指数寿命系统可靠性增长的动态Bayes评估方法  

Dynamic Bayesian evaluation method of reliability growth for exponential life system

作　　者：邢云燕[1] 武小悦[1] 刘琦 

机构地区：[1]国防科技大学信息系统与管理学院,湖南长沙410073

出　　处：《系统工程学报》2010年第5期712-716,共5页Journal of Systems Engineering

基　　金：国家自然科学基金资助项目(70571083)

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框架下指数寿命系统失效率的动态Bayes评估方法.该方法依据产品研制过程中不同阶段的试验数据,利用AMSAA模型对系统失效率的变化趋势进行描述,进而对新批次的产品失效率进行预测,并由最大熵方法给出产品失效率验前分布的确定方法,再结合产品少量的现场试验数据,对新批次的产品失效率做出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指数寿命产品的可靠性试验评估问题.Poingting at that there exist multiple-stages field data and product performance parameters have dynamic change characeristics during the process of product development,a dynamic Bayesian evaluation method of failure rate for exponential life system is proposed.The method uses AMSAA model to describe the change trend of product failure rate in terms of the existent multiple-stages field data,and makes prediction for product failure rate at new-stage.Then,the prior distribution of product failure rate at new-stage is given by using the maximum entropy method.And with small quantity of field data,the Bayesian evaluation of product failure rate is made.Numerical example indicates that the method can properly solve the problem of product reliability evaluation for exponential life system under the condition of small sample size.

关 键 词：可靠性增长 BAYES方法 AMSAA模型 指数寿命 动态评估 

分 类 号：TB114.3[理学—概率论与数理统计]