摘要
滚动轴承处于早期故障阶段时,故障冲击特征成分难以提取,为了从轴承故障振动信号中提取特征参数,对轴承故障振动信号进行变分模态分解(Variational Mode Decomposition,VMD),得到若干个本征模态分量(IMFs),计算各个IMF的能量熵与样本熵,并利用主成分分析方法(PCA)对其进行特征融合。最后利用粒子群算法(PSO)优化的支持向量机(SVM)对融合特征进行故障模式识别。轴承故障实验分析结果表明,所提方法能够有效实现滚动轴承故障诊断。
Early fault feature of rolling bearing vibration signal is weak and difficult to extract.In order to extract the characteristic parameters from a bearing fault vibration signal,the bearing fault vibration signal was decomposed by VMD,and several intrinsic mode function(IMF)are obtained.The energy entropy and sample entropy of each intrinsic mode functions was calculated and characterized by PC A.Finally,the PSO一optimized support vector machine was used to training and test the fusion feature.The experimental results show that the proposed method can be effectively used to diagnose rolling bearing fault.
作者
张龙
宋成洋
邹友军
崔路瑶
雷兵
ZHANG Long;SONG Chengyang;ZOU Youjun;CUI Luyao;LEI Bing(East China Jiaotong University,School of Mechatronies&Vehicle Engineering,Nanchang 330013,China)
出处
《机械设计与研究》
CSCD
北大核心
2019年第6期96-104,共9页
Machine Design And Research
基金
国家肖然科学基金及(51665013)
江西省自然科学基金(20161BAB216134,20171BAB216030)资助项目。
关键词
变分模态分解
样本熵
支持向量机
粒子群算法
故障诊断
variational mode decomposition
sample entropy
support vector machine
PSO
fault diagnosis
