Skip to Content

Benchmark Applications

A number of benchmark problems of varying complexity are being developed for use in creating and testing the hybrid multiscale modeling system. The two primary application problems identified in the initial proposal are briefly described here. We are currently focusing on developing the first (and simpler) of these two benchmark problems.

Calcite Precipitation Experiment

This application benchmark problem is based on intermediate-scale laboratory experiments being conducted at Idaho National Laboratory under the project entitled "Coupling Between Flow and Precipitation in Heterogeneous Subsurface Environments and Effects on Contaminant Fate and Transport" (George Redden, PI). This project is supported by DOE's Office of Science through the Environmental Remediation Sciences Program (ERSP), and was listed as one of the target applications in the SciDAC project proposal. Since several investigators are involved in both the ERSP and SciDAC projects (Redden, Scheibe and Tartakovsky), the use of this application as a benchmark example will be beneficial to both projects and will leverage programmatic funding.

The figure below shows the experimental flow cell (front view). In the experiment shown here, two solutes were injected at the bottom, one on each half of the lower boundary, causing upward flow. The flow cell is relatively thin in the third dimension, and is packed with homogeneous sand. The two solutes, containing dissolved calcium and carbonate respectively, mixed at the interface between the two flows leading to supersaturation and precipitation of calcium carbonate solids (calcite; the whitish band in the figure below).

Experimental flow cell

We have performed preliminary simulations of this experimental configuration at both pore- and continuum-scales. These simulations (reported in a recent poster presented at the Fall Meeting of the American Geophysical Union) clearly demonstrate the need for direct coupling of pore- and continuum-scale models and strongly motivate use of this problem as an application benchmark for the SciDAC project.

Continuum-scale simulations of a number of alternative scenarios are presented here. These are being used to support design of continuing physical experiments at INL by George Redden, and will provide additional example problems for testing the hybrid multiscale approach.

Field-Scale Uranium Bioremediation Experiment

This application benchmark problem is based on field-scale biostimulation experiments conducted over the past four years at the Oak Ridge Reservation under the project entitled "In Situ Immobilization of Uranium in Structured Porous Media via Biomineralization at the Fracture/Matrix Interface," (Tim Scheibe, PI). This project has been supported by DOE's Office of Science through the Environmental Remediation Sciences Program (ERSP), and was also listed as one of the target applications in the SciDAC project proposal. Again, since multiple investigators are involved in both the ERSP and SciDAC projects (Scheibe, Brooks), the use of this application as a benchmark example will be beneficial to both projects and will leverage programmatic funding.

Photo of the FRC site (tent) Photo of the FRC site (sampling station)

The key premise of the field-scale experiments is that introduction of a soluble carbon source (electron donor; here we use ethanol) will induce enhanced microbial activity, particularly at interfaces between advection-controlled (fast-flowing) and diffusion-controlled (slow-flowing) regions of the aquifer. This process introduces a strong coupling between flow, transport, and biogeochemical reactions (especially biologically-mediated metal reduction) that controls the mobility of uranium in the groundwater. These strongly coupled, highly heterogeneous, local-scale processes are difficult to represent adequately in standard continuum models but are potentially well-suited to a hybrid pore-scale/continuum-scale approach. We will also explore the incorporation of sub-pore-scale (cellular-scale) models based on an in-silico modeling paradigm into the overall modeling framework. The hybrid multiscale model application will build on an existing continuum model of groundwater flow and reactive transport at the field site; various aspects of this model have been reported in the following publications: