You must log in to edit PetroWiki. Help with editing

Content of PetroWiki is intended for personal use only and to supplement, not replace, engineering judgment. SPE disclaims any and all liability for your use of such content. More information

PEH:Immiscible Gas Injection in Oil Reservoirs

Jump to navigation Jump to search

Publication Information


Petroleum Engineering Handbook

Larry W. Lake, Editor-in-Chief

Volume V – Reservoir Engineering and Petrophysics

Edward D. Holstein, Editor

Chapter 12 – Immiscible Gas Injection in Oil Reservoirs

H.R. (Hal) Warner Jr., Warner Consulting Services and E.D. Holstein, Consultant

Pgs. 1103-1147

ISBN 978-1-55563-120-8
Get permission for reuse

This chapter concerns gas injection into oil reservoirs to increase oil recovery by immiscible displacement. The use of gas, either of a designed composition or at high-enough pressure, to result in the miscible displacement of oil is not discussed here; for a discussion of that topic, see the chapter on miscible flooding in this section of the Handbook. A variety of gases can and have been used for immiscible gas displacement, with lean hydrocarbon gas used for most applications to date. Historically, immiscible gas injection was first used for reservoir pressure maintenance. The first such projects were initiated in the 1930s and used lean hydrocarbon gas (e.g., Oklahoma City field and Cunningham pool in the United States[1] and Bahrain field in Bahrain[2][3]). Over the decades, a considerable number of immiscible gas injection projects have been undertaken, some with excellent results and others with poor performance. Reasons for this range of performance are discussed in this chapter. At the end of this chapter, a variety of case studies are presented that briefly describe several of the successful immiscible gas injection projects.

Gas injection projects are undertaken when and where there is a readily available supply of gas. This gas supply typically comes from produced solution gas or gas-cap gas, gas produced from a deeper gas-filled reservoir, or gas from a relatively close gas field. Such projects take a variety of forms, including the following:

  • Reinjection of produced gas into existing gas caps overlying producing oil columns.
  • Injection into oil reservoirs of separated produced gas for pressure maintenance, for gas storage, or as required by government regulations.
  • Gas injection to prevent migration of oil into a gas cap because of a natural waterdrive, downdip water injection, or both.
  • Gas injection to increase recoveries from reservoirs containing volatile, high-shrinkage oils and into gas-cap reservoirs containing retrograde gas condensate.
  • Gas injection into very undersaturated oil reservoirs for the purpose of swelling the oil and hence increasing oil recovery.

The primary physical mechanisms that occur as a result of gas injection are (1) partial or complete maintenance of reservoir pressure, (2) displacement of oil by gas both horizontally and vertically, (3) vaporization of the liquid hydrocarbon components from the oil column and possibly from the gas cap if retrograde condensation has occurred or if the original gas cap contains a relict oil saturation, and (4) swelling of the oil if the oil at original reservoir conditions was very undersaturated with gas. Gas injection is particularly effective in high-relief reservoirs where the process is called "gravity drainage" because the vertical/gravity aspects increase the efficiency of the process and enhance recovery of updip oil residing above the uppermost oil-zone perforations.

The decision to apply immiscible gas injection is based on a combination of technical and economic factors. Deferral of gas sales is a significant economic deterrent for many potential gas injection projects if an outlet for immediate gas sales is available. Nevertheless, a variety of opportunities still exist. First are those reservoirs with characteristics and conditions particularly conducive to gas/oil gravity drainage and where attendant high oil recoveries are possible. Second are those reservoirs where decreased depletion time resulting from lower reservoir oil viscosity and gas saturation in the vicinity of producing wells is more attractive economically than alternative recovery methods that have higher ultimate recovery potential but at higher costs. And third are reservoirs where recovery considerations are augmented by gas storage considerations and hence gas sales may be delayed for several years.

Nonhydrocarbon gases such as CO2 and nitrogen can and have been used.[4] In general, calculation techniques developed for hydrocarbon-gas injection and displacement can be used for the design and application of nonhydrocarbon, immiscible gas projects. Valuing the use of such gases must include any additional costs related to these gases, such as corrosion control, separating the nonhydrocarbon components to meet gas marketing specifications, and using the produced gas as fuel in field operations.

The sections in this chapter are presented in the following order:

  • Microscopic and Macroscopic Displacement Efficiency of Immiscible Gas Displacement
  • Gas/Oil Compositional Effects During Immiscible Gas Displacement
  • Reservoir Geology Considerations Regarding Immiscible Gas Displacement
  • General Immiscible Gas/Oil Displacement Techniques
  • Vertical or "Gravity Drainage" Gas Displacement
  • Calculation Methods for Immiscible Gas Displacement
  • Immiscible Gasflood Monitoring
  • Field Case Studies: Immiscible Gas Injection Examples
  • Miscellaneous Topics
  • Summary and Conclusions

The purposes of this chapter include listing the physical criteria that separate the successful gas injection operations from the unsuccessful ones, describing the reservoir and process variables that must be defined and quantified, and demonstrating some of the simple techniques available for predicting and evaluating field performance. Some of these calculations can be performed with spreadsheets or, more tediously, with hand-held calculators. Modern numerical reservoir simulators are commonly used to calculate the projected performance of applying immiscible gas injection to a particular reservoir. For reservoirs with several years of immiscible gas injection, these same simulators can be used to history match past performance and to project future performance under various scenarios (e.g., continuing current operations, evaluating various new producing wells options, or comparing surface facility operational alternatives). See the chapter on reservoir simulation in this section of the Handbook.

Specifically not included in this chapter is any discussion of the factors to consider in implementing a gas injection project, such as gas compression needs, gas distribution systems, wellbore configurations, and vessel selection and sizing for handling produced fluids. These subjects are covered in various chapters in the Production Operations Engineering and Facilities and Construction Engineering sections of the Handbook.

Microscopic and Macroscopic Displacement Efficiency of Immiscible Gas Displacement

The conceptual aspects of the displacement of oil by gas in reservoir rocks are discussed in this section. There are three aspects to this displacement: gas and oil viscosities, gas/oil capillary pressure (Pc) and relative permeability (kr) data, and the compositional interaction, or component mass transfer, between the oil and gas phases. The first two topics are discussed in this section; the third is discussed in the next section.

Gas/Oil Viscosity and Density Contrast

One must first understand the viscosity and density differences between gas and oil to appreciate why the gas/oil displacement process can be very inefficient. Gases at reservoir conditions have viscosities of ≈0.02 cp, whereas oil viscosities generally range from 0.5 cp to tens of centipoises. Gases at reservoir conditions have densities generally one-third or less than that of oil. Thus, gas is generally one to two orders of magnitude less viscous than the oil it is trying to displace. Regarding the fluid density difference, gas is always considerably "lighter" than the oil; hence, gas, when flowing, will segregate by gravity to the top of the reservoir or zone and oil will "sink" simultaneously to the bottom of the reservoir or zone.

Another gas/oil property that must be known for calculations at reservoir conditions is the interfacial tension (IFT) between the oil and gas fluid pair. This value is needed at reservoir conditions for the conversion of gas/oil capillary pressure data from surface to reservoir conditions. A number of technical papers discuss the calculation of IFT from compositional information about the oil and gas phases.[5][6][7][8][9] Table 12.1 from Firoozabadi et al.[8] shows several reservoir oil-gas fluid-pair IFT values (measured and calculated) as a function of temperature and pressure. As the pressure increases, the IFT values decrease, although not low enough for miscible displacement to occur. Although not illustrated in the table, it should be noted that the IFT between nitrogen and oil is higher than that between a lean natural gas and the same oil.

Gas/Oil Capillary Pressure and Relative Permeability

The gas/oil capillary pressure and relative permeability data are typically measured by commercial laboratories using routine special core analysis procedures. Gas-oil capillary pressure data can be measured with either porous-plate or centrifuge equipment. One approach for obtaining gas/oil relative permeability data is the viscous displacement method in which gas displaces oil. A second method is the centrifuge method, which is generally used to obtain capillary pressure and relative permeability information simultaneously.

In all cases, gas is the nonwetting phase in this displacement; hence, it will preferentially flow through the largest pores first. However, what is very important in the determination of the oil relative permeability is the distribution of the oil phase in the core sample because in real reservoirs connate water occupies the smallest pores. As shown by Hagoort,[10] initial water saturation has a significant effect on oil relative permeability during the gas/oil displacement (centrifuge experiments). The water phase will occupy a greater percentage of the smaller pore spaces as the connate water saturation increases. As a result, the pore structure appears more streamlined to the oil and gas phases. The oil relative permeability at higher connate water saturations is considerably higher (see Fig. 12.1 and the discussion of capillary pressure and relative permeability concepts in the chapter on those topics in the General Engineering Section of the Handbook).

The other key aspect of the oil relative permeability (kro) is the determination of its value as the oil saturation decreases. Because oil relative permeability becomes quite low but nonzero, the time to reach equilibrium in laboratory core plug measurements can be very long. Fig. 12.1 presents experimental results for cumulative oil recovery as a function of drainage time and shows that the oil continues to flow but more and more slowly (linearly as a function of the logarithm of tD); Hagoort[10] found similar behavior for the four different rock types he tested.

If the gas/oil relative permeability data were measured with the viscous displacement technique (the extended Welge technique as described by Johnston et al.),[11] extra care is needed in applying these data. First, the displacement of oil by gas is at an unfavorable mobility ratio (see discussion below) that makes the process unstable. Second, a displacement is adversely affected by capillary end effects that, for the gas/oil system, cannot be overcome by high gas throughput rates. At low oil saturations, the region of most interest, the capillary end effect is the greatest.[10]

Finally, one method developed to affect the gas/oil relative permeability and to reduce gas mobility is to inject water alternately with gas (WAG). This procedure was proposed by Caudle and Dyes.[12] Although the method was proposed for use in miscible gasfloods, the concept applies equally to immiscible gas displacements. This technique has been used in many west Texas CO2 miscible gas projects, in the Prudhoe Bay miscible flood,[13] and in the Kuparuk immiscible and miscible gas injection processes.[14][15] The three-phase gas, oil, and water relative permeabilities are calculated in numerical reservoir simulators with algorithms developed over the past several decades.[16]

Mobility Ratio

The mobility of a fluid (Eq. 12.1) is defined as its relative permeability divided by its viscosity. Mobility combines a rock property, permeability, with a fluid property, fluid viscosity. Gas-oil relative permeabilities are assumed to be dependent on the saturations of the two fluid phases and independent of fluid viscosity:


A fluid’s mobility relates to its flow resistance in a reservoir rock at a given saturation of that fluid. Because viscosity is in the denominator of this definition, gases, which are very-low-viscosity fluids, have very high mobility.

Mobility ratio is generally defined as the mobility of the displacing phase (in the gas/oil case, gas) divided by the mobility of the displaced phase, which is oil. Eq. 12.2 presents two forms of the mobility ratio equation:


Eq. 12.2 can also be written in more familiar engineering terms as the ratio of the two fluids’ relative permeability values multiplied by the ratio of the two fluids’ viscosities.


For simple calculations, the mobility ratio is calculated at the endpoint relative permeability values for the two phases. Hence, the equation that practical engineers use for the gas/oil mobility ratio is


All displacements of oil by gas are at "unfavorable" mobility ratios, with typical values of 10 to 100 or more.

Gas/Oil Linear Displacement Efficiency

The equations that characterize the mechanics of oil displacement by an immiscible fluid were developed by Buckley and Leverett[17] using relative permeability concepts and Darcy’s law describing steady-state fluid flow through porous media. The resulting fractional flow equation describes quantitatively the fraction of displacing fluid flowing in terms of the physical characteristics of a unit element of porous media. Assumptions inherent in their work are steady-state flow, constant pressure, no compositional effects, no production of fluids behind the gas front, no capillary effects, movement of advancing gas parallel to the bedding plane, immobile water saturation, and uniform cross-sectional flow (no gravity segregation of fluids within the element). Subsequent work by Welge[18] made solving the displacement equations easier.

The Welge equation for the fractional flow of gas at any gas saturation (Sg) is calculated as follows:



A = area of cross section normal to the bedding plane, ft2,
fg = fraction of flowing stream that is gas,
k = permeability, darcies,
kro = relative permeability to oil, fraction,
krg = relative permeability to gas, fraction,
M = mobility ratio, RTENOTITLE,
qT = total flow rate through area A , res ft3/D,
α = angle of dip, positive downdip, degrees,
Δρ = density difference, ρgρo, lbm/ft3,
μo = viscosity of oil, cp, and
μg = viscosity of gas, cp.

