SPE 51893 Reservoir Characterization of Ekofisk Field: A Giant, Fractured Chalk Reservoir in the Norwegian North Sea -- History Match B. Agarwal, SPE, Phillips Petroleum Company UK Ltd, H. Hermansen, J. E. Sylte, SPE, Phillips Petroleum Company Norway, L. K. Thomas, SPE, Phillips Petroleum Company Copyright 1999, Society of Petroleum Engineers Inc. This paper was prepared for presentation at the 1999 SPE Reservoir Simulation Symposium held in Houston, Texas, 14–17 February 1999. This paper was selected for presentation by an SPE Program 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 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 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, SPE, P.O. Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435. Abstract The history match of the 3-D fluid flow model constructed on the Ekofisk Field is the focus of this study. The Ekofisk reservoir is a high porosity, low matrix permeability naturally fractured chalk. Fluid flow is largely governed by the distribution, orientation and interconnectivity of the natural fracture system associated with complex structure and reservoir distribution. Because of the impact heterogeneity has on preferential fluid flow direction, significant attention was given to capturing as much of the intrinsic heterogeneity as possible, both laterally and vertically, in the new 3-D geological model. A fairly high-resolution simulation mesh was defined for the fluid flow model, and a flow-based upscaling technique was then applied to preserve the heterogeneity from the geological to the fluid flow model. Because of the complexity of the Ekofisk Field, with its numerous faults and fracture networks, anisotropy was one of the primary attributes manipulated to achieve an individual well and field history match. However, faults and fault sealing factors, vertical permeability, pseudo relative permeability curves, bubble point pressure correlations, local permeability, and rock compressibility were also key parameters in the history match, and are presented in this paper. A brief discussion on the preliminary implementation of waterinduced compaction is included. Introduction The Ekofisk Field is a prolific field discovered in 1969 and located in the Norwegian Sector of the North Sea. The reservoir consists of two fine-grained limestone producing formations, the Ekofisk formation (Danian Age) and the Tor formation (Maastrichtian Age), separated by a thin, impermeable Tight Zone. The reservoir was initially overpressured and contained an undersaturated oil at 50 Mpa (7120 psi) and 131 oC (268 oF) at a datum elevation of 3170 m (10,400 ft) subsea. The bubble point pressure was approximately 38 Mpa (5545 psig). Production started from the chalks in 1971. Current estimates from the reservoir characterization project indicate about 7 billion barrels of oil originally in place. Current production from 76 deviated and horizontal wells is 310,000 barrels of oil per day and 510 MMCFD of gas. A pilot water injection project was initiated in 1981 in the highly fractured Tor formation1 and in the Lower Ekofisk in 19862. Fieldwide water injection began in 19873. Current water injection rates are 800,000 bwpd into 37 active injection wells. Fig. 1 shows a structure map of the Ekofisk Field drawn on the top Ekofisk formation. Reservoir characterization on Ekofisk was directed at gaining a detailed understanding of reservoir hydrocarbon volumes, the architecture of the reservoir and at fully describing the heterogeneity and anisotropy of reservoir parameters4,5. Because water breakthrough has been observed in areas of the field not consistent with expectations, it is important that the highest degree of heterogeneity be represented in the flow model. This is especially significant given that the Ekofisk Field is currently undergoing a major field re-development in which 45 new wells will be drilled before the end of 2003. To date a total of 19 new wells have already been drilled and are currently on production. The history match of the Ekofisk Reservoir Characterization fluid flow model was completed in September 1997 after a period of approximately twelve (12) months of intense work. The complexity of the Ekofisk field, with its numerous faults and fracture networks, provided quite a challenge in matching the twenty-five (25) years worth of production and performance data. The heterogeneity that was captured in the 3-D geological model, and preserved in the upscaling process to the fluid flow model, proved to be the key to being able to match individual well performance. In general, a very good history match was achieved on both a field and platform basis, and on an individual well basis. 