| In this work, we develop and apply two-phase upscaling techniques for modeling well-driven, high-mobility-ratio, immiscible displacements and first-contact miscible displacements.; For first-contact miscible displacements, we propose a novel miscible upscaling technique, which comprises two key components: effective flux boundary conditions (EFBCs) and the extended Todd-Longstaff with upscaled relative permeabilities ( k*rj ) or ETLU formulation. EFBCs, which incorporate some approximate global flow information into the local upscaling problems, mitigate inaccuracies introduced by standard procedures (e.g., premature breakthrough). The ETLU formulation modifies the computation of effective fluid properties and k*rj such that bypassed oil that is immobile and unavailable for mixing is properly treated. Using synthetic 2D fields with varying permeability correlation lengths, we demonstrate that our upscaling technique is effective for partially layered systems. We also show that the technique is more accurate than standard methods over a wide range of coarsening factors and for different heterogeneity structures. The upscaling procedure is then applied to a 3D miscible gas injection field study. It is found that the original fine grid must be refined areally to achieve numerical convergence. By considering realistic production scenarios, our technique is shown to accurately reproduce the converged fine-scale solutions. Significant overall speedup factors are obtained. Our technique is thus shown to be useful for practical studies of miscible displacements.; High mobility ratios are often encountered in improved oil recovery processes due to high oil viscosities. A new two-phase upscaling approach for modeling well-driven, high-mobility-ratio displacements is presented. For a local fine region around the well, we apply near well, single-phase (NW1P) and two-phase (NW2P) upscaling procedures. The coarse-scale well indices, wellblock transmissibilities, and relative permeabilities ( k*rj ) are determined such that the fine- and coarse-scale flow rates are in agreement. Away from wells, the k*rj for each coarse block are computed by imposing EFBCs. We assess the performance of this approach by quantifying the upscaling errors over multiple realizations of synthetic 3D models with varying correlation structure and degree of spatial variability, as well as different fluid mobility contrasts and production scenarios. The overall approach (NW1P, NW2P, and EFBC k*rj ) is shown to consistently yield accurate coarse-scale simulation results. |
