Data integrity attack resilience for electric vehicle charging management centers in distributed optimal power flow with non-ideal Li-ion battery models

Abstract

With the rapid integration of distributed energy resources (DERs), power systems are becoming increasingly vulnerable to cyberattacks, particularly data integrity attacks (DIAs), due to extensive information exchange. Market participants might engage in economic-driven attacks to gain competitive edge and strategic advantages over competitors. Emerging infrastructures, such as Electric Vehicle Charging Management Centres (EVCMCs), have caught increasing attention from attackers, where successful manipulations could lead to significant financial gains or disruptions to the power grid. This paper presents a novel fuzzy-Bayesian attack-resilience mechanism that incorporates a detailed non-ideal Li-ion EV battery model to enhance cybersecurity. A fuzzy inference system (FIS)-based approach is proposed to quantitively evaluate the vulnerability of EVCMCs, and a Bayesian reputation index is introduced to identify and isolate compromised controllers. This scheme more accurately captures real-world battery behaviors, identifies the most vulnerable EVCMCs, and recovers power dispatch against DIAs. According to the simulation results: 1) Compared with traditional methods, the vulnerability of EVCMCs can be assessed quantitatively based on distinct features of each EVCMC. 2) Attackers can achieve greater financial gains and simultaneously diminish competitors' earnings without violating power system operation constraints by exploiting non-ideal battery characteristics. 3) The proposed attack-resilience scheme effectively verifies shared information among neighbors, isolates compromised controllers and recovers optimal power dispatch in the presence of DIAs.