Abstract—This paper presents the findings of an investigation into Partial Discharge signal denoising using techniques based on Empirical Mode Decomposition. The denoising techniques are based on thresholding the Intrinsic Mode Functions which result from the Empirical Mode Decomposition of a signal. The results of the tests carried out show clearly that these techniques can produce excellent results when applied to simulated Partial Discharge data. Furthermore, in the case of the data used in these simulations the techniques are able to outperform a Discrete Wavelet Transform based denoising technique.
 

Index Terms—empirical mode decomposition (EMD), signal denoising, partial discharge