Simulation of Foam Transport in Porous Media

Abstract

SPE Members Abstract Foam is an excellent fluid for achieving mobility control of gas in porous media. Practical application of foams for EOR processes, however requires a predictive model of foam displacement. Further, quantitative information on foam-flow behavior at reservoir flow rates and pressures is required as input to any field-scale modeling. An experimental and mechanistic-modeling study is reported for the transient flow of foam through 1.3 um (1.3 D) Boise sandstone at backpressures in excess of 5 MPa (700 psi) over a quality range from 0.80 to 0.99. Total superficial velocities range from as little as 0.42 to 2.20 in/day (1.4 ft/day to 7 ft/day). Sequential pressure taps and gamma-ray densitometry measure flow resistance and in-situ liquid saturations, respectively. We garner experimental pressure and saturation profiles in both the transient and steady states. Adoption of a mean-size foam-bubble conservation equation along with the traditional reservoir simulation equations allows mechanistic foam simulation. Since foam mobility depends heavily upon its texture, the bubble population balance is both useful and necessary as the role of foam texture must be incorporated into any model which seeks accurate prediction of flow properties. Our model employs capillary-pressure-dependent kinetic expressions for lamellae generation and coalescence and also a term for trapping of lamellae. Additionally, the effects of surfactant chemical transport are included. We find quantitative agreement between experimented and theoretical saturation and pressure profiles in both (lie transient and steady states. Introduction Foam is useful for controlling mobility of gases in porous media. Foam is relatively cost effective because it is mainly gas with stabilization of the gas/liquid interface provided by a relatively low concentration of surfactant (of order 1 wt%) within the aqueous phase. Since the gaseous portion of foam is dispersed, gas-phase flow mobility is greatly reduced and hence gravity override and viscous fingering, through high-permeability streaks may be reduced. However, practical implementation of foams for mobility control in enhanced oil recovery (EOR) processes has been hindered because a general understanding and a predictive model of town flow does not currently exist. Most previous studies were Eddisionian and focused upon the steady state. Although transient flow (i.e., displacement) is die most relevant to EOR, a reliable experimented data set that includes transient pressure and in-situ saturation profiles (along the length of a core) does not exist for foam flow. The most notable attempts at modeling foam flow have focused either on predicting transient flow or on predicting steady state results, but not both. P. 309^