Loading...
Exploration of solute transport mobility in 1-D and 3-D physical models, An
Foster, Allan W.
Foster, Allan W.
Citations
Altmetric:
Advisor
Editor
Date
Date Issued
2019
Date Submitted
Collections
Research Projects
Organizational Units
Journal Issue
Embargo Expires
Abstract
The advection-dispersion equation (ADE) often fails to predict solute transport in part due to the presence of less well connected, less-mobile pore space. Here, we explore the role of porosity or flowpaths of differing mobility and their physical effects on tracer transport in a series of controlled 1-D column- and 3-D tank-scale experiments. Experiments at both scales included co-located (spatially and temporally) fluid electrical conductivity (σ_f) and bulk apparent electrical conductivity (σ_b) to help parse the relationship between mobile and less-mobile domains. Systems with domains of differing mobility showed elongated tailing in observed σ_f breakthrough curves and a lag between σ_f and σ_b, phenomena not easily described by the ADE. Numerical models—STAMMT-L for the 1-D column experiments coupled with analytical graphical methods, and SEAWAT and R3t for the 3-D tank-scale experiments—were used to estimate physical parameters controlling transport. At the 1-D column-scale, three different grain packings were studied in 24.4-cm length columns, including 1) homogeneously packed #20/#30 Accusand, which contains intergranular but no intragranular pore space; 2) St. Cloud zeolite clinoptilolite, and 3) crushed amorphous silica glass, both of which contain intergranular and intragranular pore space. Experiments included stepped NaCl tracer injections at Peclet numbers ranging from 3 to 1000 to investigate flow-rate controls on physical solute transport parameters. The zeolite and amorphous silica glass results both indicated the presence of a less-mobile domain and mass transfer rates between more- and less-mobile porosity influenced by tracer injection duration and flow rate, confirmed by numerical models in STAMMT-L, whereas data from the homogeneous Accusand indicated little less-mobile pore space, as expected. At the 3-D tank-scale, a synthetic-heterogeneous aquifer was developed with four impermeable barriers installed in otherwise homogeneously packed #70 Unimin sand within a 429 cm x 244 cm x 36 cm tank. A 72-hr NaCl-pulse tracer injection at a Peclet number of 15 was performed. Results indicate solute trapping behind the impermeable barriers, creating advective flowpaths of different lengths. The orientation of the impermeable barriers with respect to bulk flow direction controls the formation of these pathways. The observation of less-mobile flowpaths from the co-located σ_b versus σ_f measurements at 12 different locations indicated that a stagnation zone forms directly downgradient of the largest barrier (perpendicular to primary flow direction), which was imaged in 3D electrical resistivity inversions. Our co-located σ_b versus σ_f data also indicate density-dependent flow that is primarily occurring within the highly permeable well conduits, and difference inversions indicate the plume increasing in thickness with depth, characteristic of density-dependent flow. Observed data and numerical simulations both indicated an influence of density-driven solute transport in the locations directly up- and downgradient of the largest impermeable barrier on observed hysteresis in σ_b versus σ_f measurements due to solute collection directly upgradient of the barrier and allowing density-differences to manifest and propagate downgradient. Numerical simulations in SEAWAT and R3t helped isolate physical transport controls within the synthetic aquifer due to the impermeable barriers, density-driven flow, or a combination of the two.
Associated Publications
Rights
Copyright of the original work is retained by the author.