Parametric Reliability Analysis

Many reliability engineers do not understand how to perform parametric reliability analysis. Yet it can save time, sample size, and money. As well, predictions can be more accurate from accelerated test data. And one does not even have to reach a failure threshold to make accurate predictions. That means time savings and lower acceleration stress can be used. Typically, what engineers alternately do is put a large number of devices on accelerated test and monitor key parameters and wait until enough devices that reach their failure threshold, say 15% tolerance change in Delta_P, then the analysis proceeds by using a traditional catastrophic approach like Weibull analysis. This seminar will provide details of how one can have another approach for doing parametric reliability analysis of test data and learn all its advantages: More efficient qualification testing Can obtain a point estimate even without failure Predict failure before it occurs testing to failure � not required Can establish aging law with a small sample size Lower stress levels More accurate predictions Quick feedback for design improvements If time permits, Alec will also talk about his new research on thermodynamic parametric reliability analsysis.