A wavelet-based method to extract frequency feature for power system fault/event analysis

Frequency is a vital factor for power system operation and protection. This paper is on extraction of frequency feature for fault/event analysis based on wavelet transform (WT) and fractal geometry (FG). The frequency signal is decomposed by WT-based multiresolution analysis (MRA) and a family of wavelet coefficients is obtained. A maxima line is constructed by connecting the maximum point in the wavelet coefficients across all the decomposition levels. A differential box counting (DBC) method based on FG is applied to compute the fractal dimension of the maxima line. It is realized that the feature of a frequency signal can be characterized by a fraction number-fractal dimension. A simulation is carried out in PSS/E to generate different faults in the power system. The proposed algorithm is implemented to extract the features of the faults. The results verify the effectiveness of the proposed method.