When gravity is negligible, this equation becomes the more familiar Buckley-Leverett equation:


Fig. 12.2 is a typical plot resulting from these calculations. The importance of the gravity term is indicated.

To relate the fraction of gas flowing to time, Buckley and Leverett developed the following material-balance equation:



L = length, ft,
Sg = gas saturation, fraction,
t = time, days, and
ϕ = porosity, fraction.

The value of the derivative dfg/dSg may be obtained for any value of gas saturation by determining slopes at various points on the fg vs. Sg curve. These slopes can be determined manually or, more precisely, using the method presented by Kern[19] for computer spreadsheets. Fig. 12.3 illustrates calculated gas-saturation distributions derived from the no-gravity and with-gravity fractional flow curves shown in Fig. 12.2. The area beneath each curve represents the gas-invaded zone. The saturation profile calculation results in lengths that first increase as saturation decreases and then decrease at lower saturations. While correct from a material-balance standpoint, it has been customary to square off the leading edge of the curve at the breakthrough saturation to account for capillary pressure that was neglected in the original derivation of the equation.

The gas/oil displacement efficiency, the percent of the oil volume that has been recovered, can be calculated for any period of gas injection by integrating the volume of the gas-invaded zone as a function of gas saturation (Sg). Hence, the fractional flow curves (Fig. 12.2) are used to generate saturation profiles (Fig. 12.3) that lead to values for the gas/oil displacement efficiency. In the next section, several of the factors affecting this efficiency are discussed.

Factors Affecting Gas/Oil Displacement Efficiency

The fractional-flow and material-balance equations discussed above are important for understanding the effects on the efficiency of the gas/oil displacement process of (1) initial saturation conditions, (2) fluid viscosity ratios, (3) relative permeability ratios, (4) formation dip, (5) capillary pressure, and (6) factors of permeability, density difference, rate of injection, and cross section open to flow.

Initial Saturation Conditions. If gas injection is initiated after reservoir pressure has declined below the bubblepoint, the gas saturation will decrease the amount of displaceable oil. If the free gas saturation exceeds the breakthrough saturation, no oil bank will be formed. Instead, oil production will be accompanied by immediate and continually increasing gas production. Laboratory investigations and mathematical analyses have demonstrated this influence of gas saturation on gas displacement performance.[20]

Swi has been shown to have no influence on displacement efficiency at gas breakthrough, but it directly affects the displaceable oil volume.[21] If Swi is mobile, the displacement equations are not directly applicable because they were developed for two-phase flow. Approximations of gas displacement performance can usually be made when three phases are mobile by treating the water and oil phases as a single liquid phase. Displacement calculations can then be made with krg and kro data determined from core samples containing an immobile water saturation. Oil recovery can be differentiated from total liquid recovery on the basis of material balance calculations incorporating an estimated minimum interstitial water saturation.

Fluid Viscosities. The effect of oil viscosity on fractional flow is illustrated in Fig. 12.4. In this plot, the Sg at breakthrough increases from 12 to 38% with a 10-fold decrease in oil viscosity.

Relative Permeability Ratios. The concepts of relative permeability can be applied equally well to complete or partial pressure-maintenance operations. Relative permeability, a characteristic of the reservoir rock, is a function of fluid saturation conditions. It is important that calculations be based on dependable data obtained by laboratory analyses at reservoir conditions using representative core samples. If possible, the laboratory-determined data should be supplemented by relative permeabilities calculated from field performance data.

Formation Dip. When formation dip aids gravity, as illustrated in Fig. 12.2, fractional flow behavior is significantly improved if permeability is high enough and withdrawal rates do not exceed gravity-stable conditions. Gravity drainage is discussed later in this chapter.

Capillary Pressure. Capillary forces are opposite gravity drainage forces and directionally decrease displacement efficiency. However, capillary forces can often be ignored as insignificant for projects with rates of displacement normally used. Only at extremely low rates of displacement, where viscous forces become negligible, is the saturation distribution controlled to a significant extent by the balance between capillary and gravitational forces. Another place where capillary forces are considered important is many of the large carbonate reservoirs of the Middle East where the matrix-blocks/fracture-system interaction can significantly affect overall reservoir performance.

Other Factors. Higher permeability, greater density difference between oil and gas, and a lower displacement rate all improve the displacement efficiency.

Unfavorable Mobility Ratio Causes Viscous Flow Instabilities

Displacements that take place at very unfavorable mobility ratios are unstable, and viscous fingering occurs. This is the situation for essentially all gas/oil displacements, especially if the displacement is occurring horizontally. The impact of such instabilities is illustrated in Figs. 12.5 and 12.6.[22][23] Both figures were drawn from technical literature concerning miscible displacement laboratory experiments using homogeneous sandpacks, but the observed effects would be the same for immiscible gas displacing oil at very unfavorable mobility ratios. Fig. 12.5 shows, in cross-sectional view, the nature of viscous fingering for two highly unfavorable mobility ratios (the two fluids have equal densities). The flood front in both cases is very unstable. Fig. 12.6 shows, in areal view, the effect of mobility ratio on the displacement process in a quarter of a five-spot pattern for mobility ratios from 0.151 to 71.5. For mobility ratios from 4.58 to 71.5 (cases D through F), the flood front is very unstable, and breakthrough occurs via narrow fingers of the injected fluid; these cases show how the process of gas displacing oil would occur.

In both of these illustrations, the cause of the viscous fingering was a slight perturbation in the flow field that grew into the viscous finger once the perturbation occurred. In real reservoir situations, there are two physical aspects that enhance the viscous-fingering phenomenon. First, real reservoirs are very heterogeneous, so a variety of styles of permeability heterogeneities can initiate viscous fingering. Second, in a cross-section immiscible gas/oil displacement process, gas is always less dense than oil. Hence, there is a gas/oil density difference, and the force of gravity causes the gas to override the oil and initiate a viscous finger of the high-mobility gas phase along the top of the reservoir interval.

If the gas/oil displacement is occurring vertically with gas generally displacing oil downward, gravity will work to stabilize the flood front between the gas and oil, although if the rate is too high, instabilities in the form of gas cones or tongues can occur.

Gas/Oil Compositional Effects During Immiscible Gas Displacement

This section briefly discusses the various mass transfer compositional aspects of the immiscible gas/oil displacement process. The implications of these compositional effects are very dependent on the oil composition, the composition of the injected gas, and the surface facilities and pipelines available in a particular field situation. The injected gas/oil composition interactions can be categorized as either swelling effects (gas dissolving into the oil phase) or stripping effects (various components from the oil transferring to the gas phase).

Swelling Compositional Effects

The most obvious compositional effect in the immiscible gas/oil displacement process is that, if the oil is not saturated with gas at the reservoir pressure or if the reservoir pressure is increased as a result of the gas injection, the volume of gas dissolved in the oil will increase until the oil is saturated at that pressure. At the same time and because of the increased volume of gas in solution in the oil, the oil formation volume factor (FVF) will increase. This phenomenon, commonly called swelling, can increase the efficiency of the gas/oil displacement process.

The significance of the swelling effect is dependent on the oil reservoir situation. For an oil reservoir in which there is a gas cap, the underlying oil column will already be fully or nearly saturated with gas at the reservoir pressure. Hence, there will be very little impact on the gas/oil displacement process as a result of the interaction between the reservoir oil and injected gas. However, for an oil reservoir in which there is no gas cap and where the oil bubblepoint pressure is very low compared with the original reservoir pressure, the swelling effect can be a very significant part of the gas/oil displacement process. The Swanson River field (Cook Inlet, Alaska) is an example of this latter situation (original reservoir pressure, 5,580 psi; oil bubblepoint pressure, 1,350 psi).[24] The change in oil FVF from a bubblepoint of 1,350 psi to being saturated at 5,580 psi is from 1.21 to ≈1.80 RB/STB. The application of immiscible gas injection to the Swanson River field is discussed in the Case Studies section of this chapter.

In some of the simple calculation techniques discussed below, the swelling effect is included. In a "black oil" type of numerical reservoir simulator, the swelling effect is taken into account because, although there are only two hydrocarbon components (gas and oil) for the two hydrocarbon phases, the swelling effect is incorporated by means of the entered table of oil pressure/volume/temperature (PVT) properties (e.g., FVF, gas in solution, oil viscosity) as a function of pressure. In other words, in this type of reservoir simulator, the gas hydrocarbon component exists either as a free-gas phase or as gas dissolved in the oil; the oil hydrocarbon component exists only as part of the liquid phase. A compositional numerical reservoir simulator will automatically take the swelling phenomenon into account in its equation-of-state phase-behavior calculations. See the chapter on numerical simulation in this section of the Handbook for detailed discussion of the various types of numerical reservoir simulators and their applications.

Stripping Compositional Effects

The other key compositional aspect of the immiscible gas/oil displacement process is the vaporizing (or stripping) by the lean injected gas of some hydrocarbon components of the oil, particularly the intermediate hydrocarbon components (C3 through C8). In most cases, the injected gas is very lean natural gas that is the residue gas from a nearby gas processing plant and composed primarily of methane. At such gas processing plants, the propane and heavier hydrocarbon components typically have been condensed from the entering produced gas; in some cases, ethane is also extracted from the gas. Such a lean injected gas will, when in contact with the oil at reservoir conditions, vaporize various hydrocarbon components from the oil until the gas and oil phases have reached compositional equilibrium.

In immiscible gas/oil displacements using nitrogen (N2), carbon dioxide (CO2), or some combination of these gases (such as flue gas, 88% N2, 12% CO2), these nonhydrocarbon injected gases can also vaporize various hydrocarbon components until gas/oil equilibrium is reached at reservoir conditions. This phenomenon has been observed in the Hawkins nitrogen injection project. Nitrogen is not as efficient as methane at stripping hydrocarbon components from the oil. Carbon dioxide, because its phase behavior is much like that of propane, can vaporize a considerable amount of hydrocarbon components from the oil at reservoir conditions.

The significance of the stripping effect depends on the oil composition. Immiscible gas/oil injection projects have been applied to reservoirs with oil gravities from 24 to 43°API or more.[14][25][26] In all cases, the stripping effect increases the recovery of hydrocarbons from the oil reservoir, but lighter oils have a much greater percentage of their components vaporized by cycling gas through the reservoir and operating at higher gas/oil ratios (GORs).

In some of the simple calculation techniques discussed below, aspects of the stripping of oil by lean injected gas can be approximated. Black-oil numerical reservoir simulators cannot handle the vaporization of hydrocarbon components from the oil into the gas phase. A compositional numerical reservoir simulator must be used to quantify this effect for a particular oil/injected-gas reservoir situation. These calculations are based on the use of an equation-of-state fluid characterization that is "tuned" to PVT laboratory data for the particular oil and potential injected-gas compositions. The compositional simulator also can quantify the effects of various surface facility configurations, including associated gas plants.

Calculation Methods. These compositional effects are too complicated to be quantified with hand-calculation methods. The black-oil numerical reservoir simulator can be used in limited ways to make these calculations. The black-oil model can reasonably handle the swelling effect, but it cannot handle the stripping effect at all.

The compositional effects of immiscible gas injection are best calculated using a compositional numerical reservoir simulator. To use such a model most accurately, considerable gas/oil PVT data need to be taken with the reservoir oil mixed with a range of gas compositions. These measurements should include a variety of special measurements, such as swelling experiments and stripping experiments, that can be used to develop a more accurate equation-of-state fluid characterization. Then the reservoir model can be used to quantify the performance of possible immiscible gas injection projects and associated surface facilities that might be built.

Surface Facility Considerations. In most applications of immiscible gas injection, during the early years of the projects, hydrocarbons are produced at about the original solution GOR. During this period, the impact of surface-facility design on the volume of produced hydrocarbon liquids is relatively limited. Later in the life of the projects when the producing GOR rises, the surface facilities are far more important in terms of both the volume of gas that can be handled and the extent to which this gas is processed. During this late period, the rate of oil production will be limited by the surface facilities’ gas-handling capacity.[27]

The gas recovered from the typical series of oilfield separators will contain a large amount of ethane through butane components and decreasing but significant amounts of the various heavier hydrocarbon components. Early generations of gas-processing plants were capable of separating out components that would condense down to temperatures of -10 to -20°F. Modern gas-processing plants operate at temperatures of -40°F or considerably lower; at these low temperatures, essentially all the ethane and heavier hydrocarbons are recovered, leaving a residue gas consisting primarily of methane. See the chapters in the Facilities and Construction Engineering section of the Handbook for additional discussion of oilfield surface separators and gas-processing plants.

