Distribution System Reconfiguration to Mitigate Load Altering Attacks via Stackelberg Games

Abstract: The integration of IoT-controllable devices in power systems (such as smart electric vehicle charging stations, heat pumps, etc.), despite their benefits, raises novel cybersecurity concerns. Vulnerabilities in these devices can be leveraged to launch load-altering attacks (LAAs) that can potentially compromise the safety of power systems. In this paper, we analyze the impact of LAAs on the voltage profile of distribution networks (DNs). We first derive closed-form expressions to quantify the attacks' impact. Using the insights derived from this analysis, we then propose a reactive defense method to mitigate LAAs based on reconfiguring the DNs. We also study optimal defense strategies that are robust to LAAs by exploiting non-cooperative sequential game theory. The proposed solution takes into account the potential uncertainties in the attack localization. Furthermore, we propose a Bayesian optimization (BO) approach to compute the equilibrium of the game, which reduces the computational burden. Our results show that attacks launched on the deepest nodes in the DN have the most detrimental effect on the grid voltage profile. Furthermore, the proposed game-theoretic strategy successfully mitigates the effect of the attack while ensuring minimum system reconfiguration.