Modelling Thermo-Hydro-Mechanical THM Interactions Affecting Evolution of Injection Induced Seismicity in Geothermal Reservoirs

Injection-induced seismicity has been one of the major obstacles in front of developing geothermal resources and a source of disturbance in populated locations. Many efforts have been made to understand underground processes leading to induced seismicity in order to control and mitigate it during thermal energy extraction. My thesis aims to investigate different factors, related to the operating and geologic conditions, leading to the occurrence of seismic events, and to derive optimum operating procedures to mitigate induced seismicity based on the proposed model. During my thesis work, I developed two conceptual models using the finite volume code Itasca FLAC3D to model thermo-hydro-mechanical processes in the reservoir. The models are verified by solving pressure transient and displacement equations analytically using the built-in FISH programming language and comparing solutions with numerical results. Heat transfer by both convection and conduction are considered, and anisotropic state of stress and uniaxial strain conditions and are applied as the initial and boundary conditions. First model is of a fault-free fractured reservoir to study the evolution of seismic cloud or what is called "triggering front". I used FISH language to introduce two sets of fracture planes with different random orientation distributions, and used Mohr-Coulomb criterion to evaluate stability at each time step, where slip along a single fracture is taken as a potential microseismic event. Simulated time ranges from hours to days to model pressure transient and poroelastic stresses changes around the injection well, and a number of simulations of long-term operation to model thermoelastic effects and temperature depletion. Results show that thermal stresses might contribute to seismic events only in the long-term operation (weeks to months) because of the low thermal diffusivity and low permeability of reservoir rocks. Furthermore, elastic moduli play a critical role in determining the poroelatic behavior of rock and fractures as they control the stress path; for example, simulations reveal that specific ranges of shear and bulk moduli used in defining the model could stabilize fractures during injection, while initiate slip during production. This reflects the importance of precise rock characterization in modeling specific reservoirs. Furthermore, it is found that seismicity triggering front is related to the hydraulic diffusivity of the reservoir and affected by the onset of seismicity, while the density of seismic events within the triggering front is found to be affected by the initial fractures status and injection rate. Mitigating induced seismicity can be achieved by establishing constraints on the down-hole injection pressure given that a robust model of underground faults and joints for a specific reservoir is available. In the second part of my work, I built coupled fluid-flow and geomechanical model to evaluate the behavior of a nearby normal fault in response to stress state perturbations due to water injection. Friction coefficient is allowed to evolve according to rate-and-state theory. Different scenarios of frictional behavior including velocity-neutral, velocity-weakening and velocity-strengthening behavior, are examined. Also, parametric analyses are performed on the injection operation parameters and the associated poroelastic and thermoelastic effects.