基于分数阶多新息卡尔曼滤波法的SOC估计

SOC estimation based on fractional-order multiple innovation Kalman filter method

针对电动汽车锂离子电池整数阶模型不能精确反映电池极化反应的问题,提出了一种基于自适应遗传算法(AGA)的分数阶模型,并采用分数阶多新息卡尔曼滤波(FOMIEKF)算法对电池荷电状态(SOC)估计。在二阶RC等效电路模型的基础上建立分数阶模型并用AGA辨识模型参数,然后用FOMIEKF算法进行SOC估计,最后与卡尔曼滤波(EKF)、分数阶扩展卡尔曼滤波(FOEKF)算法进行比较。结果表明,在混合动力脉冲测试下,模型端电压最大误差低于1%,SOC平均误差与最大误差比传统方法分别下降了0.79%、0.95%。因此,基于AGA分数阶模型的FOMIEKF方法可以有效估计SOC。

Aiming at the problem that the integer-order model of electric vehicle lithium-ion battery cannot accurately reflect the battery polarization response,a fractional-order model based on adaptive genetic algorithm(AGA)was proposed,and the fractional-order multiple innovation Kalman filter(FOMIEKF)algorithm was used to estimate the state of charge(SOC)of battery.A fractional-order model was established based on the second-order RC equivalent circuit model,and AGA was used to identify the model parameters.The FOMIEKF algorithm was used to estimate the SOC,and compared with the EKF and FOEKF algorithm.The results show that under the hybrid power impulse test,the maximum error of the model terminal voltage is less than 1%,and compared with the traditional method,the SOC average and maximum error are reduced by 0.79%and 0.95%.The FOMIEKF method based on the AGA fractional-order model can effectively estimate SOC.