Partial discharge signal extracting using the empirical mode decomposition with wavelet transform

Empirical mode decomposition (EMD) has good adaptivity for non-stationary and nonlinear signal analysis. This paper uses the advantage of EMD and combines with the wavelet transform (EMD-WT) to extract partial discharge (PD) signals in noises. The wavelet transform is a common used method for PD signal denoising. However, once the signal to noise ratio (SNR) decreases seriously, the WT method will be failed. Compare to the WT method, the EMD-WT has better performance for noise reduction. It has been verified that the EMD-WT method can preserve more information even though the SNR is low. The results show that the EMD-WT is suitable for PD denoising in a noisy environment.