ANFIS-Based Control of Resonant Converters for Optimized Charging System of Electric Vehicle (EV) Batteries

Abstract

This paper presents a resonant converter-based Electric Vehicle (EV) battery charging module utilizing Proportional-Integral (PI) and Adaptive Neuro-Fuzzy Inference System (ANFIS) control for an optimized charging system. The EV charging module is integrated with a resonant converter comprising a full bridge, HFTF, and DBR. The module utilizes a Resonant Converter which reduces the switching loss incurred during converter operation at high frequency by offering ZCS or ZVS at the switching time. A standard PI Controller manages the duty ratios of the primary full bridge switches with tuned gains. The CC and CV controllers each have their own PI Controller for current and voltage, respectively. To enhance the performance of the EV System, the standard PI Controllers in both the CC and CV control systems are replaced with ANFIS Controllers which are trained as per the data generated by the CC and CV control using an optimization technique that controls the duty ratio of the switches. The proposed ANFIS-based and PI-based control strategy provides an adaptive and flexible approach to control the battery voltage and current by intelligent adjustment of Constant Current (CC) and Constant Voltage (CV) operation modes and the passive elements switching across specific ranges of State-Of-Charge (SOC) to enhance the performance and safety of the charging system. MATLAB Simulation results demonstrated that the proposed ANFIS-based control reduces current ripple content compared to PI-based control. The ANFIS Controller improves overall battery performance, reliability, and stability, which makes it a better choice for next-generation EV charging systems.