基于无失效数据的加权E-Bayes可靠性评估方法  

Method for evaluation of weight expected-Bayesian reliability based on zero-failure data

作　　者：蔡忠义[1] 陈云翔[1] 项华春[1] 董骁雄 

机构地区：[1]空军工程大学装备管理与安全工程学院,陕西西安710051

出　　处：《系统工程与电子技术》2015年第1期219-223,共5页Systems Engineering and Electronics

基　　金：总装"十二五"国防预先研究项目资助课题

摘　　要：针对无失效数据情况下产品的可靠性评估问题,提出加权最小二乘法结合期望Bayes(expectedBayesian,E-Bayes)可靠性评估方法与模型。根据工程经验,构造出产品失效概率的先验分布;运用Bayes理论与方法,给出产品失效概率的E-Bayes估计;在威布尔分布场合下,采用加权最小二乘法来拟合产品寿命分布参数,给出产品可靠性指标的点估计和区间估计;结合算例分析,探讨模型的稳健性与方法的优越性,表明该方法具有较好的工程应用价值。Aiming at the problem of product reliability evaluation in condition of zero-failure data,a method combining the weight-least-square method with expected-Bayesian （E-Bayes）reliability evaluation is put for-ward.According to the engineering experience,the prior distribution of product failure probability is construc-ted.Then the E-Bayes estimation of product failure probability is given by using the theory and method of Bayesian.In condition of Weibull distribution,the weight-least-square method is used to seek out the parameter of Weibull distribution,and obtain the reliability indexes’point and interval estimation.At last the robustness of the model and the superiority of the method are discussed by an example which shows the value of the method in engineering application.

关 键 词：可靠性评估 无失效数据 先验分布 失效概率 期望Bayes 

分 类 号：TB114[理学—概率论与数理统计]