4 Oilfield Review

Focusing on Downhole Fluid Sampling and Analysis

Ridvan AkkurtSaudi AramcoDhahran, Saudi Arabia

Martin BowcockBG GroupReading, England

John Davies ChevronHouston, Texas

Chris Del CampoSugar Land, Texas, USA

Bunker HillRosharon, Texas

Sameer JoshiDibyatanu KunduMumbai, India

Sanjay KumarCairn Energy India Pty LtdGurgaon, India

Michael O’KeefeHobart, Tasmania, Australia

Magdy SamirAberdeen, Scotland

Jeffrey TarvinCambridge, Massachusetts, USA

Peter WeinheberHouston, Texas

Stephen WilliamsHydroBergen, Norway

Murat ZeybekAl-Khobar, Saudi Arabia

For help in preparation of this article, thanks to StephaneBriquet, David Nunez and Ricardo Vasques, Sugar Land,Texas; Kåre Otto Eriksen, Statoil ASA, Stavanger; Noriuki Matsumoto, Sagamihara, Kanagawa, Japan; MoinMuhammad, Edmonton, Alberta, Canada; Oliver Mullins,Houston; Tribor Rakela, Caracas; John Sherwood, Cambridge,England; and Dag Stensland, ENI Norge, Stavanger. Thanks also to ConocoPhillips (UK) Ltd.CFA (Composition Fluid Analyzer), InterACT, LFA (Live FluidAnalyzer), MDT (Modular Formation Dynamics Tester), MRX (Magnetic Resonance eXpert), Quicksilver Probe, Platform Express, PVT Express and SlimXtreme are marks of Schlumberger.

A new focused-sampling device allows acquisition of downhole fluid samples of

unprecedented purity, and in a fraction of the time needed with conventional sampling

technology. The method also gives superior results for downhole measurements of

formation-fluid properties.

Understanding the properties of fluids containedin a hydrocarbon reservoir requires measure -ments on fluid samples. Sample analysis helpsidentify fluid type, estimate reserves, assesshydrocarbon value and determine fluidproperties, so production can be optimized.Using fluid-analysis results, oil companies decidehow to complete a well, develop a field, designsurface facilities, tie back satellite fields andcommingle production between wells.

Fluid analysis is also important for under -standing the properties of formation water,which can have significant economic impact.Often, the most crucial goals are to identify thecorrosive properties of the water for the purposeof selecting completion materials and to measurescaling potential for avoiding flow-assuranceproblems. In addition, log analysts want toquantify the salinity of the water for petro -physical evaluation, and geologists and reservoirengineers want to establish the water source forevaluation of reservoir connectivity.

Formation-fluid samples can be acquiredusing one of three main techniques. First,wireline formation testers deployed in open holecan acquire fluid samples and also perform down -hole analysis of fluids, ensuring optimal sampleacquisition and the possibility of analyzing fluidsearly in the life of the well. These testers providea cost-effective method of acquiring early fluidsamples, with performance now often equal to orabove that achievable with the second method,drillstem tests (DSTs). In the past, DSTs,typically designed to test production and investi -gate reservoir extent, have produced sampleswith less contamination than openhole sampling.DSTs require early planning and a well comple -

tion that can withstand production pressures,and can cost much more than openholesampling, especially in offshore wells. In a thirdmethod, samples can be acquired by wirelinetools deployed in a cased, producing well.

An important aspect of fluid sampling isanalysis of the fluids at reservoir conditions. Thishelps validate sample quality during the samplingprocess, but also enables the mapping of verticalvariations in fluid properties, allowing inter -preters to determine zonal connectivity anddefine reservoir architecture early in field life.Uncontaminated fluid samples allow accuratemeasurement of fluid properties both downholeand at the surface.

After samples are acquired, they typically areanalyzed in laboratories, where they undergo aseries of tests depending on what the clientneeds to understand. Standard analyses forhydrocarbon samples include chemical compo si -tion to C30+, gas/oil ratio (GOR), density, viscos -ity, and phase properties such as saturationpressure, bubblepoint, pour point and stability ofasphaltenes.1 Several measurements can now beperformed downhole, using optical spectroscopyto characterize formation fluids under reservoirconditions.2 These include density, opticaldensity, GOR and chemical composition to C6+.

