Estimation of bus emission models for different fuel types of buses under real conditions

Urban buses are heavy vehicles that move frequently throughout the day, and most of them are propelled by heavy-duty diesel engines. For these reasons, they have energy and environmental impacts that should not be ignored. Consequently, the primary objectives of this study were to compare the changes in bus speed, acceleration, and emissions between bus stops, intersections, and road sections by applying statistical methods; and to develop a vehicle specific power (VSP)-based artificial neural network (ANN) model to estimate emissions of CO, HC, NO_X, and CO₂ for four different fuel types of buses including gas-electric hybrid electric buses (GEHE bus), compressed natural gas buses (CNG bus), EURO 4 heavy-duty diesel engine buses (EURO 4 bus), and EURO 5 heavy-duty diesel engine buses (EURO 5 bus). The results of t-tests (with p-values varying between <0.001 and 0.026, which were not >0.050) showed that the differences in emissions between different locations and between different fuel types of buses were all statistically significant. In addition, to evaluate the performance of the proposed method, a polynomial regression model using linear, quadratic, and cubic terms of transient speed and acceleration was utilized for comparison. According to the results, the proposed method had more accurate and reliable estimation, which increased the lower 10% of absolute percentage error (Lower-10% APE) by 65.2%; reduced mean absolute percentage error (MAPE) by 41.4%, root mean squared error (RMSE) by 44.9%, and mean absolute error (MAE) by 43.5%; and increased R-squared from 0.659 to 0.781.