2 B. AGARWAL, H. HERMANSEN, J. E. SYLTE, L. K. THOMAS History Match Basic Data. The basic data used in the history matching of the flow model includes the following: (1) Data from 160 wellbores; (2) twenty-five years of production history; (3) RFT and FMT data representing formation pressures for a majority of the wells; (4) production log data for most wells, showing relative contributions from sub-layers; (5) logs representing water saturation from later wells to help model the degree of water influx from the flank; (6) time-lapse neutron study to match vertical gas distribution in approximately 25 wells with data; (7) historical bottomhole pressure data. History Matching Parameters. The main parameters used in the history match of the fluid flow model were the following: (1) Anisotropy in the x-y direction to control gas and water pattern away from the gas and water injection wells; (2) fault factors to control pressure support and flow direction; (3) nonneighbor connections across faults to manage fluid movement and direct communication between sub-layers; (4) vertical permeability, kv/kh to control gas migration and pressure support; (5) flank permeability to control aquifer pressure support and water influx; (6) pseudo relative permeability curves and rock region definitions to control fluid movement; (7) endpoint krw and permeability around water injectors to match bottomhole injection pressures; (8) permeability and skin in cells with completions to adjust well productivity to match bottomhole pressure data; (9) well productivity index to match PLT contributions; (10) bubble point pressure correlation to match initial undersaturated gas-oil ratios; (11) rock compressibility to provide pressure support from the flank and to match compaction volumes. In general, the principal factor in the history match was modeling the high GOR and watercut responses in specific wells. Anisotropy was the key to achieving this and was based on the fault and fracture networks which exist in the Ekofisk field. In the following sections, we describe the various parameters and their influence on the history match. Anisotropy/Faults. Fluid flow in the Ekofisk field is largely governed by the distribution, orientation and interconnectivity of the natural fracture system associated with complex structure and reservoir distribution. There are over 300 fault planes mapped in the Ekofisk field, with a majority of the major faults defined in the fluid flow model. Faults. Non-Neighbor Connection Faults. The faults with the largest throws, primarily the NE-SW normal faults, have been designated as non-neighbor connection faults. These are faults that allow communication across different bed boundaries; for example, Layer EA2 might be feeding directly into Layer EB, or more importantly, Layer TAU1 communicates directly with Layer ED2, thus allowing a “window” through the Tight Zone. SPE 51893 Fig. 2 is a schematic of a typical non-neighbor connection fault. As is represented in this figure, non-neighbor connection faults are modeled as vertical boundaries. For this reason, in specific instances where the dip of the fault plane is large, separate definitions have been specified for the Ekofisk and Tor formations. There are a total of nine (9) non-neighbor connection faults defined in the fluid flow model. Once the fault is defined, the sealing factor across the fault can be adjusted during the history match to model the correct pressure and saturation responses. Simple Faults. Numerous faults are defined in the reservoir simulation model to control pressure and/or fluid flow direction. These faults represent, in general, the medium-sized strike-slip faults and reverse faults, whose throws are typically less than 50 feet. The fault map is “rasterized” onto the simulation mesh to determine the grid cell boundaries corresponding to the fault. Simple faults are used primarily to affect fluid flow direction in the history match, and to constrain the effects of waterflood response. Anisotropy Lateral. The remainder of the faults, not mapped directly within the flow model as either non-neighbor connection or simple faults, were defined based on anisotropy. Performance data in the field suggested that many of the fault networks act as conduits to flow, and can, therefore, be considered as enhanced transmissibility. To correctly integrate this into the reservoir simulation model, a detailed analysis was performed on the available data in which response times between injector-injector, injector-producer, and producer-producer combinations were evaluated. Based on the relative time observed to see a response in a particular well, an initial value of anisotropy was assigned to this specific area. For example, if a water injector was put on stream, and a definite response was observed in GOR in an offset producer in one to two months, this was a clear indication of direct communication between the two wells. Therefore, a relatively high anisotropy, say 10:1, was assigned to this area. This procedure was followed for all well combinations