中文摘要：

针对产品研制过程存在多阶段的试验信息和产品性能参数呈现动态变化的特点,提出一种Bayes框架下成败型系统可靠性的动态评估方法.该方法依据产品研制过程中不同研制阶段的试验数据,利用离散AMSAA可靠性增长模型对产品可靠性的变化趋势进行描述,进而对新批次的产品可靠性进行预测,并由最大熵方法给出产品可靠性验前分布的确定方法,再结合产品少量的现场试验数据,对新批次的产品可靠性做出Bayes评估.最后通过实例对比分析,表明该方法很好地解决了小子样产品的可靠性试验评估问题.

英文摘要：

Considering that there are multiple-stages field data and product performance parameters take on dynamic change trend during the process of product development,a dynamic Bayesian evaluation method for binomial system reliability is proposed in this paper.The approach uses the discrete AMSAA model to describe the change trend of product reliability according to the existent multiple-stages field data, and make prediction for product reliability at new-stage.Then,the prior distribution of product reliability at new-stage is given by referring to the Maximum Entropy Method.Along with the small quantity of field data,the Bayesian evaluation of product reliability is made by the proposed method.Numerical example indicates that the approach can properly solve the problem of product reliability evaluation under the condition of small sample size.