Copyright 2005, Society of Petroleum Engineers This paper was prepared for presentation at the SPE International Improved Oil Recovery Conference in Asia Pacific held in Kuala Lumpur, Malaysia, 5–6 December 2005. This paper was selected for presentation by an SPE Program Committee following review of information contained in an proposal submitted by the author(s). Contents of the paper, as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material, as presented, does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Papers presented at SPE meetings are subject to publication review by Editorial Committees of the Society of Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper for commercial purposes without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to a proposal of not more than 300 words; illustrations may not be copied. The proposal must contain conspicuous acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O. Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

Abstract The Ekofisk fractured chalk reservoir located in the North Sea south-west of Norway has been exploited successfully for more than three decades, largely due to injection of sea water. In a study concluded in 2004, air injection was evaluated as a method for additional hydrocarbon recovery beyond the secondary waterflood recovery. Supported by the European Commission through the fifth framework program, the Ekofisk Field owners joined forces with leading European research institutes and a contractor to investigate the potential of air injection as a cost effective IOR method. Through screening studies, extensive laboratory experiments, reservoir simulations, design of processing facilities and project feasibility evaluations, an extensive knowledge base of the air injection process for light oil fractured reservoirs was established.

In the present paper technical results will be presented. Recovery mechanisms related to an air injection process in a fractured light oil reservoir have been studied through laboratory experiments and reservoir modeling. The laboratory experiments verified air injection as a potential IOR method for a light oil fractured chalk field. Laboratory experiments were performed in order to study kinetic properties such as activation energies and ignition temperatures. In addition, diffusion coefficients were estimated through laboratory experiments and verified by numerical simulations. Potential weakening of the chalk due to heat and CO2 was evaluated based on laboratory experiments and geo-mechanical modeling. Combustion tube experiments were conducted in order to study propagation of the combustion front through porous media.

Finally, a field scale air injection feasibility study was performed. The outcome of this study, including an evaluation of potential production benefits and main cost items involved in an air injection project, is presented.

Introduction The Ekofisk oil field, located in the North Sea south-west of Norway, is a fractured chalk reservoir containing 1.3 109 Sm3 oil equivalents. Oil production started in 1971. The chalk matrix has a porosity in the order of 25-40% with permeability ranging from 0.1 to 10 mD. Due to the fracture system, the effective permeability of the reservoir is in average approximately 20 mD. The initial reservoir temperature is 130 oC, while current reservoir temperature ranges from 30 oC in the vicinity of the water injection wells to 130 oC in the areas still not affected by the waterflood.

After a period of production decline, a water injection program was initiated in 1987, which has largely improved the hydrocarbon recovery from the field.

Several improved oil recovery techniques have been screened for application to increase the recovery above what is obtained by waterflooding1. Since air injection was evaluated to have a high potential for cost efficient recovery of additional hydrocarbons a study was started in 2001. Air injection has a potential for application in several chalk fields.

Air injection has been applied successfully in many offshore oil fields. Results being relevant to the Ekofisk case can be found in the literature from field studies2-5, analog field examples6-11 and various studies on the air injection process12-

19. However, there is no experience reported in open literature for air injection offshore in low permeable chalk reservoirs. It was thus necessary to perform in depth studies to evaluate the potential for application of this technology for the Ekofisk field. The present paper reports a study which was performed during 2001-2004 to evaluate the potential, including laboratory experiments on reaction kinetics, characterization of oxidation and combustion reaction, diffusion and rock compaction, calibration of simulation tools and simulation at reservoir scale, design and cost estimation of surface facilities, and economic evaluation.

Several uncertainties for the air injection process will have to be considered if it is going to be applied to Ekofisk. Will the process work in a waterflooded chalk reservoir? Will ignition occur? An eventual air breakthrough in the production wells will be a serious safety issue, and will have to be evaluated carefully. What will be the reservoir sweep efficiency? Can the process be modeled with sufficient accuracy at reservoir scale? Also, weakening of the chalk matrix will have to be considered, mainly caused by the temperature and the production of CO2. A consequence of chalk weakening can be increased compaction and subsidence, also potentially causing well failures. Will separation of flue gas from the sales gas be

SPE 97481

Evaluation of Air Injection as an IOR Method for the Giant Ekofisk Chalk Field S. Stokka, SPE, RF-Rogaland Research; A. Oesthus, SPE, ConocoPhillips; and J. Frangeul, TOTAL

2 SPE 97481

economically feasible? These aspects were addressed in the present work.

