基于支持向量机故障诊断方法  

Support Vector Machine For Diagnosis Of Mechanical Fault

作　　者：谢芳芳 章兢 郑剑 

机构地区：[1]湖南省长沙市韶山南路740号省化肥工业公司院内 [2]不详

出　　处：《传感器世界》2006年第4期42-44,共3页Sensor World

摘　　要：支持向量机是一种基于统计学习理论的机器学习算法,它能在训练样本很少的情况下达到很好的分类效果。本文以双螺杆挤出机为例,介绍了基于支持向量机的多故障分类器,探讨了“成对分类”与“一类对多类”两种多类分类算法的应用。诊断实例表明,基于支持向量机的多故障分类器对设备故障具有很好的分类效果。The support machine（SVM） is a -learning algorithm based statistical （SLT）,which learning has vector machine on the theory desirable classification ability even if with fewer samples. In this paper, the application of the two algorithms about one-against-one model and one-against-others model is introduced. The twin screw extruder fault diagnosis by multi-fault classifier based on SVM is mainly discussed and the retest proves that this SVM really has preferable ability of classification.

关 键 词：故障诊断 支持向量机 多故障分类器 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]