摘要
为提高电池模型参数辨识的准确性和模型的适应性,文章对双极化(DP)电路模型展开研究。首先,根据模型参数不同的时变特性,通过将欧姆电阻的辨识过程分离,减少需要辨识的参数数量,进而降低待辨识参数之间的相互影响,提高递推最小二乘法(RLS)辨识精度的同时减少计算量;其次,根据模型参数在线和离线辨识对不同工况的适应性,提出一种全工况自适应等效电路模型(ECM),进一步提高模型精度;最后,以模型精度和运行速度作为指标,建立模型评价方法,验证了自适应输出ECM的优越性。仿真实验表明,相比于欧姆电阻已知的R-DP在线模型、DP在线模型和DP离线模型,基于RLS的全工况自适应ECM具有更高的精度,能够在精度和速度之间实现更好的平衡。
This study uses a dual-polarization(DP) model to improve the identification accuracy of the battery model parameters and the model adaptability.According to the different time-varying characteristics of the model parameters,the identification process of the ohmic resistance is first separated to reduce the number of parameters that must be identified by the recursive least square(RLS).The mutual influence of the parameters is then reduced to improve the RLS identification accuracy and decrease the amount of computation.Considering the adaptability of the online and offline identifications of the model parameters to different working conditions,a full working condition-adaptive output equivalent circuit model(ECM) is proposed herein to further improve the model accuracy.Based on the model accuracy and the running speed,a model evaluation method is finally established to verify the superiority of the adaptive output ECM.The simulation experiments show that the RLS-based full working condition-adaptive output ECM has a higher accuracy than the R-DP online model with a known ohmic resistance,DP online model,and DP offline model and achieves a better accuracy and speed balance.
作者
郭向伟
王晨
陈岗
许孝卓
GUO Xiangwei;WANG Chen;CHEN Gang;XU Xiaozhuo(School of Electrical Engineering and Automation,Henan Polytechnic University,Jiaozuo 454003,Henan,China)
出处
《储能科学与技术》
CAS
CSCD
北大核心
2023年第10期3230-3241,共12页
Energy Storage Science and Technology
基金
National Natural Science(52177039)
Fundamental Research Funds for the Universities of Henan Province (NSFRF210332, NSFRF230604)
Key scientific research projects of colleges and universities in Henan Province (23A470006)
Science and Technology Research Project of Henan Province (232102240078)。
关键词
电池模型
参数辨识
递推最小二乘法
模型精度
全工况
battery model
parameters identification
RLS
model accuracy
full working condition
