期刊文献+

基于EMD与多特征的支持向量机故障诊断 被引量:2

Fault Diagnosis Based on EMD and Multiple Features SVM
下载PDF
导出
摘要 针对齿轮振动信号的非线性、非平稳性与故障分类识别率低的问题,提出了一种基于EMD(经验模态分解)与多特征的支持向量机(SVM)故障诊断方法。该方法首先将采集到的信号进行小波包降噪;然后提取降噪后信号的各项时域参数指标;同时,将降噪信号经过EMD运算并提取以互相关准则选取的各本征模式分量(IMF)的能量指标;最后,将两部分特征向量组合后作为SVM的输入进行训练与预测。实验结果表明,该方法对于齿轮状态具有很好的分类精确度,能很好地应用于齿轮的故障诊断。 In order to solve the problems that the gear vibration signals is non-linear、non-stationary and low accuracy in classification,a kind of method that SVM based on the EMD and multiple features for fault diagnosis is presented. First,this method reduces the noise via wavelet-packet; Then,this method extracts the time domain parameters from the signal after reduced noise. At the same time,it decomposes the reduced noise signal using the EMD and extracts the energy indicators from the IMF that is selected according to the cross-correlation rule. Finally,the two parts of feature vectors is treated as input of the SVM for training and prediction. The experimental results show that the method has good classification accuracy for gear status and can be well applied to fault diagnosis of gear.
出处 《机械设计与制造》 北大核心 2015年第10期64-67,共4页 Machinery Design & Manufacture
基金 国家自然科学基金地区基金项目(21366017)
关键词 小波包 EMD IMF 支持向量机 故障诊断 Wavelet Packet EMD IMF SVM Fault Diagnosis
  • 相关文献

参考文献11

二级参考文献59

共引文献288

同被引文献8

引证文献2

二级引证文献8

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部