Sunflower optimization algorithm-based online identification of model parameters and state of charge estimation for vehicles lithium-ion power batteries

• Published in: Journal of power sources Vol. 652; p. 237588
• Main Authors: Zhu, Guangyao, Hu, Minghui, Wang, Xin
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.10.2025
• Subjects: Equivalent circuit model; Extended Kalman filter algorithm; Lithium-ion batteries; State of charge; Lithium-ion batteries; Equivalent circuit model; State of charge; Extended Kalman filter algorithm
• ISSN: 0378-7753
• DOI: 10.1016/j.jpowsour.2025.237588

Battery state of charge (SoC) is a hidden state parameter that cannot be directly measured. It is generally estimated by combining explicit measurable physical quantities with filter/observer information. To ensure the efficient and accurate operation of battery management systems under complex and variable working conditions, this paper designs an improved extended Kalman filtering algorithm, which incorporates the proportional-integral differential control method into the gain matrix of the extended Kalman filtering algorithm, in order to correct the a posteriori estimation of the state variables and improve the performance of the extended Kalman filtering algorithm. On this basis, a model parameter online identification and SoC estimation method for vehicle lithium-ion power battery model using the sunflower optimization algorithm is proposed, and the sunflower optimization algorithm is used to search for the forgetting factor in the online identification method of model parameters and the noise covariance matrix in the SoC estimation method in different operating conditions, so as to explore the variation rules of the forgetting factor and the noise covariance matrix with the external environmental conditions. The simulation and experimental results demonstrate that the proposed method can keep the error of voltage calculation within 19.1 mV and the root mean square error of SoC estimation within 0.77 % under dynamic conditions of various ambient temperatures, suggesting good accuracy, robustness, and potential for practical applications. •The influence of noise covariance on SoC estimation is explored•Optimization method is used to obtain forgetting factors and noise covariance•An improved EKF for estimating SoC of lithium-ion batteries is proposed•The accuracy and robustness of voltage and SoC estimation are improved