The Air Injection Process The air injection process is illustrated in Figure 1, as it is traditionally understood. Air is injected from the left in the figure. Behind the combustion zone there is a burned zone, while ahead of it there is an evaporation zone containing steam, nitrogen, hydrocarbon gases and combustion gases. Ahead of the evaporation zone is the condensation zone and then follows the water bank, oil bank and the unswept zone. The flue gas and steam generated at the combustion front are stripping, swelling and heating the contacted oil. A total consumption of 5-10% of the remaining oil in place is normally expected to maintain a propagation of the in-situ oxidation process.

In a fractured reservoir, a diffusion process is required for the air to enter the rock matrix, and the process itself might also behave somewhat different from the illustration in Figure 1. Further, the temperature increase being important for heavy oil reservoirs, will probably be less important in the Ekofisk case, as the reservoir contains light oil.

Project Tasks, Definitions and Theory The work was organized in five work packages: 1. Field pilot screening, including selection of an eventual

pilot area for a demonstration project, and evaluation of critical process factors and uncertainties. Sensitivity simulations were performed to determine which sets of parameters have the greatest influence on air-injection efficiency and safety.

2. Supporting experiments, gathering quantitative data on reaction kinetics at reservoir conditions, gas flow in a three dimensional fractured medium, and effect of the air injection process on rock properties.

3. Reservoir simulator, including calibration of the simulation tool for accurate modeling of the air injection oxidation process, thermal conditions and flow in a fractured low permeable medium.

4. Surface facilities, including evaluation of technical and economic issues of air injection and flue gas downstream handling, and selection of the most suitable flue gas handling technology.

5. Economic evaluation on field scale, including production forecasts, capital and operational expenses and risk evaluation.

From the experiments it was important to estimate the activation energy for the oxidation process. The rate of oxygen consumption is given by the equation:

Rate = dt

dn 2O = Ko .e -E/RT.(Fuel)m.(PO2)n (1)

Where KO is the Arrhenius pre-exponential factor, E the activation energy per mol, R the gas constant, T the temperature measured in Kelvin, Fuel the fuel concentration, PO2 the oxygen partial pressure, m and n the reaction orders relative to fuel concentration and oxygen partial pressure (assumed to be equal to one).

When oxygen is introduced to a rock containing oil, low temperature oxidation will start after a time depending of the activation energy. High activation energy gives longer time for starting the oxidation. As oxidation goes on at adiabatic conditions, the temperature will increase until autoignition takes place and the reaction is accelerated into a high temperature oxidation process. Results Field Pilot Screening. A suitable area for an air injection pilot was located on the east flank of the main Ekofisk field, see Figure 2. Preliminary evaluations based on simulations for a pilot area sector model gave promising results, and indicated that as much as 25% of the remaining oil after water flooding could potentially be recovered by the air injection process.

Sensitivities were run with respect to reservoir parameters believed to be essential or critical to the effectiveness and safety of the air injection process. These included effective permeability, anisotropy and pre-exponential factors. The results expose that, within variations of these parameters normally found in the Ekofisk reservoirs, air injection may be a safe and attractive oil recovery mechanism after the water flooding has been exploited to its maximum. The simulations were done in a single porosity mode. Consequently, the results are only indicative to how the dual porosity and dual permeability reservoir will behave. Supporting Experiments Oxidation experiments on crushed core. The oxidation kinetics of Ekofisk crude oil was investigated in the presence of brine and chalk, conducting three types of experiments19. One set of experiments was conducted in a small batch reactor, having a nominal liquid charge capacity of 100 ml, and a similar air header volume. Various saturations of Ekofisk oil and brine were introduced in crushed Ekofisk core material filling the small batch reactor. The experiments were operated isothermally, in the temperature range 130 – 150 ºC, and at a total pressure of 300 bars. The rate of oil oxidation, or rate of oxygen consumption, was determined by measuring the rate of decline of the total pressure versus time. The rate of reaction was influenced significantly by the temperature, and initial saturations of oil and brine. Ekofisk crude oil was sufficiently reactive in the temperature range investigated, that the oxygen charged into the reactor was completely consumed over a number of days.

