Understanding the long-term behavior of spent nuclear fuel is critical for safety assessments, and as new doped fuels are developed, additional data are needed to evaluate their performance under disposal conditions. This study presents results from a > 4-year leaching experiment on irradiated Cr + Al-doped UO₂ (ADOPT) and standard UO₂ fragments under hydrogen overpressure in simplified groundwater. Extending previous data from the DisCo project, new samplings and post-leaching characterization are included. Concentrations of radionuclides of interest reached steady state or exhibited very low concentrations and slow release rates. Ultrafiltration confirmed colloid-associated species in the leachate. SEM-WDS revealed a 10 µm secondary phase on pre-leached fragments, enriched in oxygen and partially depleted in most radionuclides. Both fuels exhibited similar long-term behavior under reducing conditions in the presence of H2. Even when secondary phases were present initially, H2 effectively suppressed matrix dissolution, as evidenced by stable radionuclide concentrations throughout the experiment.
The final experiment (full canister test, FCT) of the large-scale gas injection test (LASGIT) aimed to explore the impact of gas volume on gas transport behaviour. Unlike previous tests, the FCT involved pressurising a full-scale KBS-3 canister up to ∼7100 kPa when pressure was held, allowing excess water to drain and to establish gas entry. Once drained, pressure gradually decreased by 220 kPa, indicating gas moving into the fully saturated clay buffer. Observations showed different pore pressure and total stress behaviour compared with earlier tests, but no major changes were seen in the buffer’s response, suggesting gas migration was unaffected by gas volume. The slow pressure decay shows gas travelled through a limited number of narrow pathways, which were not formed through tensile fracturing. A gas leak early in the FCT led to depressurisation and later pressurisation of the canister, causing a 50 μm expansion in its radius. This mechanical loading on the buffer, greater than the bentonite’s drainage capacity, caused pore pressure and radial stress changes. The expansion of the canister by pressurisation or thermal effects should therefore be considered in performance assessments.
Rock spalling is a brittle failure process that occurs around tunnels excavated in hard rock under high in-situ stress states. In nuclear waste disposal, spalling in fractured crystalline rock could create connected fractures, potentially providing pathways for radionuclides. Robust numerical models are therefore needed to evaluate the extent of rock spalling so that the design and layout of a prospective deep geological repository can be optimized and made fit for purpose. With this motivation in mind, this study proposes a methodology for the numerical analysis of rock spalling based on zero-thickness interface elements with a visco-plastic-fracture constitutive law, combined with a workflow for finite element removal/excavation. To simulate spalling, zero-thickness interface elements are pre-inserted along a sufficient number of mesh lines with random orientation within the rock mass. A uniform initial stress state is generated and the excavation of the circular tunnel is performed by removing the corresponding elements, which leads to stresses in excess of the elastic limit in some of the interfaces, and subsequent visco-plastic fracture openings. A criterion for excavation of the finite elements around the tunnel is established when a block which is totally surrounded by failed interfaces is totally detached or can slide off the mesh following a kinematically admissible path. The excavation of blocks causes a stress redistribution around the tunnel and this leads the need for new excavation steps, until a new equilibrium configuration is reached. The proposed methodology is applied to assess rock spalling in the Mine-by Experiment at the Atomic Energy of Canada Limited’s (AECL’s) Underground Research Laboratory in the massive Lac du Bonnet granite. The focus of the analysis is to understand the mechanisms and the influencing factors that lead to brittle failure, and to calibrate material properties to reproduce both the final stress state of the tunnel and its spalling depth.
A microstructural rock model based on the distinct element method employing the Subspring Network contact model with rigid, Breakable, Voronoi-shaped grains (SNBV model) is proposed. The model consists of a mesh (3D Voronoi tessellation) of rigid, breakable, Voronoi blocks. The SNBV model is a microstructural rock model because it is a discrete model that can mimic rock microstructure at the grain scale. SNBV material mimics the microstructure of angular, interlocked, breakable grains with interfaces that may have an initial gap and can sustain partial damage. The model embodies the microstructural features and damage mechanisms that occur at the grain scale: initial microcrack fabric; heterogeneity-induced local tension; and intergranular and transgranular damage. The heterogeneity-induced local tension can be introduced in a controlled fashion that is not tied directly to the shape and packing of the grains and the interface stiffnesses. The synthetic material exhibits behavior during direct-tension and triaxial compression tests that matches the behavior of compact rock. The material can be calibrated to match the standard material properties and characteristic stresses of pink Lac du Bonnet granite. The material properties consist of Young's modulus and Poisson's ratio corresponding with uniaxial compression and Young's modulus corresponding with direct tension, as well as tensile strength, crack-closure stress, crack-initiation stress, secondary crack-initiation stress to mark the onset of grain breakage, crack-damage stress, and compressive strengths up to 4 MPa confinement. The model is suitable for studying the grain-scale micromechanics of brittle rock fracture.
