固体火箭发动机加延时间试验的可靠性增长Bayes分析  

Bayesian analysis on reliability-growth of solid rocket motor based on extending time test

作　　者：刘飞[1] 窦毅芳[1] 张为华 

机构地区：[1]国防科技大学航天与材料工程学院,长沙410073

出　　处：《固体火箭技术》2007年第3期201-204,237,共5页Journal of Solid Rocket Technology

基　　金：国家"863"计划(2003AA765030)

摘　　要：针对含有加延时间试验的固体火箭发动机研制过程,将Mazzuchi-Soyer可靠性增长模型用于此类研制试验的可靠性分析,扩展了模型应用范围。首先,采用威布尔分布描述发动机可靠性随试车时间变化,经过一定的数学变换后,采用Mazzuchi-Soyer可靠性增长模型融合了先验信息和各阶段试验信息,给出了研制试验结束时产品可靠性的联合后验分布;然后,利用Gibbs抽样算法解决了后验推断计算问题,最终得到了各阶段产品可靠性的Bayes点估计和区间估计;最后,给出了产品可靠性增长分析实例,表明了模型的优越性。Aiming at the development of solid rocket motor （SRM） with the extending time test, the Mazzuchi-Soyer reliability growth model is applied to analyzing the reliability of the development test, which extends application range of the model. Firstly, the change of motor reliability with test time was described by using Weibull distribution. Mter certain mathematical transform, Mazzu- chi-Soyer reliability growth model was applied to combining priori information and experimental information at all the stages, and the combined posterior distribution of motor reliability at the end of development test was given ; then posterior deduction problems were solved by means of the Gibbs sampling algorithm. Bayesian point estimation and Bayesian zone estimation of reliability for produces at all stages were obtained. Finally an analysis example on reliability growth verified superiority of the Bayesian model.

关 键 词：固体火箭发动机 加延时间试验 威布尔分布 可靠性增长 BAYES估计 

分 类 号：V435[航空宇航科学与技术—航空宇航推进理论与工程]