Laboratory and downhole fluid measurementsboth require pure, uncontaminated samples.Contamination occurs when miscible drilling-fluid filtrate that has invaded the formationmixes with the formation fluid being sampled.For instance, hydrocarbon samples are contami -nated by oil-base mud (OBM) filtrate, and watersamples are contaminated by water-base mud(WBM) filtrate.

1. The phrase “composition to C30+” indicates thatcompounds of up to 29 carbon atoms are separatelydiscriminated, with the remainder combined into afraction indicated as C30+.Pour point is the minimum temperature at which oilpours or flows.

2. Fujisawa G, Betancourt S, Mullins OC, Torgersen T,O’Keefe M, Terabayashi T, Dong C and Eriksen KO:“Large Hydrocarbon Compositional Gradient Revealed by In-Situ Optical Spectroscopy,” paper SPE 89704,presented at the SPE Annual Technical Conference and Exhibition, Houston, September 26–29, 2004.

3. Mullins OC, Schroer J and Beck GF: “Real-TimeQuantification of Filtrate Contamination During OpenholeWireline Sampling by Optical Spectroscopy,”Transactions of the SPWLA 41st Annual LoggingSymposium, Dallas, June 4–7, 2000, paper SS.

Mullins OC and Schroer J: “Real-Time Determination ofFiltrate Contamination During Openhole WirelineSampling by Optical Spectroscopy,” paper SPE 63071,presented at the SPE Annual Technical Conference andExhibition, Dallas, October 1–4, 2000.Dong C, Mullins OC, Hegeman PS, Teague R, Kurkjian Aand Elshahawi H: “In-Situ Contamination Monitoring andGOR Measurement of Formation Fluid Samples,” paperSPE 77899, presented at the SPE Asia Pacific Oil and Gas Conference and Exhibition, Melbourne, Australia,October 8–10, 2002.

4. Gozalpour F, Danesh A, Tehrani D-H, Todd AC andTohidi B: “Predicting Reservoir Fluid Phase andVolumetric Behaviour from Samples Contaminated withOil-Based Mud,” paper SPE 56747, presented at the SPE Annual Technical Conference and Exhibition,Houston, October 3–6, 1999.

Winter 2006/2007 5

To reduce contamination during samplecollection, engineers rely mostly on increasingthe volume of fluid pumped from the reservoir bypumping longer or at a higher rate. Downholeanalysis of contamination level can determinewhen fluid flowing through the sampling-toolflowline is clean enough to be collected.3

However, long pumping time increases rig timeand associated costs, and may increase the riskof downhole tool sticking. Depending on thereservoir permeability, high pumping rates cancause the reservoir fluid to drop below saturationpressure. If this happens, the downhole sampleswill not be representative of the reservoir fluid.In the case of unconsolidated formations, highpumping rate may induce sand production. Also,in settings involving high vertical permeability,even long pumping times and increased pumpingrates do not guarantee clean samples.

Fluid-analysis experts have worked to under -stand and mitigate the effects of contami nationon samples. Some methods attempt to derive thecomposition or GOR of a pure sample knowingthe composition of the OBM contami nating thecollected sample.4 However, uncer tain ties anderrors accompany fluid properties estimated in

this manner. Researchers have quantified theerrors caused by contamination on somemeasure ments. For example, the pres sure atwhich asphaltenes precipitate from solution incrude oil decreases in the presence of OBMcontamination. In one case, just 1% OBMcontamination by weight caused asphaltene-

precipitation onset pressure to decrease by100 to 150 psi [0.7 to 1.0 MPa] (left).5 Thus,measurements on contaminated samples under-estimate asphaltene-precipitation onset pressure,and may negatively affect flow-assurance andproduction predictions. These results emphasizethe need for extremely low-contamination samples.

A new sampling apparatus designed to reducefiltrate contamination focuses fluid intake sothat reservoir fluid flows into one sampling linewhile filtrate flows into a separate line. With thisinnovative tool, mud-filtrate contamination canbe separated efficiently from formation fluid inthe early stage of the sampling process. A cleanreservoir-fluid sample can be acquired muchfaster than with conventional sampling tech-niques. This article describes the advantages ofthe new, focused-sampling tool through fieldexamples of hydrocarbon and water samplingfrom the Gulf of Mexico, the North Sea, India andthe Middle East.