for which data existed. Fracture density was used as an additional basis for determining the value of anisotropy to apply to a given area. Areas with high fracture intensity were assigned higher values of anisotropy, and values with lower fracture intensity were assigned lower values of anisotropy. In this fashion, a template of x-y anisotropy that was directly linked to the fault and fracture network was created as initial input to the flow model. A study based on the stress regimes which exist in the field showed that the regional trend in anisotropy was approximately 3:1. This value was assigned to all areas with minimal fracturing, such as most of the flanks and portions of the crest of the field. Fig. 3 illustrates the anisotropy template created for the reservoir simulation model. The history match itself required tuning of these anisotropy values, but the relationship to fracture density was maintained throughout. SPE 51893 RESERVOIR CHARACTERIZATION OF EKOFISK FIELD -- HISTORY MATCH Vertical. Similar to the lateral anisotropy defined, vertical communication was also based on the fault network. Typical input values of kv/kh range from 0.0 to 0.1, with the majority of the grid cells assigned values in the 0.025 to 0.033 range. The Ekofisk is considered a tighter formation than the Tor so lower values were assigned to the top seven (7) simulation layers, and lateral variation of kv/kh was greater than in the Tor. The Tight Zone is generally a complete seal throughout the majority of the field. There are, however, areas in the field in which communication between the Ekofisk and Tor formations does occur. In previous models built on the Ekofisk field, a “window” was described in the Tight Zone where a block of cells were assigned non-zero vertical permeability. For the ERC model, however, a more geological basis was the objective of defining the “window” in the Tight Zone. This was done by initially assigning vertical permeability to cells in the flow model that contained the major faults. For the most part, these cells corresponded to the cells to which nonneighbor connection faults were defined. Although the nonneighbor connection faults allow some vertical communication through the Tight Zone in areas of large throw, the amount of fluid movement was not sufficient with non-neighbor connections alone. The values assigned to the kv/kh grids in the flow model were tuned based on the history match to control gas migration, and to match RFT pressures from individual wells. Fig. 4 is a representation of the Z-direction modifying factor grid, defined in the simulation model, for the Tight Zone. In this figure, the blue areas represent zero vertical communication and the colored areas reflect non-zero transmissibility due to the fault network. Relative Permeability/Rock Regions Mechanistic studies were performed to generate water-oil pseudo-relative permeability curves for the 3-D reservoir flow model6. The purpose of this work was to account for the fluid transfer between the high permeability fractures and low permeability matrix blocks, minimize numerical dispersion, and accurately model physical dispersion due to layering, fracturing, and matrix block geometries. Although the matrix blocks contain the bulk of the pore volume, it is the natural fracture system that dominates overall permeability. For strongly water wet areas such as the Tor and Lower Ekofisk formations at Ekofisk, the bulk of waterflood displacement is a capillary dominated process and the rate of recovery is strongly dependent on matrix block size and the amount of matrix surface area exposed to fractures, with ultimate recovery determined by the capillary, gravity, and viscous equilibrium. For less water wet areas, such as the Upper Ekofisk formation at Ekofisk, viscous and gravity forces as well as capillary forces are important. A typical water-oil pseudo relative permeability curve for the Lower Ekofisk formation is presented in Fig. 5 for a high fracture intensity region. The curves in general indicate a low relative permeability to water and high relative permeability to oil for the lower water saturation range to reflect the very favorable imbibition properties for the Tor and Lower Ekofisk 3 chalk. A fairly steep slope exists on the water and oil relative permeabilities at higher water saturations to reflect the rapid movement of water through the natural fracture system once the imbibition process is approaching completion. For the gas oil system with very low viscous gradients, the mechanistic studies have focused primarily on gravity drainage, with capillary continuity between matrix blocks and oil phase reimbibition being two important phenomena. This results in ultimate oil recoveries similar to a non-fractured block of the same height, and oil drained from the higher matrix blocks re-enters lower matrix blocks, slowing the drainage