An Adaptive V2G Capacity-Based Frequency Regulation Scheme With Integral Reinforcement Learning Against DoS Attacks

Abstract

Aggregated electrical vehicles (EVs) can flexibly serve grid frequency regulation (FR) with adjustable FR capacity (FRC) via vehicle-to-grid (V2G) technology. Therefore, current research applies integral reinforcement learning (IRL) to FR as it can easily solve difficult modeling and optimization problems. However, communication between EVs is vulnerable to denial of service (DoS) attacks, which can significantly degrade FR performance and even destabilize V2G FR systems. This paper proposes an adaptive V2G FRC-based FR scheme with IRL to improve FR resilience and mitigate FR performance degradation against DoS attacks. The proposed scheme optimizes V2G control by IRL without analytical models, integrating an event-triggered mechanism to reduce communication burden. It adapts V2G FRC to attack intensity to minimize frequency deviation and mitigate the impact of DoS attacks. The analytical estimation of convergence rate establishes the quantitative relationships among control performance, V2G FRC, and attack intensity, thereby deriving an adaptive mechanism. The theoretical analysis of the proposed scheme yields stability conditions that guarantee the FR performance. The simulations were performed on the IEEE 39-bus test system to verify the effectiveness and advantages of the proposed scheme. The results indicate that V2G FRC should be scaled down to mitigate performance degradation when intensive attacks occur.