Adaptive Multi-Objective Fast Charging of LiFePO₄ Packs via Hybrid NSGA-III–PSO Optimization

The rapid integration of Lithium iron phosphate (LiFePO₄) batteries into electric vehicles (EV) and energy storage solutions calls for smart charging strategies that optimally balance charging time, energy efficiency, and cycle life. In this paper, multi-objective adaptive multi-stage constant current–constant voltage (MCCCV) charging strategies are formulated for charging LiFePO₄ battery packs in 8S2P configuration. The charging current profile is optimized using a hybrid non-dominated sorting genetic algorithm III (NSGA-III) and particle swarm optimization (PSO) for a five-stage MCCCV charging, subject to voltage and thermal constraints. The optimized five-stage charging profiles have been validated on commercial 3.2 V 50 Ah LiFePO₄ cells. Results show that the optimized adaptive MCCCV charging achieved a 23.7% reduction in charging time over the traditional constant current–constant voltage (CC-CV) charging protocol. In addition, adaptive MCCCV showed 31.4% lower capacity fade after 500 cycles and up to 42% lower temperature rise (limited to less than 8°C), resulting in higher energy efficiency and longer battery service life. The online real-time adaptive algorithm updates the optimal charging current at each stage as a function of the state-of-charge (SOC), temperature, and internal resistance measurements. All percentages are relative to 0.5C CC-CV baseline charging (8S2P pack, 50 Ah prismatic LiFePO₄ cells, charging between 20 and 95% SOC). Contributions include: (i) formulation of a multi-objective pack-level optimization problem across time, efficiency, and degradation concurrently; (ii) development of a hybrid NSGA-III–PSO algorithm that attains state-of-the-art hypervolume metrics; and (iii) deployment of real-time adaptive control using Extended Kalman Filter with online estimation of resistances.