3 0 CASE STUDIES

STUDYING FRACTURES:
A DV A N C E D M AT E R I A L S & P RO C E S S E S | S E P T E M B E R 2 0 2 2

RECOGNIZING AND
UNDERSTANDING FAILURE MODES
Examining a failed part’s fracture surface is a great source of information
as to why and how it failed.
Shane Turcott*
Steel Image, Dundas, Ontario

T
o understand why a part or equip- link had failed and to devel-
ment failed, how it failed must op effective solutions. With
be first diagnosed. Different fail- minimal training, the failure
ure modes such as fatigue, brittle, and modes of all three failures
corrosion all have very different caus- would have been identifi- (a)
es and very different solutions. If an in- able to engineers and met-
vestigation were to guess at the failure allurgists had they been
mode and guess wrongly, the planned properly trained in fracture
remediation may not address the true recognition.
cause of failure and additional failures Or phrased differently,
may occur. had someone not studied Fatigue
fracture
The only way to diagnose how a the links’ fractures, what
features
part failed is by examining the broken are the odds that anyone
part. The various failure modes form would have guessed how
different features which in turn can be to prevent repeat failures (b)
helpful for diagnosis. Some damage for all three? Remember, if
mechanisms such as complex corrosion even one guess were made Fig. 1 — Failed chain link (a) shows signs of fracture (b) due
cracking require lab-based analysis. Yet wrong, people might have to fatigue.
for mechanical failures such as ductile, gotten hurt.
brittle, and fatigue, a lot can be learned way to cause fatigue failure of a perfect-
just by visual examination of the part. CASE STUDY 1 ly good link was from excessive, repeti-
It is possible for reliability workers, The first chain link was from a line tive loading.
mechanical engineers, and inspectors that suffered over a dozen failures in Therefore, chain link failure was
dealing with failures to learn how to di- two years. Although each failure had from cyclic loading higher than it could
agnose mechanical fractures. been taken seriously, these historical sustain. The remainder of the investi-
To demonstrate the importance of investigations were conducted without gation could then be focused on con-
identifying how something failed, con- study of the links themselves. sidering the source of elevated loading
sider three failed chain links. All three Examination of the most recent (excessive line tension, small radius
are from automotive assembly plants’ link’s fracture surface confirmed that turns, and so on). Also, considering the
conveyor chain systems. Their failures the failure mode was fatigue (Fig. 1). entire chain had experienced the ex-
represented lost production costs and With this diagnosis as to how the link cessive loading, it was likely that nu-
created safety issues for people assem- failed, the next question is naturally, merous other links had already formed
bling the cars. Although chain links are why did the link fail by fatigue? Fatigue cracks. Therefore, the chain needed to
a simple part, they can fail in a variety of failure only occurs when the repetitive be replaced after the source of elevated
manners and for very different reasons. loading exceeds the fatigue limit/du- loading was removed/reduced.
Identifying the failure modes was cru- rability of the steel. Once the link was After a dozen failures, had the lat-
cial to understanding why each chain shown to be of sound quality, the only est link been simply replaced, the chain
*Member of ASM International
 31

features to identi- Although the diagnosis as pos-

A DV A N C E D M AT E R I A L S & P RO C E S S E S | S E P T E M B E R 2 0 2 2
fy the failure mode, sible brittle fracture is well within the
it is critical to locate skill limits that reliability/mechanical
and examine where engineers should develop, to under-
(a) the cracking initiat- stand why something had been brittle
ed. The initiation site requires metal testing. Yet upon recog-
sometimes provides nizing the fracture features, at least the
critical information investigator would know what steps
relevant to how and would be required to ensure that the
why a part failed. In case was solved. For this link, its em-
Fatigue fracture the case of this sec- brittlement was determined to be from
features ond chain link failure tempered martensite embrittlement,
(Fig. 2), the fracture which is caused by tramp alloying and
features were also heat treatment, both associated with
fatigue, indicating part quality.
Forging flaw the repetitive load-
ing had exceeded the CONCLUSION
(b) strength of the steel. Components as simple as chain
Yet looking at the ini- links can fail by numerous failure
Fig. 2 — Looking at the failed chain link (a) revealed that cracking
tiation region, crack- modes, each caused by different fac-
started from a forging flaw (b).
ing had started from tors. One should never guess why
a forging flaw. This something failed because, if they did,
pre-existing, crack- they would sometimes guess wrong. To
like feature had in- understand why a part failed, one must
creased the local first accurately identify how it failed.
stresses, which then That can only be done by examining the
caused fatigue crack- failed part itself. For anyone involved
(a) ing. Ultimately, this with investigating or preventing equip-
link failed because ment failures, learning to recognize the
of a quality issue with common mechanical failure modes is
the link from original extremely helpful to equipment reliabil-
manufacture. ity efforts.
Understanding
Note
how the part failed
from study of its frac- This article is based on Decoding
ture often includes Mechanical Failures: The Definitive Guide
additional informa- to Interpreting Fractures by Shane Tur-
tion relevant to why cott, and is available in the ASM book-
it failed. If the bro- store, https://bit.ly/3zRC1Ks. This book
ken link had not been introduces fractography and how to de-
studied, it was unlike- code the fracture features of mechan-
(b)
ly that anyone would ical failures. It demonstrates how to
Fig. 3 — Failure of the third chain link (a) was consistent with brittle visually diagnose and interpret ductile,
have guessed its fail-
failure (b), later found to be caused by embrittlement. brittle and fatigue failures using numer-
ure was due to a forg-
ing quality issue. ous examples. It then explains how each
would have continued to suffer more diagnosis is used to direct the investiga-
failures. If the loading on the chain CASE STUDY 3 tion towards the root cause of failure.
had been reduced yet the chain not re- There are two additional chapters on
The fracture features of the
placed, links already cracked would advanced fatigue of rotating shafts and
third chain link were distinctly differ-
have caused more failures. If the chain static fastener failures. ~AM&P
ent because it had failed by a differ-
had been replaced without reducing ent mechanism. Its fracture features
the loading, the new chain would have were consistent with brittle failure, lat- For more information: Shane Turcott,
eventually suffered failures, too. er found to be caused by embrittlement Steel Image, 7 Innovation Dr., Suite
(Fig. 3). Due to a quality issue with the 155, Hamilton, ON L7H 7H9, Canada,
CASE STUDY 2 steel, it did not have the ability to sus- 905.745.6429, [email protected].
In addition to studying fracture tain bending loading without snapping.
 SEPTEMBER 2022 | VOL 180 | NO 6

NDT AND FAILURE ANALYSIS

Ultrasonic Defect
27 Detection

Decoding
30 Failure Mechanisms

ASM Materials Education