Another set of experiments were also done on crushed Ekofisk core material containing various ratios of Ekofisk oil and brine saturations. These experiments were run in an accelerating rate calorimeter with continuous injection of air at 300 bars pressure. The calorimeter volume was charged with a few grams of Ekofisk chalk and a fraction of ml of oil and brine. The calorimeter was operated adiabatically and self-heat rate of the crude oil during oxidation was monitored. The experiments gave information of the Ekofisk crude oil reaction features. It was observed that the time between the initial onset of the oxidation reaction and autoignition was reduced with increasing initial oil saturation. Autoignition started in the temperature range 155 – 200 oC.

A third set of experiments were done on crushed limestone containing various ratios of Ekofisk oil and brine saturations.

SPE 97481 3

The experiments were run in a combustion tube cell with continuous injection of air at 200 bars pressure. The combustion tube was 125 cm long and had a diameter of 10 cm. Experiments were done both at adiabatic and non-adiabatic conditions. The produced fluids were separated, and gas flow rate, CO and CO2 production, and O2 combustion were monitored. The non-adiabatic tests showed that stable combustion front propagation was achieved at sufficiently high air injection rate. Oil recovery was dependent on air injection rate.

The experimental results from the small batch reactor and the accelerating rate calorimeter were used as screening data, in order to guide the selection of parameters for more detailed experiments on Ekofisk core. The combustion tube tests were conducted primarily to provide a one–dimensional physical simulation of the air injection process, against which selected reaction models of the crude oil oxidation could be validated.

Oxidation experiments on Ekofisk core. Oxidation kinetics of Ekofisk crude oil in Ekofisk core was investigated in the presence of brine, conducting three types of experiments. Key parameters to be used in computer simulation of the air injection process were measured. In one experiment the rate of diffusion of nitrogen and oxygen into a matrix plug was measured under reservoir conditions of pressure, temperature and fluid saturations. Such experiments are important to quantify the gas-gas diffusion between the reservoir fractures and matrix. The experimental setup is illustrated in Figure 3. A 20 cm3 pore volume cylindrical chalk sample was mounted into a steel reactor and was sealed except at the top of the core. A narrow flow space, the diffusion chamber, allowed for flow of air past the top side of the plug. Synthetic air, composed of 20% O2 and 80% N2, was injected through the diffusion chamber. The effluent from the chamber was continuously analyzed, thus to determine the amount and type of hydrocarbons diffusing out of the core as well as the amount of O2 and N2 diffusing into the core.

The matrix plug temperature was kept close to 80 oC, and the air was injected at a constant pressure of 275 bars. The pressure was maintained by a back-pressure valve located downstream. The air injection rate was constant during the experiment at a value of 0.2 cm3/hr (pump conditions: 45°C, 275 bars).

The molar fractions of produced O2, N2 C1, C2 and C3 are shown in Figure 4, and the molar fractions of and C4, C5 and C6, are shown in Figures 5. As expected the O2 and N2 concentrations increase with time while the hydrocarbon component fractions decrease.

The experiment was simulated using diffusion coefficients based on kinetic theory. See Table 1. These results were essential for the understanding of the air injection process and formed a basis for the field simulations.

Another set of experiments was performed at simulated reservoir conditions using Ekofisk core plugs to characterize the kinetics of the oxidation and combustion reactions. The experimental setup, called an adiabatic disc reactor, is illustrated in Figure 6. An Ekofisk core plug with a volume of approximately 20 cm3 was sealed inside the reactor. Synthetic air (20% oxygen, 80% nitrogen) was injected at constant rate and constant pressure of 276 bars from top to bottom. At the

reactor inlet air was preheated to the internal core temperature to avoid thermal losses by gas convection.

In the first phase of the experiment the temperature was increased in steps via heaters wired around the core. At each step the oxygen consumption and exothermic response of the core was measured. The step-wise increase was continued until the core temperature started increasing by itself. During the second phase the core temperature evolved freely and the system was controlled to be at adiabatic conditions. Experiments were done with cores containing brine and recombined Ekofisk oil.

Results from experiments with different initial water saturations are shown in Figures 7 and 8. The temperature for start of auto-ignition increases with increasing initial water saturation, and varies in the range 160 – 350 oC. The cumulative oxygen consumption shown in Figure 8 verifying that oxidation takes place also before self-ignition starts.

The experiments were simulated, and it was possible to estimate the activation energies both for the oxidation reaction and the combustion reaction. The average values are given in Table 2. These values were used in the field simulations. It was observed that the self-ignition temperature increased with increasing water saturation. Indeed, if no water is present in the system, the heat released is transferred to the rock and the oil. However, when water is present, for the same heat rate released by the reactions, the heat transferred to the oil is smaller since part of it is transferred to the water.

