Optimal Planning of Destructive Degradation Tests

In many reliability applications, degradation of products has to be measured destructively, and destructive degradation tests are a useful tool to evaluate product reliability. Previous studies commonly obtain an optimal degradation test plan numerically with given planning values of model parameters, under a default assumption that stress acceleration is necessary. In this study, we use the general path model for product degradation and derive closed-form solutions for optimal designs of a destructive degradation test, under the planning objective of minimizing the asymptotic variance of the estimated lifetime quantile. Based on these results, we further identify conditions under which acceleration is necessary or unnecessary. It is found that the key parameters that determine the necessity of acceleration are the lifetime quantile and the parameter associated with the accelerating variable in the link function. We demonstrate the application of the results by applying them to an adhesive bond test and a seal test. Several compromise plans are investigated to robustify the test. Supplementary materials for this article are available online.