Search

Safe and permanent CO₂ storage in geological formations requires reservoir geomechanical stability. Injection of CO₂ into the subsurface changes the local pore pressure and, further, alters the effective stresses due to poro-thermo-chemo-mechanical coupled responses. Changes of pore pressure and effective stress may disrupt the host formation mechanical equilibrium. This alteration may result in geomechanical failure events such as fault reactivation and hydraulic fracturing. Such events can favor fluid migration paths for injected CO₂, induce seismic activity, and cause surface uplift. Examples of field observations during CO₂ injection include: (1) surface uplift at the In Salah project in Algeria, (2) absence if bottom-hole pressure (BHP) increase during injection in Cranfield, Mississippi, and (3) induced seismicity with magnitude M>1 in Decatur, Illinois. In this context, accurate estimations of pore pressure build up and local stress alteration induced by CO₂ injection are critical to avoid geomechanical perturbations. However, current models and predictions often assume relatively homogeneous reservoirs without taking into account compositional behavior. Further, the effects of temperature and chemical reactions have not been rigorously incorporated into the interpretation of local stress alteration and the well response to CO₂ injection. This dissertation shows geomechanical analyses of CO₂ geological sequestrations by three field case studies: Frio CO₂ sequestration pilot test in Texas, Cranfield CO₂ sequestration in Mississippi, and Crystal Geyser in Utah. Both Frio and Cranfield case studies are studied with the help of reservoir simulation and history matching of field data including assimilation of vertical heterogeneity from well-logging analysis and calibration with laboratory experiments. The Frio case study focuses on examination of reservoir capacity of a compartmentalized volume to avert fault reactivation. The Cranfield case study analyzes the influence of thermo-chemo-elastic processes on wellbore fracturing induced by CO₂ injection. The Crystal Geyser case study investigates the long-term chemical effects of CO₂-charged brine on rock mechanical properties through analyses and measurements on rock samples from the field, where a natural CO₂ leakage analog exists. The following conclusions are a result of this dissertation. CO₂ dissolution into brine reduces pore pressure build up significantly in small and compartmentalized reservoirs. Thermo-elastic and chemo-elastic effects alter local stresses and may trigger injector fracturing at bottom-hole pressures lower than expected. Capturing phase behavior, coupled thermo-chemo-mechanical processes, and reservoir heterogeneity are important factors to estimate reservoir capacity and prevent geomechanical perturbations.

Diagnostic fracture injection tests (DFIT) are used to estimate the magnitude of the minimum principal stress, which is assumed to be equal to fluid pressure at the moment of fracture closure. A small volume of water is injected into formation to create a fracture, the well is shut in, and eventually the fracture closes. The pressure during shut-in can be analyzed by several pressure transient methods to pick the time of fracture closure. Pressure at that time is taken to be the fracture closure pressure. In this study, DFIT simulations are performed with a fully numerical hydraulic fracturing simulator. Sensitivity analysis is done to investigate how reservoir parameters such as fracture toughness, permeability, fracture stiffness, and the magnitude of the minimum principal stress impact the DFIT pressure transient. Based on these insights, we use the simulator to match a DFIT pressure transient from a low permeability formation. We analyze the field data with conventional methods for picking closure. The simulation matches suggest that the conventional methods can underestimate the closure pressure in low permeability formations. Based on our results, we propose a new method for picking fracture closure based on the evolution of fracture compliance during closure. Our simulations provide insight into how the fracture compliance impacts to pressure transient. Assuming the closure pressure from our simulations matches to the data are correct, our proposed method picks the correct closure point. This study includes simulation matches to the field data with simulations that use complicated fracture geometry, which may be realistic in some formations. The multiple fractures cases have similar pressure transients and similar reservoir parameters as the single hydraulic fracture simulations, indicating that network complexity will not necessarily be evident from the pressure transient. In the future, DFIT simulations with more complex fracture geometries will be conducted to understand how fracture geometry affects the DFIT pressure transient.