• Proposes two high-efficiency Partial Power Converters (PPCs) for renewable-integrated ultra-fast EV charging. • Single-switch PPC achieves 98.7% efficiency , processing only 18.5% of power. • Multiport DAB PPC achieves 97.9% efficiency with full multi-source integration. • Total converter rating reduced from 1175 kW to ∼250 kW for 400 kW charging. • ML-based predictive control improves transient response with 35% faster settling and 58% lower overshoot . • Thermal stress reduced by 20–25°C , improving reliability. The integration of renewable energy sources with electric vehicle (EV) charging infrastructure requires efficient power conversion systems capable of handling multiple input sources. This paper investigates two partial power converter (PPC) architectures for renewable-integrated EV charging stations: a single-switch switched-capacitor topology and a dual active bridge (DAB) multiport configuration. Unlike conventional fully-rated converters that process 100% of delivered power, the proposed topologies process only 18.5% and 22.3%, thereby reducing component stress and improving efficiency. MATLAB/Simulink simulations indicate peak efficiencies of 98.7% and 97.9% under ideal conditions. A machine learning-based predictive controller is proposed to optimize switching patterns and reduce losses. The study presents simulation results comparing the proposed architectures against conventional approaches, discussing both advantages and practical implementation challenges.
Reddy et al. (Sun,) studied this question.