Integrating Level Shift Anomaly Detection for Fault Diagnosis of Battery Management System for Lithium-Ion Batteries

This study analyzes the mechanism of Internal Short Circuits (ISCs) in Lithium-ion batteries (LIBs) and identifies the factors contributing to their development. A simulation environment has been used to design a custom battery pack with a nominal capacity of $3kWh$ (total pack capacity $62.5~Ah$ ) and conduct various fault simulations to analyze the battery pack’s thermal behaviour under different conditions. The discharge rate is variable, with a value of $0.25C$ for all 18650 LIBs, and the temperature ranges from -20°C to +60°C. The collected temperature data has been preprocessed and analyzed using various anomaly detection algorithms, including the proposed LevelshiftAD method, Isolation Forest, and Elliptical Envelope. The study demonstrates that the proposed LevelshiftAD method performs better in accurately detecting temperature faults at the threshold limit faster than the other anomaly detection methods. These findings highlight the potential of the proposed approach for enhancing the accuracy with 97% and efficacy of fault diagnosis in LIBs.