基于不完备模糊规则库的信息融合故障诊断方法.PDFVIP

4 3 V ol . 4 3 N o . S 20 11 7 Journal of N anjing U niversity of A ero naut ics A st ro nautics Jul . 2011 1, 2 2 3 徐晓滨 　吉吟东　文成林 1. , , 100084; 2. , , 1000 84; 3. , , 3100 18) : 在基于模糊推理的故障诊断专家系统中, 规则库的不完备常常会引起系统的某些输入 生的输出不确定。 针对此问题, 将模糊推理与证据理论相结合提出了一种新的融合诊断方法。首先根据证据的随机集表示及随机 集扩展准则将模糊输入映射到输出, 并聚合生成故障命题的基本概率赋值 BPA ) , 以度量由规则库不完备和输 入模糊性引起的输出不确定。然后利用Demp st er 组合规则将多 个不完备规则库提供的BP A 融合, 用融合结果判 决故障, 并以电机转子故障诊断为例, 验证了所提方法可以有效地提高故障确诊率。 : 故障诊断; 模糊推理; 信息融合; 证据理论; 随机集理论 : T P 18; T P 206: A : 1005-2615 20 11) S-0055-05 Information Fusion Method for Fault Diagnosis Based on Incomplete Fuzzy Rule Bases 1, 2 2 3 X u X iaobin , Ji Yindong , Wen Chenglin 1. Depart ment of A utoma tio n, T sing hua U niver sity , Beijing , 100084, China; 2. Research Institute o f Infor mation T echnolog y, T singhua U niv ersity, Beijing , 100084, China; 3 . Schoo l o f A ut omation , Hang zho u D ianzi U niver sity , Hangzhou , 3100 18, China) : - , , Abstract In t he fuzzy reasoning ex pert syst em of fault diagnosis t he f uzzy rule base is of t en incom plet e t hu s leading t o uncer tain out put s for som e possible input s. A f u sion fault diagnosis algo rithm is present - ed based on t he f uzzy rule reasoning and the ev idence t heory . Firstly , u sing t he random set fo rmalism of evidence and t he ex t ensio n