A set of long core experiments was performed to characterize the efficiency of flue gas sweeping, the velocity and temperature of the combustion front and the air requirement to sustain a stable front. Three experiments were performed; first an air flood at reservoir conditions, then an isothermal flue gas sweeping and finally a high temperature adiabatic air flood. The experiments were performed on a composite core (20 plugs) being 85.2 cm long and having a cross section of 10.7 cm2. The core was sealed to the core holder along its length. The two first experiments were performed at 130 oC temperature and 275 bar pressure. In the third experiment the first zone was heated to 350 oC.

The reservoir core plugs showed micro-fractures after mounting in the core holders. To some extent this masked the interpretation of the experiments, but important results were obtained. During the air flood at reservoir conditions, no increase of temperature was observed, but the entire injected oxygen was consumed inside the core. The cumulative oil production and nitrogen content in the produced gases of the two first gas sweeps are shown in Figure 9. Flue gas sweep efficiency was estimated to 9 % of the hydrocarbon pore volume after 2.15 pore volume flue gas injection. It was not possible to sustain a stable combustion front in the core after the flue gas sweeping, mainly due to too high water saturation in the first zones and very low core permeability. The entire injected oxygen was consumed inside the core.

Flue gas flooding test. A flue gas flooding test was carried out on a composite Ekofisk core (six plugs) at simulated reservoir conditions. A temperature of 60 oC and pressure of 320 bars were chosen for the experiment, and the initial water saturation was 70%. Nitrogen was injected from the top of a vertically mounted core and was allowed to pass outside the core but inside the sleeve, thus simulating nitrogen flow

4 SPE 97481

through a reservoir fracture. Hydrocarbons in the core could only be produced through diffusion. The experiment demonstrated that the lighter hydrocarbons were stripped from the reservoir oil originally present in the matrix.

Subsidence and compaction. Rock mechanics experiments and subsidence and compaction modeling were performed. Two combusted Ekofisk core samples were used in pore scale analysis and rock mechanics studies. The samples came from the experiments performed in the diffusion chamber and the adiabatic disc reactor, and had been exposed to temperatures of 80 oC and 490 oC, respectively. A combination of electron microscopy and mercury porosimetry was use to characterize the grain surfaces and texture.

Figure 10 shows results from constrained pore volume compressibility measurements on combusted and uncombusted plugs, indicating that combustion has weakened the chalk. It was concluded that low temperature oxidation would not lead to any increased compaction and subsidence. For the high temperature case, the field center could subside an additional 5 m at most, with a corresponding reservoir compaction of up to 6 m. The high temperature combustion process did not alter pore geometry by melting or destroying grains other than depositing carbon rich coating on the grains. The estimated overall field subsidence and combustion is not considered to be prohibitive for an eventual air injection project at the Ekofisk field. However, near wellbore effects impacting well failures are still considered to be a serious concern.

Reservoir Simulator. In order to reduce complexity without losing accuracy the reservoir fluid thermodynamics was adapted to a 7 component fluid model for air injection. It was compared with a fluid model with 15 components, and the 7 component model reproduced the phase behavior of the 15 component model with reasonable accuracy. It was necessary to use a reservoir simulator that could handle dual porosity, gas-gas diffusion, and in-situ oxidation and combustion18.

The contact of the reservoir fluid with air (79% N2, 21% O2) and flue gas (85% N2 and 15% CO2) was simulated thermodynamically at various temperatures. The results show that the oil stripping effect is greater than the swelling effect, as the oil density increases with increased flue gas contact with the oil.

Phenomenological simulation results showed that diffusion has a great impact on recovery. The light fractions of the matrix oil are stripped when air enters into the matrix by gravity drainage and diffusion. The vaporized oil is recovered very quickly in the fracture without oxidation. Air mainly oxidizes the heavy hydrocarbons left in the matrix. Results from mechanistic simulation runs for a two dimensional model with 500 m distance between injector and producer are shown in Figure 11, clearly showing the importance of diffusion and gravity segregation for the oil production rate.

Simulations of the diffusion experiment and the adiabatic disc reactor experiment, were essential for the air injection process understanding and for the field scale simulations. The activation energies for low temperature oxidation and high temperature combustion were tuned to fit with the experimental results. The diffusion coefficients used were

based on kinetic theory of gases. Parameters used in the simulations are listed in Tables 1 and 2.