process. A two stage upscaling process was used to develop pseudo relative permeability curves for the single porosity model. The first involved detailed modelling of stacks of individual matrix blocks, which included gravitational, viscous, and capillary effects in addition to block to block reinfiltration. Then, the upscaled properties from the detailed model were applied in the field scaled single porosity cross-sections to account for intra-layer heterogeneities. History matching the GOR for Ekofisk is quite challenging primarily due to the extensive system of fractures and faults in the reservoir. Natural gas injection from early in the life of the field as well as the evolution of gas in the field during depletion below bubble point results in rapidly increasing GOR's in highly fractured areas. Also, as waterflood was initiated after the field was well below bubble point, gas collapse due to pressure support from the water injectors has resulted in significant drops in producing GOR’s. Achieving the correct balance between putting gas back into solution and maintaining a gas saturation for the higher GOR wells was a priority in this work. Bubble Point Pressure Bubble point pressure as a function of depth for the Ekofisk field was determined based on production GOR data from thirty-four (34) pre-bubble point producers. The procedure involved establishing the average pre-bubble point GOR for each producer, determining the bubble point pressure from the Rs versus pressure curve implemented in the full field model, then relating the pressure to depth based on perforation depths and PLT contributions. Fine-tuning of the correlation was required to match early-time GOR behavior. Fig. 6 represents the correlation implemented in the flow model. Results and Observations. The history match of the ERC fluid flow model was completed in September 1997 and was subsequently used in predicting new well locations for the redevelopment campaign on the Ekofisk field. Even though a good history match was achieved, it is certainly non-unique, and the ultimate test of any reservoir simulation model is its ability to predict into the future. Data gathered for new wells drilled in the field toward the later stages of the history match process were valuable in validating the performance of the reservoir model. For several of the long horizontal wells, the model proved to be quite accurate in its predictions in that it 4 B. AGARWAL, H. HERMANSEN, J. E. SYLTE, L. K. THOMAS showed high water saturation in areas not previously indicated. This lends confidence to the history match of the field. Pressure. History matching pressure in Ekofisk is complicated by the layered nature of the field as well as the extensive faulting in the reservoir. The presence of the Tight Zone providing a complete seal throughout the majority of the field, a combination of different rock properties, and aquifer support results in significant differential depletion in many areas of the field. In general, Ekofisk formation pressures are lower than those of the Tor formation. This can be observed in numerous wells around the field, such as the A-09A, C-14A and B-23A. However, a reverse trend has been observed in some wells, particularly on the eastern flank. The C-21 and C-22 are good examples of this. Fig. 7 represents Ekofisk and Tor formation pressures through time, as recorded through RFT data. Extreme rate changes due to workovers, new producers and the introduction of water flooding in a staged manner has resulted in pressure swings on the order of 1000-2000 psi periodically throughout the field. As a result, a match to within 150 psi of the measured pressure was considered good. A majority of the wells were matched to this tolerance. A good example of the RFT match is well A-05A, drilled in May, 1986, which shows a good match of both the Tor and Ekofisk formation pressures, as well as the significant differential depletion, Fig. 8. Flank well pressures often proved more difficult to match as they were much more sensitive to permeability. Too high a permeability resulted in greater water influx than actual so that a match to the measured water saturations could not be achieved. Too low a permeability resulted in severe pressure depletion causing the pressures to be much lower than those measured in RFT data. The objective was to try to achieve a balance using absolute permeability, pseudo-relative permeability and rock compressibility. Rock compressibility in the Tor formation was increased to 6.0E-06 psi-1 based on available lab data. This change had a positive impact on the overall Tor pressures, allowing a better match to be achieved. In summary, Tor flank pressure was controlled primarily by the following: 1. Vertical permeability (kv/kh) was increased in the far flank to provide more aquifer support. 