摘要
针对复杂装备传统测试性建模和评价存在故障模式结构分层、信息不确定、数据容量小的问题,提出了分层混合贝叶斯网络测试性建模和评价方法。将测试性结构模型和测试性贝叶斯网络模型相结合,利用专家评分的先验知识与小样本测试数据的后验知识,在战术导弹全寿命周期内对装备测试性进行分析和评价。应用分层混合贝叶斯网络测试性建模方法建立复杂系统的测试性网络拓扑结构,解决了故障模式结构分层问题。引入三角模糊数算法获取专家知识,计算关于故障检测率的先验条件概率分布,解决了信息不确定问题。以历史测试数据作为后验条件,由贝叶斯公式更新故障检测率的概率分布,解决了数据容量小问题。仿真结果表明,与传统测试性建模和评价方法相比,分层混合贝叶斯网络测试性建模和评价方法既考虑了复杂装备的结构分层,又融合了先验知识和实际试验结果,并且实例验证所设计的测试性建模和评价方法是有效的。
Due to uncertain information,small data capacity,layered failure mode in testability modeling and evaluation of complex equipment,hierarchical hybrid testability modeling and evaluation based on Bayesian network for tactical missiles are proposed.Prior expert knowledge and posterior small sample test data have been used to analyse and evaluate testability in the life cycle of equipment,which combine testability structure model with testability Bayesian networks.The hierarchical hybrid Bayesian network testability modeling method is applied to establish the test network topology structure of complex systems,and the fault layer structure layering problem is solved.The triangular fuzzy number algorithm is introduced to obtain expert knowledge,and prior conditional probability distribution of fault detection rate is calculated,which solves the problem of information uncertainty.Using the historical test data as the posterior condition,the Bayesian formula updates the probability distribution of the fault detection rate,which solves the problem of small data capacity.The simulation results show that compared with the traditional test modeling and evaluation methods,the hierarchical hybrid Bayesian network testability modeling and evaluation method not only considers the structural stratification of complex equipment,but also combines prior knowledge and actual test results.The numerical simulation results are addressed to illustrate the performance of the obtained results.
作者
徐星光
廖志刚
任章
胡一繁
Xu Xingguang;Liao Zhigang;Ren Zhang;Hu Yifan(Beijing Electrical-mechnical Engineering Institute,Beijing 100074,China;School of Automation Science and Electrical Engineering,Beihang University,Beijing 100191,China;Bejing Aerospace Technology Institute,Beijing 100074,China)
出处
《战术导弹技术》
北大核心
2019年第6期1-8,31,共9页
Tactical Missile Technology
基金
国家自然科学基金(61503009,61333011)
关键词
分层混合贝叶斯网络
网络拓扑结构
全寿命周期
装备故障诊断
战术导弹
hierarchical hybrid Bayesian network
network topology
full life cycle
equipment fault diagnosis
tactical missile