Quicker and CleanerTo fully appreciate the advantages of the newsampling method requires a brief overview ofconventional downhole fluid-sampling technology.In the typical scenario, overbalanced drilling intoa permeable formation will facilitate invasion ofdrilling-fluid filtrate into the formation and thecreation of filtercake on the borehole wall.During conventional formation-fluid sampling, awireline formation tester deploys a packeragainst the borehole wall to isolate the sampleprobe from borehole fluids and hydrostaticpressure. The probe is then pressed through themudcake and against the formation (left). Asuccessful seal connects the sampling tool withthe formation while isolating the tool flowlinefrom borehole fluid and pressure.

As the sampling tool withdraws fluid from theformation through the probe, the first reservoirfluid to enter the flowline is contaminated withfiltrate from the drilling fluid. The level ofcontamination, monitored in real time bydownhole spectroscopic analyzers, decreases aspumping continues. Depending on formationpermeability, anisotropy, amount of invasion,formation-fluid viscosity, and pumping time, rateand pressure drawdown, the contamination levelmay or may not decrease sufficiently to allowcollection of a fluid sample that is representativeof the formation fluid. Filtrate contaminationfrom deeply invaded zones may continue to feedinto the sampling probe, and in cases of poorlyformed mudcake, borehole fluid may continue toinvade the formation at a relatively significantrate. Achieving sufficiently low levels ofcontamination may require pumping for

6 Oilfield Review

> The effect of oil-base mud (OBM) filtrate contamination on asphaltene-precipitation onset pressure.Laboratory analysis on live oils with varying amounts of added OBM filtrate shows a decrease inasphaltene-precipitation onset pressures with increased OBM contamination. Live oils are oils thatcontain dissolved gas. Asphaltene precipitation is detected by light transmittance; precipitates scatterlight and decrease transmittance. These and similar experiments show on average that for 1% byweight OBM contamination, asphaltene onset pressure decreases by 100 to 150 psi. The 19.4%contamination sample reached saturation pressure before it reached the asphaltene-precipitationpressure. (Adapted from Muhammad et al, reference 5.)

Saturationpressure

Asphaltene-precipitationonset pressures

Contamination(live-oil basis)

2.6 % by weight7.6 % by weight14.2 % by weight19.4 % by weight

Pressure

Pow

er o

f tra

nsm

itted

ligh

t

5. Muhammad M, Joshi N, Creek J and McFadden J:“Effect of Oil Based Mud Contamination on Live FluidAsphaltene Precipitation Pressure,” presented at the5th International Conference on Petroleum PhaseBehaviour and Fouling, Banff, Alberta, Canada,June 13–17, 2004.

6. Sherwood JD: “Optimal Probes for Withdrawal ofUncontaminated Fluid Samples,” Physics of Fluids 17,no. 8 (August 2005): 083102.

7. Tarvin JA, Gustavson G, Balkunas S and Sherwood J:“Sampling Fluid from a Two-Dimensional Porous MediumWith a Guarded Probe,” submitted to Journal ofPetroleum Science and Engineering.

> Conventional formation-fluid sampling with a wireline formation tester.The tester forces a packer to seal against the borehole wall, then pressesa probe through the mudcake and against the formation (right). Formationfluid is blue-gray and filtrate is light brown. The probe (left) has a singleintake port. When pumping begins, fluid is highly contaminated (graphinset), but decreases gradually with time. However, even with longpumping times, the contamination level may not reach an acceptable limitin some formations.

Cont

amin

atio

n le

vel

Time

Acceptable sample

Winter 2006/2007 7

extended periods of time—many hours—whichcan be expensive in terms of rig time andincreased exposure to sticking in open hole.

Seeking ways to improve sample quality andreduce sampling time, researchers investigatedthe effects of different probe configurations.To test the idea that focused flow into a probecould reduce sample contamination and shortensampling time, a scientist at SchlumbergerCambridge Research in England simulated flowinto modified probes.6 The modeling resultshelped determine optimal probe size.Researchers at Schlumberger-Doll Research inConnecticut, USA, conducted 2D experiments onlaboratory models to determine the potentialbenefits in sample cleanup (above).7 Themodified probes had three openings: side open-ings, called guard probes, drew contaminatedfluid away from the central area of the probe, anda central opening, called a sample probe,collected low-contamination fluid. Experimentalresults indicated that cleanup with the guardprobes active proceeded much more quickly thanwithout, achieving lower contamination levelswith less fluid volume pumped (right).

