Real-time implementation of multi-stage constant current charging with intelligent power management for solar integrated electric vehicle chargers

Conventional electric vehicle (EV) charging methods such as constant current–constant voltage (CC–CV) suffer from prolonged charging time and poor adaptability to solar systems. Moreover, existing solar-powered EV chargers do not integrate multi-stage constant current (MSCC) charging with intelligent power management. This paper presents an intelligent solar-powered EV charging employing a MSCC strategy. In this work, five-stage charging (5-MSCC) is used by varying the charging current based on battery’s state-of-charge (SOC) which overall improves battery health. To meet this dynamic current requirement, a hybrid adaptive Artificial-Bee-Colony Reinforcement-Learning (ABC-RL) controller is designed which regulates the duty cycle of interleaved boost converter (IBC), aligning solar output to demanded battery power in real-time. The controller begins with an intelligent rule-based RL framework that adjusts the IBC duty cycle using power mismatch. To overcome its sensitivity limitations, hybrid ABC-RL algorithm is developed that combines global optimization using honey-bee intelligence with local adaptive tuning through RL-based gain scaling, enabling precise power control across the 5-MSCC charging phases. Here, five charging profiles are designed and examined, and the best-performing profile is selected based on charging time, energy loss, and threshold voltage performance. The system incorporates a dual-phase shift (DPS) controlled dual active bridge (DAB) converter to further reduce current ripples and THD (from 11.78 % to 4.54 %) compared to single-phase shift (SPS) method. MATLAB/Simulink validation through Controller-Hardware-in-the-Loop (CHIL) using OPAL-RT and controller (TI-F28379D) confirm 91–96.7 % efficiency, and stable charging across all stages, proving the effectiveness of the ABC-RL-controlled MSCC strategy for real-time solar-driven EV charging.