a snapshot of carbonate reservoir evaluation -...
TRANSCRIPT
20 Oilfield Review
A Snapshot of Carbonate Reservoir Evaluation
Mahmood AkbarBadarinadh VissapragadaAbu Dhabi, UAE
Ali H. AlghamdiSaudi AramcoDhahran, Saudi Arabia
David AllenMichael HerronRidgefield, Connecticut, USA
Andrew CarnegieDhruba DuttaJean-Rémy OlesenOil & Natural Gas Corporation-SchlumbergerJoint Research CenterNew Delhi, India
R. D. ChourasiyaOil & Natural Gas Corporation, Ltd.Mumbai, India
Dale LoganDave StiefMidland, Texas, USA
Richard NetherwoodJakarta, Indonesia
S. Duffy RussellAbu Dhabi Company for Onshore Oil OperationsAbu Dhabi, UAE
Kamlesh SaxenaMumbai, India
For help in preparation of this article, thanks to Kamal Babourand Robert Dennis, Al-Khobar, Saudi Arabia; Tim Diggs,Shell International EP, Houston, Texas, USA; Jack Horkowitz,Sugar Land, Texas; Fikri Kuchuk and participants in the firstannual MEA Carbonates Forum, Dubai, UAE; Chris Lenn,Houston, Texas; T. S. Ramakrishnan and Yi-Qiao Song,Ridgefield, Connecticut, USA; Charlotte Sullivan, Universityof Houston, Texas; W. Bruce Ward, Earthworks LLC,Norwalk, Connecticut.BorTex, CMR (Combinable Magnetic Resonance), CNL(Compensated Neutron Log), ECS (Elemental CaptureSpectroscopy), ELAN (Elemental Log Analysis), FMI(Fullbore Formation MicroImager), GeoFrame, Litho-Density,MDT (Modular Formation Dynamics Tester), PL Flagship, PS Platform, Q, RockCell, RST (Reservoir Saturation Tool),RSTPro, SpectroLith and TDT (Thermal Decay Time) aremarks of Schlumberger.
Carbonate reservoir evaluation has been a high priority for researchers and oil
and gas producers for decades, but the challenges presented by these highly
heterogeneous rocks seem to be never-ending. From initial exploration through
mature stages of production, geoscientists, petrophysicists and engineers work
together to extract as much information as possible from their data to produce
maximum reserves from the ground.
Carbonate reservoirs present a picture ofextremes. Reservoirs can be colossal thoughtheir pores can be microscopic (next page, top).Matrix permeability can be immeasurably low,while fluids flow like rivers through fractures.Evaluation techniques that succeed in sandstonereservoirs sometimes fail in carbonate reservoirs.These variations complicate both reservoir evalu-ation and hydrocarbon recovery. However,researchers are working to overcome these prob-lems because of the economic significance of oilproduction from carbonate reservoirs, especiallygiant and supergiant fields in the Middle East.
The potential rewards are great: about 60% ofthe world’s oil reserves lie in carbonate reser-voirs, with huge potential for additional gasreserves, particularly in the Middle East. In thisarticle, we examine ways to evaluate carbonatereservoirs at the scale of cores and well logs withexamples from both research and operationsgroups around the world (next page, bottom).1
Methods range from tried-and-true to experimen-tal, and represent a subset of current initiatives
rather than a comprehensive review. The resultsof borehole-scale evaluations play significant rolesin larger scale field development, simulation andmanagement efforts. We also discuss the impactof these results on ongoing research efforts.
Why the Fuss about Carbonate Rocks?Carbonate sedimentary rocks differ from silici-clastic sedimentary rocks in several ways.Siliciclastic rocks form as sediments are trans-ported, deposited and lithified, or compacted andcemented into solid rock. Most carbonate rocksdevelop from biogenic sediments formed by bio-logical activity, such as reef building and accu-mulation of the remains of organisms on theseafloor. Other types form as water evaporatesfrom shallow onshore basins or as precipitatesfrom seawater. The fragments that make up mostcarbonate rocks typically have undergone muchless transport than most siliciclastic sediments.
Siliciclastic rocks are predominantly sand-stones and shales that contain a wide variety ofminerals and particles, including quartz, feldspar,
1. For a general introduction to carbonate interpretation:Akbar M, Petricola M, Watfa M, Badri M, Charara M,Boyd A, Cassell B, Nurmi R, Delhomme J-P, Grace M,Kenyon B and Roestenburg J: “Classic InterpretationProblems: Evaluating Carbonates,” Oilfield Review 7, no. 1 (January 1995): 38-57.
Winter 2000/2001 21
> Carbonate heterogeneity. Photomicrograph pairs show three rock fabrics from the same reservoir. The images at the top are conventional plane-polarized light photomicrographs of thin sections. The cathodoluminescence photomicrographs (bottom) reveal different generations of carbonate minerals formed during diagenesis. Each rock fabric has a different response to nuclear magnetic resonance (NMR) because of thedifferent relationships of the pores within and between grains. Differences in depositional facies and stratigraphic position produced three distinctdiagenetic pathways. In the ooid grainstone (left), the cores of the ooids were dissolved early in the depositional history. Calcite cements filledboth intergranular and intragranular porosity. The fabric-retentive, dolomitized ooid-peloidal grainstone (center) initially underwent minor diagenesisduring which some skeletal grains were dissolved. Fine dolomite crystals then replaced the sediment and preserved the original fabric at an earlystage. Later, dolomite cement filled some of the large moldic pores. Sucrosic dolostones (right) represent peloidal grainstone that was replaced byfine sucrosic dolomite crystals, destroying much of the original depositional texture.
60° N
40° N
20° N
0°
20° S
40° S
60° S Reef Shelf carbonate Deep carbonate Carbonate oil province
> Distribution of carbonate rocks. The black circles denote locations of examples described in this article.
clay minerals, fragments of preexisting rocks andremnants of plants or animals. Carbonate rocksconsist of a more limited group of minerals—predominantly calcite and dolomite. Other miner-als less commonly present in carbonate rocks arephosphate and glauconite; secondary mineralsinclude anhydrite, chert, quartz, clay minerals,pyrite, ankerite and siderite.
These differences result in entirely differentclassification systems for clastic and carbonaterocks, with clastic rocks distinguished by graincomposition and size, and carbonate rocks differ-entiated by such factors as depositional texture,grain or pore types, rock fabric or diagenesis(right).2 Differentiating present-day flow unitsfrom original depositional units is becoming moreimportant than other aspects of classificationbecause optimal well placement depends onunderstanding present-day flow units.
Once deposited, sediments undergo diagene-sis, the postdepositional chemical and physicalchanges that transform the sediment into solidrock. Carbonate diagenesis can significantlymodify pore space and permeability. Carbonaterocks are highly susceptible to dissolution; grainscan be dissolved to form new pore space, anddissolution along fractures and bedding planescan produce large vugs and caves. Clastic dia-genesis normally does not involve a change inmineralogy. Carbonate diagenesis, however,commonly involves replacing the original calciteand aragonite with the mineral dolomite, a pro-cess called dolomitization, which can improvethe hydrocarbon-producing characteristics.
While both clastic and carbonate rocks areusually buried, compacted and cemented, carbon-ate sediments contain significant amounts of themetastable minerals aragonite and magnesiumcalcite; calcite itself is readily dissolved andreprecipitated by percolating pore fluids.Carbonate rocks are, therefore, more likely toundergo dissolution, mineralogical replacementand recrystallization. These effects vary according
to temperature, pore-fluid chemistry and pres-sure. Carbonate diagenesis commonly beginswith marine cementation and boring by organ-isms at the sediment-water interface prior toburial. It continues through shallow burial withcementation, dissolution and recrystallization,and then deeper burial where dissolution pro-cesses known as pressure solution may form suchfeatures as stylolites.3
When confronted with core samples or imagelogs from carbonate rocks, even casual observersnotice the tremendous variety of pore types andsizes, and the irregular distribution of pores.Pores in clastic rocks are predominantly betweenthe grains, or intergranular, and uniformly dis-tributed throughout the rock matrix. Intergranularpores are also present in carbonate rocks.Intragranular porosity may be common within
carbonate grains as a primary pore type, or maydevelop when grains, such as shell fragments,are partially dissolved. Moldic porosity preservesthe shapes of dissolved shell fragments or otherconstituents. Carbonate rocks typically have a farwider range of grain shapes than most siliciclas-tic rocks. Clearly, several types of porosity maycoexist in a carbonate reservoir, ranging frommicroscopic to cave-sized, which makes porosityand permeability estimation and calculation ofreserves extremely difficult.4
Another feature of carbonate rocks is theirsusceptibility to dissolution. At the surface, aswater and carbon dioxide form carbonic acid,dissolution can lead to impressive karst topogra-phy, including sinkholes, caves and intricatedrainage patterns like “disappearing” streams inactive karst systems.5 Inactive karst systems, or
22 Oilfield Review
Mudstone Wackestone Packstone Grainstone Boundstone Crystalline
Less than10% grains
More than10% grains
Grain-supported Lacks mud and isgrain-supported
Originalcomponents werebound together
Depositionaltexture notrecognizable
Mud-supported
Contains mud, clay and fine silt-size carbonate
Original components not bound together during deposition
Depositional texture recognizable
Pore types
Intergranular, Intercrystalline Moldic, Interfossil, Shelter Cavernous, Fracture, Solution-enlarged fracture
> Carbonate rock classification. Carbonate rocks are differentiated by depositional texture,grain types, rock fabric or other factors. The Dunham classification, published in 1962, iswidely used to categorize carbonate rocks according to the amount and texture of grainsand mud. The classification by Embry and Klovan follows the Dunham scheme, but addscategories for rocks formed by organisms that grew together, such as colonies of oysters.Describing pore types further refines rock descriptions; Lucia’s classification is widelyaccepted. (Adapted from Dunham, in Ham, reference 2, and Lucia, reference 2.)
