Simulation of WAG and Gas Injection with Potential Sweep Improvement by Application of Foam

fax 01-972-952-9435. AbstractIn principle, miscible gas injection can displace nearly all of the oil from the portions of a reservoir swept by gas. However, reservoir heterogeneity, low gas density and high gas mobility reduce sweep efficiency and drastically decrease recovery. Fortunately, the use of foam can reduce gas mobility and the effect of heterogeneity and therefore increase sweep efficiency. Based on the work of Shan and Rossen, an optimal design strategy is proposed here for surfactant-alternating-gas (SAG) foam processes and tested in a homogeneous reservoir and in two layered rectangular reservoirs. Fractional-flow theory, Stone's and Jenkins' model for gravity segregation, and an idealized model developed by Shan and Rossen are used design the process and interpret the results.In the homogeneous reservoir, fixed-pressure SAG processes result in better sweep efficiency than fixed-rate SAG processes. A one-cycle SAG process, where one big slug of surfactant is followed by one big slug of gas, gives better sweep than multiple-cycle SAG processes; sweep efficiency is higher and the process proceeds faster. Adsorption can be dealt with by adjusting the size of the first surfactant slug. The optimal SAG process gives much better sweep efficiency than a WAG process without foam. Simulations reveal factors not accounted for in Shan and Rossen's idealized model for a single-cycle SAG process; nonetheless the model accurately predicts the rate of advance of the foam front at the top of the reservoir and qualitatively predicts the shape of the foam front. A ten-cycle SAG process performs somewhat better than predicted by Stone's model for continuous injection.In heterogeneous reservoirs, a maximum-fixed-pressure SAG process also gives best results. The surfactant slug sized for a homogeneous reservoir is much too large in this case, however, and much of the surfactant wasted. Again, a onecycle process SAG performs better than multiple, smaller cycles, but sweep is not as good as in a homogeneous reservoir. A high-permeability layer at the top of the reservoir leads to better sweep efficiency at gas breakthrough than when the reservoir has a low-permeability layer at the top. A lowpermeability layer at the top allows early gas breakthrough once gas in high-permeability layers below it reaches the edge of the surfactant bank above. Compared to WAG processes, the optimal SAG process shows much better sweep efficiency.

show abstract

“…27,[38][39][40] The reservoir model consists of 200x1x40 grid blocks. 27,[38][39][40] The reservoir model consists of 200x1x40 grid blocks.…”

Section: Reservoir Modelmentioning

confidence: 99%

Success of SAG Foam Processes in Heterogeneous Reservoirs

Rossen

Renkema

2007

All Days

View full text Add to dashboard Cite

show abstract

“…Though derived to represent the coalescence regime, this model also represents the low-quality regime reasonably well, as shown below. [26][27][28][29][30][31][32][33][34][35] It includes an option for population-balance simulation, but is most widely applied using a local-steady-state option that is similar in several respects to the model of Rossen et al 18 It includes in addition algorithms for the effects of surfactant concentration and oil saturation on foam not currently included in the UT model, that are not examined in this paper. Some results presented here use the same algorithms with the chemical-flood simulator UTCHEM, 24,25 which has previously been modified to accommodate a flowing gas phase.…”

Section: Simulation Of the Two Foam-flow Regimesmentioning

confidence: 99%

Simulating Foam Processes at High and Low Foam Qualities

et al. 2000

View full text Add to dashboard Cite

fax 01-972-952-9435. AbstractFoams used for gas or acid diversion exhibit two flow regimes, depending on foam quality. Two foam simulators, one the most widely used commercial foam simulator and the other developed at our university, fit steady-state foam behavior in both regimes. A simple procedure is described for fitting simulator parameters to a set of steady-state coreflood data and examples are shown. Fitting model parameters to a single coreflood data can err by fitting this datum to the wrong flow regime.Shear-thinning reported in the "low-quality regime" can increase foam injectivity in radial flow. Foams in lowquality regime fit the same correlation for overcoming gravity override previously derived for foam in the highquality regime. The flow regime does greatly affect the effect of capillary crossflow on foam diversion between layers differing in permeability, however.Capillary crossflow harms diversion between adjacent layers, as found earlier, but the magnitude of the effect is much less for foam in the low-quality regime, and no single correlation matches all the results. Capillary crossflow can actually increase (by a small amount) diversion from adjacent layers differing somewhat in permeability to distant layers with muchdifferent permeability.

show abstract

“…Castanier and Hanssen (1995) noticed the sliding of lamellae (bubble train flow) over liquid films covering the grains in silicon micromodels. Using atomic force microscopy between the liquid/solid interface, a thick layer of surfactant was found.…”

Section: Recent Work On the Mobility And Propagation Of Foammentioning

confidence: 99%

“…One approach is purely empirical and relies on calibrating a flexible simulator by selected laboratory data, which are measured on conventional core-sample scale and believed to represent those foam effects which are most critical. The empirical foam model implemented in STARSTM, a pseudo-compositional reservoir simulator is probably the best example of this approach (Coombe, 1992;Surguchev, 1995). It has the advantage of relative simplicity, but depends heavily on obtaining "correct" values of the lab parameters and on the assumption that these parameters can be measured in a manner that is not too strongly dependent on the mode and conditions of the experiment.…”

Section: Introductionmentioning

confidence: 99%