摘要
通过传统可靠性方法得到的失效分布反映了相同受试环境下同一类产品的平均特征,并不能直接应用于系统实时定量可靠性分析。为此,针对有历史试验数据的退化失效型产品,依据局部更新的思想,利用贝叶斯方法最大可能地采用现场测试数据,并将之及时反映到实时可靠性模型中去。结果表明,随着融合现场数据的增多,产品的实时可靠性模型也逐渐全面反映出产品的个性特征。
The failure distribution from traditional reliability methods reflects the ＂average＂ characteristics of the same kind of product. In order to get the real time reliability of degradation failure product with history test data, this paper adopts the local updating ideas, uses the on-site measuring data through Bayes method and gets the real time reliability model of the product that timely reflects the on-site data, The result shows that real-time reliability model expresses the specific characters gradually with the increase of fusing onsite data.
出处
《国防科技大学学报》
EI
CAS
CSCD
北大核心
2007年第6期116-120,共5页
Journal of National University of Defense Technology
基金
国家部委基金资助项目(2005AA845023)
关键词
贝叶斯方法
实时可靠性
先验分布
性能参数
Bayes method
real time reliability
prior distribution
performance parameter
