东北大学学报:自然科学版 ›› 2020, Vol. 41 ›› Issue (11): 1543-1549.DOI: 10.12068/j.issn.1005-3026.2020.11.004

• 信息与控制 • 上一篇    下一篇

基于模型参数在线辨识技术的SOC估算方法

刘芳1, 马杰1, 苏卫星1,2, 何茂伟1   

  1. (1. 天津工业大学 计算机科学与技术学院, 天津300387;2. 北京矿冶科技集团有限公司 采矿冶金过程自动化国家重点实验室, 北京100160)
  • 收稿日期:2019-11-29 修回日期:2019-11-29 出版日期:2020-11-15 发布日期:2020-11-16
  • 通讯作者: 刘芳
  • 作者简介:刘芳(1983-),女,辽宁沈阳人,天津工业大学副教授; 苏卫星(1980-),男,辽宁盘锦人,天津工业大学教授.
  • 基金资助:
    国家重点研发计划项目(2017YFB1103003); 国家自然科学基金青年基金资助项目(61802280,61806143,61772365,41772123); 采矿冶金过程自动化国家重点实验室/北京矿冶过程自动化重点实验室研究基金资助项目(BGRIMM-KZSKL-2019-08); 天津市自然科学基金资助项目(18JCQNJC77200); 天津市教委科研计划项目(2017KJ094).

Model Parameter Online Identification Based SOC Estimation Method

LIU Fang1, MA Jie1, SU Wei-xing1,2, HE Mao-wei1   

  1. 1. School of Computer Science and Technology, Tiangong University, Tianjin 300387, China; 2. State Key Laboratory of Process Automation in Mining & Metallurgy, Beijing 100160, China.
  • Received:2019-11-29 Revised:2019-11-29 Online:2020-11-15 Published:2020-11-16
  • Contact: SU Wei-xing
  • About author:-
  • Supported by:
    -

摘要: 针对遗传算法(genetic algorithm,GA)存在收敛速度慢、易陷入局部最优以及难以实现在线应用的问题,面向如动力电池等效电路模型一类非线性较强、实时性要求高的模型辨识问题,提出一种能够快速缩小搜索空间,且有效避免陷入局部最优的在线快速搜索的优化辨识框架,实现电动汽车动力电池等效电路模型参数在线快速辨识,扩展全局搜索优化算法的应用范围.进一步,将此算法应用于电池剩余荷电状态(SOC)估算问题,提出基于改进GA参数辨识技术的无迹粒子滤波SOC估算方法(IGA-UPF).并将此SOC估算方法与基于最小二乘参数辨识技术的无迹粒子滤波的SOC估算算法(LS-UPF)作比较,结果验证了本文提出的在线快速参数辨识框架具有更好的模型参数辨识精度.

关键词: 参数在线辨识, 遗传算法, 无迹粒子滤波算法, 荷电状态, 电动汽车

Abstract: In view of the problems that genetic algorithm (GA) has slow convergence speed, be easy to fall into local optimum and difficult to realize online application, and that the identification background of power battery equivalent circuit model is with strong nonlinearity and high real-time requirements. An optimized identification framework is proposed that can quickly reduce the search space and effectively avoid falling into the local optimum for online fast search, thus realizing the online fast identification of the parameters of the equivalent circuit model of the electric vehicle power battery, and expanding the application range of the global search optimization algorithm. Further, the proposed algorithm is applied to the state of charge (SOC) estimation, based on the improved GA unscented partical filter (IGA-UPF) is proposed. The SOC estimation method is compared with the SOC estimation method based on least square-unscented partical filter (LS-UPF), which proves that the online fast parameter identification framework proposed has better model parameter identification accuracy.

Key words: parameter online identification, genetic algorithm, unscented particle filter (UPF) algorithm, state of charge (SOC), electric vehicles

中图分类号: