Support Vector Machine Based Lithium-ion Battery Electrolyte Leakage Fault Diagnosis Method

Electrolyte leakage may cause deterioration of lithium-ion battery performance, and may even lead to short circuit and cause serious safety accidents. In order to detect electrolyte leakage in time and improve the safety of lithium-ion battery, it is necessary to explore the leakage fault diagnosis method of lithium-ion batteries. In this paper, we conducted a simulation experiment on electrolyte leakage fault and analyzed the electrochemical impedance spectroscopy of the normal battery and the leakage battery by the distribution of relaxation time method during the aging process. The features of electrochemical impedance spectroscopy and distribution of relaxation time curves are extracted . The sensitivity of different features to the leakage fault is compared. A support vector machine based method is used to classify the batteries. An accurate distinction between normal and leakage batteries is achieved. The accuracy of battery classification under different cycles and SOCs is compared, and the fault diagnosis method for electrolyte leakage of lithium-ion batteries is established.