A Literature Analysis of the WAG Injectivity Abnormalities in the CO2 Process

SummaryThis paper summarizes the hypotheses and theories relating to thecauses and expectations of injectivity behavior in various CO2 andgasflooded reservoirs. The intent of the paper is to:

• Provide a concise compendium to the current understandingof the water-alternating-gas (WAG) mechanism and predictability.

• Provide a comprehensive single-source review of the causes andconditions of injectivity abnormalities in CO2/gasflood EOR projects.

• Aid in formulating the direction of research.• Help operators develop operational and design strategies for

current and future projects, as well as to input parameters for sim-ulating current and future projects.

BackgroundMoritis1 identified 94 gas improved oil recovery (IOR) projects inthe U.S. Of these, 74 are still active and 64 are CO2 miscible proj-ects. New CO2 projects start each year. Five new U.S. miscibleCO2 projects were being planned as of January 2000. Brock andBryan2 presented a summary of CO2 IOR projects and reviewedthe performance of 30 full-scale field projects and field pilots up to1987. In 1992, there were 45 active CO2 projects in the U.S.3

Because of the low oil prices following the 1985–86 price collapse,the initial industry outlook was pessimistic; however, by 1992,most projects had been shown to be technically and economicallysuccessful. In a number of projects, the production performancehas been better than anticipated.3

At the beginning of 2000, and based on 1999 production fig-ures, the U.S. production from gas-injected IOR was estimated at328,759 B/D, or approximately 5% of the total oil production inthe U.S. Oil production from CO2 activity alone contributed189,493 B/D, which is an increase of 5.8% over 1998 productionattributable to CO2 production and represents 3% of the 1999 U.S.oil production.1 This increase occurred despite the 1998–99 pricecollapse, which was deeper than the mid-1980s collapse. ThePermian Basin of west Texas and southeastern New Mexicoremains a very active area for CO2 projects. However, CO2 IORfield or pilot projects also exist in seven other states: California,Colorado, Kansas, Michigan, Mississippi, Oklahoma, and Utah.

Analysis of individual projects4 and reported problems are not pre-sented here. A review of 23 projects regarding injectivity is includedin a U.S. Dept. of Energy annual report.5 A number of reviews haveappeared in the literature.1–3,4,6 During the spring of even years, the Oil& Gas Journal usually publishes a survey of active IOR projects.

Industry’s Initial Concerns. There are two basic IOR techniquesin gasflooding a reservoir—continuous gas injection and the WAGinjection scheme. Industry initially had a number of concerns aboutCO2 injection, especially during the WAG process, in terms of con-trolling the higher-mobility gas: water blocking, corrosion, pro-duction concerns, oil recovery, and loss of injectivity. Carefulplanning and design along with good management practices haveallayed most concerns, except for loss of injectivity.

Lower injection rates of CO2 slugs and water slugs have been aconcern because CO2 field tests were conducted in the early

1970s.7 Currently, the problem is still a concern in the managementof a WAG process.4 This concern is the primary focus of this paper.

Injectivity Losses. There are two separate but related questionsregarding this perplexing issue.

• What causes the unexpectedly low injectivity during gas injection?• What is the reason for the apparent reduction in water injec-

tivity during brine injection after gas injection?Injectivity is a key variable for determining the viability of a

CO2 project. Potential loss of injectivity and corresponding loss ofreservoir pressure (and possibly loss of miscibility resulting inlower oil recovery) have potentially major impacts on the econom-ics of a gas-injection process. Many of the projects evaluated byHadlow3 showed higher CO2 (gas) injectivity than that obtained inprewaterflood water injection. However, substantial loss in waterinjectivity after CO2 or gas injection also has been seen. On theaverage, an approximately 20% loss of water injectivity can beexpected in the WAG process3; attempts to mitigate this includedecreasing the WAG ratio to decrease the mobility control, increas-ing the injection pressure, and adding additional injection wells.

Optimization of operations can improve the economics of exist-ing CO2

8 and other enhanced oil recovery (EOR) projects signifi-cantly. Three major management parameters that effect theeconomics of a CO2 or gasflood are:8

• The CO2 and water half-cycle slug sizes.• The gas/water ratio profile.• The ultimate injected CO2 slug size.

Overview of WAG Injection ProcessWAG Process Description. The WAG scheme is a combination oftwo traditional techniques of improved hydrocarbon recovery:waterflooding and gas injection.9 The first field application of WAGis attributed to the North Pembina field in Alberta, Canada, byMobil in 1957,6 where no injectivity abnormalities were reported.Conventional gas or waterfloods usually leave at least 50% of theoil as residual.10 Laboratory models conducted early in the historyof flooding showed that simultaneous water/gas injection hadsweep efficiency as high as 90%, compared to 60%10 for gas alone.However, completion costs, complexity in operations, and gravitysegregation from simultaneous water/gas injection indicated that itwas an impractical method for minimizing mobility. Therefore, aCO2 slug followed by WAG has been adopted.

The planned WAG ratios of 0.5:4 in frequencies of 0.1 to 2% PVslugs of each fluid11 will cause water-saturation increases during thewater cycles and decreasing water saturations during the gas half ofthe WAG cycle. The displacement mechanism caused by the WAGprocess occurs in a three-phase regime; the cyclic nature of theprocess creates a combination of imbibition and drainage.9

Optimum conditions of oil displacement by WAG processes areachieved if the gas and water have equal velocity in the reservoir.The optimum WAG design is different for each reservoir and needsto be determined for a specific reservoir and possibly fine-tuned forpatterns within the reservoir.12

There are a number of different WAG schemes to optimizerecovery. Unocal patented a process called Hybrid-WAG, in whicha large fraction of the pore volume of CO2 to be injected is inject-ed followed by the remaining fraction divided into 1:1 WAGratios.11 Shell empirically evolved a similar process calledDUWAG (Denver Unit WAG) by comparing continuous injectionand WAG processes.

Important technical factors affecting WAG performance thathave been identified are heterogeneity,8,9,12–17 wettability,9,12,18,19

October 2001 SPE Reservoir Evaluation & Engineering 375

A Literature Analysis of the WAG InjectivityAbnormalities in the CO2 Process

John D. Rogers,* SPE, and Reid B. Grigg, SPE, New Mexico Petroleum Recovery Research Center

*Now with the U.S. Dept. of Energy Natl. Energy Technology Laboratory (NETL).

Copyright © 2001 Society of Petroleum Engineers

This paper (SPE 73830) was revised for publication from paper SPE 59329, first presentedat the 2000 SPE/DOE Improved Oil Recovery Symposium, Tulsa, 3–5 April. Original man-uscript received for review 15 June 2000. Revised manuscript received 30 July 2001. Paperpeer approved 10 August 2001.

fluid properties,9,12,14,16,17 miscibility conditions,8,9,12 injectiontechniques,8,9,12 WAG parameters,8,9,12 physical dispersion,16,17 andflow geometry.8,12,15

Optimization of WAG. Oil recovery is enhanced if the gas andwater slugs are appropriate for a specific reservoir. Gorell,20 usinga 1D simplified model, assumed that the WAG could be analyzedas if it behaves like simultaneous solvent/water injection. Thevalidity of this assumption depends on the relative size of the injec-tion cycles. From Gorell’s study, equal WAG ratios are more effi-cient and are insensitive to assumed levels of trapping.

Wettability effects also have been shown to affect the optimumWAG ratio.21 Water-wet bead packs show an optimum WAG ratioof 0:1, or continuous gas injection. Contrarily, oil-wet packs sug-gest an optimum WAG ratio of equal, or 1:1, velocity ratios.Mixed-wet states indicate that maximum recovery is a strongerfunction of slug size in secondary CO2 recovery than in tertiaryflooding. In addition, water-wet laboratory models indicate thatgravity forces dominate, while in oil-wet tertiary floods, viscousfingering is a controlling factor.

Many operational changes have been implemented to mitigateinjectivity abnormalities and optimize the WAG process. In theWasson field22–31 of west Texas and many other CO2 projects, thecontinuous and WAG processes were used and compared. Thewater cycle of the WAG was especially affected by loss of injec-tivity. Oil response in the WAG area was slower, but the CO2 pro-duction was also lower compared to the continuous injection areas.The WAG created a considerable challenge to maintain injectionrates. The nine-spot patterns were converted to a line-drive patternin 1988 and appear to have spread the desired injection volumeamong more injectors; the desired rates were attainable in theWAG area without exceeding fracture pressures. The WAG areahad several factors that contributed to poorer IOR performance: 1)lower WAG injectivity, 2) out-of-zone injection losses, 3) struc-tural continuity, and 4) waterflood-induced fractures. The unit’spatterns were closely monitored, and WAG cycle lengths wereextended from every 6 months to once a year. Thus, the DUWAGwas suggested to inject 4 to 6 years of continuous CO2 injectionfollowed by 1:1 WAG.

The use of horizontal CO2 injection wells can increase injectionrates several-fold over the injection rates achievable with verticalwells in a five-spot pattern.32,33 This is an important considerationin low-permeability reservoirs, or where reduced injectivity devel-ops during WAG cycles. The South Cowden (San Andres) Unit isa U.S. Dept. of Energy (DOE) Class II oil program for shallowshelf carbonate reservoirs to demonstrate the technical and eco-nomic viability of using horizontal CO2 injection wells and cen-tralization of production/injection facilities to optimize CO2

project economics. Better sweep efficiencies, faster flooding rates,and/or lower injection pressures are possible with horizontalwells.33,34 Thus, the economics of EOR projects and conventionalimproved recovery methods may improve substantially with theuse of horizontal wells.

Simultaneous Gas/Water Injection. Injectivity improvementshave been seen in waterfloods in which CO2 has been present. Inparticular, improvements in waterflood injectivity were attributedin comparisons of two similar carbonate fields.35 In these cases, theinjectivity improvements were attributed to acid gases dissolved inthe produced water, with the field having the higher CO2 contentalso having higher injectivity. Injectivity deterioration occurred inhigh-rate, high-permeability areas. Ramsey and Small36 reportedimproved injectivity with carbonation of water in sandstones. Inaddition, carbonated waterfloods were suggested37 as viable IORmechanisms and have been shown to enhance conventional water-flooding. However, the use of carbonated water or “fizz floods” asIOR processes do not show significant economic impact com-pared to a full-scale miscible gasflood or WAG EOR operation,provided that miscibility can be obtained and that miscible flood-ing is economically viable.

The use of carbonated waterfloods has been suggested in lowpermeability, naturally fractured reservoirs, particularly the Austin

chalk.38–42 Additionally, the use of surfactants in carbonated waterimbibition significantly increased oil recovery in rock sampleswith mixed or oil-wet conditions.43 Injection of water with dis-solved gas and injection of water above the bubblepoint may resultin the most even distribution of gas throughout the reservoir whereoil is otherwise trapped. Improvements in the efficiency of water-flooding and tertiary CO2 flooding in heterogeneous reservoirs alsomay be achieved by the injection of water with dissolved CO2.44

Amoco conducted a small feasibility pilot study on carbonatedwaterflooding in the mid-1980s to determine potential use in thewest Texas fields.45 Bargas et al.46 published simulation resultsshowing that for the shallow, light-oil Salt Creek field in NatronaCounty, Wyoming, recovery increases under immiscible CO2

processes and significantly increases with the use of carbonatedchase water. When noncarbonated chase water is used, significantlyless oil is produced, though there is an oil increase during theimmiscible CO2 injection.

