Integrated DWT-DHT Feature Set for ABC Optimized SVM-Based PQ Classifier

Abstract

The power quality (PQ) literature contains abundant methodologies for the classification of PQ disturbances that show the significance of the problem and the necessity to design a reliable method that will accurately and quickly identify the type of disturbance. However, most of the available methods are not suitable for real-time applications as they are sensitive to noise, require more memory, computationally complex, and are also time-consuming. This work suggests an approach using Artificial Bee Colony Algorithm (ABC) for concurrent feature and classifier parameters selection to reduce execution time and enhance classification accuracy. An integrated feature set consisting of features extracted using two signal processing techniques namely Discrete Wavelet Transform (DWT) and Discrete Hilbert Transform (DHT) is proposed for clear discrimination of disturbances in contrast to the conventional single feature extraction tool commonly employed. To authenticate the effectiveness of the proposed dual set features, this method has been tested on both simulated and real-time PQ data and is found to perform better than methods using a single signal processing tool.