High-temperature surfactant foams are simulated by modifying gas-phase mobility in a conventional thermal simulator. Both surfactant-alternating-gas (SAG) and gaslliquid-coinjection processes are modeled. Foam generation by leave-behind and snap-off as well as foam coalescence and trapping mechanisms are incorporated in the model by an equation for the number density of foam bubbles; gas-phase relative permeability and apparent viscosity are modified according to the bubble density. Pressure and saturation data of laboratory corefloods are successfully history matched with simulation results. Field-scale sensitivity studies of the steam-foamdrive process demonstrate how the coalescence rate affects the extent of steam diversion. IntroductionGases (such as steam, CO 2 , and nitrogen) are injected into oil reservoirs as drive fluids in some EOR processes. Early gas breakthrough can occur at producing wells owing to override and channeling, resulting in low oil-recovery efficiency. Injecting surfactant to create foam can reduce gas mobility and improve volumetric sweep efficiency in oil reservoirs. Foams used in both mature and infant steamdrives have resulted in incremental oil production in California heavy-oil reservoirs. 1-4The behavior of foam in porous media is complex, and the mechanisms governing its flow are not yet fully understood. Laboratory and theoretical studies have investigated foam generation,5-7 bubble coalescence,8 and the effect of oil on foam stability. 9, 10 A few investigators have begun to develop a comprehensive model of foam flow in porous media, but only few experiments have been modeled. 6, II, 12This paper offers a mathematical model that includes the principal mechanisms that govern foam displacement in porous media. The effect of foam on gas-phase relative permeability and apparent viscosity is included in the model. Both static or continuousgas foams and "strong" or discontinuous-gas foams are modeled. 5 ,6 In the first case, static foam lamellae block pore throats for gas flow, decreasing the gas-phase relative permeability. In the second case, foam bubbles are displaced through the pore network and the flow behavior is controlled by the rheology and the generation, trapping, and coalescence of the flowing foam bubbles. As proposed by Falls et al. 6 and Patzek, II an equation is incorporated to calculate the flowing-foam-bubble density, which, in turn, dictates how the flowing-foam mobility is modified.Chaser SD1OOO™, a surfactant developed for steam-foam applications, was the sole chemical used in this study. Model parameters are obtained from corefloods and by history matching nitrogen foam floods in Berea sandstone cores. Two laboratory corefloods are compared to simulations with the chosen parameters. Field-scale simulations of the steam-foam-drive process are then presented for a range of bubble coalescence rates.

A percolation model of foam mobilization in porous media is developed. This model indicates that there is a minimum pressure gradient or, equivalently, a minimum gas velocity required to initiate mobilization of foam. As a result, for most foam enhanced oil recovery processes, where the surface tension is not low, deep foam penetration depends on propagation of foam formed at a high pressure gradient near the well. Low surface tension makes mobilization of CO, foams feasible, however, at pressure gradients found throughout much of the formation in a typical field application. The theory further predicts, and data confirm, that the minimum velocity for foam mobilization during steady flow of liquid and gas decreases as injected liquid volume fraction increases. The theory suggests a better strategy for foam generation: alternate injection of small slugs of liquid and gas.

No abstract

No abstract