Feature extraction of early faults of rolling bearings based on LMD and MOMEDA

JIN Jing1,2，LIU Chang 1,2，LAN Yu-tao 1,2，WANG Yan-xue 1,2
(1. School of MechanicalElectronic and Vehicle Engineering, Beijing University of Civil Engineering and
Architecture, Beijing 100044, China；2. Beijing Key Laboratory of Performance Guarantee on Urban
Rail Transit Vehicles, Beijing University of Civil Engineering and Architecture, Beijing 100044, China)

Abstract: Local mean value decomposition (LMD) was not an ideal way to extract the fault characteristics of rolling bearings under strong noise background, the multipoint optimal minimum entropy deconvolution adjusted (MOMEDA) was combined with local mean value decomposition (LMD) to process weak fault signals. First, the vibration signal of the outer ring fault bearing was reconstructed by LMD. Second, the MOMEDA filter was used to perform envelope analysis to extract the fault characteristics. Finally, the proposed method and LMD reconstituted with minimum entropy deconvolution (MED) filter fault feature extraction methods were compared. In addition, the proposed method was also used for inner ring failure analysis. The results show that the proposed method has better applicability for weak fault feature extraction. The multifrequency peak can be seen in the envelope spectrum, and there is little interference near the peak. Simulation and experimental results verify the effectiveness of the method.
Key words: local means decomposition (LMD); multipoint optimal minimum entropy deconvolution adjusted (MOMEDA); minimum entropy deconvolution (MED); rolling bearing; fault diagnosis