Another related factor is the type of pipeline networks available to transport the hydrocarbon products to market. In some geographical areas, only those hydrocarbon components that can be stabilized in a crude oil stream can be transported and sold because there are only crude oil pipelines in that area. In much of the United States and Canada, a variety of crude oil and natural gas liquid (NGL) pipelines have been built so that the lighter, liquefied hydrocarbon components can also be marketed, particularly to the petrochemical industry of the U.S. Gulf Coast. In other parts of the world, as the number of liquefied petroleum gas tankers has increased, worldwide markets for the lighter hydrocarbons have developed. As a result, more oil fields have had large-scale gas-processing projects built to recover and market propane and heavier hydrocarbons from the produced gas streams.

Reservoir Geology Considerations Regarding Immiscible Gas Displacement

Many aspects of reservoir geology interplay with the immiscible gas/oil displacement process to determine overall recovery efficiency. Because there is always a considerable density difference between gas and oil, the extent to which vertical segregation of the fluids occurs and can be taken advantage of or controlled is critical to the success of gas displacing oil. See the reservoir geology chapter in this section of the Handbook for considerably more discussion of this topic.

General Geological Considerations

As with any oil recovery process involving the injection of one fluid to displace oil in the reservoir, the internal geometries of the reservoir interval have a controlling effect on how efficiently the injected fluid displaces the oil from the whole of the reservoir. For the immiscible gas/oil displacement process, the key factors are stratigraphy and structure.

Stratigraphy. The stratigraphy of a reservoir is determined primarily by its depositional environment. First and foremost is how layered the reservoir is in terms of both how heterogeneous the various sand intervals are and the scale at which shales or other barriers to vertical flow are interbedded with the sands. Another very important aspect is how continuous the shale intervals are. With fully continuous shales, a reservoir interval should be divided into compartments that will not interact with each other. Unless the reservoir has a steep dip, such shales will negatively affect the gas/oil displacement process.

Less continuous shales can result in better distribution of injected gas without a strong negative impact on the gas/oil gravity drainage process (see Fig. 12.7 for an illustrative cross-sectional view of sands interbedded with discontinuous shales). Richardson et al.[28] analyzed the effects of such limited-size shales. First, they determined geologic factors that control shale dimensions and continuity for sandstone deposits. There is a wide range of shales whose dimensions depend on the depositional environment, with marine shales being the most extensive and flood-plain and interdistributary shales being of smaller areal dimensions (1,300 to 5,250 ft wide by 5,250 to 10,500 ft long). Next, simple 2D calculations of oil drainage off small shales were made, assuming that the various beds were horizontal. They concluded that, "The time required for oil drainage from a barrier is proportional to its width squared and viscosity, and inversely proportional to the horizontal permeability and density difference. Lateral drainage off small barriers can be rapid, and recoveries may be reduced only slightly." [28]

Structure. The structural aspects of a particular reservoir consist of several parts: the closure or vertical thickness of the hydrocarbon column, dip angle of the beds, and size and relative thickness of the gas cap compared with the oil column. Thick reservoirs (> 600 ft of oil column) are the best for application of the immiscible gas/oil drainage process with gas injection at the crest of the structure and oil production from as far downdip as possible. Dip angle is important to the efficiency of the displacement process because a higher dip angle generally means that the effective vertical permeability is increased.

The relative size of the oil column compared with the gas cap affects the performance of a particular reservoir. The gas/oil gravity drainage process has been applied to reservoirs that have, relative to the size of the oil column, very small gas caps[26] and to some with very large gas caps.[29] Success has been achieved over the full range of ratios of gas cap to oil column size. The advantage of having a large initial gas cap is that the reservoir pressure drops very slowly as the oil is produced compared with a situation with a relatively small gas cap in which the reservoir pressure falls quite rapidly until the secondary gas cap grows sufficiently.

Other Geological Factors. Within the reservoir sandstone layers, the nature of the sand layering can strongly affect the efficiency of the gas/oil displacement. In those depositional environments in which the highest-permeability sands are on the bottom of the reservoir interval, the gas/oil displacement process will be far more efficient, especially compared with the situation in which the depositional environment results in the highest permeability toward the top of the reservoir interval. The reason is that, in the first situation, the gravity override of the gas is slowed by the vertical distribution of permeability, but in the latter situation, the gas gravity override is enhanced.

Even if the reservoir were totally homogeneous, a horizontal gas/oil displacement process would not be very efficient because the gas will strongly override the oil and, because of its high mobility, will rapidly travel from the injection wells to the production wells. For reservoirs with many "random" heterogeneities, the gas/oil displacement process will be aided because heterogeneities inhibit growth of low-viscosity fingers by forcing them to travel a more circuitous path between the injector and producer.

Vertical Permeability at Various Scales. The challenge in making calculations for the immiscible gas/oil displacement process at the reservoir scale is to quantify properly the vertical permeability of the reservoir as a whole. There are several scales at which vertical permeability affects the gas/oil displacement process. The engineer has quantitative data on the vertical permeability from routine core analysis performed foot by foot on small core plugs from the reservoir interval. The next larger scale concerns the areal extent of any impermeable layers observed in the cores, be they 1 in., 6 in., or several feet thick. Geologists typically estimate the areal dimensions of these impermeable layers from their training, experience, and studies of outcrops from similar depositional environments.

Despite all the technical work performed before this process is applied to a particular reservoir, the actual effective vertical permeability and its distribution will not be fully known until some years later. The vertical permeability can be quantified by observing reservoir performance. Typically, gravity-drainage immiscible gas/oil displacements are undertaken with the assumption of good vertical permeability so that if actual reservoir performance matches the projections, then the vertical permeability is as high as previously assumed. However, if the reservoir performance is poorer than expected, a likely cause is lower vertical permeability and/or heterogeneities in the vertical permeability distribution over the reservoir.

Carbonate Reservoirs. The geologic discussion above primarily concerns sandstone reservoirs, although many of the general concepts also apply to carbonate reservoirs. Because diagenetic changes often alter the original framework of a carbonate reservoir far more than what occurs in sandstone reservoirs and because some types of carbonate reservoirs do not have sandstone equivalents, this section briefly discusses some differences in carbonate reservoirs.

One type of carbonate deposits that results in reservoirs with thick vertical dimensions, especially compared with their areal dimensions, is the carbonate reef deposit. The style of this deposit with the greatest vertical-to-horizontal aspect is called a pinnacle reef. Reef deposits typically contain large vugs. The key question is, How interconnected are these vugs? Diagenetic processes can isolate these vugs or can provide various types of pore-to-pore interconnections. For example, in New Mexico, the Abo reef trend developed on the northern margin of the Delaware basin. The original reef framework of hydrocorals, sponges, and algae has been totally dolomitized to create a pore system consisting only of vugs, fractures, and fissures.[26]

In carbonate reservoirs, the diagenetic process includes both chemical alteration, such as dolomitization, and cementing and leaching processes. Cementation with calcite, anhydrite, or other insoluble chemicals can have a significant negative impact on the reservoir’s pore system. Leaching has the opposite effect and generally enhances the reservoir quality, although leaching may increase the range of heterogeneities and lead to some superconductive flow paths in portions of a reservoir. As carbonate rocks become more brittle because of chemical alteration, fracturing commonly occurs. The geologist and petrographer must examine the cores in great detail to determine the number and sequence of cementation, leaching, and fracturing events that have altered a particular rock interval over geologic time.

Middle East Carbonate Matrix/Fracture-System Reservoirs. A particular style of reservoir in which a considerable number of immiscible gas/oil gravity drainage projects have been applied is the Middle East carbonate matrix-block/fracture-system reservoirs. Most of these reservoirs are very large folded anticlinal structures with dimensions of tens of miles long by several miles wide and with hydrocarbon columns hundreds of feet to several thousand feet thick.

In these carbonate reservoirs, the matrix is high porosity but low permeability (generally ≤1 md), and the fracture system created matrix blocks with dimensions ranging from a few feet to > 10 ft.[30][31] The fractures can be up to a tenth of an inch wide, so the effective interwell permeabilities are very high.

The geologic complications of these matrix-block/fracture-system reservoirs concern the way that the matrix and fractures are interconnected and fluid is transferred between these two portions of the pore system. This combines (1) the interaction of the fractures with the matrix along the faces of vertical fractures, (2) the interaction of one matrix block with its neighboring matrix blocks if capillary continuity exists along such surfaces (see Fig. 12.8 for the schematic oil saturation profiles for cases without and with capillary continuity), and (3) possible fluid transfer along matrix/fracture surfaces if portions of the fracture system are inclined planes and neither vertical nor horizontal. Also, the presence of cementation along some of the fracture surfaces is very important to fluid transfer because the fractures will rapidly transport fluid, but for overall high recovery efficiency, the matrix blocks must exchange oil and gas with the surrounding fracture system. The geological aspects of such matrix-block/fracture systems are difficult to quantify because their dimensions and fracture characteristics cannot be easily discerned from cores and logs. Descriptions of nearby outcrops of the reservoir formation can often be helpful in understanding the macrodimensions of the matrix-block/fracture system.

A number of technical papers have explored aspects of the geology/fluid-flow interactions of such matrix-block/fracture network carbonate reservoirs. Firoozabadi and coworkers[32][33][34][35][36][37] have developed theories, made calculations, and performed experiments to explore aspects of these types of reservoirs. Saidi[30] has discussed the physical phenomena affecting the performance of the Haft Kel field (Iran) and analyzed its performance; more discussion of the Haft Kel field is found in the Case Studies section of this chapter.

General Immiscible Gas/Oil Displacement Techniques

In this section, the general technical features of the various immiscible gas injection projects are discussed.

Types of Gas-Injection Operations

Immiscible gas injection is usually classified as either crestal or pattern, depending on the location of the gas injection wells. The same physical principles of oil displacement apply to either type of operation; however, the overall objectives, type of field selected, and analytical procedures for predicting reservoir performance vary considerably by gas injection method.

Crestal Gas Injection. Crestal gas injection, sometimes called external or gas-cap injection, uses injection wells in higher structural positions, usually in the primary or secondary gas cap. This manner of injection is generally used in reservoirs with significant structural relief or thick oil columns with good vertical permeability. Injection wells are positioned to provide good areal distribution and to obtain maximum benefit of gravity drainage. The number of injection wells required for a specific reservoir depends on the injectivity of individual wells and the distribution needed to maximize the volume of the oil column contacted.

Crestal injection, when applicable, is superior to pattern injection because of the benefits of gravity drainage. In addition, crestal injection, if conducted at gravity-stable rates—e.g., less than the critical rate (see Eq. 12.8 later in this chapter)—will result in greater volumetric sweep efficiency than pattern injection operations. There are many examples of ongoing crestal injection projects throughout the world, including some very large projects in the Middle East.

Pattern Gas Injection. Pattern gas injection, sometimes called dispersed or internal gas injection, consists of a geometric arrangement of injection wells for the purpose of uniformly distributing the injected gas throughout the oil-productive portions of the reservoir. In practice, injection-well/production-well arrays often vary from the conventional regular pattern configurations—e.g., five-spot, seven-spot, nine-spot (see the chapter on waterflooding in this section for more description of these patterns)—to irregular injection-well spacing. The selection of an injection arrangement is a function of reservoir structure, sand continuity, permeability and porosity levels and variations, and the number and relative locations of existing wells.

This method of injection has been applied to reservoirs having low structural relief, relatively homogeneous reservoirs with low permeabilities, and reservoirs with low vertical permeability. Many early immiscible gas-injection projects were of this type. The greater injection-well density results in pattern gas injection, rapid pressure and production response, and shortened reservoir depletion times.

There are several limitations to pattern-type gas injection. Little or no improvement in recovery is derived from structural position or gravity drainage because both injection and production wells are located in all areas of the reservoir. Low areal sweep efficiency results from gas override in thin stringers and by viscous fingering of gas caused by high flow velocities and adverse mobility ratios. High injection-well density increases installation and operating costs. Typical results of applying pattern injection in low-dip reservoirs are rapid gas breakthrough, high producing GORs, significant gas compression costs to reinject the gas into the reservoir, and an improved recovery of < 10% of original oil in place (OOIP). Note that gas inefficiently displaces oil in gas-swept areas. Attempts to subsequently waterflood such areas result in rapid water breakthrough and little, if any, additional oil displacement.