Humble Oil and Refining Co. first tried simultaneous water andenriched gas injection in 1963 in the Seeligson field, KlebergCounty, Texas, after injecting enriched gas since 1957 (this was amiscible flood).47 Low injection rates and high pressures wereexperienced under simultaneous injection. In 1964, alternate slugsof water and gas were injected in an attempt to increase rates anddecrease pressures. The first cycle saw increased rates; however,during the second cycle, the wells took little gas. High water satu-ration around the wellbore was the attributed cause. Other projectsusing immiscible WAG injection have shown that it can be aneffective tool in management of oil reservoirs, especially with highgas production.48

WAG With Hydrocarbon Gases. Use of hydrocarbon gases, par-ticularly liquefied petroleum gas (LPG), to develop miscibility andimprove oil recovery has been evaluated since the early stages ofIOR.49–51 In these processes, WAG was used to reduce the largemobility differences between solvent displacing oil and to improveultimate recovery. Four rich-gas, secondary pilot projects were eval-uated by Amoco49–51 in oil-wet west Texas and Canadian carbonatereservoirs. Loss of injectivity in the water injection after rich-gasinjection was observed in only one of the subject projects—the SanAndres formation. The hypothesis was that rich gas is trapped dur-ing the first cycle of water injection and causes a decrease in rela-tive permeability to water. Remedial action was unsuccessful. Theremedial actions tried were:

1. Wellbore washing for hydrocarbon cleanup such as xyleneand propane, CO2 following rich-gas injection to displace the richgas with more water-soluble CO2, or rich gas followed with leasecrude to reestablish pre-gas-injection saturations.

2. Acid treatments that may have fractured a well or that saw awell’s injectivity drop during the second cycle of water.

3. Operating changes, including reduction of the buffer slug ofgas, higher initial water-injection rates, and lower subsequentinjection rates following rich-gas injection.

Injectivity AbnormalitiesInjectivity Increases. A number of the CO2 floods have seenhigher gas injection relative to prewaterflood injection (see, forexample, North Ward Estes,52,53 Mabee,54,55 and Cedar CreekAnticline15,56). Also, some projects have had higher CO2 injectivityafter successive WAG cycles. Simulations indicate that CO2 injec-tivity is much higher in reservoirs with crossflow when account-ing for phase behavior and mixing.57 Enough CO2 solubility infollow-up brine injection has been reported during WAG cycles toraise unsaturated brine injectivity to three to five times the satu-rated brine injectivity.54 Increased brine injectivity during WAGcycles after the first slug of CO2 also has been attributed to thecombined effects of heterogeneity, crossflow, oil-viscosity reduc-tion, CO2 sweep, CO2 channeling, compressibility, and solubilityof CO2 in injected brine near the wellbore.55

Injectivity increases are not as great where vertical permeabilityis lower, pay section is thicker, or the injection well is stimulatedand production wells are not stimulated.55 The effective wellboreradius or skin and heterogeneity in the layering reduce the influ-

376 October 2001 SPE Reservoir Evaluation & Engineering

ence of the oil bank, resulting in higher injectivity.27 The effects oflow mobility in the tertiary oil bank and in the dispersive mixingzone near the CO2 displacement front are more significant for astimulated well because they pose a greater portion of the totalflow resistance when resistance is lower near the well. In addi-tion, the fronts are moving with a velocity that varies inverselywith the radius from the injector. The closer these banks are to theinjection wellbore, the more effect the banks have on the activityat the injector.

Injectivity Reduction. Injectivity reduction after CO2 injectionhas occurred frequently in west Texas,51,58 as well as in the Brentformations after hydrocarbon gas injection in the North Seaarea.9,12 The Levelland, Slaughter, and Wasson fields producingfrom the San Andres formation have all reported injectivity lossduring WAG process.26

Schneider and Owens51 studied 19 preserved cores from fouroil-wet carbonate reservoirs to evaluate injectivity in a west Texasrich gasflood. Before the rich-gas/water injection, water rates aver-aged 350 B/D; after gas injection, water-injection rates averaged100 B/D. The ratio of pre- to post-gas injection is similar in mag-nitude to the reduction observed in some relative permeabilitycoreflood tests. Efforts to improve injectivity in the field followingrich-gas injection were largely unsuccessful.50 There is no indica-tion from Schneider and Owens regarding the skin condition of thewellbore, other than well-test data indicating that the reducedinjectivity was not a wellbore or near-wellbore problem.

The Levelland Unit CO2 miscible pilot reported a 10% loss inCO2 injectivity and a 50% loss of water injection vs. the pre-gaswater injection.59 As a result of the loss of injection pressure,cycles were observed in the composition observation well.Mobility was lower after CO2 injection than before, indicating thatmobility control was good and also suggesting that reduced injec-tivity is an in-depth phenomenon rather than a near-wellbore con-dition such as skin or high gas saturation around the injector.

Factors Affecting InjectivityWettability. During a reservoir’s evolution, rock is laid down in awater environment and is initially water-wet. As hydrocarbonsmigrate through and/or accumulate in the rock’s pore structure, oilwill occupy the largest pores, while the smaller pores remain filledby water because of insufficient capillary pressure.60 Various com-pounds in the oil can chemically alter the surfaces of the pores.61

The precise taxonomy of wettability is still lacking.62 Buckley63

categorized crude oil/brine/solid interactions as polar interactions,surface precipitation, acid/base, and ion binding.64 Anderson65

defines the terminology of five different types of wettability60—fully water-wet, fully oil-wet, intermediate-wet, fractionally wet,and mixed-wet. Key factors affecting wettability of a system areaging, temperature, brine pH/composition, crude oil composition,and connate water saturation.62,64,66

Results from various labs concerning wettability are difficult toreconcile and understand.67 Variations in laboratory procedures andexperimental materials make generalized conclusions about wetta-bility almost impossible. Also, there have been concerns reported inthe literature regarding the cross-correlation of the two main meth-ods of determining the wettability of a rock (i.e., the Amott andUSBM methods).62 These two methods do not address pore-scalewettability-alteration issues,61 and only under certain conditions arethe two methods expected to be equal.68 Radke et al.69 suggest thatthe role of thin films in porous media is a very important issue inunderstanding wettability. These authors also remark that the poreshape is an important factor in determining the wetting-film thick-ness, and the pore shape and the physics of thin films are generallyneglected in the study of fluid flow in porous rocks.

Alterations in wettability are nonuniform with experimentalevidence, indicating that various components of the crude oil inter-act differently with various mineral substrates in the rock (e.g.,quartz, feldspar, and clays).66 The measurement of wettabilityalterations is difficult to determine, and contact-angle measure-ments do not address this issue adequately.61 The adhesion test issuggested as a useful measurement61,67,70 of wettability alterations.

However, the capillary force is difficult to control, and the forceapplied to the nonwetting phase will establish the measurement.

The atomic force microscope is suggested as a direct way ofmeasuring the critical capillary pressure for crude oils or the capil-lary pressure required to rupture brine films on mineral surfaces.61

Capillary pressures determine the flow in a porous medium and arerelated to the wettability of porous structures. Weak capillaryforces often operate in mixed-wet systems and are related to vari-ation in contact angles. High capillary pressures imply a water-wetsystem, whereas low values of critical capillary pressure imply thatlarge sections of the reservoir may be rendered mixed-wet becauseof brine-film instability.61

The wetting state is not the sole indicator of the type of mech-anism controlling recovery.18,19,26 Most studies were conducted toidentify the effects of trapping, water shielding or blocking, rela-tive permeability effects, and phase behavior or multiphase flow onthe miscible process. These phenomena are also important to theinjectivity problem.19

The optimum WAG ratio is influenced by the wetting state ofthe rock. Gravity forces dominate water-wet tertiary floods, whileoil-wet tertiary floods are controlled by viscous fingering.21 HighWAG ratios have a large effect on oil recovery in water-wet rock.19

High WAG ratios result in less oil recovery by extraction. TertiaryCO2 floods controlled by viscous fingering had a maximum recov-ery at WAG ratio of about 1:1. Floods dominated by gravity tongu-ing showed maximum recovery with the continuous CO2 slugprocess. The optimum WAG ratio in secondary floods was a func-tion of the total CO2 slug size.

The CO2 WAG process relative to continuous injection hashigher oil trapping owing to wettability.11 In the Brent formation ofthe North Sea, larger pores tend to be oil-wet with residing oil, andsmall pores tend to be water-wet.9 Injected gas preferentially entersthe high-permeability layers, resulting in a reduced water-injectionrate caused by the three-phase and compressibility effects.9

However, corefloods conducted by Potter58 on preservedfresh-state cores taken from the west Texas Levelland Unit, SanAndres formation, and the North Cowden Grayburg formationdid not change wettability readily when flooded with oil andCO2; any change realized was toward more water-wet character-istics. Contrarily, capillary tube visual cell (CTVC) studies71

(using surrogate solvent and refined oil) and core studies in oil-wet, intermediate-wet, and water-wet cores (using ethane as sol-vent, as well as three reservoir oils)72 show that misciblegasflooding does induce wettability alterations. The CTVC stud-ies show that miscible gasflood-induced wettability alterationsoccur, and that water-wet surfaces become strongly oil-wet whenin contact with swelling oil. The wettability changes are mani-fested in large changes in endpoint permeabilities, with relativelylower change in endpoint water permeabilities. In intermediate-wet and oil-wet systems, the in-situ wettability alterations causedby solvent flooding had a significant positive effect on miscibleflood performance. In some cases, the impact of miscible floodingwas the possible development or natural occurrence of mixed-wettability conditions.72

The distribution and flow of fluids in porous media are impactedsignificantly by the wetting properties of the pore walls.61

Nonwetting fluids occupying larger pores will have larger relativepermeabilities. Wettability is the most important cause of injectivitylosses.26 Treiber73 concluded that 84% of the carbonates that he stud-ied are not water-wet and that 90% of the west Texas/New Mexicocarbonates were at least moderately oil-wet. Thus, carbonates aremore probably oil-wet or of mixed wettability.26,74 Water wetness ischaracteristic of pure carbonate rock. Field-observed reduction ininjectivity could be related to wettability. Mixed wettability is sug-gested as a cause of low fluid mobility observed during the DenverUnit Wasson field CO2 pilot.26

Chemical Effects. Tang and Morrow64 suggest that reservoir wet-tability will change if the significant variables such as salinity andpH in the reservoir are changed. CO2 forms a weak carbonic acid inwater with a pH of 3.3 to 3.7;36 thus, changes in pH may affect wet-tability during CO2 flooding. The carbonic acid readily reverts to


CO2 with increasing temperature and decreasing pressure.28 Even atrelatively low partial pressures, pH is reduced considerably.36

Buckley67 has shown in controlled synthetic and fairly cleansandstone coreflood experiments that wettability can change withthe pH of the brine. A high pH alters synthetic cores toward morewater-wet conditions, while lower pHs have a tendency to altercores and surfaces toward less water-wet conditions; however, porecoatings may control wetting alteration in natural porous media.The subject work also states that low ionic strength NaCl brinesand asphaltic oil alter wetting to mixed-wet conditions.

There is considerable disagreement as to whether dissolution,precipitation, and particle invasion or migration occur during injec-tion of CO2 in WAG processes. One speculation is that inorganicmaterial dissolution occurs as the pressure declines, while the floodfront advances toward the producer. However, pre- and post-pilotcore studies26 and limited laboratory experimental studies25

showed negligible dolomite dissolution occurring in the WassonDenver Unit, or at least that this process had little effect on injec-tivity. Patel et al.,26 however, commented that the scope for a morecomprehensive study of this mechanism exists. Contrarily, obser-vations in other west Texas pilots and early work in the North Seaand Canadian sandstones suggested that CO2 floods could have asignificant effect on dissolution of the reservoir rock. Also, resultshave shown that oil hinders the rate of dissolution;75 thus, more oil-wet reservoirs may not have high dissolution effects.

Literature on sandstone and carbonate diagenesis emphasizesthe role of naturally occurring CO2 in leaching processes. In thesandstone reservoirs of Pembina Cardium, Alberta, Canada, CO2

corefloods initially showed a large drop in permeability, afterwhich permeability rose steadily but did not regain its initialvalue.76 Microscopic (X-ray defraction and scanning electronmicroscope) examination indicated that fines had been releasedand had migrated toward pore throats, reducing permeability. Thegradual rise in permeability noted in the experiments was attrib-uted to mineral alterations by dissolution of calcite and siderite.