2. Geologists have developed many different carbonaterock classification schemes. Some are general schemes;others are specific to a reservoir, basin or region. For moreon carbonate rock classification:Embry AF and Klovan JE: “A Late Devonian Reef Tract onNortheastern Banks Island, N.W.T.,” Bulletin of CanadianPetroleum Geology 19, no. 4 (December 1971): 730-781.Ham WE (ed): Classification of Carbonate Rocks,American Association of Petroleum Geologists Memoir 1.Tulsa, Oklahoma, USA: AAPG, 1962.Lucia FJ: Carbonate Reservoir Characterization. New York,New York, USA: Springer, 1999.
3. Stylolites are interpenetrating, sutured surfaces that formduring diagenesis.
4. For more about permeability evaluation for reservoircharacterization: Ayan C, Douglas A and Kuchuk F: “A Revolution in Reservoir Characterization,” MiddleEast Well Evaluation Review no. 16 (1996): 42-55.
Baadaam H, Al-Matroushi S, Young N, Ayan C, Mihcakan Mand Kuchuk F: “Estimation of Formation Properties UsingMultiprobe Formation Tester in Layered Reservoirs,”paper SPE 49141, presented at the SPE Annual TechnicalConference and Exhibition, New Orleans, Louisiana,USA, September 27-30, 1998.Kuchuk F: “Interval Pressure Transient Testing with MDT Packer-Probe Module in Horizontal Wells,” paper SPE 39523, presented at the SPE India Oil and GasConference and Exhibition, New Delhi, India, February 17-19, 1998.Kuchuk FJ, Halford F, Hafez H and Zeybeck M: “The Use of Vertical Interference Testing to Improve ReservoirCharacterization,” paper ADIPEC 0903, presented at the9th Abu Dhabi International Petroleum Exhibition andConference, Abu Dhabi, UAE, October 15-18, 2000.
Lenn C, Kuchuk FJ, Rounce J and Hook P: “HorizontalWell Performance Evaluation and Fluid Entry Mechan-isms,” paper SPE 49089, presented at the SPE AnnualTechnical Conference and Exhibition, New Orleans,Louisiana, USA, September 27-30, 1998.
5. Karst was first recognized and described in the Dinariccarbonate platform of Yugoslavia, also known as theKarst region. Karst is found throughout the world.
6. Lucia, reference 2: 7.7. For more on the geologic history of Indonesia:
Netherwood R: “The Petroleum Geology of Indonesia,” in Indonesia 2000 Reservoir Optimization Conference.Jakarta, Indonesia: PT Schlumberger Indonesia,November 2000: 174-227.
Winter 2000/2001 23
paleokarst, may form reservoirs dominated byangular rock fragments produced during cavecollapse. For the oil industry, karst can be amixed blessing—caves can result in catas-trophic bit drops and fluid losses during drilling,but karst also can result in extremely high poros-ity and permeability.
Given the heterogeneity of carbonate rocks, itis not surprising that hydrocarbon production fromthese formations often is heavily influenced bythe presence of faults and fractures, particularlyin older Mesozoic and Paleozoic reservoirs.Experts caution that relationships between poros-ity and permeability in carbonate rocks cannot bedetermined without understanding the pore-sizedistribution (see editorial, “Linking Petrophysicaland Geologic Information: A Task for Petro-physics”).6 Because carbonate reservoirs presentenormous challenges, they have fueled substan-tial research efforts within Schlumberger and thepetroleum industry for decades. These efforts takemany different forms as experts struggle to solvedifficult problems in carbonate reservoirs.
Carbonate Evaluation in IndonesiaIntegrated carbonate evaluation is important in allstages of exploration and production. In 1997, anoperator drilled a well in the Sibolga basin, off-shore northwest Sumatra, to evaluate a carbon-ate buildup prospect identified in seismic data(above). A full petrophysical and stratigraphicanalysis of well logs and seismic data was thenundertaken to understand drilling results andreevaluate the prospectivity of this play.
Biostratigraphic analysis of well cuttings indi-cated the buildup formed in the Middle Miocene,around 13 million years ago, in a forearc settingsimilar to that of today, with consumption ofIndian Plate oceanic crust beneath Sumatraalong the Sunda trench. This was a period ofglobal eustatic rise.7
The well was evaluated using openhole wire-line logs—gamma ray, resistivity, density andneutron—and, because mud circulation problemsduring drilling prevented conventional coring, theFMI Fullbore Formation MicroImager tool alsowas run. Integration of wireline logs, especiallythe FMI image, with seismic data was key todetermining depositional facies. Prior to forma-tion of the carbonate buildup, massive shaleswere deposited in a low-energy offshore environ-ment. Subsequent laminated shale and cross-bedded sandstone were deposited as waterdepths became shallower and depositionalenergy increased. The prograding reef-front suc-cession was produced by smaller buildups thatcoalesced to form one large carbonate platform.Eventually, relative sea level rose rapidly and sub-merged the buildup (below).
The prospect was expected to contain bio-genic gas. Further study of well logs and imagelogs, however, showed that reservoir-quality car-bonate rocks formed almost continuously in theabsence of internal sealing rocks. Top seals onthe reservoir were deposited long after gas gen-eration, so any biogenic gas that was generatedwas not trapped. As a consequence, the com-pany decided against further appraisal and wasable to redirect its resources. Nevertheless, thisexample underscores the usefulness of integrat-ing all available data to develop reasonable 3Dgeological models of reservoirs as early as possi-ble in the exploration process.
Black lagoonal fill
Wave-resistant reef facies
Back reef from storm and talus deposits
M1
M3
SW NE
> Seismic interpretation. This seismic line from the prospect was flattenedon the M3 horizon, presumably a horizontal or nearly horizontal depositionalsurface. The prograding reef-front succession was produced by initialsmaller sized buildups that coalesced to form one large carbonate platform.Eventually, relative sea level rose, submerging the buildup.
MALAYSIA
ASIA
AUSTRALIA
CentralSumatra basin
SouthSumatra basin
Bengkulubasin
Pagar Jatigraben
Singkelgraben
NorthSumatra basin
SINGAPORE
SUMATRA
Mentawai fault zone
Keduranggraben
Pinigraben
Sibolgabasin
Sumatra fault zone
0 100 200 km
0 62 124 miles
VolcanoesVolcanic rocks
Sunda trench
Sumatra forearc basin
> Location of the Sibolga basin, Indonesia.
Carbonate Evaluation in West Texas, USAIn contrast to the previous example from theexploration stage, the Permian Basin of WestTexas, USA, is renowned for vast carbonatereservoirs, many of which are now undergoingsecondary and tertiary recovery. New technologyand methods greatly enhance production byoffering interpreters a better understanding ofhow reservoir heterogeneity affects performanceand which zones contribute to flow.8 Perhaps themost significant contributions come from nuclearmagnetic resonance (NMR), borehole-image andproduction logs.
When using the CMR Combinable MagneticResonance tool in carbonate formations, engi-neers in West Texas adjust the acquisition para-meters to compensate for longer polarizationtimes relative to clastic formations.9 Typical CMRlogging speeds in this region are 90 to 140 ft/hr [30 to 40 m/hr], as opposed to speeds of 300 ft/hr[100 m/hr] in clastic rocks. Increased cutoff valuesfor T2, more than three times the T2 cutoffs usedin sandstones, have been determined from labo-ratory measurements of cores and are applied tospecific fields by local interpreters. These stepsimprove the measurement of porosity, permeabil-ity and fluid saturation in rocks whose pore sizes,shapes and pore-throat connections vary morewidely than in most clastic rocks.
In addition to adjusting log-acquisitionparameters, use of different logging suitesallows more realistic interpretation of carbonatereservoirs. In West Texas dolomite formations,high gypsum content results in overestimation of porosity when using standard crossplottingtechniques. Integrating results from CNLCompensated Neutron Log, Litho-Density andCMR logs provides better estimates of porosityand permeability. If core data are not available,which is the norm, combining these logs with theECS Elemental Capture Spectroscopy sonde forgeochemical logging also can help quantify min-eralogy to obtain more accurate porosity. Addinga borehole image log, such as from the FMI tool,further improves understanding of pores, particu-larly vugs, which commonly are distributed irreg-ularly in carbonate reservoirs (left).
Because of the maturity and marginal eco-nomics of some West Texas fields, operatorsmust minimize data-acquisition costs. Since the
24 Oilfield Review
10,399
10,400
10,401
10,402
10,403
10,404
10,405
0 100
> Contrasts in West Texas carbonate rocks. These FMI images show clean,relatively homogeneous carbonate rock (top) and fractured, vugular carbonaterock with shale filling pores (bottom). ECS data displayed in Track 1 indicatevolumes of carbonate in blue, clay in brown and quartz in yellow.
10,391
10,392
10,393
10,394
10,395
10,396
10,397
0 100
8. For more examples from West Texas: Newberry BM,Grace LM and Stief DD: “Analysis of Carbonate DualPorosity Systems from Borehole Electrical Images,”paper SPE 35158, presented at the Permian Basin Oiland Gas Recovery Conference, Midland, Texas, USA,March 27-29, 1996.
10. Logan D, Strubberg C and Conner J: “New ProductionLogging Sensors Revolutionize Water/CO2 Conformancein the Pumping Wells of West Texas,” paper SPE 59716,presented at the Permian Basin Oil and Gas RecoveryConference, Midland, Texas, USA, March 21-23, 2000.This paper also discusses use of TDT Thermal DecayTime logs with the PS Platform tool to evaluate carbon-dioxide flooding.
9. For more on NMR applications in West Texas: LoganWD, Horkowitz JP, Laronga R and Cromwell D: “PracticalApplication of NMR Logging in Carbonate Reservoirs,”paper SPE 38740, presented at the SPE Annual TechnicalConference and Exhibition, San Antonio, Texas, USA,October 5-8, 1997.
Winter 2000/2001 25
cost of acquiring core may be higher than thecost of a wireline log, interpreters have cali-brated cores and logs to make sure that interpre-tations are robust, building confidence in loginterpretations when core data are unavailable.This is especially important when evaluating thepermeability of waterflooded reservoirs. The abil-ity to distinguish high-permeability zones allowsoperators to plug swept zones and improve flood-ing of unswept zones.