Few pattern gas injection projects have been implemented in recent years because this method is not as attractive economically as alternative methods for increasing oil recovery.

Optimum Time To Initiate Gas Injection Operations

The optimum time to begin gas injection is site specific and depends on a balance of risks, gas market availability, environmental considerations, and other factors that affect project economics. When only oil recovery and improvements in reservoir producing characteristics are considered, reservoir conditions for gas injection operations are usually more favorable when the reservoir is at or slightly below the oil bubblepoint pressure, unless the bubblepoint pressure is low compared with the initial reservoir pressure. Near the oil bubblepoint pressure, nonrecovered oil represents the smallest volume of stock-tank oil, oil relative permeability is high, and oil viscosity is low.

Efficiencies of Oil Recovery by Immiscible Gas Displacement

It is customary in most displacement processes to relate recovery efficiency to displacement efficiency and volumetric sweep efficiency. The product of these factors provides an estimate of recoverable oil expressed as a percentage of OOIP. Analytical procedures are available for evaluating each efficiency factor. For the purposes of this chapter, the two components describing the overall recovery efficiency are defined as follows:

  1. Displacement efficiency is the percentage of oil in place within a totally swept reservoir rock volume that is recovered as a result of viscous displacement and gravity drainage processes.
  2. Volumetric sweep efficiency is the percentage of the total rock or PV that is swept by gas. This factor is sometimes divided into horizontal and vertical components, with the product of the two components representing the volumetric sweep.

Recovery efficiencies increase with continued gas injection, but the rate of recovery diminishes after gas breakthrough occurs as the GOR increases. The overall result is that the ultimate oil recovery efficiency is a function of economic considerations, such as the cost of gas compression and the volume and availability of lean residue gas or potentially more expensive alternatives like N2 from a nitrogen rejection plant.

Vertical or Gravity Drainage Gas Displacement

In this section, the primary manner in which the immiscible gas/oil displacement process has been used is discussed in qualitative terms. This is the use of gas injection high on structure to displace oil downdip toward the production wells that are completed low in the oil column. In many cases, an original gas cap was present, so the gas was injected into that gas-cap interval (see Fig. 12.9 for cross-sectional view of anticlinal reservoir with gas cap over oil column with dip angle α and thickness h). In this situation, the force of gravity is at work, trying to stabilize the downward gas/oil displacement process by keeping the gas on top of the oil and counteracting the unstable gas/oil viscous displacement process. If the oil production rate is kept below the critical rate, then the gas/oil contact (GOC) will move downward at a uniform rate. In the next section, the simple engineering calculation techniques for estimating the rate for stable gravity drainage for a gas/oil system are discussed.

In normal oilfield operations, producing wells are drilled and perforated as far as possible from the gas cap, both vertically and laterally. Hence, many production wells may not have any of the gas cap directly over them initially. Until the gas cap extends over the whole of the oil column and its producing wells, because of the pressure differential between the gas cap and the area around the producing wells, the downdip oil production will result in the gas cap moving rapidly along the top of the structure to provide pressure support to the whole oil column. When this happens, if the reservoir is a dipping structure, the gas cap is described as "draping" over the oil column. If the oil reservoir is an anticlinal structure, then this process is often described as the gas cap having expanded to form an "umbrella" over the whole oil column. Once the gas-cap drape or umbrella has formed, the continued downward movement of the gas cap at each location can be stabilized. Fig. 12.10 shows views of the Prudhoe Bay oil field (Alaskan North Slope) in which the original gas cap was offset to the northeast of most of the oil column and where there has been gas cap underrunning of the top of structure until it draped over much of the downdip oil column where the hundreds of producing wells were located.

There are likely to be local variations in the GOC caused by reservoir heterogeneities and near-wellbore pressure gradients. The most notable of these results in gas coning caused by high pressure gradients around the perforated interval of each wellbore. Here, the controlling factors are the oil and gas production rates, the distance from the top of the perforations to the overlying GOC, and the horizontal and vertical permeabilities.[38] In this situation, the presence of a small shale interval between the GOC and the top of the perforated interval can have a very beneficial effect on the maximum oil production rate before gas coning occurs (see Fig. 12.11). For a particular reservoir situation, gas-coning calculations are best made with a numerical reservoir simulation model.[38][39]

The gas/oil gravity drainage process is complicated if the oil column is underlain by an aquifer because the aquifer will provide pressure support to the oil column in response to any decrease from original reservoir pressure caused by oil production. If the aquifer is very strong, it will invade the lower portions of the oil column and may provide almost barrel-for-barrel voidage replacement. In this case, the original gas cap may not expand much. If the aquifer is weak or if there is a tar mat at the oil/water contact (OWC) inhibiting water influx, then the gas cap will be the primary means of pressure support for the oil column and the reservoir will perform almost as if there were no aquifer present. A problem sometimes experienced with oil reservoirs with both overlying gas caps and underlying aquifers is that the near-wellbore coning behavior is more complicated. The reason is that gas-cap gas is coning downward toward the perforated interval and aquifer water is trying to cone upward toward the same perforated interval. If water cones first into the perforated interval, then the gas coning will be more severe because, with three-phase relative permeability effects, the near-wellbore pressure gradients are greater, which causes gas coning to occur at lower oil production rates.

Calculation Methods for Immiscible Gas Displacement

Techniques described in this chapter are classic methods for describing immiscible displacement assuming equilibrium between injected gas and displaced oil phases while accounting for differing physical characteristics of the fluids, the effects of reservoir heterogeneities, and injection/production well configurations. The reservoir is treated in terms of average properties for volume of rock, and production performance is described on the basis of an average well. Black-oil-type reservoir simulation models use essentially these same techniques but, by means of 1D, 2D, or 3D cell arrays, account for areal and vertical variations in rock and fluid properties, well-to-well gravity effects, and individual well characteristics. More complex compositional models account for nonequilibrium conditions between injected and displaced fluids and can be used to describe individual well streams in terms of the compositions of the produced fluids.

The increasing capability of desktop computers and the growing amount of affordable simulation software are making it possible to use numerical reservoir simulation more often. However, results obtained from simulation will be directly dependent on the quality of data to describe the reservoir rocks and fluids. It is also important to comprehend the physics of displacement to understand the simulation results and to identify incorrect results. The fundamentals of the displacement process presented in this section are intended to provide the background needed to produce good-quality predictions of oil recoveries.

Modifications of Displacement Equations

Applicability of the basic displacement equations to a given reservoir is governed by whether the underlying assumptions are reasonable. Several authors have reported modifications that eliminate the need to make certain assumptions. Modifications that take into consideration the swelling effects experienced from injection into an undersaturated reservoir and production of fluids from behind the gas front have been presented by Welge,[18] Kern,[19] Shreve and Welch,[40] and others. Jacoby and Berry,[41] Attra,[42] and others have presented equations and simple analytical procedures for calculating performance to account for some of the compositional interchange between the displacing gas and the reservoir oil.

These works are mentioned for completeness. If significant deviations from the basic assumptions of the Buckley-Leverett method are a concern, the more practical approach is to use numerical reservoir simulation to account for reservoir heterogeneities and gravity, capillary, and compositional effects. These simulators are discussed in the numerical simulation chapter in this section of the Handbook.

Methods for Evaluating Sweep Efficiency

Some techniques for estimating the volumetric, vertical, and areal sweep efficiency of an immiscible gas/oil displacement are discussed below.

History Matching. If there are sufficient data concerning the location of the gas front and oil recovery as a function of time, past reservoir performance can be used to calculate the volumetric sweep efficiency by dividing observed recovery at various times by the theoretical recovery determined from displacement efficiency calculations.

If there are adequate data to reliably describe spatial variations in reservoir rock and fluid characteristics, numerical reservoir simulation is the best way to predict sweep efficiency, particularly after the historical production and pressure data are matched. If data or sufficient economic justification to undertake a full numerical reservoir simulation study is lacking, the following methods are presented as useful for screening studies and in situations when more detailed studies are inappropriate.

Vertical Sweep Efficiency. Several authors have presented methods for determining vertical sweep efficiency based on statistical treatments of routine core analysis data. Some of the most frequently used methods are adaptations of the Stiles[43] method for evaluating the effect of permeability variations on waterflood performance. The same assumptions and calculation procedures may be used for immiscible gas/oil displacements. The relative permeability ratio used in such calculations is considered to be a constant equal to the relative permeability to gas at residual oil saturation (krg@Sor) divided by the relative permeability to oil at initial gas saturation (kro@Sgi).

Areal Sweep Efficiency. Several investigators have shown that areal sweep efficiency is primarily a function of injection/production well pattern arrangement, mobility ratio, and volume of displacing phase injected. Various studies have confirmed what would be expected intuitively, that areal sweep efficiency increases with the volume injected and with a lower mobility ratio. Data from model studies that show the influence of mobility ratio and displacement volume on areal sweep efficiency in a regular five-spot pattern are illustrated in Fig. 12.12.[44]

Thin models containing miscible fluids of varying viscosity were used to develop these area sweepout curves. These data are considered applicable to either water/oil or gas/oil displacement. These data are presented to aid in the understanding of the effect of some factors on the gas displacement mechanism and may prove useful in preliminary studies of a potential gas injection project to predict volumetric sweep. However, the quantitative applicability of laboratory data is inherently questionable because of uncertainties in model scaling, laboratory techniques, and associated simplifying assumptions regarding no vertical gas override effects or reservoir heterogeneities. The instability of the very unfavorable mobility ratio gas/oil displacement is most difficult to quantify in laboratory experiments. All these effects can cause a smaller sweepout efficiency than presented in Fig. 12.12. Nevertheless, laboratory model studies do offer a convenient means of making quantitative estimates when simulation is not practical or justified and injected gas remains dispersed in the reservoir.

When the laboratory data are used, the common practice is to calculate a mobility ratio using the viscosity and relative permeabilities of the oil ahead of the gas front and of the gas at the average saturation behind the displacing front.

Calculating Immiscible Gas Injection Performance

Numerical simulation represents the best way to predict the performance of immiscible gas injection if there are sufficient data to characterize the reservoir rocks and fluids adequately. Even simple 2D and 3D black-oil models provide insight into the more important aspects of oil recovery for reservoirs in which compositional effects are not a major concern. When adequate data are unavailable or when screening work is being done, simple models may suffice.

The immiscible displacement of oil by gas is described with fractional flow theory. Muskat[1] presented the basics of this theory more than 60 years ago. Since then, additional work has been done to develop various mathematical calculation methods based on fractional flow theory. A few of the more recent papers discussing these techniques are listed in the References section.[45][46][47][48][49][50][51]

As discussed above, pattern gas injection is seldom used now because waterflooding performance is much better in those types of reservoirs where pattern gas injection has historically been tried. Therefore, the remainder of this section discusses a simple model used for reservoirs in which stabilized gravity drainage controls the gas/oil displacement process and increases the ultimate oil recovery.[13]

Viscous, gravitational, and capillary forces and diffusion are involved in the displacement of oil by gas, complicating technical analysis of a particular reservoir if each of these forces and flow in all three dimensions are important. Fortunately, there are instances in which one force is dominant and only one dimension is involved in the rate-limiting step. In these circumstances, engineering solutions can be direct and simple. One such circumstance is that of thick reservoirs with high permeabilities.

In steeply dipping oil reservoirs containing sands with high vertical permeabilities, gravity drainage of the oil can be more effective than is calculated from the Buckley-Leverett assumptions alone, as illustrated in Fig. 12.13.[52] When sufficient vertical permeability exists, even at lower oil saturations, oil behind the gas front can continue to flow vertically downward through the reservoir. Thus, the displacement process occurs in two steps. First, gas invades the originally oil-saturated sand as the GOC moves downdip because of oil production farther downdip. Second, oil drains vertically downward through the gas-invaded region and forms a thin layer with high oil saturation (with high kro) that drains along the base of the reservoir interval to the remaining downdip oil column.

Mathematical Model. A simple mathematical model can be used to describe the displacement of oil by gas drive and gravity drainage when the rate is less than one-half the critical rate. The critical rate is given by


where qT = total volumetric flow rate through area A, ft3/D, and k = permeability, darcies.

