Optimal stress screening strategies for multi-component systems sold under warranty: The case of phase-type lifetimes

Environmental Stress Screening (ESS) is employed to reduce, if not eliminate, the occurrence of early field failures. This paper examines the necessary trade-offs between the reduction in warranty costs and the increase in manufacturing costs associated with optimal stress screening strategies. A multi-level ESS model is presented for a multi-component electronic system. Screening can be performed at component, unit, and system levels. Components and connections are assumed to come from good and substandard populations and their time-to-failure distributions are modeled by mixed distributions. The majority of ESS models found in the literature assume that the time-to-failure distributions are exponential. The exponential distribution is used primarily to take advantage of its mathematical tractability. This paper generalizes previous work by modeling component and connection lifetimes with phase-type distributions. Phase-type distributions offer the advantage of mathematical tractability as well as versatility in the family of distributions they can represent. To date there is no significant research into the impact that the selection of a lifetime distribution for modeling the failure process has on ESS decisions. In this paper, we evaluate screening strategies for several lifetime distributions. Numerical examples are provided to illustrate the effect of various model parameters on the optimal stress screening strategy.