Research on Rolling Element Bearing Fault Diagnosis Based on EEMD and Correlated Kurtosis

In order to extract the faint fault information from complicated vibration signal of bearing, the correlated kurtosis is introduced into the field of rolling bearing fault diagnosis. Combined with ensemble empirical mode decomposition (EEMD) and correlated kurtosis, a feature extraction method is proposed. According to the method, by EEMD processing a group of intrinsic mode functions (IMFs) are obtained, then the IMF with maximal correlated kurtosis is selected, and the weak fault signal is clearly extracted. The effectiveness of the method is demonstrated on both simulated signal and actual data.