摘要
针对航天产品样本小、造价昂贵、寿命长和可靠性高等特点,提出了一种基于无失效数据来推导样本寿命的分析方法。首先,根据现有理论和试验基础,通过对环形腔He-Ne激光器的寿命特性、失效模型和失效机理进行研究,得出了对环形激光器采取加速寿命试验的结论;然后,利用试验得到的恒应力双加速寿命试验无失效数据,使用减函数法构造了特定应力条件下的多层Bayes估计,同时结合样本信息,利用Bayes方法得到了各个应力下的寿命的Bayes估计;最后,得到了激光器正常工作条件下的特征寿命。该方法将先验信息与样本信息相结合,只需通过一组特定条件下的加速应力的试验数据,就可得到激光器正常工作条件下的特征寿命,具有重要的工程意义和实用价值。
In view of the characteristics of small sample,high cost,long life and high reliability of aerospace products,an analysis method based on non-failure data to derive sample life is proposed.Firstly,according to the existing theory and experimental basis,the conclusion is drawn that the ring laser is subjected to an accelerated life test by studying the life characteristics,failure model and failure mechanism of ring cavity He-Ne laser.Then,the multi-layer Bayes estimation under specific stress condition is constructed by using the subtraction function method based on the non-failure data obtained from constant stress double accelerated life test.At the same time,the Bayes estimation of the life under various stresses is obtained by using the Bayes method combined with the sample information.Finally,the characteristic life of the laser under normal working conditions is obtained.In this method,the prior information and sample information are combined,and the characteristic life of laser under normal condition can be obtained only through a set of test data of accelerated stress under certain condition,so it has great engineering significance and practical value.
作者
李锴
邢媛
唐庆云
高春雨
LI Kai;XING Yuan;TANG Qingyun;GAO Chunyu(CEPREI,Guangzhou510610,China;Guangdong Provincial Key Laboratory of Electronic Information Products Reliability Technology,Guangzhou510610,China;Guangdong Industrial Robot Reliability Engineering Laboratory,Guangzhou510610,China)
出处
《电子产品可靠性与环境试验》
2019年第3期8-14,共7页
Electronic Product Reliability and Environmental Testing
基金
工业装备环境可靠性设计与试验创新中心能力建设项目(2017KZ010107)
广州市产业技术重大攻关计划现代产业技术专题(201802010051)资助
关键词
激光陀螺
恒应力加速寿命试验
先验分布
多层贝叶斯估计
laser gyro
constant stress accelerated life test
prior distribution
multilayer Bayes estimation