When batches of critical, very high reliability single use parts are being produced, rigorous testing is often required to qualify the parts and allow them to be used by the customer. Frequentist and Bayesian approaches are described for predicting the reliability of a subset of the batch, conditional on all of the tested parts working cor- rectly. Answers from di erent methods are compared, their strengths and weaknesses considered, and their robustness to initial assumptions examined. Some related ques- tions are explored about the impact on reliability from di erent choices of the relative size of the tested and sale units and the condition for passing the batch from both the manufacturer and customer's point of views. We describe the approach in the context of automotive airbag in ation devices, which are standard on most vehicles, but the approach is relevant for batches of single use parts which have a very high requirement for reliability and must be destructively tested.
Citation: Quality Engineering
Pub Type: Journals
destructive testing, very high reliability, incorporating expert knowledge, batch testing.