Comprehensive review of power quality disturbance detection and classification techniques

In recent decades, Power Quality Disturbances (PQD) analysis has gained significant attention due to the excessive use of non-linear power electronics. This review paper provides a comprehensive analysis of PQD detection and classification using signal processing methods for feature extraction. Methods such as Artificial Intelligence (AI), Artificial Neural Networks (ANN), Neuro-Fuzzy (NF), Genetic Algorithm (GA), and Deep Learning methods (DL), among others. Additionally, Discrete Wavelet Transform (DWT), S-Transform (ST), Multi-Resolution Analysis (MRA), and Wavelet Transform (WT) techniques are discussed. Herein, various feature extraction techniques and their combinations with intelligent methods were also evaluated for classifying PQDs. While various AI and feature extraction techniques have been examined for PQD classification, they often suffer from limitations such as high computational complexity and constraints in real-time conditions. However, experiments on datasets demonstrate an improvement in detection accuracy compared to state-of-the-art methods. A novel hybrid framework combining DL and GA methods, such as CNN with optimized DWT and MRA, aims to improve classification accuracy while maintaining computational efficiency. This framework demonstrates the potential of traditional techniques as reliable and effective classifiers compared to other algorithms.