Machine Learning-Based Techniques for False Data Injection Attacks Detection in Smart Grid: A Review

Abstract

AbstractSmart grids (SG) provide new technological solutions for optimal energy utilization and management. Along with its significance and status as a cyber-physical system (CPS), it is vulnerable to a variety of cyber-security risks. The greatest threat to smart grid security is a False Data Injection (FDI) Attack. To efficiently detect this threat, numerous machine learning-based algorithms have been proposed in the literature. This article presents the most up-to-date machine learning-based techniques and methods for bogus data injection detection. The article begins with an overview of the smart grid and a brief history of cyber-attacks on smart grids, followed by a discussion of security needs and a taxonomy of false data injection depending on delivery mode. Finally, we discuss the research that has been performed in the detection of false data injection attacks, which have been categorized according to the used learning approach.KeywordsSmart gridCyber securityMachine learningFalse data injection attackDeep learning