique unique to polymers is GPC, or gel permeation chromatography. It yields information on the molecular mass (molecular weight) of
a polymer (or the chain length). As mechanical strength is directly related to
molecular mass, it is a powerful method when investigating failed polymers.
127
Other Methods
Integrity of Analysis
When such special methods are use for analysis, care is needed to ensure that
they retain their integrity. One test for integrity is reproducibility: can
another expert perform the same test and achieve the same result? It means
128
129
130
way, FTIR, DSC and microscopy are common starting points for studying
failed polymer products.
References
1. Davies, J.R., Ed., Metals Handbook Desk Edition, 2nd ed., ASM, Materials
Park, OH, 1998.
2. Newby, J., ASM Handbook, Vol. 1: Properties and Selection: Irons, Steels, and
High Performance Alloys, 10th ed., ASM International, Materials Park, OH,
1990.
3. Newby, J., Ed., ASM Handbook, Vol. 8: Mechanical Testing, 9th ed., ASM
International, Materials Park, OH, 1989.
4. Newby, J., Ed., ASM Handbook, Vol. 9: Metallography and Microstructures, 9th
ed., ASM International, Materials Park, OH, 1989.
5. Weidmann, G., Lewis, P., and Reid, N., Eds., Structural Materials, Materials
in Action Series, Butterworths, London, 1990.
6. Brandrup, J. and Immergut, E.H., Eds., Polymer Handbook, 3rd ed., John
Wiley, New York, 1989.
7. Hummel, D.O., Ed., Polymer and Plastics Analysis, 2nd ed., Carl Hanser Verlag,
Munich, 1978.
8. Crompton, T.R., Ed., Analysis of Polymers: An Introduction, Pergamon Press,
Oxford, 1989.
9. Lewis, P.R., Premature fracture of Noryl transformer plugs, oral presentation,
FAPSIG session, ANTEC 2003, Nashville, TN, 2003.
Failure Due to
Manufacturing Faults
131
132
Figure 5.1 Section across the fractured brake lug from the motorcycle (fracture
at right-hand side).
weakened by gas porosity and shrinkage. A similar fault accounted for the
failure of a robot arm casting in aluminum alloy. The robot was designed for
use in a nuclear installation but the arm broke off the body while it was
undergoing proving tests. The seriously defective condition of the flange that
carried the arm was obvious from the state of this machined surface, so the
casting should have been rejected out of hand and certainly not built into a
robot destined to be used for handling radioactive materials. Unfortunately
the project was behind schedule, so the casting went straight into the machine
shop as soon as it was delivered and the machinist did not consider it part
of his duty to report such faults. This was a difficult casting to make and a
number of replacements were x-rayed and found to be similarly defective.
The problem was traced to gases released from the resin binder used to form
the sand molds in the foundry coupled with inadequate feeding of the heavier
sections of the body casting.
Another example, not involving a casting, concerned pivot eyes welded
to the ends of hydraulic cylinders forming part of a lifeboat davit on a
ship. Superficially, the welds appeared to be sound, but the first time they
were deployed in a proving test the eyes broke off and revealed there had
been practically no fusion between the eye and the base of the ram.
Figure 5.2 shows the ends of three of these rams; the two with the large
diameter eyes facing the camera had failed and caused extensive damage
to the installation.
133
134
Tear
involving not only the transfer of the load to another container, but also
finding an empty container and shipping it to the port, loss of earnings of
the damaged container and cost of repair. There were at least four good
reasons for requiring an urgent investigation:
1. Whether or not the owners had a case for claiming damages from
any other party such as the harbor boards or the container manufacturers
2. Identification of the fault and a satisfactory method of repair; it would
be wrong simply to repair the joints according to the original design
before vindicating the design
3. Whether further failures of similar containers were to be expected
and, if so, whether these might be avoided by reducing the recommended load carrying capacity
4. Whether partial failures of this kind might lead to complete unzipping
and failure of the containers during a voyage or when offloading, with
risk of serious accident and consequent damage to the cargo
As a preliminary to this investigation, a mechanical engineer with extensive experience in riveted stressed skin structures in aircraft had concluded
there was nothing wrong with the original design or method of construction
of these containers. This was borne out by the continuing satisfactory performance of the vast majority of these containers, some of which had been
in use for several years. This led to the view that something had gone wrong
with the manufacturing processes or material quality control, which
accounted for structural weakness in these particular failed containers.
135
Figure 5.4 Scanning electron micrographs of rivet fracture: (A) 20; (B) 55;
(C) 6000.
The side panels were sheets of aluminum alloy that were attached to one
another and the frame by riveting along vertical lap joints incorporating tophat channel sections to act as stiffeners. All the failures involved the unzipping of the vertical lap joint between the first and second sheets in from the
end of the container. The rivets had sheared across at the interface without
causing any tearing or ovality of the rivet holes. Figure 5.4 shows three SEM
views of a typical fracture surface at various magnifications. These are clearly
single shear overstress fractures. A rivet was also sectioned, polished and
etched for microstructural examination. This confirmed the shearing mode,
as evident in Figure 5.5, by the way the grain structure has flowed adjacent
to the fractured edge. The above interpretation explains the lack of any
stretching of the holes in the sheet on either side of the rivets, so the investigation concentrated on what might be wrong with the rivets.
136
The first question is whether the rivets were the right composition. An
element scan utilizing pulse counts of x-rays of characteristic wavelength for
the elements likely to be present was carried out in the SEM and compared
with that of an unused rivet of the correct composition (HR15 alloy to
BS1473). The traces were as shown in Figure 5.6. While element scans are
not precisely quantitative, the fact that the two are practically identical established that the correct alloy had been used for rivets in the failed containers.
Examination of the failed rivets under a low-power microscope revealed
no physical sign of any internal fault or of a progressive failure mode such
as fatigue, wear or corrosion. However, if one rivet near the end of a seam
had failed it would have thrown an extra load onto its neighbors and these
137
Aluminum
Number of Counts
Trace A
Magnesium
Copper
Silicon
Manganese
8
Energy keV
Aluminum
Number of Counts
Trace B
Magnesium
Copper
Silicon
Manganese
8
Energy keV
Figure 5.6 X-ray scans from SEM analysis of broken and unused rivets.
might then in turn become overstressed, causing the entire set to progressively unzip. Hence the next task was to examine the specification for the
rivets on the engineering drawing. The relevant notes and diagram are reproduced in Figure 5.7 and Figure 5.8. It will be noticed that the rivet stock
should be supplied in the H2 condition (as manufactured by the shaping and
heading operations), but before use must be solution treated and then used
within 48 h.
138
Figure 5.7 Extract from workshop drawing of rivet for container seams: drawing notes rivet with dimensions.
0.32 in.
0.160 in.
0.5 in.
0.187 in.
Figure 5.8
139
Balance
Copper
Magnesium
Silicon
Iron
Manganese
Zinc
Chromium
Titanium*
0.20.8
0.20.8
0.50.9
0.7
0.41.2
0.2
0.1
0.2
Heat-Treated
Condition
Diameter
(mm)
Tensile Strength
(MN/m2)
H2 Annealed and
cold drawn
20 40% reduction
of area
TB Solution heat
treated and
naturally aged
Up to 2
385
out before the aging gets under way, usually within an hour or so of quenching from the solution treatment. Table 5.1 and Table 5.2 are taken from
BS1473 for HR15 alloy. The tensile strength given is for the TB (naturally
aged) condition. In the H2 as-wrought condition in which these rivets were
supplied, the tensile strength would be approximately 240 MN m 2, but the
metal would have high ductility and could be worked without any risk of
cracking. In contrast, in the aged TB condition a rivet would not possess
enough ductility for it to be set without the head splitting. Thus immediately
before setting them in the container seam, the stock rivets would have to be
solution treated, quenched and then set within a time window of about 1 h
before the age hardening process began.
It is quite clear from the hardness tests on the failed rivets that they had
not responded to the age hardening process as they should have and were
consequently well below the strength levels that ought to have been achieved.
The cause was obvious they were of the correct composition but had not
been solution heat treated before setting. This was borne out by their microstructures and, in fact, had already been suspected by the investigating metallurgist from the microstructure of the failed rivet shown in Figure 5.6. If
the rivet had been solution treated the grains would have been much larger
140
Shear Strength
140
Tensile Strength
Hardness HV
120
100
80
60
40
20
40
80
120
160
200
240
280
320
360
400
440
480
Strength/MN m2
Figure 5.9 Correlation of hardness with the tensile and shear strengths for the
aluminum alloy HR155.
141
Strength means the maximum value of load the component can bear
in a given situation for example, in tension, compression or shear
without failing in a given way (for example, by fracture). Static implies that
the loads are assumed to be constant and do not vary with time. The word
nominal is used here to mean according to the specification and the design
calculations. These calculations are usually based on a simplified model of
the component, and may ignore some of the complications of service conditions. If these complications could be taken into account quantitatively, on
the one hand they would erode the nominal static strength and, on the other
hand, they would inflate the nominal static loading. The factor of safety must
be large enough to cover the uncertainties in strength and loading, which is
why it is sometimes referred to as a factor of ignorance.
It is essential that the static factor of safety be large enough that at all
times within the lifetime of the component, the true strength will exceed the
true loading. Whenever there is a service failure, the static safety factor should
be reexamined. The investigating engineer has two options. He could choose
a theoretical model of the loaded container and then analyze it to find the
loads acting on the rivets, which would provide an estimate of the nominal
static loading on a given rivet. Alternatively, the distortions in the sidewall
of a loaded container could be measured and, using this information, infer
the loads that are transmitted to the rivets. This should provide a more
accurate estimate of the static loading on a rivet.
142
143
all 48 rivets, whereas the distribution would in fact be parabolic, with the
maximum loading being near the middle of the seam. This distribution could
result in the rivets near the middle, which is where the splits appeared to
have started, being exposed to a shear stress well above the average, in which
case the factor of safety may have been almost eliminated. To take into
account these two factors would be a daunting theoretical task. Furthermore,
when the panels of a stressed skin container are loaded in shear a secondary
force field develops due to buckling, the so-called diagonal cross tension field,
which imposes additional shearing force on the rivets. An alternative
approach to the problem would have been to measure strains in a real container loaded to its full capacity and then infer from these strains what the
stresses must be. It is a technique often used to study stress distributions in
complex systems. However, the investigation did not justify this costly and
time-consuming analysis. Furthermore, it would only have supplied data for
a static container, whereas these containers split while they were being handled, so a dynamic analysis under transient loading conditions is really
required.
Whatever theoretical modeling is undertaken, what is definitely known
is that the original design of these containers had been approved and tests
on prototypes had met the requirements of all relevant standards without
reservation. Moreover, the vast majority of these containers, presumably not
constructed with defective rivets, were still giving satisfactory service several
years after the design was introduced.
Another factor that came to light during the stressing examination was
the effect of load distribution inside the container. A single heavy item, such
as a machine tool, places a different loading on the sidewalls than the same
total load distributed evenly. If the container being handled bumps into
something, then the inertia loads also add significantly to the shear stress on
the rivets and is an obvious explanation as to why the splits occurred while
containers were being off-loaded and why all those that failed had held heavy
items of machinery. It is easy to appreciate that a container packed full of
clothing, for example, would be much less vulnerable than one with a single,
heavy item fastened to the floor near the middle. Additionally, the heaviest
part of a machine tool might be at one end, which would increase the shear
forces in the riveted seams at that end of the container.
The formal report concluded that a batch of containers had been produced with a side-wall strength below that which was intended, because the
rivets had been set in the wrong condition, that is, without having been
solution treated as specified on the drawing. The factor of safety in the basic
design was adequate, but in the containers that split open it was reduced to
an unacceptable level because the rivets were too weak. The failures were
144
145
firm that bought steels of various compositions and rolled them into sections
and specific sizes ordered by a large number of small engineering customers.
Fortunately, this firm carried product liability insurance because the root
cause of the fatigue failures was that the wrong grade of steel had been
supplied, which was totally unsuitable for the type of heat treatment process
used to case harden the gears.
Steels for general engineering purposes are alloys of iron containing up
to 0.8% by weight carbon together with small percentages of other elements
such as manganese, chromium, nickel and vanadium. Without a doubt they
comprise mans most versatile engineering materials, mainly because their
carbon content allows a single product, say a chisel, to be heat treated so as
to achieve extremely high hardness at the cutting edge, a middle section
strong and tough and a softer head that is malleable. Unfortunately, when
medium and high carbon steels are heat treated to maximum hardness they
are also brittle and may sometimes crack under thermal contraction stresses
during heat treatment processes. The other alloying elements modify the heat
treatment response and impart specific properties in their own right, though
not on such a dramatic scale as carbon.
Additional carbon may be introduced into the surface of a shaped product to a controlled depth, enabling the maximum hardness and wear resistance possible to be attained in the outside skin. This is the purpose of case
hardening used to develop maximum wear resistance in parts which are in
sliding contact, such as gear teeth, while retaining less hard but much tougher
properties in the main body of the product. Typically, hardened cases are
between 0.4 and 1 mm deep and, of course, the carburization process can
only be applied to products made from steel when they have already been
manufactured to the required shape and dimensions.
The oil pump gears had been forged to shape, machined to size and case
hardened. Final grinding was used to correct any slight distortions resulting
from the case-hardening process. The preexisting cracks discovered in the
failed gears were typical of quench cracking. So on what grounds did the
gear manufacturer seek to sue its steel suppliers?
Comparison of the analyses of the steels specified to make the failed gears
and the one that was delivered by the material supplier were as follows:
Carbon (%)
Ordered
Failed gear
0.2 max
0.38
Manganese (%)
Nickel (%)
Chromium (%)
0.61.0
1.62
0.61.0
0.16
0.40.8
0.04
The significant figure is the carbon content. The specified steel has a low
carbon content and does not undergo an appreciable volume change when
146
147
Figure 5.11 Failed manhole step (left) beside deliberately broken one (right).
climbing in or out. The castings were hot-dip galvanized before being dispatched to a builders merchant, who had sold several hundred to a subcontractor constructing manholes in a sewer system.
The broken footrest end of the accident step, along with the remains of
two other steps from the subcontractors unused stock that had been broken
deliberately as a test after the accident, were submitted for examination.
Figure 5.11 shows the accident footrest on the left and one of the tested steps
on the right, viewed from the underside. The short piece of limb above the
accident step is from another of the tested footrests, to illustrate the length
of the limbs that had been left set in the brickwork. Figure 5.12 is the accident
step, viewed slightly from above, showing the fractures. There is a large
casting blowhole in the flange at the right-hand side and a smaller one at the
outside edge that is part of the fracture, as can be seen in the closer view
depicted in Figure 5.13. It occupies approximately 5% of the total crosssectional area of both fractured limbs. These blowholes had been formed by
148
Figure 5.13
Figure 5.14
75).
gases evolved at the interface between the metal and the sand mold when the
liquid iron was poured. Holes formed by gases being released within the body
of liquid metal during solidification would be much smaller and distributed
uniformly throughout the cross section.
The foundry can therefore be blamed for releasing a casting containing
visible blowholes, but the total cross-sectional area of only the one limb is
reduced by 5% or so; hence they would have minimal effect on the bending
strength. As there is not the slightest suggestion this could be a fatigue failure,
these faults cannot be responsible for the sudden fracture that occurred the
first time the footrest came under load.
149
Figure 5.15 Etched microstructure of spheroidal graphite cast iron (magnification 75).
There was no sign that either limb had deformed during the course of
fracture. This is not consistent with a spheroidal graphite cast iron conforming to BS1247. Castings made to this specification should have a minimum
of 7% elongation in a tensile test, so both limbs should have deformed before
and during the course of fracturing. A section was therefore cut from the left
side of the fracture illustrated in Figure 5.13, so that a cross section of the
fractured end could be polished and etched for microstructural examination.
This revealed the structure to consist of graphite flakes set in a mainly ferritic
matrix. Hence this casting exhibited the microstructure of a common gray
cast iron and would be brittle, not ductile and shock resistant. Figures 5.14
and 5.15 are typical microstructures of these two types. Due to the graphite
appearing in the microstructure as flakes instead of spheroids this casting
would display no ductility whatsoever in a bend test and its tensile strength
would be only two thirds of what it should have been.
In order to cause the graphite to separate as spheroids a small amount
of magnesium (0.5% by weight) has to be added to the ladle before the metal
is cast and, additionally, there are certain limits to impurity levels that must
be observed. If the magnesium is not added, or the ladle of liquid metal is
held too long before casting, then the magnesium burns off, and the metal
will solidify with a flake graphite microstructure. There would have been no
outwardly visible sign that the microstructure of this step casting was defective. The form of the graphite would only have been detectable in the foundry
by a destructive test, though a test bar could have cast at the same time if
any fault had been suspected. The subsequent galvanizing process had no
effect on the microstructure or mechanical properties.
150
It might be argued that the blowholes could have been a reason for
rejecting this step at every stage in its handling, that is, before it left the
foundry, when it was galvanized, when it was sold or even when set in the
brickwork of the manhole. However, while it was obvious that this casting
was not of high quality, no one handling it could have had any idea it was
brittle and likely to fail as it did, particularly as lettering was integrally castin to the effect that it conformed to the required specification.
Neither of the two blowholes in the one limb of the U can be held
responsible for the failure that actually occurred. The larger blowhole in the
flange was not associated with the fracture and was not subject to any significant level of stress in service. The smaller one was in the most highly
stressed part of the limb and undoubtedly would have weakened it, although
the limb would still have possessed about 84% of its intended strength; taking
into account the second limb, the step as a whole would have possessed over
90% of its intended strength.
The ladle of metal had obviously not been treated with magnesium or,
if it had been, the magnesium must have burned off before casting; in fact
the distribution and form of the flake graphite in the microstructure indicated that this is most likely what had happened. The implication is that all
castings made from the last metal in this particular ladle would have solidified
with flake graphite structures and would be susceptible to sudden fracture
when loaded. Presumably all of these must still be in service or have already
broken. The fact that this is the only step reported to have broken may be
because it had not been set in mortar so that the flange was right against the
face of the brickwork. The short, exposed parts of the limbs behind the flange
would therefore have been subjected to bending. The brittle flake graphite
iron was unable to withstand these tensile forces on the upper surface of the
limbs and broke. If the same step had been well bedded in mortar, with its
flanges flush with the brickwork, the greater part of the bending force in the
rear limb might have been eliminated. The flange would have countered the
levering action from the step and transferred the bending moment as a
compressive force in the ribs on the underside of the step. Flake graphite
irons are easily fractured by the tensile forces generated when a thin section
is subjected to bending. Hence it could be argued that whoever set this
particular step in the brickwork contributed to the cause of it failing as it
did. Notwithstanding this, if the casting had possessed a spheroidal graphite
structure, the limbs would have bent until the flange came against the brickwork and so transferred the load on the footrest to the rib beneath. It would
not have failed catastrophically.
Responsibility for the footrest having the wrong microstructure and not
conforming to the specification number integrally cast into the tread rests
151
solely with the foundry. Its insurers accepted liability and settled the claim
in full.
152
left the premises the residual hoop stress must have caused the flange to split
open, possibly due to vibration when the burner started up. The consequence
was a major conflagration as the escaping oil ignited and was continually
replenished from the storage tank.
Although some blame was attached to the service engineers employers
for the engineers use of excessive force to tighten the coupling, the major
responsibility rested with the foundry. The cause of the band of flake graphite
is a complex metallurgical effect arising from the use of iron chaplets within
a sand mold to support the cores required to form hollow sections. We need
not go into the detailed explanation as to how this results in a band of flake
graphite structure, except to say that it would have been a one-off effect in
a large number of coupling castings made at the same time and the foundry
would have had no means of detecting its presence. This one would possibly
never have failed if the service engineer had not used such excessive force.
153
to say the least because, in addition to the costs, the vehicle was not earning
its keep while it was off the road. Also, the truck owners had purchased nine
pairs of the replacement gears intending to use them for other vehicles in
their fleet.
The manufacturer suspected there might be something wrong with the
surface hardening heat treatment, and discovered the hardness of the teeth
that had broken were higher than they had specified. The customer returned
all the unused gears and the manufacturer sent them off to have the teeth
tempered to reduce their hardness. They did so in the belief that the fractures
were due to the teeth being brittle. A modified set was put in the gearbox
but this set failed after 4 weeks, every tooth breaking off the inner ring gear.
Hardness checks revealed that they had been softened to the modified hardness. The vehicle operators initiated proceedings against the manufacturer
to recover their losses.
Metallurgical investigation of the failed and unused gears found nothing
wrong with the composition or heat treatment of the steel. However, two of
the original gears that had gradually worn out over a long period of use
distinct from teeth breaking off were also produced. These provided the
answer to the premature failures. The diameter across the outer teeth is 150
mm. These outer teeth had given no trouble; the ones that had broken off
were all from the inner ring some 100 mm in diameter. In the failed gears
examined, two or three had simply worn down until they no longer engaged
with the mating teeth, which is why the gears had to be replaced.
At the time the original gears were made, all steel started off as ingots.
In the as-cast state an ingot has no directionality but, as it is worked down
to billet or bar, directionality appears as grains and inclusions are extended
in the direction of working. This is a structural effect readily demonstrated
154
Figure 5.17 Flow lines in sectioned gears. Left: Satisfactory gear from blank
made by upset forging. Right: Gear machined directly from slice of round billet.
by sectioning a wrought product, smoothing or polishing the surface, followed by etching to reveal the grain structure. The directionality is revealed
as flow lines, which resemble the grain in timber. For any given quality of
wrought steel, toughness and fatigue resistance are much better across the
flow lines than parallel with them, just like timber. The effect is equivalent
to chopping a piece of wood across the grain compared with along the grain,
where it splits easily.
The flow line patterns of the old gear and one of the new were studied
by sectioning across a diameter. The two photographs in Figure 5.17 show
how their grain flow structures were distinctly different. The original gears
had been upset forged before the teeth were cut, whereas the replacements
had been machined directly from round bar. The flow lines reveal how the
old gears had started off as a length of round billet, about twice the length
of the gear but of smaller diameter. This had been hot forged to squash it
down, so that the shape finished with a much larger diameter and shorter
length than when it started. This was done specifically to develop a flow line
pattern following the contour of the gear teeth so that all forces acting to
bend the teeth would act broadly across the grain. The teeth were as tough
as they could be for this type of steel and the inclusions were oriented in
directions that imparted the maximum resistance to fatigue initiation and
brittle fracture.
In contrast, the new gears had been machined from stock of the full
diameter for the gear required. There was no upset forging, so the grain flow
remained the same as the original billet, parallel to the gear axis. The result
was that the flow lines in the teeth were in the most unfavorable orientation
to resist fatigue and brittle fracture. This is why they broke off so quickly
under service loads, despite having the same hardness and tensile strength
155
Figure 5.18 Hook and triangular keeper for nylon securing strap.
156
Figure 5.19 Broken hook with (A) fractured ends replaced and (B) fracture face
exhibiting bright fish eyes characteristic of hydrogen cracking.
Both parts had been bright electroplated with zinc. Nylon webbing would
be stitched over the straight limbs and the straps wrapped over the load on
the vehicle, or sometimes attached to a large piece of machinery on a flat
bed trailer and then tensioned with a specially designed grip.
By far the most likely cause of premature service failures of spring steel
electroplated with zinc is hydrogen-induced cracking. Accordingly, a hook
and a keeper from the unused samples were mounted in a vice and the jaws
tightened so as to cause noticeable elastic deflection. The pressure was eased
to check that no permanent plastic deformation had taken place. The jaws
were then retightened to the same elastic deflection and left in that state at
ambient temperature. Only one limb of the hook was tensioned in this way,
the opposite limb being left unstressed. The triangular keeper was stressed
by compressing one point against the opposite base, which imposed tensile
bending stress on the inside face of the base. The stressed limb of the hook
broke completely after between 48 and 60 h. Figure 5.19A shows the broken
limb placed with the fracture faces almost in contact; while in Figure 5.19B
the hook end is turned back so as to reveal the characteristic fish-eye feature
on both sides of the fracture surfaces. This fracture is quite characteristic of
hydrogen-induced fracture.
157
Heat
Treatment
None required
8 h minimum between 190C and 220C
18 h minimum between 190C and 220C
24 h minimum between 190C and 220C
158
were made from low carbon steel wire. They did not inquire and dispatched
them as plated, without any baking treatment. However, the onus was on the
plating firm to inquire before carrying out the process and this was taken
into account when the respective insurers negotiated settlement of the substantial claims.
159
This type of failure could, and should, have been avoided if a solid rivet
incorporating a shoulder or an external spacer had been used.
160
Figure 5.22 Broken claw at the end of the scaffold guard rail.
off the side-frame. The tube followed him down to the ground, but the end
broken off the claw casting was not recovered. Figure 5.22 and Figure 5.23
show two views of the broken claw as received, when a microsection had
already been cut from the side of the fracture, and Figure 5.24 shows the
unbroken claw at the opposite end.
The claws were aluminum alloy casting welded to a length of alloy tube,
which would be of different composition as it had been extruded, not cast.
The surface of both castings exhibited numerous pin holes, caused by gas
porosity. This is a fault that should not occur with good foundry practice
but which is nevertheless sometimes found in small castings for general
engineering purposes. Unlike shrinkage cavities in a casting, porosity does
not have a marked weakening effect unless it is present in excessive quantities
161
162
163
Figure 5.25 Claw locking mechanism. (Top) Claw locked securely onto frame
tube. Crescent (1) pivots at (P) and is held in place by step abutment against
finger pull (2) by spring S. (Bottom) To release claw: drawing back the finger pull
against spring pressure allows abutments to override, enabling claw to be lifted
off frame tube. To fit the guard rail the claw is positioned over the frame tube
and pushed sharply downward, causing the abutments to ride over each other
against the spring pressure and lock the crescent against the tube. The integrity
of this position depends on the springs maintaining pressure against the back of
the crescent to keep the abutments in the same relative positions.
part was heavily rusted. The legs had rusted away, as could be seen easily
without magnification. Hence there could have been no pressure forcing the
finger pull against the crescent. Without the spring the crescent could hang
downward on its pivot and allow the locking steps to pass each other easily.
Thus the claw could not have been gripping the frame tube as it should have
164
done but was instead simply resting on top. The inability to lock the unbroken
claw was readily demonstrated by placing it over a cross tube; there was no
snap action as it was pushed down the crescent just hung vertically
and it could be lifted off again without any resistance. In contrast the spring
at the broken end was rusted but was still operative, so the two steps were
able to butt against each other and exert their normal locking action. Hence
at the time of the accident the unbroken claw would have been resting on
the frame tube while the broken claw was clamped.
The cause of the guard rail coming off while the engineer was leaning
against it and reaching upward thus had nothing to do with the weakness in
the broken claw. What must have happened is that his movement lifted the
unclamped claw off the frame and caused the rail to swing outward. This
movement applied bending force to the claw at the other end, which was still
clamped, and this caused it to break as the engineer was already falling. When
the rescuers reached the scene the guard rail was lying on the floor and the
piece broken off the claw was not recovered. But the freshly produced fracture
had exposed the porosity in the casting and the initial investigators assumed
this must have been responsible for the accident. They made no mention of
the state of the claw springs or that the legs of the one at the unbroken end
had rusted away and rendered the clamping action inoperative.
One other possibility was considered. The broken claw might have
already been cracked, making it much more likely to twist off as the guard
rail moved outward. Such partial fractures are often found when tubes are
thrown to the ground as scaffolding is dismantled, instead of handing down
the tubes to someone on the ground. If the claw lands on a hard surface endon the impact may bend it inward with sufficient force to crack it or break
it off. However, in this instance there was no staining or oxidation of the
fracture surface to suggest that the claw had been partially cracked for some
time and, unfortunately, the allegations in the Further and Better Particulars
that the hook had snapped due to having insufficient strength were not
substantiated. Because the broken end was not recovered it was not possible
to discover whether this exhibited any witness marks from an impact.
The case did not go to court. The allegations that the hook had snapped
due to a manufacturing fault and that was the direct cause of the accident
could not be substantiated. Instead, the firm that had hired out the scaffold
(who would have been the second defendant) settled the claim. They failed
in their duty to inspect the scaffolding regularly to ensure all parts were in
a safe condition. There was no case against the manufacturers of the scaffold.
Despite the fact that the casting exhibited a degree of gas porosity, it was not
a weakness that caused the accident. If it had been seriously defective, then
failure would be expected almost the first time it came under load in service
165
or, if not immediate overload, the fracture surface would have exhibited a
progressive mode such as fatigue.
166
Figure 5.26 Intact steel lever for moving the gas injection nozzle forward
within the barrel of the injection molding machine. The first arm fractured twice
by fatigue from the upper pivot hole. Scale shown by British 2p coin at right.
Other serious problems had arisen with the gas-molded products. The
new, lighter chair arms were much weaker than expected, mainly because the
attachment points for the plastic arms to the steel chassis of the chairs proved
very weak. They failed in routine quality checks, and entire batches had to
be rejected. When sectioned, the arms showed that the gas pocket created by
injection had not been formed evenly, but rather tended to be localized at
the hotter parts of the tool. This was not at all surprising, because the melt
will be less viscous where it is hottest, and the high pressure gas will naturally
move to those zones in the melt, thus creating an uneven pocket in the final
product. One of those zones was at the attachment points for the screws, and
when in use, can be expected to experience high loading from the user when,
for example, he or she lifts the chair to move it, or when simply pushing the
arms sideways. It would have been disastrous if such arms were released into
the market, and exposed to handling loads from the user. Failure could cause
167
Figure 5.27
Pristine chair arm molding made by machine maker to demonstrate feasibility of gas molding, with gate (G) at upper right. Hesitation marks
had been formed on the outer surface as indicated by the arrows.
personal injury, necessitate a recall of the product and certainly damage the
reputation of the manufacturer to whom the chair arms were supplied for
assembly into finished products.
The problem seemed intractable, and the company sued the machine
maker for loss of production, machine time and the money invested in the
new process. Experts on either side were divided and the dispute was set for
trial in the U.K. Technology Court, a specialist division of the High Court,
which tries technically complex cases with judges experienced in engineering
cases. The court is an innovator in procedure, asking for experts to meet and
prepare statements of agreement and disagreement so that argument in court
is not wasted on spurious matters. The judge himself visited the factory and
observed gas molding in action so as to be better informed about the nature
of the problems.
A key part of the arguments concerned the first samples produced in
Austria by the machine maker. They looked good externally, but there were
now serious doubts as to their structural integrity following experience at
the factory. Would the machine maker allow them to be sectioned to examine
the gas pocket? The machine maker refused, so a nondestructive method was
needed to solve the problem. We turned to x-ray inspection using soft x-rays
as used by doctors to examine limbs for fractures. It would be ideal because
flesh and bone are light materials of a density similar to the polypropylene
of the chair arms. The chair arms might be thinner (3- to 5-cm diameter)
than an adult human arm, but there would be a greater difference in density
168
Figure 5.28 X-radiograph of chair arm showing serious internal defects caused
by poor shape of gas pocket.
between a hollow interior and the solid skin. Under the watchful eyes of both
experts, the samples were duly taken to a local hospital and exposed to the
x-ray beam, one of the arms being shown in Figure 5.27 together with its
corresponding x-ray negative (Figure 5.28). The x-ray image showed big
differences in wall thickness, the inside walls tending to be much thinner
than the outside (probably due to tool temperature variation). The gas pocket
had penetrated the attachment zone at the bottom of the arm, while another
side was solid and had not been affected at all. Thus the problems of the
technique had been present from the start, which would have been apparent
had someone bothered to investigate the trial moldings produced not by the
client, but by the defendant machine makers. The additional problems with
the lever on the machine seemed indicative of the lack of adequate backup,
but did not appear directly connected with the product problems. Another
chair arm showed the same kind of problem (Figure 5.29), and the x-radiograph was checked by sectioning the arm after the action had been settled
(Figure 5.30). The case for the plaintiff was now very strong, but would it
proceed to a long and expensive court trial?
Just before trial was due to start, lawyers for the two sides met and
hammered out a surprising deal. The machine maker was suffering from a
downturn in business and was storing a large number of new machines in
warehouses. The plaintiff needed to reequip its factory and replace aging
molding machines. So the defendants agreed to sell several new machines at
a discount equivalent to the damages claimed by the factory for the failure
169
Figure 5.29 Radiograph of another design of chair arm showing inner defects.
G is the gate where hot polymer enters the tool.
of the process. The two parties settled the action then and there, saving the
cost of a full trial and, more significantly, saving goodwill between the two
parties. It is rare that such settlements are reached, but nonetheless encouraging that disputes can be settled amicably once the facts are established by
expert evidence. Both companies will, in the future, be more circumspect
170
about new technology, although it is also fair to say that gas molding continues to expand in its application, provided that thought is given to the
intended use of that product, and careful control of the gas pocket is implemented. Thus it is far better to apply the method to brand new tools designed
for gas molding, rather than retrofit existing tools, where control is much
more difficult.
Fluid Transport
6.1 Introduction
Among the most demanding roles of any material is that of containing and
transporting fluids. Why? Because escape of the contents due to failure of
the wall can create far more serious damage than the initial break. Fuel spray
from pipe leaks can cause devastating fires, while the damage from corrosive
chemicals from leaking storage tanks can be extensive. Even small losses of
water from supply pipes is serious if undetected, as many homeowners will
know much to their regret from freezing water effects in winter. In supply
systems to larger buildings, the effects are multiplied, and the losses far greater
if equipment is damaged by the leaking water. Design, rather than manufacturing or assembly, defects in central heating systems (CHS) produce a multiplier effect of such severity as to seriously affect the credibility of major
companies and corporations. So what are the requirements for containment
materials?
172
Design Criteria
The form of the chosen material is generally very simple: a pipe is simply a
hollow tube of uniform circular section. Connections between pipes are
obvious sources of stress concentration, either because of the change in
section thickness or free ends left in the joint. They may be doubly at risk if
there are any defects left over in the weld junction, a common problem in
welding of all materials. More complex types of joints can arise at T-junctions
or inserts for taps, requiring care in design to minimize the deleterious stressraising effects of sharp corners, for example.
