随机效应Wiener过程退化可靠性分析方法

Reliability analysis based on a random-effect Wiener process degradation model

针对工程中很多产品的初始退化量具有不确定性的情况,提出一种性能退化随机效应Wiener过程可靠性分析方法,同时考虑初始退化量和漂移系数的随机性.然后,根据首达时的定义,推导出了失效时间概率分布函数和概率密度函数的解析表达式.其次,结合产品性能退化测试情况,给出退化模型未知参数的极大似然估计.最后,通过对磁盘性能退化可靠性实例分析,验证了本文方法的适用性和有效性,与随机效应退化轨迹模型相比,本文方法建模过程更为合理,且评估精度高.

In engineering practice, the initial performance degradation value of many products shows indeterminacy. In such a situation, a random-effect Wiener process degradation model, which simultaneously considers the initial degradation value and the drift coefficient as random variables, is proposed for reliability analysis. Then, on the concept of the first hitting time, the analytic expressions of the probability distribution function and the probability density function are derived. Moreover, maximum likelihood estimations of the parameters in the proposed model are constructed. Finally, an illustrate example involving magneto-optical data storage disk is adopted to illustrate the reasonability and efficiency of the proposed model. Meanwhile, for comparison, the path degradation model with random effects is considered as the reference method, and the comparative results demonstrate that the estimation accuracy of the proposed model is higher and the modeling procedure is more reasonable.