Reinforcement Learning for Enhanced Control of Energy Storage and Vehicle-to-Grid Systems

The increasing penetration of renewable energy sources presents significant challenges to grid stability due to their intermittent and unpredictable nature. This study introduces a novel adaptive control framework that integrates energy storage systems (ESS) with vehicle-to-grid (V2G) technology to enhance grid flexibility and resilience. The proposed framework combines a dynamic reinforcement learning strategy and real-time forecasting models to optimize the coordination between ESS and V2G resources. This integrated approach addresses key grid challenges by providing voltage support, frequency regulation, and peak demand reduction. A distinctive feature of the framework is its adaptive reward-shaping mechanism, which dynamically adjusts reward functions based on real-time grid conditions, improving control performance in rapidly changing scenarios. The model is validated using the IEEE 33-bus system under various renewable penetration scenarios. Results demonstrate superior voltage stability, reduced power losses, and enhanced renewable energy utilization, confirming the potential of the proposed ESS-V2G to improve modern power grid performance.