Publications

Vapor diffusion in porous media in the presence of its liquid has often been analyzed like air diffusion. The diffusion rate is much lower than in free space due to the presence of the porous medium and any liquid present. However, enhanced vapor diffusion has also been postulated such that the diffusion rate may approach free-space values. The mechanisms postulated to lead to this enhancement include condensation/evaporation across isolated liquid islands in the porous media and an increased temperature gradient in the gas phase. In order to try to understand the mechanisms involved in such an enhancement, pore-scale models have been developed. Vapor diffusion in the presence of liquid islands has been evaluated for a one-dimensional pore network under a concentration gradient. The simulations show that significant enhancement of vapor diffusion is indeed possible in the presence of liquid islands, while air diffusion decreases slightly. While the present pore-scale model indicates that enhanced vapor diffusion is possible, only experimental data can confirm the relevant processes.

Capillary barriers, consisting of tilted fine-over-coarse layers under unsaturated conditions, have been suggested as landfill covers to divert water infiltration away from sensitive underground regions, especially for arid and semi-arid regions. The Hydrological Evaluation of Landfill Performance (HELP) computer code is an evaluation tool for landfill covers used by designers and regulators. HELP is a quasi-two-dimensional model that predicts moisture movement into and through the underground soil and waste layers. Processes modeled within HELP include precipitation, runoff, evapotranspiration, unsaturated vertical drainage, saturated lateral drainage, and leakage through liners. Unfortunately, multidimensional unsaturated flow phenomena that are necessary for evaluating tilted capillary barriers are not included in HELP. Differences between the predictions of the HELP and those from a multidimensional unsaturated flow code are presented to assess the two different approaches. Comparisons are presented for the landfill covers including capillary barrier configurations at the Alternative Landfill Cover Demonstration (ALCD) being conducted at Sandia.

Ross developed an analytical relationship to calculate the diversion length of a tilted fine-over-coarse capillary barrier. Oldenburg and Pruess compared simulation results using upstream and harmonic weighting to the diversion length predicted by Ross formula with mixed results; the qualitative agreement is reasonable but the quantitative comparison is poor, especially for upstream weighting. The proximity of the water table to the fine-coarse interface at breakthrough is a possible reason for the poor agreement. In the present study, the Oldenburg and Pruess problem is extended to address the water table issue. When the water table is sufficiently far away from the interface at breakthrough, good qualitative and quantitative agreement is obtained using upstream weighting.

Capillary barriers consisting of tilted fine-over-coarse layers have been suggested as landfill covers as a means to divert water infiltration away from sensitive underground regions under unsaturated flow conditions, especially for arid and semi-arid regions. Typically, the HELP code is used to evaluate landfill cover performance and design. Unfortunately, due to its simplified treatment of unsaturated flow and its essentially one-dimensional nature, HELP is not adequate to treat the complex multidimensional unsaturated flow processes occurring in a tilted capillary barrier. In order to develop the necessary mechanistic code for the performance evaluation of tilted capillary barriers, an efficient and comprehensive unsaturated flow code needs to be selected for further use and modification. The present study evaluates a number of candidate mechanistic unsaturated flow codes for application to tilted capillary barriers. Factors considered included unsaturated flow modeling, inclusion of evapotranspiration, nodalization flexibility, ease of modification, and numerical efficiency. A number of unsaturated flow codes are available for use with different features and assumptions. The codes chosen for this evaluation are TOUGH2, FEHM, and SWMS{_}2D. All three codes chosen for this evaluation successfully simulated the capillary barrier problem chosen for the code comparison, although FEHM used a reduced grid. The numerical results are a strong function of the numerical weighting scheme. For the same weighting scheme, similar results were obtained from the various codes. Based on the CPU time of the various codes and the code capabilities, the TOUGH2 code has been selected as the appropriate code for tilted capillary barrier performance evaluation, possibly in conjunction with the infiltration, runoff, and evapotranspiration models of HELP. 44 refs.

Two models for gas-phase diffusion and advection in porous media, the Advective-Dispersive Model (ADM) and the Dusty-Gas Model (DGM), are reviewed. The ADM, which is more widely used, is based on a linear addition of advection calculated by Darcy`s Law and ordinary diffusion using Fick`s Law. Knudsen diffusion is often included through the use of a Klinkenberg factor for advection, while the effect of a porous medium on the diffusion process is through a porosity-tortuosity-gas saturation multiplier. Another, more comprehensive approach for gas-phase transport in porous media has been formulated by Evans and Mason, and is referred to as the Dusty- Gas Model (DGM). This model applies the kinetic theory of gases to the gaseous components and the porous media (or ``dust``) to develop an approach for combined transport due to ordinary and Knudsen diffusion and advection including porous medium effects. While these two models both consider advection and diffusion, the formulations are considerably different, especially for ordinary diffusion. The various components of flow (advection and diffusion) are compared for both models. Results from these two models are compared to isothermal experimental data for He-Ar gas diffusion in a low-permeability graphite. Air-water vapor comparisons have also been performed, although data are not available, for the low-permeability graphite system used for the helium-argon data. Radial and linear air-water heat pipes involving heat, advection, capillary transport, and diffusion under nonisothermal conditions have also been considered.

