Wavelet Packet Energy Proportion-Based Early Warning for the Failure of Lithium-Ion Batteries

Thermal runaway propagation (TRP) is the most formidable safety concern for lithium-ion battery systems, requiring early warning signals for timely mitigation or prevention. This study utilizes the high-frequency wavelet packet energy proportion (WPEP) to examine acoustic characteristics of various prismatic batteries during the thermal runaway (TR)/TRP processes. Experimental investigations prove that WPEP signal provides a maximum interval of over 540 s, allowing a prompt solution from the battery management systems (BMSs). Additionally, noise reduction techniques, based on driving conditions before the early-warning, are proposed for electric vehicle (EV) applications. Further, quantitative analysis, including WPEP peak value -Q equalization and R ( $\Delta $ WPEP/ $\Delta {t}$ )-Q equalizations to define TR warning thresholds, are conducted. Moreover, the efficacy of the WPEP-changing trend and warning effect are verified through TRP tests, revealing the TRP and combustion mechanism. Furthermore, characterization analysis for the surface and element composition change of the battery materials is carried out before and after the test to verify the acoustic features analyzed for TR mechanism investigation. This study provides a cost-effective and dependable warning method with only one microphone attached to the battery system, which is promising for the active safety of battery systems.