Simplified reservoir scale modeling (same two dimensional model as mentioned above) led to the conclusion that there is a prevailing probability that the Ekofisk oil will commence in-situ oxidation when contacted by injected air, even in the low temperature environment in the vicinity of water injectors. It is unlikely that oxygen breakthrough will occur in the producing wells. The importance of diffusion and segregation for this result is shown in Figure 12. (The first part of the curve is believed to be masked by a modeling artifact.)

In the Ekofisk case the contribution to oil recovery by oxidation itself (heat and steam generation etc.) is probably insignificant. Most of the reservoir will be outside the oxidation region and will actually experience recovery mechanisms identical to those of flue gas injection, primarily gravity segregation and stripping of light components (C1-C4) by nitrogen. Simplified reservoir scale simulation confirmed that the oil production from air injection is close to identical to the oil production when injecting flue gas. The oxygen is being consumed by oxidation in a small reservoir volume close to the injection wells.

The schematics of the process shown in Figure 1 may therefore be somewhat misleading in this case. Evaporation and condensation will take place, but not due to the heat itself, but rather through stripping and swelling initiated by flue gases from the oxidation process.

Under favorable conditions, we consider that as much as 25 % of the oil remaining after the waterflood may be recovered through air injection. Under normal circumstances however, the most likely recovery factor is estimated to 5 % of the original oil in place, primarily owing to limitation in gas sweep efficiency (gas override). If a gravity stable gas front can be established and maintained, this recovery factor may be significantly improved.

Surface Facilities and Wells. An evaluation was made of the need for investment in new surface facilities and upgrade of wells. For full field implementation of air injection on the Ekofisk field it was concluded that it will be necessary to build new injection and gas cleaning facilities, and upgrade 30 injection and 30 production wells, summing up to a total capital expenditure of 2700 million USD. The facilities were designed to have a capacity of injecting 13 million Sm3 air per day and processing 14 million Sm3 of produced gas. The main components included in the downstream processing facilities were units for gas sweetening, dehydration, nitrogen rejection and gas recompressing. It was assumed that it will be necessary to handle 6% CO2 and 52% N2 in the total gas flow. The processing facilities were designed to separate flue gas from sales gas to obtain the current gas sales requirements. An illustration of how the downstream processing facilities can be arranged on a separate processing platform is shown in Figure 13. Economic Evaluation. Production profiles for the Ekofisk field were produced to estimate the incremental hydrocarbon production from full field implementation of air injection. The incremental oil recovery factor is shown in Figure 14, as percentage of the original oil in place, in three cases.

SPE 97481 5

1. The gas production rate was limited to 700 million SCF per day on field scale. The gas injection was controlled by injection pressure (400 bars) without rate limitation. This resulted in a stabilized injection rate slightly above 700 million SCF per day (Air_700).

2. The gas production rate was limited to 700 million SCF per day. The gas injection was controlled by injection pressure (400 bars) with rate limitation of 450 million SCF per day. Water injection was added to maintain pressure (Air_450we).

3. The gas production rate was limited to 410 million SCF per day. The gas injection rate was controlled by injection pressure (400 bars) without rate limitation. This resulted in a stabilized injection rate of 450 million SCF per day (Air_450).

The incremental recovery was 3-4 % of original oil in place, when injecting air in the period from 2017 to 2041, based on optimistic assumptions. In addition substantial amounts of associated gas are being produced resulting in incremental recovery on the order of 10% of original oil equivalents in place.

In the economic evaluation the investments were made in 2016 and 2017. Sensitivities were run on capital expenses, operational expenses, production volumes and product prices. Expected values in terms of net present values were calculated through a decision tree analysis using a combination of the above parameters. Figure 15 shows a negative expected value of the project in terms of net present value, and low probability for a positive net present value.

Conclusions Through laboratory experiments and computer simulations it was shown that air injection is a feasible process for improved oil recovery after water flooding in a light oil fractured chalk reservoir.

It was shown that in-situ oxidation and combustion will take place when injecting air in a waterflooded Ekofisk reservoir, resulting in additional hydrocarbon recovery. Activation energies for in-situ oxidation and combustion were estimated. There is a prevailing probability that the Ekofisk oil will commence in-situ oxidation when contacted by injected air, even in the low temperature environment in the vicinity of water injectors.

