A novel model has been developed for the analysis and the evaluation of average vehicle emissions in a real driving cycle (emission factors) from data in an emission database. The model assumes that emission variation can be explained by parameters determined from dynamic vehicle equation and by the frequency of acceleration events at different speeds. Because the number of resulting X-variables is large, and variables are correlated, a regression method based on principal components, the Partial Least Squares (PLS) method actually, has been adopted. In this paper, model potentiality is illustrated by an application to a case study taken from the database built within the UE V Framework Project ARTEMIS. Data are relative to tests performed under hot conditions with a sample of EURO III 1.4-2.0 l gasoline passenger cars. A set of real driving cycles was utilized as representative of urban, rural and motorway operating conditions detected in different European countries. Results for PLS model fit are good for CO\d2, less than sufficient for CO, HC and NOX; this last result, mostly due to data spread out, is analyzed in the paper by estimating the percentage vehicle's effect.
A Novel Statistical Model for the Evaluation of Vehicle Emission Factors. Application to a Euro III Gasoline Car Fleet
Sae Technical Papers
7th International Conference on Engines for Automobile ; 2005
2005-09-11
Conference paper
English
Fleet Evaluation of Alcohol/Gasoline Blended Fuels
Automotive engineering | 1984
|Vektron 6913 gasoline NOx evaluation fleet test program
Automotive engineering | 2001
|Fleet Trials Using Methanol/Gasoline Blends
Automotive engineering | 1980
|Vektron® 6913 Gasoline Additive NOX Evaluation Fleet Test Program
SAE Technical Papers | 2001
|Gasoline and LPG vehicle emission factors in a road test
Automotive engineering | 2009
|