汽车工程 ›› 2021, Vol. 43 ›› Issue (11): 1710-1719.doi: 10.19562/j.chinasae.qcgc.2021.11.017

• • 上一篇    下一篇

基于DTV-IGPR模型的锂离子电池SOH估计方法

王萍,彭香园,程泽()   

  1. 天津大学电气自动化与信息工程学院,天津  300072
  • 收稿日期:2021-04-07 修回日期:2021-08-02 出版日期:2021-11-25 发布日期:2021-11-22
  • 通讯作者: 程泽 E-mail:chengze@tju.edu.cn
  • 基金资助:
    国家自然科学基金(61873180)

SOH Estimation Method for Lithium-ion Batteries Based on DTV-IGPR Model

Ping Wang,Xiangyuan Peng,Ze Cheng()   

  1. School of Electrical and Information Engineering,Tianjin University,Tianjin  300072
  • Received:2021-04-07 Revised:2021-08-02 Online:2021-11-25 Published:2021-11-22
  • Contact: Ze Cheng E-mail:chengze@tju.edu.cn

摘要:

锂离子电池的健康状态(state of health, SOH)是保障电动汽车安全可靠运行的关键因素。现有的SOH估计方法通常忽略容量衰退过程中能够表征电池老化的温度信息。鉴于此,本文中提出一种基于电池表面温度的差分温度伏安(differential temperature voltammetry, DTV)曲线的获取方法和一种滑动平均(moving average, MA)与卡尔曼滤波(Kalman filtering, KF)结合的滤波方法以提取健康特征。同时利用组合核函数改进了传统高斯过程回归(Gaussian process regression, GPR)算法以拟合电池容量全局衰退和局部波动这两种趋势,从而建立DTV-IGPR电池老化模型进行SOH估计。在两种不同环境温度下获取的牛津数据集和NASA数据集中进行单电池和多电池验证,结果均表明,所提方法具有较高的SOH估计精度和鲁棒性。

关键词: 锂离子电池, 健康状态, 健康特征, 温度, 高斯过程回归

Abstract:

The state of health (SOH) of lithium-ion batteries is a key factor to ensure the safe and reliable operation of electric vehicles. The existing SOH estimation methods usually ignore the temperature information that can characterize battery aging in the process of capacity degradation. In view of this, this paper proposes a method to obtain the differential temperature voltammetry (DTV) curve based on the battery surface temperature and a filtering method combining moving average (MA) and Kalman filtering (KF) to extract the health feature. In addition, the combined kernel function is used to improve the traditional Gaussian process regression (GPR) algorithm to fit the two trends of overall decline and local fluctuations of battery capacity, so as to establish a DTV-IGPR battery-aging model for SOH estimation. Single cell and multi cell verification are carried out using the Oxford and NASA datasets, which are collected at two different ambient temperatures. The results show that the proposed method has high SOH estimation accuracy and strong robustness.

Key words: lithium-ion battery, state of health, health feature, temperature, Gaussian process regression