基于VMD和MPE的滚动轴承故障分析研究  

Fault Analysis of Rolling Bearings Based on VMD and MPE

作　　者：于士军[1] 刘豪睿[1] 朱恒伟[1] 荣垂霆 胡凯[1] 李天志[1] YU Shijun;LIU Haorui;ZHU Hengwei;RONG Chuiting;HU kai;LI Tianzhi(School of Computer and Information,Dezhou University,Dezhou Shandong 253000,China)

机构地区：[1]德州学院计算机与信息学院,山东德州253000

出　　处：《德州学院学报》2025年第2期42-47,共6页Journal of Dezhou University

基　　金：德州学院实验技术立项项目(SYJS24015)。

摘　　要：滚动轴承在工作中会出现非线性特征的故障振动信号,针对单一排列熵(PE)在故障特征提取过程中所产生的效果不理想及准确率比较低的情况,提出了一种基于变分模式分解(VMD)和多尺度排列熵(MPE)的滚动轴承故障诊断分析方法,提取故障振动信号的固有模态函数(IMF)的多尺度排列熵特征,并且结合支持向量机SVM和KNN进行分析。通过和PE进行诊断分析对比,表明此方法能提高检测准确率。Fault vibration signals with nonlinear characteristics may appear during rolling bearing operation.In response to the unsatisfactory performance and low accuracy of single permutation entropy(PE)in the fault feature extraction process,a fault diagnosis method for rolling bearings based on variational mode decomposition(VMD)and multi-scale permutation entropy(MPE)is proposed.This paper extracts multi-scale permutation entropy features of the inherent mode function(IMF)of fault vibration signals,and combines them with support vector machine SVM(support vetor machine)and KNN(k-nearest neighbors)for analysis.Comparing analysis with PE(permutation entropy)diagnostic,it is shown that this method can improve detection accuracy.

关 键 词：滚动轴承 多尺度排列熵 支持向量机 

分 类 号：TP391.5[自动化与计算机技术—计算机应用技术]