基于综合型模糊支持向量机的故障诊断方法及应用  

Fault diagnosis method based on integrated fuzzy support vector machine and its application

作　　者：刘冠军[1] 苏永定[1] 潘才华[1] 

机构地区：[1]国防科技大学机电工程与自动化学院,长沙410073

出　　处：《仪器仪表学报》2009年第7期1363-1367,共5页Chinese Journal of Scientific Instrument

摘　　要：设备信息和故障的不确定性、模糊性及故障样本的缺乏给故障诊断带来了较大的困难。针对该问题,分析了现有模糊支持向量机的原理和优缺点,提出了一种综合型模糊支持向量机。该模糊支持向量机既可以处理样本含有模糊信息的情况,又可以解决支持向量机分类中存在的不可分问题。然后,提出了基于综合型模糊支持向量机的故障诊断方法,并在某电路系统故障诊断中开展了应用研究。应用结果表明,该诊断方法在设备状态存在模糊性和故障样本较少的情况下,与现有模糊支持向量机诊断方法相比,实现了较准确的故障诊断。There is much difficulty in fault diagnosis caused by uncertainty, illegibility and lack of system fault sample. Aiming at the problem, the basic theory and characteristic of existing fuzzy support vector machine are analyzed. And an integrated fuzzy support vector machine is put forward. This fuzzy support vector machine not only can process the fuzzy information in data sample, but also can solve the impartibility problem in support vector machine. Then, a fault diagnosis method based on the integrated fuzzy support vector machine is proposed. Application research is implemented in the fault diagnosis of certain circuit system. Application results show that this method can realize exact fault diagnosis in the condition of illegibility and lack of sample.

关 键 词：支持向量机 模糊支持向量机 故障诊断 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]