iptc-17756-ms

16
IPTC-17756-MS Modelling Gas Injection for Enhanced Oil Recovery in Deepwater Fields John Lagasca, Robin Ozah, and Deniz Dindoruk, Shell International E&P Inc. Copyright 2014, International Petroleum Technology Conference This paper was prepared for presentation at the International Petroleum Technology Conference held in Kuala Lumpur, Malaysia, 10 –12 December 2014. This paper was selected for presentation by an IPTC Programme Committee following review of information contained in an abstract submitted by the author(s). Contents of the paper, as presented, have not been reviewed by the International Petroleum Technology Conference and are subject to correction by the author(s). The material, as presented, does not necessarily reflect any position of the International Petroleum Technology Conference, its officers, or members. Papers presented at IPTC are subject to publication review by Sponsor Society Committees of IPTC. Electronic reproduction, distribution, or storage of any part of this paper for commercial purposes without the written consent of the International Petroleum Technology Conference is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of where and by whom the paper was presented. Write Librarian, IPTC, P.O. Box 833836, Richardson, TX 75083-3836, U.S.A., fax 1-972-952-9435 Abstract Enhanced oil recovery using miscible and partially miscible gas injection processes have been utilized extensively and have been successfully applied in a significant number of mainly onshore reservoirs. Miscible and partially miscible gas injection is the most mature technology of all the EOR processes. The fact that the incremental recoveries may be significant makes it very attractive to assess and potentially deploy gas injection technology and overcome the challenges in offshore deep water settings. In this paper, we present a technique which uses established methods that simulate gas flood performance. The methods used in the study allow the use of a pseudo-miscible-black oil method in full field reservoir simulation models for initial scoping purposes. This avoids computationally expensive flash calculations resulting in a fast method to develop production profiles which allow the testing of the economic robustness of the projects. While simulations using the Todd and Longstaff (T-L) pseudo-black oil method are strictly for first contact miscible (FCM) processes, in essence, most of the published field applications for this pseudo- miscible technique have been for multi-contact miscible (MCM) processes. The T-L pseudo miscible method utilizes a mixing parameter, o that ranges from zero to one, simulating completely immiscible and first contact miscible floods respectively. A literature review indicated that models matched using field production rates yield mixing parameters within a range of 0.6-0.8, while models matched against only a corresponding compositional model yield mixing parameters closer to 1.0, a condition which assumes no fingering in a sub-grid block level (Bronchalo, et al, 2004). In order to evaluate the potential of a miscible gas flood, it was required to generate reliable production forecasts. This study was embarked upon in order to establish that simulation generated forecasts could be validated. This study utilized homogenous 2D models (box models), heterogeneous sector models with fine gridding and coarsely gridded full scale models to evaluate the impact of the mixing parameter at different scales. A fully compositional simulation was used to benchmark the results all of the pseudo- miscible modelling exercise. For the simple box models, it was found that thermodynamic effects were dominant in the determination of the mixing parameter. Below the minimum miscibility pressure (MMP), a lower mixing parameter value was required to achieve a match with a corresponding compositional model. As a result of testing the various pressure conditions which may exist in the reservoir a pressure-dependent mixing parameter has been able to be defined.

Upload: swaala4real

Post on 22-Dec-2015

6 views

Category:

Documents


0 download

DESCRIPTION

WAG Injection

TRANSCRIPT

Page 1: IPTC-17756-MS

IPTC-17756-MS

Modelling Gas Injection for Enhanced Oil Recovery in Deepwater Fields

John Lagasca, Robin Ozah, and Deniz Dindoruk, Shell International E&P Inc.

Copyright 2014, International Petroleum Technology Conference

This paper was prepared for presentation at the International Petroleum Technology Conference held in Kuala Lumpur, Malaysia, 10–12 December 2014.

