Performance Improvement of EV Charger Using AI Based Controllers

Abstract

The rapidly growing demand of electric vehicle(EV) has a detrimental impact on power quality of the utility distribution grid. The choice of EV battery charger design is crucial in reducing power quality issues. Conventional diode bridge rectifier(DBR) based converters introduce significant harmonics which also adversely affects the battery life. A DBR interfaced with a zeta converter can reduce the power supply distortions and improve the input power factor during battery charging. PI controllers are comprehensively applied in battery chargers; nevertheless, their performance tends to suffer under system dynamics. The diversity of electric vehicles (EVs) and the dynamic changes in battery operating conditions introduce numerous uncertainties, making an artificial intelligence (AI) based controller a preferable choice for EV charging applications. The performance of an isolated zeta converter with PI, fuzzy logic and Artificial Neural Network(ANN) based controllers is tested in MATLAB/Simulink environment under different operating conditions. Analysis of the results clearly indicates that AI based controllers are not only effective in reducing harmonics but also in ensuring improved performance under both steady-state and dynamic conditions.