可靠性增长的单调约束模型

大型复杂产品研制过程中每一阶段的试验数据通常具有小样本特征.对此在产品可靠性评估中,提出使用约束条件,即"下一阶段可靠度真值总是大于前一阶段",综合利用产品多阶段试验数据的方法.产品研制过程中可靠性通常是增长的,因此这种方法具有广泛的适用性.在上述约束条件下,可靠度被看作一个随机变量.首先推导出了可靠度在此条件下的分布函数表达式,然后给出了一种近似计算方法,最后通过一个实例与仅使用单阶段试验数据的评估结果作了对比.由于该方法的约束条件简单且可以求得可靠度的分布函数,因此其工程适应性强于传统的AMSAA(Army Materiel System Analysis Activity)模型,并能获得Duane模型无法给出的可靠度区间估计.

Monotone restriction model of reliability growth evaluation

The test data of large complex products presents small sample characteristic in each phase during the development process normally. The restriction, the reliability of one phase is always higher than that of the previous one, was recommended to be used for reliability evaluation based on multi-phase test data. The reliability of product is increasing during the development process normally, therefore, such method has broad applicability. The reliability was considered a random variable with such restriction. The reliability probability distribution function based on the restriction was given, and an approximative calculational method was presented. An example was given to compare the result obtained from single phase test data with that based on multi-phase test data. Since the precondition is brief and the reliability probability distribution function can be achieved, the mentioned method is more applicable to practical applications in comparison with traditional army materiel system analysis activity(AMSAA)-model, and it can gain the reliability interval estimation which can not be obtained from Duane-model.