Power Quality Disturbance Classification Using Frequency Domain Features

Abstract

Discrete Cosine Transform feature has become an effective feature extraction method in modern power system. In this paper, we used algorithm of a unique feature for power quality (PQ) disturbance signal classification. Here, we proposed the extraction of spectral features from Discrete Cosine Transform (DCT) domain. This feature extraction offers the ability to detect and localize harmonic events and it also classifies different power quality disturbance signals. A useful technique of selecting significant DCT coefficients is proposed for optimal feature selection. This process offers dimensional feature reduction. In this paper we consider seven types of power quality disturbance signals and simulate for each of the given categories. Using this extracted feature we can get not only very high classification accuracy but also a low computational burden. This feature extraction using Discrete Cosine Transform is one of the best feature extraction formula, we have ever seen.