Customized solutions in West Texas includeproduction logging below electrical submersiblepumps.10 In one field, engineers from Schlumbergerand an operating company were able to evaluatefluid entry in several wellbores by adapting the PSPlatform tool for use below the pump. Custom-built G-plates were installed above and below thepump to guide the wireline and pump cables toavoid tangling around the tubing and employing amodified wellhead assembly.
As one producing well was logged with thePS Platform tool below the pump, it becameapparent that oil entered from an interval abovethe reservoir section, and that the zone of inter-est was actually producing water. Evaluationwith the FMI tool revealed two thin, porous zonesuphole that were contributing the oil flow(above). By using a bridge plug to shut off thewater from the waterflooded reservoir section,the operator saved significant water recyclingcosts while increasing the oil production from thezone above. There were additional savings in off-set wells because acid stimulations were nolonger performed in water-prone zones. As aresult of these types of experiences, operatorsare striving for earlier identification of high-per-meability water conduits.
Carbonate Case Studies at Schlumberger-Doll ResearchResearchers at Schlumberger-Doll Research(SDR), Ridgefield, Connecticut, USA, have fol-lowed many different paths, from complextheoretical methods to simpler approaches thatemphasize well-by-well evaluation. The com-mon goal, however, has been to develop inter-pretations that can be incorporated intofield-scale solutions.
Any improvement in recovery from giantcarbonate reservoirs has a tremendous impact onoil and gas supplies. Reservoir heterogeneitycomplicates everything from drilling to well com-pletions, including petrophysical evaluation, sodeveloping a reliable log-based interpretationmethodology is essential for field development.Reservoir heterogeneity prohibits relating poros-ity and permeability directly, as might be donewhen analyzing relatively homogeneous reser-voirs. Therefore, it is crucial to distinguish car-bonate lithologies and rock fabrics to optimize
> Complementary logging results. Production logging in this West Texas well showed that oil entered from zones above the zone of interest, and that thezone of interest was actually producing water. Evaluation with the FMI tool revealed that two thin, porous zones around 4660 ft contributed the oil flow. Thedark lines in the image log are leached bedding planes.
4650
4660
4670
Water rate
Spinner
0 25
Bubble count
counts/sec
cycles/sec
0 30MD
1:400 ftGamma
ray Probe 1 R8
API deg0 1021
0
360
Caliper Y
in.3 6
Caliper X
Holdup
1
0Bubble
in.3 6
Oil rate
B/D0 1500
B/D0 1500
B/D0 4500
B/D0 300Total flow rate
Amplifiedoil rate
Temperature
°F 105
Pressure
psi75 325
Oil holdup
Oil holdup
m3/m30.75 1
WF densityg/cm30.95 1.15
4600
4650
4700
30% Porosity -10%Depth,
ft
production, whether one is deciding how to dealwith a single well or preparing to simulate pro-duction from an entire field.
Work at SDR in the 1990s led to an integratedcarbonate-evaluation methodology for theThamama formation, a lower Cretaceous reser-voir in the Middle East.11 This methodology wasapplied to studies of other carbonate reservoirsin the UAE and West Texas. Recognizing the widevariety of carbonate rocks worldwide, SDRresearchers decided, in 1997, to embark on aseries of additional studies. Several carbonatecase studies have been, or are being, conductedjointly by SDR scientists and engineers with theircounterparts from operating companies.
Investigations of two giant fields, the BombayHigh field offshore India and a field in the MiddleEast, indicate that the variety of rock types andheterogeneity within a given carbonate reservoirlend themselves to customized, formation-specific evaluations, particularly in cases ofextreme diagenetic alteration. Both studies, com-pleted in 2000, draw on techniques that rangefrom conventional petrophysical and petro-graphic analysis to the first application of a new
laboratory NMR method, called decay due to dif-fusion in internal field (DDIF).
Bombay High study—The giant Bombay High field, offshore western India, covers about1200 km2 [463 sq miles] and has more than 600 development wells. Discovered in 1974 byOil & Natural Gas Corporation, Ltd. (ONGC), thefield was brought on production in 1976. Themain pay is the Miocene L-III limestone, a reser-voir with ten hydrocarbon-bearing layers sepa-rated by shale, clay-rich limestone and tightlimestone. The layers are not continuous andhave poor vertical communication. As of April2000, the field produced 297 MMT [327 M tons]crude oil and 110 BCM [3.9 TCF] natural gas, andis now in the mature phase. A redevelopmentprogram has been prepared to improve recovery.
ONGC sought better understanding of reser-voir petrophysics to control water breakthroughin heterogeneous, layered carbonate rocks thathave undergone waterflooding since 1984.12 Theprimary reservoir generally is not fractured, soONGC suspected that a few high-permeabilityzones were contributing to water breakthrough.
The challenge, therefore, was to develop a con-sistent log-based method to identify these high-permeability zones. For the Bombay High study,61 core plugs from the N5-9 well were evaluatedalong with the logs.
Middle East study—Scientists and engineersfrom a Middle East operating company and SDRevaluated the complexities of a giant gas field thatproduces from prolific carbonate rocks. Wirelinelogs and 80 core plugs from one well form theframework for an integrated interpretation.
Researchers applied much of the same ana-lytical methodology in both cases. At the outset,both operators believed that the volume of clay(Vclay) would prove to be the key problem that thestudies should solve. Accurately quantifying theabundance of clay minerals is essential to accu-rate porosity and saturation calculations, whichin turn affect reserve estimates.
Quantitative mineralogical and chemicalanalysis of core samples conducted at SDRenhanced petrophysical analysis of the reservoirs.Mineralogy was evaluated using a proprietaryFourier transform infrared (FT-IR) technique that
26 Oilfield Review
Tota
l cla
y, w
t%
100
80
60
40
20
0
100
80
60
40
20
00 50 100 150
Gamma ray, API
Tota
l cla
y, w
t%
100
80
60
40
20
00 100 200 300
Gamma ray, API
0 5 10 15Aluminum, wt%
100
80
60
40
20
00 5 10 15
Aluminum, wt%
> Uncertainty in clay content. Because of clay-volume concerns, mineralogy and chemistry of MiddleEastern (top) and Bombay High (bottom) carbonate rocks were analyzed. Gamma ray responses, com-puted from the chemical analysis of thorium (Th), uranium (U) and potassium (K), do not correlate wellwith clay content in either case. A much better correlation can be made with aluminum (Al), and formsthe basis of the SpectroLith clay-volume computation.
Winter 2000/2001 27
relates infrared absorbance spectra to 50 mineralstandards from silicate, carbonate, clay and othermineral families.13 Chemical analysis included X-ray fluorescence, neutron activation and induc-tion-coupled mass spectrometry. All of theseresults were integrated with log data. An impor-tant result of the core analysis was that gammaray logs alone would have indicated an incorrectclay content in both reservoirs (previous page).
Therefore, a method to accurately determinemineralogy without core analysis is criticallyimportant for future reservoir characterization.
The ECS logging tool allows accurate estima-tion of mineralogy, clay concentration and lithol-ogy, and also can be used to evaluate total andeffective porosity and hydrocarbon type.14 The ECStool uses a spectrometer to measure concen-trations of certain elements—calcium, silicon,
sulfur, iron, titanium, gadolinium, sodium andmagnesium—that reflect the concentrations ofspecific minerals in the formation. The data can beanalyzed to compute mineralogy in terms of sand,clay, evaporite and carbonate minerals bySpectroLith lithology processing. In both cases,the SpectroLith-processed ECS results give a morerealistic picture of the mineralogy, as confirmed bymineralogical analysis of cores (above).
> SpectroLith-processed ECS data for accurate mineralogy that can be con-firmed by core analysis. In a Middle Eastern formation (top), the processedECS logs of carbonate, anhydrite, clay and sand content correlate well withcore data (red circles). The Bombay High results (bottom) show good agree-ment between core data and processed ECS data for carbonate, clay, pyriteand sand content, with slight discrepancies resulting from core plugs thatsampled thin shale beds not seen at the resolution of the log.
11. Ramakrishnan TS, Rabaute A, Fordham EJ,Ramamoorthy R, Herron M, Matteson A, Raghuraman B,Mahdi A, Akbar M and Kuchuk F: “A Petrophysical andPetrographic Study of Carbonate Cores from theThamama Formation,” paper SPE 49502, presented atthe 8th Abu Dhabi International Petroleum Exhibitionand Conference, Abu Dhabi, UAE, October 11-14, 1998.
12. Tewari RD, Rao M and Raju AV: “Development Strategyand Reservoir Management of a Multilayered GiantOffshore Carbonate Field,” paper SPE 64461, presentedat the SPE Asia Pacific Oil and Gas Conference andExhibition, Brisbane, Queensland, Australia, October16-18, 2000.
XX400
0 50Carbonate, %
100 0 50Anhydrite, %
100 0 50Clay, %
100 0 50Sand, %
100
0 50Carbonate, %
100 0 50Clay, %
100 0 50Pyrite, %
100 0 50Sand, %
100
XX500
XX600
XX700
XX800
Dept
h, ft
XX900
XX000
XX100
XX200
1340
1360
1380
Dept
h, m
1400
1440
1460
1420
13. Herron MM, Matteson A and Gustavson G: “Dual-rangeFT-IR Mineralogy and the Analysis of SedimentaryFormations,” paper 9729, presented at the 1997 Societyof Core Analysts Conference, Calgary, Alberta, Canada,September 7-10, 1997.Matteson A and Herron MM: “Quantitative MineralAnalysis by Fourier Transform Infrared Spectroscopy,”paper 9308, presented at the 1993 Society of Core AnalystsConference, Houston, Texas, USA, August 9-11, 1993.
14. Herron SL and Herron MM: “Application of NuclearSpectroscopy Logs to the Derivation of Formation MatrixDensity,” Transactions of the SPWLA 41st AnnualLogging Symposium, Dallas, Texas, USA, June 4-7, 2000,paper JJ.