Special problems of stress concentration can arise in another perhaps
unexpected situation, where pipe systems of different materials are present,
especially where a local system joins the grid or external water supply, for
example. Thus if the plastic pipe system in a factory is joined to the steel
pipes of the supply system outside the factory, then any local load will be
taken by the weaker plastic system, with potentially catastrophic results.
6.2.2
Fluid Transport
173
inside an enclosed system, as, for example, when steel is oxidized by dissolved
oxygen in a CHS water supply, or externally by an aggressive environment.
If dissimilar metals are used at junctions, as is common in soldering and in
many joints, then the possibility of galvanic corrosion arises. Other types of
corrosive attack will be examined in the case studies in detail, some arising
from unexpected sources. Thus when a PVC tape was used to help seal a
joint in a hot water system, thermal degradation of the PVC produced hydrochloric acid, which attacked the adjacent metal.
6.2.3
Case Studies
The case studies in this chapter include mainly metal and polymer systems,
the former being well established for bulk transport and storage of a wide
range of fluids for many years. The case studies demonstrate the problems
that can arise in CHS systems, whether in a domestic or industrial environment. Failed connections in oil lines can cause serious environmental problems, and there is a study of the special problems of vehicle brake line failures.
While metal alloys remain well used and familiar to all engineers, the recurrence of failures should act as a timely reminder of the dangers that can occur
both in new installations and older products or systems.
Plastic materials apparently offer great benefits for water supply systems
due to their great resistance to corrosion, the bugbear of steel and copper
conventional systems, but have in recent years suffered a series of failures,
especially in the U.S. There have been fewer problems with such systems in
Europe, although failures have occurred from poorly made joints, for example, a topic discussed in detail in this chapter. Different types of failure are
examined in relation to seals in systems. They are crucial but often neglected
parts of a closed distribution network, such as in hydraulic braking systems,
where failure can have immediate and catastrophic effects on users. In CHS
networks, seal failure may be slow and undetected for some time, although
consequential damage can be severe in financial terms, as one of the case
studies demonstrates.
Flexible fuel pipes are common in vehicle engines, but once again, often
overlooked as a potential hazard, perhaps because of their low cost and
primitive function. However, if failure occurs, the results can be very serious,
resulting in loss of the vehicle or worse. The investigation into a series of
such fires in new cars in the 1970s surprisingly reappeared in the early 1990s
as a direct result of a failure to learn the lessons from first failure. Indeed,
the same kind of fuel pipe (nylon) was involved in an automobile accident
that caused several subsequent accidents due to a fuel leak. The cause lay not
so much in the nylon material, however, as in a small molded junction that
connected three pipes together.
174
Fluid Transport
175
painting, electroplating or hot dip galvanizing, to superimposition of electrical potentials on a large structure such as an undersea oil platform.
Changing the battery connections on motor cars from positive earth to
negative earth in the 1950s led to a tremendous reduction in the rate of
what was then referred to as body rot.
Some metals display an inherently greater corrosion resistance than others, but even those most resistant to normal atmospheric conditions have
some reagent to which they are vulnerable. Stainless steels, for example, are
virtually unattacked under oxidizing conditions because they have a built-in
chromium content (at least 12%) which forms an electrically insulating and
self-healing protective oxide film preventing most corrosive agents being able
to set up an electrical circuit. If the oxide film is mechanically damaged it
immediately re-forms and keeps the reactive agent away from the metal.
However, if the environmental conditions reduce or exclude the oxygen necessary to re-form the chromium oxide film, stainless steels will corrode away
almost as fast as nonstainless varieties. One excellent example of this process
was when a family went away for a 6-week vacation and inadvertently left a
pork chop on the draining board of their expensive new stainless steel sink
unit. When they returned and scraped off the offending mess they discovered
a mass of pinholes through the black stain of the chops outline. Bacteria
from the decomposing meat had established chemically reducing conditions
and had consumed all the available oxygen, so there was none available to
re-form the chromium oxide film. The family tried to sue the manufacturer
for allegedly not making the sink unit out of stainless steel!
A tremendous variety of corrosion failures can lead to litigation, some
in connection with causation of an accident and others concerning product
quality and fitness for purpose. A corrosion expert is usually required to deal
with these. However, here are three good examples to illustrate the forensic
aspect when only straightforward corrosion mechanisms were involved.
6.3.1
A food processing plant was fined a large amount for causing environmental
damage due to long-standing leakage of oil from a pipe crossing a roadway.
The plant had a large storage tank on one side of the road to which tanker
deliveries were discharged, and a steel pipe carrying the oil through a duct
under the road to their heating installations. The oil was a heavy, highly
viscous variety that had to be heated in the winter months to keep it flowing
at the rate necessary to keep all the heating units working. A heating element
spiral was therefore wrapped around the pipe beneath the thermal lagging
material. The duct was a concrete channel covered with steel plates to carry
the normal road traffic.
176
Figure 6.1 Failure of a heated oil pipeline showing how corrosion follows the
line of the heating spiral.
When the environmental damage was discovered the source of the contamination was traced back to the channel under the road. Lifting the steel
plates and removing the lagging exposed a series of holes in the walls of the
pipe, following the line of the heating spiral, as shown in Figure 6.1.
A number of similar holes through the pipe wall were found and flaky
rust could be lifted off to reveal extensive areas of corrosion following the
line of the spiral. The heating element comprised a resistance wire running
through the center of the copper-covered spiral. There was no intentional
electrical connection between the spiral and the pipe, but in places the spiral
had been touching the pipe.
Given that the standard electrode potential (see Table 2.1) of iron (Fe/
Fe2+) is +0.44 V and that of copper (Cu/Cu2+) is 0.34 V, giving a corrosion
cell potential of 0.78 V, how was the electrical circuit completed? Through
the lagging material, which must have become damp and probably contaminated with brine that resulted from de-icing the roadway with salt during
the winter months. The iron was the anode (gave up electrons) with respect
to the copper (cathode) and moisture dissociation supplied the hydroxyl ions
(OH) that produced the rust. The corrosion was most severe wherever the
copper spiral made electrical contact with the iron.
There was nothing wrong with the materials or the method of heating
the pipe to keep the oil flowing. The cause of the attack was that the thermal
lagging should have been made impervious to water. Even though the duct
was drained, over a period of time storm water and debris washed from the
road would be likely to build up and block the drainage. Measures should
have been taken to keep the lagging dry and there should also have been
some thin, electrically insulating coating or wrap on the surface of the pipe
to prevent any electrical contact with the spiral. It was easily foreseeable that
a duct across a roadway would inevitably become flooded from time to time.
So the blame rested with the architect and the subcontractor who installed
the pipeline.
Fluid Transport
6.3.2
177
178
wrong with the tubing and it does offer a cheaper way of making small
diameter pipe that is ductile and easily formable compared with the more
expensive cupro-nickel alloys. The tubing starts off as a long strip of low
carbon steel of half the wall thickness of the pipe and a width of roughly
twice its circumference, with scarfed edges. This is plated with copper on
all surfaces and then roll shaped to form a double-walled tube, as indicated
in Figure 6.3. It is then passed through a furnace which melts the copper,
forming a brazed joint and also leaving a coating of copper on the outside.
It finally goes through a bath of molten lead-tin alloy to produce a corrosion-resistant coating of terne plate, rather similar to soft solder. The
tubing is produced in long coils; to form a brake pipe the required length
is cut off, fittings slid on and the ends expanded so as to make a secure
fluid tight fixing to the brake cylinder or an appropriate union.
Figure 6.4 shows a section across the wall of the brake pipe, taken
through one of the spotted areas evident to the left of the perforation.
These spots were referred to earlier as freckles, which is an apt description
as they only appear to be surface blemishes. However, the terne coating has
disappeared, and there is a cap of copper at the surface apparently following
the original contour of the pipe; beneath it, a cavity has appeared extending
almost half way through the wall. This is the result of the cell potential set
up between the iron as anode and the copper, tin and lead cathodes in the
coating. When the cavity reaches midway through the wall a further layer
of copper will become exposed, causing the pit to penetrate rapidly to the
inside wall. Figure 6.5 is a cross section through the perforation that shows
how the center layer of copper braze had lined the developing pit as it
approached the inside wall.
Fluid Transport
179
Copper-coated steel
strip with scarfed edges
Figure 6.3 Steps in the manufacturing process to make a cheap, small diameter
tube with high ductility, suitable for forming into fuel and brake pipes of motor
vehicles.
The terne coating on this short length of the pipe had been damaged by
abrasion and spray from the road wheels, or there may even have been a pin
hole in the terne metal coating that had exposed copper and possibly small
areas of iron. Once the iron had become exposed at the outside pitting, the
corrosive attack would have been very rapid. When the perforation first
appeared the brake fluid would be ejected only when the brakes were applied
and it would be squirted onto the road instead of dribbling down over the
chassis or wheel hubs. The only warning sign to the owner would be the need
to frequently top up the brake fluid reservoir.
Notwithstanding the unlikely possibility of the owner discovering the
leak, the perforated pipe should have been detected in the MoT examination
only a week before the accident. Fluid was being lost at every application of
brake pressure because within a day of purchasing the car the new owner
took it back to the dealer who immediately recognized the brake fluid reservoir was almost empty. Although the brakes were claimed to have been bled
at this time, it was by no means certain that they were or that anything at
180
Figure 6.4 Cross through a section of pipe surface pit. Notice the cap of copper
on the outside.
Figure 6.5 Cross section of pipe through spotted area. Notice thin flakes of
copper braze resting inside of the hole toward the bottom. Terne plate provides
very good corrosion resistance as long as it remains continuous over the outside
of the pipe. Unfortunately, if it is not continuous or becomes damaged so as to
locally expose a small area of iron, the subsequent rate of pitting corrosion can
be very rapid because iron is anodic to all three metals in the coatings, i.e., the
copper, lead and tin in the terne metal.
Fluid Transport
181
all was done apart from topping up the brake fluid, with the leak again going
unnoticed.
The widows claim that the brakes had been maliciously tampered with
was clearly without foundation, but the dealer and the garage who issued the
MoT certificate were prosecuted and fined heavily.
6.3.3
Instructions were to ascertain the cause of failure of a pipe clamp that split
open after a period of service on a pipe carrying pressurized hot water at
98C, allowing an escape of water that led to extensive flooding damage. A
large claim had been lodged against the firm who had installed the clamp
and the firm, in turn, alleged that the failure had nothing to do with its
fitting, but was the fault of the clamp manufacturers.
Figure 6.6 is a general view of the clamp, which is essentially a sleeve of
thin (0.7 mm) type 316 stainless steel with a T junction, fitted to the pipe
like a saddle, tightened by bolts passing through end pieces. It is lined with
a soft gasket to make a water-tight seal. The T branch projecting from the
side is not an off-take but rather it allows monitoring devices to be inserted
without interfering with the flow of water. The clamp bolts were still tight
at their original positions.
In what direction would the greatest tensile stress be acting when the
clamp was attached to a pipe? It is the hoop stress, normal to the axis of the
pipe.
The stainless steel had split from top to bottom, releasing the clamping
force and allowing the water to force its way under the gasket and escape to
outside. The bottom end of the split is shown in Figure 6.7A and at the top
182
Figure 6.7 Closer views of the split. (A) Bottom of split. (B) Top of split showing
PVC tape between the gasket and stainless steel.
Fluid Transport
183
applications. The manufacturers claimed that PVC tape was used in all of
these and no service failures had been reported to them. If this is true then
it would suggest that the water in those applications was not as hot as that
in the pipe where the failure occurred or it is also possible that the clamps
were not done up so tightly.
Many alloys are susceptible to this type of attack, some of which otherwise have excellent corrosion resistance. It occurs when a high level of tensile
stress exists or acts upon a component where there are traces of some specific
chemical ion in the environment. In the absence of the particular ion or
when the stress level does not exceed a particular threshold, there is no attack.
For stainless steels the troublesome ion is chloride. The stress in this instance
was derived from the clamps pulling the sleeve tight around the pipe, producing a tensile hoop stress. Cracks grow at right angles to the tensile stress,
so this is why they all tended to lie parallel to the axis of the sleeve. The time
taken to complete failure depends on the level of stress, the concentration of
the aggressive agent, the wall thickness and the temperature where the process
is taking place.
The source of the chemical ions was the strip of PVC tape used to
temporarily locate the gasket inside the sleeve at the time of fabrication by
the manufacturer. It served no purpose in the design or function of sealing
action of the gasket. PVC degrades thermally at temperatures above about
72C, one of the products being HCl (hydrochloric acid). The brittleness of
the tape inside this sleeve shows that it had degraded under the service
conditions on the hot water pipe. The chloride ions were thus present at the
interface of the tape and the stainless steel, which allowed the stress to
produce the stress corrosion cracks.
It would be only a matter of time in service on the hot water pipe before
the failure took place. It is unlikely that routine inspection would have
revealed the deteriorating state of the metal. At best all that might have been
detected on the outside of the sleeve would be microscopically small cracks
shortly before the catastrophic rupture took place.
This failure was clearly the responsibility of the manufacturer, who used
PVC tape at the time the sleeve was assembled. It was not caused by the
silicone gasket material, the adhesive, or the way the sleeve had been attached
to the pipe. It was foreseeable that close contact at moderately elevated
temperatures between PVC and type 316 stainless steel was likely to result in
stress corrosion cracking. The problem with keeping the gasket in alignment
during assembly could have been easily overcome in other ways, for example,
using mechanical means or a jig. The failure would probably not have
occurred, or would have taken very much longer to do so, if the pipe had
not been carrying such hot water. Use of the PVC at temperatures approaching 100C caused the rapid deterioration of the PVC.
184
The same form of stress corrosion cracking that appears in the sleeve
occurred in a fatal accident that involved failed stainless steel bolts from
one of the early PVC manufacturing plants. The chloride ion on its own
is a common cause of stress corrosion for some types of stainless steels, for
example, fittings on ocean-going boats when an inappropriate grade of
stainless steel for contact with salt water has been used. There was another
case in which hollow spring pins used for hinges in a kidney dialysis
machine failed by stress corrosion resulting from the salt solution used for
cleaning the machine after dialysis.
Both a corrosive agent and a high level of tensile stress are necessary for
stress corrosion. The stress either can be externally applied as in this instance
or, quite frequently, it can be internal residual stress from a cold forming
process. The mechanism of stress corrosion cracking was, in fact, first studied
more than 100 years ago when brass cartridge cases were found to be cracking
in humid monsoon conditions when traces of ammonia were in the atmosphere. The remedy was simply for the manufacturer to subject the (empty)
cases to a low temperature stress relief anneal after the deep drawing process,
which reduced the residual stress to a safe level.1
Fluid Transport
185
Tap
Hot
Cold
Bench
Upper Hot Valve
Spindle
Spacer Support
Bracket
(3-mm-thick sheet
steel)
Lower Cold Valve
To Cold Tap
Partial Fractures
of Plastic Joint
Cold Supply
to Heater
Cupboard
Cold
Feed
Single Screw
Attachment
to Wall
Hot Feed
from Heater
Cold Water
Mains
Water Heater
Floor
Figure 6.8 Pipework with direct cold water feed to tap. Hot water is fed from
a water heater below the sink via a three-way junction, where the failure occurred.
The heavy water heater was attached to the wall by only a single screw,
which had come dislodged from the wall, and placed the surrounding copper
pipes under bending loads (Figure 6.8). The break in the plastic fitting was
just below the lower bend in the steel bracket, where the rising main supplied
incoming cold water to the system. Since the bracket was substantial, any
extra load would have been supported by it rather than being transmitted to
the plastic junction (Figure 6.9).
186
Bench Top
Brass Coupling
Plastic
Top Valve
Tap Spindle
Spacer Support
Bracket Not
Intended to be
Fixed Other Than
to Valves
Plastic Bottom
Cold Valve
Cold Supply
to Heater
2 Fiber Washers
PTFE Tape
Brass Union
Boss White
Raising
Main
Brass Coupling
for Plumber
Figure 6.9 Details of the junction assembly showing the lower acetal resin
thread where the fracture occurred.
6.4.1
Clearly, there must have been another reason for the failure. The next suggestion was that the junction had been overtightened by the plumber. The
system had been fitted during renovation work 4 years earlier. Now this is
an interesting suggestion, because it seems to imply that the fitting could be
overtightened, and might thus have been a faulty design in allowing overtightening. Nevertheless, overtightening is of course a common way of producing extra stress on a joint, but what did the evidence show? If screw joints
are overtightened, there might be supporting evidence from the fiber washers
used in the system. They bear any excess stress at the end of tightening, so
if the joint had been overtightened, one might expect some damage to them.
In fact, there was no evidence that the washers had broken or distorted
excessively at all (Figure 6.10). But this is where the problem lay when it came
Fluid Transport
187
Figure 6.10 The washers on the failed joint showed little permanent set or
cracking, so it was unlikely that they had been overtightened during fitment.
to assessing who would pay for the very substantial damages to the computers
below the flood. The loss adjusters expert blamed overtightening as the cause
of failure, possibly exacerbated by the faulty screw attachment to the wall.
Thus the loss could be recovered from the plumber or company who installed
the system, and possibly the architect for planning the system in the first
place.
The plumbing company and the architect (or rather their insurers)
refused to accept the blame, and so proceedings were issued by the University
against them. At this point, the defendants naturally needed their own expert
to examine the failure and produce a report (preferably someone with expertise in the failure of plastics materials). The failed acetal fitting was critical,
but had not been examined in detail by the first expert. Why not? The reason
or reasons are not known. It seems so obvious that the evidence at the heart
of a failure, the cracked fitting in this instance, should be subjected to detailed
analysis. However, some experts fail, for one reason or another, to make the
leap into detailed inspection, perhaps because they do not have the expertise
or laboratory equipment, or perhaps because of casework pressure.
6.4.2
The fracture surface of the acetal fitting is shown in Figure 6.11 with an
interpretative map in Figure 6.12. The fracture surface showed several very
old fracture origins, now concealed by debris deposited from the water supply, so inferred from flaps left by crack growth (Figure 6.13). The detail shown
in the latter figure suggests that there were at least five different growth
regions, with crack growth directions as indicated by the bold arrows. The
188
Figure 6.11 Side view of the broken fitting with the brittle fracture surface at
right.
Sprue
(Injection Point)
Mold Parting Line
5
Thread Apex
Origin
Fresh Fracture
Old
Fracture
Under
Deep
Discoloration
Zones
Old Fracture
Surface at
Base of Thread
1
Notches in Weld Lines
Old Fracture Under
Sequence of Inner
Flaps All Oriented
in One Direction
A
2
Figure 6.12 Fracture surface map to show multiple crack initiation and slow
crack growth.
white areas represent the fresh fracture induced by the loss adjusters expert
when the sample rolled off his desk! When the areas adjacent to the fracture
surface were examined in detail, one significant feature emerged: there were
numerous subcritical cracks that appeared to have grown from weld lines in
the threads and on the inner surface of the fitting (Figure 6.14). The weld
lines appeared to be associated with flow lines in the outer surface of the
molding, suggesting that the molding may have been faulty. Flow and weld
Fluid Transport
189
Figure 6.13 Close-up of side of the fracture surface with flaps due to crack
branching, from which crack growth direction could be inferred.
Figure 6.14 Close-up of the crack tip showing weld lines ahead of the crack.
lines are indicative of cool tools, so that flow of molten polymer in the mold
tool is inhibited. The fact that the fitting had been screwed tightly to produce
a closure stress on the threads was indisputable as Figure 6.15 shows. So what
effect would such weld lines have when the thread was screwed tightly by the
plumber?
6.4.3
Water Pressure
Tiny weld lines could concentrate the tightening load if any were situated at
the root of a thread. One common observation of failures from screw threads
is that failure is often situated at the next-to-last thread, a feature confirmed
by theoretical analysis of the stress-concentrating effects of thread forms.
Thread failures are themselves rather common simply because threaded joints
and connections are some of the most common ways of joining components
together. However, there was no evidence of an excessive closure stress made
by the plumber when finally tightening the joint, and the cause of the failure
190
remained unknown at this stage of the investigation. What was clear was that
there was no evidence supporting the proposal that the joint had been overtightened (and that therefore the plumber was to blame).
One additional factor is the internal pressure from the water supply.
Although it had not been measured directly, the specification of the fitting
allowed for a pressure of up to 25 m head of water. The pressure will, of
course, impose a hoop stress on the wall of the fitting, but what magnitude
would it be? The pressure can be calculated using the simple equation:
P = rhg = 1000 25 9.81 Pa or Nm2
So that
P ~ 0.25 106 = 0.25 MPa
Fluid Transport
191
The hoop stress (the largest stress acting on the pipe wall from the water
pressure) in the system is then simply given by
sH = PR/t = 0.25 19/1.6 ~ 3 MN m2
This stress is small compared with the normal tensile strength of acetal
material, of about 70 MN m2. Although molding would reduce the strength
somewhat, it was difficult to see how the combined effects of internal pressure
and screw thread stress raiser could initiate crack growth and, ultimately,
catastrophic failure.
6.4.4
Molding Features
What could have created the weld lines in stressed areas of the molding? A
detailed survey was made of new moldings supplied by the manufacturer.
The moldings were smeared with graphite powder to highlight the flow
and weld lines, and it was determined that there were considerable variations between the 12 moldings examined (6 of which are compared in
Figure 6.16). It was concluded that there was a pattern to the markings,
especially in the flow lines. The set had probably been made in two batches,
each producing a different flow line pattern. Weld lines did exist but were
much less serious than in the failed sample. Moreover, the failed sample
showed weld lines where none could be found on any of the new moldings,
192
especially in the critical threaded areas of the lower inlet pipe. The failed
sample had thus been made in a different batch, and may even have been
a maverick sample.
6.4.5
One failure mechanism that has been known for many years is environmental
stress cracking (ESC). Certain fluids can cause brittle cracking of plastic
products, even at very low imposed stresses. One of the first examples to be
discovered occurred in polyethylene exposed to very strong detergents (such
as an ionic soap known as Igepal). Stress corrosion cracking (SCC), where
the fluid interacts chemically with the polymer surfaces to which it is exposed,
was also an alternative cause of failure. In order for attack to commence in
either form, there must be a driving force to open the crack. It can take two
main forms:
1. Small stresses or strains imposed on the system
2. Frozen-in strain (oriented chain molecules)
It is also known that the more aggressive the attacking agent, the lower the
critical strain needed to grow cracks.
6.4.6
Acetal Polymer
Fluid Transport
6.4.7
193
A U.S. Problem
While the various expert reports were being digested by the several parties
now in the action (university, plumbers, architects), we happened to come
across a report in a technical newspaper that suggested that there could
indeed be a problem with chlorine attack from the water supply. A news
item in a U.S. journal3 reported the settlement of a case (prior to trial)
involving domestic hot water supply pipes, where polybutylene (also
known as polybutene) pipes had been equipped with acetal copolymer
fittings. The fittings had failed and caused floods, and the paper actually
reported a class action, where many affected families sued the manufacturers and suppliers of the materials. Both the pipe and fittings had suffered extensive cracking on their inner surfaces, a problem expert opinion
considered to be caused by chlorine and also probably by dissolved oxygen
in the water.
A further class action in Texas came to trial in 1992, and resulted in a
large settlement for the numerous plaintiffs (Babb v. U.S. Brass, Shell, HoechstCelanese, DuPont). It was this case that attracted our attention (acting on
behalf of the plumbers in the U.K. case). Contact by the defendants attorney
with the plaintiff s attorneys produced not only copies of the expert reports
in the case, but also the transcripts of the trial itself (when the expert evidence
is tested in public). It was clear from those expert reports that even low levels
of chlorine down to 0.3 ppm (parts per million) in the water could cause
serious stress cracking of acetal copolymer.4 Such levels can occur in drinking
water, so it seemed as though the problem of the cracked acetal fitting may
have been caused directly by chlorine. A check with the local water board
produced extensive analyses, both from 1993 and 1988 (when the fitting
failed). The records showed that free chlorine was very variable at different
sampling points, but could rise as high as 0.9 ppm. One reason for high levels
was given as the slugs of chlorine added when there were breaks in the
pipes as a precautionary measure to prevent ingress of contamination.
A direct experiment was then carried out on the failed fitting using the
EDAX facility of the scanning microscope. Analysis of the surface showed
the presence of chlorine in significant quantities. Presentation of the new
evidence to the other parties in the dispute produced a rapid settlement of
the action, with the parties all walking away (and bearing their own costs).
The U.S. actions were appealed, but the appeal was lost in the Texas Court
of Appeal, and the many plaintiffs received full restitution of their damages.
This was incidentally one of the largest class actions (outside automobile
actions) in U.S. legal history and ultimately cost the companies many millions
of dollars in damages awarded and the costs of the actions.
194
6.4.8
Aftermath
Fluid Transport
195
Injection molding
of valve body
Date unknown
Date unknown
Closure stress
applied to main
thread; cracks
initiated
Date unknown
Creep rupture
of main thread
Total assembly
fitted to pipework
by plumber
Fracture of
final thread
Catastrophic failure of
valve thread, possibly
caused by water pressure
surge; serious flood in
Physics Dept.
occurred in the late 1970s, and the incidents raised issues of consumer safety,
quite apart from the warranty damage involved. One such fire occurred when
a new car owned by the British actor Sir John Gielgud was traveling along
the Embankment in London. Smoke was seen coming from the engine compartment, and although the fire was extinguished, substantial damaged had
been caused (Figure 6.18). Examination of other fire-damaged cars by alert
insurance inspectors showed that the fires were probably caused by a leaking
fuel pipe near the engine.
The case was apparently solved with a recall, but reappeared in the late
1990s in a case before the High Court in Dublin. It raises the problem of
dissemination of failure studies, especially in an international market where
identical or very similar products are supplied to different national markets.
It also focuses on the importance of quality control and assessment during
product design and development, especially where consumer safety is paramount.
196
Figure 6.18 Ozone cracking in NBR rubber fuel pipe. The cracks have not yet
grown to criticality.
6.5.1
Figure 6.19 The engine compartment of Gielguds car showing the smooth
rubber fuel intake pipe at T (below center) and the nylon return pipe at O (above
center). The fire started at O.
Fluid Transport
197
Figure 6.20 Abrasion marks on nylon return pipe where it came into contact
with the hot metal of the engine.
penetrate into the interior of the tubing, resulting in loss of gasoline under
pressure from the fuel pump. As soon as the liquid hits the hot engine
manifold, it vaporizes, and any small sparks from the electrical circuitry can
initiate an explosion followed by fire. In addition, a problem with another
part of the fuel system revealed that a thermoplastic material had been used
for the return pipe that passed over the engine, returning unused fuel to the
gasoline tank. Due to its poor abrasion resistance, this pipe could also leak
if any part contacted the engine manifold, where abrasion caused a hole to
develop (Figure 6.20, aa). The pipe was an extrusion of nylon 11, and
possessed a melting point of about 176C.
Since several cars had been damaged by such fires, it seemed clear that
both unprotected NBR fuel pipes and nylon return pipes had been fitted
widely, and so a recall of the model was carried out.
6.5.2
A Problem in Eire
The recall was carried out in the early 1980s, so it came as some surprise
when the original investigator was approached by lawyers involved in a High
Court claim for damages in the early 1990s. A serious fire in the Irish Republic
in the late 1980s in a rather old Fiat Mirafiori had caused serious injuries to
two young children who were strapped in the rear of their mothers car. She
had been shopping, and upon arriving back home, had removed the ignition
keys to open her front door. As she turned back to remove her children, the
car interior burst into flames. Fortunately, the children were rescued by
neighbors, but not before they had suffered serious burns. The resulting court
action claimed substantial damages for their continuing care, as well as for
the pain and suffering.
198
The accident was unusual in that the fire had occurred in the interior of
the car rather than the engine compartment (where the original problem had
been discovered). So what was the likely cause? Unfortunately, most of the
evidence had been lost in the fire, because the car was nothing more than a
burnt-out shell when it was examined later. There were two possible causes: a
faulty heater near the dashboard, and the plastic return pipe to the gasoline
tank at the rear of the vehicle. The seats were filled with highly inflammable
polymer foam, but the rapidity of the fire suggested a gasoline leak and ignition
from a spark. Ignition could have been from the faulty heater although it is
well known that static sparks can easily ignite gas or fumes. The fuel return
pipe passed through the interior of the car, and could have leaked gasoline just
before the fire. It then only needed a small spark, perhaps when the woman
left the car or closed the door, to ignite the fumes. A small fire became an
inferno as the liquid gasoline caught fire, and the fire grew into the seats. Such
a scenario seemed feasible, but lacked corroboration in detail.
6.5.3
By the time the investigators original report was requested, the case had
reached an impasse. In personal injury actions, the plaintiff often faces the
problem of discovering details of manufacturing processes, QC procedures
and previous incidents, from the defendant companies. The problem of
discovery from Fiat Spa in Ireland had led to a striking-out claim against
the company, because Fiat either could not or would not release the required
information to the claimants solicitors. The latter thus pursued the information independently, and came across the public information of the recall
in the early 1980s. The trail led to the insurance company, which had unfortunately deleted the original independent report on which the recall was
based. However, the insurance company did have a record that the report
had been commissioned, so it was able to make contact with the original
investigator. Fortunately, the investigator had retained all samples and reports
from the first accidents, and was able to supply them directly to the legal
team in Ireland. The report was supplied as an affidavit to the Court, and
the defendant witnesses (heads of litigation and engineering from Italy) were
cross examined on the evidence of the fuel line problem.
Their answers were rather evasive and lacked credibility this was somewhat surprising because they had some time to consider the questions in
court, since they were provided with a translator. On the flight from London
the previous day, the investigator heard them speaking fluent English with
their attorney! The claimants succeeded in their action, but the case went to
appeal on the legal principle of striking out the defense. It effectively meant
that Fiat could not succeed when and if the main claims went to full trial.
Although Fiat won the appeal, the claim was settled before trial with a
substantial award to the injured children.
Fluid Transport
6.5.4
199
Aftermath
Due to the extensive publicity in Ireland,7 the case revealed a series of fires
in this model, and further claims were lodged after the successful settlement
of the Murphy children. In hindsight, it seems remarkable that greater care
was not taken during testing of the new model. The radiator case in Chapter
7 shows how new car models are now driven in realistic conditions before
launch onto the wider market, so as to reveal any latent or hidden defects
that are not readily apparent to the design team. This is why FMEA methods
are so important during development, and why it is important to know what
previous accidents have occurred in such simple and low-cost items as fuel
pipes. We followed up the original study by examining fires in new Ford
Cortinas, for example. Car fires continue to plague many new models, perhaps because the engine compartment is so vulnerable. However, reinforced
fuel hose made from resistant and strong materials is readily available to
designers, and there is no reason why such hose should not be specified
during the design phase. Elastomeric hose that is also ozone resistant is also
widely available, so this particular problem is not insurmountable. Many
synthetic rubbers, which do not possess vulnerable double bonds, have been
made available commercially within the past couple of decades. Car fires are
particularly distressing in their effects, and the wider problem was recently
highlighted in the U.S., when the largest ever damages were awarded by a
jury in Los Angeles.
The case study of the Fiat Mirafiori illustrates the importance of cooperation between regulatory bodies in different countries, so that they are all
aware of design problems with specific products and can take appropriate
action. It also illustrates why it is so important that the basic engineering
message of choosing components fit for their function is taken seriously so
that problems do not recur, and the consumer does not suffer as a consequence of pure ignorance by the design team and the regulatory bodies. An
incidental issue is one of preservation of records and samples, so that the
lessons of particular investigations are not lost forever. Indeed, publication
achieves just this purpose, because then the information is recorded permanently for posterity.
200
Figure 6.21 The failed Y molding (arrow) plus connecting pipe (at right) and
repair tube as received from the Forensic Science Service.
opposite lane. The car collided with an oncoming truck, causing serious
injuries to the driver. She subsequently spent considerable time in intensive
care. The police followed the trail of diesel fuel some 5 miles to a local golf
club, where the garage recovery vehicle was found. The spill had caused
several other accidents, although none so serious as the first. A leak had
occurred in the diesel return pipe, and the vehicle was quickly repaired. The
remains (Figure 6.21) were examined initially by Strathclyde Police Forensic
Labs in Edinburgh, which concluded that the line had been cut by a vandal.
The case was referred to us to re-examine the remains, because the part
concerned was extremely difficult to access in the engine compartment, and
was thus unlikely to have been vandalized.
6.6.1
The parts of the pipe removed from the vehicle showed that the pipe itself
had not failed, but rather a plastic Y-connection linking three pipes together
(Figure 6.22). FTIR and DSC analysis showed that the connection had been
injection molded from nylon 6,6 while the pipe itself was a nylon 12 extrudate. Although the break was flat and smooth, thus appearing cut, there were
no traces of other failed attempts near the break. It is a difficult part to cut
cleanly at one stroke, and why would a vandal try to cut the harder part?
Traces of spray paint on the pipe and junction showed the parts to be original
fittings (Figure 6.22). An intact return pipe from a separate wreck of a similar
vehicle showed that the failure had occurred near or at the junction between
the pipe and the nylon 6,6 fitting (Figure 6.21). Comparison also showed
abrasion marks on the pipes at similar positions (lower arrows in
Fluid Transport
201
Figure 6.22 Comparison of failed and intact fuel lines from diesel recovery
vehicles. The upper arrow shows one of the three junctions and the lower arrows
show abrasion marks where the pipes made contact with hot engine parts.
Figure 6.23 The failure surface from the nylon 6,6 molding showing eroded
edge (chamfer, Ch) at the lower right, several crack arrest lines (arrows) and a
cusp at top left of the sample (C).
Figure 6.22), a problem caused by wear against hot metal parts of the engine,
itself a serious failure mode in car fires, as discussed in the previous section.
