Automated Classification of Power-Quality Disturbances Using SVM and RBF Networks

Abstract

The authors propose a new method of power-quality classification using support vector machine (SVM) neural networks. Classifiers based on radial basis function (RBF) networks was, in parallel, applied to enable proper performance comparison. Both RBF and SVM networks are introduced and are considered to be an appropriate tool for classification problems. Space phasor is used for feature extraction from three-phase signals to build distinguished patterns for classifiers. In order to create training and testing vectors, different disturbance classes were simulated (e.g., sags, voltage fluctuations, transients) in Matlab. Finally, the investigation results of the novel approach are shown and interpreted.