Enhancing Flexibility in V2B Applications with Renewable Energy Resources

Resumen

The incorporation of EV parking within vehicle-to-building (V2B) frameworks signifies not only a technological evolution but also a pivotal step towards constructing smarter and environmentally friendly urban environments. This initiative actively contributes to the optimization of system resources while also enabling the incorporation of renewable energy resources. In this study, we propose the development of reinforcement learning (RL) algorithms for the management of smart parking lots, aiming to minimize building energy purchases from the grid while ensuring efficient charging of EVs. The proposed methods obtained a 15% to 17% improvement in the evaluation reward in comparison with rule based method as a benchmark. In the realm of grid energy, they saved 9 to 11% in average purchase cost. In essence, these algorithms, after training, make more efficient decisions than more traditional control methods while ensuring electric vehicle (EV) charging.