Graduation Semester and Year




Document Type


Degree Name

Doctor of Philosophy in Civil Engineering


Civil Engineering

First Advisor

Melanie L Sattler


Developing cost effective control strategies for ozone has been a challenge to air quality modelers. Conventionally, the control strategies are applied across-the board to the region. The main aim of this research was to develop a Decision-Making Framework (DMF) for evaluating and optimizing the selection of ozone control strategies. Conventional across-the-board reductions conduct emission reductions uniformly throughout the region and throughout the day. By contrast, this dissertation studied targeted reductions, in which emission sources of various types are reduced at various times and locations. The proposed DMF comprised of four phases: (1) Initialization, (2) Mining, (3) Metamodeling, and (4) Optimization. This DMF was tested on a DFW 2009 future case episode which was based on a 10-day episode from August 13-22, 1999. 612 emission variables were identified in three source categories viz. point, area (includes non-road) and line (on-road). The emission control regions and time periods along with ozone monitoring regions and time periods were defined. The control strategy emission reductions and costs were also identified in this stage. Initially a Latin hypercube experimental design was setup to organize 30 sets of emission reduction scenarios to be modeled using the photochemical model CAMx. Data mining reduced the number of variables to a maximum of 126. A second Latin hypercube was setup to organize another 30 emission reduction scenarios for the significant variables identified by data mining. Metamodels were developed for ozone from the 60 CAMx runs using linear regression models constructed with the stepwise model selection method. Stepwise regression further reduced the number of variables. The metamodels were implemented in optimization as a surrogate for time-intensive CAMx modeling. Appropriate constraints were calculated for each metamodel to ensure that it satisfied EPA's MAT. The optimization was formulated to find the most cost effective combination of targeted control strategies that brings the region into attainment for the 8-hour ozone. Each day was optimized individually in sequence. In order to demonstrate applicability of the DMF 5 days (August 15, 16, 17, 18 and 19) of the episode were optimized. Although the optimization identified the key sources, time periods, and control strategies, the existing controls on these sources were not adequate to bring the region in attainment. Further reductions at these sources beyond the existing list of TCEQ/NCTCOG control strategies were required. Further modifications in the DMF for DFW were suggested to improve its performance.


Civil and Environmental Engineering | Civil Engineering | Engineering


Degree granted by The University of Texas at Arlington