基于Gassing模型的锂电池SOC估计与参数辨识

LIB SOC estimation and parameters identification based on gassing model

具有非线性、时变性的Gassing锂电池模型具有以下两大特点:模型考虑了析气现象,使其适用范围拓广到SOC>85%的临界情况;模型考虑了温度的动态变化对系统状态变量的影响。针对该模型,基于相关的试验数据,采用双Kalman滤波算法(DKF),同时实现了对模型参数的辨识和对SOC的在线估计。台架试验和实车验证表明,该算法在非临界情况和临界情况下均可以较准确地在线估计SOC(估计误差在4%以内),并且对SOC的初值误差具有较好的鲁棒性。

The nonlinear and time-varying gassing lithium battery model motioned in reference [1] had two main characteristics： the model considered the phenomenon of gassing in the battery to extend its application to the critical situation when SOC was over 85%; the model considered the dynamic change of system temperature and its effect on the system state variables. In view of this model, based on the related tentative data,the double Kalman filtering （DKF） algorithm was used, which implemented the model parameter identification and the SOC online estimation simultaneously. The bench test and real vehicle experiment results indicate that this algorithm may online estimate SOC accurately both in non-critical condition and critical condition （error of estimation less than 4%）, and has a better robustness to the SOC error of the initial value.