Characteristics Analysis and Modeling of Emissions for Bus with Liquefied Natural Gas Fuel System in Real World Driving

Abstract

To estimate accurately the emission rates of buses fueled with a liquefied natural gas (LNG) system, a method capturing the emission characteristics of LNG buses is proposed. In previous studies, vehicle specific power (VSP), a good indicator for instantaneous power demand on engines, has been mostly used in analyzing the emission characteristics of traditional vehicles such as passenger cars, diesel buses and heavy-duty trucks, with less effort on new energy buses, especially LNG buses. Based on the characteristic analyses in this study, it does not accurately describe the emission characteristics of LNG buses if the operating condition is classified only according to VSP. In this paper, we identify that the factors significantly affecting emissions are both speed and VSP using correlation analysis. Then a two-dimensional algorithm is proposed to divide operating conditions into several operating bins based on speed and VSP, and the validity of this method is verified by comparing the measured values of emissions between a training dataset and a test dataset. Based on the analysis of emission characteristics in two-dimensional operating bins, a regression model for estimating emission rates of LNG buses is established. The feasibility and accuracy of the model are validated by using a test dataset. The results indicate that all the relative errors between measured values and estimated values are within 16%, and the values of normalized mean square error (NMSE) are close to zero, indicating that the model established based on both speed and VSP can be applied to the estimation of LNG bus emissions.