Intelligent and thermally-conscious on-board EV charging using hybrid genetic optimization and neural-adaptive control process

Abstract

The increasing usage of electric vehicles (EVs) has amplified the demand for smart, thermally efficient, and battery-aware onboard charging systems. Conventional charging techniques often do not take into consideration balancing their delivery of energy with thermal stress and battery degradation, which lowers the operational efficiency and useful life of the battery. Current methods primarily focus on an isolated aspect, be it power optimization or thermal optimization; there is no combination of the two to arrive at any adaptive, real-time control based on battery metrics such as health. An integral optimization-control framework is proposed in this work that encapsulates algorithm-driven intelligence and neural adaptation into a single construct for on-board charge EVs. This paper proposes an intelligent and thermally conscious on-board EV charging framework that integrates efficiency-centric optimization using a genetic algorithm (ECO-GA), neural network-based adaptive charging control (NNACC), and metabolic inspired three-stage charging control (MET-C3). The initial phase, known as "ECO-GA" or "efficiency-centric optimization via genetic algorithm", clears a multi-variable fitness function based on charging voltage, current, battery temperature, and cycle life after generating control parameters. Together, this collection greatly improves efficiency in charging, reduces adverse effects caused by heating and lengthens cycles in which batteries are used, paving a strong path towards the charging infrastructure of EVs.