基于随机效应Wiener退化模型的剩余寿命预测

针对退化率较高的产品具有不稳定的退化路径以及产品个体差异对退化过程的影响，建立了一种新的随机效应退化模型，即漂移参数和扩散参数均为随机变量且两者之间呈线性关系的Wiener退化过程模型.基于该模型获得了产品剩余寿命分布与可靠度函数，同时设计了估计模型参数的EM（expectation maximization）算法.最后，通过分析钛合金疲劳裂纹数据以及与现有模型结果的比较，验证了所建模型的有效性和准确性.

Residual life prediction based on a stochastic effect Wiener degradation model

Since the products with high degradation rate don't conform to a stable degradation path, a new random effect degradation model namely Wiener degradation process model is established accounting for the influence of individual differences of products on the degradation process. Both drift parameter and diffusion parameter of the model are random variables, and the relationship between them is linear. Then, the residual life distribution and reliability function of the products are obtained based on the model, simultaneously, the EM algorithm to estimate parameters of the model is designed. Finally, the validity and accuracy of the model are verified by analyzing the fatigue crack data of a kind of titanium alloy and compared with the existing model.