Cyclic gas injection tests on a sand/bentonite mixture

Document available in extended abstract form only. Sand/Bentonite (S/B) mixtures are currently investigated for backfilling underground structures and for the seals in the Swiss concept for L/ILW (low/intermediate level waste) repositories. One important aspect regarding the performance assessment of these repositories is the consequence of the gas generated by microbial degradation of organic material and by anaerobic corrosion of the waste canister on the clay host rock. The main aim of repository plugs and the repository seal are to increase the gas transport capacity of the backfilled underground structures without compromising the radionuclide retention capacity of host rock and the engineered barrier system. Initially the generated gas is evacuated for diffusion through the pore water, then, once the gas concentration exceed the gas solubility the pressure will increase and other transport mechanisms may appear and become predominant (two phase flow and gas breakthrough). As a consequence the gas pressure will be realised from the emplacement cavern. Due to the progressive nature of the gas generation, gas pressure may increase again with time and penetrate once more through the material. In the present work, the results of cyclic gas injection tests on S/B mixture are presented. Several gas injection tests were performed in a conventional triaxial apparatus at the Laboratory of Soil Mechanics (LMS) of EPFL. Cylindrical specimens of 55 mm of diameter and 15 mm of height were used. The material used in the experiments is a mixture of quartz sand and MX-80 bentonite in proportion of 80/20 in dry weight. MX-80 (Wyoming) bentonite is a Na bentonite which contains around 80% of montmorillonite. A special procedure was established, and systematically applied, to obtain homogeneous samples with desired dry density. The two components of the mixture were first sieved at 0.5 mm and mixed together until getting a homogeneous mixture. Samples were statically compacted at a constant axial displacement rate of 0.5 mm/min and to the target water content of 11% (corresponding to the optimum water content for energy of compacting equal to 3.4 J/cm3). Once the sample was compacted, it was placed inside the triaxial cell and wrapped in two latex membranes in order to prevent any leakage of either water or gas from the lateral surface. The tests composed of two main phases, the saturation and the gas injection. The sample was saturated with distilled water under a low pressure gradient by injecting water from both top and bottom drainages by means of two pressure volume controllers (PV controllers). In this phase the water volume exchange was recorded by the data acquisition system connected to the controllers and the saturation of the sample was assessed by analysing the water volume exchange through the sample. The saturation phase was carried out under a confining pressure of 300 kPa which is higher than the sample swelling pressure. The gas penetration tests were performed on fully saturated samples by injecting dry air, maintaining a constant confining pressure of 300 kPa. In the used configuration the gas was injected from the bottom of the sample while the water pressure was kept constant at the top (20 kPa). The gas pressure was initially set slightly higher than the initial water pressure in the sample and then increased in steps until a significant change in the gas penetration rate was observed. The downstream gas pressure was imposed by means of a pressure regulator whereas the upstream water volume exchange was monitored by using the PV controller. Once the gas pressure reached 55 kPa, a sudden volume increase was monitored by the PV controller connected to the top of the sample and air bubbles were observed traversing the tube connected to this controller. This indicates the creation of a continuous gas pathway through the sample. Once the first cycle was ended, the sample was separated from the gas injection system and it was re-saturated. Then a second cycle of the gas penetration test was performed in similar conditions as the first cycle. It was observed that for the second cycle the sudden increase in gas flow rate was obtained at a pressure value very close to the previous test result. This result suggests that the previous gas penetration as a continuous phase did not modify in a significant way the gas pressure value needed for a sudden change of the gas penetration regime. For a better interpretation of the obtained results, the experimental set up is being modified in order to assess the volume change of the sample during the gas propagation. Tests are underway for a systematic analysis of the effects of previous gas injection cycles on the gas pressure thresholds