Development of vehicle emission rates based on vehicle-specific power and velocity

Abstract

The US Environmental Protection Agency developed the MOtor Vehicle Emission Simulator (MOVES) operating modes, which defines modal emission rates according to vehicle speed and vehicle-specific power using binning method. However, as MOVES was based on emissions data for vehicle fleets in the US, it is used primarily to estimate US emissions. To adopt this approach in other regions, here, we take into account regional conditions, such as vehicle fleet composition, emissions regulations, and driving environments. Real-world emissions test data for 17 light-duty gasoline and diesel vehicles mainly sold in Korea were used to develop CO2, NOx, and CO emission rates. Typically, the vehicle experiment and data acquisition are costly and time consuming, the amount of data needed to develop robust emission rates were considered. In addition, we studied how a re-binning of vehicle-specific power and velocity could lead to better emission rates estimates from on-road vehicles. To compare the estimates by different binning methods and real-world emissions, root mean square error (RMSE) and R-squared (R2) values were adopted. The comparison result shows that the re-binning method-based emission predictions were more accurate than MOVES prediction results under the real-world condition. The R2 of CO2 and NOx predictions were increased from 075 to 0.78 and from 0.17 to 0.2, respectively. The CO prediction accuracy was slightly increased. These findings provide the re-binning method is advantageous for developing modal-based emission rates using real-world emissions test data.