Diffusion of air into a reservoir sample, and stripping of lighter hydrocarbons, was demonstrated, and diffusion coefficients from kinetic theory of gases were confirmed. The stripping of light oil components is greater than the swelling of the oil. Diffusion has a significant impact on recovery.

It was demonstrated that a stable combustion front can form and propagate under favorable reservoir conditions (low water saturation), and that additional hydrocarbons will be produced. It is unlikely that oxygen breakthrough will occur in the producing wells.

Most of the reservoir will be outside the combustion region and will experience recovery mechanisms being identical to those of flue gas injection, primarily gravity segregation and stripping of light components (C1-C4) by nitrogen.

Additional subsidence and compaction caused by air injection and in-situ oxidation and combustion was estimated, and is not considered to be prohibitive for an eventual air

injection project at the Ekofisk field. However, possible well failures caused by compaction might be a serious concern.

Substantial investments will be necessary to implement an air injection process at the Ekofisk field, including an air injection platform, a produced gas processing platform and wells upgrading.

An additional oil recovery can in an optimistic case be as high as 5% of the original oil in place, when injecting air in the period from 2017 to 2041, but the project economics was estimated to be negative, when taking into account capital and operational expenses and project uncertainties.

Even optimistic assumptions resulted in unattractive economics. Consequently, further work to reduce uncertainties both on production and cost profiles were not considered.

Presently, competing technologies, e.g. depressurization, CO2 injection, and continued water injection are considered to be more likely options to further increase recovery from Ekofisk than air injection.

Acknowledgements The authors acknowledge the European Commission, ConocoPhillips and the Ekofisk Coventurers, including TOTAL, ENI, Hydro, Statoil and Petoro, for financing the work and for the permission to publish this paper.

The project partners IFP, University of Bath and Fabricom are acknowledged for their contribution to the work.

IFP is also acknowledged for making their ATHOS software available for the simulation work.

Abbreviations IFP Institute Francais du Petrole USD US Dollar SCF Standard cubic foot Ci Hydrocarbon molecule with i carbon atoms

Conversion factors 1 foot = 0.3048 m 1 SCF = 0.028317 Sm3

References 1. Jensen, T.B., Harpole, K.J. and Oesthus, A: ”EOR Screening for

Ekofisk,” paper SPE 65124 presented at the 2000 SPE European Petroleum Conference, Paris, Oct. 24-25.

2. Fraim, M.L., Moffitt, P.D. and Yannimaras, D.V.: “Laboratory Testing and Simulation Results for High Pressure Air Injection in a Waterflooded North Sea Oil Reservoir,” paper SPE 38905 presented at the 1997 SPE Annual Technical Conference and Exhibition, San Antonio, Oct. 5-8.

3. Sakthikumar, S., Madaoui, K. and Chastang, J., “An Investigation of the Feasibility of Air Injection into a Waterflooded Light Oil Reservoir,” paper SPE 29806 presented at the 1995 Middle East Oil Show, Bahrain, March 11-14.

4. Clara, C., Durandeau, M., Quenault, G. and Nguyen, T.H.: “Laboratory Studies of Light Oil Air Injection Projects: Potential Application in Handil Field,” paper SPE 54377 presented at the 1999 SPE Asia Pacific Oil and Gas Conference and Exhibition, Jakarta, April 20-22.

5. Glandt, C.A., Pieterson, R., Dombrowski, A. and Balzarini, M.A.: “Coral Creek Field Study: A Comprehensive Assessment of the Potential of High-Pressure Air Injection in a Mature

6 SPE 97481

Waterflood Project,” paper SPE 52198 presented at the 1999 SPE Mid-Continent Operations Symposium, Oklahoma City, March 28-31.

6. Fassihi, M.R., Yannimaras, D.V., Westfall, E.E., and Gillham, T.H.: ”Economics of Light Oil Air Injection Projects,” paper SPE 35393 presented at the 1996 SPE/DOE Improved Oil Recovery Symposium, Tulsa, April 21-24.

7. Kumar, V.K and Fassihi, M.R.: “Case History and Appraisal of the Medicine Pole Hills Units Air Injection Project,” paper SPE 27792, SPE Reservoir Engineering Journal 10 (1995), 198-202.

8. Fassihi, M.R., Yannimaras, D.V. and Kumar, V.K., “Estimation of Recovery Factor in Light-Oil Air-Injection Projects,” paper SPE 28733 presented at the 1994 International Petroleum Conference and Exhibition of Mexico, Veracruz, Oct. 10-13.

