Fault Diagnosis of Rolling Bearing Based on EEMD and MMTS

Abstract

Aiming at the non-stability of vibration signals and the inaccuracy of signal extraction of motor bearings, a fault diagnosis method based on the ensemble empirical modal decomposition (EEMD) and Multiclass Mahalanobis-Taguchi system (MMTS) is presented. Firstly, the original vibration signal is processed by EEMD and the processed signal is decomposed into a series of IMF with different characteristic time scales. Then select several IMF components in which the signal energy is concentrated to calculate the characteristic parameters of each order of IMF to construct fault eigenvector. Secondly, the fault eigenvector is used as input to construct the MMTS for fault identification. Finally, the accuracy of the method is verified by the bearing data set collected by the Electrical Engineering Laboratory of Case Western Reserve University (CWRU). The results show that the method can diagnose rolling bearing faults quickly and accurately.