The first calculation determines the gas saturation just above the GOC by using Eq. 12.5, plotting Fg vs. Sg, and finding the tangent to the curve passing through the origin, as shown in Fig. 12.2. For ease of calculation, the GOC is assumed to move at a constant rate. The next calculation determines the quantity of oil that drains from the region invaded by gas in a given time increment. For ease of calculation, this region is divided into arbitrary lengths, and the amount of oil produced by vertical gravity drainage is calculated for the average time since passage of the gas front.

For vertical drainage of oil, the rate is given by Darcy’s law, with the driving force proportional to the density difference between gas and oil. The assumption is made that resistance to flow of gas and capillary effects are negligible[53]:


where uov = vertical oil flow per unit area, res ft3/ft2-D.

From continuity considerations,


where z = vertical distance, ft; t = time, days; ϕ = porosity, fraction; and So = oil saturation, fraction.

The rate of movement of a particular saturation, (dz/dt)So, can be determined by plotting uov calculated from Eq. 12.9 vs. saturation, taking the slope to determine duov/dSo and dividing by porosity, as indicated in Eq. 12.10. The amount of oil drained from each region since passage of the gas front can be calculated by graphical integration of the height vs. saturation plot.

As a first approximation, the time for oil to flow downdip in the thin layer along the base of the reservoir interval can be neglected, as can the volume of oil in this layer. If the displacement rate exceeds one-half the critical rate, oil tends to accumulate rather than flow away along the bottom of the reservoir. More accurate calculations also include consideration of the thickness of the gas/oil transition zone arising from capillary effects above this layer of oil, especially if the transition zone is < 10 ft thick.

Recoveries calculated by this technique are quite sensitive to the values of kro at low oil saturations.[10] Ways to extrapolate measured information are discussed next. First, conventional laboratory data can be extended to low oil saturations by plotting measured values of kro vs. (SoSorg*)/(1 − SwiSorg*), in which Sorg* is the irreducible oil saturation in the presence of gas and connate water. Sorg* can be calculated by material balance for areas of a reservoir that have been invaded by gas if good data are available on GOC movement and oil recovery from the area. A second source is the oil saturation in an associated gas cap as determined in cores from that region. If the gas cap was originally filled with oil, drainage of oil over geologic time as gas migrates into the reservoir establishes an endpoint relict oil saturation. For instance, the Prudhoe Bay Sadlerochit reservoir was originally filled with oil. Gas then migrated into the reservoir several million years later, creating the gas cap.[54] Water-based mud cores from the gas cap interval showed an average routine core analysis oil saturation at discovery of 7% PV. The dip of the reservoir is 1 to 3°, but vertical permeabilities throughout the gas and oil columns are generally very high. Interestingly, oil saturations above small shale lenses in the gas cap averaged more than 7% PV, indicating that more time may be required to reach irreducible oil saturations when oil drainage is limited by the dip of this reservoir. A third source is drainage capillary pressure vs. saturation measurements. Experience has indicated that Sorg* should be < 10% PV and sometimes approaches zero. Although these endpoint saturations are seldom realized in the depletion time of a reservoir, it is important to have the correct value for predicting flow behavior and ultimate oil recovery. A benefit of even simple, multidimensional simulation models is that the inclusion of capillary effects controls the oil flow rate and conditions under which irreducible saturations are approached.

If a measured value of Sorg* is unavailable, a value is chosen to yield a straight line through the data, so for


the slope of the line n should be ≈4 according to the theory of Corey et al.[55] but may be as large as 6 and as small as 2. Recovery data can be correlated by the dimensionless parameter RTENOTITLE, which is derived by dividing the time required for vertical drainage,


by the time required for flow along the bedding plane,


where kv = vertical permeability, darcies, and hv = vertical thickness, ft.

Example Gas/Oil Gravity-Drainage Problem. The utility of this simple model can be illustrated by predicting recovery by gas drive and gravity drainage for an actual reservoir, in this case the Hawkins field in east Texas.[56]

Given: Average Hawkins Woodbine reservoir properties as presented in Table 12.2 and Fig. 12.14.

Solution: With Eq. 12.8, the critical rate for the Hawkins field is calculated to be 0.173 ft3/D-ft2. The average actual rate, qT/A, of 0.0365 ft3/D-ft2 (see Table 12.2) is 21% of the critical rate, and the simplified model should apply. With Eq. 12.5, the average gas saturation just above the GOC is found to be 46% PV by the Welge procedure, as shown in Fig. 12.2. The rate of frontal movement is


Time to gas breakthrough is 3,500/104 = 34 years. Recovery at breakthrough may now be estimated by dividing the reservoir into seven blocks, each 500 ft long and 49 ft thick. The average vertical movement of saturations in each block can be calculated from Eqs. 12.9 and 12.10. The relative permeability data for oil were extrapolated to low So values using the correlation term (SoSorg*)/(1 – SwiSorg*) discussed above. The resulting plots for two Sorg* assumptions are shown in Fig. 12.15.

In this case, Sorg* = 0 gave a slightly better fit. The resulting plot of kro vs. So is shown in Fig. 12.16. The laboratory data and a plot resulting from use of an Sorg* of 2% PV and an n of 4.5 are also shown to indicate the differences in kro at low So that result.

The average time interval since passage of the gas front is calculated from the calculated rate of frontal movement of 104 ft/yr. The residual oil left in each block is determined by graphical integration of a plot of So vs. height, such as the ones shown in Fig. 12.17. Results of these calculations are shown in Table 12.3.


Comparisons With Field Data. A recovery of 87% of OOIP was observed in the Hawkins field for an area affected by an expanding gas cap.[56] The calculated recovery of 88% compares very favorably. A 2D two-phase reservoir simulation using similar relative permeabilities predicted 87% recovery at breakthrough. The recoveries observed in the field and predicted by models that permit flow in two dimensions are > 15% greater than those calculated by conventional 1D techniques that assume flow only along the bedding planes.

Model Summary. This section has shown how a simple gravity drainage model can be readily applied to predict recoveries by gas drive and gravity drainage when flow rates are less than one-half the critical rate and permeabilities in the vertical direction are high. Some applications of the model have been unsuccessful because of lower-than-expected vertical permeabilities. As a practical matter, the simple model should be used to predict reservoir behavior only when it can be shown to match history or when applied to a field analogous to one that the model fits.

Immiscible Gasflood Monitoring

There are a variety of methods for monitoring immiscible gas injection projects. Some apply both to the pattern type of gas injection projects and to the vertical gravity-drainage type of gas injection projects; others apply only to the gravity-drainage projects. In all cases, individual well production and pressure performance as a function of time must be recorded.

The most obvious monitoring method is to track the GORs of the individual production wells as a function of time. The GOR will be approximately flat at the oil’s solution GOR until there is gas breakthrough. Then the GOR will climb. The timing of gas breakthrough and the rate of GOR climb will indicate how efficiently, or inefficiently, the gas/oil displacement is progressing. Field engineers should have made preliminary calculations, possibly using a numerical reservoir simulator, so that they have projections of what should be expected regarding gas breakthrough timing and GOR increases at the individual well locations (and as a function of the volume of gas injected at the individual injection wells).

The other methods concern primarily the monitoring of the vertical movement of the gas/oil interface (the current GOC) in gravity-drainage-type projects. Two techniques are generally used: cased-hole logging programs and monitor-well observations. Both help track the GOC movement as a function of time. By mapping the GOCs from the individual wells, engineers can determine whether the GOC is staying reasonably horizontal. By comparing the GOC movement as a function of time to the projected GOC behavior, engineers can determine how efficient the gas/oil displacement is and whether the project’s expectations are being met.

For some reservoirs, other unique techniques can be used. For example, if a reservoir has a natural oil gravity variation as a function of depth, then the production wells’ oil gravity can be tracked as a function of time.

Another technique is to take periodic gas samples and perform gas chromatograph analyses to determine the produced gas composition. To use this technique, baseline gas samples should be taken early and periodically from all wells. There are two circumstances in which gas chromatography is a useful tool for gas-injection project monitoring. The first is those projects in which flue gas (88% N2, 12% CO2) or pure N2 is injected. In that type of project, it is important to track the BTU value of the gas from each well and how the nonhydrocarbon content of the produced or residue gas changes as a function of time. This is important with respect to the use of that gas for field fuel and the marketability of the residue gas stream.

Second, this technique can be important late in the life of a gas-injection project when wells are operating at very high GORs. This technique is useful for determining whether some of the very lean injection gas is breaking through into some of the wells without becoming saturated with light and intermediate hydrocarbon components from the residual oil phase in the gas-swept region above the current GOC. Such lean gas can actually strip intermediate hydrocarbon components from the produced oil in the field gas/oil separators. If there is no gas plant as part of the field facilities, then those hydrocarbons will be lost to the downstream owner of the gas plant that processes the field gas.

One other aspect of the monitoring activities is to track from which of the perforated intervals most of the gas flow is entering the wellbores. This can be accomplished with periodic spinner or temperature surveys. Depending on the nature of the reservoir interval, it may be possible to temporarily plug off some of the perforations to reduce wells’ producing GOR.

The purpose of all these monitoring activities is to perform real-time analysis of reservoir performance and to consider any remedial actions. These actions include such alternatives as rebalancing the gas rates into the various injectors, rebalancing the flow rates of the various producers, and potentially drilling a few new injection or production wells into areas of the reservoir determined to require more drainage points or to improve the overall sweep efficiency.

Field Case Studies: Immiscible Gas Injection Examples

In this section, a number of field applications of immiscible gas injection are briefly reviewed. In each case, the following facts will be listed, to the extent that they are available.

  • Field name.
  • Structural closure.
  • Rock type.
  • Nature of pore system
  • Average permeability.
  • Source of injected gas.
  • Field size.
  • Presence of initial gas cap.
  • Oil gravity.
  • Initial reservoir pressure.
  • Oil bubblepoint pressure.

Both successful and unsuccessful immiscible gas injection projects are discussed. A number of early successful and unsuccessful gas injection projects are summarized by Muskat in his 1949 classic book Physical Principles of Oil Production.[1] Most of this section describes successful projects because those have been presented in a number of SPE technical papers. Immiscible gas injection has been used in oil fields with a wide range of characteristics.

Early Successful and Unsuccessful Immiscible Gas Injection Projects

Muskat discusses four gas injection projects in the United States from the 1930s and 1940s.[1] Two of these projects were termed successes, and two were viewed as having poor response.

Cunningham Pool (Kansas). This 1,400-acre anticlinal 31 to 36°API oil field had a maximum closure of 75 ft and 53 producing wells. The reservoir is an oolitic limestone and had an initial gas cap. Field discovery was in 1932, and gas injection began in 1936 into three to five wells after the reservoir pressure had declined from 1,115 to 424 psi. The average reservoir properties were as follows: net thickness, 8 ft; porosity, 11% BV; and permeability, 105 md. Muskat termed this project a success because of its GOR history and concluded that "in spite of the thin pay section there has been effective gas segregation, so that the injected gas largely remained trapped in the reservoir and helped to sustain the oil saturation within the oil zone." [1]

Schuler Field (Arkansas). This 3,000-acre anticlinal 34°API oil field had a maximum closure of 135 ft and 146 wells drilled on 20-acre well spacing. This sandstone reservoir had a small initial gas cap. Field discovery was in 1937; gas injection began in 1941 into six wells at the crest of the structure and after the reservoir pressure had declined from 3,520 to about 1,550 psi. The average reservoir properties were as follows: net thickness, 0 to 70 ft; porosity, 17.6% BV; and permeability, 0 to 4,000 md (355-md average). Success was achieved because the reservoir pressure was stabilized, producing GORs were decreased, and produced gas was reinjected instead of flared. Because of proration limits on production, the allowed oil rate was being produced from a few optimally located downdip wells.

Grayburg Lime (West Texas). This 750-acre section of a west Texas anticlinal reservoir is a sandy dolomite and had an average net pay of 18 ft (in a gross thickness of 130 ft). Pay porosity ranged from 8 to 14% BV, and permeabilities ranged from 2 to 10 md. Oil gravities were in the 33 to 37°API range. In 1942, 1 of the 26 producing wells was converted to gas injection after the reservoir pressure had declined from 1,800 to 1,275 psi. Approximately a year after the start of gas injection, there was a sharp rise in the producing GOR. Because the oil production rate was considerably increased at about the same time as the GOR rise, Muskat reasons that, although the gas may have traveled rapidly through "a substantially continuous and intercommunicating fracture system," the GOR rise may have resulted from the inability of the reservoir matrix to supply oil to the fracture system at the higher oil rate.

