Detection Method for Leakage Faults in Lithium-Ion Batteries Based on Electrochemical Impedance

Battery thermal runaway is a critical factor limiting the development of the battery industry. Battery electrolytes are flammable, and leakage of the electrolyte can easily trigger thermal runaway. Currently, the detection of leakage faults largely relies on sensors, which are expensive and have poor detection stability. In this study, firstly, the leakage behavior of lithium-ion batteries is simulated, and the evolution characteristics of the battery's electrochemical impedance spectroscopy (EIS) are analyzed. Then, the battery's EIS curve is fitted using an equivalent circuit model, and the model parameter evolution patterns of normal and leaking batteries are comparatively analyzed. Characteristic parameters are proposed to represent leakage faults. Finally, using random forests for model training and validation, we achieve rapid diagnosis of leakage faults. The leakage fault detection method based on electrochemical impedance spectroscopy improves the reliability of fault detection and is of great significance for diagnosing electrolyte leakage faults in batteries.