The failed surface showed several features that gave vital clues as to the
way the junction had failed in service (Figure 6.23). Part of the outer edge
showed signs of erosion in the form of a chamfer (Ch), followed by a smooth
region and a set of crack arrest arcs (arrows). The final part of the surface
showed a cusp of material rising above the plane of the surface (C). Detailed
optical inspection revealed at least 11 crack arrest arcs, as shown by the lines
202
Final Cusp
Ductile
Tear
Crack Arrest
Fronts
Growth
Crack
O
Crack
Growth
Erosion
Trough
Erosion
Zone
Figure 6.24 Fracture surface map showing detailed interpretation of crack features, with crack growth from eroded edge to final cusp.
in Figure 6.24, the fracture map of the surface. However several of the arcs
represented single cracks separated by the central bore, and there were only
seven major crack growth events. In the same study, some much smaller
cracks could be seen growing parallel to the long axis of the tube. Interpretation suggested that the sample failed from an origin (O) in the eroded edge
of the junction and grew in a series of steps to the final cusp. But what had
initiated the sequence of events? The junction was just below the battery of
the vehicle, and a possible cause of the erosion may have been attack by
spilled battery acid.
6.6.2
It is well known that nylon 6,6 is sensitive to attack by acids, especially sulfuric
acid,8 as indicated by Figure 6.25. A series of experiments with 40% sulfuric
acid (battery acid) showed that a new junction failed suddenly by brittle
cracking when exposed to acid, although only after several hours exposure
under a very light self-load. The nylon 12 material of the pipe itself remained
quite unaffected by the same fluid. It thus seemed feasible that acid had
dripped onto the junction and attacked the edge. This degraded the material
and allowed a crack to grow from the inner corner of the junction edge,
growth which stopped at the first crack arrest line more than halfway down
Fluid Transport
203
40C (T)
80
60C (T)
Tenacity (%)
100
60
25C
(N)
40
20
40C
(N)
60C
(N)
15
20
25
10
Time of Exposure (Weeks)
30
35
Figure 6.25 Loss of tenacity of Terylene PET fibers (T) and nylon 6,6 fibers (N)
on continuous exposure to 10% sulfuric acid at various temperatures.2
Figure 6.26 ESEM image of the final cusp on the fracture surface, with inner
bore at the top of the picture. Crack growth direction is shown by arrows. Several
major and minor striations are shown by the numbers at right.
204
loads. EDAX analysis showed traces of sulfur on the eroded part of the
surface, supporting the acid attack hypothesis. However, the traces of sulfur
disappeared further into the surface, suggesting that the crack grew mainly
by simple strain of the pipe system later in the failure. So what loads led to
crack growth?
6.6.3
The return pipe was not clamped to the engine manifold, but rather lay freely
against it, as shown by the abrasion zone of Figure 6.22. Ties are normally
available to prevent contact, but the present pipe examined appeared to have
lost its ties some time ago. Flexible fuel pipes can move freely against rigid
parts of the vehicle chassis, movement occurring when the engine is turned
on and thereafter with driving. So what did the major and lesser striations
represent in terms of the history of the vehicle? It was possible to suggest that
the major crack arrest lines (Figure 6.24) represented start-up of the engine
from cold, while the minor striations represented engine acceleration/deceleration during driving. Diesel fuel would thus have been leaking under low
pressures (about 48 psi) from the return pipe with the engine running and
the junction partially severed. The fuel lubricated points of abrasion, and
reduced the load on the failing junction while the vehicle was moving. However, when stopped overnight, the diesel would have evaporated away, leaving
a dry junction by the time the vehicle was started up again the next morning.
Friction at such junctions would have been greatest during start-up. Such
leakage increased with time as the hinged crack grew, with final separation
just before the major spillage onto the road near Dundee.
6.6.4
Conclusions
The most likely explanation of the accident was therefore as follows. A small
leak of battery acid occurred during routine maintenance of the recovery
vehicle, with acid falling onto the nylon junction immediately below the
battery. Chemical erosion of the edge led to stress corrosion cracking of the
junction (Figure 6.27). One crack grew radially across the section, major
growth occurring at engine start-up. The seven such major crack arrest lines
represent seven engine start-ups from cold. Since garage vehicles are generally
kept running through the day, each crack arrest front probably represents
growth in 1 day, so the total sequence lasted about a week. Diesel fuel would
have leaked slowly down the chassis and onto the nearby road, and would
have been visible to a vigilant driver. Regulations in the U.K. stipulate daily
checking of commericial vehicles for potential hazards, for example. Diesel
fuel leaves a very slippery deposit on the driving surface, when the lighter
Fluid Transport
205
Battery
A & D Leak
to First Joint
Diesel
Spray
onto
Pipe
Stage 1
Stage 2
Leak to
Ground
Figure 6.27 Failure mechanism of the pipe junction from SCC of nylon 6,6 by
leakage of battery acid.
fractions of the liquid have evaporated away after a few minutes, equivalent
to driving on black ice. The leak from the breakdown vehicle caused a total
of four accidents that day, the most serious of which was experienced by the
claimant in this action. Liability was denied by the insurers for the garage
who owned the vehicle, but the injured woman driver received substantial
damages for her very serious injuries. The new molded junction tested in
our labs turned out on analysis to be acetal rather than nylon, a salutary
reminder of the sensitivity of many engineering plastics to common chemicals found in car engines. Designers should be aware of the problem of stress
corrosion cracking, and recommend appropriate materials for the function
of the component parts of safety-critical products. Fuel lines are especially
critical parts of engines, and the best possible materials are chosen for their
construction. The consequences of poor design and maintenance of critical
car parts can be devastating for the users.9
206
Figure 6.28 Front view of the failed radiator with rust damage at the upper
left air vent plug, and blanking plug at the lower right.
Figure 6.29 Plugs from the Stratford radiator showing brittle radial cracks.
At the time we became involved in the case, other investigators had already
studied the extensive failures found in a wide variety of properties, with old
peoples homes and schools among the most common properties affected.
Some domestic houses had also been affected. Plug and vent seals in thermoplastic polyester-polyether had failed and the leaking water had damaged floors
and fabrics. The washers had been introduced in some cases only a few months
before failing, the original intent being to improve washer lifetime (compared
with older fiber washers). Failures were recorded from 1994 to 1996 before
serious attention was paid to the problem. The engineer acting for the loss
adjusters extracted several radiators (Figure 6.28) and examined the seals
(Figure 6.29), concluding that the plugs had been fitted either too tightly or
too loosely by the plumbers. This seemed unusual as many different plumbers
had been involved, suggesting widespread incompetence at performing a very
Fluid Transport
207
Figure 6.30 Failed washer showing several subcritical radial cracks (arrows).
Figure 6.31 Same washer from Figure 6.30 rephotographed from engineers
report showing identity of crack paths.
simple task. The engineers report to the insurers stated that one of the washers
had broken during extraction from the radiator (Figure 6.29). Moreover, one
of the photographs shown in that report exhibited traces of several radial cracks
when inspected under a magnifier (Figure 6.30). Although the breakage had
been mentioned in the report, the subcritical microcracks had been ignored.
The state of the same seal was confirmed by rephotographing the picture in
208
Figure 6.32 Blanking plug from Benenden School with failed washer and evidence of extrusion on the outer side in the section at left (arrow).
the original report (Figure 6.31). Between our inspection and the original
exmaination, a period of many months, there had been no change in the state
of the seal. The conclusions of the initial report, therefore, effectively ignored
vital evidence about the actual cause of the failures.
6.7.1
Failed Seals
The failed radiator taken from the Stratford old peoples home (Figure 6.28)
displayed rusting at the upper vent plug, showing the effects of long-term
leakage. The plugs are shown in Figure 6.29, each with a single radial crack.
Closer examination revealed many fine subcritical radial cracks visible as
faint black lines against the red of the material (Figure 6.30), confirming their
existence at the time of extraction (Figure 6.31). Other failed washers showed
extrusion at the inner and outer edges, suggesting permanent set from exposure under compression loading when fitted (Figure 6.32 and Figure 6.33).
The fracture surfaces showed crack growth from edge defects in the inner
surface of the washers where they made contact with the steel of the radiator
(Figure 34). Some of the cracks in the many washers examined showed the
faint striations characteristic of fatigue growth (Figure 6.35). Other samples
showed fast crack growth paths, probably induced when removing the washers.
6.7.2
Experimental Analysis
Fluid Transport
209
Figure 6.33 Air vent plug with cracked washer, showing traces of paint on the
seal. Evidence of extrusion on the inner side is shown by the arrow.
Figure 6.34 Section of fractured washer with origin at upper right corner.
gradually increasing heats of fusion, from 32.3 mJ/mg for raw granules, 36.2
mJ/mg for new washers and 43.9 mJ/mg for failed washers.
Since the failed washers were all much stiffer than new washers, trial
experiments were organized by an expert working for RAPRA in the U.K.,
to measure the set induced by exposure to hot water.10 Five stacks of four
new washers were compressed between steel plates, and washers fitted to
210
Figure 6.35 Close-up of lower right corner showing fatigue striations, probably
produced by thermal cycling.
Discussion
Since the change in material properties observed on the failed washers could
be reproduced experimentally, it was concluded that exposure to heat caused
a critical loss of sealing ability of the material used in the injection molded
washers. The temperatures to which they were exposed in service were not
measured directly, but probably lay within the range 75 to 85C, the higher
temperatures being primarily used in old peoples homes. CHS in those
locations also tend to have longer hot periods, so the time to failure will have
been lower, assuming a linear rate of material hardening. The loss of sealing
capacity thus could be explained by material changes, probably by enhanced
crystallization of the material at high water temperatures. 11 But how could
Fluid Transport
211
Temperature, F [C]
122
[50]
77
[25]
158
[70]
212
250
[100] [120]
100 000
72D HYTREL
10 000
Years
20
72D HYTREL (90 parts)
+ HYTREL 10 MS (10 parts)
10
5
1 000
2
1
100
72D HYTREL
Alone
10
3.4
3.2
3.0
2.8
2.6
2.4
Figure 6.36 Hydrolytic stability of Hytrel 72 with and without added stabilizer
as a function of time and temperature (DuPont).
cracking be explained? It is known from the commercial literature that thermoplastic polyether-ester elastomers are susceptible to hydrolytic degradation, stabilizers being available from the manufacturers.12 Figure 6.36 shows
the change in tensile product half-life as a function of temperature for a hard
grade of Hytrel (72D), with and without added stabilizer (known as MS). It
is clear from the lower curve that unstabilized Hytrel hydrolyzes fairly rapidly
even at 60C, having a half-life of only 100 days continuous exposure. There
was no evidence that stabilizer had been added to the polymer washers.
More recent commercial literature13 suggests that no grades of Hytrel
are suitable for exposure to water temperatures greater than 50C, so the
radiator manufacturers should have been advised not to use the material
as seals where temperatures were likely to be greater than this upper limit.
The injection molders had previously developed washers for hot water taps
in domestic supply systems successfully, but the temperatures of tap hot
water are usually much lower, and exposure time is very much shorter than
in CHS radiators. Washer breakage was probably caused directly by stress
corrosion cracking, by hydrolysis, effectively confined to the crack tips.
212
Conclusions
The new material should have been tested thoroughly under expected service
conditions before introduction onto the market, a conclusion agreed to by
all the experts at a pretrial meeting. Ultimately, the substantial damages to
property were paid by the insurers without a trial. The large production of
several million unused washers was scrapped, and a new design using EPDM
elastomer substituted. The case emphasizes the need to examine evidence
objectively, and not to jump to hasty conclusions. If the first investigators get
it wrong, the litigation machine is started and rolls blithely onward until
halted by subsequent reinvestigations. Meanwhile, substantial costs have
accrued, which must be paid by the losing side, usually the insurers. It is one
thing to perform the research and development with new materials, and then
ignore them (as in the case of the acetal fittings), and another thing not to
perform routine tests at all.
References
1. Grimstone, F.S., Season-cracking of small-arms cartridge cases during manufacture, J. Inst. Metals, 39, 255278, 1928.
2. Vogl, O., Polyaldehydes, Edward Arnold/Marcel Dekker, London, 1967.
3. Anon., Plastic pipe is expensive for industry, Chem. Rep., March 18, 1991.
4. Armstrong, James and Duvall, Plaintiffs Exhibits in Chris and Diane Babb v.
Shell Chemical Co. et al., Matagorda County Court, Texas, 1992.
5. Lewis, P.R., Designing with Plastics, RAPRA-REVIEW Reports No. 64, 1993;
e.g., Zhou, W., Chudnovsky, A. and Niu, X. SPE ANTEC 96, 3265.
6. Lewis, P.R., Degradation of an acetal plumbing fitting by chlorine, oral presentation, Fapsig session at ANTEC 2000, Orlando, 2000.
7. Anon., Court asked to strike out Fiat firms defence in car fire case, Irish
Times, October 12, 1994.
8. ICI Industrial Fibres Manual, 1981.
9. Lewis, P.R., Stress corrosion cracking of nylon 6,6, Fapsig session of ANTEC
2002, San Francisco, 2002.
10. Lewis, P.R. and Brown, R.P., Failure of TPE washers in central heating systems,
Fapsig session of ANTEC 2002, San Francisco, May 2002.
Fluid Transport
213
Failure of
Storage Vessels
This chapter deals with the problems of storage vessels, which must be
correctly designed in order to perform their function of containment of a
fluid safely over a reasonable lifetime. The design problem is of course very
similar to that of pipes, as described in Chapter 6. The consequences of failure
can be catastrophic if the stored contents are released into the environment,
as occurred in some of the examples described in this chapter. Historically,
a variety of materials have been used to store fluids but the choice nowadays
is narrow, with metals and plastics the main contenders. Large glass vessels,
for example, were once very common, but they have been displaced by
tougher materials capable of resisting imposed loads safely. Materials and
manufacturing costs limit the choice further for mass-produced products
such as domestic water storage tanks, where polypropylene has largely displaced galvanized steel. Widely different shapes can be produced quickly in
plastics by injection molding or rotational molding, although manufacturing
defects must be eliminated by effective quality control to prevent failure when
the tank is put into service. Galvanized tanks suffer corrosion problems in
time or through faulty installation, and can have equally dire consequences
when failure occurs suddenly.
Composite materials have been widely used in specialty applications,
such as header tanks for car radiators, the basis of the second case study in
this chapter. Thermoplastic tanks have been in use in industry for a long
period of time, at least from the early 1970s, and continue to expand in usage
due to the ease of fabrication, the excellent resistance to corrosive chemicals
and the low cost of the material. The speciality tanks are made individually
to order by hand and are constructed with fairly simple equipment, so manufacturing costs are low. The increased usage has encouraged smaller companies to enter the field to provide competitive pricing, but the only standard
215
216
covering their design, use and testing is German, and apparently not generally
available to potential users.
Plastics materials are universal in specialist areas such as battery containers, due to their insulating properties and ease of shaping. However, problems
have arisen when first introduced into safety-critical products such as miners
lamps. Other problems have been caused during installation, as a case involving the Hong Kong transit system describes. Another problem arose when
traction batteries for fork lift trucks failed prematurely due to UV degradation. Because thermoplastics can be shaped so easily, they are extensively used
for domestic bathtubs, but problems can arise if the material is not capable
of resisting applied loads. Two case studies show what can happen if either
the material or the shaping process are not carefully monitored. Larger storage tanks are susceptible to much larger loads in service, and the consequences of failure are dire if the design cannot resist either external or internal
loads. As in the last chapter, the failure of a conventional metal product is
presented first.
Tin 0.49 V
217
The tank was a header tank for a heating system in a government laboratory
in the U.K., and it was topped up from the public water supply when evaporation and loss from the system demanded. The tank was set high in the
roof of the building at the gable end, double glazed from top to workbench
level. It was painted black on the outside and faced almost due south. The
tank was not shadowed by any other building. On the hot summers day
when an inspection was carried out in the early afternoon, the water temperature in the replacement tank had reached 76C. It had failed suddenly,
with water cascading into the workshop below and causing a great deal of
damage. The workshop was packed with electronic equipment, so a large
insurance claim was filed. When the seam of the old tank had given way, the
mains water supply entered at full bore and, Murphys Law acting as usual,
the incident had occurred after work had ended on a Friday and went unnoticed until the following week.
It was a galvanized steel header tank which, for some unknown reason,
had been painted with white gloss on the inside. This was still intact over
most of the inside but corrosion was severe over an area of the base near one
corner where the bottom had eventually perforated. The tank had not been
mounted quite level, so this corner was the lowest point where any solid debris
would tend to accumulate. There were particles of metal swarf projecting
from the paint just beyond the corroded area. When the hole had been cut
out for the downpipe fitting, particles of metal swarf had been left lying on
the bottom of the tank and had been painted over but, as the swarf particles
were thicker than the paint film, the tips were standing proud and had started
to go rusty. These had initiated corrosion of the zinc, as some were still at the
center of accretions of zinc corrosion product. The zinc coating had completely disappeared in the perforated corner of the tank, though there were
still substantial areas of zinc coating in the surrounding area.
The explanation of the tank failure thus became clear. As the upstanding
steel swarf particles in the paint rusted, the paint film was broken and exposed
the galvanized coating, which then took over and acted as the anode. As the
zinc surrounding these particles corroded away, the underlying steel was
218
exposed, but this was still protected by the adjacent anodic areas of the
remaining zinc for most of the year. However, every summer day when the
water in the tank was heated to high temperatures by sunshine being absorbed
by the black outer surface, the rate of corrosion here had increased dramatically and eventually corroded away the wall of the tank.
Basically it was the fault of the plumbing subcontractor when the heating
plant was installed some 15 years or so before: first, for placing the tank high
up in the gable end right against the window, painting it black and applying
white gloss to the inside and, second, for not cleaning out the swarf when
holes were drilled for the pipework. A plastic header tank would have been
far better, but this system was installed in the days when galvanized steel
tanks were still widely used in cold water systems.
Material Requirements
219
Figure 7.1 New radiator assembly showing the black composite tank attached
to the aluminum heat exchanger vanes of the car cooling system.
Failure Investigation
As in so many such investigations, the critical failed part was the focus of
initial attention, and was inspected macroscopically first. This enables the
investigator to stand back from the actual failure zone, and place the whole
component in its context. This in itself can prompt key questions that lead
to the solution to the problem.1
The part failed by catastrophic failure of the tank, which was fitted to
the cooling system, comprising just a half-shell (Figure 7.2A and B). The box
was 41 cm long and 11 cm wide. Comparison of a failed product with an
intact equivalent is always useful and, in this case, showed that the failed part
revealed substantial distortion (Figure 7.3). The center part of the box had
contracted significantly when compared with a new box. The same photograph also shows the way the part had been made, by injection molding from
a central gate, leaving a remnant sprue in the center of the inside of the
component. The sprue is circular and rather large because a wide gate is
needed to allow the highly viscous molten polymer to penetrate all parts of
the metal tool during molding.
220
Figure 7.2 (A) Failed and new radiator tanks compared. The upper tank failed
after only 500 miles in service. (B) Failure of the tank by slow leakage through
a brittle crack (arrows).
221
Figure 7.3
The inner surfaces of the new (top) and failed tanks (bottom), the
latter showing severe longitudinal distortion. The failed tank appears lighter in
tone due to treatment with chalk dust.
Figure 7.4 DSC curve of the glass-filled nylon 6,6 material used in the radiator
tank, showing a melting point of 265C and heat of fusion of 37.2 Jg1.
7.2.3
The critical crack was found at one end of the box, as shown by the arrows
in Figure 7.5. It was a single brittle crack that adjoined a corner of an external
222
Figure 7.5 External view of the critical crack running along an external corner,
the ends shown by arrows.
Figure 7.6 Internal view of the crack showing weld line at left.
bracket used for attachment of other engine parts. The arrow at left indicates
the tide marks formed when the leakage occurred. Because they ran in one
particular direction from one end of the crack, it could be inferred that the
box was oriented vertically in the engine compartment with the lower end
at the right in Figure 7.3. The crack was about 65 mm long, running through
the wall of the tank. Inside the box, the crack possessed the same shape and
dimensions as externally, but ran from or into a different feature known as
a weld line (Figure 7.6). Weld lines are regular features found in moldings,
normally where the molten polymer is forced around a core (to form a hole
in the final product, for example) and the streams have to rejoin. The unusual
position of this weld line is that it falls nowhere near such a design detail.
The fact that the crack also lay very close to the weld line must surely provide
some further clue to its origin.
223
Figure 7.7 Inner surface of tank with sprue at extreme right with nearby cold
slug (arrow), and weld lines at left (arrow).
Further information on the reasons for the weld line became evident
when the material of the box tank was examined. It was important to see
how the weld line had been formed and whether or not there were any more
structural irregularities. Several methods were used to examine the failed
sample, of which dusting of exposed surfaces with a white powder, chalk
dust, turned out to be most useful, as Figure 7.7 shows. The reason is that
molded composites often show great surface roughness due to the orientation
of the fibers. The picture shows several prominent flow lines (precursors to
weld lines) denoted by the arrow at the left, and a totally unexpected feature,
a fragment of an original granule from the molding process (arrow at the
right in Figure 7.7).
7.2.4
Microscopy
224
Figure 7.8 (A) Main crack near buttress corner showing witness marks from
fluid spills and weld line visible on lower side of fracture. (B) Detail of fracture
surface.
is a line of smooth material on the inner side of the fracture. This could
represent unfused polymer present within the weld line already detected. Such
indeed proved to be the case when the sample was examined in the SEM
(Figure 7.9). Due to the much greater resolution of the method, vague or illdefined features from the optical microscope could be seen with much greater
clarity. The picture already shows substantial detail in the fracture, with a
three-dimensional view unequaled by optical microscopy.
One hidden bonus of SEM, at least in this case, was the need to coat
with gold to give a conducting surface (essential with polymers or nonmetals) to bleed away electrons from the main beam. When the sample was
reexamined optically, the contrast and definition was much improved, as
Figure 7.10 shows. The weld line surface was now shown to be running for
a large distance along the fracture surface. But more significantly, there
were numerous visible cold slugs of partly melted granules embedded in
the surface (shown by the upper arrows in the photograph). Their presence
seemed to confirm a problem of cold molding in the sample. This occurs
when the molding machine is started up ready for production. The barrel
heaters are still warming up, and the polymer granules are not fully melted
and homogenized before injection into the tool.
What about the mechanical strength of the material? It was important
to test the material directly, and compare new and failed polymer samples.
Tensile testing would provide basic mechanical properties, which could, for
example, be used for comparison with the specification provided by the
material supplier.
225
Figure 7.9 (A) Low magnification SEM picture of main crack near buttress
corner. (B) High magnification shot showing weld line on inner surface and fiber
pull-out.
Figure 7.10 Gold-coated fracture surface showing cold slugs embedded in composite (upper arrows) and weld line (lower arrows).
226
7.2.5
Mechanical Tests
s = 84 MN m2
s = 81 MN m2
s = 80 MN m2
s = 55 MN m2
They show that the best results fell below the ideal suggested by a specification from the material supplier (84 MN m2 compared with 140 MN
m2). In one case the tensile test showed the sample to fall well below the
ideal value when tested laterally to the flow marks in the sample. The material
was brittle, although of high modulus. The results thus seemed to confirm
that the product could be weakened seriously by flow or weld lines present.
Could mechanical analysis shed any light on the reason or reasons for failure?
7.2.6
Mechanical Analysis
The difficulty is attempting to estimate the net effect of all these factors
working together. Using the known dimensions of the buttress corner, it was
possible to use a stress concentration diagram2 to estimate a Kt value of about
7.2. The model chosen to represent a void at a cold slug, for example, was
that for a penny-shaped crack.3 However, the evidence for the existence was
rather weak, since distinguishing a void from a crack growth region was
difficult from the micrographs. Nevertheless, a value of Kt of about 6 emerged
from the analysis, so that a net stress concentration of about 25 could have
227
been working to weaken the wall near the buttress corner. The final factor
was the distortion in the box noted earlier. It was possible to calculate the
effect it could have produced from a tensile curve made during mechanical
testing, together with the observation of the widening of the crack in the
whole failed sample box. The analysis gave the surprisingly large value of
about 20 MN m2. Thus the net effect on the potential initiator was for a
total stress of about 59 MN m2. This value is now comparable with the
weakest sample tested, so it seemed reasonable to conclude that the combination of defects initiated one or, more likely, a series of cracks that created
the final leak.
7.2.7
Aftermath
The results of the investigation were clear cut. The problem was caused by a
combination of defects in the material of the wall of the box. They in turn
had been mainly caused by faulty molding. One scenario by which they were
produced has already been suggested: the failure box was made during startup of the injection molding machine, when molding conditions had not been
established. The failed box was therefore a single maverick. If this was the
case, then several conclusions could be drawn.
In the first place, the radiator manufacturers could be reassured that the
basic design was not at fault, and they could not expect to see widespread
failures on new cars. Secondly, the suppliers could be asked to re-examine
their quality policy. It is normal and good practice for molders to supply
each operator on the machines a diagram of likely defects to look for in
moldings as they are made. This is the first line of defense in quality, and
perhaps the most important, because every component is (or should be)
examined individually. The faulty box must have been mistakenly accepted
as a good box at this stage, and forwarded through to the manufacturer. The
molder should be asked to confirm that such quality checks were indeed in
place, so that the user might be assured of specification-compliant products.
The molder generally has responsibility to supply quality products, but it is
also true that development designs supplied in short batch runs are not always
examined as closely as is really necessary.
Several useful and practical suggestions to improve the situation were put
forward for consideration by the investigation. Inspection could be improved
at the audit stage in QA, by lightly dusting with chalk to show up any surface
defects (such as cold slugs), which could indicate whether there was a problem
in molding. The design could be improved by increasing the radius of the
buttress corner. Sharp radii are easy to ameliorate by simply polishing the
corners of the core which produce the corner in the tool. The effects are often
dramatic, especially when the material of construction is brittle (as in this
case).
228
Speciality batteries are made for such applications as miners safety lamps
used during work underground (right-hand lamp in Figure 7.11). Because
the cases are subjected to wear and impact blows by contact with rock
surfaces, a tougher-than-normal insulating material is needed. Polycarbonate
is a tough polymer that found widespread and successful application for this
product in the U.S. when introduced in the 1960s. U.K. manufacturers had
traditionally used ebonite or rigid vulcanized rubber, but that material is
rather brittle and required very thick walls (approximately 4 mm) to resist
imposed loads. The top was sealed in place after insertion of the battery
plates, and a screw fitting was used on the front of the product to allow
topping up the contents with distilled water. By contrast, polycarbonate
would allow a transparent case to be injection molded, and could incorporate
a leak-proof open vent to allow topping-up. To avoid design or patent
infringement, the British battery case would be completely redesigned.
The first product (right center lamp in Figure 7.11) was developed with
very little testing, and was unfortunately introduced into U.K. coal mines too
rapidly in the early 1970s. Failures occurred very quickly, with the majority
of new lamps failing within 3 months of introduction (Figure 7.12). Many
mines had been supplied and their old stock completely replaced, with widespread failures bringing large coal mines to a halt due to the danger in which
miners would be placed by loss of light underground. So what were the main
failure modes? Cracks developed from several different areas of the sealed
casing, including the leak-proof vents, the lid-case seal and at stress concentrations either present externally or within the case. Several leaks developed
around the vents, where an organic solvent (a mixture of methylene and
ethylene chloride) was used to create the final seal. Brittle cracks also developed at untreated parts of the case, such as a mold parting line on the side.
229
Figure 7.11 Various designs of miners lamp battery cases, with the oldest at
the right in aluminum (now banned because of the possibility of thermite sparks).
The other lamps are made in polycarbonate, red filled at the right, and transparent
at the center and left. The polycarbonate molding is at the extreme left.
100
80
60
40
Normal Life
of a Battery
20
0
1
Lifetime/Years
Figure 7.12 Failure records of first design of polycarbonate battery case in U.K.
coal mines.
The crack here had started at the sharp corner at a small lip created by the
joint in the tool surfaces. In other examples, cracks had started from no
visible external stress raiser, but stripping the affected product showed that
cracks could be initiated all too easily from very sharp internal corners (radii
of curvature approximately 0.1 mm). What could be the cause of the several
modes of failure?
Polycarbonate is extremely strong in the form of extruded sheet, finding
use in safety shields and bullet-proof glazing. However, when it is injection
molded into complex shapes, a lower molecular mass grade must be used in
order to penetrate all the intricacies of the molding tool. It is almost universally true that polymer strength is very sensitive to molecular mass, falling
230
rapidly with lower chain length. Moreover, the material can possess a high
degree of frozen-in strain, especially if the molding tool is cool. The polymer
chains are effectively frozen into nonequilibrium shapes, frozen-in strain
which will be relieved by crack formation if the part is constrained. Unconstrained parts will distort as shown by the maverick radiator tank in the
previous section. The material manufacturers (General Electric Corp.,
U.S.A.) recommend that the tools be kept at a temperature of at least 80C
to minimize the problem. In addition, product strength will also be adversely
affected by geometric stress concentrations like sharp corners, which can also
be zones of high frozen-in strain as a result of constrained flow of the melt
around them.
On further investigation on behalf of the National Coal Board, high levels
of frozen-in strain were found at or near the crack initiation sites due to
faulty molding practice. Figure 7.13 shows the high levels of chain orientation
in the back of a cell, further increased during solvent welding. The solvent
itself can initiate crazing and cracking because the polymer is swollen and
weakened by absorption of the organic fluid. The existing design also possessed numerous serious stress raisers, which exacerbated the problem. Internal corners, for example, produced a severe stress concentration. Thus
recommendations to the manufacturer included changes to molding procedure, adoption of a stronger transparent grade of material, quality checks
using the procedure illustrated in Figure 7.13 (strain birefringence) and amelioration of sharp corners by tool modifications. The sharp corners could be
easily modified by simply polishing the corresponding molding tool edges,
a low cost way of improving product strength by an order of magnitude. The
improved product is shown at the center in Figure 7.10. The failure rates
experienced in many U.K. coal mines improved as a direct result of the
231
16
12
197475
Figure 7.14
197576
197677
197778
cations.
232
Figure 7.15 Base of polypropylene traction battery case showing oval burn.
Figure 7.16 Close-up of hole with brittle crack in the bottom running through
to the interior.
was quite unexpected, but could be explained by the presence of an aluminum-silicon alloy stud in the container holding the traction batteries. If the
stud had been present when the battery was first inserted, it is likely that the
base impacted the stud and initiated a brittle crack. When the crack penetrated the wall, sulfuric acid would have leaked and reacted with the light
metal, igniting the plastic at the base (Figure 7.17). The small fire was quickly
extinguished by the liquid acid, leaving the shape of the stud impressed in
the base. The problem could thus be attributed to the installer rather than
the battery maker. The design was not at fault, and greater care should have
233
Figure 7.17 EDAX spectrum showing presence of aluminum and silicon with
explanatory diagram of cause of fire below (schematic of probable failure mode
from aluminum stud that impacted against battery base).
been used when installing large and heavy traction batteries to ensure that
their containers were clear of any debris.
7.3.3
UV Degradation
234
The problem discussed here occurred during one of the many military
campaigns by the Israeli Army during the 1970s and 1980s. Fork lift trucks
used during vehicle maintenance are often electrically powered, but the large
lead-acid traction batteries must be recharged regularly to maintain their
capability in service. Normally such charging would occur at night, but
circumstances were such that charging occurred during the day, leaving the
batteries exposed to intense sunlight. The tops of many such batteries suffered
severe degradation as a direct result, creating much embarrassment to the
battery supplier (Figure 7.19). It was observed that the heat-welded tops were
most severely affected in the weld zone, the material becoming bleached by
the suns rays (an effect known as chalking). The effect on the molecular
mass of the polymer was confirmed using a special analytical method known
as gel permeation chromatography (GPC), as shown in Figure 7.20. Polymer
from the inside or hidden surface of a battery showed a much higher molecular mass than that on the upper exposed surface. FTIR spectroscopy showed
the presence of oxidized chains (carbonyl groups) in the exposed material,
precursors of chain cleavage. The fact that degradation was more severe in
the thermally affected weld zone was caused by excessive temperatures used
235
Figure 7.19 Traction battery tops degraded by exposure to sunlight, with battery at left chalking in thermal weld zone.
Lid
Lid
(Upper Surface)
Weight Fraction
(Lower Surface)
Mc
Mn Mn Mn Mn
(Upper)(Lower) (Upper) (Lower)
102
103
104
105
106
107
Molecular Mass
Figure 7.20
battery top.
236
237
Figure 7.21 Failed bathtub showing large hole caused by stress from the user
at the unreinforced zone of the plastic surface.
It was clear that a molding grade of PMMA had been used in the clear
glaze applied to the ABS sheet, lowering the strength of the layer to such an
extent that cracks began when the user sat in the bath. Further crack growth
occurred at every session, probably exacerbated by extra thermal loads from
the bathing water. Polymers have large coefficients of thermal expansion and
even small rises in temperature will create stress in the layer. The strength of
PMMA with cycling rate is shown in Figure 7.24, showing the difference in
behavior between molded and cast PMMA, with a typical stress in a pointloaded zone of about 7000 psi (about 50 MN m2). The molded material
would fail at only about 500 cycles compared with nearly 300,000 cycles for
a cast high-molecular-mass material. There were several hundred fatigue
striations, confirming the diagnosis. All the baths were replaced, and the
process redesigned to incorporate a much higher molecular mass polymer.
238
239
12000
Stress (lb/in.2)
10000
A
8000
Zone Stress
6000
4000
2000
10
102
103
104
105
106
107
108
Number of Cycles N
Figure 7.24 S-N curve for two cast and molding grades of PMMA with lifetimes
for typical zone stress produced by point load from user.
240
(Figure 7.25). The tanks seemed unfit for the purpose due to the reduced
volume for the contents and potential interference with ballcocks and other
control devices. So what was the problem?
Although thermoplastic products are easy to make, they need careful
thought concerning the loads imposed on their walls over the lifetime of the
product. With tanks buried below ground, the walls must be capable of
resisting the external hydrostatic pressure from ground water present in the
subsoil. The pressure, P exerted on the walls is given by the equation
P = rdg
(7.1)
where r is the density of water, d the depth of any part of the tank wall below
the top of the water table and g the acceleration due to gravity, a universal
constant. When many existing installations were inspected, it became clear
that those tanks that had been buried deep had suffered more internal distortion than those buried at shallow depths. Allowance was also needed for
variation in the water table height as well as the time of exposure.