9. Clara, C., Zelenko, V., Schirmer, P. and Wolter, T.: “Appraisal of the HORSE CREEK Air Injection Project Performance”, paper SPE 49519 presented at the 1998 Abu Dhabi International Petroleum Exhibition and Conference, Abu Dhabi, Nov. 11-14.

10. Watts, B.C., Hall, T.F. and Petri, D.J.: The Horse Creek Air Injection Project: An Overview,” paper SPE presented at the 1997 SPE Rocky Mountain Regional Meeting, Casper, May 18-21.

11. Gillham, T.H., Cerveny, B.W., Turek, B.W. and Yannimaras, D.V.: “Keys to Increasing Production Via Air Injection in Gulf Coast Light Oil Reservoirs,” paper SPE 38848 presented at the 1997 SPE Annual Technical Conference and Exhibition, San Antonio, Oct. 5-8.

12. Fassihi, M.R. and Gillham, T.H.: “The use of Air Injection to Improve the Double Displacement Processes,” paper SPE 26374 presented at the 1993 SPE Annual Technical Conference and Exhibition, Houston, Oct. 3-6.

13. Yannimaras, D.V. and Tiffin, D.L.: “Screening of Oils for In-Situ Combustion at Reservoir Conditions by Accelerating-Rate Calorimetry,” paper SPE 27791, SPE Reservoir Engineering Journal 10 (1995), 36-39.

14. Le Gallo, Y., Le Romancer, J.F., Bourbiaux, B. and Fernandes, G. : “Mass Transfer in Fractured Reservoirs during Gas Injection : Experimental and Numerical Modeling,” paper SPE 38924 presented at the 1997 SPE Annual Technical Conference and Exhibition, San Antonio, Oct. 5-8.

15. Turta, A.T. and Singhai, A.K.: “Reservoir Engineering Aspects of Oil Recovery from Low Permeability Reservoirs by Air Injection,” paper SPE 48841 presented at the 1998 SPE International Oil and Gas Conference and Exhibition in China, Beijing, Nov. 2-6.

16. Ren, S.R., Greaves M. and Rathbone, R.R.: Air Injection LTO Process: An IOR Technique for Light-Oil Reservoirs,” paper SPE 57005, SPE Journal 7 (2002), 90-99.

17. Surguchev, L.M., Koundin, A. and Yannimaras, D.: “Air Injection – Cost Effective IOR Method to Improve Oil Recovery from Depleted and Waterflooded Fields”, paper SPE 57296 presented at the 1999 SPE Asia Pacific Improved Oil Recovery Conference, Kuala Lumpur, Oct. 25-26.

18. Okamoto, M. and Bourbiaux, B.: “A Review of the Challenging Reservoir Engineering Aspects of Modeling the Recovery from a Light Oil, Fractured Reservoir by Air Injection”, paper presented at the 2005 European Symposium on Improved Oil Recovery, Budapest, April 25-27.

19. Greaves, M., Bentaher, A. H. and Rathbone, R.R.: “Air Injection into Light Oil Reservoirs – Oxidation Kinetics and Simulation”, paper presented at the 2005 European Symposium on Improved Oil Recovery, Budapest, April 25-27.

SPE 97481 7

Component Mol fraction Mol.

Weight (g/mol)

Gas-Gas diffusion coefficients

(10-3 m2/day) N2 0.0010 28 4.8 O2 0.0000 32 4.9

CO2 0.0083 44 4.0

CO 0.0000 28 3.4

C1 0.4406 16 5.5

LITE 0.1210 35 2.2 MEDIUM 0.0844 80 1.8 HEAVY 0.3445 254 1.0

Table 1. Diffusion coefficients used in simulations

Activation energy from diffusion experiment 36-40 kJ/mol

Activation energy from adiabatic disc reactor experiment 36-40 kJ/mol

Table 2. Activation energy estimates from simulation of experiments

Figure 1. Schematic of the air injection process.