Stratigraphic units of the Salado Formation at the Waste Isolation Pilot Plant (WIPP) disposal room horizon includes various layers of halite, polyhalitic halite, argillaceous halite, clay, and anhydrite. Current models, including those used in the WIPP Performance Assessment calculations, employ a ``composite stratigraphy`` approach in modeling. This study was initiated to evaluate the impact that an explicit representation of detailed stratigraphy around the repository may have on fluid flow compared to the simplified ``composite stratigraphy`` models currently employed. Sensitivity of model results to intrinsic permeability anisotropy, interbed fracturing, two-phase characteristic curves, and gas-generation rates were studied. The results of this study indicate that explicit representation of the stratigraphy maintains higher pressures and does not allow as much fluid to leave the disposal room as compared to the ``composite stratigraphy`` approach. However, the differences are relatively small. Gas migration distances are also different between the two approaches. However, for the two cases in which explicit layering results were considerably different than the composite model (anisotropic and vapor-limited), the gas-migration distances for both models were negligible. For the cases in which gas migration distances were considerable, van Genuchten/Parker and interbed fracture, the differences between the two models were fairly insignificant. Overall, this study suggests that explicit representation of the stratigraphy in the WIPP PA models is not required for the parameter variations modeled if ``global quantities`` (e.g., disposal room pressures, net brine and gas flux into and out of disposal rooms) are the only concern.

The natural dip of the Salado Formation at the Waste Isolation Pilot Plant (WIPP), although regionally only about 111, has the potential to affect brine inflow and gas-migration distances due to buoyancy forces. Current models, including those in WIPP Performance Assessment calculations, assume a perfectly horizontal repository and stratigraphy. With the addition of buoyancy forces due to the dip, brine and gas flow patterns can be affected. Brine inflow may increase due to countercurrent flow, and gas may preferentially migrate up dip. This scoping study has used analytical and numerical modeling to evaluate the impact of the dip on brine inflow and gas-migration distances at the WIPP in one, two, and three dimensions. Sensitivities to interbed permeabilities, two-phase curves, gas-generation rates, and interbed fracturing were studied.

Simulations of soil-heated vapor extraction have been performed to evaluate the NAPL removal performance as a function of borehole vacuum. The possibility of loss of NAPL containment, or NAPL migration into the unheated soil, is also evaluated in the simulations. A practical warning sign indicating migration of NAPL into the unheated zone is discussed.

Measurements of capillary barrier performance have been conducted in above-grade wooden structures (boxes) configured to measure the water balance. The capillary-barrier portion of the boxes is 6.0 m long, 2.0 m wide, and 1.2 m high with a slope of 5%. A coarse-grained material was placed in the bottom 25-cm of the box with a 90-cm deep fine-grained material (local soil) on top. A region for laterally diverted water to accumulate and drain was created in the last 1.0 m of the box. The soil at the top is terraced into five, 1.4 m long, level intervals to prevent runoff when adding water. Water is added uniformly to the entire top of the box at a rate of about 66 l/day, or an infiltration rate of 1.7 m/year. The top of the box is covered with fiber-reinforced plastic to minimize evaporation of water, discourage plant growth, and prevent rainfall from contacting the soil. Five drains are spaced along the bottom of the coarse layer. These drains discretize the coarse layer into five collection regions to provide a means of identifying the breakthrough location into the coarse layer. A drain is also located in the downdip collection region of the box. Soil moisture changes were measured in the fine-grained material with a frequency-domain reflectometry (FDR) probe, which was calibrated using soil from the field site at a known moisture content and density.

The Waste Isolation Pilot Plant (WIPP) is a US Department of Energy research and development facility for the underground disposal of transuranic waste from US defense-related activities. The WIPP repository is located within the Salado Formation, which is comprised of beds of pure and impure halite with thin interbeds of anhydrite and related clay seams. This formation is brine saturated with a pore pressure of approximately 12.5 MPa at the repository horizon. The Salado Formation dips gently southeast, on the average approximately 1{degree}, with steeper dips locally. Elevated repository pressures, caused by gas generated as emplaced waste corrodes and degrades, may drive brine and gas out of the repository into the surrounding formation. Stratigraphic dip may cause increased brine inflow to the repository through countercurrent flow in the interbeds and enhanced gas migration distances in the updip direction due to buoyancy. Two-dimensional simulations of isolated WIPP repository room have been performed using TOUGH2 for horizontal and 1{degree} dipping stratigraphy. The impact of dip on multiphase flow at the WIPP may be significant. With dip, an additional mechanism for brine inflow may occur, namely the formation of a cell of countercurrent brine and gas flow in the interbeds. The additional volume of brine inflow resulting from the countercurrent flow cell may be of similar magnitude to brine inflow without dip. Therefore, dip must be included in any repository model to include the countercurrent brine inflow mechanism. Gas migration may also be significantly influenced due to dip. Gas migration distances may increase dramatically with preferential migration updip.

