A fast entropy assisted complete ensemble empirical mode decomposition algorithm

Empirical mode decomposition (EMD) is a simple and real-time procedure to adaptively decompose a signal into a set of oscillation scales, but it faces the serious problem of mode mixing. The improved complete ensemble EMD with adaptive noise (Improved CEEMDAN) can successfully eliminate the mode mixing by adding white noise´s IMFs and utilizing an ensemble and average procedure, but it does not satisfy the real-time processing requirement. In this paper, a new fast entropy assisted CEEMD (FEACEEMD) approach will be explained, in which the permutation entropy (PEn) index that marks an IMF´s randomness and intermittence characteristic is used to control the fusion usage of both Improved CEEMDAN and EMD in order to bring in the good things from both sides. Artificial experiments showed that the new method is much more effective, real-time and robust than the original improved CEEMDAN. Additionally, experiments using ship recorded data showed the algorithm´s engineering application potentiality.