基于ISSA-KELM的锂离子电池组SOC预测  

SOC Prediction of Li-ion Battery Pack Based on ISSA-KELM

作　　者：张英达 马鸿雁 窦嘉铭 王帅 李晟延 胡璐锦 ZHANG Yingda;MA Hongyan;DOU Jiaming;WANG Shuai;LI Shengyan;HU Lujin(School of Electrical and Information Engineering,Beijing University of Civil Engineering and Architecture,Beijing 100044,China;Institute of Distributed Energy Storage Safety Big Data,Beijing 100044,China;Beijing Key Laboratory of Intelligent Processing for Building Big Data,Beijing 100044,China;School of Geomatics and Urban Spatial Informatics,Beijing University of Civil Engineering and Architecture,Beijing 100044,China)

机构地区：[1]北京建筑大学电气与信息工程学院,北京100044 [2]分布式储能安全大数据研究所,北京100044 [3]建筑大数据智能处理方法研究北京市重点实验室,北京100044 [4]北京建筑大学测绘与城市空间信息学院,北京100044

出　　处：《电源学报》2024年第6期217-224,共8页Journal of Power Supply

基　　金：北京建筑大学博士基金资助项目(ZF15054);北京建筑大学2021年度研究生创新资助项目(PG2021056)。

摘　　要：针对锂离子电池组荷电状态SOC(state-of-charge)难以预测的问题,提出改进麻雀搜索算法ISSA(improved sparrow search algorithm)优化核极限学习机KELM(kernel extreme learning machine)的SOC预测模型。首先,引入Logistic混沌映射改进标准SSA,获取最优种群个体;其次,采用改进算法优化KELM的核函数参数S和惩罚系数C,建立ISSA-KELM预测模型;最后,利用某储能设备的历史数据进行仿真研究,对比分析ELM、KELM和ISSA-KELM模型的预测结果,并选用其他工况数据验证模型的鲁棒性。结果表明,SOC预测均方根误差和平均绝对误差分别减小至2.06%和1.54%,证明所提模型的预测精度提高,具有良好的收敛性、泛化性及鲁棒性。To address the difficulty in predicting the state-of-charge(SOC)of a Li-ion battery pack,an SOC prediction model based on kernel extreme learning machine(KELM)optimized by the improved sparrow search algorithm(ISSA)is proposed.First,Logistic chaotic mapping is introduced to improve the standard SSA and acquire the best population individuals.Second,the improved algorithm is used to optimize the kernel function parameter S and penalty coefficient C of KELM to create an ISSA-KELM prediction model.The simulation is carried out utilizing the historical data from an energystorage device,and the results predicted by ELM,KELM and ISSA-ISSA-KELM models were compared and analyzed.In addition,the robustness of the model was verified using data under other working conditions.Results show that the root mean square error and mean absolute error of predicted SOC decreased to 2.06%and 1.54%,respectively.The proposed model improved the prediction accuracy,and its convergence,generalization and robustness were also satisfying.

关 键 词：锂电池组 荷电状态 核极限学习机 算法优化 

分 类 号：TM912[电气工程—电力电子与电力传动]