Electrochemical-thermal-aging coupled model for parallel lithium-ion battery packs: Evolution of thermal and aging inconsistency under cell-to-cell variations

Cell-to-cell inconsistencies in parallel lithium-ion battery packs, originating from manufacturing and connection variations and progressively amplified during operation, lead to uneven current sharing, thermal behavior, and aging trajectories, thereby reducing pack lifetime and operational safety. To address this issue, this work develops an electrochemical-thermal-aging coupled model incorporating electrochemical reaction kinetics, thermal characteristics, and solid electrolyte interphase (SEI) growth-induced degradation. An engineering-oriented equivalent-loading coefficient is introduced to represent the combined initial effects of capacity, internal-resistance, and connection-resistance variations using a compact branch-level parameter, while a true-parallel electrical constraint enforces a common bus voltage and total-current conservation, allowing branch currents to redistribute dynamically with SOC, polarization, temperature, and aging-induced impedance. Simulations show pronounced operating-condition-dependent thermal inconsistency. At the end of discharge, the maximum difference between the average cell temperatures is 0.10 °C at 0.5C, 3.41 °C at 2C, and 2.96 °C at 4C. Under the separate 10C temperature-field analysis, the overall maximum temperature difference reaches 3.8 °C. At 1C, the temperature difference remains approximately 0.3 °C from −20 to 20 °C but increases to 2.4 °C at 40 °C. During cycling at 1C and 25 °C, the mean SOH decreases to 0.95915 after 500 equivalent cycles, while the cell-to-cell SOH difference increases from 3.54 × 10 −4 at 100 cycles to 6.65 × 10 −4 at 500 cycles, demonstrating the progressive conversion of current and temperature differences into nonuniform SEI growth and capacity degradation. Increasing the cycling rate from 0.5C to 2C increases the final SOH difference by approximately 53%, whereas the 4C condition exhibits a nonmonotonic evolution because the upper-voltage limit attenuates and redistributes branch currents. These results quantitatively demonstrate that thermal-aging coupling forms an operating-condition-dependent feedback pathway rather than a simple monotonic amplification process. The proposed framework therefore provides an efficient means of representing multi-source initial inconsistency while preserving physically realistic parallel current sharing, and offers guidance for identifying highly loaded branches, improving thermal management and balancing control, and enhancing lifetime evaluation of battery systems for new-energy locomotives.