基于屏蔽数据多重定数截尾下系统部件的可靠性估计  

Reliability estimation of components in system under multiple type-II censoring samples using masked data

作　　者：张萌[1] 陆山[1] 杨扬[2] 

机构地区：[1]西北工业大学动力与能源学院,陕西西安710072 [2]西北工业大学航空学院,陕西西安710072

出　　处：《系统工程与电子技术》2013年第5期1122-1127,共6页Systems Engineering and Electronics

基　　金：国家自然科学基金(70471057;71171164);陕西省教育厅自然科学专项基金(03JK063)资助课题

摘　　要：利用屏蔽的系统寿命数据,在多重定数截尾样本下研究串联系统中部件的可靠性估计问题。首先利用概率元分析方法,推导样本似然函数。然后基于样本似然函数,给出部件参数和可靠度函数的极大似然估计,同时分别在平方损失、线性指数损失和广义熵损失下,得到了部件参数和可靠度函数的Bayes估计。最后通过数值仿真研究屏蔽水平和截尾因子对估计精度的影响,并对各种估计进行了比较。The reliability estimation of the components in series system is investigated using masked system life data under multiple type-II censoring samples. At first, the sample likelihood function is deduced using the probit analysis method. Then based on the sample likelihood function, maximum likelihood estimations of corn- ponent parameter and reliability function are obtained. and realiability function are given under squared error, Meanwhile, Bayes estimations of component parameter linear exponential and general entropy loss function, re- spectively. Finally, numerical simulation is introduced to study the influence of masking level and censoring fac tot on the accuracy of the estimations and compare these estimations.

关 键 词：屏蔽数据 多重定数截尾 概率元 损失函数 极大似然估计 BAYES估计 

分 类 号：O213.2[理学—概率论与数理统计]