Date of Award

12-2017

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Environmental Engineering and Earth Sciences

Committee Member

Dr. Ronald Falta, Committee Chair

Committee Member

Dr. David Freedman

Committee Member

Dr. Lawrence Murdoch

Abstract

Secondary sources of contamination, such as dissolved chemicals in low permeability zones result in plume persistence and limitations for plume remediation as a consequence of the process of matrix diffusion (Mackay and Cherry, 1989; Chapman and Parker, 2005). Existing numerical transport simulators are capable of modeling matrix diffusion; however, they require fine discretization, resulting in large computational effort. An alternative approach was developed combining numerical and analytical modeling to simulate matrix diffusion effects. The semi-analytical/numerical (referred to as semi-analytical for short) approach used here was adapted from geothermal reservoir modeling of transient heat conduction in low permeability cap rocks (Vinsome and Westerveld, 1980). The semi-analytical method discretizes the high permeability parts of the aquifer in the numerical model and the matrix diffusion flux is approximated at the sub-grid scale without modifying the grid. The objective of this research is to test the semi-analytical method for the simulation of matrix diffusion effects in groundwater transport. To achieve this goal the semi-analytical method was used to simulate laboratory-scale studies and the results were compared to experimental data. In addition, various test scenarios representing heterogeneous environments were developed and compared to results from a commercial numerical simulator. Two implementations of the matrix diffusion analysis were tested in this research. Initially, a Visual Basic program in Excel® was compared to experimental results from two published studies from University of Florida and Colorado State University. Results from the Visual Basic code were also compared to fine-grid numerical simulations of two-layer systems. A FORTRAN version of this program, called REMChlor-MD was evaluated by comparing to results from large fine-grid numerical models (approximately 3 million gridblocks) with highly heterogeneous material distributions. The results indicate that the semi-analytical method matches both experimental data and fine grid numerical simulations, even for systems with highly complex heterogeneities. Besides the visual comparison, coefficients of determination were estimated for the cases studied, obtaining values from 0.724 to 0.998, demonstrating good accuracy of the matrix diffusion semi-analytical method for most practical purposes. The semi-analytical model is highly efficient, requiring only a fraction (approximately 1/10000) of the run times of the fine grid numerical simulations used as comparison basis. This evaluation is one of the stages of the project funded by the DoD's Environmental Security Technology Certification Program (ESTCP) and supported in part by the Department of Energy. The project aims to develop and implement a new generation of the screening level transport model REMChlor that considers matrix diffusion in the plume: REMChlor-MD.

Share

COinS