Development of Predictive CFD Models Based on an Analytical Approach for Modeling and Simulation of Lithium-on Battery Fires

Lithium-ion batteries play a crucial role in the propulsion of electric vehicles and stationary energy systems. However, due to over-discharge, mechanical damage, or thermal abuse, they can undergo a thermal runaway process, leading to fire incidents. Hence, there is a pressing need to investigate this phenomenon to implement prevention and mitigation strategies. This study focuses on comparing an analytical approach for modeling a lithium-ion battery fire against the experimental characterization proposed by the Federal Institute for Materials Research and Testing (BAM). The investigated battery, consisting of 2 modules, is of NCM111/C type with a SOC of 100% and an energy per module of 2.5 kWh. It was placed in the center of a calorimeter with dimensions of 4 m x 4 m x 2.5 m. The fire was simulated in FDS 6.7.1 using a predefined heating ramp, employing a regular mesh with a size of 0.025 m and a duration of up to 120 s. With this mesh size, 2.5 million cells. The air temperature was measured through a sensor located at 2.1 m from the battery. Simulations based on the analytical model yielded temperatures around 450 °C, which fully correspond to the experimental values presented by BAM. In contrast, simulations conducted by BAM for the same case study in ANSYS Fluent using RANS models (k-ω SST) resulted in temperatures in the range of 800 - 900 °C for points where much lower temperatures were expected. The agreement between the temperatures simulated by this study and the experimental ones is partly attributed to the use of an LES model, which allows for a more accurate representation of flame behavior.