Multi-Source Data-Driven Active Balancing Study of Ultra-Fast Charging Loads

Against the backdrop of rapid advancements in multi-source data fusion and intelligent control technologies, ensuring load balance and safety during ultra-fast charging of electric vehicles has become a critical challenge. This paper focuses on the issue of active load balancing in ultra-fast charging and proposes a safe, lithium-plating-free fast charging strategy driven by multi-source data. The proposed method constructs an equivalent circuit model of the lithium-ion battery electrodes and incorporates a dual closed-loop control structure consisting of an anode potential observer and a charging current controller. This enables accurate perception of battery states and dynamic regulation of charging current, effectively suppressing lithium plating and improving charging efficiency. Experimental results show that under various operating conditions, the average observation error of the anode potential is less than 5 mV, with a maximum error not exceeding 10 mV. During current regulation, the anode potential reaches and stabilizes within +1 mV of the set threshold in 26 s, with a steady-state error below 0.05 mV, and no overshoot or oscillation. Compared to the traditional 1C constant current–constant voltage (CC–CV) strategy, the proposed method reduces charging time by approximately 46%. Post-disassembly inspections of the battery revealed no signs of lithium plating, confirming the method's effectiveness and feasibility in enhancing ultra-fast charging performance, ensuring battery safety, and achieving active load balancing.