Electrochemical-thermal coupled models of different scale and dimension were developed to examine various short-circuit scenarios. All are based on the work of Newman et al1,2. To facilitate convergence at high discharge currents a radial mass transport coefficient for solution diffusion is included3. Furthermore, different reduction approaches for the computation of the solid diffusion are adapted4.
Simulation results for a simple high current discharge are validated with a surface temperature distribution measurement experiment. Therefore a commercially available 63Ah pouch bag cell with a nickel manganese cobalt oxide cathode is equipped with a matrix of temperature sensors attached to the top surface and discharged at 8 C.
The 2D5and 3D model setup allows analyzing the spatial propagation of the short from its first occurrence until the exothermic side reactions initiate. This is particular interesting for large pouch cell or prismatic battery formats, as they show an inhomogeneous current distribution that affects the temperature development.
Considering that the cell’s horizontal dimensions are large compared to its vertical dimensions, temperature and current are likely to show greater in plane gradients than cross plane gradients. Therefore it is reasonable to start with a model of a single cell layer, or electrode sandwich, consistent of the negative current collector, the anode, the separator, the cathode and the positive current collector. This reduces the computational effort and provides sufficient results for short-circuit scenarios, where all layers are affected in the same way, for example nail penetration of the complete cell.
The model of the entire cell compromises 70 electrode sandwich layers. It allows a more detailed analysis of internal short circuits, like particle induced shorts. Additionally it represents more realistic thermal boundary conditions. Thus potential cooling strategies can be investigated. With models of this scale it is also possible to study the influence of the short location or the number of affected layers on the intensity of the short.
All these models are very sensitive to parameter and setup changes and often lack convergence. This, to some extent, explains the poor reproducibility of experiment results.
Nevertheless, numerical simulation of internal short circuit in large-format battery cells provides good insight into the mechanisms and processes during the abuse and thus helps to evaluate safer battery operation strategies and cell design options. First results of different nail penetration scenarios show that most of the heat in the beginning results from the ohmic heating in the nail. However, the nail soon turns into a heat sink and most heat is produced in the cell domain further away from the nail due to electrochemical heating. A variation of the penetration depths illustrates that an increase in the number of penetrated layers causes a faster temperature rise in the beginning, but results in a lower overall peak temperature. In any case, irrespective of the number of penetrated layers, thermal runaway is triggered. Different cooling strategies are tested on a half penetrated cell. None of them, not even strong cooling, is able to prevent thermal runaway.
References
- M. Doyle, T. F. Fuller and J. Newman, J. Electrochem. Soc., 140, No. 6 (1993)
- D. Bernardi, E. Pawlikowski and J. Newman, J. Electrochem. Soc., 132, No. 1 (1985)
- J. Mao, W. Tiedemann and J. Newman, Journal of Power Sources, 271, 444-454 (2014)
- O. Iliev, A. Latz, J. Zausch, S. Zhang, Berichte des Frauenhofer ITWM, 214 (2012)
- N. P Volk, S. Arnold, A. Jossen, Poster SGBM 2014 (2014)