数据-模型融合驱动的高倍率短时脉冲电池模型  

High-rate short-time pulse battery model driven by data-model fusion

作　　者：要宇辉 孙丙香[1,2] 张慧敏 马仕昌 赵鑫泽 鲁诗默 朱振威 YAO Yuhui;SUN Bingxiang;ZHANG Huimin;MA Shichang;ZHAO Xinze;LU Shimo;ZHU Zhenwei(National Active Distribution Network Technology Research Center,Beijing Jiaotong University,Beijing 100044,China;Key Laboratory of Vehicular Multi-Energy Drive Systems,Ministry of Education,Beijing Jiaotong University,Beijing 100044,China;Chemical Defense Institute,Beijing 100191,China)

机构地区：[1]北京交通大学国家能源主动配电网技术研发中心,北京100044 [2]北京交通大学载运装备多源动力系统教育部重点实验室,北京100044 [3]防化研究院,北京100191

出　　处：《电池》2025年第2期232-237,共6页Battery Bimonthly

基　　金：国家自然科学基金(52177206)。

摘　　要：高倍率短时脉冲工况下,电池的极化特性差异大、温度上升快、极化电压消退不彻底,导致常规等效电路模型仿真效果不佳。参数辨识和分段均方误差分析发现,高倍率脉冲工况下模型在极化消退部分仿真误差较大,导致下一脉冲极化电压初始值失准。提出基于一阶等效电路模型和前馈神经网络的数据-模型融合驱动模型。相较于常规等效电路模型,该模型在20 C的短时脉冲工况下,能更精确地模拟电池的电压响应,均方根误差降低了61.29%。The polarization characteristics of the battery vary significantly,the temperature rises quickly,the polarization voltage is not completely subsided under the high-rate short-time pulse conditions,resulting in poor simulation effects of conventional equivalent circuit models.The parameter identification and segmented mean square error analysis reveals that the model has a large simulation error in the polarization decay part under high-rate pulse conditions,resulting in the inaccurate initial value of polarization voltage for the next pulse.A data-model fusion-driven model based on the first-order equivalent circuit model and feedforward neural network is proposed.Compared to the conventional equivalent circuit model,this model can simulate the voltage response of the battery more accurately under short-time pulse conditions at 20 C,reducing the root mean square error by 61.29%.

关 键 词：锂离子电池 高倍率短时脉冲工况 等效电路模型 前馈神经网络 数据-模型融合驱动模型 

分 类 号：TM912.9[电气工程—电力电子与电力传动]