To encourage a wider use of electric vehicles, Lithium-ion (Li-ion) batteries are required to handle high electric currents, generating great heat loads which deteriorate their performances and lifespan.
Immersion cooling technology is a promising solution in terms of heat transfer performances.
Multi-physics and complex processes govern those systems, from the internal chemistry of Li-ion cells to the heat transfer at the battery pack scale powering the electric engine.
The present work aims to develop numerical models for these cooling systems considering the uncertainties coming from the physical parameters.
Firstly, a fast low fidelity model is developed, including the thermal and electrical phenomena of the immersion cooling problem.
Uncertainty quantification methods like Bayesian calibration and Sensitivity Analysis are applied and coupled with experimental data, providing a deeper knowledge on the overall behavior of the system.
Secondly, a more precise approach is performed using a high fidelity Computational Fluid Dynamics (CFD) model, solving the transient conjugate heat transfer in an immersed battery pack.
Uncertainties coming from the internal resistance of the batteries are taken into account.
A Bayesian calibration of the internal resistance model is performed, using the information provided by experimental data.
This whole process allows to reduce uncertainty by almost 95% for some predicted temperatures of interest and overall to increase significantly the predictive character of the model.
Dr. Elie Solaï
Ingénieur de Recherche, LIFSE