锂离子电池荷电状态全参数自适应估计

Full parameters adaptive estimation for state of charge in lithium-ion batteries

中文关键词  分数阶模型  扩展卡尔曼滤波  荷电状态  初值补偿  自适应估计

英文关键词  fractional-order model  extended Kalman filter  state of charge  initial value compensation  adaptive estimation

中文摘要

      考虑到锂离子电池荷电状态(SOC)估计中, 初始SOC值的不确定性对估计精度有显著影响, 提出了一种融 合初值补偿机制的自适应分数阶扩展卡尔曼滤波(AFEKF)方法. 依据电池的分数阶特性, 构建了一个包含两个恒定 相位单元的分数阶等效电路模型, 并对描述电池充放电全程的分数阶等效电路模型方程进行了离散化处理. 为了 提升SOC估计在复杂工况下的适应性, 采用了线性卡尔曼滤波器对测量方程中的系数进行在线辨识. 此外, 为了解 决离散化状态方程中参数、分数阶阶数、等效电路模型初值以及噪声不确定性问题, 引入了Sage-Husa滤波器和带 有初值补偿的AFEKF方法. 最后, 通过对比实验分析了带有初值补偿的AFEKF与不带有初值补偿的AFEKF的性能 差异, 并在不同工况下进行了带有初值补偿的AFEKF的SOC估计实验. 实验结果表明, 所提出的SOC估计方法在复 杂工况下具有较强的适应性.

英文摘要

      Considering the significant impact of initial state of charge (SOC) uncertainty on estimation accuracy in SOC estimation for lithium-ion batteries, an adaptive fractional-order extended Kalman filter (AFEKF) approach with initial value compensation mechanism is proposed. According to the fractional-order characteristics of batteries, a fractionalorder equivalent-circuit model with two constant phase elements is constructed, and the equation of the fractional-order equivalent-circuit model describing the entire charging and discharging process of battery is discretized. In order to improve the adaptability of SOC estimation under complex operating conditions, the linear Kalman filter is used to identify the coefficients in the measurement equation online. In addition, in order to solve uncertainties in parameters, fractional-order dynamics, initial values of the equivalent circuit models and noises in the discretized state equation, the Sage-Husa filter and AFEKF approach with initial compensation are introduced. Finally, the performance difference between AFEKF with initial value compensation and AFEKF without initial value compensation is analyzed by comparative experiments, and SOC estimation experiments of AFEKF with initial value compensation are carried out under different working conditions. The experimental results show that the proposed SOC estimation approach exhibits strong adaptability in complex operating conditions.