Publications

The thermo-hydrologic behavior of the potential repository at Yucca Mountain, Nevada, has been simulated to investigate the effects of infiltration. Transient temperatures, liquid saturations, and liquid mass flow rates through the fractures and matrix were simulated using several different steady infiltration rates ranging from 0.3 to 30 min./year. The lower infiltration rates resulted in higher temperatures near the repository element, but the overall transient temperature profiles were similar. The hydrologic response near the repository (liquid saturations and fluxes) was found to be very sensitive to the infiltration rate. Increased infiltration rates reduced the time to re-wet the simulated repository during cooling, and an infiltration rate of 10 mm/year was sufficient to completely eliminate the dry-out zone around the repository.

The effects of heterogeneities on the performance of capillary barriers is investigated by simulating three systems comprised of a fine soil layer overlying a coarse gravel layer with homogeneous, layered heterogeneous, and random heterogeneous property fields. The amount of lateral diversion above the coarse layer under steady-state infiltration conditions is compared between the simulations. Results indicate that the performance of capillary barriers may be significantly influenced by the spatial variability of the properties. The layered heterogeneous system performed best as a result of horizontal features within the fine layer that acted as additional local capillary barriers that delayed breakthrough into the coarse layer. The random heterogeneous system performed worst because of channeled flow that produced localized regions of water breakthrough into the coarse layer. These results indicate that engineered capillary barriers may be improved through emplacement and packing methods that induce a layered system similar to the layered heterogeneous field simulated in this study.

A review of mechanisms, models, and data relevant to the postulated phenomenon of enhanced vapor-phase diffusion in porous media is presented. Information is obtained from literature spanning two different disciplines (soil science and engineering) to gain a diverse perspective on this topic. Findings indicate that while enhanced vapor diffusion tends to correct the discrepancies observed between past theory and experiments, no direct evidence exists to support the postulated processes causing enhanced vapor diffusion. Numerical modeling analyses of experiments representative of the two disciplines are presented in this paper to assess the sensitivity of different systems to enhanced vapor diffusion. Pore-scale modeling is also performed to evaluate the relative significance of enhanced vapor diffusion mechanisms when compared to Fickian diffusion. The results demonstrate the need for additional experiments so that more discerning analyses can be performed.

The thermodynamic environment surrounding a heat-generating waste package can play an important role in the performance of a high-level radioactive waste repository. However, rigorous models of heat transfer are often compromised in near-drift simulations. Convection and radiation are usually ignored or approximated so that simpler conduction models can be used. This paper presents numerical simulations that explicitly model conduction, convection, and radiation in an empty drift following emplacement of a heat-generating waste package. Temperatures and relative humidities are determined at various locations within the drift. Comparisons are made between different models of heat transfer, and the relative effects of each heat transfer mode on the thermodynamic environment of the waste package are examined.

In accordance with the Nuclear Regulatory Commission regulation regarding groundwater travel times at geologic repositories, various models of unsaturated flow in fractured tuff have been developed and implemented to assess groundwater travel times at the potential repository at Yucca Mountain, Nevada. Kaplan used one-dimensional models to describe the uncertainty and sensitivity of travel times to various processes at Yucca Mountain. Robey and Arnold et al. used a two-dimensional equivalent continuum model (ECM) with inter- and intra-unit heterogeneity in an attempt to assess fast-flow paths through the unsaturated, fractured tuff at Yucca Mountain (GWTT-94). However, significant flow through the fractures in previous models was not simulated due to the characteristics of the ECM, which requires the matrix to be nearly saturated before flow through the fractures is initiated. In the current study (GWTT-95), four two-dimensional cross-sections at Yucca Mountain are simulated using both the ECM and dual-permeability (DK) models. The properties of both the fracture and matrix domains are geostatistically simulated, yielding completely heterogeneous continua. Then, simulations of flow through the four cross-sections are performed using spatially nonuniform infiltration boundary conditions. Steady-state groundwater travel times from the potential repository to the water table are calculated.

The performance of waste packages containing high-level nuclear wastes at underground repositories such as the potential repository at Yucca Mountain, Nevada, depends, in part, on the thermodynamic environment immediately surrounding the buried waste packages. For example, degradation of the waste packages can be caused by corrosive and microbial processes, which are influenced by both the relative humidity and temperature within the emplacement drifts. In this paper, the effects of conduction, convection, and radiation are investigated for a heat-generating waste package in an empty-drift. Simulations explicitly modeling radiation from the waste package to the drift wall are compared simulations using only conduction. Temperatures, relative humidities, and vapor mass fractions are compared at various locations within the drift. In addition, the effects of convection on relative humidity and moisture distribution within the drift are presented.

