新能源汽车行驶里程关联数据属性分析与预测模型研究  

Analysis of the Attributes of the Mileage Correlation Data of New Energy Vehicles and the Study of Prediction Model

作　　者：陈娅鑫 陈昕[1] 张丽 阮永娇 孙承臻 CHEN Ya-xin;CHEN Xin;ZHANG Li;RUAN Yong-jiao;SUN Cheng-zhen(School of Automobile and Traffic Engineering,Liaoning University of Technology,Jinzhou 121001,China;Jinzhou Public Transport Co.LTD,Jinzhou 121000,China)

机构地区：[1]辽宁工业大学汽车与交通工程学院,辽宁锦州121001 [2]锦州市公共交通有限责任公司,辽宁锦州121000

出　　处：《辽宁工业大学学报（自然科学版）》2023年第1期6-10,19,共6页Journal of Liaoning University of Technology(Natural Science Edition)

基　　金：辽宁省先进装备制造业基地建设工程中心项目(LNTH2020122E)。

摘　　要：以新能源汽车监管平台数据为基础,深入分析电池荷电状态(state of charge,SOC)、电池温度、电池电压、电流、车速、总电流、总电压与行驶里程的关联关系,找出关联数据属性,以关联数据为输入参数,研究行驶里程预测。建立多元回归机器学习预测模型,并基于新能源汽车行驶历史数据,构建预测模型的数据集,用python编程,实现行驶里程预测模型机器学习训练和模型检验。结果表明,预测模型预测精度为95%,能够快速准确地预测出行驶里程,可为新能源汽车行驶里程预测提供参考依据,具有实用性和可靠性。Based on the data of the new energy vehicle supervision platform,the correlation between the state of charge(SOC),battery temperature,battery voltage,current,vehicle speed,total current,total voltage and mileage is analyzed in depth,the associated data attributes are found,and the mileage prediction is studied with the associated data as the input parameter.A multiple regression machine learning prediction model is established,and a dataset of the prediction model is built based on the historical driving data of new energy vehicles,and python programming is used to realize the machine learning training and model testing of the mileage prediction model.The results show that the prediction accuracy of the prediction model is 95%,which can quickly and accurately predict the mileage,which can provide a reference basis for the mileage prediction of new energy vehicles,and has practicality and reliability.

关 键 词：新能源汽车 行驶里程 关联分析 机器学习 电池荷电状态 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]