This paper was selected for presentation by an IPTC Programme Committee following review of information contained in an abstract submitted by the author(s).Contents of the paper, as presented, have not been reviewed by the International Petroleum Technology Conference and are subject to correction by the author(s).The material, as presented, does not necessarily reflect any position of the International Petroleum Technology Conference, its officers, or members. Paperspresented at IPTC are subject to publication review by Sponsor Society Committees of IPTC. Electronic reproduction, distribution, or storage of any part of this paperfor commercial purposes without the written consent of the International Petroleum Technology Conference is prohibited. Permission to reproduce in print is restrictedto an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of where and by whom the paperwas presented. Write Librarian, IPTC, P.O. Box 833836, Richardson, TX 75083-3836, U.S.A., fax �1-972-952-9435

Abstract

Enhanced oil recovery using miscible and partially miscible gas injection processes have been utilizedextensively and have been successfully applied in a significant number of mainly onshore reservoirs.Miscible and partially miscible gas injection is the most mature technology of all the EOR processes. Thefact that the incremental recoveries may be significant makes it very attractive to assess and potentiallydeploy gas injection technology and overcome the challenges in offshore deep water settings.

In this paper, we present a technique which uses established methods that simulate gas floodperformance. The methods used in the study allow the use of a pseudo-miscible-black oil method in fullfield reservoir simulation models for initial scoping purposes. This avoids computationally expensiveflash calculations resulting in a fast method to develop production profiles which allow the testing of theeconomic robustness of the projects.

While simulations using the Todd and Longstaff (T-L) pseudo-black oil method are strictly for firstcontact miscible (FCM) processes, in essence, most of the published field applications for this pseudo-miscible technique have been for multi-contact miscible (MCM) processes. The T-L pseudo misciblemethod utilizes a mixing parameter, �o that ranges from zero to one, simulating completely immiscibleand first contact miscible floods respectively. A literature review indicated that models matched usingfield production rates yield mixing parameters within a range of 0.6-0.8, while models matched againstonly a corresponding compositional model yield mixing parameters closer to 1.0, a condition whichassumes no fingering in a sub-grid block level (Bronchalo, et al, 2004).

In order to evaluate the potential of a miscible gas flood, it was required to generate reliable productionforecasts. This study was embarked upon in order to establish that simulation generated forecasts couldbe validated. This study utilized homogenous 2D models (box models), heterogeneous sector models withfine gridding and coarsely gridded full scale models to evaluate the impact of the mixing parameter atdifferent scales. A fully compositional simulation was used to benchmark the results all of the pseudo-miscible modelling exercise. For the simple box models, it was found that thermodynamic effects weredominant in the determination of the mixing parameter. Below the minimum miscibility pressure (MMP),a lower mixing parameter value was required to achieve a match with a corresponding compositionalmodel. As a result of testing the various pressure conditions which may exist in the reservoir apressure-dependent mixing parameter has been able to be defined.

Page 2: IPTC-17756-MS

For the purposes of this study a specific geologic facie was simulated using small grid cells was in botha fully compositional and a pseudo-black oil simulator. The mixing parameter as a function of pressurethat resulted from the 2D models was reduced by the theoretical limit derived from Fayers (1992) wassufficient to produce a match in the fine scale model. This method also resulted in a sufficient match inthe coarse scale simulated grid that models the entire reservoir.

IntroductionDynamic model simulations of gas injection processes, ranging from first-contact (FCM), multiple-contact (MCM) and, partially miscible or immiscible floods, have been modeled historically by repre-senting the PVT behaviour through the use of compositional, pseudo-miscible and modified black oilmethods. Production forecasting methods using compositional and pseudo-black oil simulators arecompared in this study in order to understand difference in the output between them. We use Composi-tional and Todd-Longstaff methods in our work. Todd-Longstaff proposed an empirical method to modelmiscible flow by modifying the black oil properties. The method employs a mixing parameter, �, torepresent the degree of mixing of oil and solvent in a grid. Dispersion effects for coarse gridded modelsare captured by �. As T-L based production forecasts require less simulation run times this comparisonwas necessary to validate the selected modelling technique. Production forecasts are required in anintegrated reservoir, production and facility design and optimization and, will affect key results rangingfrom reserves to decisions regarding facility design and compressor sizing, well placement and patternselections to name a few. Industry has generally concluded that if fine scale heterogeneity is thecontrolling mechanism for recovery, black oil simulators are more appropriate while if phase behavior ismore significant, then compositional methods are likely to be more accurate. The mixing parameter canbe used either as a history matching parameter or as a parameter that can be used to investigate the degreeof uncertainty in the production forecasts. In their original paper, Todd and Longstaff (1972) obtained ahistory match using an omega value of 2/3. Generally, industry prefers to use a value between 0.5-0.8(Brinkman et al, 1998, Hallam et al, 1995, Nishikiori et al, 2008, Whitten, 1989, Youngren and Charlson,1980, Maclean, 1989 and Al Shammari et al, 2012) if there are no other available data.

