Maximizing photovoltaic energy harvesting in electric vehicles: an adaptive control approach with NBO-ANFIS MPPT and DNN-based SoC estimation

Abstract

The development of electric vehicles (EVs) has increased demand for advanced Energy Management Systems (EMS) that optimizes energy distribution, intelligently handles multiple power sources. One of the primary issues is the intermittent, unpredictable nature of renewable sources which leads to fluctuating power availability, makes real-time energy balancing difficult. Additionally, EMS must handle nonlinear system behavior, rapid load variations requiring fast, accurate control strategies. Hence, an AI-enhanced EV EMS that uses intelligent algorithms, adaptive control techniques is presented in this work. The proposed solution integrates a solar Photovoltaic (PV) array with voltage boosted using an innovative High Gain Coupled Inductor Extended Boost (HiG-CI-X Boost) Converter before being sent onto a common high-voltage bus powering the Brushless DC (BLDC) motor. In order to ensure high performance, energy flexibility, the architecture also incorporates battery, supercapacitor energy storage devices, which are interconnected via bidirectional DC-DC converters. Deep Neural Network (DNN)-based State of Charge (SoC) estimator is incorporated to predict the battery state. A three-phase Voltage Source Inverter (VSI) converts the DC bus voltage to AC in order topower BLDC motor. Accurate control over torque, speed is ensured by a feedback loop. MATLAB simulation results show that the AI-integrated EMS significantly increases energy efficiency, reduces the load on individual storage units, maximizes transient response. Moreover, the proposed converter results with 97.63% efficiency with 1.46% increase from existing topologies, with NBO-ANFIS MPPT algorithm foreshowing tracking efficiency of 99.85% with settling time at 0.14 s, respectively. The proposed MPPT achieves approximately 9–10% higher tracking efficiency reduces the settling time by nearly 78–80%.The proposed system is a viable option for intelligent, ecologically friendly electric mobility because of its exceptional scalability, robustness.