Barnaby Padraig Fryer
The stress state of the subsurface has been shown to have an influence on a number of key processes. For example, the criticality of the stress state indicates how large stress changes need to be before a fault begins to slip, the mean effective stress controls compaction and permeability loss in hydrocarbon reservoirs, the differential stress has been seen in seismology to have a strong influence on the size distribution of earthquake catalogues, with stress state further influencing the affinity of a fault to slip either seismically or aseismically, and the stress state strongly influences the propagation of a hydraulic fracture, with stress jumps being capable of completely halting a fracture's propagation. For reasons such as these, it has long been recognized that the state of stress is an important parameter for subsurface industrial operations such as hydrocarbon production, Enhanced Geothermal System (EGS) stimulation, carbon storage, and hydraulic fracturing.
Generally, however, the state of stress does not remain constant during many of these operations. Decreasing effective stresses have led to induced seismicity during injection, fluid production has led to total stress changes which have induced seismicity, even in areas where pore pressure has decreased, temperature-induced stress changes have led to shear stimulation, and changes in total stress have led to so-called "frac-hits" where hydraulic fractures propagate towards depleted zones. These examples demonstrate an operator's ability to alter the state of stress, yet deliberate attempts to alter the stress state for the benefit of future operations have only been suggested a handful of times.
Here, this idea of altering the state of stress is extended through a series of investigations. It is shown that, because production-induced seismicity is caused by total stress changes associated with the gradient of the produced fluid pressure, a hydraulic fracture which reduces the gradient required for production will also reduce the resulting seismicity. Conversely, compaction which leads to permeability loss will result in higher pore pressure gradients and therefore more induced seismicity. Both of these studies, which also have implications for optimal horizontal wellbore orientation, are investigated through the use of a poroelastic reservoir simulator developed during the thesis. In a following study, it is suggested that high stress path reservoirs are attractive targets for fluid injection, as total stress changes will result in an increased stability despite increasing pore pressure. Further, the concept of stress preconditioning is introduced, such that operators alter the stress state of an EGS reservoir to promote more favorable earthquake distributions. These concepts, which make up the first three chapters of the thesis, all are forms of reservoir management that reduce the risk associated with induced seismicity. Following this, in a numerical study, the idea of stress preconditioning is extended to allow for directed stimulation treatments in EGS's. Finally, the idea of inducing stress jumps through production such that hydraulic fractures do not propagate vertically is evaluated by means of scaling analyses and numerical simulation, with implications for carbon storage.
More than the individual propositions, this thesis promotes the mindset that, where appropriate, attempts should be made to alter the state of stress for the benefit of future operations.
EPFL2020
