Water Coning, Water, and CO2 Injection in Heavy-Oil Fractured Reservoirs

Abstract

Summary In this work, we investigate challenges related to the recovery of heavy viscous oil from reservoirs with a dense network of fractures and vugs but with a tight matrix. Gravitational drainage of oil from the tight matrix through water injection is ineffective because of the high oil viscosity and density. To further complicate matters, we consider a strong underlying aquifer, and there is considerable risk of water coning around producing wellbores caused by the low water viscosity. To model potential recovery strategies, we carry out simulations with our higher-order finite-element (FE) compositional multiphase-flow reservoir simulator. Discrete fractures are represented through the crossflow equilibrium (CFE) approach. Phase behavior and phase-split computations are modeled with the cubic-plus-association equation of state (EOS). Fickian diffusion, facilitating species exchange between gas in fractures and matrix oil, is modeled through chemical potential gradients. First, we validate our simulator by modeling a set of laboratory experiments in which water is injected in a fractured stack saturated with oil. The experiments investigate the effects of capillary pressure and injection rates on oil recovery, and show that, at low injection rates, capillary imbibition of water from the fractures into the matrix blocks is extremely efficient. Simulations with our 3D discrete-fracture model show excellent agreement with the experimental results without parameter adjustments. Next, we consider the detrimental effect of water coning when oil is produced without injection by carrying out a parameter study investigating the impacts of different (1) water–oil mobility ratios, (2) matrix and fracture wettabilities, (3) matrix permeabilities, (4) domain sizes, (5) production rates, (6) well types and placement, and (7) a local viscosity-reduction treatment around producing wellbores. We find that the only approach to partially mitigate coning is to produce at low rates from perforated (and potentially multilateral) horizontal wells. As an alternative production strategy, we then model carbon dioxide (CO2) injection in two and three dimensions, and compare to results from a commercial dual-porosity simulator. CO2 has a high solubility in this oil, and dissolution leads to volume swelling and a large reduction in oil viscosity. In combination with the much higher density difference between the phases, the latter improves gravitational drainage. We find that a significant amount of matrix oil can be produced in addition to oil from fractures and vugs, and with a lower risk of water coning.