Characterizing Thermal Runaway of Lithium-ion Cells in a Battery System Using Finite Element Analysis Approach 2013-01-1534
In this study, thermal runaway of a 3-cell Li-ion battery module is analyzed using a 3D finite-element-analysis (FEA) method. The module is stacked with three 70Ah lithium-nickel-manganese-cobalt (NMC) pouch cells and indirectly cooled with a liquid-cooled cold plate. Thermal runaway of the module is assumed to be triggered by the instantaneous increase of the middle cell temperature due to an abusive condition. The self-heating rate for the runaway cell is modeled on the basis of Accelerating Rate Calorimetry (ARC) test data. Thermal runaway of the battery module is simulated with and without cooling from the cold plate; with the latter representing a failed cooling system. Simulation results reveal that a minimum of 165°C for the middle cell is needed to trigger thermal runaway of the 3-cell module for cases with and without cold plate cooling. During thermal runaway, the cell heat propagates through all the available thermal paths, including the cooling plates that are in contact with the cold plate, the thermal pads that separate the battery cooling units, and the busbars that connect the terminal tabs of the cells. With the functioning cooling system, heat transfer to the neighboring cells is slower and thus it takes a longer time for these cells to runaway. However, the cold plate cooling could not prevent the thermal runaway of these cells because the cooling capacity of the cold plate is designed only for dissipating the cell ohmic heat during normal module operations. The simulation results demonstrate that the FEA model developed in this study can provide useful information on the cell thermal runaway, such as the onset of thermal runaway, maximum runaway temperature, and thermal propagation from the runaway cell to the neighboring cells. This information can be used to predict gas venting from the cell and to provide guidance in the design of safety features of a battery pack enclosure. The analysis approach developed in this study can also be used for cells with different chemistry if the ARC data are available.