Optimal reliability growth program for repairable and warranted products

Abstract

Considering a two-dimensional and non-renewing free repair warranty (NFRW) policy, this paper proposes a modified reliability growth model to support the decision making of the manufacturer on the choice of a reliability growth test program for repairable products with a two-fold Weibull lifetime distribution, which has a bathtub-shaped failure rate function. Excess usage is used as a stress to trigger failures and the effect of usage acceleration on product reliability is modeled by the Accelerated Failure Time (AFT) approach. Learning from test failures is incorporated into periodic fixes through which the overall product failure intensity is reduced through discrete steps. The optimal test duration and the optimal number of fixes are both obtained such that the expected total cost per product to the manufacturer is minimized while achieving the stated reliability growth target at the same time.