基于模糊遗忘因子递推最小二乘法锂电池参数辨识方法研究  

Parameter identification method of lithium battery based on fuzzy forgetting factor recursive least square method

作　　者：胡鑫 谭功全 黄波 廖振 罗春兰 HU Xin;TAN Gongquan;HUANG Bo;LIAO Zhen;LUO Chunlan(School of Automation and Information Engineering,Zigong,Sichuan 643000;Key Laboratory of Artificial Intelligence in Sichuan Province,Sichuan University of Science&Engineering,Yibin,Sichuan 644005,China;Yibin Cuiping District Bureau of Agriculture and Rural Affairs,Yibin,Sichuan 644000,China)

机构地区：[1]四川轻化工大学自动化与信息工程学院,四川自贡643000 [2]四川轻化工大学人工智能四川省重点实验室,四川宜宾644005 [3]四川省宜宾市翠屏区农业农村局,四川宜宾644000

出　　处：《内江师范学院学报》2024年第2期64-70,111,共8页Journal of Neijiang Normal University

基　　金：四川省科技厅项目(2020JDJQ0075)。

摘　　要：针对遗忘因子递推最小二乘法的遗忘因子为固定值而造成锂电池参数辨识结果稳定性和锂电池等效电路模型精度无法同时兼顾的问题,提出一种将模糊算法与遗忘因子递推最小二乘法相结合的融合算法——模糊遗忘因子递推最小二乘法,使得遗忘因子在模糊控制器的作用下实现动态可变.通过选取一阶RC模型作为锂电池的等效电路模型,并基于固定遗忘因子递推最小二乘法和模糊遗忘因子递推最小二乘法,分别对一阶RC模型进行参数辨识,然后分别将参数辨识结果带入模型中进行模型端电压误差计算.仿真结果表明,相较于遗忘因子取值为1的固定遗忘因子递推最小二乘算法,基于模糊遗忘因子递推最小二乘法的锂电池模型端电压平均绝对误差值下降了0.00085 V,最大误差绝对值下降了0.0742V.相较于遗忘因子取值为0.9的固定遗忘因子递推最小二乘法,模糊遗忘因子递推最小二乘法参数辨识结果的稳定性有非常显著地提升.Aiming at the problem that the stability of the lithium battery parameter identification results and the accuracy of the lithium battery equivalent circuit model cannot be achieved simultaneously due to the fixed value of the forgetting factor for the forgetting factor recursive least squares method,the paper proposes an integrated algorithm that combines the fuzzy algorithm with the forgetting factor recursive least square method——the fuzzy forgetting factor recursive least squares method——thus to help make the forgetting factor dynamically changeable under the effect of the fuzzy controller.The first-order RC model is selected as the equivalent circuit model of Li-ion battery,and the fixed forgetting factor recursive least squares method and the f uzzy forgetting factor recursive least squares method are both adopted to perform parameters identification for first-order RC models,and then the results of the parameter identification are respectively input into the model for the calculation of the model end-voltage error.The simulation results show that compared with the fixed forgetting factor recursive least squar es algorithm with a forgetting factor value of 1,the average absolute error value for the end voltage of the fuzzy-forgetting-factor-recursive-least-squares-based lithium battery model decreases by 0.00085 volts,and the absolute value of the maximum error decreases by 0.0742 volts.Compared with the fixed forgetting factor recursive least squares algorithm with a value of 0.9,the stability of the parameter identification results of the fuzzy forgetting factor recursive least squares algorithm is significantly improved.

关 键 词：模糊遗忘因子 递推最小二乘算法 锂电池 一阶RC模型 参数辨识 

分 类 号：TM910.1[电气工程—电力电子与电力传动]