Chapter-8 Engineering Failure Analysis
Chapter-8 Engineering Failure Analysis
Chapter-8 Engineering Failure Analysis
8.1
ENGINEERING FAILURE
ANALYSIS
8-1 | P a g e
service to meet its functional requirements specifying the procedures of start-up and
shut down as well as the maintenance procedure and its schedule.
8.2 FAILURE OF ENGINEERING PRODUCTS
When an engineering product ceases to perform on e or more of its functions well
before its expected service life, it is said to fail. Even though designers and
manufacturers continue to strengthen the links among design, manufacturing, and
performance, failures still occur and will continue to occur for one reason or
another. Failures can cause loss of lives, unscheduled shut downs of plants,
increasing maintenance and repair costs as well as very costly litigation and
damages . Any type of failure can be expected to occur at the weakest link in the
chain of the continuous process of engineering design, fabrication and performance
as demonstrated b y the following examples.
8.3 DEFINITION AND OBJECTIVE OF FAILURE ANALYSIS
INVESTIGATIONS
A failure analysis is an investigation carried out to determine the cause of failure of
a certain product or equivalently the mistake in the continuous process of
engineering design manufacturing performance in order to prevent its recurrence in
the future. After the investigation is completed, its findings are documented in a
report. It is important to realize, however, that failure analysis investigations can
serve other useful purposes, as described below. Results of failure analysis
investigations can be extremely useful to designer s of the same or similar products.
Knowledge of prior incidents of failure and their causes aid the designer in
developing a product of improved reliability and durability. Frequently, however,
such information never reaches the designer for one reason or another. Because of
the legal aspects of failures, the results of failure analysis investigations can be used
as the basis for litigation and insurance claims. Manufacturers can also use the
results of failure analysis investigations for marketing purposes to promote a new
product with better performance capabilities. Furthermore, by identifying the
deficiencies of certain structural materials through failure analysis investigations, it
is possible to develop new materials or improve the properties of existing materials.
8.4 TYPICAL DEFECTS
A wide variety of defects can be found in a given engineering component. These
flaws may result from such sources as material imperfections, defects generated
during service, and defects introduced as a result of faulty design practice.
Regarding the first source mentioned, defects can be found within the original
material supply or can be introduced during the manufacturing process. Typical
material defects include porosity, shrinkage cavities, and quench cracks. Other
micro-structural features can trigger crack formation if the applied stresses exceed
some critical level. These include nonmetallic inclusions, unfavorably oriented
forging flow lines, brittle second phases, grain-boundary films, and micro-structural
features resulting from 300°C and temper embrittlement. The list of manufacturing
defects includes machining, grinding and stamping marks (such as gouges, burns,
8-2 | P a g e
tears, scratches, and cracks), laps, seams, delaminations, decarburization, improper
case hardening, and defects due to welding (e.g., porosity, hot cracking, cold
cracking, lack of penetration, and poor weld bead profile).
Defects can be introduced into the component during service conditions as a result
of excessive fretting and wear. Environmental attack can also cause material
degradation as a result of general corrosion damage, liquid metal and hydrogen
embrittlement, stress corrosion cracking, and corrosion fatigue. Surely, cyclic
loading can initiate fatigue damage without an aggressive environment and may lead
to serious cracking of a component.
Finally, defects can be introduced into a component through faulty design. These
human errors include the presence of severe stress concentrations, improper
selection of material properties and surface treatments, failure to take remedial
actions (such as baking a steel part after it has been cadmium plated to remove
charged hydrogen gas), inadequate or inaccurate stress analysis to identify stress
fields in the component, and improper attention to important load and environmental
service conditions as they relate to material performance.
8.5 FAILURE ANALYSIS PROCEDURE
Generally, the exact procedure for failure analysis is dependent upon the particular
case. However, any successful and complete failure analysis investigation should
provide answers to the following questions:
1. What is the mode of failure?
2. What is the most probable cause of failure?
3. How can failure be prevented in the future?
It is important to realize that any failure results from some sort of a mistake causing
a weak link in the chain of the continuous process of engineering design-
manufacturing- performance. Therefore, to provide clear and straight answers to the
above questions, any failure analysis investigation must address three main issues:
(i) service conditions, (ii) manufacturing, and (iii) design. By a process of
elimination, it is possible to narrow down the most probable cause of failure and
then identify the weakest link in the above process. Subsequently, it becomes
possible to determine possible means by which future failures can be prevented.