Another key goal of these integrated studiesis to identify and understand the various poretypes, including micropores, mesopores andmacropores, and the effect of their distribution onproduction (above). Micropores, with pore-throatdiameters less than 0.5 micron, usually containmostly irreducible water and little hydrocarbon.Mesopores, with pore-throat diameters between0.5 and 5 microns, contain significant amounts ofoil or gas. Macropores, with throats measuringmore than 5 microns in diameter, are responsible
for prolific production rates in many carbonatereservoirs, but often provide pathways for earlywater breakthrough, leaving considerable gas andoil behind in mesopores. NMR logging hasimproved evaluation of porosity, pore-size distri-bution and bound fluids (next page).
NMR logging tools, such as the CMR device,use large magnets to strongly polarize hydrogennuclei in water and hydrocarbons as they diffuseabout the pore space in rocks. When the magnet
is removed, the hydrogen nuclei relax. The relax-ation time, T2, depends on the pore-size distribu-tion; larger pores typically have longer relaxationtimes. Tar and viscous oils relax more quicklythan light oil or water. The variations in relax-ation time produce a T2 distribution from whichfluid components and pore sizes are interpreted.
An important outgrowth of the studies wasthe ability to classify pores in the three size cat-egories using NMR. This success resulted fromthe discovery that, in contrast with the earlier
28 Oilfield Review
1
2
3
4
> Understanding the distribution of micro-, meso- and macropores, a key goal of integrated reservoirstudies. Photomicrographs and scanning electron microscope (SEM) images of thin sections of BombayHigh carbonate rocks illustrate the three pore types. The enlarged thin section view (top) shows thelocations of the numbered SEM images. Blue epoxy has been injected into the sample to highlight theporosity. SEM images include 400-micron scale bars, except for the 25-micron scale in the bottomright image. The SEM image of Location 1 (middle row, left) reveals a black macropore at lower left,and meso- and micropores in the dark gray region. The SEM image of Location 4 (middle row, center)shows mesopores. The next image (middle row, right) includes the lower left portion of Location 1, asseen by the macropore surrounded by euhedral calcite crystals—crystals whose growth has not beenrestricted by adjacent grains. The bottom left image includes the upper right part of Location 3, butshows leaching around the edge of the dark macropore, and micropores are barely visible as darkdots. The bottom right image is an enlargement showing details of the micropore system in Location 3.
Winter 2000/200129
carbonate rocks studied, T2 decay-time distribu-tions are directly useful for interpretationbecause diffusive coupling is not a problem.Diffusive coupling results when spinning protonsmove between micro- and macropores during themeasurement, blurring the T2 distribution.15
A new technique developed at SDR allowsresolution of the three common pore sizes usingspectra quantified in size rather than relaxation
time. The new method, DDIF, delivers a quan-titative pore-size distribution that is especiallypowerful in carbonate rocks.16 A laboratory mea-surement technique with its own associatedprocessing, DDIF is separate and distinct fromthe conventional NMR T2 experiment. The newinsight derived from DDIF studies is that the con-ventional T2 distributions resemble the DDIF
0.0010.0000
0.0020
0.0040
0.0060
0.0080
0.0100
0.0120
0.01 0.1 1 10 100 1000 0.0010.0000
0.00020.0004
0.0006
0.00080.0010
0.0012
0.0014
0.0016
0.0018
0.01 0.1 1 10 100 1000
0.00010.0
1.02.0
3.04.0
5.0
9.0
0.001 0.01 0.1 1 10 100
6.0
7.0
Ampl
itude
Mer
cury
dis
tribu
tion
Pore-throat diameter, microns Pore-throat diameter, micronsM
ercu
ry d
istri
butio
nAm
plitu
de
8.0
0.00010.0
1.0
2.0
3.0
4.0
5.0
6.0
0.001 0.01 0.1 1 10 100T2 relaxation, sec T2 relaxation, sec
> Bombay High samples containing meso- and macropores (top left) and micro-, meso- and macropores(top right). The plots show distributions of pore-throat diameters and T2 distributions for each sample.Pores are assigned to pore types by their pore-throat diameter as measured by mercury injection(top two plots). Pores to the left of the red lines are micropores, those between the red and blue linesare mesopores, and macropores lie to the right of the blue lines. Comparison with the T2 distributions(lower plots) shows that the porosity partitioning can be accomplished using T2 cutoffs, a valuableapplication of NMR logging in carbonates.
15. For more on NMR technology, including permeabilitytransforms and NMR in carbonate rocks: Allen D, Flaum C,Ramakrishnan TS, Bedford J, Castelijns K, Fairhurst D,Gubelin G, Heaton N, Minh CC, Norville MA, Seim M,Pritchard T and Ramamoorthy R: “Trends in NMRLogging,” Oilfield Review 12, no. 3 (Autumn 2000): 2-19.
16. Song YQ, Ryu S and Sen P: “Determining Multiple LengthScales in Rocks,” Nature 406, no. 6792 (July 13, 2000):178-181.
distributions. This confirms that there is no diffu-sive coupling, so T2 distributions are valid fordistinguishing pore sizes (above).
Scanning electron microscope (SEM) imageshelped explain the absence of diffusive couplingin both cases (above right). Thus, in both forma-tions, the T2-distribution shape is similar to pore-size distribution determined by mercury injectionand DDIF. Conventional T2-based NMR analysis,discussed in detail below, was applied to determine both pore-size distributions and per-meability. An important result of the studies ismore realistic calculation of permeability using CMR logs.17
In the Bombay High field, CMR data confirmedgenerally low permeability with numerous high-permeability streaks in leached, macroporous
zones. The Timur-Coates permeability transform,which uses total porosity and the ratio of free-fluid volume to bound-fluid volume to com-pute permeability, was selected to determinepermeability using CMR data because it correctlypartitions the pore network found in theseleached, macroporous limestones. Because high-permeability streaks are so important to produc-tion and because hydrocarbon signal masksmacropores on CMR logs, FMI data were incor-porated into the log processing (next page).
The SDR equation, which relates permeabilityto the logarithmic mean of T2 and total porosity,was used to determine permeability from CMRdata for the Middle East well. In dolostones,more realistic permeability estimates were made
using NMR T2 values from the log and the corerather than using a porosity-permeability rela-tionship alone. Permeability estimates in thelimestones, which had more variable pore sys-tems than the dolostones, also improved, but notas dramatically. The more accurate permeabilitycalculations used a correction factor based onthe temperature sensitivity of NMR T2 values ineach formation.
Three different NMR T2 cutoff values wereused in this well, allowing NMR logs to be usedto determine micro-, meso- and macroporosity.The ratio of NMR T2 values to pore-throat diam-eter determined by mercury injection (NMRT2/throat) in 22 samples also showed three dis-tinct NMR T2/throat classes that correspond tofabric classes observed in thin-section analysis.
30 Oilfield Review
> Pore size analysis by DDIF, NMR T2 distribution and mercury injection. In the top row, DDIF spectra(red) are used to determine whether the NMR T2 distribution (blue) truly reflects the pore-size distributionby comparing the two spectra. The horizontal axis of the T2 distributions has been multiplied by 100 tofacilitate the overlay. For these three samples, the match is excellent. The two first samples are dolo-grainstones; the third is a sucrosic dolomite. The lower plots compare the mercury-injection distributions(blue) with the DDIF distributions (red). Mercury porosimetry uses mercury injection to determine thecapillary pressures of the connected pore space. The plots obtained from these data are interpretedas the pore-throat sizes. On the other hand, DDIF measures the pore openings, including pore bodiesand throats. Superimposing the two results reveals the connectivity of the pore network. For thesucrosic dolomite (right), the overlay reveals a network consisting of pore bodies with a diameter of20 microns connected by throats of 1 to 2 microns. For the two grainstones, the pore-body size is largerand covers a broader range. They share a network of pore throats with a diameter of 2 microns;however, the second sample (middle) exhibits a bimodal system, with a very fine network of porethroats with 0.1 micron diameters.
> SEM image showing a macropore (large blackarea at lower left) within peloidal grainstone(gray area). Micropores appear as speckledpatches in peloids. Inverted-v-shaped cementseparates the intergranular pore from microporesand produces a distinct NMR signature becausethe cement prevents diffusive coupling. The scalebar measures 50 microns.
17. Allen et al, reference 15: 7-8.
0
0.1
0.2
0
0.1
0.2
0
0.1
0.2
0
0.1
0.2
0
0.1
0.2
0.1
0.001 0.01 0.1 1 10 100 1000
1 10 100 1000
0
0.1
0.2
Dist
ribut
ion
Dist
ribut
ion
Pore diameter, microns, and shifted T2, msec
Pore diameter, microns
DDIFT2
DDIFT2
DDIFT2
DDIFMercury
DDIFMercury
DDIFMercury
Winter 2000/2001 31
Improved prediction of permeability optimizeswell placement and production, particularly fordirectional or extended-reach wells. The ability todistinguish pore types allows successful comple-tion of producible hydrocarbon zones. The methodalso helps engineers anticipate which layersmight be prone to early water breakthrough.
Integration of ECS and CMR logs with con-ventional log suites and core data resulted in themost rigorous interpretations of Middle Easternand Bombay High carbonate fabrics and diage-netic histories to date. More importantly, thedetailed joint studies provide an improved frame-work for ongoing interpretation problems in bothregions. The study groups recommend that newwells be evaluated similarly to the wells in thejoint studies. The optimal logging suite includesCMR and ECS logs in addition to routine resistiv-ity, gamma ray, density and neutron logs.
Confidence in log-based interpretation willcontinue to grow as more wells in these fieldsand other fields that produce from similar forma-tions are evaluated. Greater confidence in single-well interpretation is critical to ongoing reservoircharacterization and simulation because it is noteconomically viable to acquire core samples fromevery well. Integrated core-log studies providesignificant benchmarks for analysis of field wellslacking cores.
