Application of Machine Learning Techniques for Predicting False Data Injection Attacks in Smart Grid

Attacks that involve False Data Injection (FDI) provide a substantial risk to the safety and dependability of contemporary power systems. The weaknesses that are present inside the system's infrastructure are exploited by these attacks, which frequently circumvent the conventional measures that detect bad data. Because of the sophistication of these attacks, which are characterized by their ability to circumvent conventional protections and their targeted nature, it is necessary to develop novel techniques to detection and mitigation. For the purpose of identifying, preventing, and predicting FDI threats, this research studies the application of machine learning approaches. In order to improve the security posture of power systems, the strategy that has been developed intends to make use of the potential that data-driven models possess. According to preliminary findings, the approach that was developed demonstrates a high level of accuracy in identifying anomalies within the data of the system, which suggests that it has the potential to be an effective countermeasure against FDI threats. In particular, Extra Trees and XGBClassifier are quite efficient when it comes to identifying attacks that include the injection of bogus data into power systems.