Any effective failure analysis investigation must involve three main stages: (i) clear
definition of the problem, (ii) complete search of the literature to develop a good
technical background in the problem area as required, and (iii) development of an
experimental program which must be well oriented and coordinated to address the
problem. Clear definition of the problem is the most important of the above three
aspects. It is evident that the uninformed engineer or scientist cannot make a
contribution to solving a problem that is not understood. Developing the proper
technical background, as well as the experimental program necessary to solve the
problem, is critically dependent upon a clear definition of the problem. Since the
time required to complete a failure analysis investigation can be very important in
many cases, the effective analyst must be capable of recognizing problems and
8-3 | P a g e
communicating these problems to others with technical knowledge in a timely
fashion.
8.6 DEFINITION OF THE PROBLEM
A failure analysis investigation begins by the proponent calling the analyst to
investigate a given problem. Initially, the analyst meets with concerned personnel
such as process engineers, maintenance engineers, etc. to discuss the problem.
Typically, the discussions begin by the proponent giving some description of the
problem; e.g. during regular maintenance, cracks are detected in a certain
component, a part has undergone fracture, leaks are detected in storage tanks, etc. It
is the responsibility of the analyst to ask relevant and important questions
concerning the history of the component involved in the failure, such as (i) process
involved, (ii) the application, (iii) type of material, its specifications, shape,
dimensions, and processing technique, (iv) design parameters, (v) service
conditions, (vi) maintenance history, (vii) sequence of events preceding the failure,
(viii) frequency of failure and whether it is of localized nature, etc. Sometimes,
some of this information may not be available. However, it is very important that the
analyst collects as much information as possible about the history of the component
or part involved in the failure.
Inspection of the failure site is a very important aid in defining the problem. A
thorough examination of the condition of the failed component can provide a great
deal of information. Depending upon the particular case, attention must be given to
such features as location of cracks, fracture characteristics, perforations, deposits,
wall thinning, sagging, detached pieces, etc. Also, whenever possible, the analyst
must obtain as much information as possible from eyewitnesses about any abnormal
conditions preceding the failure. In some cases, visual inspection of the failed
component can reveal information which cannot otherwise be obtained. For
example, if fracture is involved in the failure, visual inspection can provide
information about the type of loading, load path, and fracture path. From such
information, it is possible to reconstruct the direction of stresses.
Usually, the fracture path is normal to the direction of maximum principal stress.
Photographic documentation of the failed component is almost always necessary for
future reference and further inspection after leaving the site. It is important to
photograph all features of the failed component. A clear definition of the problem is
largely dependent upon the depth of information obtained, site inspection, as well as
the skill and judgment of the analyst.
8.7 TECHNICAL BACKGROUND
Once the problem is clearly defined, the analyst must be able to identify the
technical background required to solve the problem. Depending upon the skill and
judgment of the analyst, it is possible to define the problem as environmental,
mechanical, or a combination of both. Furthermore, through discussions with
concerned personnel and visual inspection, the experienced analyst can narrow
down the various possibilities. Also, in some cases, the failure may resemble a
previous case. In this manner, the analyst can identify the technical literature
8-4 | P a g e
relevant to the particular problem, e.g., high-temperature corrosion, fatigue,
hydrogen embrittlement, stress corrosion cracking, etc. To illustrate this
methodology with an example, one analyst determines that most evidence points out
that the most probable cause of failure is high temperature carburization attack. It is
by large deposits of coke. This identifies the technical background required to solve
the problem.
8.8 EXPERIMENTAL PROGRAM AND ANALYSIS
Having defined the problem, an experimental program is developed to determine the
cause of failure. Experiments must be coordinated and tailored so that they provide
complementary information leading to identifying the cause of failure. Evidently,
both the nature and number of experiments carried out can significantly differ from
one case to another. Samples used to carry out those experiments must be carefully
selected such that they represent as much as possible all features of the failed
component. Deposits or pieces detached from the component can provide valuable
information. Whenever possible, samples removed from sound sections, as well as
samples never used in service, must also be collected for comparative purposes. It is
also important that the analyst realizes that stress analysis and fracture mechanics
are very valuable tools in failure analysis investigations. Given below is a guideline
for the type of experiments and analysis involved in failure analysis investigations.