The tanks had been provided with very thick walls, so what was the
problem? The material used was low density polyethylene, which creeps if
subjected to a continuous load. Thus an exposure depth of only 3 m of ground
water will impose a pressure of about 3000 Pa (about 3 kN m 2) on the
sidewall. The walls had been given 12 buttresses around the perimeter, but
they were about twice the thickness of the panels in between (Figure 7.26).
The solution to the problem lay in redesigning the tanks to give much
greater resistance to hydrostatic loading, although many users reverted to
fiberglass tanks, which provides a much greater wall stiffness and greater
241
Figure 7.26 Sectioned tank showing inward creep on walls and thicker buttresses.
resistance to creep. The extra cost of the material and labor-intensive processing is more than justified by the longer life of a product, which could
prove extremely expensive in replacing after fitment and collapse.
242
Outside Wall
Pipework
Overflow
Compressor
Equipment
W4
W1
Caustic
Jet
W3
W2
Short
Panel
Sight
Tube
Bund Wall
Partition
Wall
Figure 7.27 Plan of caustic soda tank welded together from polypropylene
panels to form a 2.7-m-diameter tank. The jet of caustic soda emerged from the
weld line of the lowest unreinforced panel and crossed the bund wall intended
to inhibit escape of fluid from the tank.
243
Figure 7.28 Face view of the failed panel with the large white arrow showing
the origin of the brittle crack that eventually grew to the edges of the panel (small
white arrows).
7.6.1
On inspection, the evidence of failure was actually very slight. The tank was
intact, with only a small gap in the middle of a welded seam in the center of
the panel to show that a very serious accident had occurred (Figure 7.28).
The crack had been opened up (deliberately by other investigators), so that
it traversed almost the entire panel when inspected several days after the
accident. The crack also occurred at the dead center of the weld, where the
panels had been joined together thermally (Figure 7.29).
244
Top
Buttress
Upper Single
Panel
Center
Buttress
W1
Lower Single W3
Panel
Base
Buttress
W2
Jet
Bund
Wall
Factory Floor
Figure 7.29
critical crack.
Figure 7.30
It was clearly vital to examine the fracture surface, apparently the only
piece of forensic material evidence available after the incident. It had to be
removed using a circular saw (a rather lengthy procedure with such a large
tank, although cutting the soft polymer was like slicing cheese), and the key
245
(Surface) 0
Pressure/Pa
fracture surface is shown in Figure 7.30. So what did it show? The fracture
surface was very simple, with:
1.
2.
3.
It was easy to suggest that the boundary to each zone represented a period
of slow crack growth following each complete fill of the tank. Notice also
how the size of each zone increases with each fill, a feature one would expect
from a growing crack, where the stress concentration at the crack tip becomes
larger as the crack depth increases. The origin (O1) appeared to be a small,
elongated pit, or pin-hole, in the outer surface of the panel.
7.6.2
Subcritical Cracks
The interpretation of the fracture surface was of a weld failure, but did this
mean that the weld itself was faulty, or was the material substandard, or
perhaps the material had been attacked by the caustic soda? The design of
the tank might also be at fault.
Further inspection of the tank was needed to check the other welds,
especially those welds at the same level on the overall structure. The reason
for examining those welds is that the pressure in a tank is given by an equation
similar to that applicable to buried tanks:
P = rhg
(7.2)
246
where P is the hydrostatic pressure of the contents, r the density of the fluid,
h the height from the top fluid surface and g the acceleration due to gravity.
Thus the pressure should vary with depth as shown in Figure 7.31 for fluids
of varying density. In this case, since the density is fixed, and g is a constant,
the only factor controlling pressure is the distance from the top level of liquid.
In other words, if a vertical weld had failed at a particular height, then there
could be evidence of other cracks at nearby welds at the same height if failure
had been caused by the internal pressure of the contents.
The hunch proved correct, but only in one of the four possible welds
(W4 as shown in Figure 7.27). The weld was obscured by its position, making
inspection very difficult (looking for a dark crack on black material in the
dark), although when fine talc was used to dust the weld, it revealed several
subcritical cracks (Figure 7.32). What did it indicate? It might show that the
welding was faulty, but it did not explain the uncracked welds (W 2 and W3).
247
Were there any other clues to the cause? Another factor could be overstressing:
was there any distortion visible in the tank? After all, the distortion visible
in the radiator box was a significant clue to that problem. If Figure 7.28 were
re-examined, such distortion does indeed exist, best seen by tipping the page
and viewing the edges of the far side and the crack at a slanting angle. Both
features can be seen to be bowed outward, in a convex manner, the cracked
weld showing the greater effect. The distortion in the wall was confirmed by
direct measurement of the circumference at several heights, a difference of
20 mm in a total of 7.55 m being measured for the center of the cracked weld
compared with a higher panel. Although the effect was very small, it could
be indicative of overstressing, producing creep of the panel material.
7.6.3
Material Testing
The quality of the welds and the sheet material also needed to be assessed
independently. Samples from both intact and cracked welds were sectioned,
and the sections polished and etched to show any internal structure
(Figure 7.33). No serious problem with weld structure could be found, apart
from the presence of tiny pin holes. The tensile strength of the welds also
proved reasonably consistent with one another and manufacturers figures.
Chemical analysis of the weld and bulk materials using DSC proved negative,
the melting points all being close to one another. FTIR showed no obvious
anomalies (such as an oxidation peak).
Figure 7.33 Weld design in final closure of panels to form hoop, with hot blade
inserted between panel edges. The weld forms when the two ends are pressed
together.
248
7.6.4
Weld Defects
Although weld strengths appeared reasonably high, there was a small difference of about 4% between the tensile strength of welds from intact and
cracked panels. Could mechanical analysis be used to check the stress-concentrating effect of the pinholes? The first point to check was the appropriate
formula needed to calculate the hoop stress in the vessel. Fortunately, the
stress situation in a static pressure vessel is very simple. Where the ratio of
tank radius to wall thickness is greater than about 10, then the tank can be
treated as a thin walled vessel.7 The ratio is in fact 1.35 103/12 = 112.5, so
the hoop stress, sH is given by the equation
sH = PR/t
(7.3)
where P is the hydrostatic pressure, R the radius of the tank and t the wall
thickness. Knowing R and t, and calculating P using Equation 7.2 with a
known density, then the hoop stress at the failure crack was evaluated as
sH = 3.4 MNm2
It was now possible to calculate the stress concentration (see p. 50) that would
have been acting at the critical pinhole,8 prior to the first growth of the crack,
since
Kt = weld failure stress/actual stress at origin
(7.4)
Hence Kt = 20.4/3.4 = 6.
Thus, from mechanical tests of welds, it was possible to calculate that
the real stress acting at the pinhole in the center of the lower single-thickness
panel was about six times the nominal hoop stress from the hydrostatic
pressure of the contents.
7.6.5
Tank Structure
249
Dam
Design
Barrel
Design
Figure 7.34 Schematic sections of two contrasting tank designs; at left, a dam
design with increasing wall thickness to resist increasing hydrostatic pressure
from contents and at right, the barrel design of the failed tank.
in thickness from top to base to resist the water pressure, and the same principle
should apply to any fluid reservoir.
By adding hoops, the top band could be redundant, and the lower bands
might not be sufficient to resist the much greater pressures toward the base
of the structure. Such a hypothesis would explain why the failure occurred
in a lower, unreinforced panel. It would not explain why only two welds failed
from pinholes, but the calculation of hoop stress above shows that the singlethickness panel is having to resist a large hoop stress. Doubling the wall
thickness would halve the applied stress, while having three panels here would
give a stress of only about 1.15 MN m2. Even if pinholes occurred in the
weld, failure would be much less likely with such substantial lowering of the
hoop stress. The general conclusion of the stress analysis was that the design
itself was faulty: in order to resist hydrostatic pressure, the tank should have
been designed like a dam wall, rather than like a barrel (Figure 7.34).
But why should just two of the four welds have shown cracks? The answer
to the problem came when the welding stage was inspected directly. The hoop
of panels for such tanks is made sequentially by hot fusion welding, that is,
by melting the surfaces of two panels and pushing them together. This is fine
for three of the welds where the flat panels are joined, but difficult for the
final joint when the ends have to be brought together by bending the sheet
into a cylinder (Figure 7.35). It would certainly explain the lower quality of
one weld, and the low quality of another weld was probably caused by similar
problems in bringing two large flat sheets together. The quality of such welds
is tested for through-the-thickness holes using a spark tester, a method that
will not detect partial pinholes. One rather disturbing aspect of the process
is that the hoop so formed is under a bending stress, so the outer surface of
the final wall will be in tension. This will of course make failure much more
likely, and is akin to the frozen-in strain problem of the radiator box already
250
PP Wall
PP Wall
Hot Blade
Figure 7.35 Adjacent barrel tanks holding acidic and alkaline contents at a
wire-making plant.
described. In this case, however, there was a frozen-in stress rather than strain.
It was possible to calculate the effect from simple bending theory. 7 The effect
probably added about 1.5 MN m2 to the hoop stress, so that the above
calculation underestimated the stress.
7.6.6
Aftermath
We concluded that the basic design of the tank was faulty, having been made
like a barrel rather than a dam. Failure was inevitable, and the shape explains
why failure occurred so early in the life of the structure. Because the walls
were exposed to excessive stresses, it was inevitable that the tank would fail
quickly. The stress would seek out the weakest welds in the most exposed
panel, and failure was not caused by faulty welding at all. That is not to say
that the welding process could not be improved, for example, by controlled
bending of individual panels thermally before final welding.
But the failure should never have occurred in the first place, because
there was a standard for such thermoplastic tanks, published by the German
Welding Institute, DVS 2205. Only partial translations existed in published
form at the time of design of the tank, but the design philosophy was perfectly
clear. The design procedure6 makes allowance (using safety factors, or derating factors) for holding dangerous chemicals, pinholes in welds and so on.
The manufacturer of the tank, a small business man with entrepreneurial
skills, had hired an engineer to check and approve the design. The engineer
unfortunately performed misguided calculations, leaving the insurers to
shoulder the considerable expenses of the clear-up.
Having explained the failure, what were the consequences of the investigation? As is usual, several investigators had been instructed by loss adjusters acting for the two injured parties. The above investigation was carried
out on behalf of the tank manufacturer, and since liability was accepted,
their role was limited but other investigators agreed with the general analysis. A much more serious problem (which arises whenever design defects
FeCl3
251
NaOH
Bund
Wall
Figure 7.36 Storage tanks for water treatment chemicals. Tank leakage could
have caused a fire.
are discovered) was the state of other tanks made to the same design. How
many had been built and installed? What fluids were they holding? For how
long had they been installed?
7.6.7
A range of tanks built to the same principle had indeed been installed, but
only shortly after installation of the failed tank. The original investigator was
asked by the manufacturer and insurers to inspect all such installations, which
ranged from tanks for fruit juices, soap solution and ferric chloride (FeCl 3),
an acidic fluid used for water treatment.
Although the tanks were relatively new, few were found with microcracks,
essentially because none of the tanks had been filled to capacity. The cracks
in the welds that were found were far from critical, being only millimeters
in size. The economy at the time was depressed, and the companies concerned
had been on short-time work. Some of the other tanks inspected were much
smaller than that at Warrington, and this had not been stressed to the same
extent. The most serious potential problem was found at a steel wire works
in the Midlands, where there were two tanks adjacent to one another
(Figure 7.36). One held caustic soda, the other ferric chloride. If the tanks
had split, the two chemicals would have reacted together with great evolution
of heat, probably causing a fire. The tanks were in a confined space, close to
manned equipment, so the consequences of failure would have been more
serious than at Warrington because of the enclosed space. The bund walls
had only been designed to accept the contents from a floor leakage, and
would have been no help in a crisis such as the Warrington failure. No
microcracks were found, however, because the tanks had never been more
than half-full since installation, due to the fact that the tanker could not
approach the site. The factory was set in a narrow gorge and the only access
252
Weld
Weld
Figure 7.37 Tank with internal cone for holding raw paint. It collapsed because
of the poor support for the inner cone. The base of the tank was supported by
the main weld and four vertical panels, two of which are shown. A steel support
frame should have been used instead.
to the tank building was over a small bridge incapable of carrying a full
tanker. Needless to say, this and many other tanks had to be replaced by tanks
of the correct design.
7.6.8
Such tanks continue to fail, however; one recent failure in 1998 involved a
paint tank. It had been designed with an inner sloping floor, as shown in
Figure 7.37. This was essentially to allow all the contents to be mixed within
the tank and then drained away for further processing. The tank failed the
first time it was used, collapsing in a misshapen heap on the factory floor,
the contents draining away over the factory floor. The weakness of the design
shown is that all the mass of the contents is taken by a single weld, with
support for the inner cone only provided by four plastic panels. They failed
quickly and the weld had peeled away, failing progressively, so that almost
all the weld had disappeared. The design was again at fault, with the absence
of reliable support recommended by DVS 2205, such as a steel frame resting
on the ground, and bearing the weight of the cone and contents. Fortunately,
the design was a one-off, made by a small company for a specific contract,
and had not been repeated elsewhere.
Thermoplastic storage tanks have displaced not just metal or alloy equivalents but other polymer types, especially glass fiberreinforced vessels.
Although much more expensive, the reinforcement from the glass fiber makes
for a much stiffer wall, a factor that allows tall silos to be constructed.
Nevertheless, failures have been reported by Meyer Ezrin,9 some of which
have actually been caused by the glass used. In another case reported by Derek
253
Hull, a large milk tank failed catastrophically after 10 years of reliable service.10 The collapse demolished an electricity substation and created substantial physical damage to its surroundings. It was used for storing milk
products, and was made from random mat GRP fitted with a polypropylene
liner. Failure started at a butt weld in the latter, from a combination of cyclic
loading (filled and emptied daily) and nitric acid attack, suggesting stress
corrosion fatigue. The slightly acidic contents seeped into the GRP and
caused further SCC, ultimately leading to the sudden failure.
7.7 Conclusions
The case studies in this chapter have highlighted the critical role of materials
and their use in containment products. Although plastics materials have
found extensive application for storing fluids, their use requires a considerable degree of caution both in terms of design philosophy and installation.
When changes to conventional and traditional designs are made, allowance
must be made for any changes in the material of construction. If the properties are quite different, then substantial alterations must be made to the
design geometry. Thermoplastics, in particular, possess much lower tensile
moduli than steel and most structural metals (by a factor of about 100, or
two decades) as well as lower strengths. On the other hand, they do offer
considerable resistance to hazardous chemicals, and they can be used for
chemical storage, for example. They can also be processed to shape very easily,
using methods such as injection or rotational molding or even by simply hot
welding flat panels together. As with many structural designs, the zones where
subassemblies fit together are usually the weakest parts of the structure and
will fail from here if overloaded. For structural integrity, all design work must
start at the weakest parts, and make allowance for any unpredicted problems
at these zones.
The importance of testing prototypes before release into the market is a
recurring theme in this chapter. There is no substitute for testing products
under realistic conditions to assess lifetime before launch, rather than relying
on the user to suffer the consequences of hasty marketing. It is especially
true for storage tanks holding dangerous fluids, where the results of failure
can endanger life and property. A simple water test on the large storage tanks
of the last section would have shown very quickly that the walls would creep
outward very rapidly. Testing underground tanks would be more difficult,
but could have exposed the problem of variable wall thickness. On the other
hand, no amount of testing would have shown the weakness of maverick and
faulty radiator header tanks. Here the problem is one of quality control,
assuring the user that all products are structurally sound and fit for their
intended purposes.
254
References
1. Lewis, P.R., Premature fracture of a composite nylon radiator, Eng. Failure
Anal., 6, 181195, 1999.
2. Pilkey, W., Ed., Notch in bending, Chart 2.29, Petersons Stress Concentration
Factors, 2nd ed., Jossey-Bass, San Francisco, 1997.
3. Pilkey, W., Ed., Penny shaped void, Chart 4.71, Petersons Stress Concentration
Factors, 2nd ed., Jossey-Bass, San Francisco, 1997.
4. Crawford, R.J., Rotational Molding of Plastics, Wiley/Research Studies Press,
New York, 1992.
5. Lewis, P.R. and Weidmann, G.W., Catastrophic failure of a polypropylene
tank. I. Primary investigation, Eng. Failure Anal., 6, 197214, 1999.
6. Weidmann, G.W. and Lewis, P.R., Catastrophic failure of a polypropylene
tank. II. Comparison of the DVS 2205 code of practice and the design of the
failed tank, Eng. Failure Anal., 6, 215232, 1999.
7. Roarks Formulas for Stress and Strain, 6th ed., McGraw-Hill, New York, 1989,
p. 516.
8. Pilkey, W., Ed., Stress concentration diagram for elliptical hole in infinite
panel in tension, Petersons Stress Concentration Factors, 2nd ed., Jossey-Bass,
San Francisco, 1997.
9. Ezrin, M., Plastics Failure Guide, Hanser-SPE, Munich, 1996, pp. 346349.
10. Hull, D., Fractography, Cambridge University Press, London, 1999, section
10.4.3, p. 335 ff.
Accidents
in the Workplace
The Accident
256
257
ensued, which actually blew the metal away, so it appeared to run off like
rain dropping from a waterproof jacket or trousers. All safety equipment was
kept in a locker room at the side of the foundry, about 20 m from this
particular melting furnace. Prominently displayed warning notices were on
the walls of the foundry and on the charging platform of each furnace. At
the opposite end of the foundry building to the locker room was a two-story
suite of offices for the foundry manager, foreman and clerk at floor level and
three rooms above housing an analytical laboratory. The manager, foreman
and laboratory staff frequently went into the melting area and the foreman,
in particular, often assisted the furnacemen at busy times. If their visit was
brief and not directly onto the furnace charging platforms they would wear
hard hats and safety spectacles only. The foreman usually wore his protective
trousers but would not put on the jacket and face protection unless he
intended to work on the charging platform.
The accident had occurred in mid-morning as the furnaceman began
charging for the second cast of his shift. The scrap bales had been brought
into the foundry approximately 2 h earlier and rain water was said to have
run out of some of the bales into the bottom of the stillages. The furnaceman
upturned a stillage onto the conveyor and ran the first bales up toward the
furnace platform, but stopped the load just short of the furnace mouth to
allow any moisture to dry off. He went behind the furnace to stamp numbers
on test bars for the laboratory from the previous cast. The furnace had a
fume extraction canopy, but this had been swung aside to allow access for
charging.
The foreman had been outside in the stockyard and as he returned
through the foundry he noticed that the metal level in the furnace was low
and getting very hot, and he was surprised to see bales waiting at the top of
the conveyor. Thinking that the furnaceman might have gone to the washroom, he started the conveyor. He was not wearing his protective jacket. The
noise of the conveyor alerted the furnaceman, who came toward the charging
platform from the opposite side to the conveyor. As it would be some time
before he was due to pour the metal he was not wearing his protective jacket
and had taken off his face shield because, he claimed, he could not see clearly
enough to stamp numbers on the test bars. It is believed that two or three
bales had dropped into the furnace when a violent explosion blew metal all
over the place. A partly melted bale of scrap was found over 10 m from the
furnace. Both men suffered severe burns to their upper bodies.
8.1.2
Investigations
An investigation was carried out by the Factory Inspectorate the day after
the accident. The inspector was unable to establish the precise reason for
the eruption but he pointed out that neither of the injured men had been
258
The Pleadings
259
260
plaintiffs negligently chose not to wear the jacket and helmet visor
provided on the occasion of his accident.
7. Failed to exercise reasonable care for their own safety and unnecessarily exposed themselves to the risk of such injury as actually occurred.
When the parties are represented by equally experienced legal advisers,
they recognize the strengths and weaknesses of each others cases and usually
reach some agreement by negotiation. It is only when they cannot agree, or
if they agree on liability but cannot agree on damages, that the case actually
goes to trial before a judge.
8.1.4
Outcome
261
The Accident
This particular accident occurred when two men were operating a machine of
the latter type making castings in an aluminum alloy. Both were experienced
262
Figure 8.1 Pair of aluminum alloy die castings on common sprue, as removed.
Notice the broken skin in the end face at the right-hand side where liquid metal
has burst out after dies opened.
workers, very familiar with the machine, so much so that they were left to work
on their own, virtually unsupervised, during a night shift. One of them suffered
severe facial burns when liquid metal squirted out from between the dies as he
was waiting with tongs ready to remove the finished casting as soon as the dies
opened. He and his colleague who had poured the shot into the injector
reservoir claimed he was wearing a visor, but a jet of metal struck his clothing
just below this and ricocheted up into his eye.
The casting in question, illustrated in Figure 8.1, was to form a pair of
matching spacers for a motor mount, the general view with the injector sprue
shown at the bottom. The spacer at the right shows where liquid alloy had
burst out from the face at the sprue end. The narrow blocks at the far end
are sinks to catch any dirt or residues swept before the liquid as it is injected.
These would be broken off the finished casting and put back into the melting
pot, along with the sprues. A microsection taken through the casting revealed
extensive voids due to air entrapment within in the body of the casting. This
was confirmed by x-radiography before sectioning, as illustrated in
Figure 8.2. The metallographic structure revealed the pressure of this
entrapped air had forced liquid alloy through the skin of the casting. For this
to occur the dies must have been partially opened in order for the internal
263
Figure 8.2 X-radiograph of right side casting showing excessive porosity caused
by air entrapped as liquid metal was injected.
air pressure to rupture the unsupported skin. This was clearly the source of
the liquid metal that squirted out as the dies were unclamped. The position
of the dies where this metal was ejected would have been just below the eye
level of someone standing at the side of the machine waiting to remove the
finished casting as soon as the dies opened.
On the morning after the accident it was discovered that the timing cycle
on the machine had been altered from the setting when these particular dies
had been installed the previous day. The number of castings made up to the
time the accident occurred was about 10% greater than what should have
been possible in the number of hours the two men on the night shift had
been working. It appears that they had decreased the cooling time of the
machine cycle in order to produce their expected quota of castings in less
time and thus take a longer break. It was the first time they had make this
particular casting. In view of the evidence of the way the skin had burst open
it seems they were unlucky with this particular casting in that more air than
usual had been carried in by the turbulence as the metal was injected and
the dies had opened before the skin was thick enough to withstand the
internal pressure.
8.2.2
Outcome
As in the previous case, the operators were to blame for the cause of the
accident but the management was held responsible on the grounds that:
1. The two workers were inadequately trained and had been left on their
own to work unsupervised for the entire 8-h night shift.
264
2. The time cycle setting controls should have been locked or otherwise
made tamper-proof so that adjustments could only be made by authorized personnel.
Figure 8.3 Head of chisel. The triangular eye splinter has been replaced in the
mushroomed edge in the center field. A diagonal line left of center at about 80
is one edge of the splinter; a diagonal line in the center at about 170 is another
edge, coming down to just right of center to meet at a sharp point.
265
Figure 8.4
Old gearbox shaft used as dummy for changing clutch plate. The
splined portion fits through center of clutch and the spigot end must locate in
the flywheel bearing.
266
Figure 8.5 Open-ended spanner used to tighten bolts around clutch cover.
plate to a light commercial vehicle; although fatique may not have been a
factor, he was keen to finish work for the day. Working from beneath the
vehicle he was using an old gearbox shaft as the dummy to center the clutch
plate. He needed to extract the dummy shaft before refitting the gearbox and
started to do so by tapping the projecting portion with an open-ended
spanner. Figure 8.5 shows the end of this spanner, exhibiting several deep,
angular indentations similar to the profile of the teeth of the gear. He was
looking toward the back of the engine, with his face close to where he was
striking, and swinging the spanner toward his face. He was not wearing any
eye protection. He suddenly felt a blinding pain in his right eye.
A particle was removed from his eye at the hospital, described as a sharp
flake of metal (magnetic, therefore steel). He eventually lost the sight of the
eye and commenced litigation. His lawyer sent the particle to a local college
seeking a report. Unfortunately they lost it while taking photographs. (There
is a moral here: when manipulating such a small item of evidence, it should
never be moved or turned during visual examination except when held with
a magnetic probe or tweezers over a piece of white card with the edges folded
up.) Fortunately, the colleges investigator had measured it and was able to
describe what it looked like. Instructions from the insurers of the mans
employers were to examine the gearbox shaft he was using as the dummy
and establish whether the particle might have flown from one of the teeth
and, if so, what caused it to break away.
Close examination of the outsides of the gear teeth on this dummy shaft
revealed several chipped teeth; a recent one is illustrated in Figure 8.6. The
size of this cavity matched the dimensions of the particle insofar as these
could be relied on. However, several possible cavities were found on the
outsides of other teeth in a similar position. Quite clearly the corners of
several teeth had been caught by glancing blows and small fragments had
been broken away. The scalloped form of the cavities is characteristic of a
267
hard brittle case broken by an oblique impact. As the outer corners of these
gear teeth were extremely hard (>900 HV), under relatively light blows the
exposed corners would be likely to emit tiny, sharp edged particles at high
speed.
Such a small flying particle would not have caused a penetrating eye
injury if eye protection had been worn. The injured man had been issued
safety glasses but he claimed the glasses did not stay in position when he was
working beneath a vehicle and spots of oil and greasy finger marks smeared
the glass and made it difficult to see what he was doing. There was a foreseeable risk of eye injury given the tools he was using and the way he had to
work with his face so close to the object he was striking. However, no amount
of legal argument about whether or not the risk of injury was reasonably
foreseeable could ever compensate for the loss of his eye, although damages
awarded would be substantially reduced on the grounds of contributory
negligence.
The sad fact is there would have been nothing wrong with using the old
shaft as the dummy and the injury need never have occurred if the casehardened gear teeth had first been softened by tempering.
8.3.2
Lath Axe
268
Figure 8.8 Fragment of metal removed from eye resting on a 1-mm-square grid.
owned. He was on the roof using the hammer end to trim brickwork when
he felt something hit his eye. Thinking it was a piece of the brick he rubbed
it and continued working but the pain became so intense he had to stop
working.
The tiny fragment of steel removed from his eye at the hospital, measuring 3.5 mm 2 mm, is illustrated in Figure 8.8. It has a sharp rim and
the uppermost face exhibits what is referred to as a conchoidal (shell-like)
fracture, characteristic of a piece of hardened steel displaced by a gouging
impact. The other faces resting on the graph paper were worn and bore
features consistent with the outside of this worn hammer head. The confirmatory test is to place the particle in likely cavities on the suspect tool
and see how well the surface features match the surrounding ones on the
tool. In this instance, not only did the suspect location provide a good
physical fit in the cavity, but it also displays conformity of external contour
269
and surface markings. In Figure 8.9 the particle is shown resting in the
matching cavity at the top of the head of the lath axe.
The tool had been manufactured overseas and, although well finished
externally, the hardness of the striking face was considerably above the range
specified in the British Standard 876 for hammers. This standard specifies
hardness in Rockwell numbers, but because a Rockwell test cannot be carried
out on such a tiny particle the Vickers hardness test has to be used instead.
The converted Rc hardness range is equivalent to 520 to 640 on the Vickers
scale. The rim of this hammer was 880 HV and a test on the particle gave a
value of 900 HV. At this level any steel will be brittle and the striking face of
a hammer is likely to splinter.
The brittleness on its own would have been enough to account for the
rim splintering under a heavy, glancing blow, but in this instance there was
another dangerous fault. A microsection cut from the striking face exhibited
a network of cracks produced during the heat treatment used to harden the
rim. A group of these is shown in the etched microsection of Figure 8.10.
The way they follow grain boundaries is characteristic of cracks formed in
medium and high carbon steels when the cooling rate from the hardening
temperature is too fast. In this instance it suggests that the hammer head had
been quenched into water instead of oil. Quench cracks form at the surface
and clearly account for the way numerous pieces had flaked away from this
hammer head before the accident.
Unfortunately for the injured man, who owned and had been the sole
user of the tool since new, there had been ample warning of its dangerous
condition in the way several pieces had broken off the outer rim during
270
previous use. The risk of further splintering was therefore foreseeable; so the
hammer should have been taken out of service long before the accident. A
further difficulty for the injured man was that he had purchased the tool
from a market stall and had no receipt. It appeared to have emanated from
China, a fact that raises formidable problems in litigation.
8.3.3
A young man working for a small printing firm lost an eye while helping
carry a drying frame up a narrow staircase. The frame consisted of a tall stack
of 50 hinged trays mounted in a steel frame with castors. The individual trays
were flat and made of steel wire mesh, so that sheets of paper coming from
a silk screen printer could be laid on them one at a time while the color dried.
Each tray was hinged so that it could be stacked practically vertical until
needed, and then folded down to a horizontal position when a fresh poster
came from the printer. Coil tension springs at the pivot ends held the trays
in the raised position and were stretched when the tray was pulled down.
Figure 8.11 is a photograph of the upper trays in two adjacent dryers: the
one at the left has the trays in the horizontal position, the one on the right
has them in the raised position. Attention is drawn to how far the springs
are extended when the trays are horizontal.
The side handrail of one of the six drying frames on the first floor of the
building had been damaged, so the frame was sent to a local fabricating firm
who formed a new tubular handrail and welded it into position. The top of
271
Figure 8.11 Side views of two drying frames to show how far the springs extend
when the trays are folded down to the horizontal position. In the frame at the
left all but the top two trays are down, whereas in the one at the right every
frame is still in the up position.
this handrail and the weld attaching it to the back frame can be seen in the
left-hand drying frame of Figure 8.11. Notice how far the spring is stretched
when the top tray is in the raised position and also how close this spring is
to where the welding was carried out. When the frame came back from repair
it had to be carried up a narrow staircase to the first floor as there was no
elevator in the building. One man pushed from the bottom, while the young
man was pulling at the opposite side, so at times his face was quite close to
the top weld. As they reached the turn near the top of the staircase, the
uppermost spring suddenly broke and the hooked end shot into the young
mans eye, injuring him so severely that his eye had to be removed.
The fractured spring was sent for examination to establish the cause of
its sudden breakage. As it had necked down at the point of fracture it was
clearly not a fatigue failure, confirmed by examination of the fracture that
exhibited all the features of a ductile overload. Even if it had been fatigue it
would have been difficult to explain why it had broken in the middle of the
straight portion rather than where it was bent. Springs of this type are formed
from cold drawn patented steel wire, which have a strongly oriented fiber
structure in the longitudinal direction and, significantly, have a carbon content of the order of 0.8%, which puts them in the range of steels that respond
to heat treatments.
The polished and etched longitudinal section of the length removed from
the injured mans eye is shown in Figure 8.12, with the fracture at the left.
This reveals localized areas where the steel has undergone heat treatment;
272
Figure 8.12 Polished and etched section of spring adjacent to fracture. The
wire diameter is 0.9 mm.
273
The key evidence was thus the broken tip (Figure 8.13). It showed a brittle
fracture in an unidentified, relatively rigid plastic. The fracture surface was
very fresh, and there was no sign of old or subcritical cracks. FTIR analysis
showed that the material was in fact a copolymer of polypropylene, which
should normally be tough and resilient. But what did the fracture reveal, and
were there any obvious features that could represent defects exposed in the
surface?
The central void is clearly a stress concentration with a stress concentration value of at least two. However, analysis of the crack surface itself showed
the origin to lie elsewhere, at an external corner produced as a feature of the
design of the tool used to injection mold the component, known as a mold
parting line. It is simply where two mating parts of the steel tool meet, and
often show a mismatch owing to mold tool wear (Figure 8.13B). The
274
Figure 8.13 The parts of the accident ladder. (A) End of ladder. (B) Fracture
surface of ladder tip. (C) Intact and broken tips compared.
conjunction of this small corner and the larger external corner thus produced
a stress concentration larger then elsewhere on the corner. Although only
one of the two tips had broken, both showed severe abrasive damage to the
parts in direct contact with the brick wall against which the ladder had been
leaned just before the accident.
The possible causes of the fracture at this stage of the investigation were:
Poor material
Faulty manufacture
Poor design
But there was little support for poor material, such as degradation, from
either DSC or spectroscopic analysis. Although the void represented poor
design and molding practice, it did not initiate the crack that broke the tip.
275
The sharp corner in the molding had initiated the crack and it represented
a possible design defect if it had failed and caused the accident. Sharp internal
corners are commonly present in many molded products, but are not necessarily defects unless it can be shown that these features cause product failure
and hence accidents.
8.4.2
Ladder Stability
How was the ladder being used at the time of the accident? The key variables
are the angle of repose of the ladder and the coefficient of friction between
the feet and the ground. A visit to the site of the accident was essential, so
the witness statement could be checked, and any further evidence that could
clarify the circumstances could be examined directly. The visit showed visible
evidence of the intermittent contact between the tips of the ladder and the
wall below the window, and the traces of an impact with the small sill above
the patio doors. The trace contact marks corroborated the witness statement,
showing how the tips and hence the ladder structure itself had oscillated from
side to side as it slipped down the wall. No traces of marks from the feet on
the concrete slabs of the patio itself could be found. The angle of repose and
coefficient of friction of the feet remained unknown, however. A reconstruction of events was essential to measure or calculate the data using the actual
ladder itself.
8.4.3
Reconstruction
276
N
mN
Ladder Foot
Inclined Concrete
Slab
M
q
Coefficient of Friction
The first task was to estimate the static coefficient of friction () of the
polypropylene feet.2 A simple way of doing this for very light loads is simply
to place the feet on a surface similar to that being used for the real ladder
and tilt it until the feet just slip (Figure 8.14). In this case, the surface is
simply a concrete slab. Theory shows that the critical condition for slip on
the inclined plane is
= tan
(8.1)
where is the angle of inclination of the slab to the level ground. Two
experiments established that the coefficient of friction for the feet (mass
about 100 g) was
(feet) ~1.00
Such a value is typical for the elastomeric polymer concerned acting on
smooth concrete. But would the feet as fitted to a ladder exhibit the same
value? Direct determination from the (unladen) accident ladder showed that
it slipped at an angle of about 40, so that neglecting friction at the tips, the
coefficient of friction is about 0.6, considerably lower than the value estimated from an inclined plane experiment. There are several reasons for this
situation. In the first place, the coefficient of friction of polymers is known
to be dependent on load, decreasing as the load increases. Secondly, the feet
of the accident ladder had been designed for a repose angle of about 75
277
52
(Figure 8.15), any other angle lowering the contact area of the serrated surface
against the ground. This was demonstrated by recording the footprint of the
feet at several angles of repose. The edge in contact with the ground thus
becomes even more heavily loaded over a much smaller area of contact.
