|dc.description.abstract||Foam has a wide range of applications in the petroleum industry. Historically, foam has been used in acid diversion in well-stimulation, air/gas drilling, and liquid removal in gas wells [1-3]. Furthermore, foam is an interesting EOR (Enhanced Oil Recovery) method as it significantly reduces the gas mobility in porous media due to gas entrapment and increases the gas apparent viscosity. Foam is complex in terms of understanding its generation, destruction, stability and dynamics in porous media. Thus, visualizing the foam behavior in glass micromodels enables us to achieve better understanding both at classic laboratory conditions of pressure and temperature and at reservoir conditions of the micromodel experiments when done at HT/HP, which is possible thanks to the advancements in microfluidic technology and the availability of high-resolutions cameras.
In this work, foam generation, propagation, and destruction mechanisms were visualized and analyzed in 2D glass micromodels at ambient conditions. Two micromodel types were used in this study where the first has a simple pore geometry (MT1) and the second has a more realistic pore-pattern resembling a Clashach sandstone (MT8). The latter s pore network was designed using Total s in-house SISMAGE software and using multipoint statistics (MPS) to analyze a manually modified 2D image taken from a micro-CT scanner (a training image of size 800px by 800px; thanks to SISMAGE the initial training image was converted into a final image of size 11,000px by 2,000px). The used technique conserved the pore level statistics of the training image. The size of this final pore pattern image was therefore 4.5cm by 1cm which was eventually etched on a glass micromodel chip (MT8).
In addition, we verified foam generation by co-injecting gas (nitrogen) and synthetic seawater brine containing 0.5% foaming solution and without pre-generating the foam ahead of the micromodel entry (a common experimental procedure in the literature but which does not allow to actually observe foam creation in the real porous medium). The injection flowrate was 2l/min to approach reservoir flowrates, i.e., approximately, 1 ft./day. These rates are much slower than what was used in the literature. The first experiments were mainly conducted on the water-wet micromodels. However, wettability has been reported to affect the creation of foam bubbles, mainly by the snap-off mechanism in the case of oil-wet rocks. Thus, the wettability of a sandstone-like micromodel was altered to make it hydrophobic and results showed a higher apparent viscosity compared with the hydrophilic micromodel. Furthermore, the bubble count of MT8 water-wet micromodel was analysed at different foam qualities (gas fractions), i.e., at 90%, 50%, and 30% and for the MT8 oil-wet micromodel at 50%. We observed that the oil-wet MT8 micromodel chip had a lower bubble count despite resulting in a higher apparent viscosity compared with the water-wet micromodel. In addition, for oil-wet MT8, the bubble creation at the middle of the micromodel (2.5cm from co-injection point) did not increase compared with the bubble count at the entry of the micromodel (1cm from co-injection point), which was instead observed in the case of the water-wet MT8 micromodel.||en