Laboratory coreflooding experiments under reservoir condi-tions on North Sea core material showed that dissolution could bea serious problem during CO2 flooding.75 Unlike carbonate forma-tions that consist primarily of carbonates, sandstones contain smallamounts of carbonaceous material, primarily as cements, consoli-dating the sand grains and creating the pore structure. A relativelysmall change in the pore framework owing to dissolution couldsignificantly affect the total permeability.

Thin-section examinations of post-pilot core from the WassonDenver Unit did not show evidence of dolomite dissolution, thoughanhydrite dissolution is seen but not statistically significant.25

Carbonic acid is an effective agent in increasing the solubility ofdolomite; thus, the lack of increase in porosity was attributed to thepossibility that a substantial pore volume contains CO2 (not car-bonic acid), so any trapped water will remain trapped in the smallerpores. The trapped water will come to equilibrium with the CO2

and will form carbonic acid and dissolve the dolomite. This waterwill not be a mobile phase to any great extent during the CO2 flood;thus, no significant transport mechanism exists to remove the Ca�2

and Mg�2 and HCO3� ions. The brine postflood should form car-

bonic acid from trapped CO2. There is evidence that this occurs,though sufficient pore volumes were not available to significantlyincrease total porosity.25 However, continued cycling of CO2 water,such as that occurring in the WAG process, does not appear to haveoccurred in the pilot.

Anhydrite can be dissolved in brine undersaturated with CaSO4

in the preflood. The presence of NaCl and CaCl2 brine increasesthe solubility of anhydrite.58 During the SACROC pilots, evidenceobtained in falloff and pulse testing suggested that dissolution ofdolomite occurred, and investigators postulated that precipitationof gypsum close to the injector77 was caused by CO2 injection.Tests run on North Cowden cores saw significant anhydrite disso-lution owing to brine composition, but nothing was mentionedabout the effects of CO2 on dissolution of the dolomite cores. Useof MgCl2 and MgSO4 stabilized the water/rock reactions.58

The Levelland pilot59 indicated possible effects from rock dis-solution, as evidenced by a dramatic increase in bicarbonate con-

tent. The total dissolved solids concentration was substantially ele-vated in the water at both composition observation wells comparedto the injection water, indicating that carbon dioxide was dissolv-ing in the water and forming carbonic acid. Ion concentration waslower in the water from the injection wells than it was in the waterfrom the observation wells, which suggests that the injected waterwas not in equilibrium with the formation. In addition, the authorssuggest that the oil films of the intermediate oil-wet reservoirshielded the high salinity connate water from mixing with andbeing displaced by the lower-salinity waterflood water. The injec-tion of carbon dioxide could remove the oil film and expose con-nate water to the water cycles that follow.59

Entrapment. Entrapment has been suggested as a cause of injec-tivity losses. Mechanisms found to affect trapping in miscible dis-placements at the laboratory scale are solvent diffusion, oilswelling, water saturation, and solvent contact time. Someresearchers have found that the amount of oil bypassed is sensitiveto flow rate and core length,78–80 while others have not.20

Bypassing increases as the solvent/oil viscosity ratio decreases.Major bypassing mechanisms are capillary-induced bypassing, dis-persion, and macroscopic bypassing.19 Experimental observationsof flow-rate and core-length effects can give some indication of therelative importance of each type of bypassing.

• An increase in recovery with flow rate indicates that capillarypressure effects dominate, while a decrease indicates that disper-sive bypassing or fingering is dominant.

• Recovery independent of core length shows that either capil-larity or dispersive bypassing dominates.

• Viscous fingering and dispersive bypassing increase with oilviscosity, while capillarity bypassing is a much weaker function ofoil viscosity.

Dispersive bypassing results from a distribution of pore sizesand occurs in single-phase flow.19 The distribution of pore radiigives rise to a distribution of path lengths and velocities. Mixing isnot complete at the pore junctions with laminar velocity distribu-tion in the pores, thus resulting in a broader distribution of resi-dence time, especially at high flow rates and short contact time.

Capillary entrapment occurs when the oil saturation in a porousmedium becomes low and the oil-phase network loses its continu-ity. At this point, viscous and gravitational pressure gradientsbecome insufficient to mobilize the remaining oil, which is trappedagainst capillary barriers within the porous medium.81 This bypass-ing phenomenon occurs in tertiary displacement because the sol-vent must displace water to mobilize and recover oil. The Laplaceequation, applied to an oil drop in a constriction through whichwater flowed, accounts for most of the pressure drop caused byfrictional loss and wall effects.82 If viscous drag forces are largeenough, the drop is mobilized and induces the snap-off process.Capillary entry pressure is higher in small pores and is affected bythe wetting nature of the rock. In water-wet rock, solvent displaceswater from the largest pores first because their entry pressure islower. In mixed-wet rock, the solvent will enter the smallest oil-wet pores first. Thus, capillarity-induced bypassing may depend onrock wettability, but it can occur in both mixed and water-wet rock.As viscous forces in the solvent bank increase relative to capillaryforces, the capillary-induced bypassing will be reduced.19

Macroscopic entrapment or fingering results from macroscopic-scale heterogeneities coupled with the mobility contrast betweensolvent and oil. Trapped-gas saturation is one of the key parametersin determining injectivity and displacement efficiency in a miscibleCO2 WAG injection project.83 Trapped-gas saturation influenceswater injectivity and the amount of water diversion in the WAGprocess. There is extensive trapping of gas in the high-permeabilitylayers, which diverts water to lower-permeability layers.84 Gas trap-ping plays an important role in the mobilizing and displacement ofresidual to waterflooding oil. The degree of oil-saturation reductionand the amount of gas trapping depends on the initial gas saturationbefore waterflooding, as well as the rock wettability.

Evidence in the literature shows that trapping behavior and rel-ative permeability depend on the ratio of flow rate to interfacialtension (IFT).84 The Prudhoe Bay laboratory data indicate that


trapped-gas saturation is essentially independent of residual-oilsaturation.84 This is contrary to what was evidenced in earlier flowstudies by Schneider and Owens51 conducted in 19 preserved westTexas reservoir cores. Here, the trapped oil and gas acted to lowerthe maximum attainable water saturation and resulted in lowermobility of subsequent water injection. Laboratory data supportedthe field evidence of reduced water injectivity for the process.

Laboratory corefloods on native state cores from the SouthCowden CO2 flood showed that trapped-gas saturations in the mainreservoir from 20 to 25% pore volume (PV) could develop duringWAG cycles.83 The South Cowden study also saw reduced waterrelative permeability after CO2 injection and determined that theCO2 relative permeabilities were lower than oil relative permeabil-ities at comparable water saturations.

Factors governing the trapping of residual oil include85 watersaturation, wettability, reservoir heterogeneity, capillary forces,and dynamics of water/solvent injection processes. If the water sat-uration is reduced, part of this trapped oil may reconnect and makeit more accessible to solvent.86 In a mixed-wet system, the amountof trapping was found to be a function of solvent throughput.86 Inmixed-wet and oil-wet cores, the amount of retained oil was foundto be insignificant after prolonged water/solvent injection and wasattributed to the dendritic (dead-end pore) oil in cores recovered bydiffusion mechanisms.86

Trapping is significant in laboratory-scale corefloods but rapidlydecreases at field scale if larger water barriers (compared to water-film thickness for laboratory scale) exist. Diffusion of solventthrough water films and the resultant swelling of trapped oil (andpossibly the rupture of the water film) can significantly affect thetrapped oil saturation in coreflood experiments. The trapped oilsaturation in solvent processes, therefore, has to be considered afunction not only of the water saturation but also of the solventcontact time or the flooding state.87

Water-blocking measurements with refined oil overpredict theextent of water blocking for reservoir fluid displacements whenwettability alteration occurs in various types of rocks, includingBerea, Alberta sandstone, and west Texas carbonate.71 Waterblocking was found to be more severe for Berea sandstone than forreservoir materials, even strongly water-wet Alberta reservoir core.For high WAG injection ratios in Berea water-wet rock, shieldingdominates the displacement process, and the type of miscibleprocess doesn’t matter, whether it is first-contact or multiple-con-tact miscible.88 The trapping function and oil and solvent mobilityin Berea water-wet rock control water shielding. Water-blockingestimates for reservoirs should be based on measurements in reser-voir cores using reservoir fluids.88

Mobile water does not change the mass transfer process bywhich miscibility develops,85 but the wetting conditions do affectmiscible displacement. Displacement by the nonwetting phase isaffected by highly mobile wetting-phase saturation, whereas dis-placement by the wetting phase is not significantly affected.Injecting below the optimum WAG ratio produces a high concen-tration profile directly behind the oil bank and creates mobility orviscous instability, while injecting above this ratio improves theratio and tends to improve or stabilize the process but substantiallyreduces displacement efficiency owing to trapping and prolongedproduction. The optimum WAG ratio seems to be fairly insensitiveto any assumed level of trapping of the oil phase.85 The water/solventinjection phase creates little trapping of oil but substantial trappingof solvent.

Microscopic influences caused by heterogeneity in the porousmedia include increased mixing effects owing to the tortuosity of thepore structure, longitudinal transverse dispersion of solvent, and masstransfer of solvent.85,89 Macroscopic influences such as channelingand crossflow are caused by permeability and wettability effects. Insecondary CO2 floods, local mixing caused by high water saturationsreduces recovery only slightly because high-mobility CO2 mixes withand displaces the oil before injected water arrives and creates signif-icant dendritic and trapped saturation.89 Additionally, in tertiary CO2

floods, oil recovery is slowed and reduced by restricted local mixingbecause high water saturations cause significant dendritic and trappedfractions throughout the flood.

Relative Permeability. Relative permeability is an importantpetrophysical parameter, as well as a critical input parameter, inpredictive simulation of miscible floods. However, relative perme-ability is a lumping parameter and includes the effects of wettingcharacteristics, heterogeneity of reservoir fluids and rock (IFTs),and fluid saturations, as well as other micro- and macro-influences.The importance of accurate determination of this parameter hasbeen known since the beginnings of improved and enhanced recov-ery processes. Research programs to collect relative permeabilitiesand attempts to model these parameters are replete in the litera-ture. IFTs have been shown to have an important effect on relativepermeability curves.

Early programs done at room temperature concluded that the oiland gas saturations present act to lower the maximum attainablewater saturation, resulting in reduced water mobility during subse-quent periods of water injection.51 Data from laboratory tertiaryflooding studies at representative reservoir conditions are becomingavailable in the literature.90 These data include water/oil relative per-meability when water saturation is decreasing, residual oil saturationin a miscible flood, and residual CO2 saturations. These parametersinfluence predictions of oil recovery, CO2 production, and break-through times. Large differences in CO2 and oil relative permeabili-ties can generate large differences for predicted injectivity.90 CO2

relative permeabilities can be very small in representative west Texascarbonates—as much as 100 times smaller than the oil endpoint rela-tive permeabilities.19,54,83 Reduced CO2 permeability affects gas pro-duction and injectivity more than oil recovery. Prieditis andBrugman90 showed that normalized injectivity can be predicted if thesolvent and oil endpoints are not assumed to be equal.

Defining CO2 relative permeability as equal to the oil relativepermeability will not predict the above behavior. Roper et al.16,17

have shown through simulation that a sharp injectivity reduction atthe start of the brine cycle can be associated with relative perme-ability reduction near the well and then can gradually experiencean increasing injectivity trend throughout the rest of the cycle. Thereason is suggested to be caused by two-phase flow of gas andbrine initially near the well; as the cycle proceeds, the saturationsand the relative permeabilities change.

