基于多新息理论的电池模型参数辨识及SOC估计  

SOC estimation and parameter identification of battery model based on multi-innovation theory

作　　者：聂伟民 乔文超 李鑫 NIE Weimin;QIAO Wenchao;LI Xin(th Research Institute of China State Shipbuilding Co.,Ltd.,Shanghai 201100,China)

机构地区：[1]中国船舶集团有限公司第七二六研究所,上海201100

出　　处：《电气应用》2024年第6期16-23,共8页Electrotechnical Application

摘　　要：准确的锂离子电池荷电状态(SOC)估计对于保障电池管理系统(BMS)安全稳定运行具有重要意义。以三元锂离子电池为对象进行电气特性测试,基于此数据建立一阶RC等效电路模型,再采用引入遗忘因子的多新息最小二乘法(FFMILS)进行模型参数在线辨识,联合加权多新息扩展卡尔曼滤波算法(MIEKF)进行锂离子电池的SOC估计。通过美国联邦城市行驶工况(FUDS)对算法进行验证,结果表明FFMILS-MIEKF算法的SOC估计准确性和稳定性相较于传统EKF与UKF算法均有不同程度提高。并且该算法在不同初始SOC值下均能快速收敛,具有较好的鲁棒性,能够适用于实际电池管理系统之中。Accurate lithium-ion battery state of charge(SOC)estimation is of great significance to ensure the safe and stable operation of battery system(BMS).In this paper,the electrical characteristics of the ternary lithiumion battery are tested,and based on this,we established the first-order RC equivalent circuit model for lithium-ion battery.Then we adapted the recursive least squares method with a forgetting factor which introduces the multiinnovation identification theory to identify the battery model parameters online and combine with weighted multiinnovation Kalman filter to estimate the SOC of the battery.The algorithm is verified under FUDS condition.The results show that SOC estimation accuracy and stability of the FFMILS-MIEKF algorithm are improved to varying degrees compared with the traditional EKF and UKF algorithms.Moreover,the algorithm can converge quickly under different initial SOC values,which means good robustness.As a result,the proposed SOC estimation algorithm in this paper can be well applied to actual BMS.

关 键 词：锂离子电池 电池管理系统 多新息理论 参数辨识 SOC估计 

分 类 号：TM912[电气工程—电力电子与电力传动]