Abstract
This paper discusses the viscous fingering that occurs when water or a surfactant solution displaces oil in a porous medium. Such floods were visualized in an porous medium. Such floods were visualized in an oil-wet porous medium composed of fused plastic particles. The flow structure changed significantly within the range of capillary numbers between 10 -4 and 10 -3 . The addition of surfactant resulted in narrower fingers, which developed in a more dispersive fashion.
Introduction
In describing fluid/fluid displacements in porous media, a useful dimensionless quantity is the capillary number,
(1)
which corresponds to the ratio of viscous forces to capillary forces. Here, v is the specific fluid discharge or Darcy velocity, it is viscosity, and o is interfacial tension (IFT). It has been shown that the recovery of oil from an underground reservoir increases significantly if the capillary number can be increased beyond the range of 1 × 10 -4 to 2 × 10 -3 during water flooding (see Larson et al. 1 ). To this end, surfactants are used extensively in tertiary oil recovery operations with the objective of reducing IFT and consequently mobilizing the oil ganglia which otherwise would remain trapped.
This paper is concerned with the viscous fingering that occurs when water displaces oil in a porous medium, and we present a brief consideration on the effects that surfactants have on fingering. Noting that Peters and Flock have visualized fingering within the range of capillary numbers between 1.6 × 10 -6 and 7.2 × 10 -4, we present here visualizations at capillary numbers of 7.7 × 10 5 and 1.0 × 10 -3. Both our visualizations and the experiments of Peters and Flock involve large viscosity ratios so that only the viscosity of the more viscous fluid is considered when determining the capillary number. In particular, it is observed that as the capillary number increases, ganglia or blobs of displacing fluid are created at the displacement front in correspondence with the capillary numbers at which trapped ganglia are mobilized. This creation of ganglia at capillary numbers above 10 -3 was noted briefly in a previous paper 3 in which heptane displacing glycerine previous paper 3 in which heptane displacing glycerine was described.
A secondary objective of this work was to test the Chuoke et al. theory for predicting the wavelength of fingers, wavelength being the peak-to-peak distance between adjacent well-developed fingers.
Experimental Procedure
The apparatus for these studies was described in Ref. 3. Basically, it consists of a slab of consolidated plastic particles 1.34 × 0.79 × 0.0 1 8 ft [0.44 × 0.26 × 0.006 m] with particles 1.34 × 0.79 × 0.0 1 8 ft [0.44 × 0.26 × 0.006 m] with a porosity of 0.43 and a permeability of 7, 100 darcies. This high permeability, when compared with that of reservoir rocks, should not be important for this study since capillary numbers and residual saturations are independent of pore size. Water (viscosity 1 cp [1 mPa s]) was used to displace paraffin oil (viscosity 68 cp 168 mPa s] at 77F [25C]). The water was dyed with methylene blue (which acts as a mild surfactant). Without the dye, the oil/water IFT was 42 dyne/cm [42 mN/m]. The addition of dye lowered this value to 36 dyne/cm [36 mN/m] for the concentration of dye used. For the surfactant flood, a 1 % sodium alkyl aryl sulfonate solution was used, giving a surfactant-solution/paraffin-oil IFT of 3.0 dyne/cm [3.0 mN/m].
Water Displacing Oil
To compare our experiments with previous investigations of fingering, the displacement of paraffin oil by water at an average specific fluid discharge of 1.34 × 10–4 ft/sec [4.1 × 10 -5 m/s], corresponding to a capillary number of 7.7 × 10 -5, was studied (Fig. 1). Chuoke et al .4 and later Peters and Flock 2 have presented a formula for predicting the wavelength of presented a formula for predicting the wavelength of finger, lambda m :
(2)
where k is permeability, C is a dimensionless parameter which Peters and Flock call the wettability number and suggest would have the value 25 for an oil-wet porous medium, and mu o and mu ware viscosities of the displaced oil and displacing water, respectively. It was observed that the plastic particles of the porous medium considered here were oil wet because of adsorption of oil.
SPEJ
P. 325
Influence of Wetting on Viscous Fingering Via 2D Lattice Boltzmann Simulations
