基于TVFRLS和SVD-UKF的锂离子电池SOC估算

SOC estimation for Li-ion battery based on TVFRLS and SVD-UKF

在汽车复杂运行工况下,传统离线参数辨识方法辨识准确度低,无迹卡尔曼滤波(UKF)算法在估计荷电状态(SOC)过程中,协方差矩阵非正定,导致算法估计SOC失败。提出采用时变遗忘因子递推最小二乘法(TVFRLS)与奇异值无迹卡尔曼滤波(SVD-UKF)算法进行联合在线SOC估计,提高复杂工况下算法的准确性与鲁棒性。通过城市道路循环(UDDS)工况对算法进行验证,TVFRLS与SVD-UKF联合算法模拟仿真的最大绝对误差(AEE)为1.31%、平均绝对误差(MEA)为0.56%、均方根误差(RMSE)为0.75%。相较于传统UKF算法,MEA与RMSE分别降低了60.0%和51.9%。

The accuracy of traditional off-line parameter identification methods under the influence of complex vehicle operating conditions was low,the unscented Kalman filter(UKF)algorithm was vulnerable to encounters issues that the non-positive-definite covariance matrix during the estimation of state of charge(SOC),resulting in SOC estimation failures.The joint online SOC estimation was carried out using the time-variable forgetting factor recursive least squares method(TVFRLS)and singular value decomposition unscented Kalman filter(SVD-UKF)to improve the accuracy and robustness of the algorithm under complex conditions.The algorithm was verified by urban dynamometer driving schedule(UDDS).The absolute estimation error(AEE)of the combined algorithm of TVFRLS and SVD-UKF was 1.31%,the mean absolute error(MEA)was 0.56%and the root mean square error(RMSE)was 0.75%.Compared with the traditional UKF algorithm,its MEA and RMSE were reduced by 60.0%and 51.9%.