Modelling 18650-type lithium-ion battery fires using surrogate fuels

Abstract

This study proposes a simplified combustion modelling approach for 18650-type lithium-ion battery fires in thermal runaway by using surrogate fuels instead of the complex fuel mixture of the vent gases. Twelve different fuels were compared through 1-D premixed and counterflow diffusion flame simulations to identify the most suitable surrogate fuel for batteries with LCO, LFP, and NMC cathodes. The investigation focused on batteries at 100% state of charge (SOC), which pose a higher risk of thermal runaway and greater subsequent hazards compared to lower SOCs. For 1-D flames, CH${}_4$/CO${}_2$/CO showed the best performance for all the battery chemistries studied. The battery flame was further validated with 3-D simulations and it was shown that the use of surrogate fuel significantly reduced CPU time by at least 55% compared to the two-step kinetic modelling of the vent gas, which is the common approach used in the literature. This efficiency was achieved while maintaining acceptable levels of accuracy, with maximum errors of 12.2% and 9.5% in flame scalar fields and radiative heat flux, respectively. These findings demonstrate that this approach holds promise for improving computational efficiency in CFD simulations of lithium-ion battery fires in thermal runaway, potentially enhancing the prediction of such events.