基于Bayes理论的小子样维修性验证方法研究

Research on Small Sample Maintainability Demonstration Method Based on Bayes Theory

针对现有经典维修性指标验证方法的不足,提出了一种小子样的维修性验证方法,即采用Bayes理论对维修性指标进行验证。给出了应用该方法的一般步骤,探讨了维修时间总体分布的确定方法,采用随机加权法对验前分布进行确定并利用ML-II方法实现了多源信息的融合;进一步分析推导了基于Bayes理论的维修性验证方法的检验判据和样本量的大小。最后通过实例分析表明,该方法通过综合利用各种验前信息,可以有效地降低现场试验样本量,节省试验投入,缩短试验周期,且能保证较高的可信度。

Aiming at the deficiency in classics maintainability demonstration,a small sample maintainability demonstration method was brought forward,that's to adopt Bayes theory to carry on demonstration for maintainability index.The common approaches to apply the method were presented,firstly the population distribution of maintenance time was discussed,secondly the prior distribution was determined by using random weighted method and the information fusion of multiple sources was realized by using ML-II method,again the test criterion and sample size of this small sample maintainability demonstration method were analyzed and computed based on Bayes theory of hypothesis test.In the end,an example was analyzed to prove that by means of the comprehensive utilization all kinds of prior information,this method could effectively reduce test sample size,save test costs,shorten the development cycle and could ensure sufficient reliability.