Research on Orderly Charging of Electric Vehicles Considering Renewable Energy Consumption

Optimizing and scheduling electric vehicles is crucial for achieving system-wide multi-energy complementarity and economic operation. However, renewable energy's intermittency and the uncertainty of user energy demand result in random fluctuations in the system's supply and demand sides, making it challenging for traditional scheduling methods to adapt to the actual environment's dynamic changes. To address this issue, this paper proposes a real-time dynamic scheduling strategy that takes into account electric vehicles and renewable energy. The strategy minimizes operating costs by considering user charging demand, renewable energy output, and dynamic time-of-use electricity prices in a joint operation model. We use the proximal policy optimization algorithm (PPO) to select appropriate state and action spaces for the real-time scheduling problem and design a reward function to train the agent offline. Finally, we test the strategy using the offline-trained model and observe significant reduction in the operating cost of the distribution network and improved absorption of renewable energy.