交互式多模型的锂电池荷电状态估计  

Interactive Multi-Model Method for Li-ion Battery State of Charge Estimation

作　　者：贺智佳 陈则王[1] 高唱 孙灿飞[1,2] 王友仁 HE Zhijia;CHEN Zewang;GAO Chang;SUN Canfei;WANG Youren(Nanjing University of Aeronautics And Astronautics, Nanjing 211106, China;Testing and Verification Center, Aviation Key Laboratory of Science and Technology on Fault Diagnosis and Health Management, Aviation Industry Corporation of China, Shanghai 201601, China)

机构地区：[1]南京航空航天大学自动化学院,江苏南京211106 [2]中国航空工业集团有限公司,故障诊断与健康管理技术航空科技重点实验室试验与验证中心,上海201601

出　　处：《测试技术学报》2020年第4期289-294,310,共7页Journal of Test and Measurement Technology

基　　金：航空科学基金资助项目(20183352030);国家自然科学基金资助项目(61371041);中央高校基本科研业务费专项资金资助项目(NS2017019);南京航空航天大学研究生创新基地开放基金立项资助项目(kfjj20190309)。

摘　　要：针对锂电池工作在温度不恒定的环境中时单个模型无法捕捉电池参数动态变化,导致电池模型误差增大,荷电状态估计精度下降的问题,提出使用交互式多模型平方根无迹卡尔曼滤波算法来估计电池荷电状态.以戴维宁电池模型为基础建立了一组电池温度子模型,并实现参数辨识;在平方根无迹卡尔曼滤波的基础上,利用交互式多模型方法估计电池荷电状态,通过实时调整模型概率来实现各个模型间的软切换.实验结果表明,交互式多模型方法融合了不同温度下的电池信息,增强了荷电状态估计的精度和鲁棒性.A single battery model cannot capture the dynamic change of battery parameters in the working environment where the temperature is not constant.Due to the error of the battery model increases and the accuracy of the state of charge estimation decreases.An interactive multi-model square root traceless Kalman filter algorithm was proposed to estimate the battery charge states.Firstly,a set of battery temperature sub-models based on the Thevenin battery model was established and parameter identification was realized.Then,based on SRUKF,the interactive multi-model method was used to estimate the battery SOC,and the soft-switching between the models was realized by adjusting the model probability in real time.The experimental results show that the interactive multi-model method combines battery information at different temperatures,which enhances the accuracy and robustness of SOC estimation.

关 键 词：锂电池 交互式多模型 平方根无迹卡尔曼滤波 荷电状态估计 

分 类 号：TB971[一般工业技术—计量学]