A Bilevel Attacker-Defender Model for Enhancing Power Systems Resilience with Distributed Generation

Abstract

Electric transmission and distribution systems are subject not only to natural occurring outages but also to intentional attacks. These lasts performed by malicious agents that aim at maximizing the load shedding of the system. Intentional attacks are counteracted by the reaction of the system operator which deploys strategies to minimize the damage caused by such attacks. This paper presents a bilevel modeling approach for enhancing resilience of power systems with high participation of distributed generation (DG). The model describes the interaction of a disruptive agent that aims at maximizing damage to a power system and the system operator that resorts to different strategies to minimize system damage. The proposed mixed integer nonlinear programming model is solved with a hybrid genetic algorithm. Results are presented on a benchmark power system showing the optimal responses of the system operator for a set of deliberate attacks. It was observed that the higher the participation of DG the lower the impact of the attacks was. The presence of DG also influenced the optimal strategies of the attacker which in some cases deviated from optimal attack plans to suboptimal solutions. This allows concluding that the presence of DG benefits the power system in terms of less expected load shedding under intentional attacks.