Intelligent Power Management for Hybrid Energy Storage Clusters in EV Charging Stations

The rapid growth of Electric Vehicles (EVs) has highlighted the urgent need for more efficient power distribution at Charging Stations (CS). Traditional Hybrid Energy Storage Systems (HESS) typically rely on two batteries, which limits their ability to adapt to varying energy demands. This research presents an innovative two-cluster HESS configuration, consisting of a main cluster with redox flow batteries and a support cluster with lithium-ion batteries. Leveraging Deep Reinforcement Learning (DRL), the system dynamically adjusts power allocation within and between the clusters based on real-time conditions. A custom reward function is developed to prioritize using Renewable Energy (RE) while ensuring grid stability and user satisfaction. Validated through a comprehensive simulation that mirrors real-world EV charging scenarios, the results show that the cluster-based power management system enhances operational efficiency, reduces costs, and maximizes RE usage. This flexible, adaptive, and sustainable solution provides a superior alternative to traditional HESS systems in EV charging infrastructure.