含有屏蔽寿命数据的贝叶斯可靠性分析

Bayesian Reliability Analysis for Masked Lifetime Data

研究了含有屏蔽寿命数据的可靠性评估问题 .在许多系统中 ,能够获取的可靠性数据往往是系统的寿命数据 ,而引起系统失效的单元却是未知的 ,此时如何对单元的可靠性进行评估是一个亟待需要解决的问题 .首先分析了屏蔽寿命数据的似然函数 ,并应用贝叶斯方法得到了单元可靠性分布参数和屏蔽概率的验后分布 ,由于计算上的复杂性 ,采用两种数值算法———Gibbs抽样和EM算法 ,得到了单元可靠性分布参数和屏蔽概率的验后估计 .

Reliability assessment for masked lifetime data is studied. In many systems, the available reliability data are often lifetime data of the system, whereas the exact component that causes the systems failure is unknown, how to analysis the components reliability is an urgent problem to solve. the likelihood of masked lifetime data is analyzed, and the posterior distribution of distribution parameters of component reliability and masking probabilities is obtained adopting Bayesian approach. Becaues of complexity of computation, two numerical algorthms as Gibbs sampling and EM algorithm are adopted to obtain posterior estimates of distribution parameters of component reliability and masking probabilities. The simulation results show that the method proposed is reasonable.