Adaptive Clustering Framework for Energy Optimization and Management in Electric Vehicle Propulsion System
Abstract
The propulsion systems of electric vehicles (EVs) need to have precise energy control and battery healthcare in dynamic operating environments. However, it is difficult for traditional equivalent circuit models to model the nonlinear electrochemical characteristics of batteries mainly when there is dynamic load‐dependent behavior; thus, they provide inaccurate predictions. Thus, this work proposes an Adaptive clustering framework for optimized energy management and battery health monitoring. In this, an adaptive noise‐filtered feature‐driven fuzzy neural network (ANF‐FNN) is introduced to improve the accuracy of battery performance and degradation estimation. Furthermore, a Quantum‐inspired Thermal Enhanced Temporal Particle Clustering approach (QTEPC) is employed to enhance real‐time estimation of State‐of‐Charge (SOC) and State‐of‐Health (SOH), which accounts for analyzing complex relationships with voltage, current, and temperature. Moreover, to improve regenerative braking energy, a novel Resilient Regenerative Dempster–Shafer Gain Scheduling (RRDGS) method is integrated for enabling dynamic energy distribution with minimal wastage. The experimental findings show that it works better with 99.2% accuracy, 0.85 RMSE rate, 0.18% SOC estimation error, lower prediction time and better regenerative energy use which proves its better performance in the automotive industry.