Immiscible Viscous Fingering at the Field Scale: Numerical Simulation of the Captain Polymer Flood

Abstract

Immiscible fingering in reservoirs results from the displacement of a resident high viscosity oil by a significantly less viscous immiscible fluid, usually water. During oil recovery processes, where water is often injected for sweep improvement and pressure support, the viscosity ratio between oil and water (µo/µw) can lead to poor oil recovery due to formation of immiscible viscous fingers resulting in oil bypassing. Polymer flooding, where the injection water is viscosified by the addition of high molecular weight polymers, is designed to reduce the impact of viscous fingering by reducing the µo/µw ratio. A considerable effort has been made in the past decade to improve the mechanistic understanding of polymer flooding as well as in developing the numerical simulation methodologies required to model it reliably. Two key developments have been (i) the understanding of the viscous crossflow mechanism by which polymer flooding operates in the displacement of viscous oil; and (ii) the simulation methodology put forward by Sorbie et al. (2020), whereby immiscible fingering and viscous crossflow can be simply matched in conventional reservoir simulators. This publication extends the work ofBeteta et al. (2022) to conceptual models of a field case currently undergoing polymer flooding – the Captain field in the North Sea. The simulation methodology is essentially “upscaled” in a straightforward manner using some simple scaling assumptions. The effects of polymer viscosity and slug size are considered in a range of both 2D and 3D models designed to elucidate the role of polymer in systems both with and without “water slumping”. Slumping is governed by the density contrast between oil and water, the vertical communication of the reservoir and the fluid velocity and, when it occurs, the injection water channels along the bottom of the reservoir directly to the production well(s). It is shown that polymer flooding is very applicable to a wide range of reservoirs, with only modest injection viscosities and bank sizes return significant volumes of incremental oil. Indeed, oil incremental recoveries (IR) of between 29–89% are predicted in the simulations of the various 2D and 3D cases, depending on the slug design for both non-slumping and slumping cases. When strong water slumping is present the performance of the polymer flood is significantly more sensitive to slug design, as alongside the viscous crossflow mechanism of recovery, a further role of the polymer is introduced – sweep of the ‘attic’ oil by the viscous polymer flood, which is able to overcome the gravity driven slumping and we also identify this mechanism as a slightly different form of viscous crossflow. In slumping systems, it is critical to avoid disrupting the polymer bank before sweeping of the attic oil has been performed. However, as with the non-slumping system, modest injection viscosities and bank sizes still have a very significant impact on recovery. The conceptual models used here have been found to be qualitatively very similar to real field results. Our simulations indicate that there are few cases of viscous oil recovery where polymer flooding would not be of benefit.