Vibration Signal Decomposition Based on an Improved Ensemble Empirical Mode Decomposition Method

An improved ensemble empirical mode decomposition(EEMD)method based on varied steplength pattern search was proposed aimed at the optimization of the two critical parameters including the amplitude of added white noise and the number of ensemble trials.This method takes the desired decomposition accuracy of EEMD as the target,uses the amplitude ratio of the white noise derived from adaptive EEMD,and automatically obtains the appropriate parameters of EEMD in terms of different vibration signals.The mode mixing problem can be solved by the proposed method.Simulation experiments and a case study validate the effectiveness of the proposed method.The results show that compared with the traditional EEMD and adaptive EEMD,the improved EEMD method achieves a higher decomposition accuracy and faster computational efficiency.