The objective of this research is to widen the applicability of gas flooding to shallow oil reservoirs by reducing the pressure required for miscibility using gas enrichment and increasing sweep efficiency with foam. Task 1 examines the potential for improved oil recovery with enriched gases. Subtask 1.1 examines the effect of dispersion processes on oil recovery and the extent of enrichment needed in the presence of dispersion. Subtask 1.2 develops a fast, efficient method to predict the extent of enrichment needed for crude oils at a given pressure. Task 2 develops improved foam processes to increase sweep efficiency in gas flooding. Subtask 2.1 comprises mechanistic experimental studies of foams with N{sup 2} gas. Subtask 2.2 conducts experiments with CO{sup 2} foam. Subtask 2.3 develops and applies a simulator for foam processes in field application. Regarding Task 1, several results related to subtask 1.1 are given. In this period, most of our research centered on how to estimate the dispersivity at the field scale. Simulation studies (Solano et al. 2001) show that oil recovery for enriched gas drives depends on the amount of dispersion in reservoir media. But the true value of dispersion, expressed as dispersivity, at the field scale, is unknown. This research investigates three types of dispersion in permeable media to obtain realistic estimates of dispersive mixing at the field scale. The dispersivity from single-well tracer tests (SWTT), also known as echo dispersivity, is the dispersivity that is unaffected by fluid flow direction. Layering in permeable media tends to increase the observed dispersivity in well-to-well tracer tests, also known as transmission dispersivity, but leaves the echo dispersivity unaffected. A collection of SWTT data is analyzed to estimate echo dispersivity at the SWTT scale. The estimated echo dispersivities closely match a published trend with length scale in dispersivities obtained from groundwater tracer tests. This unexpected result--it was thought that transmission dispersivity should be greater than echo dispersivity--is analyzed with numerical simulation. A third type of dispersive mixing is local dispersivity, or the mixing observed at a point as tracer flows past it. Numerical simulation results show that the local dispersivity is always less than the transmission dispersivity and greater than the echo dispersivity limits. It is closer to one limit or the other depending on the amount and type of heterogeneity, the autocorrelation structure of the medium's permeability, and the lateral (vertical) permeability. The agreement between the SWTT echo dispersivities and the field trend suggests that the field data are measuring local dispersivities. All dispersivities appear to grow with length. Regarding Task 2, two results are described: (1) An experimental study of N{sup 2} foam finds the two steady-state foam-flow regimes at elevated temperature and with acid, adding evidence that the two regimes occur widely, if not universally, in foam in porous media. (2) A simulation finds that the optimal injection strategy for overcoming gravity override in homogeneous reservoirs is injection of large alternating slugs of surfactant and gas at fixed, maximum attainable injection rates. A simple model for the process explains why the this strategy works so well. Before conducting simulations of SAG displacements, however, it is important to analyze the given foam model using fractional-flow theory. Fractional-flow theory predicts that some foam processes will give foam collapse immediately behind the gas front. In simulations, numerical dispersion leads to a false impression of good sweep efficiency. In this case simply grid refinement may not warn of the inaccuracy of the simulation.
