A comprehensive review of anomaly detection in smart grids: A machine learning approach to cyber threat identification

Abstract

Security of smart grids (SG) has gained more attention in the past decades as one of the crucial problems. Implementing Machine learning (ML) has been the most capable solution for strengthening cyber security in SGs. Considering the significance of ML for anomaly detection, this review paper investigates the role of ML models in ensuring SG security. Various research works have discussed the issue of SG security and have reviewed different technologies used including the deployment of ML for attack detection in SG. However, the continuous evolution of cyber threats in the dynamic SG environment motivates this review to contribute to the existing works. This review discusses the types of anomalies and cyber threats which threaten the security of SG. In addition, this review examines the role of different ML models along with datasets used for training ML models. On the basis of the study, this paper identifies the challenges and research gaps and outline possible future research directions.