电器与能效管理技术 ›› 2021, Vol. 0 ›› Issue (10): 78-82.doi: 10.16628/j.cnki.2095-8188.2021.10.013

• 电动汽车充电桩 • 上一篇    下一篇

基于机器学习的电动汽车续航里程预测

李晓宇1, 陈炫锴1, 李嘉栩1, 林梓瀚2   

  1. 1.深圳大学 物理与光电工程学院, 广东 深圳 518060
    2.深圳大学 经济学院, 广东 深圳 518060
  • 收稿日期:2021-06-27 出版日期:2021-10-30 发布日期:2022-01-25
  • 作者简介:李晓宇(1988—),男,副研究员,研究方向为电动汽车能源管理,机器学习、机器视觉在智能车辆系统中的应用技术。|陈炫锴(1999—),男,研究方向为数据格式与数据清洗。|李嘉栩(1999—),男,研究方向为神经网络与多维数据。
  • 基金资助:
    国家自然科学基金项目(51807121);广东省大学生创新创业训练计划项目(S202010590055)

Driving Range Prediction of Electric Vehicles Based on Machine Learning

LI Xiaoyu1, CHEN Xuankai1, LI Jiaxu1, LIN Zihan2   

  1. 1. College of Physics and Optoelectronic Engineering,Shenzhen University,Shenzhen 518060,China
    2. College of Economics,Shenzhen University,Shenzhen 518060,China
  • Received:2021-06-27 Online:2021-10-30 Published:2022-01-25

摘要:

电动汽车动力电池退役高峰下,电池梯次利用备受关注,车辆续航里程是界定电池衰老程度的重要参考指标。为对车辆续航里程及电池健康状况进行准确建模,以电动汽车续航能力为研究对象,在长短期记忆网络基础上,实验分析并确定电池系统电压、电流、温度、SOC等特征参量对神经网络训练结果的影响程度。最终选取等效里程、放电循环耗电量、电压平均值三维特征作为网络模型的输入量,获得最优电动汽车续航能力预测模型。在测试集上预测结果的均方根误差为10.81 km,相对误差9.818%,可为电动汽车能源管理和动力电池性能衰退情况评估提供依据。

关键词: 电动汽车, 锂离子电池, 续航里程, 梯次利用, 长短期记忆网络(LSTM)

Abstract:

At the peak of electric vehicle power battery decommissioning,the gradient utilization of batteries has attracted much attention.The mileage of car is an important index to define the aging degree of the battery.To accurately model the range and battery health of the vehicle,the mileage of the electric vehicle is regarded as the research object.On the basis of long short-term memory network,the influence of characteristic parameters including battery system voltage,current,temperature,SOC on the training result of neural network is analysed.Finally,the three-dimensional characteristics of equivalent mileage,discharge cycle power consumption and voltage average value are selected as the input of the neural network model.The best prediction model for the endurance of electric vehicle is obtained.The root mean square error of the prediction result on the test set is 10.81 km.The method is beneficial to electric vehicle energy management and power battery performance degradation evaluation.

Key words: electric vehicles, lithium-ion battery, vehicle range, gradient utilization, long short-term memory network (LSTM)

中图分类号: