Convective and radiative heat transfer in a battery pack: modelling, simulation and experimental validation

Thermal management inside a battery pack is a crucial issue because the temperature affects its performances both in the short and in the long run. Battery packs grow in size and energy density as, for example, it is the case of automotive applications. Therefore there is an always stronger need to know the temperature distribution along the stack. Purpose of this work was to reach a deep understanding of the cell-to-cell thermal interactions via convection and radiation, and their dependency with the stack layout and the coolant speed and direction. In this thesis, a thermal model for a battery pack was first developed and implemented in Modelica, a powerful object-oriented programming language. Afterward, an experimental ap- paratus was designed and built. Finally this was used to carry out an experimental campaign in order to validated the model by comparing the measurements with the simulation results. The focus of the modeling phase has been the convective and radiative heat transfer taking place between cells and between the cells and the environment for the case of a four by thirteen cell module. The single cells were modeled with the lumped system analysis, thus having a uniform value of temperature. Several alternative convective heat transfer models were implemented. The view factors between each cell and between cells and walls of the pack were added to the thermal model as well. The simulation results indicated a strong dependency of the convection coefficient, and the temperature trend along the stack, from the air speed and the air flow directions. The arrangement of the cells and the distance between them, both transverse and longitudinal, appeared to have none or little influence on the Nusselt number. Radiation turns out to be very important for a single cell, but almost negligible for big battery pack, with the exception of the cells at the extremities of the stack. The experimental apparatus that was designed consisted in an air duct made of Styrofoam hosting fifty-two dummy cells. These were bodies geometrically similar to the real batteries, but electrically much more simple. An extensive measurement campaign was carried out: the arrangement of the cells, the longitudinal or transverse pitch between them, the air speed, and the power dissipated by the cell were varied. According to the experimental results, the thermal model implemented for the battery pack correctly predicts the temperature of the various cells, often within few degrees Celsius. The biggest discrepancy between measurements and simulations results, that cannot be attributed to the experimental setup, is the different slope of the temperature trend along the stack. This is thought to be caused by the use of a global Nusselt number for the entire stack, as suggested by the correlations found in the literature, rather than local values of convection heat transfer coefficient.