Both studies promoted close collaborationbetween research and operations personnel thatstrengthened working relationships, makingfuture joint research more likely. The improvedunderstanding of the reservoirs resulting fromthe research efforts can be applied to operationsimmediately. Based on research findings, toolsdeveloped for oil reservoirs can be tailored foruse in evaluating gas-filled rock.
Some results of carbonate case studies at SDRcan be carried over to clastic reservoir studiesbecause there are analogies between carbonaterocks and certain clastic reservoirs. For example,ongoing work in sandstones confirms the presenceof micropores associated with grain-coating claysand partially dissolved grains. Clearly, researchersand operations personnel may benefit from shar-ing nonconfidential results of their work.
Ongoing studies in the Middle Eastern reser-voir include seismic imaging with the single-sensor Q system to better characterize thereservoir and optimize drilling targets.
Benefits of the Bombay High study includegreater understanding of the L-III reservoir, par-ticularly heterogeneity and its effects on fluidtransport; development of a rigorous petrophysi-cal approach; and evaluation of the applicability
of the new methodology to older, less extensivedata sets. ONGC has recognized the importanceof ECS and CMR data for clay estimation. Theseresults will be incorporated into future produc-tion strategies.
m3/m3
m3/m3
BFV.3_le
m
MD1:240
1355
1360
1365
1370
1375
1380
1385
1390
1395
Total porosityProd.Wtr
0.5 0 msec0.3 3500 mD
Permeability FMI image
1 1000 Resistive Conductive
m3/m3
Effective porosity
0.5 0
1 0
mD
Minipermeameter
1 1000
mD1 1000
m3/m3
Vug %
0.5 0
0 120 240 360
m3/m3
CMR T2 Dist.
0 29
m3/m3
Volumetricanalysis
0 1 m3/m3 00.5
m3/m3 00.5
Hg Macro
Core perm Core perm Horizontal scale:1:14.835
Orientation north
CMR T2 Dist.
T2 cutoff
0 29
Clay-bound water
Macroporosity
Mesoporosity
Microporosity
Irr.Wtr
Clay
Boundwater
FMI macro
> Comparison of L-III log and core data for identification of high-permeability streaks. The lithology in the first track—clay, quartz, calcite and dolomite—is computed using ELAN analysis software withECS data as the key input. The fluids are reported as oil that has not been moved by invasion (green),oil that was moved (orange), irreducible water contained in micropores (blue with black dots) andmobile water (white). NMR data help distinguish irreducible and mobile water. The second track is the porosity broken down into clay-bound water from ECS data, microporosity from CMR data andmeso- and macropores from CMR and FMI logs. The third track contains the T2 distributions from theCMR log. The solid blue permeability curve in Track 4 is calculated from ELAN volumes. The light bluedots are permeability measured on core plugs. The black line is permeability measured with 1-cmsampling on the core slab face using a minipermeameter. The macroporosity computed from FMI data(shown in Track 6) is displayed in red in Track 5. Light blue dots indicate macroporosity determined by mercury injection in core samples. The black line represents vugular porosity measured on theface of the slabbed core.
Integrated Carbonate Evaluation at the ONGC-Schlumberger Joint Research CenterCarbonate reservoirs in India present importantinterpretation challenges to scientists and engi-neers working at the Joint Research Center(JRC), a combined effort by the Oil and NaturalGas Corporation, Ltd., (ONGC) and Schlumberger.The JRC, located in New Delhi, was establishedin the 1980s to investigate formation evaluation,reservoir description, production, completion andreservoir-monitoring problems experienced byONGC and to find solutions to those problems.Several noteworthy carbonate reservoirs arelocated offshore Mumbai, India, including theNeelam field, which JRC personnel have studiedsince its discovery and first production in 1990.
At the JRC, petrophysical, geophysical andgeological evaluations of carbonate reservoirsprovide the basis for an integrated reservoir solu-tion. The ultimate goal is to maximize oil recov-ery and production efficiency by understandingand modeling the reservoir. This approach alsominimizes the number of well interventions andthe number and location of wells required so thatall commercially viable oil pools are drained. Bybuilding a numerical simulation model of thefield, geoscientists and engineers can extrapo-late field behavior over time and evaluate “whatif” scenarios, such as how a given workover pro-gram might affect overall field performance and
production, or whether the failure to drill specificdevelopment wells might leave compartments ofundrained hydrocarbon.
In the case of a mature field like Neelam, thefirst phase in building a simulation model is tocalibrate it to reproduce the historical productionbehavior of the reservoir—history matching.Since this stage conditions the reservoir model tothe dynamic data, such as well-production ratesand changes in pressures and saturations, thehistory-matched model becomes a much morerepresentative description of the reservoir thanthe input static model.
To correctly model flow behavior in carbonatereservoirs, it is essential to understand the per-meability profile. Standard log data—density,neutron, sonic, gamma ray, SP and resistivitylogs—evaluated by conventional methods all toooften indicate a homogeneous reservoir. Porosityvariations are not a reliable indicator of perme-ability variations because changes in carbonatetexture affect permeability more than changes inporosity affect permeability. The time-honoredmethod of using core data to derive a porosity-permeability relationship associated with a par-ticular reservoir fails when reservoir-rock texturevaries. Although the technique is fundamentallycorrect, it should be carried out separately foreach carbonate depositional rock type or texture.In fact, previous studies of the Neelam field haveshown that permeability increased as porosity
decreased—a difficult conclusion for petrophysi-cists to reconcile with their interpretations.
Many carbonate reservoirs contain localizedor extended layers of mud-supported rock, wherepermeability is appreciably reduced, but com-plete barriers to vertical fluid migration are rare.During the millions of years of reservoir evolu-tion, fluids segregated, resulting in a water tableat the bottom, a transition zone where both oiland water volumes are movable, and an oil zoneat the top, where the water is entirely capillary-bound and only the oil is movable. Pressures alsoequalize in the reservoir over this period.
It is only by careful inspection of core data, orthrough innovative evaluation of NMR or bore-hole image logs, that the texture of the carbonatereservoir becomes apparent as distinct zoneswith varying degrees of carbonate-mud supportand fluid-transport properties. Grainstone, oftenthe least porous, generally yields the highest per-meability among carbonate rock types. As mudcontent increases, resulting in packstone orwackestone, total porosity usually increases, butpermeability is perhaps 10 to 100 times lowerthan in grainstone due to the increased impor-tance of microporosity in associated muds.
These texture differences do not necessarilycreate true fluid-flow barriers over geologic time.However, when reservoir fluids are subjected to“instantaneous” withdrawal from the forma-tion—for example production over perhaps 5 to20 years as opposed to the millions of years ittook the reservoir to form—the resulting pres-sure pulses create separate flow units within thereservoir separated by those zones of significantpermeability reduction. This usually results inlarge pressure differences between flow unitsand complete breakdown of the smooth water-to-oil transition with decreasing depth. Fingers ofedge water propagate laterally, at any depth, intothe most permeable sections.
To complicate matters further, the permeabilityof a carbonate reservoir often is severely affectedby tectonic and diagenetic phenomena. For exam-ple, extremely high-permeability layers, called“super-k” layers, commonly result from diageneticalteration. Most of the available data for theNeelam reservoir imply that super-k layers are cre-ated by dissolution and leaching of the rock fabricby meteoric water during periods of low sea level,when the carbonate was exposed to water fromthe atmosphere and at the Earth’s surface.
Having an accurate permeability descriptionsignificantly accelerates the history-matchingprocess and greatly improves the reliability ofpredictions from the history-matched model.
32 Oilfield Review
Packstone
Grainstone
Packstone
Wackestone
Grainstone
> Texture and permeability analysis in an openhole log from the Neelam field. Track 1 displays ELANeffective reservoir porosity output, including immovable oil (green), movable oil (orange), movablewater (white) and capillary-bound water (light blue). Track 2 adds lithology analysis to the effectiveporosity output, scaled from 1 to 0. Gray represents shale; mud-supported limestone is gray-blue;grain-supported limestone is light blue; and other carbonate material is shown in deep blue. Track 3displays CMR T2 signal distributions, which correlate with carbonate textures (photomicrographs at left).
Winter 2000/2001 33
Because history matching is a complex, multi-variate process, it is sometimes possible toachieve what appears to be a satisfactory matchof historic data with an inaccurate model of thereservoir permeability distribution. In this case,the model will deliver inaccurate predictions.Only by unraveling the general reservoir perme-ability distribution can a realistic and usablereservoir-simulation model be built.18
Geoscientists and engineers at the JRCdecided to focus on mapping permeability using four complementary approaches. Whileeach approach begins at a wellbore, the resultsfrom each well must be integrated into a three-dimensional model of the field to deliver maxi-mum value to the operator. These approachesinclude the following:• NMR analysis to evaluate rock texture and
permeability profiles• cased-hole saturation logging to compare
original fluid saturations with saturationsafter some period of production to develop adepletion profile
• use of proportion curves and other geostatisti-cal tools to highlight hidden correlations thatcan be confirmed in key wells by either of thetwo previous methods
• geostatistical analysis of water breakthroughin historical well production data to evaluatehigh-permeability layers that conduct reser-voir or injection water.
The geostatistical techniques are still in experi-mental stages.
Texture and permeability analysis with open-hole logs—During field development or infilldrilling, operators have the opportunity to acquirenew openhole data. In the past, carbonate geol-ogists relied on image logs to reveal carbonate
textures, from which they infer permeability.More sophisticated techniques now are beingadded to image-log analysis to assess perme-ability. Confirming findings shown previously in alaboratory setting and from computer modelingby Ramakrishnan and others, JRC geoscientistsobserved that the width of the borehole NMR T2
signal distribution is strongly correlated to car-bonate lithology.19 Petrographic and core analy-ses corroborate JRC findings (previous page).20
This information can be used to calibrate theNMR permeability response to obtain an accu-rate, continuous permeability profile.
