基于改进DEKF锂离子电池容量与SOC协估计

Lithium-ion Battery Capacity and SOC Co-estimation Based on Improved DEKF

为解决矿井复杂环境下电机车锂离子电池的过充或过放对电池产生不可逆的容量损耗问题,提出一种基于有限差分双扩展卡尔曼滤波(FDDEKF)矿用电机车锂离子电池容量与荷电状态(SOC)协估计框架。针对扩展卡尔曼滤波器(EKF)线性化误差问题,采用有限差分改进DEKF。基于二阶电阻-电容(RC)模型,利用遗忘因子递推最小二乘法(FFRLS)进行模型参数辨识,基于时变的模型参数,利用改进DEKF对容量和SOC在分离时间尺度上实时在线估计。不同工况下实验结果表明,该方法能有效提高电池SOC的估计精度,与恒流工况相比,该方法在动态电机车工况下更适用,SOC估计均方根误差降低了0.65%,SOC估计最大绝对误差降低至0.38%。

In order to solve the problem of irreversible capacity loss of lithium-ion battery caused by overcharge or overdischarge of lithium-ion batteries for mining electric locomotives in complex mine environments,a co-estimation framework of lithium-ion battery capacity and state of charge(SOC)for mining electric locomotives based on finite difference double extended Kalman filter(FDDEKF)is proposed.Aiming at the linearization error problem of extended Kalman filter(EKF),the finite difference method is used to improved DEKF.Based on the second-order resistance-capacitance(RC)model,the forgetting factor recursive least squares(FFRLS)method is used to identify the model parameters,based on time-varying model parameters,the improved DEKF is used to estimate capacity and SOC online in real-time on a separate time scale.Conduct comparative experiments under different operating conditions,the results show that this method can effectively improve the estimation accuracy of battery SOC,compared with constant current operating conditions,this method is more suitable for dynamic electric locomotive operating conditions,with the root mean square error of SOC estimation is decreased by 0.65%,and the maximum absolute error of S0C estimation is decreased to 0.38%.