Understanding of shear reactivation of pre-existing discontinuities for brittle host rocks is an area of considerable interest for radioactive waste disposal. In particular, the potential for existing features to undergo shear displacements and related changes in permeability as the result of coupled thermal, mechanical, hydraulic and chemical effects can all have significant impacts on repository safety functions (e.g., creating permeable pathways or, for very large displacements, mechanical damage of waste packages).
The purpose of Task G under DECOVALEX-2023 is (see Fig. 1 for Task structure):
The emphasis of this task is at the laboratory scale, using well-designed experiments to link micro-scale THM(C) effects acting on fracture surfaces and asperity contacts with emergent fracture properties such as permeability.
The experiments will be conducted on typical granite samples with pre-existing well-characterized discontinuities (Fig. 2). The experiments are complementary allowing inclusion and exclusion of hydraulics permitting a clear separation of key processes. The experimental work will focus on characterizing displacements and permeability changes resulting from different thermal, mechanical and hydraulic loads on the system. It is expected that research teams will apply and develop existing constitutive models for fracture characterization and hence improve fundamental physical understanding of these complex processes as well as improving modelling predictive capabilities (Fig. 3).
The program of HM experiments and modelling challenges has been built up with increasing complexity both in terms of the processes involved, and in terms of moving from more simple homogeneous artificial samples through to real rock samples with clear fabric and texture. Initially the mechanical response of the fractured samples are tested, the deformation measured using a circumferential fibre optic strain gauge. Then the sample permeability relationship to the 3D stress field is experimentally determined, as the sample is placed under various controlled principal stress magnitudes and orientations creating different normal and shear stress across the fractures. Fracture profile is provided by detailed fracture topography scans. Finally a fractured heterogeneous granite is placed under quasi-realistic stress conditions found in a generic repository and the teams are asked to blind predict the mechanical and couple hydraulic mechanical results using the knowledge gained from the completion of the earlier parts of this task.
Laboratory TM experiments are conducted at KICT. The TM experiment program is designed in order find an answer to the following key scientific question: Can shear slip of a stressed fracture be induced by long-term heat effect, and if so, what are the factors that mostly influence the shear slip process? The experiment is designed to start with simple mechanical and thermal boundary/initial conditions on a 100 mm side-length cubic granitic rock specimen with natural, sealed fracture using true-triaxial loading system. Thermally induced fracture slip is characterized by acoustic emission (AE) monitoring and micro X-ray CT imaging. The latter part of the TM experiment is designed with complex mechanical and thermal boundary/initial conditions and fracture heterogeneity. Numerical modelling will start first with a blind prediction of KICT experiment, and will be followed by test case modelling and up-scaled benchmark modelling.
The current plan of Task G includes 4 distinct steps which are displayed in Fig. 1.
Fracture characterisation and fracturing processes: High resolution fracture surface scan data before and after mechanical treatment (Pötschke et al. 2020a,b) will be used for characterising fracture properties (apertures) as input information for modelling Steps 2 and 3. Various numerical approaches for modelling mechanical shear will be evaluated and compared.
Modelling of hydro-mechanical responses of a fractured granite sample under various controlled principal stress magnitudes and orientations including permeability changes.
Modelling of a thermo-mechanical lab experiment and predictions for up-scaled benchmarks.
A combination of Step 2 and Step 3 concepts into a new THM benchmark is planned for the second half of D-2023.
The modelling campaigns for the individual steps consist of Benchmark Exercises (BE), Experimental Analysis (EA), Blind Predictions (BP), and Up-scaling concepts (UP). The BEs for all steps will be executed on predefined cube and cylinders for subsequent M>HM>TM>THM analyses (Fig. 3).
For further information, please contact the task leaders Olaf Kolditz, Chris McDermott, and Jeoung Seok Yoon.
More details may be found on the Task G planning document at Overleaf which will be continuously extended.