High-pressure core flood experiments using gas condensate fluids in long sandstone cores have been conducted. Steady-state relative permeability points were measured over a wide range of condensate-to-gas ratios (CGR; volume of condensate per unit volume of gas, both at test pressure and temperature), and the velocity and interfacial tension (IFT) were varied between tests to observe the effect on relative permeability. The experimental procedures ensured that the fluid distribution in the cores was representative of gas condensate reservoirs. Hysteresis between drainage and imbibition during the steady-state measurements also was investigated, as was the repeatability of the data.A relative permeability rate effect for both gas and condensate phases was observed, with the relative permeability of both phases increasing with an increase in flow rate. The relative-permeabilityrate effect was still evident as the IFT increased by an order of magnitude. The influence of end effects was shown to be negligible under the IFT conditions used in the tests, with the Reynolds number indicating that flow was well within the so-called laminar regime under all test conditions. The observed rate effect was contrary to that of conventional non-Darcy flow, where the effective permeability should decrease with increasing flow rate. A generalized correlation between relative permeability, velocity, and IFT has been proposed.The results highlight the need for appropriate experimental methods and relative permeability relations where the distribution of the phases are representative of those in gas condensate reservoirs.

Summary The use of WAG (water-alternating-gas) injection can potentially lead to improved oil recovery from the fields. However, there is still an incomplete understanding of the pore-scale physics of the WAG processes and how these lead to improved oil recovery. Simulating the three-phase flow for prediction of the WAG performance in oil reservoirs is an extremely complex process. The existing three-phase relative permeabilities used in simulation are very approximate and do not properly account for the effects of fluid interfacial tension and rock wettability. Network model simulators are being developed to enable the prediction of three-phase relative permeability under different wettability conditions. However, such simulators need to be verified against experimental observations. In this paper, we present experimental results and discussion of a series of capillary-dominated WAG tests carried out in glass micromodels with wettability conditions ranging from water-wet to mixed-wet and oil-wet. Pore level fluid distribution and flow mechanisms were studied, and fluid saturation, at different stages of the experiments, were measured. The results showed that, under any of the wettability conditions, oil recovery by alternating injection of WAG was higher than water or gas injection alone. WAG recovery was observed to be higher for the oil-wet model than that in the mixed-wet one, which in turn was higher than that in the water-wet micromodel. Given enough time and more cycles of WAG injection, the recovery of the mixed-wet model seems to catch up with that of the oil-wet model. Introduction WAG injection is being increasingly applied as an improved oil recovery method, particularly in reservoirs that have been waterflooded. Christensen et al.1 reported a review of some 60 field applications of WAG. Several field trials have been reported as being successful (for instance, in Kuparuk,2 Snorre,3 and Gulfaks fields4). Both immiscible4–6 and miscible gases7 have been used. A large number of coreflood experiments8–12 and analytical and numerical simulations11,13 have been carried out. A study in 1993 demonstrated that the WAG process could be used for improving the hydrocarbon recovery in gas/condensate reservoirs.14 Most of the research work conducted so far has been on either coreflooding8–10 or numerical simulation,11,12 sometimes alongside field trials. The relationship between the injection gas/water ratio (GWR) and oil recovery has been empirically investigated using core displacement experiments, often at low pressure and generally with water-wet cores.8,10 Extensive research is in progress to develop network model simulators that can predict three-phase flow in porous media with immiscible [high interfacial tension (IFT)] and near-miscible (low IFT) fluids and rocks of different wettability. These simulators need to be verified against the experimental observations. This has been carried out to some extent for water-wet systems and using core observations. In the current project, we carry out experiments with micromodels that can be used to obtain an in-depth understanding of the physical processes involved, and to use such information in development and verification of three-phase network model simulators. As far as we know, no micromodel visualization of the WAG injection process has been carried out to directly observe the physical processes taking place in the porous media, using live oil, live water in equilibrium with injection gas, and models with different wettability. Larsen et al.15 reported some results of their WAG micromodel studies. No detail of the experimental procedure and no images of fluid distributions, or recovery results from the micromodel tests, were presented. Etched-glass micromodels are useful for viewing pore level events because of their visual clarity. Micromodels were used as early as 1960 for fluid displacement studies.16 The ability to see the movement of fluid interfaces makes it possible to distinguish between a variety of mechanisms that may take place in a porous medium when more than one phase is present. Chatzis and Dullien17 presented an excellent example of the use of the micromodel to evaluate existing theories of two-phase flow in a simple geometry. Lenormand and Zarcone18 have taken advantage of the well-defined shapes of capillary tubes in their molded resin models to compare the results of calculations of two-phase flow in both drainage and imbibition processes with observation results. Micromodel observations have played a significant role in development of network models for application to multiphase flow. The procedure has been successfully applied to network modeling of two-phase flow in simple or idealized porous media using pore-scale physics identified in micromodel experiments (Lenormand et al.,19 Blunt and King,20 Blunt et al.,21 and Billiote et al.22). In recent years, several advancements in pore-scale modeling have been made. Oren, Bakke, and coworkers23–25 have developed network models based on the pore-space geometry of the rock of interest. The application of network modeling techniques to three-phase flow is considerably less developed than for two-phase flow, especially for oil-wet and mixed-wet porous media. This is because our understanding of the pore-scale physics of three-phase displacement is still incomplete. However, previous micromodel works (Oren and Pinczewski,26 Oren et al.27) suggested that it is possible to learn a great deal about physics of the three-phase displacement in order to take major steps toward developing realistic network models for three-phase flow. For water-wet media, the pore-scale mechanisms are rather well established. But the behavior of oil-wet and mixed-wet systems has a less firm experimental basis. Kovscek et al.28 have provided a pore-level scenario for wettability alteration that has been used to describe fluid configurations of two- and three-phase displacements.29–32 Van Dijke et al.33,34,35 have presented a network model simulator for modeling three-phase flow processes, in particular WAG injection cycles. Using the micromodel results that we presented in two previous papers,36,37 they showed a good agreement between simulation and experiment, in particular with respect to the displacement mechanisms during the WAG cycles. Piri and Blunt38 presented a network model of three-phase flow to capture relative permeabilities, saturation paths, and capillary pressure for media of arbitrary wettability. In this paper, we present a description of the fluid distribution and the pore-scale physics and mechanisms for flow under the WAG processes at different wettability conditions.

