使用屏蔽数据时寿命的参数估计及含有保障延误时间的可修系统可靠性指标评定

【作者】 姜红燕；

【导师】 张帼奋；

【作者基本信息】 浙江大学 ， 概率论与数理统计， 2006， 硕士

【摘要】 本文是在攻读硕士学位期间完成的，全文共分五章： 第一章为引言。 第二章考虑一个由独立、不完全相同的单元组成的串联系统，使用屏蔽数据时的参数估计。当失效率为指数函数，假定未知参数的先验分布为对称的三角分布时给出了参数的极大似然估计和Bayes估计；第三章考虑的是在一个由两个相依单元组成的串联系统中，使用屏蔽数据时的参数估计。当单元的联合生存函数服从二维指数分布时，我们给出了参数的极大似然估计和Bayes估计。在Bayes估计中，我们假定未知参数的先验分布为伽马分布，两章都通过数据模拟对极大似然估计和Bayes估计这两种方法进行了比较。 第四章考虑两个相同部件组成的并联系统可修系统在出现故障时，由于种种原因，故障不能得到及时维修，形成了一段保障延误时间。在假设工作时间、保障延误时间、故障维修时间均服从指数分布的条件下，我们利用齐次马尔可夫过程和拉普拉斯变换，得到了系统的首次故障前时间、可用度和平均故障次数等一系列可靠性指标；第五章则是在看到拉普拉斯变换及其反演的复杂性之后，转而利用傅立叶变换给出了工作时间、保障延误时间、故障维修时间均服从指数分布的单部件系统的瞬时可用度。更多还原

【Abstract】 This thesis is finished during my Master of science and it consists of five chapters.Chapter I is preface.Chapter II is on the estimation of the parameters included in the lifetime distribution of the individual components in a series system with independent and nonidentical components by using masked data.The maximum likelihood and Bayes estimates of these parameters are deduced when the failure rates of the system components are assumed to be exponential with different parameters.In Bayes approach, it is assumed that the unknown parameters have symmetrical triangular prior distributions. In chapter III, estimation of the parameters included in the lifetime distribution of the individual components in a series system with two dependent components are introduced by using masked system life data. We deduce the maximum likelihood and Bayes estimates of these parameters when the joint survival function obeys the two dimentions exponential distribution,while in Bayes approach, it is assumed that the unknown parameters have Gammar prior distributions. And numerical simulation studies are both introduced to discuss a comparison in the two chapters above.The forth chapter consideres a parallel repairable system with two identical com-ponents,because of various reasons,the system can’t get repaired in time when it fails,thereby a period of morra exists. Under the assumption that the working,morra,and repair time both obeys the exponential distribution,some important system reliability indices such as the system availability,MTTFF and average failure time etc are deduced via Markov process and laplace transform.The system instantaneous availability is considered in the last chapter using fourier transform when the working,morra,and repair time both obeys the exponential distribution in a single component system.更多还原