基于遗传算法优化扩展卡尔曼滤波的锂电池SOC估计

Lithium Battery SOC Estimation Based on Genetic Algorithm Optimized Extended Kalman Filter

准确估算电池荷电状态是电动汽车安全、可靠运行的重要保障。传统的卡尔曼滤波估计算法一方面难以克服电池模型精确性和实用性的矛盾,另一方面要求系统噪声矩阵必须服从高斯分布。为了解决上述问题,首先建立基于BIC准则的变阶RC等效电路模型,克服模型精确性和实用性的矛盾;接着采用遗传算法对EKF中的系统噪声矩阵和测量矩阵的协方差进行在线优化,以实现在模型误差最小时对SOC进行在线估计;最后搭建测试平台,验证该算法能够克服由于模型误差和测量噪声的不确定对SOC估计的影响,误差在1. 35%以内,并且具有较高的收敛性和鲁棒性。

Accurately estimating the state of charge of the battery is an important guarantee for the safe and reliable operation of electric vehicles. The traditional Kalman filter estimation algorithm is difficult to overcome the contradiction between the accuracy and practicability of the battery model.On the other hand,the system noise matrix must be required to follow a Gaussian distribution. In order to solve the above problems,a variable-order RC equivalent circuit model based on the BIC criterion is first established to overcome the contradiction between the accuracy and practicability of the model. Then the genetic algorithm is used to optimize the covariance of the system noise matrix and the measurement matrix in the EKF to achieve online estimation of the SOC when the model error is minimum. Finally,the test platform is built to verify that the algorithm can overcome the influence of model error and measurement noise uncertainty on SOC estimation and the error is within 1. 5%,and it has high convergence and robustness.