Power quality disturbance classification based on wavelet transform and self-organizing learning neural network

Abstract

A novel approach for the power quality (PQ) disturbances classification based on the wavelet transform (WT) and selforganizing learning array (SOLAR) system is proposed. Wavelet transform is utilized to extract feature vectors for various PQ disturbances and the WT can accurately localizes the characteristics of a signal both in the time and frequency domains. These feature vectors then are applied to a SOLAR system for training and disturbance pattern classification. By comparing with a classic neural network, it is concluded that SOLAR has better data driven learning and local interconnections performance. The research results between the proposed method and the other existing method is discussed and the proposed method can provide accurate classification results. On the basis of hypothesis test of the averages, it is shown that corresponding to different wavelets selection, there is no statistically significant difference in performance of PQ disturbances classification and the relationship between the wavelet decomposition level and classification performance is discussed. The simulation results demonstrate the proposed method gives a new way for identification and classification of dynamic power quality disturbances.