电磁继电器长贮零失效情况下的Bayes可靠性评估方法

Bayes reliability evaluation method for long storage zero failure of electromagnetic relay

电磁继电器作为一种高可靠、长寿命产品,在长期贮存过程中经常会遇到零失效的情况.文中采用Bayes方法结合性能退化数据来解决电磁继电器无失效数据的可靠性评估问题,利用K-S检验法得到电磁继电器的贮存寿命模型服从威布尔分布,通过Bootstrap抽样和MCMC方法分别对先验分布和后验分布进行求解.结果表明,Bootstrap抽样确定未知参数的先验分布相比于传统方法更精确,MCMC方法则解决了后验分布求解复杂的问题,并通过实例验证了Bayes方法在零失效情况下进行可靠性评估的可行性.

As a product of high reliability and long life,the electromagnetic relay often meets zero failure during the long storage process.In view of the problem,in this paper,Bayes method combined with performance degradation data is used to solve the reliability evaluation of electromagnetic relay without failure data.The storage life model of electromagnetic relay is subject to Weibull distribution by using K-S test method.Prior distribution and posterior distribution are solved by Bootstrap sampling and MCMC method respectively.The results show that the prior distribution of unknown parameters determined by Bootstrap sampling is more accurate than traditional methods.The MCMC method solves the complex problem of posterior distribution.The feasibility of Bayes method for reliability evaluation under zero failure is verified by an example.