Theoretical and experimental study of foam for enhanced oil recovery and acid diversion
This dissertation comprises three studies of foam in porous media. The first study is a theoretical model for apparent viscosity of foam in porous media. This represents the first dynamic model for the movement of bubbles through constricted tubes in 2D, accounting for the drag on lamellae (soap films) along pore walls and the capillary forces that govern bubble shape in constricted tubes. At low velocities behavior fits earlier quasi-static and approximate models. At higher velocities, behavior is complex; for instance, pressure gradient can decrease with increasing bubble velocity. This work could provide a component to a fully mechanistic foam simulator. The second study is an experimental investigation of SAG foam processes for enhanced oil recovery, where gas and surfactant solution are injected in alternating slugs. Experimental fractional-flow curves are presented for two surfactant formulations in Berea sandstone, with no oil present. Results are then scaled-up using fractional-flow theory to a hypothetical 1D field-scale application. In one case the data suggest an abrupt jump from a strong-foam fractional-flow curve to a weaker-foam curve, as has been reported elsewhere. In both cases the data suggest successful mobility control on the field scale. The third study is an experimental investigation of post-foam liquid injection, which is the key to the success of foam-acid diversion for matrix-acid well stimulation. Results indicate that with high foam quality pressure gradient can be higher during liquid injection than during foam injection; this may require reconsideration of the optimal foam quality for foam-acid diversion. It appears that the water relative-permeability function obtained during foam injection also applies to liquid injection after foam, until trapped gas begins to dissolve into injected liquid. The extent of gas trapping and liquid mobility after foam varied with foam quality here, which means earlier models that exclude this effect may need to be revised. In both experimental studies, pressure gradient was monitored along the core and liquid saturation was determined by weighing the core continuously during the experiment. Strengths and shortcomings of this technique for determining water saturation are discussed, along with suggestions for improving the technique.