State-of-charge estimation approach of lithium-ion batteries using an improved extended Kalman filter. (February 2019)

• Record Type: Journal Article
• Authors: Yu, Xiaowei; Wei, Jingwen; Dong, Guangzhong; Chen, Zonghai; Zhang, Chenbin
• Is Part Of: Energy procedia. Volume 158(2019)
• Page Start: 5097
• Page End: 5102
• Publication Date: 2019-02
• Subjects: state of charge -- extended Kalman filter -- particle swarm optimization -- battery model; Power resources -- Congresses; Power resources -- Periodicals; Power resources; Conference proceedings; Periodicals; 333.7905
• DOI: 10.1016/j.egypro.2019.01.691
• Languages: English
• ISSNs: 1876-6102

Abstract

Abstract: The extended Kalman filter (EKF) algorithm is generally employed to track SOC rapidly due to the linearization process. However, the traditional EKF suffers from divergence since the inappropriate predetermined noise information. To provide robust estimation results, a novel estimation framework based on the battery model and the improved EKF algorithm is proposed. Firstly, the Akaike information criterion is employed to establish the optimal open circuit voltage-SOC model to improve model accuracy. Secondly, a two-stage estimation algorithm is proposed to improve the robustness of estimation and reduce calculation simultaneously. The joint recursive least square (RLS) algorithm and EKF algorithm are firstly employed for online parameter identification and coarse-grained estimation to track SOC rapidly. Then, the particle swarm optimization algorithm is introduced for fine-grained adjustment to find the global optimal estimation value. Experiments are performed on the lithium-ion batteries to verify the effectiveness of the proposed method. The results indicate that the proposed method can provide a more accurate and robust estimation result and better convergent time than the EKF.