Cyber Attack Detection in PMU Networks Exploiting the Combination of Machine Learning and State Estimation-Based Methods

Todays, the security of industrial systems is one of the most salient problems, in the world. Due to the importance of power systems and some of the unique features of these networks, any disruption or security attacks on these systems can cause irreparable damages. The data output of PMUs can be one of the purposes of hackers to create cyber-attacks on power systems. These days, detection of these attacks has got a general interest among researchers. In general, machine learning-based methods applied on PMU information are among the salient approaches for attack detection. Another method is to check the residue of the estimators and observers. The goal of this paper is to use both techniques to detect cyber- attacks on power grids using the data gathered from PMUs. An algorithm is used to identify the type of attack, for instance denial of service (DoS) or man in the middle (MitM). At the end, the suggested procedure is simulated on a sample IEEE power system and the promising results which approve the efficacy of the approach, are elaborated.