Laboratory floods attempting to emulate the South CowdenCO2 flood experienced appreciable water relative permeabilityreductions with values observed before CO2 injection.32

Corefloods conducted on native-state cores showed that trappedgas saturations of 20 to 25% PV could develop in South Cowdenreservoir rock during miscible CO2 WAG operations. Gas relativepermeability curves were then constructed to yield this magnitudeof gas trapping in the simulation. In addition, the data showed sig-nificant hysteresis effects in the water relative permeabilitybetween the drainage and imbibition curves. Irreducible water satu-rations after drainage cycles were 15 to 20% higher than the initialconnate water saturation.

Water hysteresis occurs after CO2 injection. In the San Andres,water hysteresis occurs at new and higher irreducible water satura-tions.90 The injected CO2 and oil bank develop a new minimumvalue of irreducible water saturation that does not go back to theoriginal connate water saturation. Oil-curve hysteresis studiesshowed that oil relative permeabilities measured during oilfloodfollowing a waterflood were larger than the oil relative permeabil-ities measured during the initial waterflood. These new oil relativepermeabilities could be several times larger than the original. Ifhysteresis effects are not recognized or are ignored, the aqueousand oil primary imbibition and drainage relative permeabilitycurves differ.16,17 Hysteresis is seen in the water-permeabilitycurves, but not the oil curves of Levelland corefloods.58

For immiscible WAG, oil relative permeabilities remain thesame whether trapped gas was present or not. They appeared to bea function of oil saturation only. Water relative permeabilities weresignificantly lower with trapped gas present, indicating theirdependence on both gas and water saturations.91 Gas relative per-meability was noted by Akin and Demiralto92 to decrease withincrease in flow rate for three-phase flow.

Correlations to predict three-phase relative permeability fromtwo-phase data assume that gas relative permeability is a function


of gas saturation and is not dependent on the liquid phase dis-placed. In addition, most relative permeability data is obtained atambient conditions. Studies conducted by BP Research indicatethat the assumption that gas relative permeability is a function ofgas saturation alone is not valid for the reservoir sandstones stud-ied.93 The study, conducted under ambient and reservoir condi-tions, measured the saturation-distribution histories by gammaattenuation saturation monitoring with JBN analysis of the core-floods. The study suggests that unsteady-state relative permeabili-ties obtained from displacement corefloods at ambient conditionsprovide similar, though not equal, data to that obtained at reservoirconditions. Dria et al.94 concluded, using dolomite cores andsteady-state procedures, that the relative permeability of eachphase is seen to depend on the saturation of that phase only.

Three-phase-flow effects can have important influences oninjectivity, even when CO2 is injected above its minimum misci-bility point. Some researchers have suggested that the magnitudeof the water relative permeability endpoint has only a small effecton WAG recovery.14 Gas relative permeabilities measured underthree-phase-flow conditions with CO2 are much lower than withN2. This could result in lower total mobility and lower injectivitythan would be predicted if nitrogen relative permeability data wereused to calibrate the simulator.94

Roper et al.16,17 studied the sensitivity of relative permeabilityeffects and residual phase saturations on CO2 injectivity with ana-lytical and numerical compositional models. Their study showedthat the aqueous phase endpoint is important both before and afterCO2 breakthrough (note that this study has limited discussion andsimulation of brine injectivity or the WAG process). The analysisdoes not take into account crossflow, dispersive mixing, and three-phase flow, but the numerical simulation does. The CO2 solventbank was suggested as having a low resistance compared to the lowmobility downstream of the solvent bank; as such, injectivity is notsensitive to the CO2 rich-phase relative permeability endpoints.Contrarily, injectivity is suggested as being a strong function of theoil-phase relative permeability until breakthrough, when itbecomes a weakly decreasing function of the oil relative perme-ability endpoint. Injectivity was determined to be a decreasingfunction of the aqueous-phase endpoint in Roper’s study.

Roper additionally studied the relative permeability curvaturesusing the analytical model to show that the nonaqueous phase (oil)relative permeability curvature applicable to the tertiary oil bankhas the most important influence on the early injectivity, but littleeffect after CO2 breakthrough. Roper’s results also indicate that theoil and aqueous phase relative permeability endpoints and the rel-ative permeability curvatures are almost inverses of each other.This exemplifies the antagonistic and synergistic complex behav-ior that coexists in the reservoir and shows why it is difficult toascertain one individual parameter or set of parameters that havemore influence over the injectivity phenomenon than any others.The phase residual saturation effects of displacement and their con-sequences on injectivity of CO2 are discussed later.

The distribution of the wetting- and nonwetting-phase fluids isa major factor in determining relative permeability characteristics.Nonwetting fluids occupying larger pores will have larger relativepermeabilities.61 To acquire a history match, most simulations alterthe relative permeability information significantly, as was done inthe Sundown Unit of west Texas.95 These extreme adjustments aremade to compensate for more heterogeneity and are a faster andmore convenient way to match waterfloods than to change geolog-ical models. Altering the relative permeability data also allows oneto account for poorer sweep efficiency. Because of this, an excel-lent waterflood history match obtained by altering the relative per-meability does not guarantee a correct CO2 flood forecast.

Adjusting the gas relative permeability curve could compensatefor moderate differences in the reservoir heterogeneity. Drasticallymodifying the gas curve from expected values could create mis-leading results in timing and performances of the WAG and con-tinuous CO2 process.96 Relative permeability (and its effects) is astudy unto itself and again takes into account many petrophysicalparameters. To confine the scope of the present study, this issuewill not be discussed further. Additional comments on the effects

of relative permeability on injectivity are related in the IFT sectionof the present study.

Saturation Effects. The nature of the WAG process causes the sat-urations to cycle and can exacerbate the trapping occurrence. Thevolume of the trapped phase depends on the initial saturationbefore the flood.12 High water saturation acts to reduce the amountof extraction that occurs in both water-wet and mixed rock.19

Roper et al.,16 in their analysis of the causes of tertiary injec-tivity abnormalities, investigated saturation effects with an analyt-ical model and a compositional numerical simulator. The methodof study they used compared an analytical model that made noallowance for dispersion or vertical communication with the com-positional simulator that had considerably more complexity in thedefinition of the reservoir and fluids.

The analytical solution showed that injectivity was not sensitiveto the CO2-phase residual saturation in the solvent bank. However,analysis by numerical compositional simulator predicts higher CO2

injectivity early in the displacement for lower CO2 residual satura-tions. Suggested reasons for this are increased mobility in the mix-ing and increased crossflow because of reduced pressure dropwithin the mixing zone in higher-permeability layers. After CO2

breakthrough, injectivity is a weakly decreasing function of CO2-phase residual saturation. The authors reasoned that:

• A lower value of saturation means an increased total mobilityratio in the mixing zone.

• There is an increasingly unfavorable local mobility ratiowhere the multiphase mixture displaces compositions at the rear ofthe oil bank.

• More oil is bypassed because of crossflow and instability.• Increased bypassing of oil reduces the flow area available for

CO2 cycling through the reservoir.A result from the compositional simulator17 showed that injec-

tivity is very sensitive to the oil-phase saturation, in which the sat-uration results from displacement by the aqueous phase. Theanalytical solution showed an increasing injectivity trend for higherresidual saturation, but the numerical solution showed a reversal inthe presence of dispersion and vertical convective mixing. Higheroil-bank mobility is present in both the numerical and analyticalsolutions. Higher oil-bank mobility reduces the driving force forcrossflow, which dominates other mechanisms and is responsiblefor causing injectivity to go from a decreasing function of thisparameter to an increasing function before CO2 breakthrough. Lesscrossflow improves sweep in high-permeability layers. After break-through, injectivity becomes an even more strongly increasingfunction of the oil-phase saturation to displacement by the aqueousphase. The residual oil-phase saturation to displacement by an aque-ous phase is one of the most important petrophysical parametersregarding injectivity, and the magnitude of its influence depends onthe amount of crossflow.

The oil-phase saturation to displacement by the CO2-rich phasedoes not enter into the analytical model, but the numerical modelindicates that it is one of the most significant influences on early-time injectivity and one of the most important parameters in oil-recovery predictions. In the presence of dispersion and verticalcommunication, early-time injectivity is an increasing function ofthe oil saturation displaced by the CO2-rich phase. This is coun-terintuitive because oil-phase relative permeabilities at intermedi-ate saturations are associated with higher residual saturation.More efficient oil-bank transport increases CO2-specific velocityin high-permeability layers; phase behavior effect, oil viscosity inthe low-permeability layer, and instability effect mechanismscombine and contribute to a larger oil saturation displaced by theCO2-rich phase.

A higher value of oil-phase residual saturation displaced byCO2 increases the oil-phase saturation in the mixing zone and inthe solvent bank within the high-permeability zone. This causes anincrease in pressure drop in these zones and a subsequent decreasein the driving force for crossflow of oil in the tertiary oil bank fromthe high-permeability zones to the low-permeability layers. Theincrease in oil-bank transport in the high-permeability layer ismore efficient and increases injectivity. Also, with less crossflow


of CO2, the solvent bank in the high-permeability layer travelsfaster relative to the oil bank in the low-permeability layer. As aresult, a greater fraction of the length of the oil bank in the low-permeability layer is contacted by CO2 and experiences oil-viscosityreduction. This scenario can increase injectivity but, as pointed outby the authors, these are complex interactions, and competingmechanisms can reverse both early- and late-time trends with otherpossible outcomes.17

The aqueous-phase residual saturation sensitivity analysisshows that early injectivity before CO2 breakthrough is a verystrongly decreasing function of this parameter. Higher residual sat-uration reduces mobility in the dispersive mixing zone and com-plements the mobility reduction in the tertiary oil bank by loweringaqueous-phase relative permeability at intermediate saturations.After CO2 breakthrough, sensitivity to this parameter is reduced.Reduction of mobility in the lower-permeability layer also con-tributes to the relaxing of the sensitivity to this parameter afterbreakthrough. The presence of miscible residual oil saturation sub-stantially reduces predicted oil recovery and reduces CO2 relativepermeability, and they should be applied together.90

Heterogeneity, Anisotropy, and Stratification. Stratification maystrongly influence the water/gas displacement process.12

Horizontal fluid flow in vertically communicating porous strata areinfluenced by flow perpendicular to the bulk flow caused by vis-cosity forces, capillarity forces, gravity forces, and dispersion.97

Capillary crossflow can lead to significant improvement in verticalsweep. Typical oil reservoirs have a Dykstra-Parson coefficient of0.6 to 0.8.57 WAG recovery is more sensitive to reservoir hetero-geneity than is waterflooding.98 Unfavorable mobility miscible dis-placements lead to crossflow from the low-permeability layer to anadjacent higher-permeability layer and tend to reduce frontaladvancement in the lower-permeability layer.14

Pizarro and Lake99 studied the effect of heterogeneity on injec-tivity through geostatistical analysis and autocorrelation of thereservoir permeability distribution. They modeled injectivity as afunction of 10 parameters in heterogeneous reservoirs: permeabil-ity in the x and z directions, viscosity, pressure at the well, lengthand width of a rectangular reservoir, bottom and top of the perfo-ration interval, reservoir thickness, and flow rate.

