Early Internal Short Circuit Detection and Quantitative Diagnosis of Lithium-ion Batteries Based on Voltage Correlation Coefficient and Extended Kalman Filter

Li-ion batteries are extensively utilized in energy storage and automotive fields due to their high energy density, long lifespan, and low cost advantages. However, thermal runaway caused by internal short circuits in Li-ion battery cells occasionally happens. Early internal short circuit detection and warning are crucial for ensuring the safe and stable operation of lithium-ion battery systems. Accurate battery fault diagnosis techniques play a significant role in both theoretical research and practical applications of battery cells. In this paper, a voltage correlation coefficient and extended Kalman filter-based method for early internal short circuit (ISC) detection and quantitative diagnosis of lithium-ion battery cells is proposed. Firstly, the voltage correlation coefficient between different cells within the battery pack is used to identify normal cells and short-circuited cells. Then, Thevenin models are established separately for normal and short-circuited cells, and model parameters are estimated based on the forgetting factor recursive least square (FFRLS). The extended Kalman filter (EKF) is utilized to conduct the estimation of the state of charge (SOC) for both normal and faulty cells, and the short circuit resistance is quantified by calculating the SOC difference. This allows for quantifying the severity of the fault. The proposed method is applied to battery packs with different degrees of faults, and its effectiveness is demonstrated.