Weighted Entropy for Data Compression in Smart Grid

Compression is an important issue in smart grid communication to transmit and store large volume of power system data. In this work, a data compression method for smart grid signals is presented based on wavelet packet transform (WPT). Minimum description length (MDL) algorithm is employed for the selection of the best wavelet function and decomposition level. Additionally, a modified Shannon entropy is used for the selection of the best basis presentation of a signal out of the full WPT. Finally, a threshold criteria is applied to the WPT best basis coefficients to obtain a finely controlled and flexible compression ratio (CR). A set of power system signals from different measurement devices of Smart Grid are used for the performance evaluation of the presented technique. The presented method significantly reduces the size of the data after compression. Comparative results are presented with a fuzzy transform based method which is utilized for the compression of electric signal waveforms for smart grid.