As gas flooding becomes a more viable means of enhanced oil recovery, it is important to identify and understand the pore-scale flow mechanisms, both for the development of improved gas flooding applications and for the predicting phase mobilisation under secondary and tertiary gas flooding. The purpose of this study was to visually investigate the pore-level mechanisms of oil recovery by near-miscible secondary and tertiary gas floods. High-pressure glass micromodels and model fluids representing a near-miscible fluid system were used for the flow experiments. A new pore-scale recovery mechanism was identified which significantly contributed to oil recovery through enhanced flow and cross-flow between the bypassed pores and the injected gas. This mechanism is strongly related to a very low gas/oil interfacial tension (IFT), perfect wetting conditions and simultaneous flow of gas and oil in the same pore, all of which occur as the gas/oil critical point is approached. The results of this study helps us to better understand the pore-scale mechanisms of oil recovery in very low-IFT (near-miscible) systems. In particular we show that in near-miscible gas floods, behind the main gas front, the recovery of the oil continues by cross-flow from the bypassed pores into the main flow stream and as a result almost all of the oil, which has been contacted by the gas, could be recovered. Our observations in high-pressure micromodel experiments have demonstrated that this mechanism can only occur in near-miscible processes (as opposed to immiscible and completely miscible processes), which makes oil displacement by near-miscible gas floods a very effective process.

TX 75083-3836, U.S.A., fax 01-972-952-9435. AbstractIn reservoirs that have been waterflooded or gas injected, it is still possible to recover a significant amount of the remaining oil by water-alternating-gas (WAG) injection. WAG injection has been successfully implemented in some waterflooded reservoirs. However, the physical processes underlying the complex three-phase flow in WAG have not been well understood.A series of WAG experiments has been conducted, using high-pressure glass micromodels, with high quality images of the oil recovery processes being video recorded. The experiments were performed using water-wet, oil-wet and mixed-wet micromodels. The authors presented the results of the experiments of water-wet models in SPE ATC&E 2000 (SPE 63000) 1 . This paper presents experimental results of oilwet and mixed-wet models and demonstrates the difference in flow mechanisms of the WAG process under different wettability conditions.

Copynght 1995, Stecting Commttee of the European IOR-Symposiuin. 1hie paper was preented at the Sth. European IOR-Sympoaium in Vienne AustriL May 15-17. 1995 Thie paper was selected for prsentaton by the Steenng Commlttee. foliowing revew ol inforrnallon contalned in an abstract submitted by the author(8). The paper. es preaented has not bean reviewed by the Stearing Committee,