Buses in urban have environmental problems because they are mostly having higher emission factors and pollution levels. This study analyzed the contributing factors on bus emissions including NOX, CO, HC, and CO2 and further evaluated the impact degree of these factors. A back-propagation neural network (BPNN) was applied, and the results showed that the composition of pollutant emissions for different fuel types was various. BPNN can be utilized to solve the multifactor, uncertainty, and nonlinearity problems without making any prior presumptions about the data distribution. Among them, diesel buses under EURO-IV and EURO-V emission standards were more likely to produce higher emissions of CO and NOX. By contrast, the emission level of CO and NOX for compressed natural gas bus was lower, but the emission level of CO2 and HC was heavier. In this study, nine variables, namely, speed, acceleration, passenger load, past speed, past acceleration, acceleration time, delay time, stops, and location were selected to investigate their effects on bus emissions. The results showed that factors delay time, location, and stops had the strongest impacts on bus emissions. By contrast, bus emissions were not sensitive to past speed and passenger load. In addition, to fully understand the influence of contributing factors, the impact degree of all these factors on bus emissions was summarized in this study.
Keywords: bus emissions; compressed natural gas; contributing factors; fuel type; gas-electric hybrid electric; impact degree.
© 2021. The Author(s), under exclusive licence to Springer-Verlag GmbH, DE part of Springer Nature.