Reliability assessment for nonlinear degradation data with individual variation

Abstract

Aiming at nonlinear degradation data on occasion of Constant-Stress Accelerated Degradation Test (CSADT), a reliability assessment method of nonlinear accelerated degradation based on Wiener process is verified in consideration of individual performance degradation variation. Wiener process is used to describe the degradation process of products. The time scale model is used to change nonlinear data into linear data. Drifting coefficient of Wiener process is treated as a random variable. The reliability model and accelerated degradation model are built in consideration of individual variation. The Two-Step Maximum Likelihood Estimation method (TSMLE) is used to obtain the unknown parameters in the reliability model. An example is given to show the accuracy and superiority of the presented method.