低温阀门无失效数据环节可靠性建模方法研究

A Reliability Modeling Method for Non-failure Data Link of Cryogenic Valve

针对无失效数据环节的可靠性评估时,采用单一模型难以同时得到参数的点估计和置信区间估计,而采用不同方法分别进行点估计和区间估计会造成结果的一致性问题,本文针对在无失效数据情形下,对低温阀门填料环节进行可靠性分析,提出了一种运用多层bayes方法得到产品可靠度的点估计,综合bootstrap抽样法从产品的寿命概率分布中重新抽取样本,与矩法结合得出低温阀门填料环节的可靠度区间估计的可靠性建模方法。该方法根据低温阀门实际运行工况,确定多层bayes方法中取值上限参数c的值,进而得到低温阀门填料环节的寿命概率分布曲线,再利用bootstrap法从寿命概率分布中重新抽取新样本,新样本采用矩法获得低温阀门填料环节可靠度的区间估计,并与置信限方法得到的区间估计进行对比。结果表明:在威布尔分布条件下,新模型提高了可靠度区间估计精度,为无失效数据环节可靠性评估提供了理论基础。

In evaluating the reliability of a non-failure data link,it is difficult to estimate the points of parameters and their confidence interval at the same time by using a single model,and different methods for point and interval estimation may cause the inconsistency of results.This paper carried out the reliability analysis of the packing link of a cryogenic valve under the condition of non-failure data.A point estimation method of product reliability was proposed with the multilayer Bayes method.Samples were extracted from product life probability distribution with the bootstrap sampling method.Combined with the moment method,a reliability modeling method for estimating the interval reliability of the packing link of a cryogenic valve is obtained.The method was based on actual operation conditions to determine the value of the range parameter c with the multilayer Bayes method.Thus,the service life of the cryogenic valve′s packing link probability distribution curve was extracted from the life probability distribution of new samples through reusing the bootstrap method.New samples of the cryogenic valve′s packing link reliability interval estimation were obtained with the moment method.The results are compared with the interval estimation by using the confidence limit method.They show that the new model improves the accuracy of reliability interval estimation under the Weibull distribution,providing a theoretical basis for the reliability evaluation of non-failure data links.