基于迭代扩展卡尔曼滤波算法的电池SOC估算  

Battery SOC Estimation Based on Iterative Extended Kalman Filter Algorithm

作　　者：吴铁洲[1] 罗蒙[1] 张琪[1] 蒋小维[1] 

机构地区：[1]湖北工业大学,太阳能高效利用湖北省协同创新中心,湖北武汉430068

出　　处：《电力电子技术》2016年第5期95-98,共4页Power Electronics

基　　金：国家自然科学基金(51247004)~~

摘　　要：采用卡尔曼滤波算法估算动力电池的荷电状态(SOC),其估算精度与SOC初值无关,但与动力电池的等效模型有关。为进一步提高SOC估算精度,充分考虑温度对电池模型参数的影响,改进电池的二阶RC等效电路模型,建立电池的非线性状态空间模型;为保证SOC估算结果的收敛性,将迭代滤波理论引入到扩展卡尔曼滤波(EKF)算法中;采用Levenberg—Marquardt(LM)方法优化迭代过程,并将其应用于动力电池SOC的估计。实验结果表明,与EKF和迭代EKF(IEKF)算法相比,采用改进的电池等效模型和优化算法,具有较好的收敛性,且提高了估算SOC的精度。Kalman filter algorithm estimate battery state of charge （SOC）,the estimation accuracy is no related with the initial value of SOC, but related with the battery equivalent model.To enhance the SOC estimation accuracy, fully con- sider the impact of temperature on the battery model parameters ,improve the battery second-RC equivalent circuit model, establish the battery nonlinear state space model.To ensure the convergence of the SOC estimation results, in- troduce the iterative filter theory to extended Kalman filter（EKF） algorithm, use the Levenberg-Marquardt（LM） method to optimize the iterative process,and apply it to the battery SOC estimate.Experimental results show that compared with EKF and iterated EKF（IEKF） algorithm, base on the improved battery equivalent model and optimization algori- thm to estimate the battery SOC, with good convergence, and enhance the SOC estimation accuracy.

关 键 词：荷电状态 状态空间模型 卡尔曼滤波 

分 类 号：TM912[电气工程—电力电子与电力传动]