论文部分内容阅读
针对一类不能获得足够数量的试验和现场使用数据的复杂系统,提出Monte Carlo(MC)实时可靠性估计方法.以系统部件失效时间序列作为样本空间,应用MC法仿真产生更多的虚拟样本,建立了虚拟样本序列下系统实时可靠性估计的一般框架,并研究了MC估计的期望和方差与先验信息之间的关系,得到了具有包含关系的信息条件下,基于较少先验信息的估计的方差不超过拥有更多先验信息的估计的方差.考虑所有部件失效时间序列和导致系统失效的部分部件失效时间序列的先验信息,分别导出了系统实时可靠性模型.最后,通过算例分析,验证了本文不同先验信息下可靠性模型的有效性和方法的合理性.
Aiming at a kind of complex system which can not obtain a sufficient number of test and field data, a Monte Carlo (MC) real-time reliability estimation method is proposed. With the system component failure time series as the sample space, MC method is used to generate more virtual samples, The general framework of real-time reliability estimation of system under virtual sample sequence is established, and the relationship between expectation and variance of the MC estimation and prior information is studied. Under the condition of information with inclusive relation, based on less prior information The estimated variance does not exceed the estimated variance with more prior information.According to all the component failure time series and the a priori information of the failure time series of the partial components that cause the system failure, the real-time reliability models of the system are derived respectively. An example is given to verify the validity of the reliability model under different prior information and the rationality of the method.