# 基于双时间尺度扩展卡尔曼粒子滤波算法的 电池组单体荷电状态估计

China Mechanical Engineering - - 中国机械工程 -

Cell SOC Estimation of Battery Packs Based on Dual Time-Scale EKPF

LIU Zhengyu1，2 TANG Wei1 WANG Xuesong1 LI Panchun1

1.School of Mechanical Engineering，Hefei University of Technology，Hefei，230009 2.Engineering Research Center of Safety Critical Industry Measurement and Control Technology，

Ministry of Education，Hefei，230009

Abstract： In order to accurately estimate the SOC of the battery packs，a enhance self correcting （ ESC） model was established for the lithium battery packs，and then an average battery model and a SOC difference model for each battery were established according to the ESC model of the lithium bat⁃ tery. The dual ⁃ time ⁃ scale EKPF algorithm was used to estimate the average SOC of the batteries and the differential SOC of each cell，so as to obtain the SOC of each cell in the batteries. The SOC estima⁃ tion experiments of 12 lithium battery series were carried out . The results show that the SOC estimation method based on the dual time ⁃ scale EKPF algorithm may achieve accurate estimate of the cell SOC. And it is proved that the dual⁃time⁃scale EKPE algorithm has higher estimation accuracy than that of the dual time⁃scale EKF algorithm and EKF algorithm.

Key words： extended Kalman particle filter（EKPF）；cell state ⁃ of ⁃ charge（SOC）estimation；dual time⁃scale；battery pack

0 引言

［］ 1载电压法、电化学阻抗谱法、内阻法、神经网络法和基于电池模型的卡尔曼滤波算法 以及状态观

［］ 2⁃3

［］ 4

［］ 5

［］ 6 man filter，REKF）等改进的卡尔曼滤波算法，以及改进的观测器算法，包括H∞观测器算法 、自适应

［］ 7 Luenberger 观测器算法 及 PI观测器算法 等。

［］ 8 ［］ 9但以上方法主要研究如何对单个电池进行精确的SOC估计，而实际应用中电池组中电池单体由于制造工艺的不一致和使用环境的不一致，或多或少会导致单体间SOC的不一致 。在电池组单体

［］ 10

SOC不一致时，如果使用以上方法则只能对电池组单体SOC进行逐个估计，从而导致计算复杂度太高。

1，2 1 1 1

1.合肥工业大学机械工程学院，合肥， 230009 2.安全关键工业测控技术教育部工程研究中心，合肥， 230009

DOI：10.3969/j.issn.1004⁃132X.2018.15.010 开放科学(资源服务)标识码(OSID) ：