Classification of power quality disturbances at power system frequency and out of power system frequency using support vector machines

Abstract

In this paper, firstly it is tried to classify pure sine and power quality disturbances (PQD) such as voltage sag, voltage swell, voltage with harmonics, transients and flicker at power system frequency (50 Hz). Wavelet transform (WT) is used to extract distinctive features. Wavelet energy criterion is applied to wavelet detail coefficients. It is seen that classification performance of support vector machine (SVM) used as classifier is well. Then pure sine and PQD, that are out of power system frequency, are tried to classify. Curve fitting approach is used for estimating frequency. It is observed that SVM classifies PQD signals well when frequency of pure sine is updated with the frequency of PQD even if they deviate from 50 Hz. © 2013 Wydawnictwo SIGMA - N O T Sp. z o.o. All rights reserved.