A computational fluid dynamics based analysis of coupled thermal and electrochemical processes in lithium-ion battery cells

This study presents a computationally efficient lumped electrochemical–thermal solver implemented in OpenFOAM, decomposing cell voltage into open-circuit, ohmic, activation, and concentration overpotentials with Arrhenius temperature dependence and Fickian diffusion in idealized particles. The implicit Backward Euler discretization with Thomas Algorithm achieves 44–53 × speedup over explicit schemes, enabling 100 × larger time steps while maintaining stability. Validation against a commercial benchmark under 4C discharge demonstrates excellent agreement, with mean absolute percentage errors of 0.17% for average state of charge, 0.07% for cell voltage, and 0.39% for total heat generation. Further experimental validation against literature data for a 48-cell battery module confirms the solver’s scalability to system-level simulations (11 h). Parametric analysis across 1C–4C rates revealed that solid-phase diffusion constitutes the primary performance bottleneck, accounting for 63.3%–65.7% of voltage loss and 52%–62% of heat generation. At 4C, cell temperature increases from 293 K to 333 K, highlighting elevated thermal risk. Cooling simulations for 21700 Li-ion cells using two experimentally-reported anisotropic thermal conductivity models reveal that optimal cooling strategy depends on the anisotropy ratio ( k a x i a l / k r a d i a l ): for high anisotropy ( ≈ 27 ), end cooling reduces thermal gradients by 40% compared to side cooling despite lower efficiency (44.9% vs. 54.1%); for moderate anisotropy ( ≈ 14 ), both strategies yield comparable gradients (22–23 K). The work establishes a versatile and validated framework that bridges the gap between high-fidelity electrochemical modeling and scalable pack-level simulations, delivering a physics-informed and computationally robust tool for optimizing battery thermal management in electric vehicles.