|
Fault Pattern Recognition of Rolling Bearing Based on EMD-SVD Model and SVM
WU Hu-Sheng;Jian-Xin LV;Lai-Hong LAI;Lu-Shan WU;Yu-Rong ZHU
2011, 31(2):
89-93.
DOI: 10.3969/j.issn.1006-1355-2011.02.022
According to the non-stationarity characteristics of the vibration signals from rolling bearing and the difficulty for obtaining enough fault samples, a comprehensive fault diagnosis method based on Empirical Mode Decomposition (EMD),Singularity Value Decomposition (SVD), Renyi-entropy and Support Sector Machine (SVM) is proposed. Firstly, the denoised vibration signals are decomposed into a finite number of Intrinsic Mode Functions (IMF). Secondly, some IMF components are selected according to the criterion of mutual correlation coefficient between IMF components and denoised signal. Thirdly, the phase space of the selected IMF components is reconstructed so as to obtain the attractor orbit matrix. Fourthly, with the SVD method, singular value sequences are obtained, and then Renyi-entropies of these sequences are calculated as faulty eigenvector. Finally, the eigenvector serves as input of SVM classifier so that the faults of rolling bearing are recognized. Practical rolling bearing experiment data is used to verify this method, and the diagnosis results and comparative tests fully validate its effectiveness and generalization ability.
Related Articles |
Metrics
|