摘要
为了提高滚动轴承的故障诊断率,提出了一种经验模态分解(empirical mode decomposition,EMD)结合时域分析后使用主成分分析(principal component analysis,PCA)融合特征量的特征提取方法。首先,通过EMD分解得到前5个本征模态函数(intrinsic mode function,IMF)分量的上、下包络值矩阵的奇异值;然后,对轴承原始信号进行时域分析得到各种时域特征参数;最后对奇异值和时域特征参数使用PCA降维融合后输入到多分类支持向量机(support vector machines,SVM)中进行分类。通过实验仿真验证,融合后的特征量诊断准确率达到了98.6%,该方法能充分地提取出轴承故障特征信息,诊断效果良好。
In order to improve the fault diagnosis rate of rolling bearings,a feature extraction method using empirical mode decomposition(EMD)combined with time domain analysis and then using principal component analysis(PCA)to fuse feature quantities was proposed.First,the singular values of the upper and lower envelope value matrices of the first five eigenmode function(IMF)components were obtained through EMD decomposition.Then,the original bearing signal was analyzed in time domain to obtain various time-domain characteristic parameters,Singular values and time-domain feature parameters were fused using PCA dimensionality reduction and then input into a multi-class support vector machine(SVM)for classification.It was verified by experimental simulation that the accuracy of the fusion feature quantity diagnosis reaches 98.6%.This method can fully extract the bearing fault feature information,and the diagnosis effect is good.
作者
汪峰
周凤星
严保康
WANG Feng;ZHOU Feng-xing;YAN Bao-kang(School of Information Science and Engineering, Wuhan University of Science and Technology, Wuhan 430081, China)
出处
《科学技术与工程》
北大核心
2022年第6期2351-2356,共6页
Science Technology and Engineering
基金
国家自然科学基金(51975433)
湖北省自然科学基金(2019CFB133)。
关键词
轴承故障
经验模态分解
主成分分析
支持向量机
bearing fault
empirical mode decomposition
principal component analysis
support vector machine