Literature ReviewTodd and Longstaff (1972) aimed to represent unstable frontal advance through the application of amixing parameter, �, which accounts for the inherent viscous instability, geological heterogeneity andother dispersive forces that are lost through the use of reservoir simulation scale grids. The mixingparameter also adjusts the density and the viscosity of the solvent and resident oil mixture. Neglectingthese dispersive forces will lead to the dampening of flow instabilities and lead to an optimistic forecast.The T-L models assume a system that will achieve FCM, through scaling of the reservoir rock and fluidvia the mixing parameter but this method does not account for any thermodynamic effects explicitly.While Todd and Longstaff (1972) have stated that these assumptions may “preclude the use of themiscible simulator for forecasting displacement where multiple contacts are required before miscibility isobtained”, many authors have reported excellent matches even in MCM applications. In contrast, authors(Bronchalo et al, 2004, Hui et al, 2006, and Stalkup, 1983) have expressed reservations on the use ofcompositional simulators due to their tendency to overestimate production as it neglects viscous fingeringunless extremely fine grids are used in Figure 1.

The mixing parameter can be a physical manifestation of a variety of phenomena, which may includebut is not limited to, viscous fingering, gravity override and/or tonguing, numerical dispersion, immis-cibility of solvent with resident oil or of chase gas to oil. A mixing parameter value of � � 0 will resultin a case of complete immiscibility and the solutions would be similar to an immiscible gas flood whilea mixing parameter of � � 1 will result in complete mixing. While there is no universally agreed uponmethod for the prediction of the mixing parameter, a comparison of the Koval method (analogous to the

2 IPTC-17756-MS

Page 3: IPTC-17756-MS

Buckley-Leverett method but adjusted for an unstable miscible flood) yields a theoretically boundingmaximum value that depends only on the mobility ratio between the solvent and the oil (Fayers, 1992,Blunt et al, 1994, and Blunt and Christie, 1993) (Eq. 1). Some authors (Fayers, 1992, Warner, 1977 andBronchalo et al, 2004) have proposed several adjustments based on the overall heterogeneity of thereservoir (Eq. 2 and 3) or based on the fluid properties. As the reservoir of interest is considered to behighly heterogeneous in nature these were not used in this study as they do not give realistic values forhighly heterogeneous porous media.

(1)

(2)

(3)

Several studies involving the T-L scheme have been done previously and are summarized in Table1(appendix).

Figure 2 is a cross plot of the mixing parameter used by various papers in Table 1 (History Match,x-axis) versus the mixing parameter that would have been predicted by Equation 1 based on the provided

Figure 1—Illustration of Viscous Instability for Different Models (Bronchalo et al, 2004).

Figure 2—Comparison of the History Matched and Predicted Mixing Parameters

IPTC-17756-MS 3

Page 4: IPTC-17756-MS

fluid data. Stalkup (1983) suggests that a value of 0.5-0.7 should be used as a first estimate. As Figure 2indicates, there is a narrow band between the history matched values of 0.6-0.8 that may lead credenceto this suggested range. Coarsely gridded compositional models will generally reduce viscous instabilityand attempting to match a compositional model with its corresponding T-L model will lead to a mixingparameter closer to one under stable FCM conditions (Bronchalo et al, 2004).