8.4.5
Repose Angle
The next question to be addressed was the likely angle at which the ladder
had been leaned against the wall. The claimant thought that it had been
leaned against the sill of the upper window, so that then the angle of repose
could be calculated for two situations: an unextended ladder and the ladder
extended by one rung. Greater degrees of extension would create progressively lower angles of repose. As a working hypothesis, it was thus assumed
that the ladder had been used either unextended, or with one rung extended.
Leaning the ladder against the sill would be given by
sin = vertical height to sill/ladder length
(8.2)
With the known length of the unextended ladder being 4.46 m, extended
by one rung being 4.71 m and with the adjacent sill being 3.69 m from the
level ground surface, then = 56 (unextended) or 52 for an extension of
one rung. The situation for the ladder is shown schematically in Figure 8.16
and Figure 8.17, with the plastic tips leaning against the window sill. In this
position, a slight movement of the ladder would allow the ladder to jump down onto the adjacent wall, where either the aluminum tips of the lower
section of the ladder or the plastic tips on the upper section would make
contact with the brick wall. Both sets of tips showed abrasion, most visibly
obvious in the case of the plastic tips (Figure 8.18). In addition, matching
the ladder to the wall with known lengths and heights showed that the
unextended ladder had most likely been used. The angle of repose of 56 was
well below the recommended angle of repose of 75 for this design of ladder.
278
Upper
Window
Sill
Brick
Wall
French
Window
56
Patio
Figure 8.16 Side view of unextended ladder leaning against window sill (schematic).
Buffer
Sill
Upper
Section
First Contact
with Wall
Lower
Section
But what could explain the curious set of marks below the window and the
comment about walking down the wall?
8.4.6
Stability Experiments
The final phase of experiments involved simulating the weight of the user
when working near the top of the ladder. Both the actual weight and height
279
Figure 8.18 Top of failed ladder tips showing wear on surfaces, with contamination by brick dust (arrows).
of the user were known, and it was most likely that he was standing on the
fifth step from the top when the ladder slipped. The user could thus be
simulated by simply suspending a fixed mass of 72.6 kg (726 Newtons) to
represent his weight on a rope from the upper rungs. The mass of the ladder
of about 20 kg (200 N) was leaned at an angle of 56 against the wall,
representing the unextended ladder (Figure 8.19). The exact position of the
suspended mass was also important, because when ascending or descending
a ladder, the user shifts his weight from foot to foot. The user could not
remember which rung he was standing on at the time of the accident, but it
was likely that the fourth, fifth and sixth rungs from the top were in use at
the time. Since the user was cleaning the right-hand pane of glass, his left
hand was probably holding the left-hand part of the uppermost rung. By
leaning over, his weight will have shifted to the right-hand part of the rung
he was standing on, so the right-hand part of the fifth rung was initially
chosen as the point of attachment of the suspended mass.
280
Figure 8.19 Reconstruction with weight hanging from ladder rung (safety ladder at left).
One feasible stimulus for the accident could have been momentary loss
of contact of the left-hand tip with the sill against which it was resting, by
reaching over to the right, for example. A spring balance was used to determine how much force was needed to pull the left-hand stile at right angles
from the sill (about 10 kgf). The torque at the top of the ladder simultaneously pulled the left-hand foot out of contact with the ground, and the
loss of contact causes the right-hand tip to slip down to a slightly lower
position against the wall. Repetition of the effect led to progressively lower
positions against the wall until the whole ladder became dangerously unstable
281
Figure 8.20 Stick-slip marks made on white board by ladder tips after uncontrolled slipping of the ladder. It occurred with the user mass hung from the fourth
rung down from the top.
and the experiment was halted. The effect is known as stick-slip motion
(Figure 8.20).
So what did the reconstruction show? First, it suggested that the ladder
had been leaned against the wall at an angle of a maximum of about 56,
well below the recommended safe angle of repose. Second, it confirmed that
at relatively low angles of repose of this particular design of ladder, stick-slip
motion could occur after momentary instability (even if the angle of repose
was above the critical angle of repose). That instability was produced by a
torque load that moved one of the tips of the ladder from contact with the
wall. Provided the user was near the top of the ladder, catastrophic and
uncontrollable loss of the ladder was inevitable.
8.4.7
Conclusions
The reconstruction of the accident showed that the user initially leaned the
ladder at too low an angle for a reasonable safety margin. Although it was
above the critical angle of repose, it was susceptible to stick-slip instability
when the user was near the top of the ladder. The visible contact evidence
from the wall confirmed stick-slip movement of the ladder tips. But the
question of the fractured tip remained unanswered. It could still have caused
the accident if the tip had broken at the sill, and initiated stick-slip motion.
However, both the broken and intact tips showed abrasion against brickwork,
so it is more likely that it survived for some distance down the wall. The final
piece of evidence was an impact mark on the small wooden sill above the
patio doors. This was probably caused when the tip hit the sill, and the tip
282
broke from its weakest point. The fracture was thus the result, and not the
cause of the accident.
Given this conclusion, it was decided that the case had no chance of
success at trial. The victim of the accident did not receive any compensation.
The moral of this forensic story is that initial perceptions of failure may not
always survive critical scrutiny. Investigation should always examine the circumstances surrounding a product failure: the fracture is just one more piece
of evidence to build up a picture of events based on the most reasonable
interpretation of all the evidence. Ladder accidents are very common, and a
site visit is essential to gather more evidence of the incident. In this case, the
contact evidence on the wall was important for construction of working
hypotheses to explain the accident. In addition, practical reconstruction of
the accident using the failed ladder demonstrated stick-slip motion above
the critical angle of repose, and helped explain the contact evidence. However,
the answer to the problem lay not in the tips but the feet of the ladder, since
it is the feet that play the crucial role in the stability of ladders.
8.4.8
Workplace Accident
The second accident occurred when a painter was working alone in a factory,
painting the interior walls. He was found lying on the ground partly on the
fallen ladder, unconscious and with severe head injuries. There were no
witnesses, so the reconstruction was vital in establishing what had happened.
Unlike the previous case, the ladder had marked both the floor and the wall
during the fall. It had slipped back perpendicular to the wall, so it was
essential to determine original the angle of repose. Marks had been left by
the feet on the quarry tile floor (Figure 8.21), showing exactly where it had
originally had been leaned. The exact ladder length was also known. It turned
out to be at an angle of about 75, the correct angle for safe work, a fact
established by reconstruction shortly after the accident by his supervisor. It
was also clear that the painter had been near the top, as shown by the extent
of painting that had been completed (Figure 8.22).
What had caused the accident? The marks on the floor provided the clue
to what had happened (Figure 8.21). When soft rubber slips against a hard
surface, the only marks should be from abraded rubber, but the marks were
gray. Examination showed them to be more likely to have been formed from
the aluminum feet of the ladder. Replacing the accident ladder as in
Figure 8.23 showed that the rubber feet had worn to such an extent that they
were pushed into the hollow aluminum feet by the weight of the ladder above.
A user on the ladder would exacerbate the situation. The coefficients of
friction between aluminum or rubber and quarry tile were measured by direct
experiment and turned out to be about ~0.4 and 1.0, respectively. Static
analysis shows that the stability of any ladder decreases as the user ascends,
283
Figure 8.21 Fallen ladder on quarry tile floor with mark from foot at center of
picture.
and is most unstable when the user is near the top. So the painter had used
the ladder successfully at least once, as shown by the painting near the top
of the wall (Figure 8.22). At the second attempt, the ladder became unstable
when the painter was at or near the top, and started to slip uncontrollably,
perhaps when he was reaching out on one side to paint the wall. As the ladder
slipped the painter lost his balance and fell, his head hitting the floor and
the fallen ladder.
The painter won substantial compensation for his injuries because
employers have a responsibility to provide employees with safe and reliable
equipment needed to perform their work. The amount of compensation was
reduced by an element of contributory negligence, however. The painter
should have checked the state of the ladder, and he should not have used it
without someone footing the ladder to provide extra stability. All such points
284
Figure 8.22 The accident scene showing the unpainted wall above (top arrow)
and the slipped ladder (lower arrow). Note damage to the duct in center of wall
caused by the falling ladder.
would have been discussed at trial, but the substantial costs of trial were saved
by the agreed settlement before the case proceeded.
8.4.9
After car accidents, one of the most common causes of death and serious
injury both in the U.K. and the U.S. is accidents in which the victim falls. 3,4
285
Figure 8.23 Reconstruction with ladder feet placed on floor mark. The ladder
was erected at the correct angle, but the rubber feet have been pushed into the
ladder recesses.
So what can be done to lower the accident rate? Informing users of the
dangers of ladders and the best way to use them is one way of warning the
user of the potential dangers. Indeed, new ladders have some of the most
elaborate warning notices posted on any product. Due to the instability of
lean-to ladders, they should be stabilized wherever possible, by tying the feet
or tips to a solid support, or by having a colleague standing on the feet to
prevent slip. Ladders should never be used when the floor is wet, as this
lowers the coefficient of friction, and the floor should also be level so that
both feet are in contact. Ladders must be inspected at regular intervals, and
the critical feet and tips renewed when worn or damaged, a service many
ladder manufacturers actually provide for free. But can ladder design be
improved so as to lower the chances of slippage accidents of the kind considered here?
One of the problems of the accident ladders is that both possessed
fixed attitude feet, so that any deviation from a repose angle of 75 lowered
the contact area with the ground. Since the feet are the key to the stability
of ladders, one solution is to provide feet that adjust as the angle of repose
is changed (Figure 8.24). Such articulated feet are thus always in complete
contact with the ground and provide an extra margin of safety. The design
of the plastic tips and feet can also be improved for ladders of the specific
design considered here. Although the tip broke during the first accident
rather than causing it, the design possesses several stress concentrators:
one by design (sharp corners), the other from manufacture (internal void).
Both could easily be eliminated, the first by simply smoothing the corners
of the metal molding tools, the other by more careful control of the
molding process. Both procedures would increase the strength of the
286
References
1. Specification for portable aluminum ladders, steps, trestles, and lightweight
stagings, BS 2037, 1994.
2. Hepburn, H.A., Portable ladders. I. The quarter length rule, Br. J. Industrial
Safety, 4, 155, 1958.
3. Goldsmith, A., The role of good design and manufacturing in preventing
ladder accidents, J. Products Liability, 8, 127, 1985.
4. Betzen, F.E. and Jones, P.L., Potential of ANSI safety requirements for portable
metal ladders, J. Products Liability, 6, 7, 1983.
Failure of Medical
Implements
9.1 Introduction
In society today access to adequate health care is considered paramount, with
extensive use being made of state-of-the-art technological methods, novel
instruments, medical aids and biocompatible implant materials and devices.
Such technological support has spread to almost all sections of medical
science, together with areas as diverse as prevention, diagnosis, therapy and
rehabilitation.
The introduction to Chapter 2 likened the work of the forensic engineer
to the medical pathologist, drawing analogies between failed mechanical
components and the human body. This relationship can be further reinforced
when considering the relatively new and fast developing area of engineering
in medicine. Here, engineering/mechanical principles are employed to solve
problems that arise in areas of biology and medicine. In short, engineered
components are being designed and used to replace parts of living systems
in the human body. The range and variety of engineered components utilized
in medicine are enormous and include devices such as turbines to pump
blood, heart pacemakers, drug delivery systems, joint replacement devices,
etc. Materials used for the fabrication of such components are designed to
operate in intimate contact with living tissue and have minimal adverse
reaction with the body system. However, there are only limited ranges of
materials that exhibit characteristics desirable for in vivo service and these
materials have been termed biomaterials. There are ranges of relevant standards (ASTM, BS and ISO, for example) that describe required material
characteristics and mechanical response, with the focus firmly on corrosion
resistance of metallic materials. Nonetheless, biomaterials are not limited to
287
288
metallic systems: polymeric and ceramic materials are also utilized for in vivo
service. Components or devices engineered from biomaterials are referred to
as biomedical or implant devices.
289
Performance
Functional characteristics
Intended in-service conditions of use
Design Attributes
290
When a patient has been diagnosed with a diseased or damaged hip joint,
total hip replacement is accepted as a standard method of treatment. However, complications may arise from in-service (usually fatigue) failure of the
291
Figure 9.1 Total replacement hip device. Sketch of device (left) and x-ray of
device in situ, with expected location of fracture arrowed (right). The wire was
added during the surgical procedure to prevent the femur from splitting.
292
Figure 9.2 Optical micrograph of fracture surface on femoral stem of total hip
device. Crack growth from medial corner (A), and clamshell markings emanating
from lateral corner (B).
and exhibited clam-shell markings typically found on fatigue fracture surfaces (Figure 9.2). The boundary between the two crack fronts was identified
by a step near the central portion of the trapezoidal cross section.
Based on microscopic observation and stress analysis, Rimnac proposed
the following sequence of events to explain the fracture of the femoral stem.
Residual tensile stresses induced by large compressive overload on the medial
side of the stem resulted in the growth of a fatigue crack prior to a second
crack nucleation on the lateral side. With crack growth, there is a progressive
reduction in residual tensile stresses, along with an increase in the extent of
crack closure, ahead of a fatigue crack subjected to cyclic compressive loads.
As a result, the medial cracks slowed down or stopped after a certain amount
of crack advance. The nucleation of medial cracking increased the stress level
to which the lateral side of the stem was exposed. With time and cyclic loading
in service, the two crack fronts eventually met up. This coalescence of crack
fronts produced a step on the fracture surface. Rimnac et al. concluded that
the failure of the femoral stem was a consequence of the combined effects of
the trapezoidal cross-sectional design and of the materials properties of the
stainless steel used to fabricate the component.
9.4.2
Total hip prostheses do have a finite lifetime within the body environment
and will therefore simply wear out with time. This can be aptly demonstrated
by the following case where a person falling had apparently caused breakage
of a total hip prosthesis.
293
An elderly woman had a fall when she went to answer the front door of
her council house. Her foot went through a floorboard that had been weakened by dry rot, causing her to tumble. When help arrived, it was quickly
recognized that the woman had damaged her right hip and required urgent
medical treatment. At the hospital, it became apparent that the patient had
been the recipient of a total hip replacement some 10 years earlier and the
device had broken. A surgical procedure was promptly undertaken to retrieve
the broken prosthesis and insert a replacement device. After a full recovery,
the patient instigated legal proceedings against her landlord (the local council) for her injury and suffering. Although the woman stated that her fall had
not been particularly heavy, she believed that her injury was a direct result
of the fall. The legal action was for a claim of negligence in maintaining her
floor to a reasonable standard, resulting in a fall that broke her hip prosthesis.
The broken femoral stem of the device was submitted for investigation
Figure 9.3A, and is shown with the surgical extraction tool having been left
in situ. Figure 9.3B, shows an optical micrograph of the fracture surface,
where striations associated with a fatigue failure mechanism are clearly visible. Fatigue is a mechanism that often involves many hundreds, if not thousands, of loading cycles to reach the point of failure (Chapter 2, Section 2.4)
and is not an overload mechanism. If, as claimed, the hip prosthesis had
suffered an overload failure resulting from a fall, the loading pattern would
present a totally different fracture surface (Chapter 2, Section 2.3).
To date, the life expectancy of a total hip prosthesis is approximately 10
years. When the age of the device in question was taken into consideration and
combined with evidence of failure resulting from fatigue, it was determined
that the device had simply reached the end of its useful service lifetime and
failure was not a direct result of the fall. The loading experienced at the instant
of impact would in all probability have been withstood by the device if it had
not been considerably weakened by the presence of a growing fatigue crack.
However, the woman had suffered a fall that was a result of poor maintenance practice by her landlord. The litigation was settled out of court by
the landlords insurers, with the woman receiving substantially reduced damages for the incident.
9.4.3
294
Figure 9.3 Broken femoral stem of total hip prosthesis. The extraction tool
protruding vertically out of fracture surface (A), and fracture surface clearly
showing fatigue striations (B).
providing a more secure bond. Oil residue that was thought to have originated from the manufacturing process had not been completely removed
prior to the total hip replacement surgery.
The shell is implanted into the upper part of the hip (acetabulum) and
under normal conditions, the bone would form an integrated bond with the
shell. However, when surface contaminants are present, bone does not always
bond with the shell.
With mineral oil as the contaminant, small amounts of the residual
lubricant will be broken down and eliminated by the body. At the time of
this writing, the only known problem resulting from such contamination is
295
Figure 9.4 In situ acetabular ball and shell of a total hip replacement device
(center).
the inability of the bone to bond with the affected hip implant. However, the
result of such a deficiency is a device that will come loose from the bone.
Reported symptoms would include severe groin pain and the inability to bear
weight on the affected leg. The only solution that would provide effective
relief for the patient is revision surgery.
At the time of the recall, it was anticipated that some 17,500 patients had
been fitted with the suspect shells but that there were only some 61 reported
cases of patients experiencing loosening of the shell. However, in June 2001
it was reported that the number had increased to more than 1,850 revision
surgeries.
Thus a simple process failure, machining oil contamination at the production stage, can have far-reaching and dire consequences for a large number
of individuals. What was initially considered possible contamination of a small
number of implant shells was subsequently found to be a major failure of
performance and design attributes of the device. Although this case had not
been resolved and not all the facts have been made public at the time of this
writing, to the inquiring mind it does raise some obvious questions concerning product integrity. Furthermore, this type of failure must also bring into
question the reliability of quality control procedures in place on the production line.
9.4.4
296
of the stem was changed from matte to polished. During 1992 the product
range was extended to include the option of a polished conventional stainless
steel stem, and a polished high nitrogen stainless steel stem was added to the
range in 2001.
In the U.K., the Medical Devices Agency (MDA) had become aware of
28 revisions, due to recipient pain, where corrosion was confirmed in the
polished titanium alloy femoral stem. The incidence of these revisions was
spread across nine different centers, and represented a confirmed failure rate
of 0.14%. In addition, the MDA became aware of eight additional possible
cases, of which five were considered by the clinicians involved to have been
resolved without the need for further surgery. Since 1999 the number of
reported cases of corrosion of the polished titanium stem has continued to
increase.
The most common clinical symptom signaling potential problems was
severe atypical pain at 2 to 5 years post-implantation, along with an absence
of overt infection or femoral loosening. However, it would appear that pain
relief could be achieved upon movement. Additionally, corrosion has been
seen along the full length of the explanted stems, but in some cases it is
focused on the distal half of the stem.
A most interesting aspect of the problem is that extensive testing in five
independent biomaterials laboratories has failed to establish categorically
why the polished titanium alloy modular femoral stem sometimes corrodes
in this manner. Furthermore, the MDA has no information of any case of
corrosion in the original matte-finished titanium alloy stem or in either of
the stainless steel variants. Although presenting manufacturing difficulties,
titanium alloy is generally considered to be highly corrosion resistant in vivo.
One probable answer is that the matte surface would bond better with the
bone and resist any physical actions that would damage the TiO 2 surface film.
It is this film that gives titanium its corrosion resistance, as the metal itself
is highly reactive chemically.
In addition, the elastic modulus of titanium is a closer match to that of
bone than the range of alternative bio-alloy systems. Consequently, from a
purely mechanistic view, titanium alloys are gaining popularity for implant
device fabrication. Nevertheless, this particular case does raise questions as
to the performance attributes of polished titanium prosthetic devices.
A review of the published literature24 describes sporadic cases of corrosion of other models of polished and unpolished cemented titanium alloy
hip implants. However, the polished stem in question is currently the only
titanium alloy hip implant repeatedly exhibiting this corrosion problem in
the U.K. Accordingly, the MDA issued a device alert MDA DA2002 (04),
dated May 2002 for the device in question.
297
298
Figure 9.5 ESEM image of failed distal end of TPE catheter. The failure surface
cut across the drug infusion hole nearest the proximal end shows both brittle
behavior and ductility.
Figure 9.6 ESEM micrograph of fracture surface at the proximal end of the
catheter, showing completely brittle and ductile behavior.
Extensive reports of breast implant failures arose in the U.S. in the late 1980s.
Most of the failures occurred from permanent implants made from silicone
rubber filled with silicone gel. In the case we were asked to investigate, a silicone
breast tissue expander failed suddenly after several infusions of saline solution,
299
Ductile Shear
Lips
Inner
Lip
Longitudinal
Scratches
Tear
Zone
Brittle Fracture
Zone
Hole
Ductile Fiber
Main Ductile
Fracture Tip
Shear
Lips
Ductile
Fiber
Brittle
Fracture
Zone
Possible Origin
(Mirror Zone)
Figure 9.7 Fracture surface map of failed distal end of catheter, describing main
surface features and possible origin (sketch magnification: 50).
Figure 9.8 SEM fractograph of a cut surface on a new TPE catheter, the cut
being deliberately made with a Tuohy needle (a hollow steel needle with sharp
tip for insertion into the back). Defects in the edge of the needle leave striations
in the catheter surface.
causing great distress to the patient.6 The device in question was designed to
stretch muscle and other chest tissues prior to implantation of a permanent
prosthesis. The operation is commonly performed after mastectomy for
300
Figure 9.9 General view of breast tissue expander, with dome at the left for
filling via hypodermic needle.
301
Volume of Saline
Added, ml
Comments
Aug. 8
250 cc
Aug. 22
50 cc
Sept. 5
100 cc
Sept. 19
170 cc
Oct. 3
100 cc
Oct. 10
150 cc
post-Oct. 17
expander bag, it was surmised that all saline fills after this event would
simply have leaked into the chest, and been dissipated into the patients
body (Table 9.1).
This appeared to be a clear-cut case of product liability, with a faulty
device unknowingly being implanted by the consultant. However, medical
cases where liability seems at first sight to be straightforward tend to become
less clear as the cases proceed through litigation. In particular, defendants
will usually allege that medical negligence was involved. In this case, it might
be claimed that the consultant had wrongly placed the device, putting excessive load on the device. There was no evidence in fact for this allegation.
302
Upper
Shoulder
C1-C3
Cn Formed Here
Main Fracture
Surface
Microcracks
Cp
Lower
Shoulder
Crack in
Adhesive
Layer
Tube
Microcracks
1000 mm
Figure 9.12 Finer cracks at the interface of the tube and the shoulder (ESEM).
Nevertheless, it was determined that the consultant had made a simple arithmetical error in her notes (Table 9.1). Furthermore, the consultant was
apparently unaware of any lack of expansion of the bag, even when the added
saline exceeded the nominal fill volume, some 21 days prior to termination
of the procedure. However, the patient had developed a strong relationship
with her consultant, and refused to accept the implications of the facts of the
case as shown by the hospital records. The case thus did not proceed, and
the patient received no compensation for her undoubted trauma.
303
Figure 9.13 (A) Recovered needle with waisted section arrowed. (B) Closer view
of threading eye and tapering point of circular cross section mounted on Blu-Tack.
304
Regular Eye
Spring Eye
Swaged Eye
Suture
Sewing Eye
Figure 9.14 Morphology of different needle eyes. Regular, spring and swaged
are of a surgical variety, and sewing is the variety used by seamstresses.
extreme tip was slightly bent, and external surfaces had originally been coated
with a bright metal; patchy areas of corrosion were visible all along the body
of the needle. X-ray analysis in the SEM revealed the needle to be unalloyed,
plain carbon steel, with no chromium or nickel within its bulk that would
classify it as a stainless variety. However, there were patches of a nickel
electroplated surface coating commonly used to form bright, protective coatings on articles made from plain carbon and low alloy steels. This was a
surface coating, and not a significant ingredient in the steel from which the
needle was made.
Surgical needles are invariably made from stainless steel. Many are curved
and some are semicircular. Points are usually of a triangular cross section to
produce razor sharp edges having a cutting action for piercing tissue. The
needle in question had none of these features.
Suturing threads in surgical needles are usually single filaments secured
by swaging into the end of the needle, although some do have eyes
(Figure 9.14). The needle under observation had an elongated eye with
grooves on either side to minimize the effect of the double thickness of thread
when passing through material. Eyes of surgical needles that do not have
swaged threads are far less elongated, almost circular, as they only have to
accommodate a filament as distinct from a woven thread. Also, the surfaces
on either side of the eye are ground flat so that the thickness of the eye is
less than the full diameter of the body of the needle. Again, the needle in
question did not exhibit the features associated with a surgical needle.
Since surgical needles are made from stainless steel, there is no need for
any protective coating. However, unalloyed plain carbon steel rusts rapidly
in moist air and so ordinary sewing and darning needles are electroplated
with nickel to maintain their bright appearance. It has to be emphasized that
stainless steels retain this property only when there is a free access of oxygen
305
Outcome
In the first instance, it had been suggested to the patient that monetary
compensation for her perceived injury would be in the region of 50,000 (ca.
1998) sufficient motive to confuse an ordinary sewing needle with the
surgical variety. However, when reviewing this case, it became perfectly clear
that design attributes required of surgical needles were a decisive factor when
clarifying the exact nature of the needle in question. As a final afterthought
to the case, it transpired that the female patient had been an enthusiastic
seamstress for a number of years, thereby providing a possible scenario as to
how a sewing needle could have entered her thigh.
306
A
Air Out
Sight Tube
Float
B
Air from
Patient
Figure 9.15 Cross section of the breathing apparatus (A), and transparent sight
tube (B) showing molding defects in shadow.
Outcome
307
2. Some marks on the molding could only have been caused by a blunt
instrument (such as a screwdriver tip) levering against the precision
tool surfaces (Figure 9.16). However, these markings could in no way
be responsible for the variation of process conditions observed.
The expert for the plaintiff refused to consider evidence contained on
the sample moldings, and reiterated his original views. The case collapsed
after the plaintiff was cross-examined on the witness stand on the third day
of the trial.
After specializing in production of steel-based surgical instruments, a manufacturer decided to introduce a new range of titanium microsurgical tweezers. The production lines were installed and commissioned, and the new
instrument range introduced to the market place. Fabrication of the tweezers
required an initial stamping of blanks from sheet stock, followed by a pressing
sequence to form a dog-leg (double bend) toward one end (Figure 9.17A).
This dog-leg section was then riveted to a mirror image blank, to form the
final tweezer profile. In addition, a dog-leg geometry provided the spring
necessary for normal operation.
Production and sales went well for the first year and, in order to increase
output, it was decided to introduce larger titanium sheets for the initial
blanking process. Although this pressing operation had been undertaken on
308
Figure 9.17 (A) Press formed double bend on part fabricated titanium surgical
tweezer blanks. (B) In-plane crack in the arm thickness of the partly fabricated
tweezers.
many occasions over the previous year, and with no unforeseen problems,
blanks from the larger titanium sheet cracked in an in-plane mode during
the pressing operation (Figure 9.17B).
A through-thickness section, taken at the region of cracking on one of
the tweezer blanks, revealed the microstructure to be typical of heat-treated
Ti-6Al-4V alloy. The location and extent of cracking can be seen in
Figure 9.17A. Cracking was limited to within the confines of a double bend
and approximately midway through the material thickness. Bands of oriented
beta (b) structure were visible over the section (Figure 9.18A), and the inplane cracking had occurred along one such band (Figure 9.18B).
At some point in its history, the initial billet, from which the Ti-6Al-4V
sheet had been rolled, was held for too long at a temperature above its
alpha/beta (a/b) transus. This resulted in formation of large b grains. Subsequent working (forming the sheet stock) had rolled the b grains into long
pancakes. The Widmanstatten basket weave structure of the b grains became
oriented in the rolling direction, causing planes of weakness. As such, a b
structure has low ductility/formability and the double shearing mechanism,
309
310
weakness in the form of large pancake b grains. It was the shearing action
of the bending process that introduced in-plane cracks within the partially
fabricated tweezers.
From a general engineering viewpoint, the observed cracks could be
considered cosmetic and would be unlikely to lead to failure in normal use.
However, when considering the intended performance and design characteristics, in addition to product liability, it was considered prudent to scrap any
partially or fully formed tweezers that had been fabricated from the titanium
sheet in question.
It transpired that the tweezer manufacturer had not specified a desired
material grain size to his stockist. It may well have been that he had no
knowledge of the need for such a specification. However, had the manufacturer requested the stockist to advise and provide appropriate material for
such an application-specific use, he might have had cause for complaint in
that the sheet supplied was not suitable for his particular purpose.
9.8.2
311
Figure 9.19 (A) Part fabricated ophthalmic tweezers broken at section change.
(B) Sketch of crack initiation points and propagation paths.
312
313
314
Figure 9.21 Site of fracture of plastic connector near junction of two aluminum
tubes.
Figure 9.22 Defect in plastic connector (top arrow) and wear pattern (lower
arrow) on aluminum crutch tube.
315
The injured woman received compensation for the trauma and injury
after the manufacturer accepted the findings of our investigation.
Heat Shrink
Tubing
Signal Cable
Figure 9.23 Needle electrode schematic, showing heat shrink tubing over a
soldered joint between needle probe and signal cable.
316
to protect the solder joint were minimal, relying on polymeric tubing that
had been heat-shrunk over the rear of the needle and approximately the first
20 mm of lead wire. As a solution to the problem of strain relief, it could be
considered minimal. There are far more efficient mechanical routes for negating the transmission of load to a solder joint.
When considering soldering as a fabrication route, it should be remembered that the process has been in use since Roman times. More recently, the
1950s saw the start of an explosion in the use of soldering as a fabrication
process it was the advent of todays electronics industry. The point being
made here is that soldering is a tried and tested manufacturing process.
Furthermore, when working as a designer or manufacturing engineer, training will develop awareness of the normal or expected location of a failure
in any type or part of the system (the weakest link principle). Any deviation
from the normal or expected location will have been caused by additional
(unknown) factors. When considering design and manufacture of the needle
probe, the weakest link will be its soldered joint, this being true for virtually
all assemblies having soldered connections. It would be safe to say therefore
that the probe failed at an expected location.
9.10.1 Outcome
The defense by the manufacturer was based on the so-called development
risks, where new technologies carry an increased probability of unknown
effects or failure. Without this form of protection, manufacturers would not
be willing to shoulder the burden of introducing new technologies to the
market place. However, although the technology of magnetic nerve stimulation may be considered novel, it was not the technology that failed, and
therefore the technology was not at issue. What was at issue was a design or
manufacturing failure. The crux of the matter was whether or not the solder
joint failure could have been foreseeable.
As a soldered joint should be recognized as a weak link by both designer
and manufacturer, the needle failure would be considered a foreseeable event.
It was not the technology of nerve stimulation that failed during the surgical
procedure; rather, it was the tried and tested manufacturing route of soldering that was victim to failure. Based on these observations, any defense
focused around development risk was considered null and void. The manufacturer conceded, paying the majority of costs involved before the case
went to trial.
317
References
1. Rimnac, C.M. et al., Failure analysis of a total hip femoral component: a
fracture mechanics approach, in Case Histories Involving Fatigue and Fracture
Mechanics, Special Technical Publication 918, ASTM, Philadelphia, 1986,
pp. 377388.
2. Willert, H.G. et al., Crevice corrosion of cemented titanium-alloy stems in
total hip replacements, Clin. Orthopaed. Relat. Res., 333, 5173, 1996.
3. Schll, E. et al., Osteolysis in cemented titanium alloy hip prosthesis,
J. Arthroplasty, 15(5), 570575, 2000.
4. Shahgaldi, B.F. et al., Effect of corrosion on tissue in contact with orthopaedic
implants, J. Bone Joint Surg., 82-B, 2000; Orthopaedic Proceedings Supplement II, 122.
5. Lewis, P.R., Fracture of a TPE catheter, oral presentation, Failure Analysis and
Prevention Special Interest Group of the SPE, ANTEC, Dallas, 2001.
6. Lewis, P.R., Fracture of a breast tissue expander, oral presentation, FAPSIG
at ANTEC, San Francisco, 2002.
Component Failure
in Road Traffic
Accidents
10
319
320
Figure 10.1 Four exhibits from the inquest showing damage to the truck and
one of the wheels.
Figure 10.2
321
Side view of steering box with broken end of sector shaft resting
in position.
Figure 10.3
Side view of steering box with broken end of sector shaft removed.
outside with a roughened area at the center (Figure 10.4) and the break had
occurred immediately below the bottom bush in the steering box where
bending forces would have been greatest. Figure 10.5 shows how the steering
box was mounted on the chassis. Otherwise the vehicle had been well maintained and was in a fully roadworthy condition prior to the accident.
322
Figure 10.4 Fracture surface of splined shaft seen from four different directions, showing distortion of every spline at edge of fracture.
Figure 10.5 Arrangement of linkages and drop arm when steering box is
mounted on chassis. Notice the drop arm and drag link project below chassis.
The drivers father stated that his son was 29 years old and had worked
as a heavy truck driver for 5 years, during which time he had never been
involved in any serious accident or had any convictions. The day before he
made the trip he had played in a cricket match but had gone to bed early
because he had a long journey the next day. He occasionally drank a pint of
beer for refreshment, but was not a heavy drinker and valued his license too
323
324
for the vehicle leaving the road as the driver would have had no steering
control. If the braking system had been damaged by collision with the crash
barrier it would not have been possible to stop the vehicle before it reached
the bridge. There were no tire marks to suggest the brakes had been applied
prior to collision with the barrier.
A legal representative of the vehicle manufacturers questioned this witness and asked if he had noticed any twisting of the splines of the steering
shaft. The witness replied that he had observed they were bent noticeably on
both sides of the fracture (see Figure 10.4); if the splines had been in this
condition before the accident they would have altered the position of the
steering wheel in the straight ahead position. When asked if it was likely
that this shaft had been twisted for months or weeks, the witness replied that
if it had, he would have expected someone driving the vehicle regularly to
have noticed it, unless the steering had been completely reset immediately
after the splines had been bent.
This witness was further questioned about the possibility that the shaft
might have been cracked before the accident. He replied that if the shaft had
been held together by the small area at the axis he would have expected it to
have failed suddenly on the road, even if the vehicle were unladen. He would
not expect such a small segment of metal to be able to withstand heavy
steering loads such as those experienced in maneuvering a fully laden vehicle.
Subsequent to his examination at the scene, he had taken the wheel and tire
to the forensic science laboratory for expert examination, along with the
steering box and broken sector shaft.
