Bayesian Online Reliability Estimation Based on Two-Stage Wiener Process Degradation Model

Deteriorate process of products with the effect of various degradation mechanism or complex components often shows a multiphase degradation phenomenon. Misspecification of single-phase degradation model will lead to poor performance. To make full use of degradation signals, a two-stage degradation model considering random effects on deteriorate rate and change time point is established to enhance the accuracy of degradation description and reliability assessment. Online parameter updating and reliability estimation are applied to reflect the equipment status and capability. We also analyze the reliability estimation error that a product with two-stage degradation signals may introduce when change point time has not shown off. Simulation study is conducted to verify the proposed method.