A novel two-stage game theory-based data-driven coordination control for efficient power allocation and EV charging management

In response to the rising energy demand, electric vehicle (EV) charging stations are increasingly incorporating distributed renewable energy sources (DRES) and battery energy storage systems (BESS) to enhance energy management and reduce reliance on the distribution grid. However, integrating multiple DRES requires advanced coordination strategies to ensure efficient EV charging. This research proposes a two-stage coordination control technique, utilizing game theory and data-driven methods, to optimize power allocation and EV charging. A deep convolutional neural network (DCNN) is used to address uncertainties in energy demand and minimize grid load fluctuations by leveraging Nash equilibrium in a noncooperative environment. The first stage optimizes power allocation from sources like photovoltaic systems (PV), BESS, and EVs, while the second stage efficiently dispatches the optimized power to EVs. Validation through simulations and hardware-in-the-loop (HIL) experiments shows that the proposed method significantly improves grid performance. Compared to existing methods, it reduced grid power fluctuations to 1.8 kW (from 9.2 kW in dynamic pricing schemes and 7.4 kW in rule-based systems) and reduced grid peak-to-average power to 2.8 kW (compared to 20.5 kW and 14.5 kW in dynamic pricing and rule-based systems, respectively). Additionally, the proposed method charged 20 EVs within 24 min, outperforming existing methods, which charged 14 and 12 EVs in 10 and 18 min, respectively. These results confirm the effectiveness and reliability of the proposed data-driven control strategy in optimizing EV charging and energy distribution.