Canal Field(California). This 1,100-acre dome oil field had a maximum closure of 150 ft with wells drilled on 20-acre well spacing. The reservoir is a sand of variable character with silt/shale streaks. There was no original gas cap; the original reservoir pressure was 3,550 psi; and the oil bubblepoint pressure was 2,800 psi. Porosity ranged from 15 to 32% BV with an average of 22% BV, and permeability ranged from 10 to 1,000 md with an average of 200 md. Field discovery was in 1937. Gas injection began in 1942 into a single crestal well, but within a year, two other injection wells located down the flanks of the structural axis were added. Within 6 months, the ethyl mercaptan tracer was spotted in one well, and within a year, two other wells showed tracer responses. It was concluded that "the appearance of the tracer at the producing wells definitely proves gas channeling through high-permeability streaks, rather than a uniform drive through the sand as a whole." [1]

In these four situations, the first two were deemed successful applications of immiscible gas injection; the last two were deemed failures. All these fields were relatively small compared with those discussed below. In small fields, there is less opportunity to optimize well placement and make changes in the course of the project life.

Hawkins Field (East Texas).[4][25][56][57] The Hawkins field contains > 1.3 billion bbl OOIP and 430 Bcf gas in the original gas cap. The reservoir consists of two high-quality sandstone intervals (27% BV porosity and 1- to 3-darcy permeability), the Lewisville and the Dexter; the Dexter, the better-quality sandstone, contains 70% of the oil. The structure caused by a deep-seated salt dome has 1,200 ft of closure and is extensively faulted (see structure map in Fig. 12.18). The reservoir is divided into two areas separated by a major fault. The eastern area contains 20% of the OOIP and 7% of the original gas cap gas and is underlain by the active Woodbine aquifer that covers much of east Texas. The western area contains the rest of the oil and gas. The western area has a tar mat that varies in thickness from 50 ft on the north to 100 ft on the south. This tar mat impeded aquifer influx until a decline in reservoir pressure resulted in water influx in the north that constituted a strong waterdrive that tilted the gas cap to the south, where aquifer influx did not occur. The average formation dip is 6 to 8°. Oil gravity averages 24.2°API gravity and varies somewhat vertically (21 to 26°API range). Oil viscosity averages about 3.7 cp (with values as high as 15 cp observed near the OWC on the east side), and its FVF is 1.22 RB/STB.

The 10,000-acre field, discovered in 1940, was developed on 20-acre well spacing. Unitization was completed on 1 January 1 1975 after 536 × 106 bbl of oil production. Before unitization, oil production was supported by gas-cap expansion and aquifer influx in the east. A small gas injection project was used in the west to stabilize the GOC and to prevent oil migration into the gas cap because of the aquifer influx from the north.

Extensive laboratory testing was conducted on reservoir core samples to quantify the ability of both water and gas to displace oil.[25] The results of these tests are shown in Table 12.4. These tests showed that gas was more efficient at displacing oil from the reservoir rock than water and that gas would recover at least 10% PV more oil. From the laboratory data, engineers calculated that, in the field, waterdrive would leave a residual oil saturation of 35% PV, whereas gas drive would leave an average residual oil saturation of 12% PV; the difference results from the lower density difference between the oil and water.

The field was unitized to facilitate the implementation of a gravity drainage project using crestal gas injection. Gas injection began in 1975. Two types of gases were injected. All produced gas, less fuel and shrinkage, was reinjected into the gas cap areas, and beginning in 1977, 120 MMcf/D of flue gas (88% N2, 12% CO2) generated at a nearby plant was also injected.[4] More recently, pure nitrogen from a cryogenic nitrogen rejection plant has been injected.

In 1987, a tertiary immiscible gas-drive process was started in the East Fault Block where the aquifer had invaded a large portion of the oil column. This tertiary process has been called the double displacement process (DDP).[25][57] In this process, the invading aquifer is being displaced to the original OWC so that the gas-drive gravity drainage process can remobilize much of the waterflood residual oil all the way down to this depth. Although the DDP is working, it is working more slowly than expected because of "higher viscosity oil (note the higher viscosity oil downdip discussed above), significant targeted oil volume found in lower-quality rock (in bypassed-oil zones), and lower-than-expected oil relative permeability." [57] With the success of the DDP in the east, a similar project was implemented in the west.

Overall Hawkins’ recovery efficiency from the gas-drive mechanism is about 87% in the gas-swept areas or > 20% better than estimated for the waterdrive process. Overall reservoir performance resulting from immiscible gas injection is considered excellent.

Prudhoe Bay Field (North Slope, Alaska).[13][27][38][54][58] The Prudhoe Bay field, the largest oil and gas field in North America, was discovered in 1968. The main Permo-Triassic reservoir is a thick deltaic high-quality sandstone deposit about 500 ft thick with porosities of 15 to 30% BV and permeabilities ranging from 50 to 3,000 md. The field contains > 20 × 109 bbl of oil overlain by a 35-Tcf gas cap. The reservoir, a monocline structure dipping 1 to 2° to the south and southwest, is bounded on the north by major faults and on the east by a major lower Cretaceous truncation. The oil averages 27.6°API gravity and has an original solution GOR of about 735 scf/STB. Under much of the oil column area, there is a 20- to 60-ft-thick tar mat located above the OWC.

Because of its remote location, first production did not begin until 1977 after an oil pipeline across Alaska to the southern port of Valdez was built and extensive oilfield facilities were installed. The initial facilities had a gas-handling capacity of about 2 Bcf/D to separate and compress the produced gas for reinjection into the updip gas cap. For the regulated oil production rate of 1.2 × 106 BOPD, this was sufficient for more than the first decade of field operations. As of 2004, no gas pipeline has been built from the North Slope of Alaska.

Immiscible gas/oil displacement has been the production mechanism at work over much of the Prudhoe Bay oil column (Fig. 12.10), and waterflooding and miscible WAG processes have been used in the very downdip portions of the oil column. More recently, horizontal wells are being used nearer the GOC to exploit the thin oil columns that cannot be drained with existing wells. During the first decade of operations, the original gas cap had expanded to drape over much of the oil column area, but the producing GOR was kept low by perforating the production wells as far as possible, both vertically and laterally, from the gas cap. By the mid-1980s the engineers determined that maintaining the oil production rate depended on expanding the field’s gas-handling capacity. Two major projects were undertaken. First, a central gas facility was constructed to strip the produced gas of many of the hydrocarbon components for sale as blendable NGLs and for use as miscible injectant. This project also increased the field gas handling capacity to 3.3 Bcf/D. The second project was a sequence of gashandling facility expansions that increased overall capacity to > 7.5 Bcf/D.

The reinjection of residue gas served two purposes. First, it maintained reservoir pressure, which helped increase oil recovery; the reservoir pressure declined only by ≈1,000 psi (from 4,300 to 3,300 psi) during the production of the first 10 × 109 bbl of oil. Second, the very lean residue gas has vaporized large amounts of hydrocarbon components from the relict oil saturation in the original gas cap[44] and from the remaining oil behind the gas/oil front. Simulations showed that the vaporization mechanism would contribute recovery of an additional 4 to 6 STB/MMcf of additional gas produced.[13] During the past 15 years, the average API gravity of the marketed Prudhoe Bay hydrocarbon liquids has increased by > 5°API.

Haldorsen et al.[58] have studied the efficiency of the Prudhoe Bay gravity drainage mechanism by means of laboratory experiments and field cased-hole log evaluations. They concluded, "The ‘most likely’ displacement efficiency, through the stochastic approach, was 68 percent after three years and 76 percent after 30 years of gravity drainage."

The ultimate Prudhoe Bay oil recovery has increased from initial estimates of 9.6 to 13 × 109 bbl oil, much of which is related to exploiting the immiscible gas/oil gravity drainage and oil stripping mechanisms. This has been accomplished by massive expansions of the gas handling facilities and extraction of the maximum volume of blendable NGLs from the produced gas stream. Application of the immiscible gas processes at Prudhoe Bay has been aided by the lack of alternative uses or markets for the produced gas.

Empire Abo Field (New Mexico).[26][39] Another example of the application of the immiscible gas/oil gravity drainage process is found at the Empire Abo field. This field covers ≈11,000 acres (12.5 miles long by 1.5 miles wide) and contains approximately 380 million STB of OOIP. This reservoir is a dolomitized reef structure (Fig. 12.19) with a dip angle of 10 to 20° from the crest toward the fore reef. The oil column is approximately 900 ft thick, but the average net pay is only 151 ft thick. The pore system of this reservoir is a network of vugs, fractures, and fissures because the primary pore system has been so altered by dolomitization; the average log-calculated porosity was 6.4% BV. Numerical simulations of field performance and routine core analysis data have indicated that the horizontal and vertical permeabilities are about equal. The Empire Abo field has a small initial gas cap (< 1% of the hydrocarbon PV), so the oil was gas saturated at the original reservoir pressure of 2,360 psia. The oil gravity was 43°API, and viscosity at reservoir conditions was approximately 0.4 cp.

The field was discovered in 1957 and unitized in 1973. At that time, high-GOR back-reef wells were shut in, only the downdip solution-GOR wells continued to produce, and gas-plant residue gas was reinjected into some of the former crestal production wells. The unit added 56 infill producers on 20-acre well spacing in the downdip area during the mid-1970s (see structure map in Fig. 12.20). In the late 1970s, 100 additional producers on 10-acre well spacing in the downdip area were drilled.

Because of the concentration of producing wells downdip, the gas cap draped downdip along the top of the reservoir. The Empire Abo field overall performance was excellent because of its very high vertical permeability; however, gas coning was a major issue, with the overall relatively low reservoir permeability of approximately 50 md. In 1979, Empire Abo was the site of one of the first applications of horizontal wellbores to minimize gas coning.[59] The field is currently being blown down to recover as much gas as possible. On the basis of production data through 2002, on a stock-tank-oil basis, approximately 74% of the OOIP has been recovered from this field by application of the immiscible gas/oil gravity drainage process.

Heft Kel Field (Iran)[30] An example of the application of immiscible gas/oil displacement in a large Middle East matrix-block/fracture-system reservoir is the Heft Kel field. Its Asmari reservoir structure is a strongly folded anticline that is 20 miles long by 1.5 to 3 miles wide with an oil column thickness of approximately 2,000 ft. The most probable OOIP was slightly > 7 × 109 STB with about 200 million STB in the fissures; numerical model history matching resulted in a value of 6.9 × 109 STB. The matrix block size determined from cores and flowmeter surveys varied from 8 to 14 ft. The numerical simulation model considered matrix permeabilities from 0.05 to 0.8 md. The overall horizontal and vertical permeabilities are approximately equal. There was an initial gas cap on the oil column. The oil gravity is approximately 37°API. The IFT at the bubblepoint pressure (1,412 psi and 116°F) is approximately 9 dynes/cm.

The field was discovered and put on production in 1928. It was produced on primary production from then until 1976 with a plateau rate of 200,000 BOPD for several early years. In 1976, gas injection began at a rate of 400 MMcf/D using gas from the nearby NIS gas dome. Recently, the field has been producing at approximately 35,000 BOPD.

Saidi[30] describes the many oil recovery mechanisms at work in this oil-wet reservoir as gravity drainage at constant IFT and reservoir pressure; oil swelling in the present gas-invaded zone because of the increase in reservoir pressure; oil swelling in the present oil zone through thermal convection/diffusion process; oil imbibition within the oil column; oil gravity drainage from the partially saturated blocks within the gas-invaded zone; and oil gravity drainage from the fully oil-saturated block in the oil zone and the blocks between that and the present GOC.

The flow behavior developed from the history match is that the oil-drainage performance follows that of stacks of discontinuous blocks, supporting practically no vertical capillary continuity between the matrix blocks (see Fig. 12.8).

By going to immiscible gas injection, oil recovery is increased by about 500 × 106 bbl by returning to the original reservoir pressure and could be increased by another 100 × 106 bbl if the reservoir pressure is increased an additional 100 psi because of the reduction in gas/oil IFT with increasing reservoir pressure.

Overall, the application of immiscible gas injection to the Haft Kel field has been considered a success. The estimated displacement efficiency by water was 17%, whereas that estimated for immiscible gas displacement was 32%.

