Advancements in Arc Fault Detection for Electrical Distribution Systems: A Comprehensive Review from Artificial Intelligence Perspective
Kriti Thakur, Divyanshi Dwivedi, K. Victor Sam Moses Babu, Alivelu Manga Parimi, Pradeep Kumar Yemula, Pratyush Chakraborty, Mayukha Pal
TL;DR
This paper surveys arc fault detection in electrical distribution systems from an artificial intelligence perspective, outlining traditional signal-processing approaches (time-, frequency-, and time-frequency-domain methods) and contrasting them with AI-driven techniques. It highlights the superior performance and scalability of AI-based methods, while detailing models from ML to DL (SVM, RF, DNN, RNN, CNN, etc.) and representative hybrids, such as HTFNN and transformer-inspired approaches. The review emphasizes arc fault models and detection challenges across linear, nonlinear, AC, and DC loads, identifies gaps in generalization and data requirements, and proposes pathways like edge AI and data-efficient learning to enable real-time, robust protection. The work aims to guide researchers and practitioners toward safer, more reliable electrical distribution networks by mapping current capabilities, gaps, and future research directions in arc fault detection. It underscores the practical impact of AI-enabled arc fault detection for reducing fire risk, improving reliability, and accelerating protective actions in modern grids.
Abstract
This comprehensive review paper provides a thorough examination of current advancements and research in the field of arc fault detection for electrical distribution systems. The increasing demand for electricity, coupled with the increasing utilization of renewable energy sources, has necessitated vigilance in safeguarding electrical distribution systems against arc faults. Such faults could lead to catastrophic accidents, including fires, equipment damage, loss of human life, and other critical issues. To mitigate these risks, this review article focuses on the identification and early detection of arc faults, with a particular emphasis on the vital role of artificial intelligence (AI) in the detection and prediction of arc faults. The paper explores a wide range of methodologies for arc fault detection and highlights the superior performance of AI-based methods in accurately identifying arc faults when compared to other approaches. A thorough evaluation of existing methodologies is conducted by categorizing them into distinct groups, which provides a structured framework for understanding the current state of arc fault detection techniques. This categorization serves as a foundation for identifying the existing constraints and future research avenues in the domain of arc fault detection for electrical distribution systems. This review paper provides the state of the art in arc fault detection, aiming to enhance safety and reliability in electrical distribution systems and guide future research efforts.