Improvement of the excavation damaged zone in saliferous formations. Phase II. Final report

In Germany, salt formations are considered to be suitable to host a deep geological repository for radioactive waste. However, local stress changes adjacent to man-made openings lead to the evolution of an excavation damaged zone (EDZ) during and after excavation. Such an EDZ can have a major impact on the operation of a radioactive waste repository since it represents a region where progressive failure occurs. This decreases the material strength and thus increases the permeability of the originally tight host rock. The objective of this investigation is determined by the need to develop a modeling strategy that can be applied to simulate the permeability increase due to mechanical deterioration of rock salt, in particular that occurring in the EDZ. The identification of material parameters is often conducted by back-calculation of laboratory experiments. However, standard laboratory tests, e.g. compression tests, are only applicable to provide information about the macroscopic deformation. For this purpose, further laboratory tests were conducted to derive material parameters used in the constitutive models. Combined acoustic emission and uniaxial compression tests as well as microstructural analyses were carried out to dissolve the macroscopic behavior micromechanically. The information obtained was then used for parameter identification utilizing optimization methods. The objective was to identify the best estimate of the micro-parameter values that can be applied to simulate the laboratory results performed. To make a qualitative comparison between the numerical analysis and the acoustic emission (AE) testing, the onset of failure at contacts was equated with events detected by AE testing. The onset of failure is identified at stress levels above 3 MPa. The number of tensile fractures increases continuously and the maximum is reached between 10 and 12 MPa axial stress. At the beginning of loading, tensile fractures appear more frequently than shear fractures, which is in good agreement with laboratory results. It also has been shown that the correct mechanical behavior is mainly achieved by an adequate parameter set controlling grain deformation and deformation along the grain boundaries. The calibrated parameter combination was then used to generate realistic fracture networks occurring in the EDZ. For this purpose, the excavation of a drift is simulated with continuum mechanical approaches and the zone stresses are recorded in certain parts and for certain depths of the EDZ. The zone stresses are then transferred as boundary stresses on a discontinuum model that comprises polyhedral elements to simulate the fracture development at grain scale. The results of the DEM modelling show a fracture-induced anisotropy and fractures are oriented parallel to the maximum principal stress. It could be shown that the most damaged region is at a depth of 0.5 m from the drift contour. Between 2-3 m depth the fracture density decreases considerably. The permeability increase in the EDZ can compromise the sealing function of the originally tight rock salt. Therefore, the specification of adequate permeability values is mandatory since it provides information about major pathways for fluid flow. However, the larger number of fractures makes the discontinuum approach less efficient for large-scale modelling and methods are needed to upscale the information gained from discontinuum modelling. The models must be calibrated which is generally similar to the procedure used for calibrating the mechanical input parameters: fracture transmissivities are varied until conformity between measured and simulated results is achieved. However, it is very difficult to obtain relevant hydraulic information, and there was no data available that was suitable for calibration. The transmissivity through fractures is based on the cubic law that considers fluid flow between smooth-walled plates with a given aperture. For a first estimate, the aperture of fractures was therefore adjusted based on information coming from microscopic investigations. The anisotropy of the fracture pattern is also reflected in the hydraulic conductivity tensor. An anisotropy factor of 1.73 can be specified at a depth of 0.5 m from the drift contour. For depths of 0 m and 2 m no tensor ellipsoid could be specified since the simulated fracture connectivity is below the percolation threshold. In the first phase of the project, in situ experiments were conducted in order to verify the sealing injection technique using liquid silica. In situ experimental injection tests of liquid silica were performed in a small area of the EDZ in the mine Niedersachsen Riedel in Germany. The liquid silica was mixed with the fluorescent dye Uranin to visualize the silicate in rock salt. Once the work was completed and after a waiting period of three months, drill cores were taken from the injected host rock and thin sections were prepared to inspect the fracture system microscopically. The last part of the study is about the long-term stability of liquid silica improved rock salt samples. A laboratory research program was conducted during the first phase of the project which considered an injection of rock salt with sodium silicate solution. The influence of the contact with NaCl and MgCl₂ solutions on the mineral content and the composition of the solutions were investigated. These studies have been completed by analyzing the solubility of amorphous silica in saturated NaCl solutions with MgCl₂ of varying concentration. Additionally, it was tested whether amorphous silica is able to act as a chemical barrier by analyzing the reaction behavior of selected trace elements in contact with amorphous silica. In the course of the experiment, a decrease in the concentration of several elements is observed. This means that amorphous silicate may act as a chemical barrier for specific elements in the saline environment.