The forensic scientist was a highly qualified chemist with 20 years experience in forensic science and an expert in tire examination. He had examined
the parts submitted to the laboratory. He stated his findings could not possibly support the theory that premature failure or deflation of the tire caused
the accident. This witness had also examined the steering box and sector
shaft. The extract from his report read to the court was as follows:
Rocker shaft. The fracture surfaces of the two pieces were
examined and found to exhibit characteristics typical of
fatigue failure. It was apparent the components had been
operating for some time on the central quarter-inch of
metal, but after a cursory examination it is not possible
to state whether the final failure was spontaneous or the
result of impact.
A verdict of accidental death was recorded. The witnesses statements
all supported the view that the driver lost control because the steering arm
suffered a fatigue failure.
325
326
327
Figure 10.6 Fracture surface of splined shaft failed in torsional fatigue (right),
alongside one that failed in bending (left).
the radius so that, when there is only a small area of sound metal near the
center, the stress level under normal loads is much greater than when the
cracking is still in the outer, splined region. It could be argued that these
failures were the result of poor design, in that the shafts were not strong
enough to withstand foreseeable loadings for the life of the vehicle; however,
the age and uncertain history of the vehicles by the time such fractures began
to be experienced, together with the evidence of poor maintenance, did not
provide a case strong enough for a successful claim against the manufacturers.
Splined shafts with square section splines follow a similar pattern but
the width of the splines gives the final fracture a more block-like appearance
and, because some of the first splines start to fatigue in two positions the
blocks are sometimes pulled out of line when the final, ductile torsional
overload occurs at the axis.
The fracture of the steering shaft at the left in Figure 10.7 is a bending
fatigue failure of a long shaft that directly connected the steering wheel to
the steering box on a vintage car (from the days when a frontal collision
could drive the steering column like a spear into the drivers chest). When
restoring the car the owner had managed to obtain a steering box and shaft
from a similar vehicle that was being cannibalized for spares. He fitted the
parts to his car and subsequently exhibited it at meetings and took part in
road rallies. In one of these the steering column suffered a fatigue fracture
while rounding a bend at the bottom of a hill. A spectator was badly injured.
The shaft must have been slightly bent when fitted, with the result that every
time the steering wheel was turned through more than 180 the middle of
the column went through a cycle of compression and tension. Eventually the
inevitable fatigue failure occurred.
328
329
maneuvering or moving slowly, when the bending loads on the inboard end
of the axle are considerably greater than when it is traveling normally on the
road. A typical fracture surface of a steel stub axle shows two fatigue crack
initiation sites, a larger area radiating from the bottommost fiber and a much
smaller one radiating from the uppermost fiber. The final break is a ductile
fracture occupying a band across the middle.
The reason for this unequal development of the fatigue cracking is that
tensile stresses acting on the bottommost fibers are caused by the weight of
the vehicle tending to bend the axle upward plus cornering forces acting to
push the bottom of the wheel outward. Both are of a cyclic nature when the
vehicle is traveling on the road but the cornering force only becomes significant when the weight of the vehicle is transferred to the axle as it changes
direction. However, the stresses on the uppermost fibers are always lower,
because there is no external gravitational force bending the axle upward as
there is along the bottom of the axle; rather it is the opposite effect
gravitational force produces compressive stress in the uppermost fibers and
so tends to counteract tensile forces when the vehicle is cornering. The net
result is that the fatigue initiates first at the bottom of the axle and propagates
upward and it is only when the cross-sectional area has been reduced and
the applied loads generate higher stress levels that the cornering forces are
able to initiate a fatigue crack at the top of the axle. This is also the reason
why the final break almost always takes place when cornering, usually at low
speed, because the upward bending forces are greatest when a heavily laden
vehicle is maneuvering.
Torsional overstress which twists the retaining nut off the end of an axle
is a common cause of wheel detachment when the vehicle is traveling at high
speed and usually occurs shortly after wheel bearings have been adjusted
during servicing. In the majority of motor vehicles wheels are mounted on
their axles by a pair of taper roller bearings that must to be critically adjusted
to allow a small amount of free play without becoming sloppy. Some front
wheel drive vehicles are fitted with a pair of ball bearings separated by a
tubular spacer clamped tightly together and the only provision for adjustment is to place shims at the ends of the spacer. The retaining nut has to be
tightened to a high torque. Some wheel detachments have occurred simply
because a mechanic has mistakenly tightened taper roller bearings.
Taper bearings are retained on the axle by a washer and castellated nut
which has to be carefully set to allow a small amount of free play and is
prevented from coming unscrewed by a cotter pin or C clip passing through
a hole in the screwed end of the axle and lining up with opposite castellations
of the nut. The method of making this adjustment is to fit the roller bearings
and tighten them until they bind, then release the nut by one sixth of a turn
and check that the wheel turns freely with minimum end float. The nut is
330
then locked in this position, or the nearest pair of castellations in the unscrewing direction, by inserting the cotter or spring pin through the castellations.
When disc brakes are fitted to a vehicle the friction pads and associated
parts may be examined without removing the wheel bearings. However, in
order to inspect the friction linings and expander of drum brakes the drum
must be removed. On small cars this is usually integral with the hub as a
single casting, which necessitates taking off the castle nut and withdrawing
the outer taper roller bearing. After the brakes have been attended to, and
usually the bearings regreased, the hub/drum unit is replaced and the bearings have to be adjusted to the required amount of free play. This is where
the problem is likely to arise. If the mechanic replaces the nut and turns the
castellation in the tightening direction to line up the castellations with the
hole in the axle (or, as has been known to happen, tightens them fully as if
they were ball bearings with a spacer), then the bearings will heat up when
the vehicle goes back on the road. Within about an hour or so of high-speed
driving, the bearing on the right side of the vehicle seizes and twists off the
end of the axle to produce the torsional shear failure illustrated in Figure 10.8.
The reason that it is the wheel on the right-hand side of the vehicle that
detaches is because the direction of rotation in forward motion tends to
tighten the castle nut as the bearings expand under frictional heat. This
quickly results in the castle nut binding tightly as the free play is reduced and
the friction may then cause the castellations to shear through the cotter. When
this happens the bearing locks solid, the threaded end of the axle shears under
torsional overload and the wheel detaches, carrying the hub with it.
Although similar frictional heating may occur if bearings on an axle at
the left side of the vehicle have insufficient free play, the direction of wheel
331
rotation as the friction increases acts to unscrew the nut, which tends to relax
the frictional force compared with positive tightening on the right axle. The
driver feels no immediate effect as the wheel detaches but first becomes aware
of the problem when a wheel overtakes the side of the car! A loose wheel
rolling into the path of another vehicle traveling at high speed often has fatal
results, certainly in countries that drive on the left side of the road. The clues
as to the cause of the axle shearing are readily apparent in the state of the
outer bearing, the nut and cotter and the signs of frictional heating in the
vicinity of the bearing, as illustrated in Figure 10.9.
Wheel detachment on both right and left sides of the vehicle may also
occur if worn or faulty bearings fail and pieces of a broken roller or parts of
the cage jam and cause seizure. This can usually be traced by careful examination of the debris to a worn out or damaged bearing. The detachment
illustrated in Figures 10.8 and 10.9 occurred on a small car that had undergone a 20,000 mile service only the day before, and this had required removal
of the rear brake drums to inspect the linings. The young mechanic who had
carried out the work was not supervised so full liability for the accident was
accepted by the garages insurers.
332
Figure 10.10 Pipe connecting brake expander inside front brake drum to
hydraulic circuit.
unnoticed, all the fluid eventually escapes and the brakes become inoperative.
The brake pipe failure discussed in Chapter 6 (Section 6.3.3) was an example
of corrosion-initiated failure. Figure 10.10 shows another example of external
corrosive attack. The following case is an example of fatigue failure.
10.3.1 Brake Failure by Fatigue
The failure mechanism outlined in Section 6.3.3 was due to a corrosion cell
being set up between metals of markedly different electrode potentials. This
problem may be overcome quite simply by making the pipe from a homogeneous alloy with substantially better resistance to corrosion and thus needs
no protective coating. Such an alloy has been used for many years in upscale
cars and is as formable as Bundy pipe (see Figure 6.3), although rather more
expensive. However, this alloy is just as vulnerable to fatigue, with failure
usually occurring where the ends of the pipe are swaged within couplings to
fixed components in the braking system, such as expander cylinders and the
pistons of disc brakes. Fatigue is likely to occur if there is relative movement,
which sets up cyclic bending or torsional stresses at the root of the flange
inside a union where the pipe is connected.
Figure 10.11 shows a short length of pipe from a buggy-type vehicle
used for towing strings of luggage trailers at an airport. The brakes suddenly
failed, causing severe injury to one of the loaders and some damage to an
aircraft. It was found that all the hydraulic fluid had been lost from the
reservoir during the 72 h since the level had last been checked. The maintenance log at that service recorded the level to be slightly low, so a minor top
up had been required. Cursory inspection of the underside of the vehicle at
that time had revealed no signs of leakage.
After the accident it was discovered that fluid had been escaping between
the pipe and the union connecting it to the brake expander mounted at the
333
Figure 10.11
Figure 10.12
Close-up of swaged end with crack just below the curved bulb.
top of the brake plate of the front wheel. The escaping fluid had simply run
down the pipe and dripped off the bend at the bottom, so there was practically
nothing to show how much fluid had been lost. The escape had occurred
through a circumferential crack in the pipe immediately below the swaged
flange, as marked by the arrows in Figures 10.11 and 10.12. The crack
extended approximately 240 around the circumference, exactly following
the ridge left by the swaging tool when the flange was formed at the end of
the pipe. A section along a diameter was removed as shown in Figure 10.13,
which enabled the completely cracked part of the flange to be separated. The
fracture face exhibited characteristic features of fatigue, showing the mark
left by the swaging tool to be the initiation site. On the other half of this
section still attached to the pipe, the lower limb in the figure is cracked right
through but the opposite side shows the end of the crack penetrating only
about 15% of the wall.
334
Figure 10.13 Section through fatigue crack. It can be seen as a faint line below
the bulb running across the wall.
When the expander cylinder mounted on the inside of the brake plate
was inspected it was discovered that the retaining nut was loose. This caused
the cylinder to be twisted sideways every time the brake shoes came into
contact with the rotating brake drum, whether the wheel was rotating forward
or backward. Cyclic stresses on the pipe were concentrated at the groove left
by the swaging tool directly behind the flange. A fatigue crack was initiated
and began to propagate, following the groove. Although the stress cycles
would have been essentially torsional in nature at first, the sharp change in
section immediately below where the flange was clamped dominated the
direction of crack propagation. Later on, as the one-sided crack development
continued, bending stresses from the internal hydraulic pressure would cause
build up of axial stress and accelerate the rate of crack propagation.
The fatigue cracking must have started some time before the accident,
but brake fluid would only have begun to escape when the crack penetrated
the wall. Gradual reduction in the cross-sectional area of the pipe wall would
result in ever-increasing stress levels, causing the crack to propagate faster
and an ever-increasing escape of brake fluid. As it was a power-assisted
braking system the driver would have been unable to detect any difference
in the feel or movement of the brake pedal. Hence the loss of all the hydraulic
fluid from the system only 3 days after topping up the reservoir is entirely
consistent with the way the fatigue cracking had propagated. The only prior
warning would have been if fluid had been noticed underneath the vehicle
but with continual use day and night this most probably would have been
spread all over the loading and baggage handling areas of the airport.
335
336
337
the end of a welding rod. Quite often the filament may not be stretched
enough to fuse like these examples but, instead, the filament is thrown forward or backward and distorts because it is so weak and ductile when incandescent. Consequently, a filament that is found to be badly distorted after a
collision, although not melted as illustrated above, must have been incandescent at the moment of impact because otherwise, if it had been cold, it
would not have possessed the necessary ductility to deform plastically.
Filament (B) in Figure 10.15 exhibits a wasted, tapering end, as distinct
from a fracture of the full cross section or a molten blob, and there is a tiny
white particle adhering near the point. This is a fragment of glass that has
fritted onto the surface. A color photograph would show this end of the wire
to be coated with a yellowish-white film of tungsten oxide. It is typical of a
filament that was incandescent when the glass envelope was broken and
allowed an ingress of air.
One small part of the filament begins to react and the oxide instantly
volatilizes, reducing the cross-sectional area carrying the current. As a
consequence this small area gets hotter than the rest of the filament and
so oxidizes ever more rapidly. The oxide condenses on some nearby cooler
surface, but meanwhile the filament continues to waste away faster at this
one spot until eventually all the metal has been converted to oxide, producing these characteristic conical-shaped ends on each side of the break.
338
The particle of fritted glass is also evidence that the filament was hot enough
to fuse it to the surface after the glass envelope had broken.
Figure 10.15D exhibits a characteristic brittle fracture when the filament
is below its brittle/ductile transition temperature. When the filament is broken in this way the coils are undistorted and there is no sign of melting,
oxidation or any reduction in cross-sectional area at the point of fracture.
Longitudinal score marks from the wire-drawing process extend right up to
the edges of the fracture. Hence when this filament was broken there was no
current flowing and its temperature must have been below 300C. It was
certainly not incandescent when the lamp suffered some kind of mechanical
shock. When filaments have been in service for a long time grain growth
occurs, which makes the filament more susceptible to mechanical shock; in
some old filaments a single grain may occupy the entire cross-sectional area.
10.4.2 Motorcycle Accident
Figure 10.16 and Figure 10.17 show the filaments from bulbs from a motorcycle headlight and a car involved in a fatal accident. In her statement to the
police the female driver of a large car said she had paused at the gateway at
the end of her driveway to check that there was no approaching traffic on
the poorly lit urban road. Seeing nothing she pulled out into the road but
was almost immediately struck at the rear by a solo motorcycle. The rider
was thrown onto the back of her car and killed instantly. The motorcycle was
wrecked. She claimed that it must have been traveling very fast and could
not have been showing lights or she would have seen it approaching.
Figure 10.16 shows the main beam and the dipped beam filaments inside
the broken bulb of the motorcycle headlight. The dipped beam filament has
a brittle fracture adjacent to the terminal post, but is bowed and stretched
to a limited extent. The main beam filament is broken and bent into a much
more pronounced curvature. There is light oxidation adjacent to the ductile
fracture, which has rounded blobs at the ends, but no significant wasting
away of the cross section. A tiny particle of glass, visible as the white spot
(arrowed), was fritted onto the filament.
The above evidence clearly establishes that the main beam filament was
incandescent when it was subjected to a force that caused it to bow into a
broad horseshoe shape. The filament fractured while it was still hot but the
glass envelope must also have been broken at the same instant. This allowed
ingress of air which oxidized the tungsten but, as the temperature of the
filament was falling, it was by then too low for extensive volatilization of the
oxide. It was however still hot enough to fuse the fragment of glass to the
surface. The brittle fracture of the dipped beam filament shows it was not
incandescent when it experienced the force but the middle part of the coil
must have been above 340C for it to bend in a ductile manner. The center
339
Figure 10.16 Bulb from motorcycle headlight showing fragment of glass adhering to filament.
part of this filament must have been heated above this temperature in the
middle of the coil by radiation from the main beam filament but the junction
at which it connected to the support post where the fracture occurred had
been below the ductile/brittle transition temperature.
The above evidence thus establishes conclusively that the motorcyclist
had been riding on main beam when the collision occurred.
Figure 10.17 shows the bulb from the rear headlight of the car, found
not to be working when the lights were tested after the accident. This is
photographed against a black background in order to show up the yellowish/white tungsten tri-oxide coating toward the ends, but there is also dark
blue di-oxide over the center left-hand side. This evidence shows that the
filament had been incandescent when the glass envelope was cracked and air
entered, but the filament cooled rapidly and did not break until its temperature had fallen below 340C. Inertia force from the collision had caused the
340
341
Figure 10.18 Damage to the front of the car caused by collision with the trailer.
Figure 10.19 Extension bracket and side arms to which the towball was attached
at the rear of the towing vehicle.
passed him to have been in the region of 60 mph and was well toward the
middle of the road, although on its own side of the central white line.
Figure 10.18 is a view of car showing its position immediately after the
collision. Figure 10.19 shows the tow ball mounting on the back of the
Landrover. This was an extension bracket from the main chassis member
to bring the tow ball level with the bumper and there were two side supports
of steel angle section, now hanging loose. The tow ball itself had been bolted
onto the extension bracket by the two high tensile steel bolts shown in
Figure 10.20. It will be noticed that both bolts are bent but the bracket
itself was only slightly deformed. The left side link seen hanging loose in
Figure 10.19 was found to be buckled when the extension bracket was
removed.
342
Figure 10.20 Bolts from the towball plate showing remains of threads stripped
from nuts. The arrow shows a piece of sheared thread.
At the inquest a police vehicle examiner said he believed the tow ball had
broken away because the bolts were of high tensile steel and had been grossly
overtightened, causing them to strip the threads in the mounting bracket
(Figure 10.20). These finally pulled out under the normal slewing action of
the trailer as it was being towed with an insecure and unevenly distributed
load. The trailer wandered over the centerline and the oncoming car was
unable to avoid a frontal collision. Figure 10.21 is the underside of the head
of one of the bolts, showing sharply defined, unrusted witness marks from
the teeth of the lock washer. The other bolt displayed similar marks, thus
establishing that both must have been tight before the accident. Slivers of
machined threads referred to by the vehicle examiner are noticeable in the
roots of the bolt at the right-hand side in Figure 10.20. A complete section
of a thread from the plate can be seen five turns below the head in the lower
bolt.
However, an independent investigator, while agreeing that the high tensile bolts had sheared threads from the softer steel bracket, disputed this
explanation for the cause of the collision on the grounds that if the forces
343
Figure 10.21 Witness marks (arrows) from the lock washer on the underside
of the bolt head.
had been solely side-to-side slewing when the bolts finally pulled out, this
would not account for the way the plate itself and one of the side links had
buckled, nor for the bending of both bolts in the same direction just under
their heads. Many other Landrover vehicles (including those operated by the
same police force for recovery purposes) were fitted with this same type of
towing bracket and no similar failure had ever been encountered. He considered this damage could only be accounted for if the front corner of the
trailer had been pushed back violently while still coupled to the towing vehicle
and reasonably in line with the direction of travel.
This was the view accepted in the subsequent litigation between the
insurance companies, where the major issue was whether or not the trailer
had detached before the collision and had veered across the road into the
path of oncoming traffic. It was agreed that because the trailer was 40 cm
wider than the towing vehicle the front corner would have been projecting
almost to the middle of what was a fairly narrow main road. The oncoming
car was traveling fast and its front wheel collided with the corner of the trailer,
pushing it back violently and tearing the tow ball plate off the bracket. This
accounts for the buckling of the left side arm and why both bolts bent in the
same direction as they ripped threads of the nuts retaining the towball. The
collision forces must have been greater than any previously applied up to
that moment of impact. The damage to the car was all along its offside, none
near the center of the hood as would have occurred in a frontal collision.
Loose scaffolding on the trailer had been projected through the windshield
on the drivers side only. The offside wheel of the trailer had been torn from
its axle but there was no significant damage to the tow hitch still attached at
the front.
344
In light of this evidence it was decided that the driver of the car had been
primarily responsible for this accident. There was a suggestion he might have
been using a mobile telephone as he went by the vehicle in front of the
Landrover.
10.5.2 Tow Bolt Fatigue
The second case concerns a large twin-axle skeletal trailer carrying a portable
workshop that was being towed by a long wheel base truck on a motorway.
(A skeletal trailer is a flat chassis frame fitted with tow hitch and the normal
road-running equipment, onto which box-shaped containers may be
attached by means of turnbuckles.) Witnesses stated that the trailer was
traveling at about 60 mph when it began to swing violently from side to side,
detached from the truck, overturned, slid across the roadway on its side,
overturned again as it struck the crash barrier and came to rest in the
overtaking lane of the opposite roadway. Two cars collided with it, one caught
fire and there were several fatalities. The cause of the detachment was that a
high tensile steel bolt inside the hydraulically damped hitch had broken,
allowing the socket end clamping the tow pin to pull out.
Maintenance records revealed a history of problems with this particular
trailer. On three separate occasions cracks found in the A frame had been
repaired by welding. These were attributed to twisting and bottoming out
of the trailer as it was towed over rough ground on building sites. The owners
had been warned that the truck used for towing this trailer was unsuitable
as it had a long overhang behind the rear axle that placed excessive forces on
the tow hitch, as illustrated in Figure 10.22. A new tow hitch had been fitted
only 10 weeks before the accident, yet 6 weeks later two of the bolts attaching
it to the frame were found to be loose and wearing and had been replaced
by drilling out the holes and fitting a larger diameter bolt. Even more disShort Overhang
Long Overhang
Figure 10.22 Diagrams showing how tow ball overhang affects loading on the
tow hitch when traveling over uneven ground.
345
Figure 10.23 Parts of the hitch involved in the accident. The socket end still
clamping the ball on the tow pin had pulled off the damper tube in the hitch body.
Figure 10.24 Closer view of parts that separated. The fractured end of the bolt
is visible through the hole in the damper tube at right, which fit into the tube
at left. The nut and missing end were not found after the accident.
turbing was a report that one of the bolts inside the hitch had broken and
had been replaced on site.
Figure 10.23 shows the hitch and tow pin as recovered after the accident.
The bolts which can be seen at the right are those that held it to the trailer
frame and those at the left were from the towpin attached to the truck.
Attention is drawn to the as-new appearance of the aluminum castings and
the bored-out holes for the larger bolts at the right-hand side. The tow pin
is dual purpose in that it has a ball at the top for a socket clamp and a pin
below for attaching a trailer having an eye ring. There are two other bolts in
the socket end of the hitch, marked by the large arrow, which pulled out at
the time of the accident. These two components are shown in Figure 10.24
and were held together by two high tensile bolts, one vertical (at the right)
and the other (at the left) horizontal. The nyloc nut (a locking nut with nylon
insert) is still present on the horizontal bolt but the vertical bolt has broken
level with the inner surface of the tube and the slightly rusted fracture is just
346
visible inside the hole. Figure 10.25 is a closer view of this fracture and reveals
the characteristic form of double bending fatigue initiating at the right-hand
side with secondary initiation at the left.
The interpretation of these fractures is that both bolts had been subject
to cyclic shearing forces as the trailer was being towed. The vertical oscillations of the hitch led to fatigue cracking across the threads of the vertical
bolt, which eventually broke completely and the nut was lost. The head
portion was loose inside the damper tube, so all the towing forces were now
transferred to the horizontal bolt. This could act as a pivot in the vertical
plane but side-to-side movements in the horizontal plane caused it to bear
alternately against the holes in the steel tube and gradually wear them away.
Eventually there was so little section left that a penultimate tug tore through
one side and the final tug as the trailer swayed in the opposite direction pulled
through the other, allowing the socket end of the hitch to pull off the damper
tube. Thus the trailer detached and the exposed end of the damper tube
(Figure 10.26) revealed the tearing of the holes on both sides caused by the
horizontal bolt having to take over the work of the fatigued vertical bolt,
both manifestations that the hitch had been persistently overloaded and was
indeed unsuitable for towing this trailer by a truck with a large overhang.
Much has been omitted from the findings of the investigation into the
cause of this accident; for example, checks on the trailer after the accident
revealed that the nose weight on the towbar was nearly twice what it should
347
Figure 10.26 Tearing of hole in damper tube as horizontal bolt was pulled out
(close-up of lower center part of the damper tube shown in Figure 10.24).
have been and no proper record had been made of the bolt replaced only 3
weeks earlier, presumably the vertical one, and no attention whatsoever was
given to discover why it had failed so soon after the new hitch had been fitted.
The trailer was an accident just waiting to happen. There was no suggestion
there was anything wrong with the hitch. It was unsuitable for the particular
type of trailer and the way it was being used.
11
Fraudulent
Insurance Claims
349
350
the premises, the redecoration had been completed, new carpet laid and a
new, larger radiator fitted. The homeowner produced the original radiator,
which had a hole near the bottom on the side facing the wall. The hole was
part way up the rear panel of the radiator, well above the bottom section
where sediments collect and pitting corrosion eventually leading to a perforation tends to occur. Suspecting that this was not a genuine failure, the loss
adjuster sought a metallurgists opinion as to the origin of the hole. A piece
was cut out so that the inside wall of the radiator could be examined. The
wall exhibited a degree of internal corrosion consistent with several years
service, but there was no significant pitting. The damning evidence, however,
was that the way the metal had deformed revealed the hole had been made
using a 3-mm-diameter punch with a sharp point, almost certainly an engineers center punch. In addition, although the punch had penetrated the wall
at an angle it would not have been possible to drive it in while the radiator
still hung on the wall. Not surprisingly the claim was rejected and the insurance policy canceled.
351
sometimes running into tens of thousands. Often it is found that the keys
surrendered fail to exhibit a degree of wear commensurate with the use
claimed. In many cases the wear is so slight that the keys can only have been
the spare set that had never, or hardly ever, been used. In such cases the
insurers substantially reduce the settlement of the claim on the grounds that
the insured failed to secure the vehicle.
All the necessary observations can usually be made by visual examination, supplemented by use of moderate magnification of the critical areas. It
is sometimes helpful to compare the surrendered keys with a set for the same
make and age of vehicle with a similar mileage by visiting a dealer for the
make of car. The critical areas for assessing the degree of use are:
1. Finger grime and dust in the crevices around buttons and icons,
particularly in the small lettering molded into the plastic case. General dulling wear on the areas rubbed by the fingers. Light scratches
on the exposed rounded corners. Scoring of the key-ring slot and
rounding of its sharp edges caused by sliding along the metal key
ring.
2. Effects similar to those listed above appear on the plastic finger grip
of the key. It is useful to check whether it is a manufacturers original
or a copy key, which will bear a different name and logo integrally
molded into the plastic. The metal blade is particularly revealing. If
it has milled external edges, the peaks soon wear and become rounded
by pushing against the spring loaded plungers inside the lock barrel.
If the key has a profile machined into the sides, similar rounding on
the internal corners soon develops. Lock barrels are greased internally.
When a vehicle is new this is a clear, golden color but with use it
gradually becomes darkened by fine wear debris that collects in the
crevices along the key blade. After very many insertions into the
ignition lock the electroplating on the outside of a key begins to flake
or wear away, particularly near the entry nose.
3. Many sets of keys are mounted on a split steel ring attached to a
leather fob. When new the leather has square cut edges that do not
always line up, so the mismatch may soon begin to curl and become
rounded. The stitching is tight when new but, after a period of use,
strands begin to break and the stitching around the outside loosens.
A leather fob that has been carried about in a pocket or purse for
a lengthy period loses its original firmness and the edges in particular develop a rounded and worn appearance. Transfer lettering
may also have worn off, leaving only faint indentations on the
leather.
352
353
Figure 11.2 Hole in the end plate of a filter cartridge showing a piece of wood
trapped between outside folds. (A) View of base where hole is exactly in line with
that on outer case (Figure 11.1). (B) Side view of piece of wood showing it is the
edge of a planed board, not a piece of tree pruning.
of the paint or abrasion. After the bottom of the canister had been sawn away,
it could be seen that the flap of metal pushed inside had an indentation at
the center, consistent with the canister having been pierced by an 8-mmdiameter conical steel punch with a nose 1 mm in diameter. The same tool
had continued through the space between the canister and the filter cartridge
and punched a circular hole 8 mm in diameter in the end plate, as illustrated
in Figure 11.2A. A sliver of wood was trapped between the folds of the paper
filter but no shreds of this could be found on the sharp edges of either hole.
Instead of following the line of the two holes, the sliver of wood had entered
through the hole in the canister but then changed direction and passed
354
through the space between the inside of the canister and the top of the
cartridge, to finish up trapped between the paper folds on the outside, as can
be seen in this photograph. When this sliver was examined it was found to
be a strip some 5 cm long broken off the edge of a machined tongue-andgrooved floorboard. It was of roughly square cross section and could be
passed through the hole in the canister without touching the edges.
The substantial claim was initially rejected by the insurers as fraudulent
and they threatened prosecution, but were impressed by their insureds genuine distress upon learning how the hole had been made and that he had
only guessed that it was caused by a piece of tree pruning. (Laboratory tests
had shown the thickness of the steel was such that all attempts to drive a
wooden dowel into the end of the filter case resulted in the wood splintering
and only causing a slight dent in the steel.) He said the tractor had been left
in the orchard overnight and he had only been running it for about 20 min
when the clutch failed. He had gotten off the tractor and had seen the hole
in the filter and a train of oil leading back to where he had started that
morning. There were no metal punches or anything similar that could have
formed the hole in the tractor toolbox. The insurance inspector made inquiries of the local police who said several instances of malicious damage to
machines and tractors left out overnight had been reported in that vicinity
but they had been unable to find the culprit(s). Accordingly, the insurance
company met the claim in full.
355
356
Figure 11.3 Fatigue fracture of connecting rod big end, initiating at the oil
passageway hole.
combustion zones would be the first to suffer and, again, there would be
blueing and galling wear over the sides of the cylinders swept by the piston.
None of the signs usually associated with oil starvation could be found.
Turning to the front of the crankcase cover, Figure 11.5, the hole and the
nearby indentation look more like they have been made with the end of a
screwdriver blade than with a piece of steel reinforcement wire sticking out
of concrete. There are several wavy lines scratching the paint in different
directions behind the hole. How could these random scratches have been
produced if the car was moving forward over the piece of concrete at 40 mph?
They would obviously have been straight and parallel or, at most, exhibit
only slight differences in direction.
357
Figure 11.4 Side view of broken big end. Notice the right-hand bolt (arrow) is
bent, the other straight.
The owners handbook for the vehicle gave the engine oil capacity as 11
pints (5.75 L). The leakage rate through the hole was tested by pouring 2 L
of hot engine oil into the sump and measuring the time for 1 L to run through
the hole. This was found to be 22 sec. Hence assuming the sump contained
358
5 L at the time of the incident, the engine could have run for over 110 sec
before experiencing significant distress. During that period the car running
at 40 mph would have covered just over 1 mile, so the engine could hardly
have been starved of oil if the sump was perforated only about 1/4 mile before
the car was brought to rest, as the plaintiff stated.
The claim was rejected. Although the insurance company could have
taken legal action it canceled the policy and advised the insured that he would
need to obtain coverage in the future from a different underwriter.
359
Figure 11.6 Road plan showing alleged positions of three cars in collision.
car somersault at least once, but saw it had landed on its wheels so he ran
over and released the driver from his seat belt. The driver was dazed, but able
to walk. The driver of car B said he realized another vehicle was approaching
because he could see its headlights, but he could do nothing to warn the
oncoming car as he was sideways on to it. After the collision he telephoned
for his daughter to come out in her car and she drove all three to their
respective homes. Her husband was in the motor repair business and he went
out with a recovery vehicle and took the three damaged cars to his premises,
where they were later classified as beyond economic repair by all three
insurance companies. Since no one was injured the police were not informed.
A staff engineer for the insurers of car C inspected the damage 4 days
after the accident. His suspicions were aroused so he visited the scene but
could find no marks on the road or the soft verges, nor could he find any
debris such as broken glass or plastic. However, a claims inspector for the
insurers of car A visited the scene 3 weeks later and found several pieces of
headlight glass and the entire rear bumper and light cluster from car B,
conspicuously present on the grass verge.
360
Figure 11.7 Positions of car A and car B at the moment of collision, reconstructed from bodywork damage.
An independent engineer and a crash reconstruction expert later examined all three vehicles together and compared the damage to the circumstances described in the drivers statements. Many discrepancies were found.
The most significant were:
1. With car B set to make a left turn into a narrow road the steering
would have been turned to the left, but the rear damage caused by
car A was concentrated toward the right-hand side and indicates the
relative positions at the moment of contact must have been as illustrated in Figure 11.7. Moreover, the trunk top of car B appears to
have been open, as its edges had left clear impressions on the front
hood of car A. The most severe compression of the brackets securing
the rear bumper to car B was at the right-hand side and the left side
bracket appeared to have been broken by pulling away the bumper
with an upward twist.
2. Car C had registered two separate frontal impacts against the drivers
door of car B about 25 cm apart. This evidence is clear from the
photograph in Figure 11.8 (top) showing the front door (drivers side)
of car B taken in the recovery breakers yard. The front crash-resisting
bumper structure of car C had dented the sill below the door in two
places. (It was this deformation of the floor pan and buckling of the
side members that necessitated car B being written off.) In the first
impact the rim of the right side headlight of car C had lightly dented
the middle of the door panel. This shows toward the right side of the
photograph where at the 1 oclock position there is paint damage
showing that a single break in car Cs headlight glass had cut into the
door panel (Figure 11.8, bottom). The second impact a little nearer
361
Figure 11.8 (Top) Damage to the sill below the right front door of car B revealing two separate impacts by the bumper of car C, the second below and slightly
to the rear of the first. (Bottom) Indentations in door caused by broken glass in
the headlight of car C. In the first there is only a slight mark from broken glass
near the 1 oclock position of the headlight but in the second the glass is broken
around the entire circumference and the indent in the door panel is deeper.
the rear of car B had made a deeper indent in the sill slightly below
the first, and the door panel paintwork is now marked by broken glass
in a complete circle the same diameter and the same height above
ground as the headlight rim, revealing that all the headlight glass must
have been broken when this second impact occurred. A specialist in
crash reconstruction estimated the closing speed for the first impact
would have been approximately 12 mph and the second 16 mph.
3. There was hardly any damage along the left side of car C, not even
at the front as would be inevitable if it had struck car B as shown
in Figure 11.6. Its right wing was buckled by compression from the
broken headlight back almost as far as the drivers door and the
glass windshield flattened and completely shattered. The roof was
362
Figure 11.9 Roof and windshield damage to car C. Notice the left wing exhibits
very little damage and the headlight glass is unbroken whereas the right wing is
compressed and the headlight glass broken.
Figure 11.10 Relative positions of cars B and C at first impact (full outline)
and second (dotted outline) as reconstructed from damage.
363
road lit up by its headlights, yet the filament of the headlight in the
offside of car C exhibited a brittle fracture, showing it was not incandescent when it broke. (See Chapter 10, Section 10.4.)
