Implementation of solar-powered DC fast charging stations for electric vehicles using neural network-based maximum power point tracking and model predictive control

This paper addresses the surging demand for Electric Vehicle (EV) fast-charging infrastructure driven by governmental incentives, environmental concerns, and economic viability. Traditional grid-based charging stations pose sustainability challenges, contributing to carbon footprint and straining the power grid. To address these issues, this study proposed a novel approach: implementing DC fast charging stations for EVs powered by solar energy. The proposed system featured a step-up DC-DC converter with a tuned input duty cycle to achieve optimal performance. To enhance efficiency further, a neural network (NN) trained with RMSprop was utilized for Maximum Power Point Tracking (MPPT). Additionally, Model Predictive Control (MPC) is employed to regulate the converter's output, ensuring stable and efficient charging. The primary objective of this system was to facilitate quick and consistent EV charging while minimizing adverse impacts on the power grid, especially during peak demand periods. The proposed solution was validated for its performance, stability, consistency, and resilience through comprehensive MATLAB simulations. Hence, this study presented a unique and innovative approach to developing an environmentally friendly and efficient DC electric vehicle charging network, addressing key challenges in the transition to sustainable transportation.