基于信息融合的诊断贝叶斯网络研究  

Diagnostic Bayesian Networks Research Based on Information Fusion

作　　者：宫义山[1] 高媛媛[1] 

机构地区：[1]沈阳工业大学信息科学与工程学院,辽宁沈阳110178

出　　处：《计算机技术与发展》2009年第6期106-108,共3页Computer Technology and Development

基　　金：辽宁省博士启动基金资助项目(521101302)

摘　　要：针对设备故障诊断技术中存在的固有不确定性问题,通过分析传统故障树模型存在的局限性以及传统贝叶斯网络建造困难的特性,提出了一种融合于故障树和传统贝叶斯网络的新方法——诊断贝叶斯网络,并阐述了故障树和贝叶斯网络的故障诊断策略优化方法的基本思想和具体算法。通过比较分析,综合考虑了故障树和贝叶斯网络在诊断推理和模型表达方面的特点及仿真结果,提出的新方法可以使二者优势充分发挥,在故障诊断领域中具有实际的应用价值。To the intrinsical uncertainty of fault diagnosis, the limitation of traditional method of fault tree and the difficulty to construct traditional bayesian networks are analyzed, a new method - diagnostic Bayesian networks which combines fault tree and Bayesian networks is proposed. It also expatiated the philosophy and algorithm to fault diagnosis strategy optimization method of fault tree and Bayesian networks. Through the comparative analysis and the integrative consider the characteristic and simulation result of fault tree and Bayesian networks in the diagnosis inference and the model expression. The new method could enable the superiority of them to display fully, it has the actual application value in the fault diagnosis domain.

关 键 词：故障诊断 不确定性 故障树 贝叶斯网络 诊断推理 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]