Discontinuum-based numerical analysis of thermoshearing in planar and rough granite fractures: A comparison with laboratory observations

In deep geological repositories (DGRs), decay heat from spent nuclear fuel raises the temperature, causing thermal stress and shear slip in pre-existing fractures in crystalline rocks, a phenomenon known as thermoshearing. An in-depth understanding of the thermoshearing mechanism in fractured rock masses is crucial for assessing the safety and stability of DGRs, as irreversible increase in permeability due to fracture shear slip can elevate fluid flow containing radionuclides and can compromise the integrity and performance of engineered barrier systems. This study presents a numerical reproduction of the laboratory thermoshearing experiments conducted on granite samples with planar and rough fractures using the three-dimensional discontinuum-based coupled code, TOUGH-3DEC. The calculated and measured temperatures at the heater and rock samples show reasonable agreement. The numerical models effectively capture the characteristic fracture shear behavior related to fracture roughness, as observed in the experiments. Irregular fracture roughness and relatively greater differential stress conditions applied in the rough fracture (RF) model induce shear slip before heating, differentiating the initiation timing of shear slip in the planar fracture (PF) and RF models. After heating, the maximum shear displacements for the PF and RF models are 68 μm and 74 μm, respectively, occurring near the top and bottom boundaries, where the temperature increase is the greatest. Among the five locations analyzed by digital image correlation, the largest shear displacement but the smallest dilation occurs at a specific location, indicating that intensified stress concentration at local protruding contacts can restrain the fracture dilation. The seismic events calculated using the developed seismic analysis algorithm show similarities with the measured acoustic emission events in terms of relative size and temporal distributions.