A wavelet-based universal data compression method for different types of signals in power systems

Abstract

Wavelet-based data compression methods are widely used for the data compression in power systems. There are two major problems of wavelet-based data compression techniques: there is no standard for the selection of best wavelet functions and decomposition scales; most of the existing wavelet-based data compression methods are only suitable for disturbance signals with abrupt change. This paper proposed a wavelet-based universal data compression method for different types of signals in power systems, including oscillations, such as low-frequency oscillations (LFOs) and sub-synchronous oscillations (SSOs), and disturbances. Based on compression ratios (CRs) and distortion rates (DRs), this paper constructed a compression distortion composite index (CDCI) for the evaluation of compression methods, with a balanced consideration of compression and accuracy. Based on CDCI, the best wavelet functions and decomposition scales for different signals can be selected from the candidates. The universality of the proposed method was verified by analyzing the compression of different typical actual recorded signals. The results indicated that the proposed method can provide high compression ratios and low distortion rates.