Methodology and Application to Field “A”The T-L or pseudo-black oil model is usually much faster computationally compared to a fully compo-sitional simulator. However, there is significant uncertainty on the value of the mixing parameter thatshould be used, especially if it involves an MCM displacement or if the reservoir pressure falls below theMMP. This study aims to develop a simple method to determine the appropriate mixing parameter for agiven pressure using 2-D models and the correlation of Fayers (1992).

In order to arrive at the appropriate omega value and corresponding dependence with pressure thefollowing procedure was used. First, a simple 2-D homogeneous model at different pressures wassimulated using the T-L method with the aim of matching an equivalent compositional simulation. Themixing parameter and the residual oil saturation to miscible flooding (Sorm) were adjusted in order tomatch the resultant oil production forecasts between the two models. At pressures at or above the MMP,a mixing parameter approaching 1.0 was required to achieve a match between the two models. This wasconsidered too optimistic. However, if the Koval equation (Fayers, 1992) can be considered the uppertechnical limit for the T-L parameter, then it was hypothesized that all of the mixing parameters shouldbe scaled by the Koval factor. This would ensure that that the upper technical limit is honored at abovethe MMP. This was tested in both a fine scale three-dimensional sector model and a coarse grid modelfor a reservoir.

Table 2 summarizes the sub-sector and the full field properties for Field “A”. For this study, nitrogenwas assumed to be the miscible injectant. The PVT characterization was achieved by regression of alumped 7-component model with a standard primary depletion experimental set including constantcomposition expansion, differential liberation experiments and separator tests. Using the method ofextrapolation recommended by Orr (2007), it was estimated that the MMP is about 20,000 psia. Since theinjection pressure is limited by available compression discharge pressures which allow for 15,000 psiainjection at the sandface, the process is not miscible. It is thought that this Nitrogen MMP of 20,000 psimay be overstating the actual value; no experiments were available to tune the data and validate thisestimate. However, regardless of the actual value of the MMP, the subsequent workflow and the insightsremain valid in this study.

Homogeneous 2D and 3D Box ModelingThere are three dominant parameters impacting forecast production performance and oil recovery in themodelling exercise. These are namely pressure (thermodynamic effects), density number and grid size

Table 2—Sub-Sector and Full Field Summary

Property Sub-Sector Field-A Full Field

Reservoir Pressure 19,500 psia 19,358 psi

Reservoir Temperature 244 F 245-258 F

Average Permeability 20 mD 20 mD

Average Porosity 20% 18-20%

Oil Viscosity at Reservoir T and P 4.5 cP 3-5 cP

Oil Gravity 30.4 24.3 – 27.8 API

Gas Oil Ratio 181.2 scf/bbl 170 scf/bbl

Bubble Point Pressure 1054.57 psia 950-1014 psia

4 IPTC-17756-MS

Page 5: IPTC-17756-MS

effects. These parameters impact miscibility, gravity override and dispersion which all affect oil recovery.The miscibility effect was investigated by the use of isotropic 2D models and handled through theappropriate choice of omega. The gravity number and the grid size effects (vertical layers) wereinvestigated through the use of isotropic 3D models.

Investigation of the Pressure and Miscibility Dependent Nature of the Mixing Parameter Using a2-D Box ModelThe 2D fine gridded isotropic model used an average porosity and permeability within a 100 x 100 x 1grid blocks each of 1 m3 bulk volume and was initialized using the 7-pseudo component lumped PVT.Both the production and the injection wells were controlled through a constant bottomhole fluid rate, theaim being to equalise injection and withdrawal thus maintaining a relatively constant pressure until 1.0HCPV was injected. The mixing parameter and Sorm were varied until the cumulative oil production ofthe T-L model matched the compositional model.

Generally speaking, the T-L model forecast earlier gas breakthrough when compared to the compo-sitional model as illustrated in Figure 3. It can be seen that there is an earlier increase in gas productionwith a corresponding decline in oil production. Furthermore, the T-L model produces more oil late in lifeas compared to the compositional model. The early breakthrough occurs due to the inability of the T-Lmodel to produce a sharp composition shock front illustrated in Figure 4.