2. Flank absolute permeability was increased to provide more lateral support. The constraint of water influx was rigorously enforced, however. 3. Bottom aquifer support was increased by adjusting permeability and kv/kh in the bottom layer of the flow model. This was a sensitive parameter and required careful tuning to match water production. 4. Tor rock compressibility was increased to a higher value based on laboratory data. Ekofisk Formation pressures initially tended to be a little on the high side. Pseudo-relative permeability (kro) was used in conjunction with flank permeabilities to bring pressures back SPE 51893 in line with measured data. Also, the fact that Tor pressures were low and Ekofisk were high, suggested either downward fluid/pressure migration or limited upward fluid/pressure migration through the Tight Zone. Tuning of the kv/kh grids helped to control this. Gas-Oil Ratio. Gas-oil ratio (GOR) was one of the major matching parameters in the fluid flow model. The intent was to match measured and simulated GOR on a field and platform basis, and on an individual well basis, to the best extent possible. Given the high GOR’s in some of the early wells and some of the later Tor wells, this was not a trivial task. One of the main challenges was to maintain a reasonable gas saturation in the Tor formation past commencement of the waterflood. The natural result was for the gas to be rapidly pushed back into solution with an increase in pressure from water injection. A good example of this can be seen with the GOR behavior of well A-14. Initially, the well produced at solution GOR with initial pressures considerably above the bubble point. With pressure depletion and possibly some impact of the natural gas injection, the GOR,s rose steadily through the years and peaked at a GOR of 12,230 SCF/STB in 1990. With the start up of the offset water injection well K-03, a sharp and dramatic drop in producting GOR is seen as displayed in Fig. 9. The model produces a very reasonable match of this behavior. An excellent match was obtained for both the full field and individual platforms, Figs. 10 through 13. In these figures, the solid line represents the response calculated from the flow model, and the dashed line reflects the actual historical values. Watercut. Watercut was another important matching parameter for the Ekofisk field. With the re-development of the field ongoing, and approximately 45 new wells planned, determining the waterflood pattern development away from injectors was key. Ekofisk is a very complex field with the fault network acting as a major conduit to flow. Given the heterogeneity in the field, water does not move radially away from the water injection wells. Instead, fingers and channels exist which are normally very difficult to duplicate in a reservoir simulation model. By directly linking permeability and anisotropy to the fracture system in the field, and by defining a fairly high grid resolution, the ERC flow model is better able to model the water pattern development. An excellent match was achieved for the full field and individual platforms as presented in Figs. 14-17. Good examples of individual well water cut matches can be seen on the match of well C-06 as presented in Fig. 18. This well had direct communication to the offset water injection well K-19. Nearly instantaneous breakthrough was seen with the start of water injection at K-19. This well was subsequently sidetracked a short distance away and came in essentially water free producing from the same intervals as the original well. The excellent match of this water cut performance could only be achieved with extreme flow channels resulting from the fracture and fault system. Another good example can be seen with the results of the C-23 well, which was drilled close to a fault system connecting the well with the offset water SPE 51893 RESERVOIR CHARACTERIZATION OF EKOFISK FIELD -- HISTORY MATCH injection well K-04. Watercut has increased steadily in this well and is currently at approximately 90 percent, Fig. 19. Tracer tests, performed in 1994 by injecting tracer into the K-04 well and monitoring offset producing wells, confirmed a strong interaction with the injection well. This match could only be achieved with a high quality reservoir model. One area of difficulty was in the vicinity of the Charlie platform itself, the C-11 well. This well is located in a vicinity of the field where seismic imaging is very poor, and as a result the fault system poorly defined. This well has historically been used as a compaction monitoring well. In this regard, time lapse data on water saturation in the wellbore existed. The data indicates that the thin zone just beneath the Tight Zone showed high water saturation very early, followed by the zone just above the Tight Zone. These two zones are typically considered the high compaction zones. Given the timing and completion schedule of the water injectors