%0 Journal Article
%T 车辆行驶工况的开发和精度研究
%A 高建平
%A 孙中博
%A 丁伟
%A 郗建国
%J 浙江大学学报(工学版)
%D 2017
%R 10.3785/j.issn.1008-973X.2017.10.020
%X 为了反映某一地区车辆的真实油耗和排放，通过实车试验数据的采集，以短行程分析法为主体构建本土化的行驶工况.采用因子分析选取15个表征道路特性参数中最具代表性的3个，结合组合优化算法优化的k均值聚类算法对各运动学片段完成聚类，构建的行驶工况与采用主成分分析和组合优化算法优化的k均值聚类算法及不优化k均值聚类算法分别构建的行驶工况相对实际行驶工况以α'=0.008 33水准的检验无显著差异.利用AVL-Cruise平台搭建的整车模型加载各工况，开展油耗和排放性能仿真实验，其中以实际行驶工况与采用因子分析和组合优化算法优化的k均值聚类算法构建的行驶工况油耗和排放相对误差最小，分别为1.15%、1.17%、1.8%和1%.结果表明，构建的行驶工况能够反映这一地区实际的交通特征，提高了工况拟合精度.</br>Abstract: The localized driving cycle was constructed by using the method of short stroke analysis through the collection of real vehicle test data in order to reflect the real fuel consumption and emission of the vehicle in a certain area. The factor analysis was conducted to select the 15 parameters of three most representative in the characterization of the road. The k-means clustering algorithm of combinatorial optimization algorithm was combined to complete clustering on kinematic sequences. The driving cycle of k-means clustering algorithm of principal component analysis and combinatorial optimization algorithm optimization and not optimization of k-means clustering algorithm that were constructed relative to the actual driving cycle to test the α'=0.008 33 level had no significant difference. AVL-Cruise platform was used to load driving cycle of the vehicle model in order to conduct the fuel consumption and emissions performance simulation experiment. The factor analysis with real driving cycle and combinatorial optimization algorithm was used to optimize k-means clustering algorithm in order to construct minimum fuel consumption and emissions of driving cycle of relative error, 1.15%, 1.17%, 1.8% and 1% respectively. Results show that the driving cycle of the construction can reflect the actual traffic characteristics of this area and improve the fitting precision of the driving cycle.
%U http://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2017.10.020