基于蒙特卡洛和SH-AUKF算法的锂电池SOC估计*
SOC Estimation of Lithium Battery Based on Monte Carlo and SH-AUKF Algorithm
-
摘要: 针对锂离子电池荷电状态(State of charge,SOC)估计精度低的问题,将Sage-Husa自适应算法与无迹卡尔曼滤波算法相结合,提出了一种可以对系统噪声进行不断更新和修正的自适应滤波新算法——SH-AUKF算法。在动态应力测试(Dynamic stress test,DST)工况下,采用无迹卡尔曼滤波(Unscented Kalman filter,UKF)、自适应无迹卡尔曼滤波(Adaptive unscented Kalman filter,AUKF)和SH-AUKF三种算法分别对SOC进行估计。结果表明,SH-AUKF算法估计SOC的误差最小,估计精度最高。与UKF相比,SH-AUKF算法的估计精度提高了45.4%;与AUKF相比,SH-AUKF算法的估计精度提高了14.3%。为了进一步降低噪声干扰的偶然性和突发性对SOC估计的影响,在估计过程中加入了蒙特卡洛采样方法。结果表明,融合了蒙特卡洛方法的SH-AUKF算法估计SOC时,估计误差区间仅为±1×10-3,有效提高了估计精度。Abstract: Aiming at the problem of low estimation accuracy of SOC of lithium battery, a new adaptive filtering algorithm, SH-AUKF algorithm is proposed by combining Sage-Husa adaptive algorithm with AUKF method. SH-AUKF algorithm can update and modify system noise continuously. UKF, AUKF and SH-AUKF algorithms are used to estimate SOC under DST conditions. The results show that SH-AUKF algorithm has the lowest estimation error and the highest estimation accuracy. Compared with UKF, the estimation accuracy of SH-AUKF algorithm is improved by 45.4%. Compared with AUKF, the estimation accuracy of SH-AUKF algorithm is improved by 14.3%. In order to further reduce the influence of accidental and sudden noise interference on SOC estimation, Monte Carlo sampling method is added in the estimation process. The results show that the error range of SH-AUKF algorithm combined with Monte Carlo method is only ±1×10-3, which effectively improves the estimation accuracy.
下载: