The multifractal spectra of porespace of different types of porous sedimentary rocks are obtained using optical microscopy and a boxcounting technique. The computed f(α) functions were based on the canonical formalism, and can be used as a tool for rock characterization.

In the Canadian nuclear fuel waste management concept, a number of engineered barriers, such as the bentonite-sand buffer which surrounds the waste container in the emplacement boreholes, are used to inhibit the transport of radionuclides. The buffer material is also required to effectively conduct heat from the fuel-waste containers to the surrounding rock. To a large extent, in situ buffer performance will depend on the degree of moisture within the buffer. The moisture content will in turn depend on temperature, temperature gradients, and buffer initial and moisture flux boundary conditions. Modelling of coupled heat and moisture transport in the buffer before resaturation is necessary to assess in situ buffer performance. This paper describes the results of a parametric study using the Philip and de Vries coupled heat and moisture transport model to assess the effects of variations in the moisture diffusivity parameters and the boundary conditions on buffer performance.

The results show that the thermal performance of the buffer is affected by heat-induced moisture movement. In particular, the thermal vapour diffusivity, DTvap, has the most significant effect on thermal drying in a closed system. Work is currently underway to improve our capability to model coupled heat and moisture transport in buffer. Laboratory experiments are in progress to more accurately define the moisture diffusivity parameters and the model is being modified to include the effects of boundary moisture fluxes and pressure potentials so that the resaturation process may be modelled. A full scale buffer/container experiment is currently being planned for conduct in AECL's Underground Research Laboratory to assess further the effects of scale and in situ boundary conditions on buffer performance and to qualify the model.

Clay-based buffers are often proposed as engineered barriers for underground disposal vaults of used nuclear fuel. Thus, in the Canadian conceptual vault design, each used-fuel container is emplaced in a vertical borehole in rock and surrounded by a compacted buffer material, made up of a mixture of sodium bentonite and sand. There is some evidence, however, that the buffer may be susceptible to cracking due to, for example, cementation or moisture depletion.

In this paper we estimate numerically the consequences of cracking on the diffusion of uranium through a finite buffer layer. In particular, we study how the effects of cracks depend on the rock conditions at the edge of the vault.

For swept-away boundary conditions at the rock/buffer interface (e.g., representing a major fracture zone in granite), the total flux through the cracked buffer system is greater than through the corresponding uncracked buffer, particularly for buffers with a very small effective diffusion coefficient. On the other hand, for the case in which the rock at the rock/buffer interface is intact and, thus, the mass transport of material from the buffer into the rock is small, the effect of the cracks on the total release flux is negligible.

Cements and concretes are often considered as components of barriers for the containment of radioactive waste. The performance of such materials as mainly physical barriers to the transport of dissolved radionuclides depends on the mass transfer characteristics of the material. In particular the diffusion and sorption behaviour of the radionuclides and the water permeability are important. These parameters also influence how the chemistry of the concrete is imposed on the repository. In addition, the transport of gas through concrete controls the way in which gases escape from the repository.

Diffusion and gas transport have been measured in a variety of cementitious materials, covering both structural concretes and cementitious backfills; all possible repository construction materials. Measurements have been made using aqueous iodide, strontium and caesium ions and tritiated water as diffusants. The results show that the diffusion of tritiated water is more rapid than that of other species, whilst the transport of strontium and caesium is hindered by sorption; particularly in materials containing blast furnace slag. The transport of gas in these materials has been found to be very sensitive to the degree of water saturation and is extremely low in fully saturated structural concretes. Cementitious backfills have, nevertheless, been identified that have appreciable gas transport even when almost water saturated.

The consequences of the results for the performance of cementitious barriers are discussed.

The effective diffusivity of carbon dioxide and iodine through “G” tunnel tuff were determined using a steady-state method and an unsteady-state method respectively. Results show that the effective diffusivity of carbon dioxide and iodine through dry tuff increased with temperature. The effective diffusivity of carbon dioxide decreased as the moisture content of the “G” tunnel tuff increased. An emprical correlation was obtained to estimate the effective diffusivity of carbon dioxide as a function temperature and the percent saturation. Specific surface area and pore volume of tuff was determined using a mercury porosimeter. A scanning electron microscope was utilized to further characterize the porous structure of the tuff samples.

We predict the transport of 14C from the proposed nuclear waste repository at Yucca Mountain using a porous medium model. Use of this model is justified if the Peclet number, which indicates equilibrium between gas in fractures and liquid in rock pores, is much less than unity. For the assumed release rates, maximum predicted concentrations of 14CO2 in rock near the ground surface are comparable to the USNRC limit for unrestricted areas. Furthermore, dilution near the ground surface as the 14CO2 enters the atmosphere will lower the concentrations by several orders of magnitude. Travel times from the repository to the surface are predicted to be hundreds to thousands of years. For a wide range of the parameters, the release rate from the source has negligible effect on the maximum concentrations at the ground surface.

A series of trial assessments are being undertaken within the United Kingdom to rehearse procedures for post-closure radiological risk analysis of low and intermediate level radioactive waste disposal sites. These trial or “Dry Run” studies draw together workers from a variety of scientific disciplines allowing simultaneous rather than isolated development of assessment procedures. This paper illustrates the use of chemical codes and data within a probabilistic risk assessment framework through a demonstration exercise performed on a hypothetical repository located beneath Harwell, Oxfordshire. The implications for risk assessment practices reliant on the assumption of “conservative” parameters are highlighted by comparison with a more mechanistic approach which takes into account competitive effects.

The Vault Model has been developed to assess the performance of engineered barriers in a conceptual geological disposal vault for used nuclear fuel. It represents container failure, release of radionuclides from used fuel and mass transport of released radionuclides through the clay-based sealing materials surrounding the waste containers. This paper focusses on mass-transport processes represented by the Vault Model, including diffusion, convection and retardation.

In particular, we present results of several scoping calculations carried out with the Vault Model. We consider cases where the clay-based barriers are represented by either a one- or a two-layer system adjacent to an intact rock and a case where the two clay-based barriers are adjacent to a highly fractured rock. These calculations provide insight into the model and produce test cases for comparison with both relatively simple analytical estimates and similar computer codes, as they become available. The analytical estimates generally support the Vault Model results and thus enhance our confidence in the accuracy of the Vault Model calculations.

A mechanistic model describing a dynamic mass balance between the production and consumption of silicic acid was coupled to a near-field mass transport model to predict the dissolution kinetics of a high-level waste glass in a deep geologic repository. The effects of interactions between an iron overpack and the glass are described by a time-dependent precipitation reaction for a ferrous silicate mineral. The kinetic model is used to transform radionuclide concentration-versus-reaction progress values, predicted from a geochemical reaction path computer code, to concentration-versus-time values that are used to calculate the rate of radionuclide release by diffusive mass transfer to the surrounding host rock. The model provides for both solubility-limited and kinetically limited release; the rate-controlling mechanism is dependent on the predicted glass/groundwater chemistry.