摘要
作为描述汽车行驶工况汽车行驶的速度-时间曲线,体现了汽车道路行驶的运动学特征,是汽车行业一项共性的技术指标,同时也是车辆能耗/排放测试方法和限值标准的重要基础。为了建立反映我国实际道路行驶的测试工况,基于参与数据采集汽车的行驶特征,建立能够合理描述该特征的汽车行驶工况曲线。首先,根据所采集的40多万条汽车行驶数据,通过数据预处理实现车辆行驶工况数据的重构。其次,判断每一个运动学片段起点和终点的特征,将重构后的数据划分成独立的运动学片段。再次,根据每个运动学片段的速度、加速度等参数特征,结合主成分分析、聚类分析等方法建立长度为1300 s的代表性汽车行驶工况曲线。最后,借助误差分析与相关系数检验验证所构造曲线的合理性。基于数据驱动方法得到的汽车行驶工况曲线,有助于相关部门制定更加合理的测试工况,进而提高汽车能耗的管理水平。
The speed-time curve of a vehicle,embodying the kinematic characteristics of road driving,is used to describe the vehicle driving condition.It is a common technical index of automobile industry and an important basis for vehicle energy consumption/emission testing methods and limit standards.In order to establish the test conditions that reflect the actual road driving conditions in China,this paper establishes a driving cycle curve which can reasonably describe the driving characteristics of vehicles participating in data collection. Firstly,based on more than 400,000 pieces of data collected by automobile driving,the reconstruction of vehicle driving cycle data was realized through data processing. Next,the characteristics of the starting point and end point of each kinematic sequence were judged,and the reconstructed data were divided into independent kinematic sequence. Then,according to the parameters such as speed and acceleration of each kinematic sequence,a representative driving cycle curve with length of 1300s is established by combining K-means cluster and PCA. Finally,the rationality of the curve is verified by error analysis and Pearson correlation coefficient test.The driving cycle curve based on the data-driven method is helpful for the relevant departments to formulate more reasonable test conditions and improve the management level of automobile energy consumption.
作者
温晴岚
赵丽雅
张霰月
周陈裕
李鹏
申厚雪
金晶亮
WEN Qinglan;ZHAO Liya;ZHANG Xianyue;ZHOU Chenyu;LI Peng;SHEN Houxue;JIN Jingliang(School of Sciences,Nantong University,Nantong 226019,China;School of Electrical Engineering,Nantong University,Nantong 226019,China)
出处
《系统仿真技术》
2022年第3期162-167,共6页
System Simulation Technology
基金
国家自然科学基金项目(71603135)
江苏省研究生科研与实践创新计划项目(KYCX22_3327)。
关键词
行驶工况
运动学片段
主成分分析
聚类分析
相关性检验
driving cycle
kinematics sequence
principal component analysis
cluster analysis
correlation test
