Motor Speed Signature Analysis of Bearing Fault Detection Based on SK and Adaptive Signal Reconstruction with EEMD

摘要

Local bearing faults will lead to impulses in motor speed which is already available in the vector controlled AC motor drives. Due to the limited sampling frequency of control system, the extraction of fault feature is difficult. Therefore, a signal processing scheme which combines spectral kurtosis (SK)with the adaptive signal reconstruction based on ensemble empirical mode decomposition (EEMD)is proposed. Firstly, motor speed is decomposed by EEMD into a finite number of intrinsic mode functions (IMFs)which reflect the local characteristics of original signal. Then, the correlation coefficient is employed to eliminate the irrelevant components, and reconstruct the speed signal with a higher signal-to-noise ratio adaptively. Finally, the reconstructed speed is analyzed with SK to extract fault features. The proposed scheme is verified on the motor bearing outer ring fault detection platform, and its effectiveness over the conventional SK is also shown.