基于改进UKF算法的动力电池SOC在线估计

Online Estimation of SOC in Power Cell Based on Improved UKF Algorithm

为了更准确的估计电池荷电状态(SOC),我们考虑了影响SOC估计精度的主要因素和传统SOC估计方法的缺点,运用递推最小二乘法进行电池模型参数辨识,实现了模型参数的实时修正,提高了模型精度。我们采用次优无偏MAP时变估计器对噪声协方差矩阵进行实时更新,结合无迹卡尔曼滤波算法进行SOC在线估计,改进了传统UKF算法,降低了噪声对SOC估算的影响。试验和仿真结果表明,改进后的UKF算法相比传统UKF算法具有更高的估算精度。

In order to estimate the state of charge of the battery, the main factors affecting the accuracy of SOC estimation and the shortcomings of the traditional SOC estimation method are considered. Recursive least square (RLS) method is used to do the online identification of the parameter of the battery model, which can realize the real time correction of model parameters and improve the accuracy of the model. The suboptimal unbiased MAP time-varying estimator is adopted in this paper to real time update the noise covariance matrix. And it’s combined with the Unscented Kalman filter (UKF) for SOC online estimation, which improved traditional UKF algorithm and reduced the impact of noise on SOC estimation. Finally, the experiment and simulation show that the improved UKF has better accuracy in SOC estimation than the traditional UKF estimation.