Vertical conformance of WAG displacement is strongly influ-enced by conformance between zones. In a noncommunicating-layered system, vertical distribution of CO2 is dominated bypermeability contrasts.7 Flow into each layer is essentially pro-portional to the fraction of the overall system kh (flow capacity,where k�permeability, md, and h�height of layer or zone, ft) andis independent of WAG ratio. There is a tendency for more CO2 toenter the high-permeability zone with increasing WAG ratio.7

Because the WAG behavior is cyclic, the most permeable layerresponds most quickly and takes more fluid than is relative to itspermeability-height contribution. When water is injected, itquickly displaces the highly mobile CO2, and all the layers attainan effective mobility nearly equal to the initial value. The higher-permeability layers always respond first. WAG will reducemobility not only in the high-permeability layer but also in thelow-permeability layer, resulting in a larger amount of CO2 enter-ing the highest-permeability layer.7

The ratio of viscous to gravity forces is the prime variable fordetermining the efficiency of WAG injection, and it controls ver-tical conformance and displacement efficiency of the flood.Crossflow or convective mixing can substantially increase injec-tivity even in the presence of low vertical-to-horizontal perme-ability ratios.16 Transport of CO2 is enhanced significantly by thehigh-permeability layers establishing a highly conductive pathparallel to the low-permeability layer. With crossflow, CO2 istransported through the highly permeable layer and reaches down-stream locations in the low-permeability layer that, without cross-flow, would have to flow through the low-permeability layer toreach the downstream locations. Thus, crossflowing will increaseinjectivity of CO2. High permeability “thief” zones in the Mabeefield could be the cause of high injection rates.55 Heterogeneousstratification causes physical dispersion, reduces channeling of

CO2 through the high-permeability layer, and delays break-through.16,17 This is attributed to permeability contrast and mobilityratio contrast caused by different growth rates in the mixing zone inregions of low oil saturation for CO2-swept regions in each layer; itis thus unfavorable.

Transport Considerations. A study using surfactants and a singlecapillary constriction investigated mass transfer at the porelevel.100 The effects of mass transfer on oil trapped in pore throatsindicate that the mass transfer is sufficiently slow that equilibriumis not necessarily attained. This is contrary to a previous study byRaimondi and Torcaso,101 who concluded that mass transfer inporous media takes place at equilibrium conditions and that misci-bility in a reservoir is attained instantly because of equilibrium.Though mass transfer through porous structures and packed bedshave been studied for some time in other disciplines, these twostudies are early attempts to apply the mechanisms of mass trans-fer to a petroleum-recovery perspective.

More recent studies indicate that contact time on the develop-ment of miscibility has not been resolved.19 Miscibility developswhen light crude oil components mix with the solvent. Contacttime can have a strong influence on flow performance. Someresearchers suggest that miscibility can require 32 ft or more.Others suggested that the core length is not a significant factor.Miscibility is most certainly dependent on the compositionalmakeup, and micro- or local heterogeneity in the cores probablylends to the discrepancy. Slim-tube measurements give values ofresidual oil saturations close to zero, which is far different fromfield and coreflood measurements that give considerably highersaturations. In-situ emulsification is a natural consequence of sys-tems that produce low IFT in the converging/diverging porousmedia, and the Marangoni instability has been noted.

As noted previously, the mixing phenomena can significantlyinfluence injectivity. At the reservoir scale, physical dispersion cansignificantly reduce injectivity, though the oil bank mobility is notlow, the longitudinal dispersion is scaled to reservoir conditions,and the mixing zone is a small fraction of the porous media.16,17

Thus, analytical assumptions that dispersion and associated phasebehavior can be neglected may not be justified. Injectivity is alsoreduced because dispersive mixing reduces channeling of CO2

through high-permeability layers and delays CO2 breakthrough.Neglecting volume change of mixing and the presence of mixingzones, the associated mobility reductions are rendered increas-ingly more important in the high-permeability layer as the dis-placement progresses and the solvent bank grow in proportion tothe cumulative throughput in each layer. Injectivity is lowerbecause a negative overall volume change upon mixing slowsgrowth of the region invaded by low-viscosity CO2 in both thelow- and high-permeability layers.16,17

Early investigators speculated that multiphase flow could signif-icantly affect the injectivity of a field project.102 Henry andMetcalf103 measured a slight injectivity drop across cores that hadmultiphase flow in CO2/oil systems. There is no clear experimentalevidence that multiphase behavior effects result in field-observedfluid mobilities. Patel et al.26 concluded that phase behavior doesnot necessarily create injectivity decreases by itself. Phase behavioreffects have been shown to reduce fingering.57 Accounting forphase behavior and mixing, CO2 injectivity appears to be muchhigher for reservoirs with crossflow.

The CO2/oil system is dynamic and forms multiple phases.Grigg and Siagian104 have shown that in low-temperature CO2

floods, four phases can exist—three nonaqueous phases and a solidasphaltene phase. Additionally, four liquid phases and a solid phasecan co-exist in a CO2 flood—an aqueous phase, liquid hydrocar-bon, liquid carbon dioxide, and gaseous carbon dioxide. The sys-tem can move in and out of miscibility; thus, dynamic phasebehavior should be considered in modeling a system.16,17,105

IFT. Unlike conventional gas and oil or water and oil, the flowbehavior of low-IFT fluids occurring in most IOR processesdepend on IFT, viscosity, and flow rate, as well as rock propertiesof pore distribution and wettability.106,107 The IFT is another criti-


cal parameter in IOR process and more recently has received anincreased effort in gas-condensate reservoirs.108 The effect of lowIFT is to increase flow rates and lower the residual saturations, cre-ating conditions for improved hydrocarbon recovery.109 The IFTdetermines the curvature of the relative permeability curves. In acompletely miscible process, the IFT is zero, and relative perme-ability is a linear function of the fluid saturation with a slope of1.110 This parameter is the most sensitive and the most easily mod-ified in the capillary number. The capillary number is a function ofvelocity, viscosity, and IFT. Orders-of-magnitude change in thecapillary number is normally required to produce significantdecrease in residual oil saturation and, with gas injection, the IFTcan be lowered significantly. In-situ mass transfer dictates the levelof IFT reduction. Considerable decrease at relatively low cost isthe benefit of miscible flooding. Pore-size distribution also affectsthe IFT because it will dominate if pore throats are small.

Limited knowledge of the effects of low IFT on relative perme-ability has been available since the 1950s; however, there has beenlittle documented mathematical correlation other than empiricaldeductions between IFT and gas/oil relative permeability.108,111

Typical early unsteady-state gas/oil relative permeability experi-ments did not span the range of IFTs that would be present in amass-transfer-dominated system. Thus, they are inadequate inassessing the mobility/IFT interaction.109 As a result, the couplingof IFT, mobility, and pore-size distribution has been omitted frommany analyses.

More recent studies into the effects of IFT on relative perme-ability curves have been conducted and published. The most recentpublications report on studies conducted on gas/liquid IFT effectson relative permeability in gas condensate reservoirs, but they maybe extended to near-miscible/miscible EOR projects. In a miscibledisplacement process, Harbert107 suggested that both water and oilrelative permeability curves were found to shift upward, indicatingthat the two phases interfere less with each other as IFT is reduced.In addition, he suggested that flow tests on representative reservoir-rock samples are necessary to describe low IFT relative perme-ability for field process performance calculation.

The relative permeability has been shown to vary for gas-condensate fluids when the velocity changes at a fixed IFT or whenthe velocity is fixed and the IFT is changed.110,112 Therefore,numerous relative permeability curves are necessary to cover therange of flow rates and IFT values within a reservoir’s differentflow regimes. Henderson et al.110 suggested that the capillary num-ber could be used to correlate the gas relative permeability and thegas velocity and gas viscosity parameters. Further development inthis concept is required. Fulcher et al.113 suggested that the capil-lary number is not important in correlating relative permeabilityand residual saturation, but the IFT and viscosity individuallyaffect the flow rate. They also observed that increasing the capil-lary number reduces hysteresis effects in the relative permeabilitycurve. Below a surface tension of 2 mN/m, the surface tension hasa significant effect on relative permeability.113 One interesting phe-nomenon is that high IFT ultimately caused condensate relativepermeability to decrease with increasing condensate saturation,and condensate immobile under gas injection could be recoveredby water injection.112

McDougal et al.114 described gas/oil flow studies over a rangeof IFTs of three orders of magnitude. The gas relative permeabilitycurves showed a marked increase, while very little change wasobserved in the oil curve with decreasing IFT. The gas curvebecomes a linear function of gas saturation as IFT tends to zero.Additionally these authors state that to predict a priori the direc-tional trend of the relative permeability curves with varying IFT isvery difficult because it is also linked intrinsically to the viscosityratio. Therefore, the effect of IFT on relative permeability curvescan be understood only by accounting for both the capillary num-ber and the viscosity ratio’s role in determining phase distributionsduring displacement.

Thomas et al.108 illustrate a method in which one may deter-mine when one should be concerned more with controlling mobil-ity and when IFT optimization is justified. Additionally,interactions between the pore-size distribution, IFT, and viscosity

will determine if miscibility is important to recovery. Low IFT isgenerally a necessary condition for efficient recovery from mostreservoirs, but in many cases, zero IFT is unnecessary unless thepore-throat-size distribution is extremely tight and the rock is oil-wet. Spontaneous imbibition tests106 conducted on water-wet Bereasandstone with high oil saturation imbibed conventional waterovernight. When low IFT fluids replaced the conventional fluids,neither the oil nor brine phases were imbibed after 1 week.

Hanniff and Ali109 suggest that the capillary number plays animportant role in controlling residual saturation and that changes inthe capillary number are caused almost entirely by changes in IFT.But these authors also demonstrate the importance of gravitationalforces compared to capillary forces and suggest that the Bondnumber is more appropriate in interpreting data. However, a grav-ity-dominated system does not exist when the IFTs are below acritical value. Pope et al.115 believe that it is not correct to modelthe relative permeabilities strictly or directly as a function of IFT.These parameters should be modeled as a function of the combinedeffects of pressure gradient, buoyancy, and capillary forces using ageneralized form of the capillary number and Bond number into atrapping number.

The IFT between water and CO2 is high at low temperaturesand pressures (e.g., approximately 70 mN/m at 25�C and0.1MPa)116 and decreases (20 to 27 mN/m) as the temperature andpressure increase. At higher pressures, the IFT is largely inde-pendent of pressure. This is attributed to the solubility increase ofCO2 in water with pressure. At higher pressures, the free-energydensity of the CO2 becomes more liquidlike and closer to that ofwater. At low pressures (i.e., �3.5 MPa) the higher-temperatureisotherms have IFT values that are lower than lower-temperatureisotherms. As the pressure is increased, the IFT isotherms con-verge (cross over) at 2.5 to 3.5 MPa, and thereafter, the higher-temperature isotherms have higher IFT values compared to thelower-temperature isotherms.

In the proximity (�10 to 20�C) of the critical point of CO2, theIFT decreases markedly and the surface tension of CO2 isapproached, creating a dip or cusp in the IFT vs. pressure. At highertemperatures, this dip or cusp is less of an effect, eventually disap-pearing with temperature. This cusp is attributed to an increase inthe excess adsorption owing to attraction of CO2 to the interface.The rise in IFT after the dip with increasing pressure implies a des-orption process.116 A very small amount of a third-phase interme-diate in composition between the CO2- and H2O-rich phases isobserved to occur at the minimum of the dip. The IFT between theCO2 and water systems at high pressures is approximately 20 to 25mN/m and is lower than that for water/hydrocarbon systems.117

This is attributed to the higher miscibility of CO2 and water vs.hydrocarbons and water.

IFTs of the ternary CO2/water/alcohol mixtures are lower thanthe binary CO2/water system. Methanol has the least effect on IFTreduction with progressively increasing effect to isopropyl alcohol(higher-molecular-weight alcohols). This leads to the possible useof surfactants to lower the IFT between water and carbon diox-ide.117 The IFT between carbon dioxide and water can be loweredfrom �20 mN/m to 2 mN/m with several surfactants. One surfac-tant, PFPE COO�NH4

�, has been shown to reduce the IFT to 0.8mN/m with a critical micro-emulsion observed. This could allow achange in the relative permeability and could increase injectivityduring WAG. The solubility of polymers in CO2 is generally lowand is a function of the surface tension and molecular weight of thepolymer.118 However, the effects on the purpose of the flood (oilrecovery) and the economics are not known.

