UC Riverside

ABSTRACT OF DISSERTATION

Heavy Duty Diesel Particulate Matter and Fuel Consumption Modeling for Transportation Analysis

by

George Alexander Scora

Doctor of Philosophy, Graduate Program in Chemical and Environmental Engineering

University of California, Riverside, March 2012

Dr. Matthew Barth, Chairperson

One of the most important issues concerning transportation is the impact of vehicle emissions on air quality and human health. Vehicle emission modeling is used to predict and evaluate the relationship between transportation activity and transportation emissions for many applications. The first part of this dissertation builds on previous emission modeling work and focuses on the development of a microscale HDD emission model for particulate matter (PM), one of the primary diesel pollutants of concern. In this work, a hybrid approach is used in which a physically based fuel and emission model is coupled with a statistical model to produce PM estimates at the microscale level. The microscale model is calibrated using measured on-road real-time data from UC Riverside's Mobile Emissions Laboratory (MEL). Microscale modeling errors are less than 3% for fuel consumption and less than 17% for PM over a 2.5 hour validation cycle. PM emissions from compression release braking events were also observed, quantified in the data set, and modeled to improve the overall PM emissions estimation.

A key factor in HDD fuel use and carbon dioxide estimation is the operational variability associated with heavy duty truck use. With a better understanding of this variability, it is possible to alter vehicle operation in order to reduce CO₂. The second part of this dissertation focuses on operational parameters and the development of a mesoscale fuel consumption and emission model that accounts for road grade and vehicle weight in addition to velocity. Model development and validation for this portion of work are based on simulated data from the microscale HDD model of measured activity and various road grade and vehicle weight combinations. Mesoscale modeling errors for the validation data set are less than 2% for fuel consumption and less than 12% for PM emissions.

In the final portion of this dissertation, the usefulness of this new mesoscale fuel and emissions model for transportation applications is demonstrated by its implementation in an environmentally friendly navigation (EFNav) application. The focus of EFNav is to decrease the amount of wasted energy (or increased emissions) due to poor routing choices. Routing today is typically based on minimizing the distance or duration traveled. With the inclusion of the new mesoscale fuel/emissions model, vehicle routing can now be based on fuel consumption or emissions that vary with average real-time link speed, average link road grade, and vehicle weight. The application of the mesoscale model is supported by a digital roadway map that integrates real-time traffic data from multiple sources. Vehicle testing of the mesoscale model with EFNav demonstrates the sensitivity of the system to road grade and vehicle weight and the ability of the system to accurately predict fuel consumption and emissions, making it a useful tool for HDD routing.