by Sohrab Gheibi, BIGCCS PhD Student, NTNU
The huge amounts of CO2 that we are emitting to the atmosphere require rapid action to mitigate climate change. A promising option to reduce emissions of this greenhouse gas is storage in deep geological formations. CO2 storage can also serve for utilization purposes, like Enhanced Oil Recovery (EOR). Safe storage of CO2 is a key aspect of carbon dioxide sequestration projects. Public concerns related to possible leakages and induced seismicity due to fault reactivation are of the main factors that can lead to accomplishment or failure of CCS projects, especially in large scales.
Supercritical CO2 is injected into the deep geological formations. In a simple language, the stresses inside and outside the reservoir are carried by both the rock (solid phase) and the existing brine (liquid phase). It is generally assumed that the formation and the brine are in equilibrium. Injection of CO2 increases the pore pressure of the formation which consequently decreases the portion of stress carried by the solid phase. Therefore, the compressed reservoir rocks unload and expand. Expansion of the rocks induces changes in the total stresses and both the reservoir and its surrounding rocks including cap rock experience a new stress distribution. The new stress distribution could improve or decline the stability condition both inside and outside of the reservoir. Different forms of instabilities may occur in CO2 storage reservoir and cap rock system. Hydraulic fracturing (tensile failure) or shear failures like reactivation of pre-existing faults/fracture or even creation of new faults/fractures can occur both inside and outside of the CO2 reservoir.
The aim of my PhD work is to investigate the likelihood and conditions of occurrence of each of the failure forms and geomechanical modeling of CO2 storage reservoir and generally fluid injection into the geological formations. To do so, an in-house code called “MDEM” is used. This code is a hybrid FEM/DEM code which is capable of modeling continuum in elastic and discontinuum after failure.
One of part of my studies is the investigation of the stress changes in different reservoir conditions. There are a number of analytical solutions to estimate the stress changes due to pore pressure change inside the reservoir.
Faults and generally discontinuities are sources of weakness and changing of pore pressure can lead to their reactivation and creation of leakage pathways and probably felt seismic activity.
Due to the injection, fault planes can deform and the deformation does not necessarily mean failure. Since there is pore pressure contrast between the reservoir and the cap rock, the fault plane which might have cut both the reservoir and the cap rock will deform differently, meaning that the portion of the fault plane in the cap rock may deform either in different direction or different amount compared to its portion inside the reservoir. This contrast may lead to a complicated stress path around the fault in the interface of the reservoir and the cap rock.
However, it is generally assumed that the faults do not affect the stress changes. Our analyses on the other hand have shown that presence of faults may influence the stress redistribution and it is significant in some cases. The intensity of its effect depends on the insitu stress regime, geometry of the reservoir, geometry and mechanical properties of the fault, mechanical properties of the reservoir and cap rock etc.
Fig. 1 indicates the model used in the analysis. The reservoir is a rectangular reservoir with 40 m thickness crossed by a 600 fault at its center. Analysis indicates that the pressure build-up can significantly change the total and effective stresses. These stress changes are more severe when faults exist in the formation (Fig. 2).
Pore fluid pressure variation due to injection changes the total stress; therefore, the shear stress acting on the fault/joint planes also changes. This causes deformation of fault face with respect to each other, both in normal faulting and reverses faulting stress regimes. Due to the pore pressure difference between the cap rock and the reservoir, the elastic/plastic deformation of the fault plane induces a different stress path around the fault (Fig. 2).
The plot of the Mohr circle as the stress state representor vs. failure criterion which is Coulomb criterion in most of the cases is used to simply check the safety of the storage. Like all circles, Mohr circle has a center point indicating mean effective stress and a radius indicating the deviatoric or shear stress. Each point on the circle shows the acting normal and shear stresses on a plane with a specific angle. The blank colored line is Coulomb failure criterion. If the circle touches/cuts the failure line it means that failure has occurred theoretically. In other words, closer the Mohr circle to the failure line means the greater risk of failure. Therefore, if we project the information of the stress values in Fig. 2 into Fig. 3, it is clear that, the reservoir and the cap rock experience a greater decrease in mean effective stress (translating to the left in Fig. 3) and increase in deviatoric stress (Mohr circle enlargement) in the footwall and hanging wall of faults in reverse and normal faulting stress regimes, respectively (Fig. 2). The stress path brings the fault vicinity to a different stability condition both in the reservoir and the cap rock (Fig.3). it is to be noted that the Mohr circles belong to the most unstable region of the model.
What Fig. 3 basically indicates us is that; the reservoir and the cap rock are more unstable if we consider the effect of faults on stress redistribution. Our analysis indicates that exclusion of this effect can mislead us either to be more optimistic or pessimistic in some conditions. One should be aware that a fault itself may be stable but its presence as a geometrical feature can affect the stress path causing the rock around the fault to be unstable.
Sohrab Gheibi is a PhD student at NTNU, his work is linked to BIGCCS Task 3.4, and he is planning to finalise his thesis by the end of 2016.