Finite element modelling of Cell Swelling for Traction Batteries

Abstract

The ongoing electrification of the vehicle fleet creates an increasing demand for high performance traction batteries for electric vehicles. The most common battery type used in electric vehicles today is lithium-ion batteries. During operation the cells in these traction batteries swell. To design a complete battery pack with high volumetric energy efficiency, the cells need to be placed in an optimal manner with minimised margins. In order to achieve this, cell swelling needs to be predicted in the development process. The main purpose of the project is to study methods on how to empirically model the battery cell swelling behaviour using the Finite Element Method. An established Finite Element model on battery cell swelling open possibilities to evaluate the effects swelling has on complete vehicles, e.g. in crash simulations. Thus, numerical studies were performed with Finite Element models and numerical predictions were compared to provided physical test data (no physical tests were performed in this project). Different modelling techniques were developed and evaluated. Two types of cell swelling, and a complete module was modelled. Swelling forces on the battery cell were modelled using thermal expansion of the jelly rolls and gas pressure build-up in the cell was modelled using fluid expansion in the cell casing. A complete module model containing multiple cells was developed to evaluate how the swelling affect the module structure. In addition, studies on how the models can be simplified, to reduce complexity and computational demand, were conducted. The predicted results from the swelling force and gas pressure models coincided well with the physical test data. However, several assumptions were made which need to be validated in future studies. The module model did not coincide well with the available test data. Due to convergence issues some simplifications were made to the model. Thus, further investigation on how to model a complete module is needed. Although some of the models did not yield the desired response, this project is a good foundation for further studies on battery cell swelling models.