Compound Fault Diagnosis for Bearings Based on Morphological Component Analysis

The ICA is widely used in blind source separation of linear mixture model,which has two important limitations: statistically independent sources and non-Gauss distribution.The morphological component analysis(MCA) is a novel decomposition method based on sparse representation of signals and images.The basic principle of MCA is introduced and then illustrated with simulations,and the method is used to diagnose the bearing with three types of faults.Finally,the fault features are found out and the faults are distinguished successfully,which verify the validity of the method in fault diagnosis of bearing.