A comprehensive review on deep learning techniques in power system protection: Trends, challenges, applications and future directions

Abstract

The new edged and multi-source integrated electric power systems (EPSs) with increasing complexity necessitate advanced protection mechanisms to meet the demand for reliability. The primary goal of this review is to explore the potential of deep learning (DL) as a transformative tool in power system protection (PSP). The initial objective is to perform an across-the-board bibliometric study to recognize trends, key contributors, and research hotspots in DL applications within this arena. This analysis has tapped data from primary scientific databases, centering on publication trends, citation patterns, and collaborative associations. Furthermore, this study analyses numerous applications of DL in PSP, including automatic fault recognition, differential and distance protection, anomaly detection, adaptive mechanisms, and cybersecurity. The findings show a rising interest and investment in DL expertise for PSP, with distinguished support from a diverse range of scientists and academicians. Extensive case reports of DL applications reveal substantial developments in fault detection and location accuracy, tripping times, and system adaptability, outperforming conventional PSP approaches. This comprehensive survey highlights the potential of deep learning to improve the resilience and safety of utility systems, leading to a more stable and efficient energy distribution network. As a result of this advancement, several societal impacts can be observed, such as reduced power outages, upgraded safety, and a more sustainable energy supply, benefiting both consumers and the industry landscape.