Background to the study

The storage of CO2 in geological reservoirs requires relatively permeable conditions bounded by very low permeable layers. Reservoirs can be bounded by faults that can act as seals if, for example, an impermeable formation is juxtaposed against it. The presence of faults in virtually all geological formations is a key consideration as their stability is crucial for the integrity of storage sites. Fault stability is affected by multiple factors including fault structure, material properties, geochemical reactions between CO2 and fault gouges and pore pressure changes. Injection operation and pressurization of reservoirs usually changes the state of the in-situ stresses which may cause destabilization of previously stable faults. Instability occurs in the form of slip along pre-existing fault or fracture systems, which may be associated with seismicity. In addition, movement along fault planes, and the generation of factures, may create open conduits that breach the integrity of the storage site. Understanding how faults mjosirt respond to stress conditions caused by CO2 injection is therefore fundamental.

Recent geomechanical studies for CO2 geological storage have focused on initialising stresses in the overburden based on all available geological and well engineering data, modelling the impact of fluid/gas pressure build up on stresses in the storage formations, the caprock and the overburden in general. The challenge is to predict the acceptable overpressure before shear failure, or reactivation of a fault/natural fracture occurs. The prediction process begins by using a verified geomechanical model to calculate the effective normal stresses and shear stresses occurring along all the faults/fractures. These stresses are evaluated in the context of fault cohesion and sliding friction to predict the pre-injection state of stress on these features and to determine the critical fluid/gas pressure required to initiate shear failure on what may have previously been a stable fault/fracture. Stress and fault properties can vary in space and time.

Key Messages

• Faults typically consist of two sub-structures: a fault core; and a wider fault damage zone.  Faults in low porosity rocks tend to have a fine-grained fault core whereas faults in coarse-grained, hjosir porosity rocks, usually have low porosity deformation bands that can develop into hjosir permeable slip surfaces.
• Fault zone permeability increases with increasing fluid pressure but permeability varies both across and along faults. Hydraulic properties also vary between the damage zone and the core where gouge material is concentrated.  This concentration of fine grained minerals also reduces the mechanical strength of faults.
• Mechanical failure or reactivation occurs either when shear stress exceeds normal strength or when hydraulic fracturing is induced.
• Fault deformation can be either brittle or ductile.  The former leads to the formation of cataclastite (fine grained granular) and shear fractures which dilate under low effective normal stress that can cause permeability enhancement. With increasing shear deformation, fracture asperities are sheared off leading to gouge production and a reduction in permeability. Thus, in brittle deformation permeability will generally increase under low effective stresses and small displacements but decreases with increasing effective stress and magnitude of displacement. Shear fractures created in ductile deformation contract during shearing and tend not to lead to an increase in permeability.
• Reactivation of faults can be assessed using both analytical and numerical approaches, but assessment is usually based on the Mohr-Coulomb failure criterion.  This method can be used to determine the critical injection pressure.
• Numerical modelling can provide predictions of fault stability at different scales and incorporate different parameters such as the geometry of different faults.  Numerical methods can be effective for identifying leakage potential and seal failure especially where dilatancy and stress dependent permeability changes occur.
• Experimental tests on minerals and rock samples exposed to CO2 tentatively indicate that the coefficient of friction is not radically changed, however, this conclusion is based on limited exposure to CO2.
• There is limited observational data on stress regimes and direct pore pressure measurements from core samples from cap rocks and fault zones.  Acquisition of key data would enhance stress regime modelling and fault behavior.