Bearing Fault Diagnosis Based on Wavelet Transform and ICA

The key to fault diagnosis of rolling element bearing is how to find typical characteristic frequencies from low SNR mixed signals. Jointing Continuous Wavelet Transform (CWT) with Independent Component Analysis (ICA)，this paper proposes a method to select wavelet scales with iso-interval frequency and analyze envelope spectrum of independent signal to diagnose the fault of rolling element bearing. Finally, the effectiveness of this method has been verified by practical signal of rolling element bearing.