(191a) Multi-Scale Experiments and Simulations for Foam Mobility Control

Authors: 
Hirasaki, G. J., Rice University

Mobility control can be an important factor for success in enhanced oil recovery (EOR) by chemical flooding, miscible or immiscible gas injection, and steam flooding.  Mobility control by foam is an alternative for polymer in chemical flooding and can be achieved for gas or steam injection with addition of surfactant.

            The challenge facing foam mobility control technology is its predictability.  Foam is a complex fluid consisting of gas dispersed in liquid with a mobility that depends on a number of variables.  Here we describe the multi-scale experiments and simulations used to improve our ability to predict the performance of foam mobility control for EOR.

            Micromodels (with moderate viscosity oil in oil-wet medium describing fine and coarse porous media in parallel with a fracture) demonstrate that the lack of mobility control with waterflooding, gas injection or WAG injection can be resolved with a properly formulated foam flood.

            One-dimensional (1-D) sand packs or core floods are generally used to characterize foam.  Homogeneous 1-D systems have a minimum pressure gradient for foam generation.  A 1-D sand pack with heterogeneities in series demonstrate that foam can be generated by capillary snap-off, independent of flow rate if the permeability contrast is greater than a factor of 4.  Two-dimensional (2-D) sand packs and parallel plate models demonstrate the effects of heterogeneity in parallel and of gravity.  Visual observation show cross-flow from high to low permeability behind the foam from and cross-flow from low to high permeability ahead of the foam front.  A three-dimensional (3-D) sand pack demonstrate that foam parameters obtained from 1-D experiments may underestimate the foam strength in 3-D predictions.

            Numerical simulation is used to interpret laboratory experiments and scale-up to field applications.  The foam parameters for numerical reservoir simulators are estimated from appropriate laboratory experiments. The surfactant can be injected in the CO2 slug for CO2 soluble surfactants.  The dependence of foam propagation on the partition coefficient for the surfactant between the CO2 and aqueous phases is demonstrated.