一种基于恒流充电阶段电压数据的锂离子电池健康状态估计模型

An estimation model of the state of health of lithium-ion batteries based on the voltage data in the constant current charging stage

    摘要
  • 摘要: 为解决储能系统电池管理中健康状态（State of Health, SOH）估计算法需兼顾高精度与轻量化的需求，本研究提出一种局部电压拟合（Local Voltage Fitting, LVF）方法。该方法将等效电路模型参数和待估计SOH值作为联合待定系数，构建电池恒流充电阶段电压与充电时间的映射关系模型，进而通过非线性拟合直接求解SOH估计值。基于18650型磷酸铁锂（LiFePO4）电池的循环寿命实验数据，对所提出模型的性能进行了验证。实验结果表明，该模型展现出了良好的SOH估计效果，最大相对误差（Maximum Relative Error, MRE）为0.009489，均方根误差（Root Mean Square Error, RMSE）为0.001249。与BP（Back Propagation, BP）神经网络和梯度提升回归树（Gradient Boosting Regression Tree, GBRT）算法进行对比分析发现：LVF模型的MRE相较于BP和GBRT算法分别降低了60.27%和56.80%，RMSE分别降低了83.04%和79.12%。此外，还利用了同济大学公开的电池实验数据对该模型进行验证，同样取得了良好的SOH估计效果，并且发现LVF的估计精度不受电池一致性的限制。本文所提出的LVF模型具有轻量可维护且SOH估计精度高的优点，适用于工程实际应用。

     

    Abstract: This paper addresses the application requirements for Lithium Iron Phosphate (LiFePO?) batteries State of Health (SOH) estimation algorithms in energy storage systems, which demand both high accuracy and model simplicity. A Local Voltage Fitting (LVF) model is proposed and validated. First, combining a second-order equivalent circuit model, the basic mathematical expression for SOH estimation is derived based on the relationship between terminal voltage and charging time during the constant-current charging phase of the battery; To enhance the model's applicability for fitting lithium iron phosphate batteries, this study selected segmented voltage data within the State of Charge (SOC) range 0.30, 0.90 from the constant-current charging curves of each cycle, excluding the difficult-to-fit regions at the ends of the curves, thereby establishing the local fitting framework for the LVF model.

     

    • HTML全文
    • 参考文献(0)
    • 相关文章
    • 施引文献
  • 资源附件(0)

/

返回文章
返回