Detection and localization of early-stage abnormal sounds in Lithium-ion battery thermal runaway

Abstract

Lithium-ion batteries are widely used in energy storage stations and power grid transmission because of their high energy density, fast discharge capability, and low maintenance requirements. However, they are susceptible to thermal runaway, which can result in fires or explosions under extreme conditions, posing a significant challenge to their large-scale deployment. Currently, there is a lack of accurate methods for early warning and localization of thermal runaway at the single-cell level. To address this, we propose an acoustic-based detection and localization system that can effectively detect early-stage abnormal sounds from the battery’s safety valve. Our findings demonstrate that the system can accurately detect and localize these early acoustic signals with high precision. Specifically, it provides an early warning up to 101 s before the safety valve opens and 791 s before full thermal runaway, achieving a localization accuracy of 0.05 m for abnormal sound signals. This work underscores the potential of acoustic-based monitoring for early thermal runaway detection, offering a proactive safety measure for lithium-ion battery applications.