Analysis of VHF-wideband electromagnetic noises from partial discharge using discrete wavelet transform

In this paper, a new application of discrete wavelet transform to analyze the wide-band electromagnetic waves from partial discharge (PD) is introduced. The discrete wavelet transform has a special feature to perform the signal decomposition and reconstruction by multi-resolution analysis. Using this method, one can decompose the detected signal into various frequency sub-band signals, and then estimate the true shape of PD signal by mixing the appropriately-selected decomposed signals. The threshold processing of sub-band signals is applied to reduce the interference of broadcasting signals. It is shown that the proposed method is advantageous to obtain almost the true shape of PD signal, whose characteristics can be used for the purpose of PD classification. Experimental results obtained by applying high voltage on faulty insulators are used to show the effectiveness of the proposed method