An improved method for EMD modal aliasing effect
Shao-Wei Feng, Kai Chai
Abstract
In order to eliminate the modal aliasing effect of the empirical modal decomposition (EMD) caused by signal interruption (singular event, cycling wave, local impact signal) or the interaction of each component, two elimination methods are proposed, namely, high frequency harmonic method and mask signal method. Firstly, two causes of EMD mode aliasing are analyzed: one is the local oscillation in the signal; the other is the amplitude-frequency relation of the signal. Secondly, the local impact mode aliasing is solved by the high frequency harmonic method. Finally, mask signal method is used to eliminate modal aliasing of similar frequencies. The simulation results show that the two methods of eliminating EMD modal aliasing, which are collected high frequency harmonic method and mask signal method, can quickly and effectively curb the modal aliasing phenomenon on the basis of ensuring the adaptive characteristics of EMD, and provide a novel method to expand the real-time online application of EMD.