Previously, deriving permeability from NMRwas challenging because of the variable and ill-defined T2 cutoff between capillary-bound andfree fluids. The JRC-developed method first usesthe Schlumberger-Doll Research permeability for-mulation, usually referred to as kSDR. This rela-tionship, also used in the Middle East studydescribed earlier, defines permeability as a func-tion of porosity and the log-mean value of theNMR T2 distribution independent of a T2 cutoff.JRC scientists observed a well-defined depen-dence of the premultiplier in this relationship onrock texture, so they introduced a texture-relatedterm into the KSDR relationship. They confirmedthe accuracy of the method by comparing thetrend of NMR-derived permeability with brine-corrected core permeability data. The agreementbetween the texture and permeability estimatesfrom this technique and the results of an exten-sive core study is reasonable given the uncer-tainty in the permeability results caused bycarbonate heterogeneity.
Meaningful reservoir production forecastsrequire an accurate knowledge of the respec-tive volumes of free oil and free water, so JRC
engineers obtained free water by inverting theTimur-Coates permeability relationship and equat-ing it to the texture-related permeability measure-ment. This splits the total water—defined simplyas effective porosity minus hydrocarbon volume—into free- and capillary-bound waters.
Saturation in carbonate reservoirs cannot bederived from a simple Archie relationship. It iscommon to encounter oolithic molds or solutionvugs that affect the cementation factor m used inthe Archie relationship.21 For years, carbonateenthusiasts have known that a “variable m”approach is required. The difficulty resides inproperly partitioning the total porosity betweenprimary, matrix and vugular porosity.
A method first introduced by Brie and others in1985 makes use of an acoustic-scattering modeldeveloped previously by Kuster and Toksöz to eval-uate this partitioning.22 The technique uses a totalporosity from density, neutron, or both logs, andcompressional and shear velocities from soniclogs. An iteration technique adjusts the amount ofvugular porosity necessary to minimize the errorbetween expected theoretical sonic compres-sional and shear transit times and measured val-ues. Once partitioning of the porosity is evaluated,an equivalent approximation for electric propertiesprovided by the Maxwell-Garnett model is used toassess the effect of conductive or isolated inclu-sions on the cementation factor.23 A variable mvalue is obtained to use in ELAN Elemental LogAnalysis calculations to obtain a much more accu-rate volume of hydrocarbon. While other studieshave used variable values for m, this is perhapsthe first study in which the method has been vali-dated against individual core measurements of min the laboratory (above).
2.30
2.20
2.10"m
" cor
e
"m" log
2.00
1.901.90 1.95 2.00 2.05 2.10 2.15 2.20 2.25 2.30
> Cementation factor. Values of the cementation factor, m, derived from well-logdata using the Kuster-Toksöz model and core plugs measured in the laboratory,vary from 1.95 to 2.20. Laboratory measurement of m confirms that log-derivedvalues are reasonable and ultimately result in more accurate predictions ofhydrocarbon volume.
18. A full discussion of reservoir simulation is beyond thescope of this article, but will be covered in a futureOilfield Review article.
19. Ramakrishnan et al, reference 11.20. Olesen JR, Dutta D and Sundaram KM: “Carbonate
Reservoirs Evaluation with Advanced Well-Log Data,”presented at the 4th International Petroleum Conferenceand Exhibition, New Delhi, India, January 9-12, 2001; and also Extended abstract, presented at the AAPGInternational Conference and Exhibition, Bali, Indonesia,October 15-18, 2000.
21. Ooids are small, round grains of calcium carbonatelayers around a sandy nucleus. Oolithic molds are thespherical holes that remain when ooids dissolve.
22. Brie A, Johnson DL and Nurmi RD: “Effects of SphericalPores on Sonic and Resistivity Measurements,” Trans-actions of the SPWLA 26th Annual Logging Symposium,Dallas, Texas, USA, June 17-20, 1985, paper W.Kuster GT and Toksöz M: “Velocity and Attenuation ofSeismic Waves in Two-Phase Media: Part I, TheoreticalFormulations, Part II, Experimental Results,” Geophysics39, no. 5 (October 1974): 587-618.
23. Maxwell-Garnett JC: “Colours in Metal Glasses and inMetallic films,” Philosophical Transactions of the RoyalSociety of London 203 (1904): 385.Sen PN, Scala C and Cohen MH: “A Self-Similar Modelfor Sedimentary Rocks with Application to the DielectricConstant of Fused Glass Beads,” Geophysics 46, no. 5(May 1981): 781-795.
The petrophysical evaluation resulting fromthe combination of the improved capillary-bound,free-water and oil-volume derivation has beencompared with the results of an extensive, MDTModular Formation Dynamics Tester-derived pres-sure-profile analysis and well-test data (above).
Analysis of depletion profiles—In developedfields, operators commonly acquire new datathrough casing.24 In these cases, JRC members
have taken advantage of the RSTPro answer-product line to improve remaining oil-saturationestimates from the RST Reservoir Saturation Tooldevice to a level of accuracy that allows directcomparison with original openhole saturation.25
This permits derivation of a depletion profile thatclearly defines three kinds of reservoir zones:those with no apparent depletion, which areprobably low-permeability, mud-supported rocksthat separate flow units within the reservoir;
partially depleted zones that consist of “normal”reservoir rock; and zones of extreme depletion,which may be super-k layers or zones containingmassive solution channels.
A demonstration of these three zones comesfrom the Bassein reservoir in the Neelam field,where the zones were correlated over a distanceexceeding 6 km [3.7 miles]. In every well studied,a combination of the RST depletion profile and aborehole temperature survey acquired duringproduction highlights producing zones in perfo-rated intervals and reveals a reservoir separatedinto three major flow units (next page, left). Alarge amount of oil is still present in the upper-most unit, but virtually no production occurs fromthis unit because pressure depletion here is moresevere than in the lower units.
To improve cased-hole determination ofremaining oil volume in carbonates using the RSTtool, it is crucial to understand RST sensitivity tocompletion, especially cement, conditions. In sili-ciclastic rocks, cement thickness has a smalleffect on the length of the segments of the RSTsaturation-evaluation quadrilateral displayed incrossplots of carbon/oxygen ratios (C/O) fromnear and far detectors (NCOR versus FCOR). The quadrilateral and the crossplotted C/O ratiosare used to determine fluid saturations (nextpage, right).
Borehole geometry, formation lithology,porosity and the carbon density of the hydrocar-bon define the end points of the saturation-evaluation quadrilateral. The lower left corner,WW, is where both the borehole and the forma-tion are water-bearing. Travelling clockwise, theWO point indicates water-bearing borehole andoil-bearing formation. At the upper right is theOO point, oil in both borehole and formation.Finally, the OW point represents oil in the bore-hole and water in the formation. The exact posi-tion of these four points is obtained in controlledlaboratory conditions.26
In carbonates, the entire evaluation quadri-lateral is translated due the additional carbonand oxygen in the carbonate matrix. The amountof translation of the evaluation quadrilateral isrelated to the amount of carbonate rock sur-rounding the tool and also the distance betweenthe tool and the carbonate material.
Intuitively, the effect will be greatest in asmall borehole, with the tool and the rock matrixseparated only by the casing. As hole size andcement thickness increase, the tool is lessaffected by the carbonate rock. At the limit, for a
34 Oilfield Review
> Pressure profile at the top of the transition zone. Improved petrophysical evaluation techniquespredict capillary-bound, free-water and oil volumes more realistically and have been compared withthe results of an extensive, MDT-derived, pressure profile analysis and proven against a set of well-test data. In this example, CMR-derived fluid predictions (right) were confirmed by oil production during well tests.
> Pressure profile at the top of the water zone. Conventional log calculations using a constant valuefor m indicated an oil-bearing zone, but pressure-profile evaluation using both the MDT Modular Formation Dynamics Tester tool and well-test results proved that the zone is water-bearing, aspredicted by the JRC petrophysical evaluation methodology (right).
True
ver
tical
dep
th, m
Reservoir pressure, psi3500
XX00
XX50
X100
X1503550 3600 3650 3700 3750
Oil-water transition pressuregradient 1.294 psi/m, equivalent to0.910-g/cm3 fluid density
Formation water pressure gradient1.436 psi/m, equivalent to 1.01-g/cm3
fluid density, or 22 ppk salinity
Test 2: 1930 BWPD with1/2-in. choke; 23.4 ppk
Test 1: 1500 BWPD with1/2-in. choke; 23.4 ppk
True
ver
tical
dep
th, m
Reservoir pressure, psi
X170
X120
XX70
XX20
3200 3250 3300 3350 3400 3450 3500
Test 10: Most prolific producer,but water only, salinity 23.4 ppk.
Test 9: Before acidization,1/4-in. choke: 140 BOPD,24 BWPD, channelingsuspected as water ratechanges with choke size.After acidization producedwater only.
Test 7: Before acidization, 1/2-in. choke, flownot measurable. After acidization, 1/2-in. choke,1715 CMGPD, 514 BOPD, 37.5° API, 91 BWPD, 29.3 ppk.
Test 11: Produced water only,salinity 24.5 ppk.
Test 8: Before acidization, 1/2-in. choke1745 CMGPD, 858 BOPD, 38° API, WC<1%. After acidization, 1/2-in. choke,4084 CMGPD, 2593 BOPD, 38° API, WC <1%
Packstone
Boundstone
Wackestone
0.50
0.45
0.40
0.35
0.30
FCOR V ILL0.25
0.20
0.15
0.10
0.05
0.00
0.1
0.00.00 0.10 0.20 0.30 0.40 0.50
NCOR0.60 0.70 0.80 0.90 1.00
WO
WW
OO
OW
NLM2-2NLM2-4
NLM4-2
NLM5-9North
3
2
1
Winter 2000/2001 35
cement thickness larger than the radius of inves-tigation of the tool, the carbonate rock has noeffect on the measurement because the toolsamples only the cement.
In the past, if an openhole caliper was avail-able, it was incorporated in the RST data set toevaluate cement thickness from the differencebetween openhole radius and casing outer radius.Use of caliper data assumes that the borehole hasnot been enlarged from the time openhole logswere run to the time casing was cemented, thatthe casing was perfectly centered in the boreholeand that the borehole was perfectly round ratherthan oval-shaped. The latter assumptions arehighly unlikely, especially in deviated wells. WithRSTPro technology, and the acquisition of anadditional RST pass in sigma mode, it is possible
to compute an optimized cement-sheath thick-ness that will result, after diffusion correction, ina minimal discrepancy between formation cap-ture cross-section (sigma) measurements derivedfrom the far and near detectors.
