Design and simulation of a 5 KW solar-powered hybrid electric vehicle charging station with a ANN–Kalman filter MPPT and MPC-based inverter control for reduced THD

Abstract

This study focuses on the control of an OFF-board electric vehicle (EV) charging station, providing a cost-efficient solution for managing high grid demand periods. By integrating a Kalman filter with Artificial Neural Networks (ANN) for Maximum Power Point Tracking (MPPT), the system optimizes energy capture from photovoltaic (PV) panels, even in severe weather conditions and partial shading. Unlike traditional MPPT methods, which face challenges with multiple peaks in the Power–Voltage (P–V) curve, the hybrid algorithm enhances tracking accuracy, reduces errors, and cuts tracking time by up to 99.93%. This ensures a reliable and sustainable power source for EV charging, reducing grid dependency during peak demand. On the inverter side, an innovative Model Predictive Control (MPC) strategy, using a K+2 step approach, is implemented to efficiently regulate the inverter. The system achieves a Total Harmonic Distortion (THD) of just 0.56%, boosting charging speed while minimizing harmonic distortion costs. Controlled by Texas Instruments’ TMS320F28379D digital signal processor, this system offers stable, low-cost EV charging by prioritizing solar energy use, even under harsh weather conditions, over reliance on grid power.