Estimation of Parameters for the Simulation of Foam Flow through Porous Media. Part 1: The Dry-Out Effect

Abstract

We develop a method of estimating foam simulation parameters from laboratory experiments. Gas and surfactant solution are coinjected into sand packs at different experimental conditions in two systems. At steady state, saturation of the aqueous phase is shown to be relatively constant over a wide range of foam qualities. In order to obtain an accurate model fit at the transition foam quality, the difference between the foam model parameter fmdry and the transition water saturation S w t is identified and a method to precisely calculate S w t is developed. By superimposing contour plots of the transition foam quality and the foam apparent viscosity, one can estimate the reference mobility reduction factor (fmmob) and the critical water saturation (fmdry) using the STARS foam model. The parameter epdry, which regulates the abruptness of the foam dry-out effect, can be estimated by a transient (continuous gas injection) foam experiment.

Effect of Permeability on Foam-model Parameters and the Limiting Capillary Pressure

“…These parameters are presented in Table 4. In addition, using the data reported by Ma et al (2013), who studied the effect of surfactant concentration on foam in sandpacks, we find a trend of increasing P c * with increasing surfactant concentration ( Table 5). The smaller values of the P c * values in Table 5 compared to those reported in Figure 7 are due to very high permeability of the rock used by Ma et al experiments.…”

“…We focus our study on one steady-state implicit-texture foam model, that in STARS (Computer Modeling Group, 2012;Farajzadeh et al, 2012;Ma et al, 2013Ma et al, , 2014. In that model gas mobility is scaled by a function, FM, when foam is present:…”

“…We have modified the procedure explained by Ma et al (2013Ma et al ( , 2014 by adding shear-thinning function because in their approach smaller values can be assigned to epdry, although as shown later choosing very large values of epdry (as described by Cheng et al, 2000;Boeije and Rossen, 2013), does not alter the general conclusions of this paper. The procedure used here, described in Appendix B, seeks a least-squares fit to all the app f  data subject to weighting and constraints supplied by the user.…”

“…PB models can be restructured to assume local equilibrium (LE) between the dynamics of lamella creation and destruction; all current IT models assume LE. These models contain adjustable parameters that are estimated by fitting the model to experiments (Cheng et al, 2000;Boeije and Rossen, 2013;Rossen and Boeije, 2013;Ma et al, 2013Ma et al, , 2014Lotfollahi et al, 2015). Several different foam models can fit the steady-state mobility of strong foam in a given porous medium to a satisfactory degree (Lotfollahi et al, 2015).…”

Understanding aqueous foam with novel CO2-soluble surfactants for controlling CO2 vertical sweep in sandstone reservoirs

“…Here, we chose 100 for fmmob, which is less than these employed by earlier researchers, such as 1000 (Rossen et al 2006), 3000 (Rossen and Shen 2007;Jamshidnezhad et al 2008a), or 5000 (Cheng et al 2000;Rossen and Renkema 2007;Kloet et al 2009), because we believe too strong foam used previously masked some details in the foam process, which is crucial for the novel foam. Meanwhile, it is also much less than the result of coreflood matching (Ma et al 2013) because the foam in 3D is much weaker than in 1D, which has been demonstrated by Li et al (2006). In simulations, we made fmsurf equal to the injected surfactant concentration (3.34 9 10 -5 , molar fraction) (Hanssen et al 1994;Rossen and Renkema 2007).…”