Sensitive Features Extraction of Early Fault Based on EEMD and WPT

A method based on ensemble empirical mode decomposition (EEMD) and wavelet packet transform (WPT), which is used for extracting early failure sensitive features, is presented. An early fault diagnosis model is also built. According to this method, firstly vibration signals from the working site are decomposed by using EEMD into different IMFs(intrinsic mode function); the IMFs of the maximum related coefficient of IMF components and the original signal are then chosen to form the new information; Oriented IMFs, WPT decomposition is carried. each node of wavelet coefficient is obtained. They are decomposed by using WPT to obtain the wavelet coefficient of each node. Finally the envelops of wavelet packet coefficient are calculated by using Hilbert transform and the power spectrum can be used to obtain the early fault sensitive features. Effectiveness of the proposed method is verified through simulation. This method is also used for rolling bearing inner ring, outer ring faults and fault diagnosis of rolling elements. The diagnosis results indicate that the method of extracting sensitive features is effective and realizes fast and accurate fault diagnosis.