It is well known that depleting a hydrocarbon reservoir can redistribute the in-situ stresses sufficiently to reactivate and induce slip in nearby faults, which otherwise usually evolve over geologic periods. In this paper, we develop a method to calculate the redistributed stresses around a depleted, heterogeneous reservoir with non-uniform pressure distribution. Using this method, we analyze the stability of nearby faults and show that different reservoir depletion strategies can affect the fault stability differently. The reservoir is modeled as a thin poroelastic inclusion in an elaStic matrix while the fault gouge obeys Mohr- Coulomb failure criterion. The developed method allows considering the reservoirs and faults to be of arbitrary shapes and matedHal properties while the pressure inside the reservoir does not have to be uniform. After the fault is reactivated, it is modeled as a shear (mode II) fracture. Displacement discontinuity between the sides of this fracture gives an estimate of the fault slip magnitude.
Geological discontinuities such as faults are inherent in most petroleum formations. It is well known that changing fluid state in a hydrocarbon reservoir can redistribute in-situ stresses sufficiently to reactivate and induce slip in nearby faults, which otherwise usually evolve over geologic periods. This work centers on analyzing the stability of faults due to reservoir depletion. From the engineering view point, the consequences of fault reactivation can range from shearing of the boreholes (drilled through the fault zone), to dynamic release of the stored elastic energy and induced seismicity, to drastic change of the formation permeability and the production-depletion strategies. From the scientific viewpoint, the addressed phenomenon represents a clear and robust demonstration of the importance? of poroelastic effect (in this context first suggested by Geertsma, 1966). Indeed, the stability of the fault is not affected at all if the poroelasticity of the reservoir material is not taken into account.
Fault reactivation due to stress redistribution caused by natural resource recovery is not unusual and represents a typical example of human-induced seismicity. In mining operations, it is well known that man-made underground openings (cavities) redistribnte remote stresses in their vicinities affecting the stability of near-by faults and triggering seismic events (e.g., see Ortlepp, 1997). Similarly, in the petroleum industry, fracture reactivation is sometimes attributed to small-scale failure associated with drilling boreholes (e.g., Mokhel et al., 1996; Thiercelin and Atkinson, 1996). In both cases, open cavities probably represent the extreme case of real situation maximizing the stress disturbance. However, filled cavities (inclusions) can also redistribute stresses sufficiently for the fault to slip. This is of importance for hydrocarbon recovery when the solid matedHal is not actually removed (mined), but the initial state of stress is disturbed by the withdrawal of subsurface fluids (e.g., Geertsma, 1966; Segall, 1989; Addis et al., 1998; Rudnicki,
1999).
Probably the largest seismic events triggered by gas extraction are three major earthquakes (with magnitude M> 7) reported near the Gazli (Uzbekistan) gas field in 1976-1984 (e.g., Simpson and Leith, 1985). The largest registered earthquake triggered by the oil withdrawal is probably the 1983 Coalinga (California) earthquake with M=6.5 associated with production from the Anticline Ridge oil fields (McGarr, 1991). Detailed description of these and many other examples of extraction- induced seismicity can be found in the reviews of Yerkes and Castle (1976) and Nicholson and Wess
