柴油乘用车实际道路排放的多项式回归模型  

Polynomial Regression Model of A Diesel Car in Real World Emission

作　　者：胡志远[1] 秦艳[1] 孙鹏举[1] 谭丕强[1] 楼狄明[1] 

机构地区：[1]同济大学

出　　处：《汽车技术》2014年第10期13-17,35,共6页Automobile Technology

摘　　要：以车载排放试验数据为基础,对某柴油乘用车实际道路行驶速度、加速度、比功率、发动机转速、负荷、空燃比、排气流量、排气温度、燃油消耗量等影响因素进行了主成分分析,所提取主成分的累积方差贡献率达95%。通过多项式回归建立了柴油乘用车实际道路THC、CO、CO2、NOx排放与综合影响因素之间的回归模型。结果表明,建立回归模型的THC、CO、CO2和NOx排放预测误差分别为1.9%、6.2%、2.4%和2.7%,所建立模型的准确性较好。Based on vehicle emission test data, we analyze the factors affecting emission of a diesel car in real world driving conditions such as vehicle speed, acceleration, vehicle specific power（VSP）, engine speed, load, air-fuel ratio, exhaust flow, exhaust temperature, fuel consumption etc. for principal component analysis and the principal components extracted have cumulative variance contribution ratios of up to 95 %. Then we establish a regression model between THC, CO, CO2 and NOxunder real world driving conditions and comprehensive influential factors based on polynomial regression. The results show that the prediction errors of THC, CO, CO2 and NOxin the established model are1.9 %, 6.2 %, 2.4 % and 2.7 % respectively, indicating accuracy of the model.

关 键 词：乘用车 柴油车 道路试验 排放模型 多项式回归 

分 类 号：U467.48[机械工程—车辆工程]