The increased recovery in the T-L model post-breakthrough can be explained in the following manner.A fundamental concept in the T-L formulation is that the gas injectant maintains a state of first contactmiscibility with the resident oil As such there is no consideration for any PVT changes that may haveoccurred previously with the nitrogen vapourising the lighter components from the oil as occurs in theEOS model. Near the injector, the compositional model left behind high viscosity oil, while the remainingoil in the T-L formulation retains its original viscosity. As a result, later in life in the EOS model the oilis not so favourable to being vapourised resulting in lower oil recovery during this period as compared tothe T-L formulation.

Further analysis of the thermodynamic properties of the system illustrates the shortcomings of the T-Lmodel from a fundamental perspective. Figures 5 and 6 compare the density and the viscosity of acombined oil and solvent system above and below the FCM pressure. Figure 5 illustrates this behaviourfor a methane-oil system whilst Figure 6 compares this for a nitrogen-oil system. The principal differencebetween the two systems is that the methane-oil system is FCM whilst the nitrogen-oil system is not. Itcan be seen that the T-L method closely reproduces the FCM system, for the below FCM condition it doesnot fully reproduce the divergence of the two-phase properties.

Notwithstanding the mis-match in thermodynamic properties a match on the cumulative oil as per thecompositional model was made by varying the � and the Sorm values (Figure 7). While the thermody-namic properties cannot be rigorously matched, the mixing parameter can be decreased to promote earlier

Figure 3—Oil Rate and Gas Rate Comparison between Compositional and Black-Oil Models

IPTC-17756-MS 5

Page 6: IPTC-17756-MS

breakthrough and the Sorm can be increased to enable the model to leave behind a greater proportion ofoil thus mimicking the compositional production profile. In the plots that follow, matches were achievedbetween the compositional (heavy blue line) and the T-L match (thin red line) using an assisted historymatching program. All the other light blue lines are other possible outcomes for different combination ofparameters.

From these history matches, it was then possible to construct a plot of �o and Sorm versus pressure.From Figure 8, the mixing parameter became equal to 1.0 if the pressure was above the MMP. This

is reasonable since the level of dispersion modeled at the same grid block resolution should be the samebetween the compositional and the black oil models if their thermodynamic displacement behaviors areconsistent which was demonstrated in Figure 5. Above this MMP, the Sorm was required to be decreasedin order to match the EOS model predicted cumulative oil. As the pressure was decreased, the mixingparameter was also decreased to model the earlier gas breakthrough times, reflecting the viscous tonguespresent in a partially miscible flood.

Figure 4—Oil Saturation as a Function of Distance from the Producer Injector to the Producer

Figure 5—Comparison of Density and Viscosity Values for a FCM System (Methane and Field-A Oil)

6 IPTC-17756-MS

Page 7: IPTC-17756-MS

Investigation of the Effect of the Gravity Number on the Mixing ParameterMany miscible injection schemes are gravity unstable, meaning they are susceptible to gravity overridedue to the gravity forces dominating the viscous forces. This could result either through a low gas injectionrate, large well spacing or high vertical permeability. As a test, 3D models were run with varying increasesin vertical permeability that would change the gravity number by at least two orders of magnitude. Thecompositional model was run and as such provided a base production profile that the pseudo black-oilmodels were matched against. In this case, there is no functional dependence between the gravity numberand the mixing parameters as seen in Figure 9. It is possible that these are model dependent due to theexistence of many shale layers in the simulation.

It was significantly more difficult to get an acceptable match between the compositional and the T-Lmodel at high gravity numbers since the gravity effects exacerbate gas segregation as illustrated in Figure10. This suggests that the T-L model may not be a good model for systems where gravity override is amajor problem. While (Figure 9 suggests that it may be possible to achieve matches on the cumulative oilusing similar omega values, Figure 10 cautions that the in-situ saturation profile may not be accuratelymodeled when gravity forces dominate the simulation. This is exacerbated by the thermodynamiclimitations illustrated by Figure 6). The compositional model produced a heavy oil phase that segregateddownward while the T-L model tends to produce more mobile oil that can more easily migrate to theproducer. The result is a more uniform gas conformance in the T-L model compared to the compositionalmodel in Figure 10. A similar trend in the omega values as a function of the density number was foundin Warner (1997).