This cement thickness and the outer casingdiameter can be used to generate an “RSTcaliper” for input to the RST oil-volume evalua-tion module. Previous RST logs of carbonaterocks offshore India exhibited remaining oil pro-files that were difficult to justify. The new tech-nique has produced credible logs since itsintroduction early in 2000. Changes in saturationprofile of as much as 20 saturation units com-monly occur between evaluations, with or withouttaking optimized cement thickness into account.
> Depletion-profile analysis. In these fourwells, a combination of the RST depletionprofile and a borehole temperature deriva-tive acquired during production shows thethree major flow units in the reservoir.The base of Zone 1 is the original oil-watertransition. Zone 2 includes the major pro-ducing horizons. Reserves remain in Zone 3,but reservoir simulation suggests that majorpressure depletion has occurred.
> Overlay of the RST saturation-evaluation quadrangle with near-detector C/O ratio (NCOR) to far-detector (FCOR) data from India.Color coding in the Z-axis represents shale volume (VILL); red isclean limestone, blue is 10 percent shale. The data were recordedin a 22-porosity unit limestone reservoir, in an 8.5-in. borehole with7.0-in. casing. The borehole geometry, formation lithology andporosity, along with the carbon density of the hydrocarbon, fullydefine the end points of the characterization locus. The data pointscluster along the WW-WO line, indicating a water-filled borehole.The formation oil saturation varies from 0 to 40%.
24. For more on production logging: Akhnoukh R, Leighton J,Bigno Y, Bouroumeau-Fuseau P, Quin E, Catala G,Silipigno L, Hemingway J, Horkowitz J, Hervé X,Whittaker C, Kusaka K, Markel D and Martin A: “KeepingProducing Wells Healthy,” Oilfield Review 11, no. 1(Spring 1999): 30-47.
25. Olesen JR and Carnegie A: “An Improved Technique forReservoir Evaluation Through Casing,” paper IRS2k-O228,presented at the Improved Recovery Symposium,Institute of Reservoir Studies, Ahmedabad, Gujarat,India, July 27-28, 2000.
26. More than 3000 combinations of hole sizes, lithologies,porosities, borehole and formation saturations havebeen measured. Interpolation between endpoints isobtained by nuclear modeling of the tool, formation andborehole conditions, with model endpoints calibratedwith laboratory data.
Geostatistical analysis—Innovative use ofstatistical tools at the JRC extends NMR analysisand RST results from key wells to the entire fieldmodel, which previously comprised only conven-tional log data and downhole production data.These new techniques involve proportion curvesand water-conduit tracking.
A vertical proportion curve plots a histogram atevery stratigraphic level within the formation(right). In this example, logs of porosity categoriesare displayed with depths relative to the top of thereservoir—the conforming surface. The verticalsampling interval is 1 meter [3 feet] in these wells.The RST survey discussed earlier was run in anearby well. The logs have been projected onto anorth-south line shown on the map. The verticalproportion curve is generated by showing the rel-ative percentages of the different porosity cate-gories at every stratigraphic level. The curvebecomes narrower with increasing depth becausethe wells have different depths of penetration.Nevertheless, it can be inferred that the formationprobably consists of two relatively high-porosityzones separated by a low-porosity zone.
This example demonstrates that the proportion-curve technique uses grouped or categorical datarather than continuous data. Traditionally, verti-cal proportion curves have used depth-dependentlithofacies data to understand depositionalcycles and to constrain geostatistical realiza-tions.27 But, as shown at the JRC, these curvescan have strong application in diagnosing reser-voir flow behavior and relating this behavior tothe reservoir characterization developed fromopenhole and cased-hole saturation logs.
To understand this technique more clearly, theexample can be taken a step further by including avertical proportion curve derived from productionlogs (right, lower right of figure). Productivity, orflow rate, is categorized as high, medium, low orno flow. At the top of the reservoir, flow rates arehigh, which implies that a thin layer of high per-meability exists at the top of the formation.Subsequent analysis showed that this is a super-klayer. Farther down, there are two other majorzones of high flow rate mixed with lower flowrates. Those zones should have medium to highpermeability. These observations support infer-ences made on the basis of RST data describedearlier, since the proportion curves capture theaverage behavior of the region.
Comparing the production log and porosity pro-portion curves shows that porosity alone is anincomplete descriptor of permeability in the regionand, consequently, that better openhole perme-ability-based descriptors are needed—such asthose derived from MDT or NMR data.
The proportion-curve technique has beenapplied elsewhere in the reservoir, and to otherforms of dynamic and static data, to derive severaluseful results rapidly and efficiently. For example,it is possible to map throughout the field the lat-eral and vertical extent of high-permeability zones
that have flowed water for an extended period bycombining an openhole gamma ray with a cased-hole gamma ray run later in the life of a well intoa proportion curve. Comparing openhole gammaray logs to density-neutron curve separationenables detection of weathered zones wheremeteoric water has created super-k layers throughdiagenetic alteration.
Proportion curves allow quick, efficient analy-sis of vast amounts of data, an important consid-eration when interpreting and synthesizing datafrom an entire field, which may include openhole,
36 Oilfield Review
> Creation and application of proportion curves. Porosity data from ten wells were sampled at 1-m [3-ft]vertical intervals and categorized to form porosity proportion logs (upper left). Well locations andthe north-south projection line are shown in the map (upper right). The porosity proportion logs arecombined to form a porosity proportion curve (lower left). The bottom of the curve narrows becausethe wells have different penetration depths. A similar curve can be generated using production logsfrom the wells (lower right). Zone 1 includes only high productivity values; Zones 2 and 3 have somehigh flow rates mixed with lower flow rates.
27. For more about proportion curves and depositionalcycles: Jain AK and Carnegie A: “Value AdditionThrough Stochastic Evaluation of Gamma Rays—A Geostatistical Approach to Geological Modeling andCharacterization of the Reservoir,” presented at theAAPG International Conference and Exhibition, Bali,Indonesia, October 15-18, 2000.For more about proportion curves and geologicalrealizations: Klauser-Baumgartner D and Carnegie A:
“Geostatistical Modeling of Delta Front Parasequencesby Indicator Kriging,” Information Processing andModeling in Geology, 10. Kontaktwochenende,Sedimentologie, Aachen, 1995.
28. For more on the water-conduit tracking method:Carnegie A: “Techniques to Optimize the Efficiency ofHistory Matching in Integrated Studies,” paper 402,presented at the Improved Oil Recovery Symposium,Institute of Reservoir Studies, Ahmedabad, Gujarat,India, July 27-28, 2000.
29. For more on the RRT method: Russell SD, Akbar M,Vissapragada B and Walkden G: “Small-ScaleHeterogeneity and Permeability Estimation fromDipmeter and Image Logs for Reservoir Rock Typing:Aptian Shuaiba Reservoir of Abu Dhabi,” Bulletin of the American Association of Petroleum Geologists(in press).
D1
D2
D3
Porosity < 8%
Porosity 8-16%
Porosity 16-24%
Porosity > 24%
D4
D9
D8
D7
D6
D5
C2
Porosity proportion logs Wells supplying the logs
N-S
pro
ject
ion
line
No flow
Low productivity
Medium productivity
High productivity
Zone 1
Zone 2
Zone 3
Winter 2000/2001 37
cased-hole and historical production data fromseveral hundred wells. Proportion curves can begrouped to gain local insight about specific partsof a field. They also offer a high level of immunityto incorrect or low-quality data, since the “noise”created by such data sets tends to cancel itself,and since the amount of high-quality data farexceeds the amount of questionable data withinthe entire data set. A proprietary PC-based soft-ware package, refined at the JRC, performs inter-active database management, computation, and2D and 3D visualizations of these proportioncurves. The package is compatible with theGeoFrame Application Builder program, whichfacilitates database access.
Water-conduit tracking method—Detection ofhigh-permeability conduits, such as faults orsuper-k layers that conduct reservoir or injectionwater, can be improved by performing a networkanalysis of water breakthrough times on historicalwell-production data. A PC-based software pack-age written at the JRC helps users detect thewater path interactively. The water-breakthroughinformation comes from historical well-produc-tion data loaded into a production-management
database. This tool allows quicker, more objectivediagnosis of the progress and evolution of waterbreakthrough over an entire field than traditionalmanual analysis.28 This method, known as water-conduit tracking (WCT), makes assessing thevalidity of multiple scenarios more efficient thanin the past.
By conducting the vital analyses of rock tex-tures, permeability, depletion profiles and pro-duction data, and judiciously integrating all theresults with other field data using geostatisticaltechniques, the JRC is creating models thatresult in more realistic reservoir simulations.These simulations provide more reliable guid-ance for development and production decisionsthan analyses performed in isolation.
Evaluating Carbonate Heterogeneity in Abu DhabiLocal operations teams augment the contributionsof formal research efforts to understand carbonaterocks. Scientists and engineers in Abu Dhabi, UAE,have developed new techniques to evaluate het-erogeneous carbonate reservoirs by integratinggeological, openhole and production-log data.
Characterization of the small-scale hetero-geneities within reservoir rocks has led to a clas-sification of 17 reservoir rock types (RRT) in theShuaiba formation. Reservoir rock types are basedon lithofacies, wireline log data, core porosity andpermeability, capillary pressure and pore-size dis-tributions derived from mercury-injection analy-ses, and production data.29 RRTs can be used tobetter correlate zones within reservoirs in theabsence of cores.
An oil field in Abu Dhabi has been producingfrom the Lower Cretaceous Shuaiba formationsince 1962. Within the field, the Shuaiba formationvaries from shallow-water shelf to deepwaterslope sediments, with four distinct reservoir facies.RRTs range from nonproductive rocks to those withup to 30% porosity and 20-Darcy permeability(below). These significant heterogeneities must beconsidered when planning well trajectories, wellcompletions and production strategies.
