Design and Experimental Validation of Wireless Electric Vehicle Charger Control Using Genetic Algorithms and Feedforward Artificial Neural Network

Abstract

Integrating electric vehicles (EVs) into the transportation ecosystem is crucial for environmental protection. With the increasing demand for sustainable mobility solutions, wireless power transfer (WPT) systems present a promising method to facilitate the adoption of EVs while reducing carbon footprints. This paper presents a control strategy for the primary side of a WPT charger utilizing a genetic algorithm (GA) combined with a feedforward artificial neural network (ANN). The aim is to optimize charging in constant current (CC) mode and enhance energy transmission efficiency. The proposed approach employs a GA to control the WPT charger, enabling real-time adaptation of charging parameters. The ANN estimates the system’s efficiency, ensuring optimal performance during the charging process. The developed control strategy significantly improved energy transfer efficiency and system stability. Simulation results demonstrate the effectiveness of this new approach, achieving an efficiency of 89.32% in challenging situations of loss of communication with the vehicle. To validate the design procedure, an experimental prototype was constructed, operating at an operational frequency of 85 kHz. Experimental results confirm the proposed design methodology.