Abstract:

Degradation data are becoming more and more important in reliability assessment of products with small samples, high reliability and longevity. Aiming at reliability assessment of multi-stage electronic products with calibrations in the practice of engineering, a degradation model is developed in terms of degradation signals of multi-stage systems based on Wiener diffusion processes under pre-specified periodical calibrations, and the drift coefficient in the degradation model is inferred by the Bayesian method. Then, a smooth function is built to solve the problem of discontinuity resulted from pre-specified calibrations, based on which a method for Bayesian reliability assessment is presented for multi-stage system with calibrations in terms of the distribution of first passage time. Finally, a numerical example is given to validate the validity and practicability of the model, method and analysis results.