A “Toolbox” for Forensic Engineers

Failure Due to Manufacturing Faults 141
investigation was in progress, a mechanical engineer was addressing the
complementary questions of what level of stress the rivets were subjected
to when they failed and why only containers used to transport machinery
had “unzipped.” At first sight this may appear to be a rather academic
question; the behavior of the rivets themselves may be thought to provide
a sufficient answer. Because they fractured in shear they must have experienced
a set of shear stresses with little or no normal stress. Furthermore,
the magnitude of this stress must have exceeded the shear strength of the
rivet — the material failed by overstressing. On the other hand, fully loaded
containers without the defective rivets did not fail, so the shear stress on
their rivets must have been less than the shear strength of the rivets in the
specified condition.
The shear stress is therefore known to lie between these limits, but it is
not known accurately, so the magnitude of the safety margin is not known.
To allow for the unknown, it is usual in design to use a static factor of safety
(or load factor) of a component, which is defined as
nominal static strength of the component
factor of safety =
nominal static loading of the component
“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.