
The case
As part of a research program conducted with Ford Motor Company and the developers of LS-DYNA (LSTC), Dynas+ developed and validated several modeling strategies to simulate the mechanical behavior of Li-Ion cells during an automotive crash. The goal was to faithfully reproduce the deformation mechanisms leading to internal short circuit while developing methodologies efficient enough to eventually be integrated into full vehicle crash simulations. The work was then extended to leverage the new features introduced in LS-DYNA R10.1.
The solution deployed
Dynas+ engineers developed several complementary methodologies in LS-DYNA to reconcile physical accuracy with numerical performance:
- comparison of solid element and Thick Shell formulations for representing multilayer stacks;
- optimization of the mesh and element aspect ratio to substantially reduce computation times;
- development of an indentation model reproducing the experimental tests on Li-Ion cells;
- implementation of new contact strategies (MORTAR) to improve stability under very high compression;
- evaluation of an approach based on PART_COMPOSITE_TSHELL, allowing the dozens of internal layers to be represented through the integration points of the thick elements, with strong potential for reducing computation costs.

The results achieved
This work made it possible to:
- identify the modeling strategies best suited to multilayer Li-Ion cells;
- significantly improve the robustness of indentation simulations thanks to MORTAR contacts;
- demonstrate the potential of Thick Shell formulations to reduce computation times while maintaining representative mechanical behavior;
- pave the way for the future integration of battery models into multiphysics vehicle crash simulations.
This work was the subject of two scientific papers presented at the 11th European LS-DYNA Conference (Salzburg): "Mechanical modelling Li-Ion cell crush experiments using LS-DYNA" and at the 15th International LS-DYNA Users Conference (Detroit): "Li-Ion Battery Modeling strategies for EV crash applications" (Matthieu Seulin, Charlotte Michel, Vincent Lapoujade, Pierre L'Eplattenier).



