Life Prediction Approach by Integrating Nonlinear Accelerated Degradation Model and Hazard Rate Model

摘要

Accelerated degradation testing and modeling is an effective way to predict the lifetime when the failure data are rare and the testing is costly. Under this consideration, we first proposed a novel method of lifetime estimation by integrating nonlinear accelerated degradation model and hazard rate model. Then, the parameters of acceleration model are estimated based on the accelerated degradation data, which can be used to obtain the relationship between acceleration model parameters and stresses. Further, we provide a lifetime estimation approach fusing the failure data and censored data based on Cox proportional hazards model. In addition, to facilitate the practical usage of our model, we introduce a two-step method of model identification based on MLE (Maximum Likelihood Estimation). Finally, a practical case is adopted for illustration.