Thermal runaway (TR) refers to a hazardous phenomenon where a chain of exothermic reactions spontaneously increases the temperature of battery cells. This is often triggered by internal malfunctions such as short-circuiting or external occurrences such as nail penetration. Different types of lithium-ion cells behave differently during TR, and even seemingly identical cells tested under the same conditions can show differences in heat generation. Understanding the variation in thermal behaviour statistically is important for battery safety management. However, it requires large-scale experimental testing, which is not only time-consuming but also necessitates rigorous data collection and analysis. Now, Donal Finegan and colleagues across the USA, UK, and France present an open-source dataset, referred to as the Battery Failure Databank, containing hundreds of abuse tests designed to induce TR across various commercial Li-ion cell designs and testing conditions.

In contrast to previous datasets on thermal behaviour during TR, Finegan and colleagues collected their data using a technique called fractional thermal runaway calorimetry. This method measures the total heat released, as well as the fractional breakdown of the heat released through the cell casing and ejected materials during TR. Additionally, the researchers recorded high-speed synchrotron radiography videos to observe the internal dynamic response of cells during TR. The resulting databank provides a valuable resource for evaluating and comprehending battery thermal behaviour. For example, correlations were revealed between the fraction of mass ejected from cells, the total heat output, and the cell energy and power density. This allowed the researchers to explain the thermal variation during TR among cells.

Original reference: J. Power Sources 597, 234106 (2024)