Experimental and computational study of self-heating ignition and calorimetry of Lithium-ion batteries during storage

Abstract

Fire accidents involving Lithium-ion batteries (LIBs) threaten the safety of their storage facilities, where there are thousands of open-circuit cells stacked together forming ensembles. Ignition of ensembles could be triggered by self-heating but this important ignition phenomenon has received little attention in the literature. A few studies have investigated the self-heating of a single cell, but do not account for the effect of heat transfer. However, a large-size open-circuit LIB ensemble during storage can develop temperature gradients and therefore ignition is affected by both heat transfer and chemistry. In this thesis, I conducted ignition and calorimetry experiments using a commercial type of prismatic LiCoO2 cell to quantify self-heating conditions and find the chemical kinetics and thermal properties. Results show that self-heating ignition is possible when cells are stacked together and that the critical ambient temperature decreases with the number of cells. A computational model, based on open-source code Gpyro, is used to understand and predict ignition in different ensemble sizes and storage conditions. I used both ignition experiments and model predictions to quantify and compare two critical temperatures: the cell thermal runaway temperature defined in standard SAE-J2464, and the critical ambient temperature triggering ignition. I find that the cell thermal runaway temperature is insensitive to size, but the critical ambient temperature decreases with size. This shows that the critical ambient temperature should be used to design safe storage rather than the SAE standard. I further use the experiments and computational model to predict LIB ignition during storage with different states of charge and cathode materials. In order to understand whether the accelerating rate calorimetry (ARC) can properly quantify self-heating ignition, for the first time, I quantify the uncertainty caused by ignoring heat transfer in experiments. ARC can generally measure the onset of self-heating, but underestimate the heat of reaction, the maximum temperature and cannot measure critical ignition temperature. The results in this thesis help improve the safety of the open-circuit LIB storage and provide a scientific understanding of self-heating hazards and guidance for better standards.