A novel early warning method for thermal runaway of lithium-ion batteries based on mechanical stress signal

Thermal runaway constitutes a critical safety hazard in lithium-ion batteries, posing substantial risks to both human and property safety. Therefore, early warning is essential for mitigating these risks. Current monitoring strategies predominantly rely on voltage, current, temperature, and gas signals. However, these parameters exhibit limitations: electrical signals (voltage and current) and temperature demonstrate fluctuations significantly only before thermal runaway, while gas detection becomes feasible only after the opening of safety valve. These constraints result in insufficient warning time. This study systematically investigates the mechanical stress evolution and distribution on the surface of 280 Ah lithium iron phosphate batteries during overheating-induced thermal runaway across various stages of charge. The stress evolution is divided into four stages: exponential growth, linear growth, abrupt drop, and rebound. Subsequently, the underlying mechanisms driving each stage are elucidated through comprehensive analysis of thermal runaway. Notably, we established an innovative early warning method by identifying the transition point from exponential to linear growth, enabling thermal runaway prediction over 30 min in advance, which is also prior to the opening of safety valve. Furthermore, the stress distribution pattern when the safety valve opens is analyzed to provide guidance for optimized pressure sensor placement, thereby reducing implementation costs.