Internal short circuit fault diagnosis for automotive lithium-ion batteries using an adaptive parameter identification model

This paper develops a model-based diagnostic method for internal short circuit (ISC) faults in lithium-ion batteries. This method utilizes a second-order equivalent circuit model (ECM) combined with a recursive least squares method with an adaptive forgetting factor for online parameter identification, with initial parameters optimized by a particle swarm optimization (PSO) algorithm. An extended Kalman filter (EKF) is then employed for state estimation to generate residuals, which are analyzed using the refined cumulative sum (CUSUM) statistical method to detect and diagnose ISC faults. Experimental results show that this method effectively identifies ISC faults in automotive power batteries under various initial state of charge (SOC) levels and fault-free conditions. At 100% initial SOC, fault detection times are 104 seconds for a $\text{15}\Omega$ fault, 77 seconds for a $\text{10}\Omega$ fault, and 27 seconds for a $\text{1}\Omega$ fault; at 90% initial SOC, detection for a $\text{10}\Omega$ fault takes 97 seconds, demonstrating high accuracy and robustness.