A novel numerical analysis is described, in which the steam-assisted gravity drainage (SAGD) recovery process in bituminous oil sand is studied. A geomechanical/reservoir simulator was modified to incorporate the absolute permeability increases resulting from the progressive shear dilation of oil sands. The objective was to obtain a realistic prediction of shear dilation, as the oil sands approached failure and beyond, and the concomitant increases in permeability. Changes in the in situ stresses that caused this dilation were due to the combined effects of reduced effective stress with high-pressure steam injection, and increased deviatoric stress with thermal expansion under lateral confinement. The resultant volumetric strains were used to modify the absolute permeability characteristics of the oil sands as the SAGD process progressed. The spatial and temporal growth of enhanced permeability zones resulted in an accelerated steam chamber growth.
The relationship between volumetric strains and absolute permeability changes was obtained from existing laboratory data on quality specimens of non-bituminous Athabasca oil sands. The source sample was obtained from an outcropping of the McMurray Formation, thus avoiding most of the sample disturbance associated with unconsolidated core obtained conventionally. Under triaxial loading, the resultant volumetric strains increased absolute permeabilities by a factor of 4 to 6.
The analysis is innovative in that the model used an effective stress approach, and used the volumetric strains to modify absolute permeabilities. Thus, the encroaching SAGD steam chamber was found to modify the stress regime, which in turn modified the permeabilities within the reservoir. Geomechanical enhancement of the SAGD process was found to be a significant beneficial effect, and would be increased by operating the SAGD process at higher injection pressures.
Conventional reservoir simulations of thermal recovery processes in heavy oil and bituminous oil sands do not explicitly incorporate geomechanics. However, they implicitly include geomechanics since input permeabilities are air permeabilities obtained from core plugs. Due to the unconsolidated structure of these sands, core plugs are highly disturbed1. Porosities are typically 120% to 130% of the in situ porosities, as determined by petrophysical logging and other means. Specimens tested at overburden stress indicate that this disturbance results in permeabilities that are four times higher than in situ liquid permeabilities, on average.
Fortuitously, these inflated permeabilities are comparable to those in situ, once the oil sand reservoirs have been disturbed. This is due to the shearing and dilation resulting from stresses altered by the recovery process itself. This is evident from history matches of SAGD projects, such as the UTF2, which used permeabilities obtained conventionally. Note that the limited number of operating SAGD projects have all been operated at relatively high injection pressures, relative to the reservoir depth, which result in low effective stresses. Therefore, while the implicit use of geomechanics, through the use of high absolute permeabilities, does work for the existing SAGD projects, this methodology cannot necessarily be extrapolated to other reservoirs with different operating conditions.
Although history matches and predictions have been successfully conducted without geomechanics, these have been done without a complete accounting of the underlying physics. This is acceptable practice if simulating reservoirs and operating conditions comparable to those of current successful SAGD projects. However, a more rigorous approach was taken in this study, in which the geomechanics of the SAGD process were explicitly included in a combined geomechanics and reservoir simulation.