Optimal feature retrieval for classification of non-stationary Power Quality disturbances

Abstract

Since last few decades, Power Quality (PQ) issues has drawn the attention of both the utilities and the customers. This paper presents one of the most advanced signal-processing techniques i.e., Wavelet Transform (WT) to extract some of the important useful features of the non-stationary PQ signal. The features are then used to classify the nature of the PQ disturbance. The feature dimension is further reduced by selecting the optimal set of features using Genetic Algorithm (GA) to achieve a higher classification accuracy. The optimal features obtained using GA are used to train a Support Vector Machine (SVM) classifier for automatic classification of the Power Quality (PQ) disturbances. Nine types of PQ disturbances are considered for the classification purpose. The simulation results show that the combination of WT and SVM can effectively classify different PQ disturbances.