Improved Hybrid Electrochemical Model and Variable Forgetting Factor Recursive Least Squares Full-Parameter Identification for Lithium-Ion Batteries

During lithium-ion battery operation, internal parameters are influenced by various factors, including temperature, state of charge (SOC), and operating current, which can significantly degrade modeling accuracy. To improve parameter estimation under complex operating conditions, this study proposes an enhanced hybrid electrochemical model based on the Shepherd–Unnewehr framework. The model integrates a Variable Forgetting Factor Recursive Least Squares (VFFRLS) algorithm to enable online adaptive parameter identification. Simulation results confirm that the proposed model reduces the mean voltage estimation errors to 18.7 mV (BBDST) and 12.7 mV (DST), achieving 47.8% and 61.6% improvements over the original Shepherd–Unnewehr model. These results demonstrate the proposed approach’s effectiveness and feasibility in improving parameter identification accuracy for lithium-ion battery models.