Intrinsic mode function determination of faulty rolling element bearing based on kurtosis

Feature extraction of faulty rolling element bearings (REBs) is important in condition monitoring and faults diagnosis of rotating machinery. The empirical mode decomposition (EMD) is a useful tool for non-stationary signal analysis, by which the analyzed vibration generated by the faulty REBs can be decomposed into a set of intrinsic mode functions (IMFs). However, how to determine an interesting intrinsic mode function (IMF) is often difficult in the applications of EMD. To address this issue, a kurtosis based scheme for the feature extraction of faulty REBs has been introduced in this paper. In the scheme, the vibration generated by REBs is acquired firstly. Then, the EMD is performed to decompose the vibration into a set of IMFs. Lastly, the kurtosis is calculated to determine the interesting IMF, which has the maximum kurtosis value. Simulation and test results supported the introduced scheme positively.