Intelligent assessment of power quality disturbances: A comprehensive review on machine learning and deep learning solutions

Abstract

As the global focus on clean energy and smart grids intensifies, detecting power quality disturbances (PQDs), caused by energy instability, has become increasingly critical for achieving sustainable development goals by ensuring stable and reliable power quality. Power quality disturbances, such as voltage sags or harmonics, are disruptions in the electrical supply that can affect everything from household appliances to industrial machinery, making their detection and management essential for a stable power system. They can cause significant damage to power grid infrastructure, leading to energy inefficiency, restricted electricity generation and consumption, equipment malfunction, and industrial process failures. The incorporation of artificial intelligence (AI) has transformed PQD classification, providing substantial advancements in monitoring and managing electrical systems. This paper presents a systematic review of the existing literature, focusing on the integration of machine learning and deep learning techniques for PQD detection. It analyzes high-quality studies on PQDs detection and classification, categorizing them based on the AI techniques employed. Additionally, it emphasizes the role of digital signal processing (DSP) techniques in extracting features, with studies segregated based on the incorporation of DSP and non-DSP approaches. A case study demonstrates the practical application and effectiveness of AI techniques in real-world contexts, with the Bagged Trees classifier achieving the highest testing accuracy of 96.6 %. The insights provided aim to support researchers and practitioners in navigating the evolving landscape of power quality assessment, ultimately improving the reliability and accuracy of PQD detection systems.