Improved Multi-Innovation Least Squares Method for High-Accuracy Identification of Lumped Parameters in Lithium-Ion Batteries

With the advent of renewable energy and electric vehicles, lithium-ion battery state modeling and online parameter estimation face the challenge of higher accuracy and robustness. This paper proposes an improved Multi-innovation Least Squares (MILS) algorithm for high-precision lumped parameter identification of lithium-ion battery model. The parameters are modified by combining multiple innovations in the time window to enhance the adaptability of the algorithm to nonlinear and dynamic conditions. Experimental validation across three distinct operational regimes demonstrates the Improved MILS algorithm’s superior performance against two benchmark identification methods, particularly during transient conditions. Normalized parametric analysis confirms robust convergence characteristics across varying operating scenarios, verifying the method’s consistency and physical interpretability. The proposed approach offers a good potential to improve the accuracy and reliability of battery management systems.