Power Quality Faint Disturbance Identification Using Wavelet Packet Energy Entropy and Weighted Support Vector Machines

Abstract

Wavelet packet energy (WPE) features are usually used to detect power quality disturbances (PQDs) in recent literatures, but by computing, we find that WPE features are not valid to detect faint disturbances. In this paper, wavelet packet energy entropy (WPEE) and weighted support vector machines (WSVMs) are utilized to detect and classify PQDs automatically. Electric power quality is an aspect of power engineering that has been with us since the inception of power systems. The types of concerned disturbances including voltage sag, swell, harmonics, flicker and interruption. Wavelet packet are utilized to decompose the signals into different special frequency bands, then to obtain five common features for sampling PQ disturbances including energy entropies of frequency bands. WSVMs are designed and trained for making a decision regarding the type of the disturbances. Simulation illustrates the algorithm feasibility and effectiveness.