锂离子电池自适应参数辨识与SoC估算研究

Research on adaptive parameter identification and SoC estimation for Lithium-ion battery

传统的锂离子电池SoC估算以固定参数的电池模型为基础,不能反映电池内部的真实状态,进而影响了估算精度。为此,提出一种模型参数自适应辨识与SoC估算相结合的协同估算方法。首先建立电池二阶RC等效电路模型,通过脉冲充放电实验获取开路电压与SoC关系并进行分段线性化,得到模型状态空间表达式。然后采用限定记忆递推最小二乘法进行模型参数自适应辨识与逐步更新,并设计PI观测器以实现SoC估算,两者协同从而提高估算精度。理论分析证明了PI观测器的鲁棒性。实验和仿真结果表明,该算法具有很高的SoC估算精度,估计误差范围在1.5%以内,平均估算误差只有1.28%。

Conventional SoC estimation methods for Lithium-ion battery based on models with fixed parameters can- not reflect the real internal state of battery, thus affects estimation accuracy. Therefore, a synergetic estimation scheme which combines adaptive parameter identification with SoC estimation is proposed. First, second-order RC equivalent circuit model is established, the relationship between open circuit voltage and SoC is obtained through pulse charge/discharge test and then is piecewise linearized, and state-space equation of battery model is got suc- cessively. Then, recursive least-square method with restricted memory is employed for adaptive parameter identifi- cation as well as step-by-step update, a PI observer is also designed for SoC estimation, the both cooperates in or- der to improve estimation accuracy. The robustness of the PI observer is confirmed by theoretical analysis. The ex- perimental and simulative results demonstrate that the proposed algorithm has high estimation accuracy, the SoC es- timation error is managed within 1.5% range, and the mean estimation error is 1.28%.