Validation of the Method Using Sector ModelsFigure 11 illustrates the proposed mixing parameter as a function of pressure for the reservoir propertiesassociated with Reservoir A. This was created by multiplying the matched 2D mixing parameter at variouspressures (Figure 8, blue line) and the theoretical maximum mixing parameter of 0.67 as predicted by Eq.1. The theoretical maximum assumes a first contact miscible process and does not account for fingeringdue to the immiscibility that occurs when the pressure is decreased. This is likely the main contributingreason to the lower � value that the EOS T-L matching process has yielded. This composite line is referredto as the combined or composite mixing parameter.

Figure 12 summarizes the resultant cumulative oil comparisons between the compositional and the T-Lforecast for the values chosen to be used in the T-L modelling. The case of ��0.0 is displayed to illustratean immiscible displacement between the oil and the solvent where the incremental oil production is due

Figure 6—Comparison of Density and Viscosity Values for a Partially Miscible System (Nitrogen and Field-A Oil)

IPTC-17756-MS 7

Page 8: IPTC-17756-MS

solely on pressure support. Similarly the technical limit is shown for a fully miscible case. It can be seenthat there was only a 3.3% difference in the final cumulative oil produced between the composite case andthe compositional model case.

The oil rate comparison plot is shown in Figure 13 while the gas rate comparison plot is shown inFigure 14. Several observations can be made from these plots. Qualitatively, the profiles were matchedadequately. Consistent with the 2D runs, the gas rate increased more rapidly for the T-L case comparedto the compositional model. After the initial gas breakthrough, the T-L slightly over performs over thecompositional since there was very limited stripping that occurs in the latter.

Figure 7—History Match of 2D Models at Different Pressures

8 IPTC-17756-MS

Page 9: IPTC-17756-MS

Full Field ModelingThe goal of this study was to demonstrate that a method to quickly find appropriate values for the mixingparameter and Sorm for use in the pseudo-miscible T-L model was possible and reliable. The sectormodeling above validated this method and therefore was then applied to field scale model. The pressure

Figure 8—Adjustable Parameters as a Function of Pressure

Figure 9—Mixing Parameter as a Function of the Density Number

Figure 10—Oil Saturation Profiles for the 3D Models

IPTC-17756-MS 9

Page 10: IPTC-17756-MS

dependent mixing parameter with the appropriate values of omega and Sorm above was then applied tothe full field model.

A potential field development and corresponding well layout is illustrated in Figure 15. Illustratedbelow is a possible field development employing injectors and producers as denoted by I and P wellsrespectively.

Figure 16 compares the cumulative oil profiles as forecast from a compositional and T-L formulation.The compositional and the composite T-L model were within 1.1% of each other in terms of cumulativeoil recovery forecast. Similar oil rate profiles can be found, as illustrated in Figure 17. The averagereservoir pressure of the T-L model was within 500 psia of the compositional model, as illustrated inFigure 18. Note that once again at late life, the compositional model has a lower production rate comparedto the T-L model due to the lack of late life stripping (Figures 17 and 18).

The impact of each of the elements of the composite mixing parameter can be identified. For example,any increase in oil recovery over that of pure depletion realized when the mixing parameter was reducedto zero thus mimicking an immiscible flood would be attributable mostly to pressure support while theremaining incremental oil was attributed to a combination of interfacial tension reduction, viscosityreduction and near-miscible displacement.

Figure 11—Mixing Parameter Comparisons

Figure 12—Cumulative Oil Comparisons between T&L Models and the Compositional Model

10 IPTC-17756-MS

Page 11: IPTC-17756-MS

Figure 13—Oil Rate Comparisons Between the T&L and Compositional Models

Figure 14—Produced Gas/Solvent Rate Comparisons Between the T&L and Compositional Models

Figure 15—Grid Top and Well Locations (ft.)

