Investigation of Detection Mechanisms Against False Data Injection Attacks Based on Machine Learning Approaches

Due to the proliferation of advanced communication technologies, power systems have evolved from being solely physical to becoming cyber-physical entities. While this transformation has improved system speed, accuracy, and efficiency, it has also exposed various vulnerabilities. Among these vulnerabilities, cyberattacks stand out as one of the most devastating risks that power systems face. Among these cyberattacks, one of the most insidious is the false data injection attack (FDIA). In these types of attacks, attackers attempt to manipulate the measurement data of the physical system, leading operators to make erroneous decisions within the energy management system, relying on falsified information. Hence, the identification of such attacks holds paramount importance. Additionally, prior research has demonstrated that traditional detection systems, such as bad data detection (BDD), commonly employed in power systems, can be circumvented by attackers. Therefore, it is crucial to introduce contemporary methods to safeguard power systems against cyber threats. In this context, numerous studies have been conducted in attack detection; however, there remain certain limitations. This chapter aims to give brief information about FDIA and also addresses existing detection mechanisms by implementing machine learning techniques, which have the potential to significantly improve the accuracy and efficiency of FDIA detection.