Research On Fault Diagnosis Method Of Rolling Bearing Based On Blind Source Separation

Nowadays, along with the development of the industry, the mechanical fault diagnosis technology has become a new subject, and most attention is focused on how to accurately and efficiently extract the fault signal characteristics. Among these methods, the most frequently used feature extract ion method is the signal processing and analysis. In reality, the collected vibration signal is often coupled with interference and the situation that the inter individual source signals mixed with each other, which seriously affects the accuracy of fault feature extraction, and blind source separation method can solve this problem effectively. The main content of this paper is the application of blind source separation in the mechanical, especially the fault feature extraction of rolling bearing.In this paper, the basic theory of blind source separation is analyzed in detail, certain typical faults and its mechanism in rolling bearing are introduced, and certain common time frequency analysis methods are presented. At the same time, the solution is proposed according to the interference and underdetermined problem in the application of rolling bearing, and this method is put into the application of actual rolling bearing diagnosis.Due to there is noise interference in the operation process of rolling bear ing, the characteristic signal is likely to be buried, According to the advantages of wavelet semi soft threshold denoising and blind source separation method, a blind source separation method is proposed to eliminate the noise before and after the blind source separation. In the simulation and experiment, the method is successfully removed the noise from the source signal, the source signals are well separated and the signal features are extracted to verify the effectiveness of the source signals in the fault feature extraction of rolling bearing.According to the underdetermined problem of blind source separation that the number of observed signals is smaller than the number of source signals, proposed a method that using variational mode decomposition(VMD) method to raise the dimension of the observed signal and then conduct blind source separation on the new observed signal. By simulation and comparison, blind source separation method based on variational model decomposition is compared with the method based on the domain mean decomposition(EMD). The results verified the superiority of the method based on variational mode decomposition. Finally, according to the single channel rolling bearing signal, using the method, it extracted the fault features of rolling bearing successfully.