Concluding DiscussionThe industry has improved techniques to minimize injectivitylosses and/or improve recovery economics for the WAG process.These techniques include flood designs based on industry expe-rience and dimensionless performance plots; WAG tapering,such as Unocal’s Hybrid WAG and Shell’s DUWAG processes;volume-based (instead of time-based) WAG; well realignment;horizontal gas-injection wells; shorter, more frequent cycles toreduce large gas swings; increased WAG cycle lengths to mini-


mize injectivity losses; and improved conformance control usingfoams, gels, and polymers.

Conclusions on factors affecting injectivity are as follows.1. Lower injectivity is not necessarily a near-wellbore effect.2. Low mobility in the tertiary oil bank significantly affects injec-

tivity, especially for stimulated injection wells with nonstimu-lated producing wells. The closer the banks are to the injectionwells, the more effect the lower-mobility banks have on theactivity at the injector.

3. Salinity and pH may change reservoir wettability. DissolvedCO2 (carbonic acid) reduces the water pH. A lower pH causesless water-wet conditions. However, pore coatings may controlwetting alterations in natural porous media.

4. Wettability is a complex critical parameter in injectivity reduc-tions. Wetting characteristics of the reservoir rock appear to bethe most controlling factors of the operating strategy for an IORprocess. Water-wet laboratory models indicate that gravity forcesdominate, and the evolved operating strategy is continuous gasinjection under these conditions. Viscous fingering is a control-ling factor in oil-wet conditions and suggests a WAG processwith an optimum of equal, or 1:1, velocity ratio. Mixed-wet con-ditions indicate that maximum recovery is a stronger function ofslug size. Injectivity losses owing to wettability effects on misci-ble flooding need to be further delineated. Wettability alterationcaused by CO2 flooding has conflicting reports in the literature.In mixed-wet rocks, injected gas enters the high-permeabilitylayers, resulting in reduced water-injection rate owing to three-phase and compressibility effects. Mixed wettability is suggestedas a cause of low fluid mobility. Low injectivity in the carbonatereservoirs of west Texas is probably caused by the oil-wet ormixed-wet behavior of these rocks.

5. Dissolution, precipitation, and particle invasion/migrationduring injection of CO2 and/or the WAG process, and theireffects on the WAG process (if any), have not been proved ordisproved conclusively.

6. Trapping and bypassing of gas, like wettability, is a complexparameter in determining injectivity, possibly because of its linkto wettability. There is a discrepancy in the literature as to theeffect of residual oil saturation on the trapped-gas saturation.Bypassed or trapped oil causes three-phase relative permeabilityreductions, resulting in loss of injectivity. Trapping behavior andrelative permeability depend on the ratio of flow rate to IFT.Mobile water in the reservoir may shield the in-place oil frombeing contacted by injected solvent. Optimum WAG ratio is fairlyinsensitive to any assumed level of oil-phase trapping. Trappingappears to be more significant at the coreflood laboratory leveland rapidly decreases at field scale. Rupturing the water film cansignificantly affect the trapped oil. Trapped-oil saturation in sol-vent processes has to be considered as a function not only of thewater saturation but also of the solvent contact time or the flood-ing state. Trapping and water-shielding of oil is significant inwater-wet reservoirs. Trapped gas creates significant hysteresiseffects and reduces relative permeability to water, especially inmixed-wet and oil-wet reservoirs.

7. Oil and gas saturations present in a miscible flood act to lowerthe maximum attainable water saturation, resulting in reducedwater mobility. CO2 relative permeability can be very smallcompared to oil endpoint relative permeability. Relative perme-abilities in miscible gas-injection systems are dependent on thesaturation of that phase only. In the Prudhoe Bay mixed-wetsystem, the oil relative permeability is a function of oil satura-tion only. Gas relative permeability, as a function of gas satura-tion alone, may not be valid. CO2 injectivity is a decreasingfunction of the aqueous phase endpoint. Injectivity before CO2

breakthrough is a strongly decreasing function of aqueous-phase residual saturation.

8. Vertical heterogeneity and high permeability/porosity ratio (k/�)layers can have significant effects on gas injectivity. Crossflowor convective mixing can increase CO2 injectivity substantially,even in the presence of low vertical-to-horizontal permeabilityratios. The dispersive mixing zone has low mobility and canreduce CO2 injectivity by augmenting total mobility and macro-

scopic oil bypassing resulting from reservoir heterogeneity. CO2

injectivity is an increasing function of increased transport inhigh-permeability layers near the injection face.

9. Mass-transfer contact time and miscibility development ratestill appear to be in contention in the literature.

10.The IFT correlation of gas relative permeability by the capillarynumber is being debated in the literature. Increasing the capil-lary number has been suggested to reduce hysteresis effects inthe relative permeability curve. Though low IFT is importantfor efficient recovery, zero IFT is unnecessary unless pore-throat size is small.

Further DirectionsKeeping in mind that laboratory injection studies alone cannot beconsidered perfect indicators of field-scale injectivity during theWAG process,119 there are a number of suggested research areas.These include the investigation of the effects of high-velocity fluidsaround (WAG) injection wells, miscible flooding on wettability,flood banks, pressure redistribution, crossflow (communicating andnoncommunicating permeability differences), wettability, dissolu-tion, precipitation, particle invasion/migration, residual oil saturationon trapped gas saturation, contact time, mass transfer, IFT, and rel-ative permeability. The use of pore-scale simulators is recom-mended to aid in understanding these effects. Additionally,field-scale verification studies and demonstration projects areneeded for proper scaling to field application.

AcknowledgmentsThe authors express their appreciation for the support receivedfrom the U.S. Dept. of Energy Natl. Petroleum Technology Office(Contract No. DE-FG22-97BC-15047) and the State of NewMexico. Acknowledgment is also extended to the many organiza-tions and individuals without whose help this project would nothave been as successful.

References1. Moritis, G.: “2000 Worldwide EOR Survey,” Oil & Gas J. (20 March

2000) 45.2. Brock, W.R. and Bryan, L.A.: “Summary Results of CO2 EOR Field

Tests, 1972–1987,” paper SPE 18977 presented at the 1989 SPE JointRocky Mountain Regional/Low Permeability Reservoir Symposiumand Exhibition, Denver, 6–8 March.

3. Hadlow, R.E.: “Update of Industry Experience With CO2 Injection,”paper SPE 24928 presented at the 1992 SPE Annual TechnicalConference and Exhibition, Washington, DC, 4–7 October.

4. Grigg, R.B. and Schechter, D.S.: “State of the Industry in CO2 Floods,”paper SPE 38849 presented at the 1997 SPE Annual TechnicalConference and Exhibition, San Antonio, Texas, 5–8 October.

5. Grigg, R.B. and Schechter, D.S.: “Improved Efficiency of MiscibleCO2 Floods and Enhanced Prospects for CO2 Flooding HeterogeneousReservoirs,” Annual Report, Contract No. DE-FG26-97BC15047,U.S. DOE, Washington, DC (October 1999).

6. Christensen, J.R., Stenby, E.H., and Skauge, A.: “Review of WAGField Experience,” SPEREE (April 2001) 97.

7. Gorrell, S.B.: “Implications of Water-Alternate-Gas Injection forProfile Control and Injectivity,” paper SPE 20210 presented at the 1990SPE/DOE Symposium on Enhanced Oil Recovery, Tulsa, 22–25 April.

8. Pariani, G.J. et al.: “An Approach To Optimize Economics in a WestTexas CO2 Flood,” JPT (September 1992) 984; Trans., AIME, 293.

9. Surguchev, L.M. et al.: “Screening of WAG Injection Strategiesfor Heterogeneous Reservoirs,” paper SPE 25075 presented at the1992 SPE European Petroleum Conference, Cannes, France,16–18 November.

10. Caudle, B.H and Dyes, A.B.: “Improving Miscible Displacement byGas-Water Injection,” Trans., AIME (1958) 213, 281.

11. Huang, E.T.S. and Holm, L.W.: “Effect of WAG Injection and RockWettability on Oil Recovery During CO2 Flooding,” SPERE (February1988) 119.

12. Surguchev, L.M., Korbol, R., and Krakstad, O.S.: “Optimum WaterAlternate Gas Injection Schemes for Stratified Reservoir,” paper SPE24646 presented at the 1992 Annual Technical Conference andExhibition, Washington, DC, 4–7 October.


13. Jones, L.G., Cullick, A.S., and Cohen, M.F.: “WAG Process PromisesImproved Recovery in Cycling Gas Condensate Reservoirs: Part I—Prototype Reservoir Simulation Studies,” paper SPE 19113 presentedat the 1989 SPE Gas Technology Symposium, Dallas, 7–9 June.

14. Cullick, A.L. et al.: “WAG May Improve Gas-Condensate Recovery,”SPERE (August 1993) 207; Trans., AIME, 295.

15. Christman, P.G. and Gorell, S.B.: “A Comparison of Laboratory andField-Observed CO2 Tertiary Injectivity,” JPT (February 1990) 226;Trans., AIME, 289.

16. Roper, M.K. Jr. and Pope, G.A.: “Analysis of Tertiary Injectivity ofCarbon Dioxide,” paper SPE 23974 presented at the 1992 SPE PermianBasin Oil and Gas Recovery Conference, Midland, Texas, 18–20 March.

17. Roper, M.K.: “Simulation of Tertiary Carbon Dioxide Injectivity,”PhD dissertation, the U. of Texas, Austin, Texas (1994).

18. Tiffin, D.L. and Yelling, W.F.: “Effects of Mobile Water on Multiple-Contact Gas Displacements,” SPEJ (June 1983) 447.

19. Stern, D.: “Mechanisms of Miscible Oil Recovery: Effects of Pore-Level Fluid Distribution,” paper SPE 22652 presented at the 1991 SPEAnnual Technical Conference and Exhibition, Dallas, 6–9 October.

20. Gorell, S.B.: “Modeling the Effects of Trapping and Water AlternateGas (WAG) Injection on Tertiary Miscible Displacements,” paper SPE17340 presented at the 1988 SPE/DOE Enhanced Oil RecoverySymposium, Tulsa, 17–20 April.

21. Jackson, D.D., Andrews, G.L., and Claridge, E.L.: “Optimum WAGRatio vs. Rock Wettability in CO2 Flooding,” paper SPE 14303 pre-sented at the 1985 Annual Technical Conference and Exhibition, LasVegas, Nevada, 22–25 September.

22. Hindi, R., Cheng, C.T., and Wang, B.: “CO2 Miscible FloodSimulation Study, Roberts Unit, Wasson Field, Yoakum County,Texas,” paper SPE/DOE 24185 presented at the 1992 SPE/DOESymposium on Enhanced Oil Recovery, Tulsa, 22–24 April.

23. Hsu, C.F. et al.: “Design and Implementation of a Grass-Roots CO2

Project for the Bennett Ranch Unit,” paper SPE 35188 presented at the1996 SPE Permian Basin Oil and Gas Recovery Conference, Midland,Texas, 27–29 March.

24. Bullock, G.W., Wood, T.B., and Korecki, M.L.: “A Brief History of theWasson EOR Project,” SPEPE (August 1990) 338.

25. Mathis, R.L. and Sears, S.O.: “Effect of CO2 Flooding on DolomiteReservoir Rock, Denver Unit Wasson (San Andres) Field, TX,” paperSPE 13132 presented at the 1984 SPE Annual Technical Conferenceand Exhibition, Houston, 16–19 September.

26. Patel, P.D., Christman, R.G., and Gardner, J.W.: “An Investigation ofUnexpectedly Low Field-Observed Fluid Mobilities During SomeCO2 Tertiary Floods,” SPERE (November 1987) 507.

27. Tanner, C.S. et al.: “Production Performance of the Wasson DenverUnit CO2 Flood,” paper SPE 24156 presented at the 1992 SPE/DOESymposium on Enhanced Oil Recovery, Tulsa, 22–24 April.

28. Fleming, E.A., Brown, L.M., and Cook, R.L.: “Overview ofProduction Engineering Aspects of Operating the Denver Unit CO2

Flood,” paper SPE 24157 presented at the 1992 SPE/DOE Symposiumon Enhanced Oil Recovery, Tulsa, 22–24 April.