Groundwater travel time (GWTT) calculations will play an important role in addressing site-suitability criteria for the potential high-level nuclear waste repository at Yucca Mountain,Nevada. In support of these calculations, Preliminary assessments of the candidate codes and models are presented in this report. A series of benchmark studies have been designed to address important aspects of modeling flow through fractured media representative of flow at Yucca Mountain. Three codes (DUAL, FEHMN, and TOUGH 2) are compared in these benchmark studies. DUAL is a single-phase, isothermal, two-dimensional flow simulator based on the dual mixed finite element method. FEHMN is a nonisothermal, multiphase, multidimensional simulator based primarily on the finite element method. TOUGH2 is anon isothermal, multiphase, multidimensional simulator based on the integral finite difference method. Alternative conceptual models of fracture flow consisting of the equivalent continuum model (ECM) and the dual permeability (DK) model are used in the different codes.

Laboratory-scale heater experiments are Proposed to observe thermohydrologic Processes in tuffaceous rock using existing equipment and x-ray imaging techniques. The purpose of the experiments is to gain understanding of the near-field behavior and thermodynamic environment surrounding a heat source. As a prelude to these experiments, numerical simulations are performed to determine design-related parameters such as optimal heating power and heating duration. In addition, the simulations aid in identifying and understanding thermal processes and mechanisms that may occur under a variety of experimental conditions. Results of the simulations show that convection may play an important role in the heat transfer and thermodynamic environment of the heater if the Rayleigh-Darcy number exceeds a critical value (= 10 for the laboratory experiments) depending on the type of backfill material within the annulus (or drift).

This paper presents the relations that describe thermodynamic equilibrium in a three-phase system. Multiple components, including air, water, and oil components, are considered in three phases: (1) aqueous, (2) oil, and (3) gas. Primary variables are specified for each of seven possible phase combinations. These primary variables are then used to determine the necessary secondary variables to completely describe the system. Criteria are also developed to check the stability of each phase configuration and determine possible transitions from one phase configuration to another phase configuration via phase appearances and disappearances.

The multiphase, multicomponent, non-isothermal simulator M2NOTS was tested against several one-dimensional experiments. The experiments represented limiting conditions of soil venting processes: (1) a through-flow condition in which air flows through the contaminated region, and (2) a bypass-flow condition in which air is channeled around (rather than through) the contaminated region. Predictions using M2NOTS of changing in situ compositions and effluent concentrations for toluene and o-xylene mixtures were compared to the observed results for each condition. Results showed that M2NOTS was able to capture the salient trends and features of multicomponent through-flow and bypass-flow venting processes.

The numerical code TOUGH2 was used to assess alternative conceptual models of fracture flow. The models that were considered included the equivalent continuum model (ECM) and the dual permeability (DK) model. A one-dimensional, layered, unsaturated domain was studied with a saturated bottom boundary and a constant infiltration at the top boundary. Two different infiltration rates were used in the studies. In addition, the connection areas between the fracture and matrix elements in the dual permeability model were varied. Results showed that the two conceptual models of fracture flow produced different saturation and velocity profiles-even under steady-state conditions. The magnitudes of the discrepancies were sensitive to two parameters that affected the flux between the fractures and matrix in the dual permeability model: (1) the fracture-matrix connection areas and (2) the capillary pressure gradients between the fracture and matrix elements.

Coupled thermal and hydrologic flow processes have been recognized as important factors in the evaluation of Yucca Mountain as a potential repository for high-level radioactive wastes. As a result, several models and numerical codes such as TOUGH2 have been used to investigate the thermohydrologic conditions near a potential nuclear waste repository. However, very few of these models have been tested through laboratory or field scale studies. This work has therefore focused on modeling well-controlled experiments of non-isothermal flow processes in porous media at different scales to serve two primary objectives: (1) identify processes that are potentially important to thermal and hydrologic transport at Yucca Mountain and (2) build confidence in models and codes through combined experimental and numerical studies of thermohydrologic behavior at different scales and conditions. In this report, three independent studies of thermohydrologic flow processes at laboratory and field scales are presented. The experiments and field studies that are presented here were performed independently of this work. The main focus of this report was to use the numerical code TOUGH2 to simulate the non-isothermal flow behavior observed in each experiment to generate understanding of the thermohydrologic processes and to gain confidence in the code. TOUGH2 was chosen due to its current use in calculations associated with Yucca Mountain and its capability of modeling the coupled transport of air, water, vapor, and heat in porous media.

Saturation profiles resulting from TOUGH2 numerical simulations of water infiltration into a tuff matrix from a saturated vertical fracture have been compared to experimental results. The purpose was to determine the sensitivity of the infiltration on local heterogeneities and different representations of two-phase characteristic curves used by the model. Findings indicate that the use of simplified (linearized) capillary pressure curves with rigorous (van Genuchten) relative permeability curves resulted in a more computationally efficient solution without a loss in accuracy. However, linearized forms of the relative permeability functions produced poor results, regardless of the form of the capillary pressure function. In addition, numerical simulations revealed that the presence of local heterogeneities in the tuff caused non-uniform saturation distributions and wetting fronts in the in matrix.