屏蔽数据并联系统的可靠性分析  

Reliability analysis for parallel system with masked data

作　　者：师小琳 师义民[2] 段俊杰 SHI Xiaolin;SHI Yimin;DUAN Junjie(chool of Electronic Engineering,Xi′an University of Posts and Telecommunications,Xi′an 710121,China;Department of Applied Mathematics,Northwestern Polytechnical University,Xi′an 710072,China;The Graduate School of Northwest University,Xi′an 710069,China)

机构地区：[1]西安邮电大学电子工程学院,陕西西安710121 [2]西北工业大学应用数学系,陕西西安710072 [3]西北大学研究生院,陕西西安710069

出　　处：《西北大学学报（自然科学版）》2018年第2期161-167,共7页Journal of Northwest University（Natural Science Edition）

基　　金：国家自然科学基金资助项目(71401134;71571144);国际科技合作与交流计划基金资助项目(2016KW-033)

摘　　要：在逐步Ⅱ型截尾模型下,研究含屏蔽数据三部件并联系统的可靠性分析。假设组成系统的部件相互独立且服从补充指数分布,利用屏蔽系统寿命数据给出了部件参数与可靠度函数的极大似然估计。鉴于极大似然法在完全屏蔽情形的局限性,通过引入潜在变量并运用Metropolis-Hastings抽样算法获得了部件参数及可靠度函数的贝叶斯估计、参数的最大后验密度置信区间。最后利用Monte-Carlo方法给出数值例子,分析了屏蔽水平对两种估计精度的影响,并验证了文中结果的正确性。Under progressively Type-II censored model,the reliability of the three-component parallel system with masked data is studied.Assuming that the components of the system are independent of each other,and the lifetime of the components follows the supplemental exponential distribution,the maximum likelihood estimation of the component parameters and the reliability function is obtained by using the masked system lifetime data.In view of the limitations of the MLE method in the case of complete masking,the Bayesian estimation of the component parameters and the reliability function are derived by introducing the potential variables and using the Metropolis-Hastings sampling algorithm.Also,the highest posterior density confidence interval of the component parameters is deduced.Finally,numerical simulation demonstrates the correctness of theoretical results by using Monte-Carlo method.In the meantime,we study the influence of masking level on the accuracy of MLE and Bayes estimations.

关 键 词：可靠性分析 屏蔽数据 并联系统 逐步Ⅱ型截尾 点估计 置信区间 

分 类 号：O213.2[理学—概率论与数理统计]