【摘要】 传统无迹卡尔曼滤波（UKF）估算电池的荷电状态（SOC）时，不仅存在其系统协方差矩阵失去正定性的风险，还存在精度不高的问题。以磷酸铁锂电池为研究对象，采用二阶RC等效电路模型，利用离线辨识方法获得电池模型的数据，在Simulink中搭建电池模型。针对传统UKF存在的问题，提出了SVD-UKF算法，考虑到噪声的变化，结合Sage-Husa自适应算法，进一步提出了ASVD-UKF算法。仿真结果表明，采用ASVD-UKF算法相较于采用SVD-UKF算法，估计SOC的平均绝对误差（MAE）由2.023%降低为0.404%，均方根误差（RMSE）从2.088%降低到0.543%，提高了SOC估算的精度。更多还原

【Abstract】 When the conventional unscented Kalman filter(UKF)is used to estimate the State of charge(SOC)of batteries,it not only has the risk of losing the positive quality of the system covariance matrix,but also has the problem of low accuracy. Taking lithium iron phosphate battery as the research object,the second order RC equivalent circuit model is adopted,and the data of the battery model are obtained by off-line identification method,and the battery model is built in Simulink. Aiming at the problems existing in traditional UKF,SVD-UKF algorithm is proposed. Considering the change of noise,combined with Sage-Husa adaptive algorithm,ASVD-UKF algorithm is further proposed. The simulation results show that,compared with SVD-UKF algorithm,the mean absolute error(MAE)of SOC estimation decreases from 2.023% to 0.404%,and the root mean square error(RMSE)decreases from 2.088% to 0.543%,which improves the accuracy of SOC estimation.更多还原