应急电源系统平均寿命Bayes估计方法研究

Bayesian estimation for mean life of emergency power system

提出了一种元器件应力分析和综合E-Bayes理论相结合的平均寿命预估方法,该方法基于定时截尾试验获取的无失效数据,利用元器件应力研究结果作为失效数据输入量,使得综合E-Bayes两个输入相互解耦,解决了失效数据对无失效数据的直接依赖以及与产品性能关联性弱的问题。本研究将该方法用于C919飞机应急电源系统,推导了系统平均寿命的Bayes预计模型,同时将预估结果与传统综合E-Bayes法进行比较。分析表明,使用本研究提出的综合E-Bayes算法得到的系统平均寿命预估值极差范围更小,对参数选择的稳健性更好。

A tool combining with components stress analysis and synthetic E-Bayes theory is proposed to deal with the mean life of complex aeronautical electronics with the intention to handle the nondecoupliug for the inputs of traditional synthetic E-Bayes. It is done with the truncation time of terminated testing as zero-failure data and the result of the components stress analysis as the failure data which is aimed to solve the direct dependence on zero-failure data and little relationship with practical performance for failure data. The computational formula of the mean life is deduced when the tool is used in the emergency power system of C919 airplane. To research the robustness, thistool is compared with the traditional synthetic expected Bayesian. As a result, the foregoing shows a better robustness.