Wind generator fault detection using end effects processing of Hilbert-Huang transform

This paper presents a novel approach to improve the end effects of Hilbert-Huang transform (HHT) for the wind generator fault detection. The proposed approach utilizes a back-propagation neural network (BPNN) to extent the end of the spectrum. HHT consists of empirical mode decomposition (EMD) and Hilbert transform (HT), on which the end effects distort Hilbert spectrum. The extension of the two ends obtained by the BPNN forms a new spectrum to improve the end effects. Experimental results indicate utilizing the proposed approach to analyze generator currents can improve the exactitude of Hilbert spectrum.