Trilevel smart meter hardening strategy for mitigating cyber attacks against Volt/VAR optimization in smart power distribution systems

Smart meter data are crucial information for applications in distribution management system (DMS) to manage the distribution system in a reliable and secure manner. However, the smart meter can be attacked by an adversary through the injection of false data into smart meters. This cyber attack can lead to malfunction of the applications that use smart meter data, thereby yielding an abnormal distribution system operation. This paper presents a multilevel optimization-based smart meter hardening strategy where a defender can mitigate the detrimental impact of cyber attacks on volt/VAR optimization (VVO) in DMS via the manipulation of smart meter data. A trilevel defender–attacker–defender (DAD) model with limited attack and defense resources is proposed to determine an optimal attack mitigation strategy. The proposed DAD model aims to identify smart meters that must be hardened from the VVO attack (upper level) while considering both the worst-case attack against the VVO (middle level) and the VVO during the attack (lower level). A nested column-and-constraint generation algorithm is applied to the DAD model to identify and prioritize the most essential smart meters that should be hardened to mitigate the impact of the VVO attack. An IEEE 33-node distribution feeder was used to quantify the performance of the proposed mitigation framework in terms of voltage deviation, solar photovoltaic (PV) penetration level, attack/defense resources, and computational complexity. The simulation results show that mean percentage errors for voltage deviation of the attack and the defense are calculated as 22.02% and 6.22%, respectively.