Modeling Thermal Runaway of Lithium-Ion Batteries at Cell and Module Level Using Predictive Chemistry

Abstract

<div>Thermal runaway of lithium (Li)-ion batteries is a serious concern for engineers developing battery packs for electric vehicles, energy storage, and various other applications due to the serious consequences associated with such an event. Understanding the causes of the onset and subsequent propagation of the thermal runaway phenomenon is an area of active research. It is well known that the thermal runaway phenomenon is triggered when the heat generation rate by chemical reactions within a cell exceeds the heat dissipation rate. Thermal runaway is usually initiated in one or a group of cells due to thermal, mechanical, and electrical abuse such as elevated temperature, crushing, nail penetration, or overcharging. The rate of propagation of thermal runaway to other cells in the battery pack depends on the pack design and thermal management system. Estimating the thermal runaway propagation rate is crucial for engineering safe battery packs and for developing safety testing protocols. Since experimentally evaluating different pack designs and thermal management strategies is both expensive and time consuming, physics-based models play a vital role in the engineering of safe battery packs. In this article, we present all the necessary background information needed for developing accurate thermal runaway models based on predictive chemistry. A framework that accommodates different types of chemical reactions that need to be modeled, such as solid electrolyte interphase (SEI) layer formation and decomposition, anode-solvent and cathode-solvent interactions, electrolyte decomposition, and separator melting, is developed. Additionally, the combustion of vent gas is also modeled. A validated chemistry model is used to develop a module-level model consisting of networks of pouch cells, flow, thermal, and control components, which is then used to study the thermal runaway propagation at different coolant flow rates.</div>