Estimating vehicles emissions at signalized intersections in the highway capacity manual

Over the past decades, motor vehicle volumes have continued to increase at a high rate. As a result, engineers in the transportation field not only need more robust knowledge of traffic operation control and transportation planning, but more attention is also needed to understand and estimate the influences that this increasing volume of vehicles has on the environment, especially the influence on air quality. The EPA has stated that reducing carbon monoxide (CO) from vehicle emissions is the most significant way to control air pollution from the transportation sector. The Highway Capacity Manual is a national and international resource that has become a guideline for evaluating the operation of roadway, transit and pedestrian facilities. The Highway Capacity Manual assesses the operation of a roadway based on the perception of its users. Performance measures are used to describe the traffic operation of the roadway. At present, no measures are provided to describe the operation of the roadway based on environmental impacts. The incorporation of air pollution estimation into the Highway Capacity Manual will allow the roadway’s operation to be assessed both from an operational and environmental aspect, ultimately creating a sustainable development for both transportation and the environment. The objective of this dissertation is to develop MOVES-like estimation models of vehicle emissions for pollutants at a signalized intersection that can be incorporated into the Highway Capacity Manual. “EPA’s Motor Vehicle Emission Simulator (MOVES) is a state-of-the-art emission modeling system that estimates emissions for mobile sources at the national, county, and project level for criteria air pollutants, greenhouse gases, and air toxics.” (EPA, 2014). A thorough understanding is needed about what parameters, and influence of these parameters on vehicle emissions. This dissertation develops two kinds of models in order to estimate emissions caused by on-road vehicles. Two modeling approaches are used to estimate four kinds of emissions including CO, NO, NH3 and NOX separately. The following summarizes the work of this dissertation: The objective of this dissertation is to develop MOVES-like estimation models of vehicle emissions for pollutants at a signalized intersection that can be incorporated into the Highway Capacity Manual. “EPA’s Motor Vehicle Emission Simulator (MOVES) is a state-of-the-art emission modeling system that estimates emissions for mobile sources at the national, county, and project level for criteria air pollutants, greenhouse gases, and air toxics.” (EPA, 2014). A thorough understanding is needed about what parameters, and influence of these parameters on vehicle emissions. This dissertation develops two kinds of models in order to estimate emissions caused by on-road vehicles. Two modeling approaches are used to estimate four kinds of emissions including CO, NO, NH3 and NOX separately. The following summarizes the work of this dissertation: (1) Two modeling approaches are used to estimate vehicle emissions including: multiple linear regression and Artificial Neural Network (ANN). In the multiple linear regression modeling, two different models were developed including one model using operation modes as independent variables and another model using traffic related parameters as independent variables. Both model approaches and independent variables are used to estimate four types of pollutant emissions. Statistically, the emission models using traffic parameters as independent HCM related parameters are capable of providing a better emissions estimate based on the higher R square value. For CO, the variables found to be significant were volume to capacity ratio and grade with an R2 of 61.56%. For NO, the variables found to be significant were volume to capacity ratio and grade with an R2 of 99.47%. For NOx, the variables found to be significant were volume to capacity ratio and grade with an R2 of 99.47%. For NH3, the variables found to be significant were volume to capacity ratio and grade with an R2 of 99.25%. This study shows that volume to capacity dominate the emissions quality at a signalized intersection. The research found that for NOx, Idling and Moderate Speed Coasting were significant. For NH3, all variables were significant except Low Speed Coasting. For CO, Braking and Cruise/Acceleration were significant. It was also found that longer delay time reduces CO emissions, but it causes the other three pollutant emissions increase. (2) The ANN modeling method using the Levenberg-Marquardt method was used to train the HCM related variables and MOVES emissions outputs. The parameters of volume to capacity ratio, and road grade are used to estimate emissions. The Validated R value of the obtained ANN model is found