基于Bootstrap的可修系统贮存可用度近似置信下限评估方法

针对整机贮存期内的区间型定检数据以及维修方案,提出了可修系统贮存可用度近似置信下限的Bootstrap评估方法。使用基于定时截尾数据的极大似然估计(MLE)方法计算参数的伪MLE值,并利用修正的似然函数法对贮存阶段的无失效数据进行了统计处理。结合Bootstrap方法,引入修如旧和修如新模型计算系统贮存可用度。计算结果表明:该方法得到的置信下限真值覆盖率接近置信水平,其评估精度满足工程的应用需求。

Estimating Method of Approximate Confidence Lower Limits of Repairable System Storage Availability Based on Bootstrap

Taking aim at the grouped data and maintenance program of the equipments in the storage period, the Bootstrap estimating method of approximate confidence lower limits of repairable system storage availability was proposed. The pseudo maximum likelihood estimate (MLE) value was calculated by MLE method based on type Ⅰ censored data, and the zero-failure data was solved by MMLE method. Storage availability of the system was calculated by the Bootstrap method and the introduced models in which the repaired systems seem as the old one or the new one. The result of the illustration computation shows that the truth coverage ratio of confidence lower limits obtained by the method approaches to the confidence level, and the estimate precision is satisfied with the demand of engineering application.