| The University of Texas miscible gas flooding, equation-of-state compositional simulator, UTCOMP, has been successfully modified to handle horizontal wellbore simulations. It was subsequently successfully used to make comparisons of field-scale carbon dioxide flooding process performance with several combinations of horizontal and vertical wells and various alternative reservoir descriptions resulting from a systematic method of reservoir characterization, including conditional simulations. Results have demonstrated that under conditions typical of tight West Texas carbonates where the tertiary injectivity of CO{dollar}sb2{dollar} and brine is commonly an important factor in the project economies, horizontal wells could be a solution. On the other hand, we have also demonstrated that horizontal wells are not always better than vertical wells.; We have compared results of coarse mesh simulations with those of fine mesh simulations, for a coreflood experiment conducted in naturally heterogeneous sandstone outcrop cores called Antolini Sandstone. We have shown that simulations using coarser physical and numerical grids and effective properties, calculated from the fine-grid characterization data, give results comparable to those of fine mesh simulations. In this example, we have shown both the huge potential for savings in computer time when a coarser grid can be used and the very large errors that can be made if the averaging scheme is applied indiscriminately.; Finally, we have examined in detail the importance of fingering during field applications of carbon dioxide flooding. In this study, the effects of gravity, physical dispersion, capillary pressure, phase behavior, and heterogeneity are also combined and simulated for carbon dioxide flooding on a field scale using stochastic permeability fields that span the full range of reservoir heterogeneity (V{dollar}sb{lcub}rm DP{rcub}{dollar} varying from 0.095 to 0.80) and dimensionless correlation in the bulk flow direction ({dollar}lambdasb{lcub}rm Dx{rcub}{dollar} varying from 0.025 to 3.0). Through this systematic study, we have shown that contrary to most previous findings and common beliefs, viscous fingering is not the dominant flow pattern for field-scale carbon dioxide flooding under secondary conditions even without using WAG, which would tend to reduce fingers even further. (Abstract shortened by UMI.)... |
