Influence of the convective boundary conditions on thermal modelling of LTO cylindrical cells using a CFD driven approach

Accurate thermal modelling of Lithium-ion batteries is a critical prerequisite for designing safer Battery Thermal Management Systems (BTMS) and preventing hazardous phenomena like thermal runaway or lithium plating. While complex numerical simulations offer detail, computationally efficient models are preferred for real-time control strategies. This study presents a simplified thermal model for a 1500 mAh Lithium-Titanium-Oxide (LTO) cylindrical cell under natural convection, integrating an Equivalent Circuit Model (ECM), the Bernardi equation for heat generation, and empirical Nusselt number correlations. The model was tested under two dynamic current profiles — representing short and long discharge cycles — reaching peak rates of 1C. Results indicated maximum average surface temperatures of approximately 23.3 °C and 24.0 °C, respectively. The simplified model demonstrated high fidelity against experimental data, achieving percentage errors within ±5%. To refine the convective boundary conditions often approximated in simplified models, a detailed Computational Fluid Dynamics (CFD) analysis was conducted, accounting for the complex geometry of the battery holder. Following experimental validation of the CFD results (error within ±5%), a comparative analysis based on the Root Mean Square Error (RMSE) of the dissipated heat rates was performed to identify the optimal correlation. The analysis revealed that distinct Nusselt correlations are required depending on the operational profile: Morgan's correlation minimised the RMSE (6.3%) for the short discharge test, whereas Churchill and Chu's correlation proved superior for the long discharge test (RMSE 5.8%). These findings provide a transferable methodological approach for selecting heat transfer correlations in reduced-order thermal models.