RRTs are defined on the basis of reservoirquality, distribution and productivity, but are prod-ucts of their depositional environment and dia-genetic history. RRTs observed in cores and logsfrom two wells in the field have been correlated
RRT 6 RRT 7
RRT 14 RRT 15
RRT 8
> Shuaiba heterogeneity. RRTs range from rudists—extinct mollusks similar to oysters—in lime mud (top left) tomixed rudists in a grainy matrix (top center) to diagenetically altered, debris-filled rudstone (top right). A pencilor fingertip in each photograph indicates the scale. RRTs from the northern part of the field comprise rudstone(bottom left photomicrograph) and fine-grained packstone or packstone (bottom right). The field of view of thephotomicrographs is 4 mm by 6 mm.
38 Oilfield Review
> Shuaiba RRTs and permeability indicator derived from core and log data. Photographs (far right) in thiscomposite plot from a well in a field in Abu Dhabi show the heterogeneity of three of the distinct RRTs. Permeability characterized by analysis of Stratigraphic High-Resolution Dipmeter Tool (SHDT) data (Track 8)shows close agreement with log and core data.
SHDT perm
indicatorCore RRTs
SHDT/log facies
Raw SHDT channels
GRDepth,
ft
XX950
X1000
X1050
X1100
X1150
X1200
X1250
X1300
Petrophysical interpretation
SHDT conductivity
Proportion of
heterogeneityCore perm
0.2 2000
Winter 2000/2001 39
with logs in uncored wells; this correlation allowsmore accurate permeability estimation in thosewells than with use of log data alone. The RRTstudy contributes significantly to the field devel-opment because the operator, Abu DhabiCompany for Onshore Oil Operations (ADCO), canuse realistic permeability estimates and upgraded3D geological models to optimize field drainage,thereby maintaining and prolonging production.
One innovative RRT characterization methodrelies on careful integration of conventional welllogs, such as gamma ray, neutron and density,
with high-resolution dipmeter and image logs.Heterogeneities in the form of conductivity varia-tions are quantified using specialized software,including BorTex and RockCell applications, toidentify RRTs and generate permeability indica-tors (previous page).30 In extremely heteroge-neous carbonates, permeability derived usingthis methodology resolves heterogeneity betterthan 1-in. core plugs or minipermeameter data(above). The higher resolution and increasedborehole coverage of imaging devices providemore accurate differentiation of RRTs than
dipmeter logs alone and facilitate identificationof flow paths between vugs and large pores.Because dipmeter and image logs are morewidely available than core, RRT analysis is apowerful tool for evaluating wells that lack core samples.
Another successful technique to evaluateporosity in the Shuaiba formation uses boreholeimages to map primary and secondary porosity.
> Integrated permeability data. Core plugs from 246 one-foot intervals (left) and 586 minipermeameter measurements at 2- to 3-in. intervals (center) from a wellin Abu Dhabi show significant scatter because of the extreme small-scale heterogeneity. On the other hand, the SHDT-derived permeability indicator (right)shows a clear trend that closely correlates with RRTs found in cores. Each color in the cored interval represents a distinct RRT of the Shuaiba formation.
30. A full discussion of BorTex and RockCell software isbeyond the scope of this article. For more information:http://www.geoquest.com/pub/prod/index.html.
XX900
XX950
X1000
X1050
X1100
X1150
X1200
X1250
X1300
XX900
XX950
X1000
X1050
X1100
X1150
X1200
X1250
X1300
XX900
XX950
X1000
X1050
X1100
X1150
X1200
X1250
X13000.1 1
Dept
h, ft
Core permeability, mD Minipermeameter core permeability, mD SHDT permeability indicator, mD10 100 1000 10,000 0.1 1 10 100 1000 10,000 1 10 100 1000 10,000
The wellbore azimuthal porosity spectrum revealsextreme porosity heterogeneity, which, in turn, isrelated to permeability (above).31
While RRT studies aid long-term reservoircharacterization and simulation studies, theresults also can impact field development in theshort term. For example, recognition of differentRRTs in a horizontal Shuaiba well allowed theoperator to optimize production rates.32 A hori-zontal oil well drilled in 1997 initially produced
6000 BOPD [953 m3/d] water-free for fourmonths, then abruptly lost pressure and was shutin. The operator needed to determine whetherthe decrease in reservoir pressure, migration offines or water loading stopped oil flow.Interpretation of high-quality PL Flagship produc-tion logging data, coupled with RRT analysis thatincorporated geological and openhole data, con-firmed that the horizontal section penetrated twovastly different RRTs (next page, left).
Engineers and geoscientists discovered that alow-permeability RRT along the central segmentof the wellbore affected flow behavior. By under-standing the sensitivity to choke settings, theoperator optimized well performance by select-ing a lower setting that gave an even drawdownalong the length of the well despite the lateralreservoir heterogeneity. This has resulted in thestable production of thousands of barrels per dayof dry oil. This reservoir-management strategy isbeing applied elsewhere in the reservoir to opti-mize placement of additional wells.
Future Research EffortsClearly, significant work remains for those explor-ing and exploiting carbonate reservoirs. Althoughthe complexity and heterogeneity of carbonaterocks present enormous interpretation and oper-ational challenges, the examples presented inthis article underscore the need for integration ofall available data and prudent selection of evalu-ation tools.
Schlumberger is addressing carbonate issuesmore aggressively by establishing a CarbonateResearch Center (CRC) at King Fahd University ofPetroleum and Minerals (KFUPM) in Dhahran,Saudi Arabia (next page, right). The proximity ofthe center to the prolific carbonate reservoirs ofthe Middle East, key operators and selectregional universities will facilitate intraregionalcollaboration. Innovative information-technologysolutions for virtual teamwork will accelerateresearch progress and dissemination of suc-cesses worldwide (see “From Reservoir Specificsto Stimulation Solutions,” page 42 ).
Key areas of activity for the CRC include landseismic-data acquisition, NMR interpretation,water management and well stimulation in car-bonate reservoirs. Research efforts will comple-ment rather than duplicate work performed inother research facilities. For example, Middle Eastcarbonate case studies will be performed in theCRC rather than SDR as appropriate. Because ofthe proximity to classic carbonate field locations,
40 Oilfield Review
> Shuaiba porosity evaluation. A new technique uses FMI images (Track 1 of upper log) to map primaryand secondary porosity by generating histograms of porosity at each depth (Tracks 2 and 3). In thelower figure, the lime mudstone at 3 feet is relatively homogeneous and microporous, as shown inthe photograph and photomicrograph to the right of the lower log. The floatstone, a type of matrix-supported limestone with some large grains, which is located above 5 feet, has extreme porosity heterogeneity, as shown in photograph and photomicrograph to the right.
31. Akbar M, Chakravorty S, Russell SD, Al Deeb MA,Saleh Efnik MR, Thower R, Karakhanian H, Mohamed SSand Bushara MN: “Unconventional Approach to ResolvingPrimary and Secondary Porosity in Gulf Carbonates fromConventional Logs and Borehole Images,” paper 0929,presented at the 9th Abu Dhabi International PetroleumExhibition and Conference, Abu Dhabi, UAE, October15-18, 2000.
32. Russell SD, Al-Masry Y, Biobien C and Lenn C:“Optimizing Hydrocarbon Drainage in a Heterogeneous,High-Permeability Carbonate Reservoir,” paper SPE59427, presented at the SPE Asia Pacific Conference onIntegrated Modelling for Asset Management, Yokohama,Japan, April 25-26, 2000.
3 ft
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.90.0
0.10 0100Porosity, % Porosity, % 100
0.90.00.10.20.30.40.50.60.7
Depth,ft
FMI dynamic image FMI porosity histogramp.u. 050
FMI/log porosity
Secondary porosity
p.u. 040
Log porosity
FMIporosity
2
3
4
5
6
7
8
Winter 2000/2001 41
an endeavor specific to the Dhahran facility will befield testing new tools and modification of existingtools to achieve the best possible results in car-bonate rocks.
Much of the emphasis so far has been on oilreservoirs, but long-term strategic emphasis ongas production is pushing even the largest oilproducers in the Middle East to pursue develop-ment of carbonate gas reservoirs. Deeper gasplays present significant interpretation, explo-ration and production challenges.
Although this article has focused on data atthe scale of a wellbore, there are larger scales atwhich Schlumberger and operators are evaluat-ing carbonate reservoirs. Interpretations fromlogs and cores are being incorporated in field-scale reservoir simulations. The simulationsallow reservoir models to be extended into thefourth dimension, time, to better predict fieldresponse and optimize performance.
New seismic acquisition methods, such as the Q single-sensor technology that has been used for data acquisition in the MiddleEast, will address imaging challenges at an even larger scale. Preliminary results from these tests suggest that significant improve-ment in data quality will advance our under-standing of carbonate reservoirs and, whenproperly integrated with other data, lead togreater success in carbonate exploration, devel-opment and production. —GMG
> Evaluating RRTs to optimize production. RRT analysis confirmed thatthe horizontal section of the well penetrated two RRTs (top). Wirelinelogs (bottom) confirm the variations in gamma ray, porosity and resistivityof the two RRTs. The invasion inferred from separation between deepand shallow resistivity curves in RRT 14 indicates higher permeabilitythan in RRT 15.
> The new carbonate research environment.The Schlumberger Carbonate Research Centeris located on the campus of King Fahd Universityfor Petroleum and Minerals (KFUPM) inDhahran, Saudi Arabia.
GR14
15
14
Deep
Shallow
PHIE
0.2 2000ohm-mShallow resistivity
Departure, ft
Dept
h, ft
sub
sea
8000 6000 4000 2000 0
0.2 2000ohm-mDeep resistivity
0 50APIGR
0 0.4ft3/ft3
Effective porosity MD:5000ft
XX000
XX500
XX000
X1500
45-fttotal
RRT 14
Well C Well A
16-fttotal
X700
X720
X740
X760
X780
X800
X820
X840
X860RRT 15