Date of Award
Doctor of Philosophy (PhD)
Computer Engineering and Sciences
Veton Z. Kepuska
Samuel P. Kozaitis
Due to the increasing number of cyberattacks on smart grid networks globally in the last decade, maintaining a stable electric power supply has become increasingly challenging. The shift from traditional means of measurement instrumentation to smarter devices in electrical substations is experiencing increasing incidents of intrusion. Defense against those intrusions is now a global trending research topic and is attracting governments’ attention. Various techniques have been developed to mitigate the effect of cyberattacks on national power grid systems. The state estimation technique has proven its capability in detecting random false data-injection attacks (FDIAs). However, attackers may have considerable knowledge of the system, as well as accessibility to smart measurement units, therefore the ability to stay hidden from state estimators. Moreover, it is important to identify the type of data as well as the corresponding buses related to compromising FDIAs to eliminate them. This research presents a physical-based method for the power system model and analysis. The approach considers a correlation-based analysis for cyberattack detection, presenting a formulation of the power system variables’ cross-correlations based on a linearized distribution grid model. The proposed detection approach has the advantage of determining the compromised system parameter, as well as its location in the grid. Multiple simulations were run using the IEEE 33-bus radial distribution system to demonstrate the efficiency of the proposed approach, and the results were compared with benchmark machine learning methods. The results showed better accuracy and other performance indices in detecting FDIAs, with a maximum of 99% accuracy. Further, the proposed method had good success in localizing FDIAs, with a maximum true localization rate of 85%.
Hathah, Anfal Yahya, "Detecting Cybersecurity and Behavioral Anomalies in Smart Grid Substation" (2022). Theses and Dissertations. 1331.