基于多阶段-随机维纳退化过程的产品剩余寿命预测方法  

Prediction of Residual Life of Products Based on Multi-stage Stochastic Wiener Degradation Process

作　　者：魏高乐[1] 陈志军[2] 

机构地区：[1]空军工程大学装备管理与安全工程学院,西安710043 [2]北京航空航天大学可靠性与系统工程学院,北京100191

出　　处：《科学技术与工程》2015年第26期27-34,共8页Science Technology and Engineering

摘　　要：针对退化失效型产品的剩余寿命预测是开展其视情维修的关键前提。考虑到一些产品性能退化过程具有多阶段非单调的特点,提出基于多阶段-随机维纳退化过程的产品剩余寿命预测方法。针对产品退化过程的每一阶段,主要考虑以高应力加速退化试验数据为先验信息,不预先给予Wiener过程参数的假定分布,而利用加速因子将高应力下的参数估计值折算到正常应力下的参数值;再通过使用Anderson-Darling方法确定参数的最优分布,作为参数的先验分布;然后通过Bayes方法利用产品正常应力下实时退化数据对参数先验分布进行更新,得到参数后验分布,进而对产品的剩余寿命进行预测;最后以某型加速度计为实例验证了所提出的方法适用性和有效性。The degradation products of residual life prediction is canied out as a crucial prerequisite for condi- tion based maintenance （CBM）, to take into account some performance degradation process with muhi-stage non monotonic features, put forward the residual life prediction method based on multi stage random Wiener degradation process. For each stage of product degradation process, the main consideration to high stress accelerated degrada- tion test data as the prior information is not given the assumed distribution of Wiener process parameters, and the use of acceleration factor for the parameters estimated value under high stress converted to parameters values for un- der normal stress , and determining parameters optimal distribution by using Anderson-Darling method to as the pri- or distribution of the parameters, and then through the Bayes method, use real-time degradation data of the product under normal stress to update the prior distribution of parameters, get the parameter posterior distribution, to pre- dict the residual life of the product; Finally, taking a case study of the accelerometer, the suitable and validity of the proposed method has been proved.

关 键 词：多阶段 WIENER过程 加速因子 Anderson-Darling方法 先验分布 后验分布 剩余寿命 

分 类 号：TB114.3[理学—概率论与数理统计]