Oil Bank

8 SPE 97481

Figure 2. Greater Ekofisk Area

Analysis

CorePV = 20 cm3

Atm PressureHigh Pressure

Diffusion chamber

0.48 cm3

Liquid trap

Figure 3. Experimental setup for diffusion experiments

Pilot area

SPE 97481 9

Molar fractions of O2, N2, C1, C2 and C3 in produced gas

0

1

10

100

0 1 2 3 4 5 6 7 8

Time (days)

Mol

ar fr

actio

n (%

)

%N2 %O2 %C1 %C2 %C3

Figure 4. O2, N2, C1, C2 and C3 molar fractions in produced gas vs. time

Molar fractions of C4, C5 and C6 in produced gas

0,01

0,10

1,00

0 1 2 3 4 5 6 7 8

Time (days)

Mol

ar fr

actio

ns (%

)

%iC4 %nC4 %iC5 %nC5 %C6

Figure 5. C4, C5 and C6 molar fractions in produced gas vs. time

10 SPE 97481

Figure 6. Experimental setup for kinetics experiment

EKOFISK ADR : Temperature variations

0

100

200

300

400

500

600

0 2 4 6 8 10 12 14 16 18 20 22

Time (hours)

Tem

pera

ture

(°C

)

Swi = 0 Swi = 13.3% Swi = 22.9% Swi = 42%

165°C

347°C

268°C

323°C

Figure 7. Temperature curves for different initial water saturations

SPE 97481 11

Cumulative oxygen consumption

0,0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0 2 4 6 8 10 12 14 16 18 20

Time (hours)

Cum

ulat

ive

Oxy

gen

Con

sum

ptio

n (m

ole)

Sw = 0% Sw = 13.3% Sw = 22.9% Sw = 42%

Figure 8. Cumulative oxygen consumption vs. time

Oil Production and N2 History

0

2

4

6

8

10

12

14

16

18

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0

Time (days)

Volu

me

- cm

3 (1

30°C

,275

b)

0

10

20

30

40

50

60

70

80

90

mol

ar fr

actio

n (%

)

Oil Prod. % N2

N2 Breakthrough

Oil produced due to Air and Flue Gas sweep

Figure 9. Oil production from long core experiment

12 SPE 97481

Figure 10. Constrained pore volume compressibility

Figure 11. Oil production sensitivity to diffusion and gravity segregation

Constrained Pore Volume Compressibility of Air-Oil Test PlugsEkofisk Reservoir Chalk, Initial Porosity ~39%

0.0E+00

5.0E-06

1.0E-05

1.5E-05

2.0E-05

2.5E-05

3.0E-05

0 1 2 3 4 5 6

Test Number

Con

stra

ined

Por

e V

olum

e C

ompr

essi

bilit

y, /p

si

Combusted Composite PlugUncombusted Intact PlugUncombusted Composite PlugsAir Diffusion Plug

AirOilMaterilProps Constrained Pore Volume Compressibility of Air-Oil Test PlugsEkofisk Reservoir Chalk, Initial Porosity ~39%

0.0E+00

5.0E-06

1.0E-05

1.5E-05

2.0E-05

2.5E-05

3.0E-05

0 1 2 3 4 5 6

Test Number

Con

stra

ined

Por

e V

olum

e C

ompr

essi

bilit

y, /p

si

Combusted Composite PlugUncombusted Intact PlugUncombusted Composite PlugsAir Diffusion Plug

AirOilMaterilProps

SPE 97481 13

Figure 12. Outflow of oxygen as function of diffusion and gravity segregation

Figure 13. Downstream processing facilities

UTILITY

GAS SWEETENING

FIREPUMP-ROOM

NITROGEN REJECTION

RECOMPRESSION PKG.

MOLECULAR SIEVE PKG.

14 SPE 97481

Increased Oil Recovery due to Air Injection

-0,5 %

0,0 %

0,5 %

1,0 %

1,5 %

2,0 %

2,5 %

3,0 %

3,5 %

4,0 %

2015 2020 2025 2030 2035 2040 2045

Oil

Rec

over

y Fa

ctor

bas

ed o

n O

OIP

, %

Air_450weAir_700Air_450

Figure 14. Increased oil recovery factor as percentage of original oil in place

Figure 15. Cumulative probability plot of expected values in terms of net present values

Cumulative Probability PlotAir Injection Value

0,0

0,2

0,4

0,6

0,8

1,0

-3000 -2000 -1000 0 1000 2000

Net Present Value

Cum

Ps

0 PositiveNegative

Cumulative Probability PlotAir Injection Value

0,0

0,2

0,4

0,6

0,8

1,0

-3000 -2000 -1000 0 1000 2000

Net Present Value

Cum

Ps

0 PositiveNegative