IPTC-17756-MS 11

Page 12: IPTC-17756-MS

ConclusionThe combination of methods used in this study allows the use of a pseudo-miscible-black oil method infull field models for initial ultimate oil recovery scoping purposes. Use of such avoids computationallyexpensive flash calculations associated with EOS models resulting in a fast method to develop productionprofiles which permits the testing of the economic robustness of potential development projects. It wasfound that the mixing parameters for different reservoir pressure conditions for the properties of ReservoirA through matching the compositional and pseudo black oil models in a 2D model can be directly usedin a full field model provided that it is decreased by the Koval technical limit.

Figure 17—Comparison of Oil Rate Profiles – Full Field Optimization Figure 18—Comparison of Ave. Reservoir Pressures (reservoir unit-1) –Full Field Optimization

Figure 16—Comparison of Cumulative Oil Profiles – Full Field Optimization

12 IPTC-17756-MS

Page 13: IPTC-17756-MS

ReferencesH. Al Shammari, B.B. Singh and K. Mishra, “Understanding the True Potential of a Major Reservoir

Sabriyah (Upper Burgan) of Kuwait Oil Company”, presented at the EAGE Annual Conferenceand Exhibition held in Copenhagen, Denmark, 4-7 June 2012.

H.L. Bilhartz, G.S. Charlson, F.I. Stalkup and C.C. Miller, “A Method for Projecting Full-ScalePerformance of CO2 Flooding in the Willard Unit”, presented at the Fifth Symposium onImproved Methods for Oil Recovery of the Society of Petroleum Engineers in Tulsa, Oklahomaon 16-19 April 1976.

F.P. Brinkman, T.V. Kane, R.R. McCullough and J.W. Miertschin, “Use of Full-Field Simulation toDesign a Miscible CO2 Flood”, presented at the 1998 SPE/DOE Improved Oil Recovery Sym-posium held in Tulsa, Oklahoma on 19-22 April 1998.

J. Bronchalo, R. Venkataraman and M. Blunt, “A Multiscale Methodology for Simulating MiscibleGas Injection Projects Applied to a North African Oil Field”, presented at the SPE AnnualTechnical Conference and Exhibition held in Houston, Texas, 26-29 September 2004.

S. Chugh, R. Baker, L. Cooper and S. Spence, “Simulation of Horizontal Wells to Mitigate MiscibleSolvent Gravity Override in the Virginia Hills Margin”, JCPT, February 2000, pp. 28–34.

F.J. Fayers, M. Haajizadeh, C.Y. Lin and J. Taggart, “Use of the 4-Component Todd and LongstaffMethod as an Upscaling Technique in Simulating Gas Injection Projects”, presented at the 2000SPE/DOE Improved Oil Recovery Symposium held in Tulsa, Oklahoma, 3-5 April 2000.

F.J. Fayers, M.J. Blunt and M.A. Christie, “Comparisons of Empirical Viscous-Fingering Models andTheir Calibration for Heterogeneous Problems”, SPE Reservoir Engineering, May 1992, pp.195–203.

M.J. Fox, V.N. Simlote and W.G. Beaty, “Evaluation of CO2 Flood Performance, Springer “A” Sand,NE Purdy Unit, Garvin County, OK”, presented at the SPE/DOE Fourth Symposium on EnhancedOil Recovery held in Tulsa OK, 15-18 April 1984.

R.J. Hallam, T.D. Ma and E.W. Reinbold, “Performance evaluation and optimization of the Kuparukhydrocarbon miscible water-alternating-gas flood”, New Developments in Improved Oil Recovery,1995, pp. 153–164.

J.R. Hervey and A.C. Iakovakis, “Performance Review of a Miscible CO2 Tertiary Project: RangelyWeber Sand Unit, Colorado”, SPE Reservoir Engineering, May 1991, pp. 163–168.

