Page 161 - Failure Analysis Case Studies II

146
Since the piston rods were welded semi-automatically with a stationary electrode and a rotating
work piece, it was suggested that applying a gas flame for two revolutions would suffice as an
intermediate solution. Induction heating was recommended as the long term solution, since the
temperature profile will be more even through the thickness of the joint, the temperature of the weld
joint is raised to the preheat temperature without overheating and oxidising the surface, and it is
faster than flame heating.
Weld penetration was improved by increasing the current by 20% and reversing the polarity to
electrode negative (DC current) for maximum depth to width ratio.

6.4. General
The above recommendations were intended specifically to remedy the problems with the weld
failures. During the investigation a number of other areas that required attention came to light.
0 En9 is a hardenable steel but the hardenability properties were not being used-instead, the
weldability was reduced. Steels with similar strengths but lower carbon equivalents and thus better
welding properties and lower preheat temperatures needed to be considered. Alternatively, by
quenching and tempering the rods, higher strengths can be achieved and thus a reduced wall
thickness might suffice, resulting in lighter components.
0 The refractory compound shielding on the adapter that was intended to prevent carburising of
the area that is to be welded was found to be unsatisfactory. Possible alternatives included copper
plating and removal of the carburised area by machining. However, if the copper plating is not
removed, this could lead to weld contamination.
0 Much of the effect of carburising was cancelled by the nitriding process, Le., the hardness of the
case was reduced from the specified 60 HRc after carburising to f 40 HRc after nitriding. Since
nitriding was a major contributor to the failure, the value of this process needed to be critically
evaluated.

7. SUMMARY
The problem of the failed actuator piston rods was analysed and solved, the solution tested and
successfully implemented into the large scale production of the actuators within two weeks. No
further weld fractures were reported.
A large number of metallurgical aspects of the weld was considered in this investigation, most of
which were not profound revelations to anyone versed in materials science and did not require
exotic testing facilities or specialised expertise. The main point that the authors wish to emphasise,
with this investigation serving as an example, is that careful consideration should be given to every
stage of a design, processing or material selection change in any product. Too often, metallurgical
aspects of such changes are ignored or glossed over, each change eroding the “metallurgical safety
factor” until the sum of the changes results in a non-viable component that enters service flawed or
has marginal mechanical properties.
Any one of the changes outlined in Table 1 would not in itself have resulted in the failures, but
their implementation left fewer options and less room for error for the following changes. For
example, changing the surface treatment (based, at the time, on the customer’s colour preferences
rather than corrosion or wear properties) raised the processing temperature from 70°C (phosphating)
to 580°C (nitriding) which had far-reaching effects on tempering, precipitation of carbides and other
brittle phases, residual stresses, etc. but may have proved insignificant if the base material had had
a lower carbon content or a low carbon/low alloy filler metal had been utilised.
This is where a well managed and frequently examined change management system such an IS0
can play a vital role in guaranteeing ongoing unwavering quality in the face of changing suppliers,
process capabilities and customer requirements-but not without the watchful guidance of a
materials engineer.