Smart Battery Abnormal High Temperature Fault Diagnosis Based on Internal Temperature Sensing

Temperature is a core monitoring parameter for lithium-ion batteries. Real-time power battery thermal characteristics monitoring and fault diagnosis ensure the reliable and safe battery operation. The novelty of this study is smart battery abnormal high temperature fault diagnosis based on multi-point internal temperature perception results. The fiber bragg grating (FBG) sensor is designed based on the requirements of the battery internal temperature perception and is implanted inside the battery with low damage. The internal abnormal high temperature is induced by discharging an overcharged battery at a high rate, and the internal temperature is sensed and then categorized by clustering algorithm. The importance analysis of features is carried out to select fault diagnostic characteristic parameters, and light gradient boost machine (LGBM) model is applied to realize the diagnosis of internal abnormal high temperature. The results show that LGBM fault diagnosis algorithm significantly outperforms other algorithms with an accuracy of more than 0.989 for abnormal high temperature diagnosis at different positions inside the battery and has low misdiagnosis and missed diagnosis rate of within 2.4% and 4.0%.