M.H. Hui, J. Kamath and K. Dehghani, “A Multi-Scale Approach to Modeling Miscible Gas InjectionSweep Efficiency in a Giant Carbonate, Light-Oil Reservoir”, presented at the 2006 Abu DhabiInternational Petroleum Exhibition and Conference held in Abu Dhabi, UAE, 5-8 November 2006.

D.J. Jethwa, B.W. Rothkopf and C.I. Paulson, “Successful Miscible Gas Injection in a Mature UKNorth Sea Field”, presented at the 2000 SPE Annual Technical Conference and Exhibition held inDallas, Texas, 1-4 October 2000.

M.L. Katz and F.I. Stalkup, “Oil Recovery by Miscible Displacement”, WPC Conference Paper.J.E. Killough and C.A. Kossack, “Fifth Comparative Solution Project: Evaluation of Miscible Flood

Simulators”, presented a the Ninth SPE Symposium on Reservoir Simulation held in San Antonio,Texas, 1-4 February 1987.

K.H. Lee and M.M. El-Saleh, “A Full-Field Numerical Modeling Study for the Ford Geraldine UnitCO2 Flood”, presented at the SPE/DOE Seventh Symposium on Enhanced Oil Recovery in Tulsa,Oklahoma, 22-25 April 1990.

D.A. MacLean, “Design of a field-wide hydrocarbon miscible flood for the Kaybob Beaverhill Lake“A” Pool”, JCPT, May-June 1989, pp. 20–28.

J.S. Moore and G.C. Clark, “History Match of the Maljamar CO2 Pilot Performance”, presented at theSPE/DOE Enhanced Oil Recovery Symposium held in Tulsa, Oklahoma, 17-20 April 1988.

IPTC-17756-MS 13

Page 14: IPTC-17756-MS

N. Nishikiori, K. Sugai, C. Normann, A. Onstein, O. Melberg and T. Eilertsen, “An IntegratedWorkflow for Gas Injection EOR and a Successful Application to a Heterogeneous SandstoneReservoir in the Southern North Sea”, presented at the International Petroleum TechnologyConference held in Kuala Lumpur, Malaysia, 3-5 December 2008.

S. Shoaib and B.T. Hoffman, “CO2 Flooding the Elm Coulee Field”, presented at the 2009 SPE RockyMountain Petroleum Technology Conference held in Denver, Colorado, 14-16 April 2009.

M.R. Todd and W.J. Longstaff, “The Development, Testing and Application of a Numerical Simulatorfor Predicting Miscible Flood Performance”, JPT, (July 1972), pp. 874–882.

H.R. Warner, “An Evaluation of Miscible CO2 Flooding in Waterflooded Sandstone Reservoirs”, JPT,October 1977, pp 1339–1347.

G.P Willhite, A.P. Byrnes, M.K. Dubois, R.E. Pancake, J.S. Tsau, J.R. Daniels and W.A. Flanders, “APilot Carbon Dioxide Test, Hall Gurney Field, Kansas”, presented at the 18th SPE Improved OilRecovery Symposium held in Tulsa, Oklahoma, 14-18 April 2012.

M.D. Whitten, J.H. Anderson and G.E. Tinker, “Miscible Flood from Inception: The Brassey FieldDevelopment, British Columbia, Canada”, presented at the 64th Annual Technical Conference andExhibition of the Society of Petroleum Engineers held in San Antonio, TX, 8-11 October 1989.

G.K. Youngren and G.S. Charlson, “History Match of the Little Creek CO2 Pilot Test”, JPT,November 1980, pp. 2042–2052.

R.S. Wu, J.P. Batycky and B.C. Harker, “A New Approach to Modeling Miscible Gas DisplacementTests”, presented at the 64th Annual SPE Technical Conference and Exhibition held in SanAntonio, TX, 8-11 October 1989.

14 IPTC-17756-MS

Page 15: IPTC-17756-MS

Appendix

Table 1—Summary of the Literature Survey

IPTC-17756-MS 15

Page 16: IPTC-17756-MS

Table 1 (Continued)—Summary of the Literature Survey

16 IPTC-17756-MS