surrounding this well, the only way to match the water saturation profile in the wellbore was to channel water directly to the area by way of defining two additional faults with very tight sealing factors. Implementation of this extreme measure helped but did not provide an exact match. The complex nature of the fracture and fault system in this area in combination with poor seismic imaging appears to be the reason for not being able to match the watercut precisely. These above examples and others are considered as validation of the Ekofisk reservoir model. The drilling of the new re-development wells has been an overwhelming success with production rates now being seen at Ekofisk that have not been achieved since the late 1970’s. The confidence in the model’s history match and prediction capabilities has been an important part of this success. Permeability. The permeability algorithm4 and upscaled permeabilities5 developed for the ERC model performed very well in that major changes to the distribution of permeability were not necessary. Some local refinement around wells, in particular water injectors, was required possibly due to in situ fracturing. Also flank permeability in certain geologic layers were adjusted to control pressure support from the aquifer. Water Induced Compaction. For the Ekofisk field, laboratory data has been used to establish a relationship between the amount of water contacting the chalk and the degree of compaction observed, Fig. 20. That is, increased compaction is observed as the water saturation in a particular block or cell increases, ie, the rock is weakened as a function of increasing water saturation. Logic was developed to incorporate these relationships in the model and the history match was re-run. Implementation of this water weakening logic in the fluid flow model had a significant impact on both the Tor formation pressure and the late GOR behavior of the wells. Furthermore, the compaction volumes of the fluid flow model much better matched the actual volumes. The increased compaction resulting from water weakening, provided more energy in the system. This increase in energy resulted in higher pressures, 5 especially in the Tor formation. This, in turn, mitigated the pressure depletion in specific compartments in the field, which prevented excess gas coming out of solution and, thus, too high GOR’s late in history. The fluid flow model with the water weakening logic implemented will be used in the future for the Ekofisk field and will be the basis for further history matching. Refinement of the relationships established from laboratory data still need to be performed and the history match subsequently finetuned. Concluding Remarks An integrated, high-resolution geological model and a fairly high-resolution reservoir simulation flow model have been developed for the detailed planning of the redevelopment of the Ekofisk Field. A number of innovative techniques have been applied within the disciplines of geoscience, petrophysics and reservoir engineering resulting in integrated models which enable team-driven decisions to be made regarding reserves optimization. The models have been developed using all data, information and knowledge available during the time span of the project. Processes have been put in place to ensure the continuous and rapid updating of the geological and fluid flow models as new data, information and knowledge becomes available. In this study, the history match of the fluid flow model is presented. Here the challenges of matching such a complex field and the approach taken to achieve the final history match are discussed. The following observations are warranted: 1. Additional reservoir characterization is obtained through the history matching process. Flow across faults, vertical permeabilities, flank characteristics, etc, are solved for in the history match. 2. The geological model is not an exact representation of the reservoir; it reflects the best interpretation possible based on the limited data available. The history match solves a number of characterization issues and the process of downscaling will ultimately reflect this back into the geological model. 3. Generally, a good pressure match was achieved in the fluid flow model. 4. GOR was well matched on both a field and platform basis and an individual well basis. 5. A good watercut match was obtained with some exception, for example around the Charlie platform where thin zones of high water production exist. 6. Further development of the relationships for the water weakening logic will enhance the history match, and allow more accurate GOR trends to be achieved late in history. Nomenclature Rs = solution gas dissolved in oil phase, scf/stb GOR = gas-oil-ration, scf/stb PLT = production logging tool 6 B. AGARWAL, H. HERMANSEN, J. E. SYLTE, L. K. THOMAS cf = rock compressibility, psi-1 Acknowledgments The authors acknowledge permission to publish the above paper from Phillips Petroleum Company Norway and Coventurers, including Fina Exploration Norway SCA, Norsk Agip A/S, Elf Petroleum Norge AS, Norsk Hydro A.S., Den norske stats oljeselskap a.s., TOTAL Norge A.S. and Saga Petroleum A.S. References 1. 