Optimal design of step-stress accelerated degradation test oriented by nonlinear and distributed degradation process

Abstract

The Step-stress accelerated degradation test (SSADT) is an efficient tool to assess the reliability of products. To ensure an accurate evaluation of reliability-related indexes with limited cost, the optimal design for SSADTs is widely applied. Conventional optimal design methods ideally assume the degradation process to be linear and distribution-free, which may cause the designed test plans to be drastically different from the optimal one. To overcome the drawbacks of the unreasonable assumption, this paper develops a new optimal design method for SSADTs, including a generalized Wiener process-based degradation model (GWPDM) considering the distribution feature and a corresponding optimization criterion aiming at accurate prediction of degradation distributions. The GWPDM is first discussed, where the influence of the nonlinearity, unit-to-unit variability, measurement error and stress-related temporal variability on the degradation process is clarified. The statistical inference of the developed GWPDM is conducted to deduce the proposed optimization criterion. Thus, the optimization model is established to determine the optimal SSADT plan. The proposed method is used to design a SSADT plan for the metal-film resistor. Finally, the necessity and the robustness of the proposed method are verified.