多元参数退化的系统相关竞争失效可靠性评估方法  

A Reliability Evaluation Method for System′s Dependent Competition Failure and Multi-Parameter Degradation Failure

作　　者：胡启国 高展 HU Qiguo;GAO Zhan(Department of Mechatronics and Vehicle Engineering, Chongqing Jiaotong University, Chongqing 400074, China)

机构地区：[1]重庆交通大学机电与车辆工程学院

出　　处：《西北工业大学学报》2019年第6期1191-1199,共9页Journal of Northwestern Polytechnical University

基　　金：国家自然科学基金(51375519);重庆市基础科学与研究专项(cstc2015jcyjBX0133)资助

摘　　要：为了解决系统相关性竞争下的可靠性问题,提出一种突发失效与多元参数退化失效相关性竞争失效的可靠性评估方法。以Wiener随机过程与Inverse Guassian随机过程描述系统多元参数退化失效过程,利用Copula函数对系统多元退化失效过程进行建模,通过二阶段极大似然法估计退化失效相关参数;构建突发失效关于退化量的条件概率,利用Bayes-Bootstrap法对极大似然估计所得的突发失效参数进行修偏,进一步建立系统的相关性竞争失效模型。以航空发动机的退化数据为例进行综合分析,研究其突发失效和多退化失效之间竞争下的可靠性问题。获得的分析结果能够反映出航空发动机的性能可靠性,验证了模型的有效性,具有较好的工程应用价值。In order to enhance the reliability of a system that has dependent competition failure,a reliability evaluation method is proposed to evaluate the dependent competition failure and multi-parameter degradation failure.The multi-parameter degradation failure process is described with the Wiener stochastic process and the inverse Gaussian stochastic process.The Copula function is used to model the system′s multi-degradation failure process.The two-stage maximum likelihood method is used to estimate the degradation failure parameters.The conditional probability of dependent competition failure in terms of degradation degree is established.The Bayes-Bootstrap method is utilized to correct the dependent competition failure parameters obtained with the maximum likelihood method and to further establish the system′s dependent competition failure model.The degradation data of an aero-engine is used as an example to analyze the reliability under competition between dependent competition failure and multi-parameter degradation failure.The analysis results can effectively demonstrate the reliability of an aero-engine′s performance and verify the validity of the model,thus having good engineering application values.

关 键 词：多元参数退化 竞争失效 COPULA函数 两阶段极大似然法 Bayes-Bootstrap法 

分 类 号：TB114.3[理学—概率论与数理统计]