Swanson River Field (Cook Inlet, Alaska)[24] A very different style of successful immiscible gas/oil displacement project is that applied to the Swanson River field’s Hemlock reservoir. Figs. 12.21 and 12.22 show an areal view of this reservoir and a type log through the Hemlock formation, respectively. This field is a north/south-trending anticlinal flexure about 6 miles long by 1 to 3 miles wide with as much as 600 ft of closure. The Hemlock formation consists of interbedded fine- to coarse-grained sandstone, conglomerate, siltstone, and coal, with numerous thin, impermeable, calcareous stringers of somewhat limited areal extent. Field experience has confirmed that these calcareous stringers are effective barriers to the vertical migration of fluids in the vicinity of producing wells. There are 10 Hemlock intervals, and the H1 through H5, H8, and H10 intervals have been engineered and managed separately (see Fig. 12.22).

The OOIP of this field is 435 million STB. This oil was very undersaturated at discovery in 1957, with an original reservoir pressure of 5,580 psi but with a bubblepoint pressure of only 1,350 psi. The oil is 37°API gravity, had an initial FVF of 1.21, and had a viscosity of 1.1 cp.

From these reservoir characteristics, it was clear that at the start of production the reservoir pressure would fall rapidly and that some form of pressure maintenance needed to be applied quickly. Fortunately for this very undersaturated oil field, a Tcf-sized dry gas field was discovered nearby for which there was no immediate or large outlets for its gas. A gas rental contract was signed between the two fields owners, and starting in 1966, 400 MMcf/D of gas was delivered for injection into crestal wells in the Swanson River Hemlock reservoir.

Laboratory studies indicated that a number of mechanisms were at work when methane gas displaced the low-bubblepoint-pressure Swanson River oil. First, if operating at a 5,000-psi pressure, the oil would swell from an FVF of 1.21 to an FVF of 1.80 on contact with the injected gas. Second, free gas would become saturated with intermediate hydrocarbon components from the fairly light (37°API) reservoir oil. Laboratory experiments show that breakthrough recovery efficiency exceeded 60%.

Although the Swanson River field has operated at high GORs (> 10 Mscf/STB) after the first several years of production, reservoir performance has been excellent, with 38% of the OOIP recovered through the first 10 years of gas injection. To show that the oil vaporization mechanism is also at work in this reservoir, during these first 10 years of production after the start of gas injection, the produced oil gravity increased from 37 to 40°API.

The Swanson River field is currently being blown down to recover as much remaining gas as possible. From production statistics through 2002 and an updated OOIP of 390 × 106 bbl, ultimate recovery from application of immiscible gas/oil displacement to this field is nearly 58% of OOIP. This has been achieved primarily through oil swelling and oil stripping, not vertical gas/oil gravity drainage.

Kuparuk River Field, Alaskan North Slope.[14][15] Another type of immiscible gas/oil project is that applied to some areas of the Kuparuk River oil field. The Kuparuk River field contains approximately 5.9 × 109 bbl of 24°API oil in two producing zones (zones A and C). The oil was slightly undersaturated at discovery (3,300-psi original reservoir pressure, with oil 300 to 500 psi undersaturated). The shallower C sand is more permeable than the A sand, but the A sand contains approximately 62% of the OOIP. The reservoir covers > 200 sq miles.

The field was discovered in the late 1960s and went on production in 1981. The maximum rate has exceeded 300,000 BOPD, producing from 42 drillsites with wells on 160-acre well spacing. The challenges in this field were that there was no initial gas cap on the reservoir into which to reinject the produced gas; the reservoir dip is very slight, and the reservoir intervals were not very thick, so gas/oil gravity drainage could not efficiently occur; the gas could not be flared; and there was no off-site location to which the residue gas could be sent. For these reasons, the gas had to be reinjected into one of the reservoir intervals for storage and to maintain the oil production rates. The challenge for the field engineers was how to reinject the residue gas in the most efficient way, given the various constraints.

After a few years of gas injection into the A sand in one area of the field in which the offsetting producers soon experienced rapidly increasing GORs, the engineers decided to go to an immiscible WAG injection process.[14] WAG could be applied at minimal cost and with few operational complications because most of the field was already being waterflooded. Calculations indicated that there were three beneficial effects to this approach. First, there would be some swelling of the oil because of the free gas. Second, the residual oil saturation could potentially be reduced by the presence of trapped gas. Third, WAG injection was used to reduce the high mobility of the gas by means of three-phase relative permeability effects (simultaneously having mobile gas, oil, and water in the pore system); also, a tapered WAG scheme helped in this regard. The overall effect was that oil recovery could potentially be increased by 1 to 3% of OOIP without resulting in significant gas cycling problems. To date, immiscible WAG injection has worked as expected and has been a satisfactory solution to the Kuparuk River gas disposal problem.


Gas Sources for Immiscible Gas Injection Projects

The first consideration in any immiscible gas injection project is where to get the volume of gas necessary for the project. Historically, produced and processed residue gas from that particular oil field has been used. This is the most satisfactory solution if the economics of the additional oil recovery justify deferred gas sales. In some locations, this is not an issue because there is no current market outlet for the produced gas. Generally, reinjection of this local gas supply is sufficient to nearly maintain the current reservoir pressure.

The next best alternative in some locations is to develop deeper gas horizons as a gas supply. This is particularly true in the Middle East where massive volumes of gas are often found in the deeper formations, such as the Khuff, underlying some major oil reservoirs. The third alternative is to look to nearby fields for a source of gas; this alternative has been used at the Swanson River field in Alaska[22] and the Heft Kel and other Iranian fields.[30]

If there is no source for lean hydrocarbon gas, then possibly flue gas (88% N2, 12% CO2) or nitrogen might be used. Both of these options require that a plant be constructed at the field to generate these gases in the large volumes needed. This approach requires the economics of the project to justify the large capital expenditure for such facilities, additional operating costs, and future impact of the produced gas becoming increasingly contaminated with nonhydrocarbon components with time.

Sometimes, the solution to the gas supply problem is to use a combination of sources to provide the required volume of gas. At the Hawkins field, both residue hydrocarbon gas and flue gas initially were injected. More recently, nitrogen from a cryogenic nitrogen rejection plant has been injected.[4][25][57]

Use of Horizontal Wells in Immiscible Gas Injection Projects

Horizontal wells have not been discussed in this chapter because the technology to drill horizontally has developed rapidly since the mid-1980s and such wells have not had widespread use in historic immiscible gas injection projects. Horizontal wells are particularly suited for use in gravity drainage immiscible gas injection projects because they maximize the distance between the producing wells’ perforations and the overlying gas cap and because such wells minimize the pressure gradients in the near-wellbore region (the cause of near-wellbore gas coning). At the Prudhoe Bay field, a large number of horizontal wells have been drilled for a variety of purposes, including the two mentioned above. For these reasons, in future immiscible gas gravity drainage projects, it is logical to consider using horizontal production wells. The chapter on fluid flow and well analysis in this section of the Handbook shows how to predict and to interpret the performance of horizontal wells.

Operating Procedures for Thin Oil Columns

One consideration in immiscible gas gravity drainage projects is the challenge of maximizing oil recovery from thin oil columns. The thin oil column may be what is found initially,[29] or in most cases, as the gas cap expands, it is all that is left to produce late in the life of the project. The field engineers have to monitor individual well performance and overall reservoir performance closely to optimize production under these circumstances. Obviously, if new wells are drilled, they should be either carefully targeted horizontal wells or wells with very limited perforated intervals.

A related consideration is when there is a thin oil column sandwiched between the expanding gas cap and the underlying aquifer.[29] In this situation, well perforations must be chosen to maximize recovery and to minimize the production of both gas and water. Although coning simulations with numerical reservoir simulators will provide insights into the best approach for a particular reservoir situation, actual field experience is necessary to optimize the operations.

Summary and Conclusions

In this chapter, the technical aspects of immiscible gas/oil displacement have been described, and several field case studies have been presented. The conclusions concerning immiscible gas/oil displacement are listed below:

  1. Immiscible gas/oil viscous displacement is an inefficient oil displacement process because gas is a highly mobile fluid.
  2. Gas-oil capillary pressure data indicate that in many situations the residual oil saturation to gas displacement is significantly lower than the residual oil saturation to water displacement.
  3. The immiscible gas/oil process becomes efficient and desirable when gravity works to keep the very-low-density gas on top of the higher-density oil and/or there is significant mass transfer of components from the oil to the gas.
  4. The most successful immiscible gas/oil injection projects are the vertical gravity drainage projects in which gas is injected into the crestal primary or secondary gas cap, with the oil wells producing from as far downdip as possible to maximize this distance from the gas cap both vertically and laterally. To maximize the efficiency of this approach, the overall oil production rate has to be restricted to the critical displacement rate.
  5. One gas/oil compositional mass-transfer effect is oil swelling. If an oil field contains a very undersaturated oil, then oil swelling by contact with the injected gas can be a very significant effect. However, if a reservoir has an original gas cap, the oil swelling effect is minimal because the oil is already fully saturated or nearly saturated with gas.
  6. The other gas/oil compositional mass-transfer effect is stripping or vaporization of intermediate hydrocarbon components from the oil by the lean injected gas. The importance of this effect increases as the producing GOR rises. Toward the end of the life of an immiscible gas injection project, the stripping effect can contribute many of the liquid hydrocarbons produced in the surface facilities and associated gas plants. This effect occurs with all types of oils but is more significant for lighter, or higher API gravity, oils.
  7. A few immiscible gas injection field projects have been undertaken that are not vertical gas/oil gravity drainage projects but in which compositional effects have led to project success. An excellent example of this approach is the Swanson River field in Alaska.
  8. Gas coning into producing wellbores’ perforated intervals occurs with thin oil columns or as the gas/oil interface moves downward. Horizontal wells are a method of further reducing the height of the remaining oil column by lowering pressure drawdown and thus minimizing the effects of gas coning.
  9. Numerical reservoir simulators are the best tool to evaluate all the technical aspects of an immiscible gas injection project, either historical performance and/or projections of future performance. Simple mathematical techniques have been developed to analyze some types of immiscible gas/oil displacements.


A = cross-sectional area, ft2
Fg = fractional gas flow, fraction
h = thickness of oil zone normal to direction of dip, ft
hv = vertical thickness, ft
k = permeability, darcies
kh = permeability in horizongal direction, darcies
kro = relative permeability to oil, fraction
krg = relative permeability to gas, fraction
kv = permeability in vertical direction, darcies
L = distance along the bedding plane, ft
M = mobility ratio, krgμo/kroμg
Ng = oil production, fraction
Pc = gas/oil capillary pressure, psi
qT = total flow rate through area A , ft3/D
Scw = connate water saturation, fraction PV
Sg = gas saturation, fraction PV
So = oil saturation, fraction PV
Sorg* = irreducible oil saturation in presence of gas, fraction PV
Swi = initial water saturation, fraction PV
t = time, days
tD = dimensionless time
tv = time for vertical drainage, days
tu = time for drainage along bedding plane, days
u = flow per unit area, ft3/ft2-D
vs = downdip free fall of oil, ft/D
V = volume of gas injected, res bbl
z = vertical distance, ft
α = angle of dip (positive downdip), degrees
μg = gas viscosity, cp
μo = oil viscosity, cp
ρg = gas density, lbm/ft3
ρo = oil density, lbm/ft3
ϕ = porosity, fraction