The Waste Isolation Pilot Plant (WIPP) is a US Department of Energy (DOE) research and development facility for the underground disposal of transuranic waste in southeastern New Mexico. The WIPP repository is located 655 m below the land surface in the lower portion of the Salado Formation, which is comprised of beds of pure and impure halite with thin interbeds of anhydrite and related clay seams. The regional dip of the Salado Formation is approximately 1{degree} southeast in the vicinity of the repository. The proposed waste storage area has eight waste disposal panels, each of which will contain seven rooms. The repository is designed to follow a single stratigraphic horizon. Due to the dip, the north end of the repository will be about 10 meters higher than the south end. Waste that is emplaced in the disposal rooms will generate gas due to microbial degradation, anoxic corrosion, and radiolysis. Brine inflow to the rooms from the surrounding Salado Formation may significantly influence the gas generation rate and the total amount of gas generated. The salt surrounding the repository will creep in response to the excavation, reducing the room volume. Gas generation in the room may increase the pressure sufficiently to drive brine and gas into the surrounding Salado Formation. Migration of gas and brine in the Salado is an important factor in evaluating the performance of the repository. The studies summarized in this paper have. been performed to evaluate brine and gas flow processes in the WIPP disposal system and to identify some of the important processes. These studies are done in support of, but are not part of, the formal Performance Assessment (PA) effort. Because of probabilistic and system-scale requirements, the PA effort uses the Sandia-developed BRAGFLO (BRine And Gas FLOw) code for multiphase flow calculations.

Detailed simulations have been performed for the TEVES (Thermal Enhanced Vapor Extraction System) Project using the TOUGH2 code considering air, water, and a single-component NAPL. A critical parameter varied in the simulations is the borehole vacuum which directly affects air flow through the system and indirectly influences soil temperatures and water and NAPL fluid masses. Contaminant migration from the heated zone into the unheated soil can occur if the borehole vacuum, or borehole flow rate, is not sufficient. Under these conditions, evaporation of liquids (water and NAPL) due to the heating can cause flow from the heated zone into the unheated soil. Insufficient air sweep may be indicated by a vapor dominated mass flow rate into the borehole, at least for the present configuration. Sufficient air flow through the heated zone must be provided to contain the contaminants within the heated zone.

Brine inflow to the Waste Isolation Pilot Plant is important in assessing the performance of the repository, and a mechanistic model is needed for performance calculations. Brine inflow experiments are being conducted, and formation parameters such as the permeability and diffusivity are inferred from these data using a simplified one-dimensional radial, uniform property, single-phase Darcy flow model. This model has met with limited success in interpreting some of the recent data. Much of the data could not be satisfactorily fit with the above model because the brine inflow rate increases with time, so a more mechanistic model is being developed based on the TOUGH and TOUGH2 computer codes. These codes are much more complex than the simplified model and include a number of parameters that have not been measured. Therefore, a one-dimensional brine inflow sensitivity study has been undertaken to evaluate the importance of a number of these parameters in influencing the behavior of brine inflow to open boreholes. In addition, two-phase conditions have been included in the study, and the sensitivity of gas inflow rates and the formation pressure and saturation distributions after 1 year are examined. These results should be helpful in determining what additional measurements are necessary to assist in the development of a more mechanistic brine inflow model.

Sensitivity studies have been conducted for the gas release from the Waste Isolation Pilot Plant (WIPP) using the TOUGH2 computer code with performance measures of peak repository pressure and gas migration distance at 1000 years. The effect of formation permeabilities including impermeable halite, two-phase characteristic curves including different models and residual saturations, and other variations was studied to determine their impact on the performance of the WIPP repository. 15 refs., 7 figs., 2 tabs.

The Strategic Petroleum Reserve (SPR) cavern fluid velocity model for natural convection uses the Modified Local Similarity (MLS) method to analyze the boundary layer behavior. In order to use the MLS approach, boundary layer velocity and temperature profiles are calculated in terms of local similarity variables based on the natural convection equations. Modifications were made to the local similarity equations enabling consideration of turbulent flow and mixed convection conditions. The details of these changes are addressed in this report. 80 refs., 43 figs., 3 tabs.

As part of the Strategic Petroleum Reserve (SPR), the Weeks Island oil storage site is a converted salt mine that contains approximately 73 million barrels of oil overlying 0.5 million barrels of brine. The oil is contained on two levels of the converted mine which are connected by a number of shafts and openings. Oil recycle exercises are periodically conducted to test the oil fill and withdrawal systems in which oil is simultaneously injected and withdrawn from two different locations in the lower level, and brine may be transported around the lower level of the mine by the movement of the oil. 11 refs., 16 figs.