29. Kittridge, M.G.: “Quantitative CO2-Flood Monitoring—Denver Unit,Wasson (San Andres) Field,” SPEFE (December 1993) 299.

30. Hsu, C.F. et al.: “Field-Scale CO2-Flood Simulations and TheirImpact on the Performance of the Wasson Denver Unit,” SPERE(February 1997) 4.

31. Burbank, D.E.: “Early CO2 Flood Experience at the South WassonClearfork Unit,” paper SPE 24160 presented at the 1992 SPE/DOESymposium on Enhanced Oil Recovery, Tulsa, 22–24 April.

32. Hallenbeck, L.D. et al.: “Innovative Approach to CO2 ProjectDevelopment Holds Promise for Improving CO2 Flood Economics inSmaller Fields Nearing Abandonment,” paper 28334 presented at the1994 SPE Annual Technical Conference and Exhibition, New Orleans,25–28 September.

33. Taber, J.J. and Seright, R.S.: “Horizontal Injection and ProductionWells for EOR or Waterflooding,” paper SPE 23952 presented at the1992 SPE Permian Basin Oil and Gas Recovery Conference, Midland,Texas, 18–20 March.

34. Lim, M.T., Khan, S.A., and Pope, G.A.: “Simulation of CarbonDioxide Flooding Using Horizontal Wells,” paper SPE 24929 presentedat the 1992 SPE Annual Technical Conference and Exhibition,Washington, DC, 4–7 October.

35. El Sheemy, M.M.: “Water Injection Performance: Case Study andRelated Spontaneous Injectivity Improvement,” paper SPE 15740 pre-sented at the 1987 SPE Middle East Oil Show, Manama, Bahrain,7–10 March.

36. Ramsay, H.J. Jr. and Small, F.R.: “Use of Carbon Dioxide for WaterInjectivity Improvement,” JPT (January 1964) 25.

37. Christensen, R.J.: “Carbonated Waterflood Results: Texas andOklahoma,” paper SPE 66 presented at the 1961 SPE Annual Meetingof Rocky Mountain Petroleum Sections of AIME, Farmington, NewMexico, 25–26 May.

38. Perez, J.M. et al.: “Improving the Potential to Produce Oil FromNaturally Fractured Reservoirs,” paper 28677 presented at the 1994SPE International Petroleum Conference and Exhibition of Mexico,Veracruz, Mexico, 10–13 October.

39. Shenawi, S.H. and Wu, C.H.: “Compositional Simulation ofCarbonated Waterfloods in Naturally Fractured Reservoirs,” paperSPE 27741 presented at the 1994 SPE/DOE Symposium on ImprovedOil Recovery, Tulsa, Oklahoma, 17–20 April.

40. Shenawi, S.H. et al.: “A New Iterative Mathematical Model for theAnalysis of Imbibition Carbonated Waterflood in NaturallyFractured Reservoirs,” paper SPE 27717 presented at the 1994 SPEPermian Basin Oil and Gas Recovery Conference, Midland, Texas,16–18 March.

41. Perez, J.M., Poston, S.W., and Edwards, C.M.: “The Effect of Micro-Fractures on Oil Recovery From Low-Permeability Reservoirs,” paperSPE 25857 presented at the 1993 SPE Rocky Mountain Regional/LowPermeability Reservoirs Symposium, Denver, 12–14 April.

42. Perez, J.M., Poston, S.W., and Sharif, Q.J.: “Carbonated WaterImbibition Flooding: An Enhanced Oil Recovery Process forFractured Reservoirs,” paper SPE 24164 presented at the 1992SPE/DOE Symposium on Enhanced Oil Recovery, Tulsa, 22–24 April.

43. Flumerfelt, R.W. et al.: “A Cyclic Surfactant-Based Imbibition/Solution Gas Drive Process for Low-Permeability, FracturedReservoirs,” paper SPE 26373 presented at the 1993 SPE AnnualTechnical Conference and Exhibition, Houston, 3–6 October.

44. van Lingen, P.P., Barzanji, O.H.M., and van Kruijsdijk, C.P.J.W.:“WAG Injection to Reduce Capillary Entrapment in Small-ScaleHeterogeneities,” paper SPE 36662 presented at the 1996 SPE AnnualTechnical Conference and Exhibition, Denver, 6–9 October.

45. Blackford, T.A.: “Carbonated Waterflood Implementation and ItsImpact on Material Performance in a Pilot Project,” paper SPE 16831presented at the 1987 SPE Annual Technical and ExhibitionConference, Dallas, 27–30 September.

46. Bargas, C.L. et al.: “Immiscible CO2 Process for the Salt CreekField,” SPERE (November 1992) 397.

47. Walker, J.W. and Turner, J.L.: “Performance of Seeligson Zone 20B-07 Enriched-Gas-Drive Project,” JPT (April 1968) 369.

48. Ma, T.D. et al.: “Simultaneous Water and Gas Injection Pilot at theKuparuk River Field, Reservoir Impact,” paper SPE 30726 presentedat the 1995 SPE Annual Technical Conference and Exhibition, Dallas,22–25 October.

49. Henderson, J.H. et al.: “A Laboratory Investigation of OilDisplacement From Porous Media By A Liquefied Petroleum Gas,”paper SPE 218-G presented at the 1952 SPE Fall Meeting of thePetroleum Branch, AIME, Houston, 1–3 October.

50. Harvey, M.T. Jr., Shelton, J.L. and Kelm, C.H.: “Field InjectivityExperiences With Miscible Recovery Projects Using Alternate Rich-Gas and Water Injection,” JPT (September 1977) 1051.

51. Schneider, F.N. and Owens, W.W.: “Relative Permeability Studies ofGas-Water Flow Following Solvent Injection in Carbonate Rocks,”SPEJ (February 1976) 23; Trans., AIME, 261.

52. Winzinger, R. et al.: “Design of a Major CO2 Flood, North Ward EstesField, Ward County, Texas,” SPERE (February 1991) 11.

53. Ring, J.N. and Smith, D.J.: “An Overview of the North Ward EstesCO2 Flood,” paper SPE 30729 presented at the 1995 SPE AnnualTechnical Conference and Exhibition, Dallas, 22–25 October.

54. Prieditis, J., Wolle, C.R., and Notz, P.K.: “A Laboratory and FieldInjectivity Study: CO2 WAG in the San Andres Formation of WestTexas,” paper 22653 presented at the 1991 SPE Annual TechnicalConference and Exhibition, Dallas, 6–9 October.

55. Roper, M.K. et al.: “Interpretation of a CO2 WAG Injectivity Test inthe San Andres Formation Using a Compositional Simulator,” paper


SPE 24163 presented at the 1992 SPE/DOE Symposium on EnhancedOil Recovery, Tulsa, 22–24 April.

56. Good, P.A. and Downer, D.G.: “Cedar Creek Anticline CarbonDioxide Injectivity Test: Design, Implementation, and Analysis,”paper SPE 17326 presented at the 1988 SPE/DOE Enhanced OilRecovery Symposium, Tulsa, 17–20 April.

57. Chang, Y.-B. et al.: “CO2 Flow Patterns Under Multiphase Flow:Heterogeneous Field-Scale Conditions,” SPERE (August 1994) 208.

58. Potter, G.F.: “The Effects of CO2 Flooding on Wettability of WestTexas Dolomitic Formations,” paper SPE 16716 presented at the1987 SPE Annual Technical Conference and Exhibition, Dallas,27–30 September.

59. Henry, R.L. et al.: “Utilization of Composition Observation Wells in aWest Texas CO2 Pilot Flood,” paper SPE 9786 presented at the 1981SPE/DOE Symposium on Enhanced Oil Recovery, Tulsa, 5–8 April.

60. van Lingen, P.P., Bruining, J., and van Kruijsdijk, C.P.J.W.: “CapillaryEntrapment Caused by Small-Scale Wettability Heterogeneities,”SPERE (May 1996) 93.

61. Basu, S. and Sharma, M.M.: “Defining the Wettability State of MixedWet Reservoirs: Measurement of Critical Capillary Pressure for CrudeOils,” paper SPE 36679 presented at the 1996 SPE Annual TechnicalConference and Exhibition, Denver, 6–9 October.

62. McDougal, S.R. et al.: “Network Analagous of Wettability at the PoreScale,” Developments in Petrophysics, M.A. and P.K. Harvey (eds.),Geological Society Special Publications No. 122, London (1997) 19–35.

63. Buckley, J.S. and Liu, Y.: “Some Mechanisms of CrudeOil/Brine/Solid Interactions,” paper presented at the 1996 Symposiumon Reservoir Wettability and Its Effect on Oil Recovery, Montpellier,France, 11–13 September.

64. Tang, G.Q. and Morrow, N.R.: “Effect of Temperature, Salinity andOil Composition on Wetting Behavior and Oil Recovery byWaterflooding,” paper 36680 presented at the 1995 SPE AnnualTechnical Conference and Exhibition, Dallas, 22–25 October.

65. Anderson, W.G.: “Wettability Literature Survey—Part 1:Rock/Oil/Brine Interactions and the Effects of Core Handling onWettability,” JPT (October 1986) 1125.

66. Buckley, J.S. et al.: “Wetting Alteration by Brine and Crude Oil: FromContact Angles to Cores,” paper SPE 30765 presented at the 1995 SPEAnnual Technical Conference and Exhibition, Dallas, 22–25 October.

67. Buckley, J.S. and Morrow, N.R.: “Characterization of Crude Oil WettingBehavior by Adhesion Tests,” paper SPE 20263 presented at the 1990SPE/DOE Enhanced Oil Recovery Symposium, Tulsa, 22–25 April.

68. Dixit, A.B. et al.: “Core Wettability: Should IAH Equal IUSBM,” paperpresented at the1998 Intl. Symposium of the Soc. of Core Analysts,The Hague, 14–16 September.

69. Radke, C.J., Kovscek, A.R., and Wong, H.: “A Pore-Level Scenariofor the Development of Mixed Wettability in Oil Reservoirs,” paperSPE 24880 presented at the 1992 SPE Annual Technical Conferenceand Exhibition, Washington, DC, 4–7 October.

70. Rao, D.N., Girard, M., and Sayegh, S.G.: “Interfacial Phenomena InMiscible Gas Processes,” paper SPE 19698 presented at the 1989 SPEAnnual Technical Conference and Exhibition, San Antonio, Texas,8–11 October.

71. Yeh, S.W., Ehrlich, R., and Emanuel, A.S.: “Miscible-Gasflood-Induced Wettability Alteration: Experimental Observations and OilRecovery Implications,” SPEFE (June 1992) 167.

72. Rao, D.N., Girard, M., and Sayegh, S.G.: “Impact of MiscibleFlooding on Wettability, Relative Permeability, and Oil Recovery,”SPERE (May 1992) 204.

73. Treiber, L.E., Archer, D.L., and Owens, W.W.: “A LaboratoryEvaluation of the Wettability of Fifty Oil-Producing Reservoirs,”SPEJ (December 1979) 531; Trans., AIME, 253.

74. Legens, C. et al.: “Carbonate Rock Wettability Changes Induced byOrganic Compound Adsorption,” Journal of Pet. Sci. and Eng., (June1998) 20, No. 3–4, 277.

75. Ross, G.D. et al.: “The Dissolution Effects of CO2-Brine Systems onthe Permeability of U.K. and North Sea Calcareous Sandstones,”paper SPE/DOE 10685 presented at the 1982 SPE/DOE Symposiumon Enhanced Oil Recovery, Tulsa, 4–7 April.

76. Sayegh, S.G. et al.: “Rock/Fluid Interactions of Carbonated Brines ina Sandstone Reservoir: Pembina Cardium, Alberta, Canada,” SPEFE(December 1990) 399.