  1. 1.0 1.1 1.2 1.3 1.4 1.5 1.6 Muskat, M. 1949. Physical Principles of Oil Production, 470-502. New York City: McGraw-Hill Book Co. Inc.
  2. Cotter, W.H. 1962. Twenty-Three Years of Gas Injection into A Highly Undersaturated Crude Reservoir. J Pet Technol 14 (4): 361-365. SPE-82-PA.
  3. Shehabi, J.A.N. 1979. Effective Displacement of Oil by Gas Injection in a Preferentially Oil-Wet, Low-Dip Reservoir. J Pet Technol 31 (12): 1605-1613. SPE-7652-PA.
  4. 4.0 4.1 4.2 4.3 Kuehm, H.G. 1977. Hawkins Inert Gas Plant: Design and Early Operation. Presented at the SPE Annual Fall Technical Conference and Exhibition, Denver, Colorado, 9-12 October 1977. SPE-6793-MS.
  5. Katz, D.L., Monroe, R.R., and Tanner, R.P. 1943. Surface Tension of Crude Oils Containing Dissolved Gases. Pet. Tech. (September): 1.
  6. Hough, E.W. 1966. Correlation of Interfacial Tension of Hydrocarbons. SPE J. 6 (4): 345-349. SPE-1565-PA.
  7. Katz, D.L. and Firoozabadi, A. 1978. Predicting Phase Behavior of Condensate/Crude-Oil Systems Using Methane Interaction Coefficients. J Pet Technol 30 (11): 1649–1655. SPE-6721-PA.
  8. 8.0 8.1 Firoozabadi, A. and Katz, D.L. 1988. Surface Tension of Reservoir CrudeOil/Gas Systems Recognizing the Asphalt in the Heavy Fraction. SPE Res Eng 3 (1): 265-272. SPE-13826-PA.
  9. Broseta, D. and Ragil, K. 1995. Parachors in Terms of Critical Temperature, Critical Pressure and Acentric Factor. Presented at the SPE Annual Technical Conference and Exhibition, Dallas, Texas, 22-25 October 1995. SPE-30784-MS.
  10. 10.0 10.1 10.2 10.3 10.4 Hagoort, J. 1980. Oil Recovery by Gravity Drainage. SPE J. 20 (3): 139–150. SPE-7424-PA.
  11. Johnston, E.F., Bossler, C.P., and Naumann, V.O. 1959. Calculation of Relative Permeability From Displacement Experiments. Trans ., AIME (1959) 216, 61.
  12. Miscible Processes, Vol. 8, 111-114. 1957. Richardson, Texas: Reprint Series, SPE.
  13. 13.0 13.1 13.2 13.3 13.4 Simon, A.D. and Petersen, E.J. 1997. Reservoir Management of the Prudhoe Bay Field. Presented at the SPE Annual Technical Conference and Exhibition, San Antonio, Texas, 5-8 October 1997. SPE-38847-MS.
  14. 14.0 14.1 14.2 14.3 Ma, T.D. and Youngren, G.K. 1994. Performance of Immiscible Water-Alternating-Gas (WAG) Injection at Kuparuk River Unit, North Slope, Alaska. Presented at the SPE Annual Technical Conference and Exhibition, New Orleans, 25–28 September. SPE 28602.
  15. 15.0 15.1 Stoisits, R.F., Scherer, P.W., and Schmidt, S.E. 1994. Gas Optimization at the Kuparuk River Field. Presented at the SPE Annual Technical Conference and Exhibition, New Orleans, Louisiana, 25-28 September 1994. SPE-28467-MS.
  16. Aziz, K. and Settari, A. 1979. Petroleum Reservoir Simulation, 30-38. London: Applied Science Publishers Ltd.
  17. Buckley, S.E. and Leverett, M.C. 1942. Mechanism of Fluid Displacement in Sands. Trans., AIME 146: 107.
  18. 18.0 18.1 Welge, H.J. 1952. A Simplified Method for Computing Oil Recovery by Gas or Water Drive. Trans., AIME 195: 91.
  19. 19.0 19.1 Kern, L.R. 1952. Displacement Mechanism in Multi-well Systems. Trans., AIME 195.
  20. Craft, B.C. and Hawkins, M.F. 1959. Applied Petroleum Reservoir Engineering, 370. Englewood Cliffs, NJ: Prentice-Hall Inc.
  21. Jr., E.L.A. 1953. Mile Six Pool - An Evaluation of Recovery Efficiency. J Pet Technol 5 (11): 279-286.
  22. 22.0 22.1 22.2 Miscible Processes, Vol. 8, 197. 1965. Richardson, Texas: Reprint Series, SPE.
  23. 23.0 23.1 Miscible Processes, Vol. 8, 205. 1965. Richardson, Texas: Reprint Series, SPE
  24. 24.0 24.1 24.2 24.3 Young, R.E., Fairfield, W.H., and Dykstra, H. 1977. Performance of a High-Pressure-Gas Injection Project, Swanson River Field, Alaska. J Pet Technol 29 (2): 99-104. SPE-5874-PA.
  25. 25.0 25.1 25.2 25.3 25.4 25.5 Carlson, L.O. 1988. Performance of Hawkins Field Unit Under Gas Drive-Pressure Maintenance Operations and Development of an Enhanced Oil Recovery Project. Presented at the SPE Enhanced Oil Recovery Symposium, Tulsa, Oklahoma, 16–21 April. SPE-17324-MS.
  26. 26.0 26.1 26.2 26.3 26.4 26.5 Christianson, S.H. 1977. Performance and Unitization of the Empire Abo Pool. Presented at the SPE Permian Basin Oil and Gas Recovery Conference, Midland, Texas, 10-11 March 1977. SPE-6384-MS.
  27. 27.0 27.1 Metz, W.P. and Elliot, R.A. 1991. Gas Handling Expansion Facilities at Prudhoe Bay, Alaska. Presented at the International Arctic Technology Conference, Anchorage, Alaska, 29-31 May 1991. SPE-22113-MS.
  28. 28.0 28.1 28.2 Richardson, J.G., Harris, D.G., Rossen, R.H. et al. 1978. The Effect of Small, Discontinuous Shales on Oil Recovery. J Pet Technol 30 (11): 1531–1537. SPE-6700-PA.
  29. 29.0 29.1 29.2 Selamat, S., Goh, S.T., and Lee, K.S. 1999. Seligi Depletion Management. Presented at the SPE Asia Pacific Improved Oil Recovery Conference, Kuala Lumpur, Malaysia, 25-26 October 1999. SPE-57251-MS.
  30. 30.0 30.1 30.2 30.3 30.4 30.5 Saidi, A.M. 1996. Twenty Years of Gas Injection History into Well-Fractured Haft Kel Field (Iran). Presented at the International Petroleum Conference and Exhibition of Mexico, Villahermosa, Mexico, 5–7 March. SPE 35309.
  31. O’Neill, N. 1988. Fahud Field Review: A Switch from Water to Gas Injection. J Pet Technol 40 (5): 609–618. SPE-15691-PA.
  32. Horie, T., Firoozabadi, A., and Ishimoto, K. 1990. Laboratory Studies of Capillary Interaction in Fracture/Matrix Systems. SPE Res Eng 5 (3): 353–360. SPE-18282-PA.
  33. Firoozabadi, A. and Hauge, J. 1990. Capillary Pressure in Fractured Porous Media (includes associated papers 21892 and 22212). J Pet Technol 42 (6): 784-791. SPE-18747-PA.
  34. Firoozabadi, A., Ishimoto, K., and Dindoruk, B. 1994. Reinfiltration in Fractured Porous Media: Part 2 - Two Dimensional Model. SPE Advanced Technology Series 2 (2): 45-51. SPE-21798-PA.
  35. Dindoruk, B. and Firoozabadi, A. 1997. Crossflow in Fractured/Layered Media Incorporating Gravity, Viscous, and Phase Behavior Effects. SPE J. 2 (2): 120-135. SPE-35457-PA.
  36. Dindoruk, B. and Firoozabadi, A. 1996. Crossflow in Fractured/Layered Media Incorporating Gravity, Viscous, and Phase Behavior Effects: Part II - Features in Fractured Media. Presented at the SPE/DOE Improved Oil Recovery Symposium, Tulsa, Oklahoma, 21-24 April 1996. SPE-35458-MS.
  37. Ghorayeb, K. and Firoozabadi, A. 2000. Numerical Study of Natural Convection and Diffusion in Fractured Porous Media. SPE J. 5 (1): 12-20. SPE-51347-PA.
  38. 38.0 38.1 38.2 38.3 Addington, D.V. 1981. An Approach to Gas-Coning Correlations for a Large Grid Cell Reservoir Simulator. J Pet Technol 33 (11): 2267-2274. SPE-8332-PA.
  39. 39.0 39.1 Killough, J.E. and Foster Jr., H.P. 1979. Reservoir Simulation of the Empire Abo Field: The Use of Pseudos in a Multilayered System. SPE Journal 19 (5): 279-288. SPE-7418-PA.
  40. Shreve, D.R. and Welch, L.W. Jr. 1956. Gas Drive and Gravity Drainage Analysis for Pressure Maintenance Operations. Trans., AIME, 207: 136-143.
  41. Jacoby, R.H. and Berry, V.J. Jr. 1958. A Method for Predicting Pressure Maintenance Performance for Reservoirs Producing Volatile Crude Oil. Trans., AIME, 213: 59-64.
  42. Attra, H.D. 1961. Nonequilibrium Gas Displacement Calculations. SPE J. 1 (3): 130-136.
  43. Stiles, W.E. 1949. Use of Permeability Distribution in Water Flood Calculations. J Pet Technol 1 (1): 9-13. SPE-949009-G.
  44. 44.0 44.1 44.2 Dyes, A.B., Caudle, B.H., and Erickson, R.A. 1954. Oil Production After Breakthrough as Influenced by Mobility Ratio. J Pet Technol 6 (4): 27-32. SPE-309-G.
  45. Johns, R.T. 1992. Analytical Theory of Multicomponent Gas Drives with Two-Phase Mass Transfer. PhD dissertation, Stanford U., Palo Alto, California.
  46. Jessen, K., Wang, Y., Ermakov, P. et al. 1999. Fast, Approximate Solutions for 1D Multicomponent Gas Injection Problems. Presented at the SPE Annual Technical Conference and Exhibition, Houston, Texas, 3–6 October. SPE-56608-MS.
  47. Falls, A.H. and Schulte, W.M. 1992. Features of Three-Component, Three-Phase Displacement in Porous Media. SPE Res Eng 7 (4): 426–432. SPE-19678-PA.
  48. Lake, L.W. 1989. Enhanced Oil Recovery, first edition. Englewood Cliffs, NJ: Prentice-Hall Inc.
  49. Juanes, R. and Patzek, T.W. 2003. Relative Permeabilities in Co-Current Three-Phase Displacements with Gravity. Presented at the SPE Western Regional/AAPG Pacific Section Joint Meeting, Long Beach, California, 19–24 May. SPE-82445-MS.
  50. Azevedo, A.V. et al. 1997. Nonuniqueness of Riemann Problems. Zeitschrift fûr angewandte Mathematik und Physik 47 (6): 977.
  51. Marchesin, D., Plohr, B., and Schecter, S. 1997. An Organizing Center for Wave Bifurcation in Multiphase Flow Models. SIAM J. Appl. Math. 57 (5): 1189.
  52. Richardson, J.G. and Blackwell, R.J. 1971. Use of Simple Mathematical Models for Predicting Reservoir Behavior. J Pet Technol 23 (9): 1145-1154. SPE-2928-PA.
  53. Cardwell, W.T. Jr. and Parsons, R.L. 1949. Gravity Drainage Theory. Trans., AIME 179: 199.
  54. 54.0 54.1 Erickson, J.W. and Sneider, R.M. 1997. Structural and Hydrocarbon Histories of The Ivishak (Sadlerochit) Reservoir, Prudhoe Bay Field. SPE Res Eng 12 (1): 18-22. SPE-28574-PA.
  55. Corey, A.T. et al. 1956. Three-Phase Relative Permeability. Trans., AIME 207: 349.
  56. 56.0 56.1 56.2 King, R.L., Stiles, J.H. Jr., and Waggoner, J.M. 1970. A Reservoir Study of the Hawkins Woodbine Field. Presented at the 1970 SPE Annual Meeting, Houston, 4–7 October 1970. SPE-2972-MS.
  57. 57.0 57.1 57.2 57.3 Langenberg, M.A., Henry, D.M., and Chlebana, M.R. 1995. Performance and Expansion Plans for the Double-Displacement Process in the Hawkins Field Unit. SPE Res Eng 10 (4): 301-308. SPE-28603-PA.
  58. 58.0 58.1 Haldorsen, H.H., Rego, C.A., Change, D.M. et al. 1985. An Evaluation of the Prudhoe Bay Gravity Drainage Mechanism by Complementary Techniques. Presented at the SPE California Regional Meeting, Bakersfield, California, 27-29 March 1985. SPE-13651-MS.
  59. Stramp, R.L. 1980. The Use of Horizontal Drainholes in the Empire Abo Unit. Presented at the SPE Annual Technical Conference and Exhibition, Dallas, Texas, 21-24 September 1980. SPE-9221-MS.

SI Metric Conversion Factors

acre × 4.046 863 E + 03 = m2
bbl × 1.589 873 E − 01 = m3
cp × 1.0* E − 03 = Pa•s
dyne × 1.0* E − 05 = N
ft × 3.048* E − 01 = m
ft2 × 9.290 304* E − 02 = m2
ft3 × 2.831 685 E − 02 = m3
°F (°F − 32)/1.8 = °C
in. × 2.54* E + 01 = cm
lbm × 4.535 924 E − 01 = kg
psi × 6.894 757 E + 00 = kPa


Conversion factor is exact.