Page 270 - Handbook of Materials Failure Analysis

266 CHAPTER 11 The reliability design of mechanical system

We also cannot guarantee the behavior of a product over 10 years under the
extreme environments. These test conditions would be appropriate to mechanical
items, like power engines, but the test conditions are not fit to assess the degradation
of paint on the automobile body. In addition to testing the new engine, we should
devise quantitative test methods for other items—new electrical components (includ-
ing batteries), electronic control units, lighting systems, or coating materials.
According to the identified failure mechanisms, testing must be conducted by
subassembly to heighten acceleration.
Without such quantitative lifetime testing, we can’t identify all the failures
influencing the product’s lifetime because there would be unanticipated failures.
For example, in my experience of prototype testing, the lifetime of a tub in a washing
machine was lengthened at first by small structural changes—a corner radius
increase, rib insertion, and so on. The final test, however, showed a weakening of
strength in the plastic due to a chemical reaction; the problem was solved by chang-
ing the release agent of the injection molding process—something no one could have
been predicted as the solution. Note that this method reveals exact failure modes,
including totally unexpected ones, that other methods, like failure mode and effects
analysis, cannot identify.
For the CFRP of the 787 Dreamliner, the failure mechanism is mainly delam-
ination, which can be found in pressure/humidity/temperature cycling and ultra-
violet irradiation testing. If the AF for testing this is calculated using an adequate
life-stress model (time-to-failure model) and the sample size is determined
according to the BX life target, then quantitative results can be derived. Note that
we should check the possibility of failure due to various overstresses, such as bird
strikes, with sufficiently degraded samples. For the electrical systems in the 787,
engineers should incorporate the same components used in other commercial air-
planes (and different combinations of them), assessing the possibilities of overs-
tress failures under reliability marginal stresses, since they can assure lifetime
reliability. But for new components like the lithium-ion battery, the failure
mechanics as well as the stresses produced in the aircraft environment have been
changed. Thus, we cannot presume what kind of failure mechanisms would occur
due to chemical reaction until the projected lifetime reaches. Generally, chemical
failure mechanisms are delicate and thus difficult to identify and reproduce,
which means that the acceleration conditionsand relatedfactors canbehardto
determine. Thus, they should test until lifetime under almost normal conditions,
and the behavior of sufficiently degraded components should be checked under
the rated stresses and overstresses. The media have reported various accidents
or disasters due to unanticipated failure mechanisms in chemical items, such as
the fires occurring in the Sony lithium-ion battery in notebook computers in
2005, Firestone tires causing Ford Explorer rollovers in the 1970s, wire bundles
incorporating silver-plated copper wire leading to fire in the Apollo 1 cabin in
1967, and so on.