文献类型：期刊文献

中文题名：基于支持向量机的火车滚轴故障诊断

英文题名：Fault Diagnosis of Train Roller Based on Support Vector Machine

作者：田景文[1];吴浩[1];高美娟[2]

第一作者：田景文

机构：[1]北京化工大学信息科学与技术学院;[2]北京联合大学自动化学院

第一机构：北京化工大学信息科学与技术学院,北京100029

年份：2007

卷号：35

期号：7

起止页码：248-250

中文期刊名：机床与液压

外文期刊名：Machine Tool & Hydraulics

收录：CSTPCD;;北大核心:【北大核心2004】；

语种：中文

中文关键词：滚轴;支持向量机;小波变换;故障诊断

外文关键词：Bearing; Support vector machine; Wavelet transform; Fault diagnosis

摘要：火车故障的60%都是由于滚轴问题引起的,现在的诊断方法都是基于知识的,但故障样本的不足从一定程度上制约了基于知识的方法在实际中的应用,针对这一问题,利用支持向量机在小样本情况下具有较强分类能力的特点,本文提出了一种基于支持向量机的滚轴故障诊断方法。该方法采用小波变换对齿轮的震动信号进行处理来构造特征向量,然后输入到支持向量机分类器中进行模式识别。
The 60% fault of train is attributed to bearing problem. The method used for fault diagnosis is based on knowledge now. But lack of fault samples restricts the application of the methods based on knowledge in practical fault diagnosis to a certain extent. In order to solve this problem, a diagnosis method of bearing fault based on a support vector machine was proposed based on the advantage that a support vector machine has strong classification ability with fewer samples. According to this method, feature vectors were extracted from bearing vibration signals after wavelet transform and they were input into amultiole-fault classifier of the support vector machine for fault identification.