Faulty Detection of Rolling Bearing Based on Empirical Mode Decomposition and Spectral Kurtosis

Rolling bearing fault is one of the major faults of rotating machinery. However, vibration generated by incident faults of rolling bearings are weaker, non-stationary and nonlinear. Therefore, the interesting components extraction from the observed vibration is important for the whole process of diagnosing analysis. In order to improve the effectiveness of fault diagnosis of rolling bearings, this paper presents a diagnosis method based on empirical mode decomposition (EMD) and spectral kurtosis. Firstly, the raw vibration signal is preprocessed by AR filtering. Secondly, the vibration is decomposed into a number of intrinsic mode functions (IMFs) through EMD. Thirdly, we can calculate factors called "Cross-correlation coefficient" which could reconstruct selected IMFs. Finally, we can calculate the cross-correlation coefficient and spectral kurtosis (SK) value for every IMF component. The results show that the SK method can be effectively improved by the EMD filtering.