Obviously, however, the exact experiments and analysis carried out are dependent
upon the particular case.
Investigative tools used in failure analysis studies can be classified into two main
types: (i) tools which can be used on site and (ii) laboratory tools.
8.8.1 NONDESTRUCTIVE TESTING
It is recalled from earlier discussions that the primary objective of nondestructive
testing is to detect either surface and internal cracks or discontinuities. Non-
Destructive tests are explained in the Chapter No. 2.
8.8.2 MATERIAL VERIFICATION
Although rarely the material of the component turns out to be different from that
specified, it is always recommended to verify the material(s) of the failed
component. In most cases, microchemical analysis by energy dispersive X-ray
spectroscopy in a scanning electron microscope provides sufficiently accurate
results.
8.8.3 THEORETICAL STRESS ANALYSIS
From knowledge of service conditions such as applied loads or pressure, and
geometry and size of the component, it is possible to calculate the principal stresses
developed in the part. Comparing the maxi mum principal stress with the strength of
the material used in the application provides important information about the
suitability of the material for the application and whether the design parameters
were strictly followed during service.
8-5 | P a g e
8.8.4 FRACTURE MECHANICS
Analysis using the principles of fracture mechanics provides valuable information
about the stress developed in the failed component at the time of fracture in relation
to the design stress, particularly in the case of crack propagation by a brittle
mechanism. The outcome of this analysis determines the next course of action in the
investigation. Although fracture mechanics can be used to determine the number of
cycles preceding fracture by a fatigue mechanism, the method based upon
measuring the spacing of fatigue striations can be more convenient provided the
striations are clearly visible on fracture surfaces.
8.8.5 MACROSTRUCTURE AND MICROSTRUCTURE OF THE MATERIAL
A careful and thorough characterization of the macrostructure and microstructure of
the material involved in the failure can frequently provide important clues.
Whenever possible, it is particularly useful to compare the structural features of
samples removed from the failed component with those representative of samples
removed from sound sections, as well as of samples never used in service. It is also
important to refer to any specific recommendations by the manufacturer about the
use of the material. Macro-structural characterization at small magnifications, e.g.,
10, can provide further information about the fracture path and the location from
which it was originated, e.g., internal surface or external surface. Selection of the
proper technique for micro-structural characterization is dependent upon the type of
information required. However, it is always recommended to begin with light
optical microscopy or scanning electron microscopy to characterize the overall
micro-structural features.
In examining the microstructure, particular attention must be given to such features
as grain size and shape, secondary precipitates and their distribution, and micro-
cracks and their location. An unusually large grain size is an indication that the
material is exposed to an excessively high temperature. Grain shape and the shape of
annealing twins provide information about the state of cold work. Elongated grains
indicate that the material is severely cold worked. Plastic deformation can be
reflected by bent twin boundaries. Fine slip lines within the grains also indicate that
the material has been plastically deformed. If secondary precipitates are detected,
particular attention must be given to their nature and morphology. Micro-chemical
analysis by either energy dispersive or wavelength dispersive X-ray spectroscopy
can be used to determine their composition. Provided the precipitates are present in
sufficient amount (.10% by volume), they can be identified by X-ray diffraction. In
certain cases, transmission electron microscopy may be required. While examining
the gross micro-structural features, particular attention must be given to whether
micro-cracks are present. For example, grain boundary cracks may result from either
creep deformation or embrittling the grain boundaries by precipitates of a secondary
phase.
8.8.6 CHARACTERIZATION OF FRACTURE SURFACES: FRACTOGRAPHY
Although a great deal of information about fracture characteristics can be obtained
by means of visual inspection, characterization of the mechanism responsible for
8-6 | P a g e
fracture usually requires micro-structural examination of the fracture surface,
particularly in a scanning electron microscope. It is very important to realize that
any fracture results from an overloading condition. In contrast, cracks can propagate
by various mechanisms. Therefore, the analyst must be extremely careful in
reaching a conclusion about the mechanism responsible for crack propagation if the
entire fracture surface is not available. Sometimes, cracks are not propagated to the
extent causing fracture. If cracks extend through the entire thickness of the part, it is
possible to induce complete separation by certain fixtures and examine the
morphology of crack surface. In some cases, particularly when the component
operates at high temperatures, the details of fracture or crack surfaces can be masked
by oxidation. Descaling in a suitable acid, e.g., HCL, can be used to remove the
oxide layer, revealing the morphology of the fracture surface.