The physical evidence thus demonstrates overwhelmingly that this accident did not happen as claimed. It had clearly been staged somewhere other
than the road junction in order to justify substantial insurance claims. The
drivers had colluded in drawing up their statements. The crushed windshield
and soft top of car C were more consistent with having something like a mini
rubbish dumpster dropped onto it than rather than having rolled over on
the road, and it is difficult to believe how the driver could have escaped very
serious head injury if he had been strapped inside the vehicle at the time this
compression of the soft roof had occurred.
At the trial it transpired that there had in fact been a close business
relationship between the drivers of cars B and C, far from their previous
claim of being casual acquaintances at a golf club. Furthermore, the writeoff values they were claiming for their respective vehicles were substantially
more than the purchase prices. However, no connection could be established
between them and the driver of car A, although it is difficult to believe that
he was not part of the conspiracy, inasmuch as he claimed to have seen car
C turn over and land on its wheels yet when he ran to open the door the
driver climbed out a little dazed but otherwise uninjured.
12
Criminal Cases
366
Figure 12.1 Edge of a counterfeit coin above a legally minted coin, showing
irregular milling marks and poorly formed lettering.
examines such details in their loose change? The face patterns are excellent
although many do not match known patterns and dates. Metals are heavy
compared with paper so the more manufacturing processes the forger has to
use to make good copies, the less the profit. The tendency has thus been
to cast the coin by an investment casting technique, using the lost wax
process in which many wax patterns copying the genuine coin are set in a
refractory mold and cast at the same time from the correct alloy. The tiny
area on each coin where the pattern was joined with others onto the tree
inside the mold is easily removed by polishing to match the rest of the rim.
However, some very good counterfeit silver (actually cupro-nickel
alloy) coins were produced by casting when decimal coins were first introduced in the U.K. Coins were physically larger than they are now and the
value of currency and the cost of living were such that there existed a stronger
motive for counterfeiting the 50p coin; the pound at that time was still a
paper note. The first 10p coin weighed 11 g and the 50p coin weighed just
over 13 g. It did not require any great proficiency in mathematics to work
out that, with some means of melting and casting, 130 worth of 10p coins
could be converted to 550 worth of 50p pieces. Molds containing upward
of 30 or so impressions of 50p pieces were mounted in a small centrifugal
casting machine, the 10p coins melted and then forced into the investment.
It was a job that could be done in a few minutes when the boss wasnt around.
There was no record of any purchase of the cupro-nickel alloy. When solidified the individual castings were broken off the sprue and the tiny area where
they had been connected was polished. Such castings of course are not quite
Criminal Cases
367
Figure 12.2 (A) Etched microsection of legally minted coin made from wrought
copper nickel alloy showing recrystallized grains with directional coring from
original cast billet from which the coinage strip was rolled (magnification: 75).
(B) Etched microsection of counterfeit coin made by casting, revealed by large
irregular grains having nickel rich cores merging into copper rich matrix due to
nonequilibrium cooling from the liquid state (magnification: 75).
so sharply defined as a coin made in a coining press from solid metal, but
they were practically impossible to detect visually in normal transactions.
There were, however, physical ways of detecting the counterfeits. Legally
minted coins start off as cast slabs that are rolled down to thin strip and the
coins then blanked out and stamped in power presses with machined dies.
This is sometimes followed by milling the edges. The counterfeit coins were
as-cast so they had a different microstructure, as illustrated in Figure 12.2A
and Figure 12.2B, which are etched sections of the wrought and the cast
cupro-nickel alloy, respectively. The cast structure (Figure 12.2B) exhibits
large grains that are cored (that is a compositional segregation due to the
first solid to appear being higher in nickel than that formed near the end of
the solidification process), which is especially pronounced in copper-nickel
alloys. Figure 12.2A shows a longitudinal section of a minted coin that has
368
been rolled from a 125-mm billet to a 3-mm-thick strip. This exhibits smaller
equiaxed grains that are drawn out in the direction of working and the coring
segregation is oriented in the direction of rolling.
A microstructural examination requires a section to be cut from the
suspect coin and to be polished and etched. Once it had been established
that the counterfeits were being made by casting, then physical tests that do
not require such preparation could be devised to quickly sort them. However,
as the value of the currency fell and the size of the coins was later reduced,
the risk and surrepticious effort were unjustified, so the quantity of forged
coins in circulation fell.
Criminal Cases
369
remelt the ingots and pour the molten metal through a sieve at the top of a
tower and, as it falls, the streams break up into droplets and solidify. Although
considered individually, quantities of impurities were close in the two groups
of pellets analyzed; when certain critical ratios were compared there were
significant differences. For example, while the lead alloy is held in the molten
state, copper, being completely insoluble and less dense than lead, tends to
rise to the surface and will be higher in the first ingots to be cast than the
last, whereas bismuth, being soluble in the liquid lead and having no tendency
to segregate, will tend to oxidize near the surface and will thus be higher in
the last ingots to be cast than the first. Individual ingots from the same cast
will thus vary slightly, particularly when their analyses are reported in parts
per million. The average copper/bismuth ratio in the 10 pellets was slightly
more than twice that in the pellets recovered from the wound. This suggests
that even though these two batches of pellets might have originated from the
same cast in the refining furnace they had been poured at different times
during the shot making process and could well have been made by different
shotmakers who had purchased their ingots from the same refiner.
A single kettle of liquid alloy at the shotmakers would hold typically 30
to 50 t of metal and would be continually replenished with additional ingots
as liquid was drawn off. A total of 50 t of alloy would be enough to make
500 million pellets if each pellet weighed 0.1 g. Under cross-examination the
analyst agreed that his data showed that the two samples of pellets most
probably came from the same source but not necessarily from the same
cartridge.
The firearms expert explained that although there would be no difficulty
in uncrimping the end of a cartridge, removing the wadding, taking out 10
pellets and then recrimping, he did not believe the chemical analysis did, or
could, prove the 10 pellets were taken from the same cartridge as the other
20. He pointed out that the statistical mean weight of the two batches differed
slightly despite exhibiting almost identical standard deviations and that there
appeared to be two distinct groups even within the batch of 20.
The retired manager from the shot manufacturing firm explained how
the shot is made by melting 50 t or so of ingots in a kettle (so called because
the metal poured is drawn from the bottom), and is poured through a screen
at the top of a tower. The streams of liquid metal break up into droplets and
solidify as they fall down and finish up in a water bath that quenches them.
They are then polished and rolled over a series of sloping plates that reject
the pellets that are not truly spherical, which are returned to the kettle to be
remelted. Ingots of different alloys are segregated in their stockyard, but there
is no certainty that all the ingots charged to the kettle on the same day will
have come from the same supplier or have been delivered at the same time.
Finished pellets are sent as small orders to various cartridge makers and they
370
Criminal Cases
371
those stolen and is much larger than the guillotined pieces that it was to have
been cut into for the electroplating process. It is included here to show the
characteristic nodules on the surface as well as along the edge at the righthand side. Freshly produced cathode is a bright salmon pink color when it
leaves the refinery but oxidizes to a dull brown after a few weeks. For use as
anode in barrel electroplating the cathode has to be reduced to small pieces
about 25 mm2, so the electroplating firm needed to send its metal to another
firm that had a powerful guillotine. The electroplating firm arranged to have
this done on a Monday but, because the consignment arrived late on Friday
afternoon, the truck was driven inside the plating firms yard and the gates
locked. It was not visible from the outside. On Monday morning when the
gates were opened up the truck was gone, and with it 9 t of copper worth
more than 9000.
The nonferrous metal trade has a system of notifying all dealers and
metal merchants of a theft and the nature of the material stolen. Some 4 to
5 days after the theft a small van drove to a scrap dealer 250 miles away from
where the truck was taken. The driver offered to sell a quarter ton of scrap
copper, saying he had more if the price was good enough. The scrap merchant
realized that in the load of mixed copper a few old hot water cylinders,
pipes, wire, and so on were pieces of cathode, so he purchased that load
and offered to buy all that the driver could deliver. He also notified the police.
Within an hour the driver was back with another quarter ton, most of which
was cathode. He was arrested and the police accompanied him back to a
small yard where they found the stolen truck containing just over 8 t of
copper cathode.
The man arrested claimed that he was just a small-scale scrap dealer, who
had bought all the metal in small lots from local housing estates and small
industrial units. It hardly needed a metallurgist to identify it as ex-refinery
cathode, which is totally different from swarf or clippings arising from manufacturing processes, or discarded items from plumbing systems and electrical wiring. The thief was obviously unaware that the only uses for cathode
copper are in electroplating and for remelting and alloying with other metals.
In no way could the metal in his possession have been collected from housing
estates and light industry. It was particularly incriminating that the total
amount was 9 t and was just starting to develop the brown oxide coating. It
could only have emanated from the one refinery still operating in the U.K.,
and its color was consistent with it having been produced some 3 to 4 weeks
before it was recovered. Further inquiries revealed that the man arrested had
been living with his family in a caravan not far from the refinery at the time
of the theft.
Another case of theft that required examination of coarse pieces of swarf
found in the soles of a suspects boots has already been referred to in Chapter
372
2. The suspect was accused of breaking into a factory and stealing metal bars
from the side of a large lathe. Pieces of swarf picked up from the floor around
the machine were compared with pieces dug out from the accuseds boots.
Not only was the composition of these pieces, as shown in x-ray scans
obtained under the scanning electron microscope, practically identical
aluminum brass containing copper, zinc and aluminum but the serrations
produced by the cutting tool matched exactly when the suspect chip was
compared with the swarf from the scene. As a cutting tool wears it develops
ridges and cavities that are unique to that particular tool, an effect similar
to the way the rifling of a gun barrel marks a bullet fired from it. The piece
picked up was large enough, and recent enough, to identify as having been
machined with the same tool as the one found in the boot.
Criminal Cases
373
Figure 12.5 Twisted links of ornamental brass chain used to secure door.
374
Criminal Cases
375
and looked as if they had been executed by the same person, who was by no
means a skilled welder.
The police had established that a considerable number of coin boxes in
different localities had been broken into with these particular devices, mainly
because the holes in the sides matched the claws formed by the cold chisels
and the locks had been pushed in by a circular object of similar diameters
to the twist drills. The engineering evidence proved complicity in their design
and manufacture and led to the conviction of all three suspects found with
the devices in their cars.
376
working. He admitted that he did this from time to time but had no idea
where the barrels came from and said this job was usually done at night
because the furnace was in use by the other firms for car parts during normal
daytime hours.
His association with the beer barrel melting became clear after forensic
examination of metal particles found on the soles of the boots and on the
clothes he was wearing when arrested.
All the suspects clothes were brushed and a number of tiny particles of
metal were collected. In addition, there were numerous particles in the soles
of their boots, such as illustrated in Figure 12.6. Examination under the SEM
revealed that all were rounded particles characteristic of splashes and droplets
such as those emitted by molten metal running into open molds. Every one
of 20 samples taken from boots and clothing of all five men were found to
be aluminum alloy containing low percentages of magnesium and silicon.
This composition spread would include the alloy used for beer barrels, but
not the ones commonly used for automobile castings, which usually contain
greater amounts of silicon. Most significantly, no particle recovered from the
cable strippers clothing or boots was copper or lead but, even if there had
been any from this source, they would have taken the form of mechanical
slivers or clippings, not solidified droplets
The clothing exhibited numerous holes formed by hot particles landing
on them, particularly below the trouser knees. Figure 12.7A shows two holes
joined together in the side pocket of a nylon jacket worn by the cable stripper.
These had been formed by one, or possibly two, splashes of hot metal striking
the fabric and melting through the nylon. Figure 12.6B shows the lower leg
of corduroy trousers, where small hot particles had landed. These particles
Criminal Cases
377
Figure 12.7 (A) A hole melted in the pocket of a dark blue nylon jacket and
(B) a hole burned in the lower leg of corduroy trousers, both caused by proximity
to splashes and hot droplets of metal emitted during the pouring and casting of
molten aluminum alloy into ingot molds.
appear to have been still adhering when the trousers were removed but had
been torn away when the trousers were brushed, judging by the torn fabric.
Although most of the above observations are essentially scientific, the
engineering dimension becomes important in identifying the particles and
clothing damage as typical of that expected when workers are standing close
to foundry operations. Few scientists have ever observed the splashing that
occurs and the distance droplets may carry when a stream of molten metal
poured from a furnace first strikes a launder and then runs into an open
ingot mold, or the fine spray of droplets that are thrown up as it solidifies
when the metal has picked up hydrogen due to being melted under reducing
conditions. In this instance, the damage to the clothing worn by the cable
stripper clearly established that he must have been standing very close to the
stream of molten metal as the ingots were being cast.
13
Intellectual
Property Cases
13.1 Introduction
Intellectual property is an area in which many forensic engineers become
deeply involved. Forensic engineers are employed because courts need to
arbitrate between parties who dispute some aspect of new technology or
engineering. Indeed, many practicing engineers also become involved
because they will need to protect new ideas or devices that they conceive or
make either for themselves or for their employers. Disputes commonly
involve inventions or designs where priority must be established, especially
in technologies where development is rapid.
The subject essentially deals with the creation of new devices and machines,
or ideas that take some practical form. An essential attribute is thus novelty;
the device must be unknown before its creation. A second attribute is that the
device has a bodily form and can be applied in some useful way. The new
device receives legal protection in the form of a document that establishes the
owner, the period for which the document is valid, the nature of the right
protected and the exact definition of the device. In the most important example, the document is a patent, but it can also be a registered design (in the
U.K.) or a design patent (in the U.S., for example). Other forms of intellectual
property include written words (copyright) and marks of approval placed on
devices (trademarks and tradenames). Confidential information, such as a
secret formula, can also be protected. The rights established are like property
rights, such as ownership to land established by a title deed. However, the
property is virtual rather than real, and has a strictly limited life, 20 years for
a patent and varying periods for other rights. Although different rules apply
in different countries, there is reasonable compatibility between the major
379
380
381
382
4 11
5
1 14
13
7
5
9
7
9
12
10
8
Figure 13.1 Early version of a domestic wheelie bin of 250-L capacity. The
diagram at right shows a metal bar under the rim to support the lifting fork.
3
2
4
10
9
7
15
14
24
10
19
23
21 22 24 8
16 13 17
Figure 13.2 Rear view of garbage truck showing the lifting bar with forks (22)
used to raise the bins, from GB 1 383 036 of 1972.
The critical technical step in the design of the wheelie bin is the creation
of a reinforced lip at the front of the bin that enables it to be hoisted automatically by a hydraulic lift onto the truck (Figure 13.2 and Figure 13.3). The
critical patent is Schneider, dating from 1976 (GB 1 588 932).
13.2.1 The Lifting Lip
Part of the design problem with using a thermoplastic material is ensuring
that the bin is rigid enough to withstand the critical lifting phase. If the
383
1
3
2
4 29 5 25 6
20
19
9
8
17
16
14
12
11
21
23
15
13
Figure 13.3 Section of the truck rear to show the lifting operation.
A container for refuse, comprising a body of substantially rectangular crosssection having a cover therefor, a depending flange which extends along one side
wall of the body in the region of an upper edge of said wall and is stiffened by
substantially vertically arranged struts connected to the body, and a rib which
extends substantially parallel with the body between the body and the depending
flange; and wherein a reinforcing member is provided for said flange, said
member extending along and projecting laterally from said flange and being
adapted and arranged to serve as an abutment for said container during emptying thereof.
384
385
2
14
3
8
9
10 7
13
12
5
1
6
5
4
12
5
Figure 13.4 Section and plans of the Schneider lifting rim with the forks shown
by the dotted lines in the upper diagram.
thus commonplace in many plastic products, and often can be used for a
secondary design function.
13.2.2 Plastic Bins
What is the material of choice for the wheelie bin? When first developed
in Germany, wheelie bins were injection molded in HDPE (high density
polyethylene), a low-cost but very tough thermoplastic. This makes for a
light product capable of withstanding much mechanical abuse, an essential
feature of the product specification. The product is manufactured in a
single operation, with hot polymer injected via a single gate in the center
of the base. The process requires considerable capital investment both in
the large molding machines and the precision tools needed to create the
shape. The molding process itself exerts influence on the geometric design
of the product. One such constraint, the re-entrant angle problem, is
imposed by the need to withdraw the core of a molding tool at the end of
386
387
Direction of withdrawal
of metal tool cores
during injection
molding.
3
2
2
14
8
9
10 7
13
12
Side core Central core
5
1
34
60
25
34 26
21
22
23
21
22
27
24
23
10
26
29
24
10
20
20
A
Rim (3)
Depending
flange (4)
Rib (7)
Reinforcing
member (12)
Side
wall (1)
Strut (8)
Strut (5)
B
Figure 13.5 (A) Manufacture of the Schneider bin showing the withdrawal
direction in the injection molding tool, assuming the gate lies at the center of
the base. (B) Key parts of the Schneider lifting rim labeled according to claim 1
of the patent.
388
Comb bar
Scale
0
10
20
30
40
50 mm
Figure 13.6 Section of the Taylor 1100-L bin with a hollow steel bar as the
lifting lip.
steel 1100-L wheelie bin. Manufacture of the basic box shape was not
difficult and could be achieved by a small number of welding steps on
preformed sheet steel. But the critical lifting flange needed redesign: multiple ribs and flanges are critical in improving the stiffness of low modulus
thermoplastics, but would be excessively expensive on a steel lifting lip.
In any case, the much greater stiffness of steel (210 MPa compared with
about 2 MPa for HDPE) implied that a much simpler design could be
introduced. The Taylor company came up with the idea of using a tube
welded to the chassis (Figure 13.6). The concept exploits the monocoque
idea, where the load is supported in the shell of a hollow member, a
principle widely used in general engineering. The flange, or comb bar as
it is called, could be made using the same dimensions needed for the
plastic version so that the forks could engage easily, and the mechanical
properties of steel would easily accommodate the lifting loads. An intrinsically strong and tough product, it would also be more resistant to crack
growth.
Naturally, the product when made and sold into the market attracted the
attention of the patentee of the plastic equivalent, and a dispute arose, which
eventually came to trial in the Patents County Court in London in 1996. The
389
large steel 1100-L bin made serious inroads into a market dominated by a
single manufacturer, and the competition harmed the sales of the plastic
equivalent. The problem faced by the patentee was simple: was claim 1
infringed by the steel product? As usual, the legal problem starts with the
claim and its technical interpretation by a skilled reader.
The difficulty the patentee faced was considerable because of the rather
narrow way in which the claims had been drafted, with several engineering
terms of inevitably narrow definition. This is a perpetual problem for patentees (or the patent agents who usually draft such documents). Draft your
claims too widely and you run the risk of catching not just infringers but
also the prior art (in which case your patent is obvious and hence invalid).
Draft the claim too narrowly and the claim fails to give you the protection
you desire for a complete monopoly of your particular product.
13.2.5 The Trial (Schneider v. Taylor)
The trial came before the Patents County Court in London in July 1996. The
trial lasted 3 days. Virtually all of the oral evidence was presented by three
expert witnesses: two were experienced in design and use of refuse equipment, and the third (the author of this chapter) was experienced in design
of polymer products. The plaintiffs case rested on a wide interpretation of
claim 1 to catch a lifting rim of quite different structure to that claimed by
the preferred embodiment. As usual, the defense attacked the validity of the
Schneider patent, using examples from the prior art to show how the device
developed in an apparently logical way from antecedents.
13.2.6 Meaning of Terms
The thrust of the counterattack by the defendants was not, however, very
convincing, mainly because the immediate prior art did not include the
specific reinforcement elements claimed by Schneider in claim 1. In other
words, the inventive concept was original and novel, and so could not be
attacked for being obvious to a skilled reader.
The main defense then shifted to construction of claim 1, and the
infringement issue. The principal integers were the focus of attention, and
the specific problem for the claimants was how a bent steel tube without
conventional ribs could be construed to include ribs. The exact meaning of
several terms in claim 1 were thus crucial to the infringement issue. Those
terms included
depending flange
strut
rib
390
Here are some technical definitions from the 1964 edition of Chambers
Technical Dictionary, a popular technical dictionary in the U.K.
flange = a projecting rim, as the rim of a wheel which runs on
rails, the top and bottom members of an I-beam
strut = any light structural member or long column which sustains
an axial compressive load
A general definition of rib was not satisfactory, but the Oxford English Dictionary states
rib = a raised band, a prominence running in a line
The components are shown in the preferred embodiment of Figure 13.5B.
13.2.7 Alternative Interpretation
In order to bring the Taylor lifting rim into the claim, some flexibility in the
meaning of the key terms was needed by the plaintiffs in the case. One of
the experts appearing on behalf of Schneider labeled the components of the
Taylor bin as shown in Figure 13.7. There are a number of problems with
this interpretation. Let us look at the common identities first: the sidewall
and struts are clearly not controversial, and cannot be confused. The problems arise with the identification of the rib, flange and reinforcing member.
Thus the rib is identified with part of the outer flange, and the reinforcing
member is divided into two separate parts, and flank each side of the rib.
In this case of identity of parts it is necessary to refer to the description
of the invention in the patent-in-suit. The relation of the parts is described
generally first at line 87 on page 1 of the patent:
Conveniently, the impact bar, the reinforcing member
and the flange together define an assembly of U-shape in
cross-section, opening towards the base of the container
The structure is described in more detail on page 2 of the Schneider patent
from line 45 onward. From line 65 on, the patentee describes the function
of the various components thus:
Where the refuse container is provided with a device for
suspension in a so-called comb bar, the spacing between
the individual struts 5 is approximately uniform as illustrated in Figure 13.4. The intervening space is however
divided up by smaller struts 8. The lower edges 9 of these
391
Second part of
reinforcing member
Upper region
of side wall
Weld-machined
joint
Section bent to
shape and welded
First part of
reinforcing
member
Weld
Rib
Weld
Flange
Webs and end
caps welded
Side wall of the
container
Vertically arranged
struts of generally
triangular form
Figure 13.7 One interpretation of the parts of the Taylor lifting rim.
struts are flush with the lower edge 10 of the rib 7. It has
been found that the flange 4 might be deformed between
two struts 5 and 8. As a result of such deformations the
area enclosed by the body part, the struts and the flange
4 can be reduced, and this will continue until the teeth
of the comb bar will no longer find sufficient room for
penetration with the result that the refuse container will
no longer be capable of being hung in the garbage truck
and will therefore become useless.
What the patentee is describing here is the problem of inadequate reinforcement of the lifting rim. The pocket in which the forks or teeth of the
comb bar sit is defined by the reinforcing member 12, depending flange 4,
the lower edges of the rib 7 and small struts 8 and the sidewall of the container
1 (Figure 13.5A). The situation of Figure 13.7 shows a rib that
1. Does not have the properties of a rib, with a free end
2. Does not form a surface against which the teeth of the comb bar abut
during lifting
392
It also seems clear that the structure of the lifting rim the patentee claims
is not just the structure shown in the preferred embodiment in Figure 1 of
the patent, but a whole range of structures, variants of the same basic principle. It will be noticed that all have structures that are capable of being
injection molded, with a withdrawal line oriented along the main axis of the
container (Figure 13.5A).
13.2.8 Standard Dimensions
Another important issue that arose during the course of the trial was the
question of what dimensions should be chosen for the lifting lip. After all,
the patent claims or specification nowhere state dimensions of the lifting rim,
and the rim could be of any reasonable size, provided it was capable of
performing the desired function of bearing the imposed load of the rubbish
within it when lifted onto the truck. This is a common situation in patent
claims, because the patentee wants the widest possible definition of the invention claimed so as to deter direct copying or other kinds of imitation.
However, it seems reasonably clear that the dimensions of the lifting
pocket on the Taylor bin are designed to accept the standard forks or teeth
of the lifting comb on the truck. After all, the Taylor bin only appeared after
the Schneider bin, and Taylor could not sell its design on the market if it did
not fit comb bars already fitted on rubbish trucks. Standardization of components is of course commonplace with engineered products and a problem
that recurs when a new idea suddenly appears and several manufacturers
enter the field with sometimes only slightly different designs. The consumer
is tied to his supplier because only the suppliers spare parts are made to the
exact dimensions of the device originally supplied to the consumer. This was
the problem faced by car exhaust makers that wished to supply replacements
for existing cars, and it is known as the must fit criterion. The current
British law (Copyright Design and Patents Act, 1988) excludes protection of
must fit dimensions for new designs (as protected by the new design right,
the replacement for copyright in design drawings). Thus it seemed clear that
Taylor had copied the existing dimensions of the Schneider pocket, but
without necessarily infringing any intellectual property (IP) rights held by
Schneider.
13.2.9 Judgment
At trial, all the witness evidence consisted of testimony from the three experts
involved. Despite attacks on credibility of some of the experts, and strained
attempts to define a rib in rather unusual ways, the patent was found to be
valid but not infringed. The patent survived attacks on its validity, but the
attack from the prior art failed, and a clear inventive step was identified.
393
From the expert evidence provided by the polymer expert, the judge decided
the patent had been written for a designer skilled in plastics injection molding, and it was difficult to see how a steel flange could have been envisaged
by the patentees. It was thus difficult to see how a steel rim could fall within
the normal construction of claim 1. Since no serious attempt had been made
at trial to attack the validity of the Schneider patent, it was held to be a
legitimate invention. The final verdict clearly justified the Taylor company in
the defense of its very successful product.
13.2.10 Afterword
It is interesting to note that the Taylor steel bin was itself patented, and used
to pursue alleged infringers of the concept. The company has proven the
commercial and technical quality of its bin, and competed very successfully
in a large market with its plastic equivalent. The smaller plastic bins of the
domestic market remain safe for the original patentees, but much of its trade
in the large bins has been lost to a product that offers extra advantages in
terms of longevity, safety and security to the user. Whether or not steel bins
enter the domestic market remains to be seen, although it seems unlikely in
the immediate future.
There are several factors involved. Small plastic domestic bins are lighter
and more user friendly than steel. On the other hand, they are more susceptible to damage (especially of the lifting lip), and they are also combustible.
Brittle cracks at the corners of ribs are quite common and, if they reach a
critical size, they will propagate catastrophically. This could cause the loaded
bin to fall while being loaded onto the rubbish truck, so damaged or cracked
bins should be replaced as soon as the damage is detected. However, the
manufacturing costs for a steel domestic bin will be higher pro rata than for
larger containers, and corrosion may occur after some time in service.
Although rusting can be inhibited by galvanizing and painting, such bins
receive a great deal of mechanical abuse in service, creating impact craters,
for example, from which rusting can develop.
The case study is a good example of how materials displacement can
legitimately be pursued by designers, especially if the original patents are too
narrowly worded.
13.3 Lawnmowers
Many intellectual property actions are settled before any trial, especially if
the potential liability in costs awarded against the losing party could exceed
the cost of a court trial. The costs of full trials are considerable, especially if
common and well-known products are the subject of patent disputes. There
394
will be much court time devoted to examining the prior art and the validity
of the patent-in-suit, largely because it is a common defense in such actions
to challenge the very basis of the patent being used to catch an infringer. On
the other hand, the resultant judgment can create severe financial problems
to the losing side. The reason for this state of affairs is that a product found
to infringe may have to be removed from the market, or alternatively, a royalty
on all sales (past, present and future) may be imposed by the patent owner.
In fact, the largest damages ever awarded in any court action worldwide were
awarded against Kodak for infringing Polaroids instant photography patents,
a case that had severe effects on Kodaks cashflow for several years. 3
The case study analyzed in this section ended with a High Court trial,
and was not appealed. There were no fundamental points of law that could
be challenged, and the case was so definite in its conclusion that the plaintiff s
patent was declared invalid on no fewer than three different counts. Patents
that do come to trial are subjected to rigorous forensic analysis, and it is only
those patents that are fundamentally inventive that survive. In many consumer product areas, patenting is extraordinarily extensive (especially by
larger manufacturers), but any particular patent faces a severe problem in
that it may be difficult to distinguish it from the many similar patents in the
prior art. The case study presented here involves an area of great commercial
activity: gardens, and specifically, lawnmowers.
Although the materials of construction are rarely specified in new inventions (so as to give much wider claims), a significant aspect of the case study
involved the use of materials, and the specific advantages of thermoplastics
for lightweight consumer products. Choosing one class or indeed a specific
member of that class of materials necessarily entails consideration of the
manufacturing method used to make the shaped product (as seen in the
previous case study). There are substantial differences between the ways in
which different materials are shaped, and such methods can limit the shape
used in the actual embodiments of a new idea.
13.3.1 History of Development
The idea of using a mowing machine to trim lawns dates back to the Victorian
period, but the irresistible rise of the internal combustion engine produced
numerous examples of powered mowers in the early part of the 20th century.
The cylinder mower was (and to some extent, still is) the principal way in
which rotary power was applied to cutting grass. The blades in such mowers
are set lengthways on a cylinder spaceframe, and, as the cylinder rotates at
the level of the lawn, it cuts through the sward. The quality of finish is high
because the blades can cut very finely, and the sward can also be cut very
close to the ground, producing a smooth and level finish. This is the traditional way of cutting grass, and the art of cylinder mower design has reached
395
396
Cowl
Motor
Hood
Fan
Blade
Figure 13.8 The hovermower shown in section and plan (U.K. 3 338 038 of
1964).
the strict property needs, but the hoods could be mass manufactured using
injection molding. This method is ideally suited to long production runs (for
large product numbers for a mass market) because of the high capital investment needed to make the metal tools needed to create the product shape.
Quality control is of course vital to ensure that the final product meets the
property specification, but when well controlled, this manufacturing method
gives exactly identical products for production runs of hundreds of thousands, and even millions. Injection molding is also a method that is less labor
intensive than manual fabrication or pressing of metal shells (for example),
and the process can be automated relatively easily.
Although there are several plastics that can meet the demanding property specification, one in particular has been widely used: ABS (acrylonitrile-butadiene-styrene). It is tough and crack resistant, relatively easy to
mold and available in a wide range of grades and colors. This is the material
of choice for the structure of the Dyson vacuum cleaner (1996).
The hovermower has been manufactured by several companies since
the expiration of the 1968 patent, and it is available to European gardeners
at low cost. For reasons of product liability, the hovermower has not been
widely adopted in the U.S. While various improvements have been made
to the basic idea, there are still several intrinsic problems with the way the
machine works.
397
398
24
23
29
22
29
8
10 14
16
13
6
23
7
2
5
3
1
4
7
6
6
6
11
38
21 18 19
39
25
17
20 12 38 25
16
10
8
14
8
21
25
6 13
12
11
6
3
38
23
2
39 18 19 17 20 38
1
4
25
over a greater area, and that area is linked together so that all parts of the
shell contribute to load bearing.
13.3.6 The Rasenkatz Patent
What does the patent-in-suit specify and, critically, what is the width of the
claims? Claim 1 of the patent (filing date June 16, 1977), which was the basis
of the infringement case against the Flymo products, reads as follows:
What We Claim Is
399
and a wall extending upwardly from the bottom wall to define the outermost
peripheral sidewall of the body; a motor equipped with an output drive shaft
having a free end, said motor being mounted in said body so as to cause said
free end to penetrate said bottom wall for carrying the cutting blade thereon;
and, a cover co-extensive with the uppermost edge of said outermost peripheral
side wall and engageable therewith to define, conjointly with the dish-like body,
a housing enclosing said motor.
The claim can be subdivided as before:
What We Claim Is
400
Cowling
Cutter
Axis
Ridge
Cowling
Flap
Motor
Deck
Deck
Section A-A
B
Ridge
Roller
Section B-A
Hinged
Flap
Cover Plate
Deck
Wheels
Rollers
A
Section A-A
B
Ridge
Twin Rollers
Section B-A
Figure 13.10 Sections of two Flymo rotary mowers, the 350S and the 420S.
401
3
17
21
2
33(38)
22
17
24
12
32
33 31
26
28
30
2
2
15
25
14
7 10 9
1
34 11
III
2
IV
16
12
21
1
17
18
13
4
2
33
15
35
20
IV
20
21
22
16
2
37
19
17
25
26
32
13
12
18
2
33
5
36
10
34
38
1
7
11 8
III
21
22
25
2
37
19
1
5
26
16
27
13
29
20
24
27
32
33
15
17
18
2
38
7
10
34 11 9
Figure 13.11 The Zundapp lawnmower showing construction of the outer case
with base (1) and mating cover (17) (D 2 030 848).
it is obvious in light of prior art, and invalid. This is normally the task of an
independent expert witness, who can advise the court in its approach to
interpretation or construction of the patent. The witness was asked initially
to propose a design sequence for a lawnmower based on general engineering
principles, ending with the design in question. One aim of such a task is to
pinpoint the steps in the sequence that are either commonplace or unusual
(and thus potentially inventive). The argument proceeded as follows; the
summary is taken from the more extensive expert report presented in evidence at the trial.
1. The first step is to consider the design for function, i.e., cutting grass
using a motor-driven blade. The motor and blade must be supported
safely and rigidly, and there must be a partly open space at the base
of the mower for the action of the blade on the grass.
402
Monocoque
mower
Platform
mower
Space frame
mower
Figure 13.12 Conceptual development of mower case, showing three alternative design options (stage 1).
5.
6.
7.
8.
403
so metals are excluded. The polymer options are also favored from
another viewpoint: vibration damping or attenuation. Vibrations
from the motor and blade could cause resonance in the body, thus
making the mower noisy in use. It could also cause fatigue cracking
in susceptible areas of the body. Polymers are favored here because
they tend to absorb or damp vibrations, while metals tend to transmit
them freely (like a bell).
Thermal resistance is important when the mower is in use: the motor
will develop moderate temperatures of up to about 70C. Metals will
easily resist such temperatures, but polymers could be affected. If a
polymer is chosen it should be able to remain rigid above this temperature to give a safe margin.
The loading criteria are critical, and a stiff material is demanded.
Polymers are generally less rigid than metals, so care would be needed
in choosing the wall thickness. The greater the wall thickness, the
greater resistance to bending and torsion. The importance of torque
and bending stiffness are directly related to the safety of the user; if
it were too low, the part of the shell next to the blade could be
distorted and make contact with the blade, causing catastrophic failure. Monocoque construction is favored because the overall product
stiffness is much greater than that of individual shells. In addition,
an external flange could increase the local stiffness of the critical outer
edge of the mower to resist external blows. The options thus become
as shown in Figure 13.13; stage 2. The favored option would be the
monocoque mower because it provides a cover for the motor, shielding it from the user.
