模糊多态贝叶斯网络在冗余液压系统可靠性分析中的应用  

Application of fuzzy multi-state Bayesian networks in redundant hydraulic system′s reliability analysis

作　　者：方玉茹[1] 阚树林[1] 杨猛[1] 谢超[1] 裴学胜[1] 

机构地区：[1]上海大学机电工程与自动化学院,上海200072

出　　处：《计算机集成制造系统》2015年第7期1856-1864,共9页Computer Integrated Manufacturing Systems

摘　　要：为解决传统方法在状态多样化且精确故障率难以获得的复杂冗余系统的可靠性分析中的局限性,提出一种改进的基于模糊多态贝叶斯网络的可靠性分析方法。该方法基于二态故障树来构建多态贝叶斯网络结构,利用模糊数描述根节点故障率,结合工程经验构造出体现冗余系统及单元多态性和多态节点故障间不确定逻辑关系的条件概率表。通过理论推算得到系统故障概率与可靠度、后验概率计算及根节点重要度的分析过程,并将其应用到船舶推进器液压系统可靠性分析实例中,结果表明该方法能较好地满足冗余系统可靠性评估、已知当前状态时的系统故障预测和实施关键单元可靠性增长的实际需要,对类似系统可靠性研究具有参考价值。To solve the limitations of traditional methods in the reliability analysis of complex redundant systems with multiple states and accurate failure rate difficult to obtain, an improved reliability analysis method based on fuzzy multi-state Bayesian networks was proposed. Through two-state fault tree, multi-state Bayesian network structure was built, and fuzzy numbers were used to describe the failure rate of root nodes. The conditional probability tables reflected the polymorphism and uncertain logical relationship between nodes were constructed according to engineer- ing experience. The analysis process was elaborated to calculate the reliability, the posterior probability and the im- portance degree, and it was applied in the reliability analysis of hydraulic system in ship propellers. The results showed that the proposed method could meet the actual needs of reliability assessment, failure prediction knowing current state and reliability growth of redundant system, which had a reference value for the reliability research of similar systems.

关 键 词：模糊性 多状态系统 贝叶斯网络 复杂冗余系统 液压系统 推进器 可靠性分析 

分 类 号：TH137[机械工程—机械制造及自动化]