基于贝叶斯网络的机载氧气系统可靠性评估  

Reliability assessment of airborne oxygen system based on Bayesian network

作　　者：王鹏[1] 陈曦[2] 阎芳[1] 赵长啸[1] 

机构地区：[1]中国民航大学天津市民用航空器适航与维修重点实验室,天津300300 [2]中国民航大学中欧航空工程师学院,天津300300

出　　处：《中国民航大学学报》2016年第2期1-5,共5页Journal of Civil Aviation University of China

基　　金：国家自然科学基金项目(U1333120);国家重点基础研究发展计划(973计划)(2014CB744902)

摘　　要：通过分析机载氧气系统架构,构建基于贝叶斯网络的系统可靠性评估模型。通过前向和后向推理得到叶节点失效率及根节点后验失效概率,结合重要度及联合重要度的大小,找出影响系统可靠性的关键因素,求解最大可能解释(MPE)。通过与故障树分析(FTA)法对比,说明该方法的优势。综合各定量分析结果,说明贝叶斯网络方法进行可靠性评估的有效性,评估结果还可对系统结构设计和故障诊断提供相关参考信息。By analyzing airborne oxygen system architecture, the system reliability assessment model is built based on Bayesian network. The failure rate of leaf node and posteriori failure probability of root nodes are obtained by the forward and backward inferences, the key factors of system reliability are identified based on importance sequence and joint importance, the most probable explanation （MPE） is given. The advantages of the method are explained by comparing with FlA. Quantitative analysis proves the effectiveness of reliability assessment by using Bayesian network, and assessment results can also provide referencial information for the system structure design and fault diagnosis.

关 键 词：贝叶斯网络 机载氧气系统 定量评估 

分 类 号：V240.2[航空宇航科学与技术—飞行器设计]