基于便携式排放测试系统与BP神经网络的大型客车排放预测  

Emission Prediction of Large Buses Based on Portable Emission Mesurement Systems and BP Neural Network

作　　者：李昌庆 谢小平[1] Li Changqing;Xie Xiaoping(State Key Laboratory of Advanced Design and Manufacturing for Vehicle Body,Hunan University,Changsha 410082)

机构地区：[1]湖南大学汽车车身先进设计制造国家重点实验室,长沙410082

出　　处：《汽车技术》2021年第1期57-62,共6页Automobile Technology

摘　　要：以大型客车为研究对象,在长沙市的不同道路工况下进行了车载排放测试,借助道路测试得到的数据,利用BP神经网络,以逐秒的速度、加速度、比功率和油耗数据为输入,建立CO2、CO和NOx的排放预测模型,并用部分试验数据进行了验证。结果表明,CO2、CO和NOx预测结果的总体相关系数R为0.9167,线性高度相关,在整体误差水平上,CO2、CO和NOx排放因子的相对误差分别为0.0378%、0.3767%、3.7860%,该模型对大型客车尾气排放的预测效果较好。Emission tests are carried out on large buses under different road conditions in Changsha.Based on the data from road tests and BP neural network,the emission prediction model of CO2,CO and NOx is established with the inputs of speed,acceleration,specific power as well as fuel consumption every second,and verified with some test data.The results show that the overall correlation coefficient R of CO2,CO and NOx predictions is 0.9167,which is highly linearly dependent;the relative errors of CO2,CO and NOx emission factors are 0.0378%,0.3767% and 3.7860% respectively.It indicates this model can better predict exhaust emissions of large buses.

关 键 词：车载排放测试 聚类分析 比功率 BP神经网络 排放预测 

分 类 号：U491.9[交通运输工程—交通运输规划与管理]