The final set of criteria deal with the manufacture of the device: it
must be capable of mass manufacture with few assembly steps. Injection molding was well known before the priority date of the B&D
patent, and tough thermoplastics such as ABS can be molded easily
into very complex shapes. Some complexity is needed because wheels
may be attached, a handle is needed to connect the user to the body
in order to steer the device, the motor needs to be fixed securely to
the body, the cable feeding electricity to the motor needs to be fixed
securely and bosses for screw fixtures will be needed during assembly.
The material is tough (impact resistant) and resistant to the environment, is thermally stable to about 100C and can be made stiff enough
for the function of the mower.
If injection molded, a single, closed monocoque is impossible to make
by molding since the method involves injection of hot polymer into
a metal tool with a shaped cavity. The cavity is produced by two
mating halves that must be capable of separation at the end of the
404
Flanged monocoque
mower
Flanged platform
mower
B
405
Bond line
Bond line
is now common for experts instructed in many U.K. product liability cases
to meet before trial starts. Those points should already be clear to the legal
teams without the need for expert interpretation, especially with a simple
mechanical patent.
13.3.8 Pretrial Litigation
In important civil actions, there are often subsidiary issues which are tried
separately before the main trial. This dispute did give rise to a prior action,
and was concerned with the process of discovery. This is the process
whereby each side provides information to the other that is relevant to the
dispute. In IP trials, such information includes design data (e.g., plans of
various models, and even real products) and internal reports concerning
design philosophy and implementation, etc. The action was important
because it set a precedent for all future actions and widened the basis of
discovery. The question at issue was simple: should a new Black & Decker
lawnmower about to be launched (and mentioned specifically in a witness
statement) be shown to the Flymo legal team?
13.3.9 Discovery
Discovery is the process of finding evidence not already revealed in a case as
it proceeds to trial. Thus a manufacturer will or should keep voluminous
documentation concerning product quality, dates and times of manufacture,
as well as financial records of sales, and components and materials purchased
for manufacture of the product concerned. In product liability cases, records
of customer returns for particular batches or products will also be important
for checking whether or not there had been a problem with the batch from
which the failed product came (especially if a design defect is suspected).
406
There may also be material evidence in the form of actual failed products
which had been kept or returned to the manufacturer by other consumers.
If they do exist, it is important that they be inspected for the nature of the
failure mode and should be compared with the actual failed product for any
parallels. All such evidence will normally be confidential to the manufacturer,
and is said to be privileged. The process of discovery is thus a normal part
of pretrial proceedings, and it is a two-way process, with each side exchanging
discovery documents and any other evidence that is directly relevant to the
case in hand. An expert will be asked not only about the nature of the evidence
which should be disclosed before discovery, but he or she will also be asked
to examine that evidence after discovery has taken place for any clues that
might shed light on the actual failure at the heart of the case.
13.3.10 The New Model
Although many documents or, in this particular case, actual products are
privileged, when they are mentioned in a witness statement they should be
revealed to the other side in a dispute. One of Black & Deckers witnesses
mentioned in his witness statement a new model about to be launched for
the 1991 mowing season, and which apparently fell within the inventive step
claimed in the Rasenkatz patent at the heart of the dispute. Such new models
are normally kept under wraps until the launch so that the competition does
not know what to expect until the model is publicized, and its various
apparent advantages demonstrated to the buying public. Such secrecy is of
course notorious in the car industry, where competition for new buyers is
extremely fierce. Just such secrecy is endemic in many other consumer product markets. Black & Decker refused to release the product for discovery by
Flymo, pleading commercial secrecy, until the dispute was brought before
Judge Hoffmann sitting in the Chancery Division of the U.K. High Court.
The judge decided that, since the new model had been mentioned in a
disclosed statement, it should be made available to the defendants for inspection. A report of the short action6 has become important because of the
significance of the result for unrelated disputes. This is thus an example of
case law not just for patent actions but for any other action involving discovery in the U.K. Inspection duly occurred in a meeting at the patent agents
acting for Flymo, and in fact the very same model was used by the plaintiff s
counsel to open the case in the Patents Court.
13.3.11 Main Trial (Black & Decker v. Flymo I)
The action between the two companies occurred over 6 days in the Patents
Court. The two witnesses of fact, the chief designer at Black & Decker and
the chief designer at Flymo, gave evidence in turn. Most if not all the design
407
408
Bolens
1965 PAT 3192692
Filed Sept. 22, 1960
Husqvarna
+
1974 PAT 1448945
filed Jun. 11, 1974
Flymo
RE 40
1985 Flymo
RE 35
Flymo
1988 RE 42
1983
Figure 13.15 Development of the Flymo wheeled rotary lawnmowers from the
hover mower.
manufacturer of safety razors. The defendant argued that his device was
obvious and only used a sequence of noninventive steps. He won the case
but Gillette swept the market with its product.7 At the time of design of the
models, the Flymo designer had used the Husqvarna as a model, and had
not even seen the Rasenkatz patent at all. Each step that the designer took
in developing the range of models may have been noninventive and thus the
Gillette defense could succeed at trial.
13.3.13 Validity
The next question that arose concerned the validity of the patent. The patent
had been challenged by the expert acting on behalf of Flymo as being obvious,
in that a simple design sequence using well-known principles would arrive
very easily at the patent mower. This was not enough in itself to destroy the
patent, because, for example, the expert admitted that he was not a lawnmower designer. However, the similarity of the structure of the Zundapp to
his preferred design was very clear, and this could support the inference that
the patent was indeed obvious. So the Zundapp patent was clearly worth
examining in some detail for its principles of construction.
The general design principles of the Zundapp mower have been briefly
mentioned above, but there were numerous other mowers of very similar
409
design (such as the Husqvarna shown in Figure 13.15) in the prior art.
However, for practical convenience, arguments at the trial revolved mainly
on the Zundapp patent. It was useful to have the device before the court for
direct examination, but a Zundapp mower itself could not be found. The
defendants therefore asked a prototype modeler to make a model of the casing
of the Zundapp, using the design ideas presented in the Zundapp patent and
the diagrams of the preferred embodiment (Figure 13.11). It showed a base
and cover modeled to approximately the same size as a real mower, and
constructed from sheet plastic bonded together. The model was then available
as a court exhibit so that the principle of construction could be examined in
detail. The point about whether or not the Zundapp cover and base could
be injection molded was discussed. The base of the model showed a reentrant angle in the base or chassis part, a feature that the modeler had
included in the prototype, but which appeared not actually to be present in
the Zundapp diagrams (Figure 13.11). As the witness explained, the reentrant to the base chassis could, however, be molded using sliding cores.
The design sequence discussed above was presented by the expert
instructed by Flymo in court and tested by cross-examination. The expert
for Black & Decker was not called to present his evidence. It was attacked
vigorously by the barrister acting for Black & Decker. However, several other
issues emerged during the course of cross-examination, including the meaning of words in the specification.
13.3.14
Wording of Claim 1
As in the previous case, the meaning given to specific words in a patent claim
are crucial to the size of the claim. If the key terms of the claim are interpreted
in a narrow sense, then the patent will not catch infringers, but if they are
given a wider meaning, then the size of the claim widens and catches infringing products (Figure 13.16). However, a wide claim poses a dilemma for the
patentee, because if too wide, the claim will catch the prior art. If this
happens, the patent must be invalid because it describes an idea that is already
known in the art. Such a problem is known as a squeeze argument, and is
central to the problem of successful patenting.
What terms of claim 1 of the patent-in-suit were important at trial? Few
of the terms of claim 1 of the Rasenkatz patent could be called technical,
apart from
output drive shaft
Terms like sidewall and chassis have a technical connotation, but
dish-shaped body is not a word found in technical or engineering dictionaries!
410
Claim 1
Prior art
Patent
Noninfringing
product
Claim 1
Prior art
Patent
Infringing
product
411
Another point that arose during the trial concerned the strict requirements of claim 1: nowhere in the claim is the need for wheels mentioned.
Although wheels are well known and thus an obvious feature, the claim could
be interpreted in such a way as to include within its terms a hovermower
hood!
13.3.15
Judgment
412
step of the Rasenkatz mower). He identified the inventive step of the patent
as being
making a chassis with upstanding walls which obtained, in part, its strength [sic] from its co-extensive
cover
This was the clam-like shell, or monocoque principle. When the designer
seated a cover on his chassis to the Flymo mowers, he was not being inventive
but simply producing a design variant. The judge could not identify any part
of the design process that could be described as inventive. The Rasenkatz had
played no part in the Flymo design sequence (Figure 13.15). If claim 1 was
construed widely to include the Flymo mowers, then it was invalid for obviousness. He went on to consider the claim, if properly construed, was also
obvious in the light of the prior art. The only features that could be described
as inventive were
1. Making a dish-like chassis with upturned walls defining the outer
periphery
2. Mounting the motor in the body
3. Making the edges of the cover mate with the edges of the chassis
The defendants submitted that such a lawnmower would be an obvious
design variant not requiring invention (Figure 13.14). The judge agreed,
bearing in mind that it was based on hindsight reasoning. So even when
correctly construed, the patent was obvious. The Zundapp lawnmower clearly
disclosed the idea of a chassis with upstanding walls and an enveloping cover,
and this was supported by the model made for the court. The judge then
returned to the Windsurfer tasks. Having already defined the inventive step,
he considered the differences with the Flymo models, which exhibited rear
grass collectors and could not be envisaged by claim 1 of the patent-in-suit.
The exact position of the joint between the chassis and cover was a matter
of design choice (from the evidence of the expert shown in Figure 13.14 and
the Flymo designer), not invention, so Rasenkatz was obvious in the light of
Zundapp.
The court thus found that the patent was obvious on three different
grounds: it included the Zundapp on proper construction, it included the
Flymo mowers (on the plaintiff s construction of claim 1) and, quite separately, it also included the Flymo hovermower, a problem for a patent known
as lack of fair basis. It was clear that wheels are not an essential integer of
claim 1 of the Rasenkatz mower, so it could be read as including a much older
patent: the hovermower. It is interesting to note that such a decision (invalid
and not infringed) was so clear-cut that it was not appealed.8 However, it is
413
not common for patent court decisions to go so far. More commonly, a patent
may be found perfectly valid, but not infringed. Another possibility is that
some of the subsidiary claims are found to be invalid and require amendment
for validity. Subsidiary claims are usually much less powerful than claim 1,
and are normally inserted to catch those intent on copying but not willing to
proceed so far as to infringe the main claim of a patent.
13.3.16
Afterword
Major patent trials are dramatic events because so much is at stake. Since the
patent was rather old by the time the dispute came to trial, there would have
been large royalties to be paid on all sales of the two Flymo models concerned
if the case had gone against Flymo. Since Flymo would have been taking Black
& Deckers ideas for many years, a royalty of 5 to 15% would have been payable
to Black & Decker for the period of manufacture and sale of the two mowers
concerned (15 and 12 years, respectively). Interest would also have been payable
for the period over which the payment was due. However, this did not happen.
The Rasenkatz patent was demonstrably weak, and too close to the prior art.
The way claim 1 was interpreted by the plaintiffs to include the Flymo mowers
meant that it also included the prior art, such as the Zundapp mower.
Both companies continue to design and produce new and innovative
products for the market, a market that appears insatiable. DIY activities, in
which Black & Decker specializes, continue to grow tremendously, although
competition from cheap Far Eastern imports, such as the Workmate, is great.
New products for gardening, such as devices for collecting leaves and other
debris, continue to be developed. Flymo has developed a collecting mower,
for example, where the grass is collected above the hood. Black & Decker
made a similar product in the early 1990s, and the two companies clashed
again over the Flymo patent. The case came to the Patents Court in 1996. To
show how similar the two designs were (in Flymos opinion), various experiments with grass collection were arranged for presentation before the court.
They did not convince the court, however, and the court found that the Black
& Decker model had not infringed the patent-in-suit. An attack on the
validity of the latter also failed, producing a verdict of valid but not
infringed.
There are still many independents on the market that make mowers in
a range of different materials. Decisions of the type made in the Flymo case
are watched very closely by other manufacturers, who are aware that if the
decision had gone against Flymo, they could be next in the frame for infringement proceedings. In the event, the decision freed the market for a wide
range of competition, clearly good news for the consumer. Although thermoplastics have penetrated a substantial slice of the market, other materials
continue to be used.
414
415
June 13, 1983, against a new design of RCD actually designed in New Zealand,
but imported to the U.K. by a large electrical manufacturer (Volex). The
interest in the dispute arises from the way the technical effect was interpreted
by the plaintiff, with expert evidence from leading electrical engineers, while
the technical effect was in fact entirely mechanical in nature. The defense
strategy focused on the mechanics of operation of the design, showing the
court that it was this feature alone that distinguished the design of its product
from the main claim of the patent-in-suit (which worked in a rather different
way). It was also an example of a provisional patent battling it out against
the recent prior art in the form of an existing and enforceable patent. Clearly,
if the prior art was valid and infringed, then the provisional patent for the
new device could not be valid. Much was at stake in the battle between the
two parties to the action.
13.4.3 Patent-in-Suit
The patent, originally developed by Technology Research Corporation
(Clearwater, FL), presents a series of figures that show how a preferred
embodiment of the device works (Figure 13.17). As in all patent cases, the
way the device works and the technical effect claimed has to be worked out
by close reading of the specification taken together with the drawings provided. Claim 1 of the patent-in-suit reads as follows:
What We Claim Is
416
17
15
19
25
27
31
35
45
57
63
37
51
11
21
23
11
13 53
41
17
55
61
37
27
31
35
3745
57
15
63 13
29
33
35
39
59
13
17
31
41
51
13
59
43
39
45
57
47
53
33
51
C
19
15
35
17
27
31
35
48
59
15
13
11
63
51
E
47
61
45
57
59
Figure 13.17 Preferred embodiment of TRC residual current device: (A), (B)
(C) unprimed state with electrical contacts open; (D) fully primed state with
elecrical contacts closed; (E) actuation with latch arm bent to one side by cam 61.
417
Main spring
Lever arm
Moveable
platform
Base
R1
R2
A
As usual, the claim is stated in very general terms with numerous technical terms, and other words of common meaning that could have a technical
connotation. The meaning of the claim is not immediately obvious, although
the specification describes in general terms how a preferred embodiment of
the device works in practice.
The sequence of figures shown in Figure 13.17 confirms that the device
operates mechanically, in a way similar to the action of a very familiar small
household machine: a mousetrap (Figure 13.18). Unlike an RCD mechanism,
the mousetrap is not a precision device, but it can perform its function
reasonably well. It is a device based on the interaction of several components,
which must be primed and baited to achieve its final effect. In its primed
state, several forces interact and are balanced in a state of static equilibrium.
The state is unstable, however, and only needs a small trigger for the device
to react to an intruder (Figure 13.19).
A mousetrap fulfills its function when the victim steps onto a springloaded platform to operate the main spring, which then swings over to hit
the mouse. The main spring is held in place by a rigid lever that, in turn, is
held in place by a hook attached to the platform onto which the mouse steps.
In the case of an RCD, however, the device catches the electrical imp whenever
a loss in voltage is detected. Any leakage of current will cause a drop in
418
voltage, tripping the set of springs that are finely balanced in a metastable
state of mechanical equilibrium. This state is maintained by a friction force
between a bent arm and a bent leaf spring, so that any small movement of
another rigid arm held in a solenoid will trip the device and cause the bent
leaf spring to open the circuit.
13.4.4 Modeling the Technical Effect
It is helpful to redraw the action sequence shown in Figure 13.17 in a simplified diagrammatic way, so that the technical effect is clear and transparent.
In addition, the balance of forces in the primed device should show exactly
how the device is held in a metastable equilibrium. This would be the starting
point, in any case, for quantitative analysis during the detailed design work
which should precede prototyping, followed by tooling up for production
models. The device is shown in schematic form in Figure 13.20, with the
balance of forces in the primed device shown in Figure 13.21. The lower
diagram identifies the principal forces acting against one another when the
contacts are touching and the external circuit is complete. The schematic
may then be compared with the situation as represented by the actual embodiment presented in the TRC patent-in-suit (Figure 13.17). Because various
parts of the framework within which the device sits have been added, there
are several extra forces shown, which are primarily reaction forces at fixing
points. (Their addition, however, does not change the basic balance exhibited
by Figure 13.21.)
A more comprehensive diagram of the TRC patent is shown in
Figure 13.22. The solenoid, which holds an armature within it when current
is flowing and the voltage is normal, is linked mechanically to the two leaf
springs which lock together at a fulcrum. One of the leaf springs also makes
an electrical contact, which is broken when the armature is released, following
Figure 13.20 Schematic of TRC device primed and showing the coil spring
and two leaf springs that power the action (inverted).
419
S1
M
F
S2
S1 = Main spring
S2 = Catch spring
S3 = Helical spring
M = Magnetic retaining force
F = Friction at latch
N = Normal reaction at latch
S3
R4
S4
R5
S2
R1
S1 = Main leaf spring
S2 = Latch arm spring
S4 = Return spring
R1 = Reaction at electrical contacts
R3 = Reaction at latch face
R4 = Reaction at button assembly
R5 = Reaction at main spring rivet
M = Magnetic holding force
F = Friction at latch
S1
R3
Solenoid
Figure 13.22 The TRC device shown with its frame and the full balance of
forces acting at equilibrium in the primed state.
420
also be totally reliable and robust, and thus constructed of wear and corrosion-resistant materials. Two metal leaf springs and a metal coil spring are
the basis of the stored mechanical energy when primed, all working within
the elastic limit of the metal. In the Hookean region of behavior, applied
strain and the stress produced in the metal are completely reversible for
thousands of cycles, so the reliability of the device is assured.
13.4.5 The Protector
The device accused of infringing the Powerbreaker patent in the action
appeared, at first sight, to be very different from the Powerbreaker. It is known
as the Protector. A notional equilibrium diagram of the set-up of the primed
device is shown in Figure 13.23. The balance of forces in the primed state is
shown in Figure 13.24. The sequence of actions exhibited in Figure 13.25 show
how the Protector works. The Protector was developed by PDL, a plastics
molding company based in New Zealand, and was the subject of a European
Patent application (EPO PAT Appl. No. 526 071). The device consists of a set
of rigid levers and coil springs, which looks quite different as a reaction
10
14
51
13 90
53
51
50
55
60
40
15
54
52
61
16
43
101
42
13
50
12
51
11
51
71
4
15
80
5
52
17
70
73
Figure 13.23 Sections of the Protector (from EP 526 071) showing various
working elements of the device.
421
Actuator
assembly
S1
Lever
assembly
R4
R2
2R1
R3
S2
S3
Solenoid
assembly
Figure 13.24 The balance of forces at equilibrium in the primed state of the
Protector.
mechanism compared to the TRC device, but in fact it has a similar end result
in cutting the electrical external circuit when a voltage drop is detected
(Figure 13.25B to D).
Clearly there is a difference in the way each device reacts to an undervoltage condition, a difference that is reflected by the different component
parts in each mechanism. The Protector is based on a set of rigid levers
primed by operating a push button that is independent of the main mechanism. The potential energy of the system is raised by coil springs only, while
the Powerbreaker uses two leaf springs and a coil spring to achieve its effect.
The trigger action is quite different: in the case of the Powerbreaker, the
armature is an integral part of the mechanism, being physically connected
into the mechanism at all times, while the armature in the Protector acts
independently. In the primed state, the solenoid is entirely separate in space
and physically unconnected with the primed levers, making physical contact
only when released by the solenoid (Figure 13.24). It achieves its effect by an
422
10
42
43
8 1
5
4
9
4
5
2
7 3
43
42
43
42
5
9
4
9
3
10
6
10
101
101
5
4
9
7
7 3
13
6
13
12
12
11
1
1
2
2
3
Figure 13.25 The sequence of events during the action of the Protector in
guarding against leakage of electricity: (A) primed, (BD) triggered, (EG) reset,
(H) unprimed.
423
424
425
13.5 Endword
Despite the publicity given to biotechnology, computing and electronics,
mechanics is still one of the most prolific areas of invention worldwide. The
refuse bin is a good example of how simple ideas can be patented, and those
ideas copied, provided they fall outside a reasonable interpretation of the
valid patent claims. The market for such bins had expanded in Europe at
such a rate as to create a demand that could be satisfied by several manufacturers. By changing the material of construction, the design changed and so
avoided infringement of the original patent. The steel bin was in turn patented, with the new patent avoiding reference to the critical nature of the
lifting lip and instead focusing on the bin lid. The same relatively small
manufacturer has more recently patented a range of related inventions,
including a compost bin and a device for weighing filled containers.
Lawnmower disputes will no doubt continue between leading and very
large global manufacturers fighting for market share. Much greater care is
needed in drafting patents, however, where the smallest advances in design
are registered. The patentees dilemma is at its critical edge due to the close
proximity of prior art. The same is true of many other consumer products
made by large companies and traded globally. A further danger lurks ever
closer: the problem of imports from the tiger economies of the East, especially
426
References
1. Battle, C., The Patent Guide, Allworth Books, New York, 1997.
2. Cornish, W.R., Intellectual Property: Patents, Copyright, Trade Marks and
Allied Rights, 4th ed., Sweet & Maxwell, London, 1999; Cornish, W.R., Ed.,
Cases and Materials on Intellectual Property, 3rd ed., Sweet & Maxwell, London, 1999.
427
3. Polaroid Corp. v. Eastman Kodak Co., Fleet Street Reports, 25, 1977; Reports of
Patent Cases, 379, 1977.
4. Lucas v. Gaedor, Reports of Patent Cases, 279, 1978.
5. Catnic Components v. Hill & Smith, Reports of Patent Cases, 183, 1982.
6. Black & Decker Inc. v. Flymo Ltd., All England Law Reports, 3, 158, 1991;
Weekly Law Reports, 753, 1991; Fleet Street Reports, 93, 1991.
7. Gillette Safety Razor Co. v. Anglo American Trading Co. Ltd., The Illustrated
Official Journal (Patents) Reports of Patent Cases, Vol. 30, 465, 1913.
8. Black & Decker v. Flymo, Intellectual Property Decisions, 14(5), 2, 1991.
Index
C
Carbon, in steel, 145146
Carbon black, 236
Car keys, 350351
Case hardening, 145
Castings
aluminum, 161
shrinkage of, 3233
Cast iron
flake graphite, 148149
gray, 9596
malleable, 97
microsection of, 96
shrinkage of, 131
threaded iron pipe coupling failure,
151152
varieties of, 95
Catheter failure and fracture
breast tissue expander, 298303
TPE, 297298
Causal analysis, 37
Cause-and-effect analysis, 47
Cause-and-effect-with-added cards, 78
B
Bathtub failure, 236239
Battery case failures
description of, 228
Hong Kong subway fire, 231233
miners backpack, 228231
429
430
D
Deformation, 64
Degradation
corrosion-related, 172173
environmental, 172173
Demise
early, 27
slow, 27
Diesel fuel pipe connector failure, 199205
Differential scanning calorimetry, 121124
Disputes
commercial, 15
Index
E
Early demise, 27
Elastic energy, 39
Elastomer seals, 47
Electromechanical devices
description of, 414
residual current devices
description of, 414
patent infringement suit
patent-in-suit, 415418
Power Breaker v. Volex et al.,
414415
Protector, 420422
technical effect modeling,
418420
trial, 423425
Energy dispersive x-ray analyzer, 103104
Engineering materials
description of, 2122
historical description of, 22
limits of, 25
mechanical properties of, 172
strength range of, 35
tensile strength of, 35
Environmental degradation, 172173
Environmental stress cracking, 58, 192
Evidence
description of, 1619
photographic, 19
Experts
description of, 16
report prepared by, 1921
Eye injuries
clutch plate replacement, 265267
contributory negligence, 264265
description of, 264
lath axe, 267270
printers drying frame, 270273
431
F
Factor of safety, 141
Failure
axle bracket in racing motorcycle
sidecar, 7684
common sense ideas for analyzing, 6
crutch, 313315
definition of, 27
diesel fuel pipe connector, 199205
example of, 2830
fatigue-related
description of, 4351
incidence of, 48
preventive maintenance to
prevent, 50
footrest, 146151
freight container
case study of, 133144
description of, 133
factor of safety, 141
investigation of, 134
load considerations, 142143
rivet fractures, 134136
seams, 133134
gear, 144146
historical descriptions of, 22
hooks, 155158
incidence of, 2526
initial approach, 8689
keeper, 155158
manhole step, 146151
mechanical overload-related
fracture surfaces, 4142
polymeric materials, 3941
tensile strength measurements,
3539
medical implements, See Medical
implements
nerve probe, during surgical procedure,
315316
previous repair as cause of, 27
prototype, 7
replacement gears for heavy truck
gearbox, 152155
rivet, in multipurpose tool, 158159
root causes of, 89
scaffolding guard rail, 159165
sudden, 27
threaded iron pipe coupling, 151152
432
traumatic, 131132
wear and tear, 25
wire hook, 155158
Failure modes and effects analysis, 810,
1314
Fatigue
brake failure caused by, 332334
cracking caused by, 46
description of, 43
factors involved in, 291
failure caused by
description of, 4351
hip joint prosthesis, 290292
incidence of, 48
preventive maintenance to prevent,
50
fractures caused by
appearance of, 4445
characteristics of, 45
double bending, 4546
initiation sites of, 4547
torsional, 5152
incidence of, 48
load bearing component design
considerations, 4851
metals, 4748
tow bolt, 344347
wheel detachments caused by, 328331
Fatigue cracking, 33
Fault-finding checklist, 34
Faults, 7376
Fault-tree analysis, 47
Femoral stem failures, 292296
Flexible pipe, 173
Fluid transport
description of, 171
materials for, 171173
plastic materials used for, 173
FMEA, See Failure modes and effects analysis
Footrest failure, 146151
Force, 62
Force-extension curves, 3637
Forensic engineering
causal analysis, 37
court action role, 426
definition of, 13
objective of, 85
Formaldehyde, 192
Foundry explosion case study
background, 255257
investigations, 257258
outcome of, 260261
pleadings, 258260
Fourier transform infrared microscopy, 297
Fractography, 100102
Fractures
breast tissue expander, 298303
double bending fatigue, 4546
fatigue-related
appearance of, 4445
characteristics of, 45
double bending, 4546
initiation sites of, 4547
torsional, 5152
overstress, 41
rivet, 134136
torsional, 5152
Fracture stress, 38
Fraudulent insurance claims
car keys, 350351
description of, 349350
engine damage caused by loss of oil,
354358
oil filter perforation, 352354
three-car road accident, 358363
Freight container failure
case study of, 133144
description of, 133
factor of safety, 141
investigation of, 134
load considerations, 142143
rivet fractures, 134136
seams, 133134
Frictional heating, 330331
Frozen-in strain, 230
Fuel pipe failure
abrasion, 196197
case study of, 194199
cracking, 196197
description of, 173
diesel fuel pipe connector, 199205
Fusion welding, 34
G
Galvanized tank leak, 216218
Gaskets, 182183
Gas molding, 165170
Gear failure in oil pump, 144146
Gear wheel failure, 153
Index
H
Hardness testing
Brinell hardness test, 108
description of, 107
Mohs hardness scale, 107
Rockwell hardness test, 109
Vickers hardness test, 107108, 157
Heat-affected zones, 98
Heavy truck gearbox, failure of replacement
gears for, 152155
Hip joint failures
acetabular hip shell contamination,
293295
description of, 290
fatigue-related, 290292
femoral stem, 292296
Hoist carrier, 8789
Hook failure, 155158
Hovermower, 395396
Hydrogen-induced cracking, 156
I
Incident scene inspection, 87
Indirect strain measurements
brittle lacquer for, 115117
description of, 115
Infrared spectroscopy, 124126
Injection molding, 165
Insurance claims, fraudulent
car keys, 350351
description of, 349350
engine damage caused by loss of oil,
354358
oil filter perforation, 352354
three-car road accident, 358363
Intellectual property cases
description of, 379380
forms of, 379380
lawnmowers, See Lawnmowers
patents, 380381
wheeled containers, See Wheeled
containers
433
Investigations
description of, 9091
incident scene inspection, 87
initial approach, 8689
macroscopic examination, 93
metallographic sectioning
cast irons, 9599
description of, 9495
observations
description of, 86
initial, significance of, 8789
optical microscopic examination, 94
photography, 9192
product and material standards, 9293
sample handling, 9192
scanning electron microscopy
battery case failures evaluated using,
231232
complexity of, 102
description of, 100102
drawback of, 102
energy dispersive x-ray analyzer and,
104
environmental, 102104
fractographs, 299
fracture striations, 238
lamp filaments, 337
radiator failure evaluations, 223225
situation assessments, 8990
Iron, electrode potential of, 176
Iron pipe coupling failure, 151152
K
Keeper failure, 155158
L
Ladder
accidents involving, 273286
description of, 6266
design improvements for, 284286
Lamp filament failure
description of, 335336
mechanisms, 336338
motorcycle accident caused by, 338340
scanning electron microscopy of, 337
Latent defects, 13
Lath axe, eye injury caused by, 267270
Lawnmowers
434
M
Macroscopic examination, 93
Magnetic particle inspection, for crack
detection, 118
Magnetic resonance imaging, 127
Manhole step failure, 146151
Manufacturing, 78
Manufacturing defects
casting shrinkage, 3233
description of, 3031
freight container failure caused by,
133144
gear failure in oil pump caused by,
144146
keeper failure caused by, 155158
Index
Metal(s)
constituent atoms in, 36
corrosion of, 174
corrosion resistance of, 175
ductile properties of, 39
electrode potentials of, 53
fatigue of, 4748
plastic deformation of, 37
solidification of, 131
stress corrosion cracking of, 5557
theft of, 370372
yield strength increases, 37
Metallography
definition of, 9495, 9798
manufacturing route revealed by, 98
Microscopic examination, 94
Mohs hardness scale, 107
Molding
gas, 165170
injection, 165
Monocoque lawnmowers, 397398
Murphy Infants v. Fiat Spa, 198
N
Necking, 38
Needle corrosion, 303305
Nerve probe failure during surgical
procedure, 315316
Nuclear magnetic resonance, 127
Nylon
6,6, 202
glass-reinforced, 218219
O
Oil delivery pipe, corrosion failure of,
175176
Oil filter perforation, 352354
Oil pump, gear failure in, 144146
Omission-related manufacturing defects,
3133
Optical microscopic examination, 94
Overstress fractures, 41
Ozone cracking, 196
P
Pareto analysis, 7
Pareto diagram, 78
435
Patents
description of, 380381
infringement suits
lawnmowers, See Lawnmowers
residual current devices, See
Residual current devices
wheeled containers, See Wheeled
containers
Photography
evidence, 19
investigational uses of, 91
Pipe
connections between, 172
corrosion failure of, 175176
coupling failure, 151152
flexible, 173
fuel, See Fuel pipe
oil delivery, 175176
plastic, failure of, 335
stainless steel, 181184
stress concentration, 172
vehicle brake, 177181
Pitting corrosion, 174
Plaintiff, 16
Plastics
environmental cracking of, 58
pipe, failure of, 335
storage vessels composed of, 216
water supply system use of, 173
Plumbing failure
acetal polymer, 192
background, 184185
fracture surface, 187189
molding features, 191192
outcome of, 194
reasons for, 186187
in United States, 193194
water pressure effects, 189191
PMMA, 236237
Polycarbonate, 229
Polyethylene, 40
Polymer(s)
acetal, 192
advantages of, 22
crystallinity of, 122
disadvantages of, 22
Polymeric materials
microcracks in, 3940
overloading of, 3941
Polypropylene, 40
436
Q
Quenching, 91
R
Radiator failure, 218227
Radiography, for crack detection, 118120
Replacement gears for heavy truck gearbox,
152155
Repose angle, for ladder, 277278
Residual current devices
description of, 414
patent infringement suit
patent-in-suit, 415418
Power Breaker v. Volex et al., 414415
Protector, 420422
technical effect modeling, 418420
trial, 423425
Risk priority number, 1314
Rivets
failure of, in multipurpose tool, 158159
fracture of, in freight container failure,
134136
Road traffic accidents, component failure in
brake pipe failures, 331335
lamp filaments
description of, 335336
failure mechanisms, 336338
motorcycle accident caused by,
338340
scanning electron microscopy of,
337
overview of, 319
steering arm from truck, 320327
tow hitches
description of, 340
tow bolt fatigue, 344347
trailer accident caused by, 340344
wheel detachments, 328331
Rockwell hardness test, 109
Root causes, 89
Rubber, flexibility of, 40
S
Sample handling, 9192
Scaffolding guard rail failure, 159165
Scanning electron microscopy
battery case failures evaluated using,
231232
complexity of, 102
description of, 100102
drawback of, 102
energy dispersive x-ray analyzer and, 104
environmental, 102104
fractographs, 299
fracture striations, 238
lamp filaments, 337
radiator failure evaluations, 223225
Screw thread failure, 189
Sealing material failure
case studies of, 205207
description of, 205
seals, 208
Secondary defects, 12
Sharp radii, 227
Shear stress, 141
Shotgun pellets, 368370
Slow demise, 27
Spectroscopy, 124126
Stainless steel
corrosion resistance of, 175
molybdenum-bearing type of, 184
pipe clamp corrosion case study,
181184
Index
T
Tensile strength
definition of, 35
437
438
U
Ultrasonic crack detection, 118
Ultraviolet degradation of battery cases,
233236
Underground storage tank failure, 239241
V
Vehicle brake pipe failure, 177181
Vickers hardness test
case study applications of, 157
description of, 107108
tensile strength and, 109111
W
Washers, 210
Weaknesses, 7376
Welding
defects related to, 3334
failures associated with, 132
Welds
defective, 132133
storage tank failure caused by, 245248
Wet corrosion, 54
Wheel detachments, 328331
Wheeled containers
description of, 381382
design of, 382
durability issues, 386
growth in use of, 386
Y
Yield strength, 37
Yield stress, 37
Z
Zinc
electronegativity of, 216
plating of steel, 157