Feature Vector Extraction by Using Empirical Mode Decomposition from Power Quality Disturbances

Abstract

Power quality assumes that voltages/currents are in rated frequency and values from transmission to distribution and their shape is pure sinusoid and except these comments the distributed electrical energy is assumed as 'not in good quality'. In this paper, the method known as empirical mode decomposition (EMD) will be used for extracting feature vectors from distorted power signal. The proposed method uses three phase normalized voltage/current signals but single phase analysis of voltage signals will be implemented in this work. Power disturbances such as voltage sag, swell, interrupt, flicker and DC component analysis are successfully decomposed to obtain feature vectors for any classification algorithm by using EMD. © 2012 IEEE.