基于最后逃逸时间的随机退化设备寿命预测方法

现有基于随机退化过程建模的寿命预测研究中, 通常用退化过程的首达时间(First passage time, FPT)来定义寿命. 但是, 这种寿命定义较为保守, 可能会导致其明显小于设备实际寿命. 鉴于此, 基于最后逃逸时间(Last exit time, LET)的概念, 给出一种新的寿命与剩余寿命(Remaining useful life, RUL)定义方式. 在该新框架下, 提出一种基于最后逃逸时间的寿命预测方法, 推导得到最后逃逸时间下基于Wiener退化过程模型的寿命与剩余寿命表达形式, 讨论了该方法与传统首达时间下寿命预测方法之间的关系. 此外, 通过数值仿真验证了该方法的正确性, 并对模型参数进行了敏感性分析. 最后, 通过轴承以及激光器的实际退化数据说明了该方法的有效性、可行性以及潜在的工程应用价值.

Lifetime Prediction for Stochastic Deteriorating Systems Based on the Last Exit Time

In existing studies of lifetime prediction based on stochastic process models, first passage time (FPT) is often used to determine the lifetime. However, this definition is generally conservative, and it may be obviously shorter than the actual lifetime. This paper provides a novel definition of lifetime and remaining useful life (RUL) based on the concept of the last exit time (LET). Under this framework, this paper proposes a lifetime prediction method based on the LET, the lifetime and RUL distributions under the concept of the LET have been derived based on the Wiener-process-based degradation model, and then the difference between the LET-based method and the FPT-based method is discussed carefully. Additionally, the proposed method is verified by numerical simulations, and the sensitive analysis is also completed. Finally, the practical degradation data of bearing and laser device are utilized to illustrate the validity, feasibility, and potential engineering value of the proposed method.