带有色观测噪声AUKF的锂离子电池SOC估计

摘要

在实际工况中,电池测量参数易受相关性较强的有色噪声干扰.仅考虑有色观测噪声满足一阶自回归模型,提出一种带有色观测噪声的自适应无迹卡尔曼滤波算法(CM-AUKF).算法对荷电状态(SOC)估计的平均绝对误差为0.0004,均方根误差为0.0003,估计精度和稳定性较高,可克服对系统噪声方差初值敏感的问题,提高SOC估计的自适应能力.%In actual conditions,measurement parameters of batteries were easily disturbed by colored noise,only the measurement noise was considered to satisfy one-order autoregressive model.An algorithm of adaptive unscented Kalman filter with colored measurement noise(CM-AUKF) was proposed.The algorithm could accurately estimate state of charge(SOC) and possess algorithm stability,for the mean absolute error was 0.000 4 and the root mean square error was 0.000 3.The sensitivity of initial covariance value of system noise was solved,the adaptation of estimating SOC was improved.