基于最小二乘法的锂离子电池参数辨识方法研究

目的 针对使用戴维南等效电路模型对锂电池进行参数辨识不够精确的问题,提出一种二阶 RC 等效电路模 型并对锂电池进行参数辨识。 方法 通过脉冲放电实验得到锂电池的相关数据,在 MATLAB 上使用最小二乘算法 对所建立的二阶 RC 等效电路进行参数辨识,并对不同 SOC(State of Charge)下锂电池各个参数的变化情况进行分 析,通过计算锂电池的端电压来判断参数辨识的精确度,最后将辨识结果与戴维南等效电路模型所辨识的结果进 行对比并分析。 结果 随着锂电池 SOC 下降,锂电池的各个参数会有轻微的波动,在锂电池的 SOC 处在较低的水平 时,锂电池的各个参数变化比较剧烈,这是由于锂电池的化学浓差极化所导致的,当将辨识的参数用来求解锂电池 的端电压时,随着时间的推移,发现锂电池的端电压的误差波动比较稳定,且最大误差不超过 0. 05 V,反观使用戴 维南等效电路模型求得锂电池的端电压误差波动比较大,且最大误差超过了 0. 08 V。 结论 在锂电池参数辨识上 二阶 RC 等效电路比戴维南等效电路更加准确,能够更好地描述锂电池的动静态特性,为后续对锂电池的荷电状 态估计提供了有力的基础。

Research on Parameter Identification Method of Lithium-ion Battery Based on Least Squares Method

Aiming at the problem that the Thevenin equivalent circuit model is not accurate enough to identify the parameters of lithium batteries a second-order RC equivalent circuit model was proposed to identify the parameters of lithium batteries. Methods The relevant data of lithium battery were obtained through a pulse discharge experiment and the least squares algorithm was used on MATLAB to identify the parameters of the established second-order RC equivalent circuit the changes of each parameter of the lithium battery under different states of charge were analyzed the accuracy of parameter identification was judged by calculating the terminal voltage of the lithium battery and the identification results were analyzed and compared with the results identified by the Thevenin equivalent circuit model. Results As the state of charge SOC of the lithium battery decreased the parameters of lithium battery fluctuated slightly. When the SOC of a lithium battery was at a low level the parameters of the lithium battery changed more drastically which was caused by the chemical concentration polarization of the lithium battery. When the identified parameters were used to solve the terminal voltage of the lithium battery it was found that the error fluctuation of the terminal voltages of the lithium battery was relatively stable over time and the maximum error did not exceed 0. 05 V. In contrast the error fluctuation of the terminal voltages obtained by using the Thevenin equivalent circuit model was large and the maximum error was more than 0. 08 V. Conclusion In the identification of lithium battery parameters the second-order RC equivalent circuit is more accurate than the Thevenin equivalent circuit which can better describe the dynamic and static characteristics of lithium batteries providing a strong basis for the subsequent estimation of the state of charge of lithium batteries.