摘要
针对滚动轴承故障特征提取与状态监测问题,提出一种基于集合经验模式分解(EEMD)、Renyi熵、主元分析(PCA)和概率神经网络(PNN)的新方法.首先,将轴承振动信号通过EEMD分解成一组本征模态函数(IMF),计算每个IMF分量的Renyi熵值作为表征故障特征的向量,采用主元分析(PCA)对特征降维,提取主元输入概率神经网络进行故障分类.通过SKF6203轴承的正常、内圈点蚀、外圈点蚀和滚动体点蚀这4类状态的诊断实验验证了方法的有效性,诊断正确率为91.7%.
In order to solve the problems of fault feature extraction and condition monitoring of rolling bearings, a novel approach based on the ensemble empirical mode decomposition (EEMD), the Renyi-entropy, the principal component analysis (PCA) and the probabilistic neural network (PNN) is proposed. The vibration signals are first decomposed into a couple of intrinsic mode functions (IMFs) by the EEMD method, and the Renyi entropy of each IMF is computed as the fault characteristic vectors. Then, the PCA is used for feature reduction, after that the principal components (PCs) are acquired and fed into the PNN for final classification. The experiment of SKF6203 rolling bearings including norm condition, inner race defect, outer race defect and ball defect proves the validity of the proposed method, and the diagnostic accuracy reaches 91.7%.
出处
《东南大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2011年第B09期107-111,共5页
Journal of Southeast University:Natural Science Edition
基金
江苏省自然科学基金资助项目(BK2009356)
江苏省高校自然科学研究资助项目(09KJB510003)
关键词
故障诊断
滚动轴承
集合经验模式分解
RENYI熵
主元分析
概率神经网络
fault diagnosis
rolling bearing
ensemble empirical mode decomposition (EEMD)
Renyi entropy
principal component analysis (PCA)
probabilistic neural network (PNN)
作者简介
窦东阳(1983-),男,博士,讲师,ddy41@163.com.