A Computationally Efficient Multi-Scale Model for Lithium-Ion Cells

Abstract

In this work, a computationally efficient multi-scale and multi-dimensional model is set up to describe the electrochemical, electrical and thermal behavior for a generic pouch cell format. As solving the model in multiple spatial dimensions would require an extensive amount of computational resources, we apply effective spatial discretization techniques, namely the orthogonal collocation and Lobatto IIIA method. In order to reduce the number of electrochemical submodels, a coupling method based on node point interpolation is introduced. The proposed model shows an improvement in solution time by a factor of up to 60 while maintaining its accuracy compared to the finite element method solution. To investigate the spatial accuracy, simulation quantities such as potential distribution and temperature distribution for constant current discharge profiles are examined. With the aid of experimental data gained from Swagelok T-Cells, the model parameters are tuned in for discharge current rates of up to 10C and projected to a 40 Ah cell design. Due to the greatly reduced computational time, the proposed reformulated model can be used for complex physics-based simulations that are typically too computationally expensive with standard modeling approaches such as online estimation and parameter optimization.