8.8.6.1 Fracture Surfaces Produced by a Cleavage Mechanism
Since most cleavage cracks are transgranular, the fracture surface appears granular
or shiny at small magnifications, and cleavage steps become clearly visible at higher
magnifications. In contrast, a quasi-cleavage fracture exhibits radial markings and
chevrons at small magnifications, and quasi-cleavage steps are observed at high
magnifications. If cleavage occurs on a single crystallographic plane, the resulting
fracture surface is smooth on an atomic scale. However, because engineering alloys
are complex multiphase alloy systems, corresponding fracture surfaces produced by
cleavage exhibit a variety of morphological features. Due to the presence of several
microstructural features, cleavage cracks can nucleate at various locations within the
grains. When these cracks propagate, the crack front can bypass various obstacles as
described earlier producing a fracture surface of complex irregular morphology, as
shown in figure 8.2.
Figure 8.2 Fracture surface produced by cleavage mechanism showing river patterns
8-7 | P a g e
8.8.6.2 Fracture Surfaces Produced by Micro-void Coalescence
Since plastic deformation is involved in producing a fracture surface by micro-void
coalescence, the fracture surface shows evidence for slip lines or bands. When a
void is created at the interface between a secondary phase particle and the matrix,
plastic deformation by slip occurs in the vicinity of the particle. With continued
deformation, the void increases in size, and eventually the interconnecting layers of
the material are separated. This results in a fracture surface consisting of cusps or
dimples. If voids are formed at the interface between the matrix and secondary
phase particles, the centers of some cusps contain particles. Figure 8.3 illustrates a
fracture surface produced by micro-void coalescence. In some cases where the
localized plastic deformation occurs in the vicinity of the voids, slip lines may be
observed at the fracture surface. Straight slip lines are observed in fcc materials if
dislocations are confined to move on single slip planes corresponding to high
stacking fault energy. Conversely, if the stacking fault energy is low and
dislocations can readily cross-slip, the slip lines become wavy. Both the density,
size, and depth of dimples observed on fracture surfaces produced by micro-void
coalescence reflect the amount of plastic deformation, which has preceded fracture.
For highly ductile materials, the fracture surface contains a smaller density of large
deep dimples. As the ductility is reduced, the dimples become finer, shallower, and
of larger density. Although in this case, localized plastic deformation can still occur
in the vicinity of voids, the amount of macroscopic deformation required to
complete the separation diminishes.
8-11 | P a g e
4. Visual inspection, including photographs of the failed component
illustrating specific features such as cracks, corrosion deposit, perforations,
pits, thinning, etc. It is useful to illustrate the locations from which samples
were removed for analysis.
5. Material verification, presenting the results of analyzing the chemical
composition of the material in comparison with the nominal composition if
provided by the proponent. This clearly shows whether the material is the
same as that specified. It is useful to include in this section some material
properties relevant to the application.
6. Stress and fracture mechanics analysis, presenting the results of analyzing
the stresses in the failed component.
7. Micro-structural characterization, presenting comparative micro-structural
features of failed and un-failed sections, as well as those of material never
used in service. Particular attention should be given to such features as
grain size, structure and morphology of grain boundaries, precipitation of
secondary phases. Also if applicable this section may include results of
characterizing fracture surfaces, corrosion deposits, etc. Results of property
evaluation such as hardness measurement may also be included in this
section.
8. Mode of failure, based upon the results presented. The mode of failure is
clearly specified, e.g., fatigue, creep, carburization attack, etc.
9. Cause of failure, combining all above information. The most probable
cause of failure is discussed in this section.
10. Conclusions, summarizing the mode and cause of failure.
11 Recommendations, presenting views about various possible solutions to the
problem.
12. References, used for obtaining information during the investigation.
8-12 | P a g e