On The Hilbert-Huang Transform and Its Application to System Identification

摘要

The essence of the Hilbert-Huang transform(HHT),a promising signal processing tool,is the combination of the so-called empirical mode decomposition(EMD)that decomposes a time domain signal into its intrinsic mode functions(IMFs)and the Hilbert transform(HT)of the decomposed IMFs that leads to a temporal-frequency amplitude/energy distribution of the acquired response.Since the Hilbert transform is directly related to the Fourier transform,it may inherit the leakage problem.The subsequently identified amplitude,phase or frequency based on the distorted Hilbert transform may yield inaccurate or even false results of system identification.In this paper,we attempt to tackle such issue by using a regressive Fourier technique and an envelope amplitude curve fitting approach.The case studies show that these methods can improve the accuracy of HHT.The implementation of the improved HHT to damage detection application is also discussed.