Deep Reinforcement Learning-Based Detection Framework for False Data Injection Attacks in Power Systems

Abstract

Numerous advantages have resulted from the increased integration of cutting-edge technologies in power systems, but it has also brought forth new vulnerabilities, mainly in the form of bogus data injection attacks. The stability and dependability of power systems may be compromised by these assaults, necessitating the creation of efficient detection mechanisms. We provide a unique Deep Reinforcement Learning-Based Detection Framework for False Data Injection Attacks in Power Systems in this academic publication. In order to learn and adapt to dynamic attack patterns, our model makes use of the power of deep reinforcement learning. As a result, it is resilient and able to recognize sophisticated attacks in real-time. We have out extensive tests on a sizable dataset acquired from a realistic power system simulation to assess the efficacy of our proposed framework. With an accuracy score of 97%, precision score of 95%, recall score of 89%, and F1 score of 92% on the test set, the results show how good our model is. The comparison table shows that the proposed framework performs better than a number of current approaches, including Linear Regression, Support Vector Machine, Random Forest, AdaBoost Classifier, and Gradient Boosting Classifier. Our model achieved an impressive ROC curve of 0.99, highlighting its capability to distinguish between normal and adversarial data with high accuracy. The advantages of our proposed model lie in its ability to detect false data injection attacks with high accuracy and its adaptability to evolving attack patterns. Moreover, it demonstrates robustness against adversarial attacks, making it a reliable defense mechanism for modern power systems. Deploying the proposed framework may considerably improve the security and resilience of power systems, assuring the continuation of consumers' access to energy. Hence, our research introduces a powerful Deep Reinforcement Learning-Based Detection Framework for False Data Injection Attacks, contributing a valuable tool for securing power systems against emerging threats. With its remarkable performance and potential for future development, this model represents a crucial step towards establishing cyber-resilient power infrastructures for the years to come.