Ensemble empirical mode decomposition (EEMD) can restrain mode mixing of EMD at a certain level,however,the calculation amount grows and the completeness loses due to the white noise unneutralized completely.Using complementary EEMD (CEEMD) adds the white noises in pairs into a target signal,the same results as those using EEMD can be obtained but the white noises added are neutralized completely.Here,based on CEEMD and permutation entropy for signal randomness detection,a modified EEMD (MEEMD) was proposed.With MEEMD,the data were decomposed using EMD after the abnormal components were separated using the ensemble and average method.It could not only restrain mode mixing effectively,but also decrease the reconstruction error and the computation amount.By analyzing simulation and actual signals,the results indicated that MEEMD is feasible and effective for restraining mode mixing of EMD.
NETL
NSTL
万方数据
维普
