Wavelets for the analysis and compression of partial discharge data

One of the major problems encountered in the on-line condition monitoring of partial discharges for high voltage plant items is the huge volume of memory required to store PD data. This paper proposes wavelet-based and wavelet packet-based compression methods for PD measurement data, which are performed through wavelet decomposition, detail coefficient thresholding and signal reconstruction. All of the compression methods associated with these strategies are treated and the compression results as applied to practical PD data are compared. Using this technique in conjunction with the transitional compression method, it has been demonstrated that data from partial discharge activity can be compressed to approximately 5% of its original volume, which is very promising for practical on-line condition monitoring.