77. Graue, D.J. and Blevins, T.R.: “SACROC Tertiary CO2 Pilot Project,”paper SPE 7090 presented at the 1978 SPE Symposium on ImprovedMethods for Oil Recovery, Tulsa, 16–19 April.

78. Shelton, J.L. and Schneider, F.N.: “ The Effects of Water Injection onMiscible Flooding Methods Using Hydrocarbons and CarbonDioxide,” SPEJ (June 1975) 217.

79. Yelling, W.F.: “Carbon Dioxide Displacement of a West TexasReservoir Oil,” SPEJ (December 1982) 805.

80. Negahban, N., Shiralkar, G.S., and Gupsta, S.P.: “Simulation of theEffects of Mixing in Gasdrive Core Tests of Reservoir Fluids,” SPERE(August 1990) 402.

81. van Lingen, P.P. and Knight, S.: “Evaluation of Capillary Entrapmentwithin Reservoir Flow Units,” paper SPE 38934 presented at the 1997SPE Annual Technical Conference and Exhibition, San Antonio,Texas, 5–8 October.

82. Arriols, A. et al.: “Trapping of Oil Drops in a Non-Circular PoreThroat,” paper SPE 9404 presented at the 1980 SPE Annual TechnicalConference and Exhibition, Dallas, 21–24 September.

83. Wegener, D.C. and Harpole, K.J.: “Determination of RelativePermeability and Trapped Gas Saturation for Predictions of WAGPerformance in the South Cowden CO2 Flood,” paper SPE 35429 pre-sented at the 1996 SPE/DOE Symposium on Improved Oil Recovery,Tulsa, 21–24 April.

84. Jerauld, G.R.: “General Three-Phase Relative Permeability Model forPrudhoe Bay,” SPERE (November 1997) 255.

85. Rampersad, P.R. et al.: “Impact of Trapping of Residual Oil by MobileWater on Recovery Performance in Miscible Enhanced Oil RecoveryProcesses,” paper SPE 29563 presented at the 1995 SPE RockyMountain Regional/Low-Permability Reservoir Symposium, Denver,20–22 March.

86. Lin, E.C. and Huang, E.T.S.: “The Effect of Rock Wettability on WaterBlocking During Miscible Displacement,” SPERE (May 1990) 205.

87. Müller, T. and Lake, L.W.: “Theoretical Study of Water Blocking inMiscible Flooding,” SPERE (November 1991) 445.

88. Walsh, M.P., Negahban, S., and Gupta, S.P.: “An Analysis of WaterShielding in Water-Wet Porous Media,” paper SPE 19652 presented atthe 1989 SPE Annual Technical Conference and Exhibition, SanAntonio, Texas, 8–11 October.

89. Dai, K.K., and Orr, R.M. Jr.: “Prediction of CO2 Flood Performance:Interaction of Behavior With Microscopic Pore StructureHeterogeneity,” SPERE (November 1997) 531.

90. Prieditis, J. and Brugman, R.J.: “Effects of Recent RelativePermeability Data on CO2 Flood Modeling,” paper SPE 26650 pre-sented at the 1993 SPE Annual Technical Conference and Exhibition,Houston, 3–6 October.

91. Reinbold, E.W. et al.: “Early Performance of the KuparukHydrocarbon Miscible Flood,” paper SPE 24930 presented at the 1992SPE Annual Technical Conference and Exhibition, Washington, DC,4–7 October.

92. Akin, S. and Demiral, M.R.B.: “Effect of Flow Rate on ImbibitionThree-Phase Relative Permeabilities and Capillary Pressures,” paperSPE 38897 presented at the 1997 SPE Annual Technical Conferenceand Exhibition, San Antonio, Texas, 5–8 October.

93. Berry, J.F., Little, A.J.H., and Skinner, R.C.: “Differences in Gas/Oiland Gas/Water Relative Permeability,” paper SPE 24133 presented atthe 1992 SPE/DOE Symposium on Enhanced Oil Recovery, Tulsa,22–24 April.

94. Dria, D.E., Pope, G.A., and Sepehrnoori, K.: “Three-PhaseGas/Oil/Brine Relative Permeabilities Measured Under CO2 FloodingConditions,” SPERE (May 1993) 143; Trans., AIME, 295.

95. Folger, L.K.: “Improved CO2 Flood Predictions Using 3D GeologicDescription and Simulation on the Sundown Slaughter Unit,” paperSPE 35410 presented at the 1996 SPE/DOE Enhanced Oil RecoverySymposium, Tulsa, 21–24 April.

96. Prieditis, J., Yang, A.P., and Wilkins, M.D.: “Simulation of a CO2 Floodin the Slaughter Field with Geostatistical Reservoir Characterization,”paper SPE 49168 prepared for presentation at the 1998 SPE AnnualTechnical Conference and Exhibition, New Orleans, 27–30 September.

97. Soucemarianadin, A. et al.: “Sweep Efficiency in Multi-LayeredPorous Media: Contrast Between Stable and Unstable Flow,” paperSPE 16955 presented at the 1987 SPE Annual Technical Conferenceand Exhibition, Dallas, 27–30 October.


98. Perez, G. and Chopra, A.K.: “Evaluation of Fractal Models To DescribeReservoir Heterogeneity and Performance,” SPEFE (March 1997) 65.

99. Pizarro, J.O.S. and Lake, L.W.: “Understanding Injectivity inHeterogeneous Reservoirs,” paper SPE 39697 presented at the 1998SPE/DOE Improved Oil Recovery Symposium, Tulsa, 19–22 April.

100. Arriols, A., Willhite, G.P., and Green, D.W.: “Mobilization of an OilDrop Trapped in a Non-Circular Pore Throat Upon Contact withSurfactants,” paper SPE 9405 presented at the 1980 SPE AnnualTechnical Conference and Exhibition, Dallas, 21–24 October.

101. Raimondi, P. and Torcaso, M.S.: “Mass Transfer Between Phases in aPorous Medium: A Study of Equilibrium,” SPEJ (March 1965) 51.

102. Shelton, J.L. and Yarborough, L.: “Multiple Phase Behavior in PorousMedia During CO2 or Rich-Gas Flooding,” JPT (September 1977) 1171.

103. Henry, R.L. and Metcalfe, R.S.: “Multiple-Phase Generation DuringCarbon Dioxide Flooding,” SPEJ (August 1983) 595.

104. Grigg, R.B. and Siagian, U.W.R.: “Understanding and ExploitingFour-Phase Flow in Low-Temperature CO2 Floods,” paper SPE 39790presented at the 1998 Permian Basin Oil and Gas RecoveryConference, Midland, Texas, 25–27 March.

105. Guzman, R.E. et al.: “Three-Phase Flow in Field-Scale Simulation ofGas and WAG Injections,” paper SPE 28897 presented at the 1994SPE European Petroleum Conference, London, 25–27 October.

106. Taber, J.J.: “Dynamic and Static Forces Required to Remove aDiscontinuous Oil Phase from Porous Media Containing Both Oil andWater,” SPEJ (March 1969) 3.

107. Harbert, L.W.: “Low Interfacial Tension Relative Permeability,” paperSPE 12171 presented at the 1983 SPE Annual Technical Conferenceand Exhibition, San Francisco, 5–8 October.

108. Thomas, F.B. et al.: “Miscible or Near-Miscible Gas Injection, Whichis Better?” paper SPE 27811 presented at the 1994 SPE/DOEImproved Oil Recovery Symposium, Tulsa, 17–20 April.

109. Haniff, M.S. and Ali, J.K.: “Relative Permeability and Low TensionFluid Flow in Gas Condensate Systems,” paper SPE 20917 presentedat the 1990 Europec, The Hague, 22–24 October.

110. Henderson, G.D. et al.: “Measurement and Correlation of GasCondensate Relative Permeability by the Steady-State Method,” paperSPE 30770 presented at the 1995 SPE Annual Technical Conferenceand Exhibition, Dallas, 22–25 October.

111. Klins, M.: Carbon Dioxide Flooding: Basic Mechanisms and ProjectDesign, Intl. Human Resources Development Corp., Boston,Massachusetts (1984) 113–115.

112. Chen, H.L., Wilson, S.D., and Monger-McClure, T.G.: “Determinationof Relative Permeability and Recovery for North Sea GasCondensate,” paper SPE 30769 presented at the 1995 SPE AnnualTechnical Conference and Exhibition, Dallas, 22–25 October.

113. Fulcher, R.A. Jr., Ertekin, T., and Stahl, C.D.: “Effect of CapillaryNumber and Its Constituents on Two-Phase Relative PermeabilityCurves,” JPT (February 1985) 249.

114. McDougall, S.R., Salino, P.A., and Sorbie, K.S.: “The Effect ofInterfacial Tension Upon Gas-Oil Relative PermeabilityMeasurements: Interpretation Using Pore-Scale Models,” paper SPE38920 presented at the 1997 SPE Annual Technical Conference andExhibition, San Antonio, Texas, 5–8 October.

115. Pope, G.A. et al.: “Modeling Relative Permeability Effects in Gas-Condensate Reservoirs With a New Trapping Model,” SPEREE (April2000) 171.

116. Chun, B.-S. and Wilkinson, G.T: “Interfacial Tension in High-PressureCarbon Dioxide Mixtures,” Ind. Eng. Chem. Res. (December 1995)34, No. 12, 4371.

117. da Rocha, S.R.P. et al.: “Effect of Surfactants on the InterfacialTension and Emulsion Formation Between Water and CarbonDioxide,” Langmuir (1999) 15, No. 2, 419.

118. O’Neill, M.L. et.al: “Solubility of Homopolymers and Copolymersin Carbon Dioxide,” Ind. Eng. Chem. Res. (August 1998) 37, No.8, 3067.

119. Kamath J. et al.: “Laboratory Investigation of Injectivity LossesDuring WAG in West Texas Dolomites,” paper SPE 39791 presentedat the 1998 SPE Permian Basin Oil and Gas Recovery Conference,Midland, Texas, 25–27 March.

SI Metric Conversion Factorsbbl 1.589 873 E �01 � m3

dyne 1.0* E �02 � mNft 3.048* E �01 � m�F (�F �32)/1.8 � �C

psi 6.894 757 E �03 � MPa

*Conversion factor is exact.

John D. Rogers is currently a petroleum engineer/project man-ager on the drilling, completion, and stimulation team and thehydrate resource technology team of the Gas Supply ProjectsDivision at the U.S. DOE/NETL facility in Morgantown, WestVirginia. e-mail: [email protected]. Previously, heworked at the Petroleum Recovery Research Center of theNew Mexico Inst. of Mining and Technology in Socorro, NewMexico as a senior research associate, studying aspects ofEOR applied in conventional and naturally fractured reservoirs.Before obtaining graduate degrees, Rogers worked 8 years forAmoco Production Co. as a petroleum production engineer,optimizing various Permian Basin waterfloods. His current areasof interest are advanced drilling systems, improved recoverysystems to maximize hydrocarbon recovery, and productionand use of unconventional natural gas resources. Rogers holdsan MS degree in petroleum engineering from Texas Tech U.and BS and PhD degrees, both in chemical engineering, fromNew Mexico State U. Reid Grigg is a senior engineer and headof the Gas Flooding Process and Flow Heterogeneities Sectionof the New Mexico Petroleum Recovery Research Center, adivision of the New Mexico Inst. of Mining and Technology inSocorro, New Mexico. e-mail: [email protected]. Previously,he worked for Conoco Production Research in Ponca City,Oklahoma, and Core Laboratories in Dallas. His areas of inter-est are phase behavior of high-pressure fluid, flow in porousmedia, gas hydrates, and gas injection for improved oilrecovery. Grigg holds BS and PhD degrees in chemistry fromBrigham Young U.

SPEREE


A Literature Analysis of the WAG Injectivity Abnormalities in the CO2 Process

Documents

Transcript of A Literature Analysis of the WAG Injectivity Abnormalities in the CO2 Process