Power quality events classification and recognition using a novel support vector algorithm

This paper presents a method of power quality classification using support vector machines (SVMs). In SVM training, the kernel parameters, and feature selection have very important roles for SVM classification accuracy. Therefore, most appropriates of these kernel types, kernel parameters and features should be used for the SVM training. In this paper to get optimal features for the classifier two stage of feature selection has been used. In first stage mutual information feature selection (MIFS) and in the second stage correlation feature selection (CFS) techniques are used for feature extraction from signals to build distinguished patterns for classifiers. MIFS can reduce the dimensionality of inputs, speed up the training of the network and get better performance and with CFS can get optimal features. In order to create training and testing vectors, different disturbance classes were simulated using parametric equations i.e., pure sinusoid, sag, swell, harmonic, outage, sag and harmonics and swell and harmonics. Finally, the investigation results of this novel approach are shown. The test results show that the classifier has an excellent performance on training speed, reliability and accuracy.