Groundwater flow at the prototype repository. Task 8e of SKB

The Task Force on Groundwater Flow and Transport of Solutes (TF GWFTS) and the Task Force on Engineered Barrier Systems (TF EBS), both established by the Svensk Kärnbränslehantering AB (SKB), have defined the so-called Task 8 to investigate the hydraulic interaction of the granitic host rock at the Hard Rock Laboratory at Äspö and the bentonite clay buffer in a deep geological repository. Subtasks 8 a-d ran parallel to the related BRIE-project (Bentonite Rock Interaction Experiment) at the Äspö Hard Rock Laboratory (HRL). These subtasks lead to a variety of predictive models with respect to flow in a buffer-rock system on the comparatively small scale of the BRIE. Task 8e related to the Prototype Repository (PR) was therefore subsequently defined in order to check these concepts on a much larger scale and under the additional influence of temperature. Task 8 encompassed obviously characterizing the groundwater flow field as well as simulating bentonite re-saturation. In the framework of Tasks 8 b-d an approach had been developed where the problem of groundwater flow in rock and buffer is solved by decoupling both aspects. Groundwater flow was simplified to a steady-state single-phase flow model including discretely described large fractures. Modelling was performed with the code d3f. This report is concerned with Task 8e addressing the aspect of groundwater flow at the PR with respect to two questions: is the conceptual approach for groundwater flow that was developed for the BRIE viable at different conditions and what is the influence of temperature on the flow field? The procedure of the PR-experiment can roughly be divided into a pre-installation phase where only the tunnel and the boreholes existed and an operational phase after installing buffer and heaters when the heaters were switched on. During the pre-installation phase groundwater flow was isothermal without interference of the buffer and a lot of effort went into characterizing the hydrogeological conditions around the PR. This phase represents therefore the most simple and, at the same time, the best known flow conditions. The pre-installation phase was therefore considered to be most appropriate to develop a well-founded flow model. Based on the model concept for the BRIE and beginning with the related material data, a first model was set up which was then calibrated against outflow data for the tunnel as well as for the boreholes. It became apparent that only moderate modifications were necessary to achieve a satisfactory match thereby confirming the model concept for groundwater flow at Äspö. To answer the second question called for modelling flow during the operational phase which included the thermal effects from heating the canisters. To investigate the impact of heating a pure heat conduction model was set up for calculation with the code COMSOL. Significant heating appeared to occur only in the vicinity of the canisters, though. Neglecting a possible influence of convective heat transport on the temperature field, the temperature evolution at each of the borehole surfaces in the model was determined as input into a thermo-hydraulic model for the operational phase. Using the option in d3f to couple heat transport with groundwater flow allowed then to have a direct comparison of the isothermal and the non-isothermal flow field. The general patterns of the isothermal flow field remained basically preserved in the thermo-hydraulically coupled (TH-)model. The flow velocities were not increased by considerably more than a factor of 2. This factor is consistent with the temperature-induced changes of density and viscosity of the water. It has thus to be concluded that heat production from the waste canisters does not result in a noteworthy change of groundwater flow as long as there is significant outflow from the rock into the deposition boreholes. The pressure gradient from the boundary towards the geotechnical openings is simply too high to allow for density-dependent flow effects. How this would change in accordance with the expected low flow rates that are imposed by the water uptake of the bentonite buffer remains to be investigated.