2. 3. 4. 5. 6. Thomas, L. K., Dixon, T. N., Evans, C. E., and Vienot, M. E.: “Ekofisk Waterflood Pilot,” JPT,February, 1987, pp. 221-232. Sylte, J. E., Hallenbeck, L. D., and Thomas, L. K.: “Ekofisk Formation Pilot Waterflood,” Paper SPE 18276 presented at the 1988 SPE Annual Technical Conference, Houston, Texas, October 2-5, 1988. Hallenbeck, L. D., Sylte, J. E., Ebbs, D. J., and Thomas, L. K.: “Implementation of the Ekofisk Full Field Waterflood,” SPE 19838 presented at the 1989 SPE Annual Technical Conference, San Antonio, Texas, October 8-11, 1989. Agarwal, B., Allen, L. R., and Farrell, H. E.: “Ekofisk Field Reservoir Characterization: Mapping Permeability Through Facies and Fracture Intensity,” SPE Formation Evaluation, December 1997, pp. 227-233. Agarwal, B., Thomas, L. K., Sylte, J. E., and O’Meara, D.: “Reservoir Characterization of Ekofisk Field: A Giant, Fractured Chalk Reservoir in the Norwegian North Sea: Upscaling,” SPE 38875 presented at the 1997 SPE Annual Fall Meeting, San Antonio, Texas, October 5-8, 1997. Hermansen, H. and Thomas, L. K.: “Reservoir Simulation Challenges at the Ekofisk Field,” Conference on Characterization and Interpretation of Fluid Flow in Fractured Reservoirs, Society of Core Analysts (SPWLA), Hamburg, Germany, October, 1998. SPE 51893 RESERVOIR CHARACTERIZATION OF EKOFISK FIELD -- HISTORY MATCH W - 0 4 01 0 SPE 51893 W W G 9 6 0 0 G G - 0 1 0 0 0 0 G 6 - 9 0 0 - 98 0 0 04 G 0 1 04 00 - Depth: 2890 m (9480 ft.) Size: 27 sq. km. (16.8 sq. mi.) 76 producing wells, 37 injectors 0 1- 0 20 Fig. 1.--Top structure map of the Ekofisk field. EA2 EA2 EB EB EC EC Fig. 2 - Schematic representing typical non-neighbor connection fault in the Ekofisk Field. 7 8 B. AGARWAL, H. HERMANSEN, J. E. SYLTE, L. K. THOMAS Fig. 3 - Lateral anisotropy map of the Ekofisk field fluid flow model. SPE 51893 Fig. 4 - Vertical anisotropy map of the Tight Zone of the Ekofisk field fluid flow model. W a t e r- O il P s e u d o R e l a t i v e P e r m e a b i l i t y Ppb vs. Depth Based on Pre-Pbp GOR Data Hi g h F r a c t u r e I n t e n s i t y R e g i o n 1 9500 0 .8 9800 k ro Depth, ft 0 .6 0 .4 10100 10400 k rw 0 .2 10700 4500 0 0 0 .2 0 .4 0 .6 5000 5500 6000 0 .8 W a t e r S a t ura t io n , f r a c t i o n Fig. 5 - Typical water-oil pseudo relative permeability curve from a high fracture intensity region of the Ekofisk field. Bubble Point Pressure, psi Fig. 6 - Ekofisk field bubble point as a function of depth correlation. SPE 51893 RESERVOIR CHARACTERIZATION OF EKOFISK FIELD -- HISTORY MATCH 9 8000 D a n i a n R F T D a ta PRESSURE, psi 7000 U . E K O S IM U L A T IO N A V E R A G E L . E K O S IM U L A T IO N A V E R A G E 6000 5000 4000 3000 2000 75 80 85 90 95 100 YEAR 8000 C r e ta c e o u s R F T D a ta PRESSURE, psi 7000 T O R S IM U L A T IO N A V E R A G E 6000 5000 4000 3000 2000 75 80 85 90 YEAR Fig. 7 - Ekofisk andTtor RFT and simulation model pressures. Fig. 8 - RFT match for well A-05A from history match. 95 100 10 B. AGARWAL, H. HERMANSEN, J. E. SYLTE, L. K. THOMAS Fig. 9 - Well A-14 GOR match. Fig. 11 - Bravo Platform GOR match. Dashed line reflects actual and solid line reflects simulated response SPE 51893 Fig. 10 - Alpha Platform GOR match. Dashed line reflects actual and solid line reflects simulated response Fig. 12 - Charlie Platform GOR match. Dashed line reflects actual and solid line reflects simulated response SPE 51893 RESERVOIR CHARACTERIZATION OF EKOFISK FIELD -- HISTORY MATCH Fig. 13 - Total Field GOR match. Dashed line reflects actual and solid line reflects simulated response Fig. 15 - Bravo Platform watercut match. Dashed line reflects actual and solid line reflects simulated response Fig. 14 - Alpha Platform watercut match. Dashed line reflects actual and solid line reflects simulated response Fig. 16 - Charlie Platform watercut match. Dashed line reflects actual and solid line reflects simulated response 11 B. AGARWAL, H. HERMANSEN, J. E. SYLTE, L. K. THOMAS Fig. 17 - Total Field watercut match. Dashed line reflects actual and solid line reflects simulated response SPE 51893 Fig. 18 - Well C-06 watercut match. Dashed line reflects actual and solid line reflects simulated response Ekofisk Compaction Data - 40% Initial Porosity 45 40 Model Porosity, percent 12 35 Sw - 0% Sw - 5% Sw - 10% Sw - 15% Sw - 20% Sw - 25% 30 25 20 0 2 4 6 8 10 Effective Axial Stress, thousands Fig. 19 - Well C23 watercut match. Dashed line reflects actual and solid line reflects simulated response Fig. 20 - Ekofisk uniaxial stress laboratory data for a 40% porosity chalk sample converted to model porosity.