Grid cyber-physical security: Dynamics-aware attacker modeling for realistic grid setups

Utilities in the U.S. and around the world are increasingly moving toward smart grid technology through broader automation, integration and remote control capabilities. While this transition has enabled grid operators to meet the exponential growth in electricity demand in an economic and reliable manner, it has also introduced inherent cyber vulnerabilities that can undermine the security of the grid. The current security landscape of power systems exhibits a juxtaposition of rapidly expanding threat surfaces against slowly-evolving deployment of defense measures. The divergence between attack and defense capabilities is in part driven by expeditious changes to the grid due to climate change mitigation - movement toward netZero. Since time is limited, we are pushing forward with new technologies such as EVs, heatpumps, batteries (all controllable remotely) without really putting safeguards in place from the viewpoint of grid cybersecurity. Hence, highly coordinated and sophisticated cyber-physical attacks on the power grid are expected and in fact have been growing in number recently. We aim to develop a set of dynamics-aware optimization techniques which will be utilized to (i) design cyber-physical attacks that maximize the instability of the system (attacker) given existing protection schemes (protector) and certain attacker capabilities, and (ii) develop mitigation strategies that can minimize the worst-case damage (defender) in a fast-enough time.