Development Of Meso-level Emission Inventory For Urban Road Networks

With the continuing growth of the automobile ownership, how to reduce vehicle emissions has become a key issue to be addressed urgently in urban transportation system. In China, since there is a lack of an officially recognized vehicle emission model, emission models from other countries are always utilized to establish the emission inventory in order to evaluate regional vehicle emission pollutions and provide analytical basis and data support for developing emission control strategies. Most of existing studies have focused on the macro-level modeling, e.g. quantifying the total regional emissions, while few studies have attempted to develop modeling methodologies for establishing the meso-level link-based emission inventory, which is more sensitive to traffic management strategies. In addition, the discrepancy exists between the results from different emission models, which leads to inaccurate and inconsistent resulting meso-level emission inventory for an urban road network. As such, it is imperative to systematically study the modeling methodologies for establishing the meso-level emission inventory in urban networks in order to reduce the uncertainties in developing the meso-level emission inventory, In this context, on the basis of a comprehensive review of the state-of-the-art, this thesis strives to identify a suitable emission model and determine the corresponding meso-level parameters, and ultimately develop a method of establishing the meso-level emission inventory for urban road networks.First, a comparative analysis is conducted for the modeling theories and meso-level parameters of the well-developed emission models of other countries. Then the parameter calibration for MOBILE and IVE models is performed at the meso-level based on the real-world emission data. In these efforts, the parameter sets of vehicle miles traveled (VMT) are found the key parameters to be calibrated for MOBILE, and the distribution of VSP Bins is found the key parameter to be calibrated for IVE. The differences between predicted and measured values are compared according to the estimation results of different models. As a result, the IVE model is selected as the meso-level emission model for urban road networks. The study shows that the average relative errors of MOBILE are 37.72%,74.04% and 29.52% in estimating the emission factors of NOx,HC and CO respectively, while the corresponding relative errors of IVE are 6.33%,7.23% and 14.72%. The relative errors of IVE model are significantly lower than those of MOBILE model for different road classes.Subsequently, in light of the vehicle type composition in traffic flows, which is another critical meso-level parameter, the study collects the data of vehicle type composition in traffic flows, traffic volume, and the relevant road characteristics by using the video identification technology. Then, an approach to estimate the vehicle type composition is determined utilizing the decision tree algorithm by analyzing the factors of the road capacity, the distance to central city, the quadrant character, the road type, the number of lanes, and the type of land use, which is further validated by using the field surveyed data. The validation results show that the proposed approach is accurate in estimating the fraction of light-duty gasoline vehicles with a relative error lower than 5%. The accuracy decreases for estimating the other vehicle types (the relative errors are still less than 30%) because of the lower absolute values of fractions of these vehicle types.Finally, a case study is conducted to estimate the meso-level emission inventory by applying the above accomplishments on the road network of expressways in Beijing. In order to meet the input requirement of IVE model on the meso-level parameters, this study determines the VSP distributions in different average speeds by analyzing the second-by-second floating car data, analyzes the vehicle technology distribution through the field investigation, and corrects the base adjustment factors based on the measured real-world emission data. This case study shows that the vehicle emissions of CO, HC, and NOx on the expressway network in Beijing are 131.29,4.22, and 20.48 tons respectively. The vehicle emission intensity presents temporally two peaks in the morning and evening, and demonstrates relatively high average emission intensities on the 2nd Ring Road and the 4th Ring Road spatially.