This thesis examines the geomechanical behaviour of heterogeneous carbonate rocks under laboratory conditions, with a focus on how contrast in lithologies influences fault reactivation and implications for induced seismicity in geothermal reservoirs. As there is significant geothermal potential in carbonate reservoirs in north-western Europe, geothermal energy extraction will be crucial for the energy transition. However, human interference with the subsurface is never without risks. In the past, numerous seismic events have been associated with subsurface activities in carbonate rocks. To better understand these risks, laboratory setups can be used to study these effects on a smaller scale.

We utilised two distinct carbonate rocks in this thesis: the Dinantian carbonate (lower porosity) and Indiana limestone (higher porosity). We tested these intact rocks in a series of uniaxial and triaxial compression experiments to characterize the rock properties mechanically. From these rocks, we created laboratory faults at an angle of 35° for each lithology and a heterogeneous configuration where the lithologies were mixed. In our displacement-driven experiments, we investigated the impact of surface roughness and juxtaposition of the mixed configurations. For the injection-driven experiments, only one kind of roughness is investigated, along with juxtaposition and pore pressure on fault stability.

The results on intact rocks in both confined and unconfined conditions learned us that the stiffness of Dinantian carbonate samples is the highest among all the samples. Hence, the matrix itself accommodates minimal deformation until failure occurs, either through the development of a shear fracture under confined conditions or via axial splitting in UCS tests. For the Indiana limestone, we observed a more compliant behaviour characterised by strain hardening and compaction under confined conditions.
In the displacement-driven experiments, this translated into the highest critical Mohr-Coulomb stresses for the Dinantian carbonate, independent of roughness. Dependent on the type of roughness, either the Indiana limestone or the mixed samples with increased roughness exhibited the lowest critical Mohr-Coulomb stresses. The results indicated that the reactivation stresses tend towards the more compliant lithology in a heterogeneous configuration.
In our injection-driven experiments, we observed an opposite trend where reactivation in the Dinantian carbonate tends to reactivate with significantly less pore pressure compared to the Indiana limestone and the heterogeneous configurations. This results from the difference in fracture flow in the Dinantian carbonate versus matrix flow in the Indiana limestone. Fracture flow results in a local distribution of pore pressure, which causes early reactivation compared to the more equal distribution along the fault in the Indiana limestone. In heterogeneous configurations, we observed again a behaviour that tends towards the compliant lithology, as the pore pressure can distribute evenly over the sample in the Indiana limestone part.

This implies that induced seismicity appears to be a higher risk in a critically stressed fault zone with less porous carbonate rocks that exhibit secondary permeability in the form of fractures.