Research of Mechanical Fault SVM Intelligent Recognition Based on EEMD Sample Entropy

Sensors & Transducers

Junwei Lou¹  Jiali Zhao¹  Yuan Liu¹  Chibing Hu¹ 

[1] School of Mechanical and Electronical Engineering, Lanzhou University of Technology, No 287, Lan Gongping Street, Lanzhou, Gansu, 730050, P.R. China;

关键词: EEMD;    Sample entropy;    SVM;    Rolling bearing;    Intelligent diagnosis.;   

DOI  :  

来源: DOAJ

【 摘 要 】

The extraction of fault information is the key of fault intelligent recognition of support vector machine for rolling bearing. Because of the non-adaptive and mode mixture of wavelet transform and empirical mode decomposition, ensemble empirical mode decomposition (EEMD) and sample entropy have been adopted to extract fault information of rolling bearing. For three kinds of conditions and pitting diameters, the vibration signal of rolling bearing has been acquired by experiment. Then by wavelet transform to reduce noise, the noise reduction signal has been decomposed into several intrinsic mode function components by EEMD, and the complexity of major components has been described by sample entropy. In addition, a SVM rolling bearing fault classification recognizer which EEMD sample entropy has been adopted as training and recognition samples is proposed. The experiment result shows that under small sample, the inner race, outer race and ball fault of bearing can be accurately recognized and the accuracy for reorganization enhance with the number of samples increasing.

【 授权许可】

Unknown