A

B

C

D

Without Guard

A

B

C

D

With Guard

> Setup and visual results of laboratory experiments simulating focusedflow. The experimental setup (top right) consisted of a 2D formation madeof glass beads, surrounded by a single oil with an optical index identicalto that of the glass beads, all held between two vertical glass plates. Abottom portion of oil was dyed red to represent the filtrate-invaded zone.Above this, the oil was left transparent. A sample and guard-probeassembly at the bottom of the formation extracted fluid (inset). A cameramonitored the cleanup in the formation directly in front of the probeassembly. After image processing, the time-lapse visual images (left) showlarge differences in the area cleaned up by the sample probe alone (left)and the sample and guard probes together (right). The sample and guardprobes clean up a large area in front of the sample probe, ensuring thatonly uncontaminated fluid enters the sample probe.

> Contamination reduction with and without guard probes. Laboratorymeasurements detected decreasing contamination levels with increasingvolume of fluid pumped, corresponding to increasing pump time. Samplingwithout the guard probe (blue) never achieved contamination levels lessthan 1%.

Dye

conc

entra

tion,

% o

f max

imum

100

10

1

0 20 40 60 80 100 120 140 160 180 200

Volume pumped, ml

Without guardWith guard

8 Oilfield Review

> Engineering experimental setup to investigate the feasibility of focused sampling. Aformation interface tester, containing a 15.5-in. diameter, 12-in. tall sandstone core, is incontact with two fluid manifolds and a laboratory-prototype focused-sampling probe.Simulated formation fluid is supplied at the base of the core, and simulated filtrate issupplied in a ring around the core. Flow into and from the tester is controlled andmonitored by pumps and valves, and contamination level is calculated from electricalconductivity measurements on the flowlines. (Adapted from Dong et al, reference 8.)

Formation and filtrate fluid(supply side)

Sampling and conductivity(measurement side)

Accumulators

Meteringvalves

Meteringvalves

Flowmeters

Flowmeters

Conductivitymeters

> Focused-sampling experimental data. The decrease in contamination inferred fromelectrical conductivity measurements on the guard and sample flowlines demonstratesfluid cleanup in this sampling test. The contamination level in the sample flowline (green)decreased rapidly, while the contamination level in the guard flowline (red) decreasedgradually. Summing the flow from both flowlines produces the total flow (blue), the flowthat a traditional probe would have measured. (Adapted from Dong et al, reference 8.)

Cont

amin

atio

n, %

10

20

0

30

40

50

60

70

80

90

100

–10 0 10 20 30 40 50 60

Elapsed time, min

Guard flowlineTotal flowSample flowline

Further engineering tests at Sugar LandTechnology Center in Texas extended the 2Dresults in simulated formations to threedimensions and actual rock formations. In theseexperiments, a downhole probe prototype of anew focused-sampling tool drew fluids from alarge sandstone core in a test apparatus (left).8

The 15.5-in. diameter, 12-in. tall core containedaqueous sodium chloride [NaCl] formation fluidand mud-filtrate fluid of different knownconductivities. Fluid flow to the guard andsample flowlines was controlled and measuredwith metering valves and high-pressureflowmeters. Calibrated electrical conductivitymeters on the flowlines leading from the guardprobe and the sample probe recorded thecleanup history of each sampling test. Withfocused sampling, the contamination levels ofthe fluid in the sample flowline decreasedrapidly, while the contamination level in theguard flowline decreased gradually (below left).A traditional probe would have measured thecombined flow, and would not have achievedcontamination less than 10%.

The key to acquiring such low-contaminationsamples is the focusing effect achieved by themulti-intake probe.9 This innovative design hasbeen implemented in the Quicksilver Probewireline sampling tool, a new module of the MDTModular Formation Dynamics Tester tool. Insome ways, the configuration of the QuicksilverProbe module is similar to that of traditionalsamplers, in that a packer seal isolates the fluid-sampling zone from the borehole. However,within the fluid-sampling zone, a cylindricalguard probe on the periphery of the samplingzone surrounds the innermost sampling area(next page, left). An additional packer sealseparates the guard intake from the sampleintake. The inner and peripheral areas areconnected to separate flowlines, called thesample and guard flowlines, respectively. Twopumps in the tool, one above the probe and onebelow, can draw fluid into the two flowlines atdifferent rates, and spectroscopic analyzersdetermine the composition of fluid in eachflowline (next page, right). The focusing effect ofthe method is somewhat analogous to the waylaterolog devices use guard electrodes to focuscurrent into a formation to measure resistivity.10

The Quicksilver Probe focused-sampling toolpumps fluid from the formation through thecentral and peripheral areas of the samplingzone simultaneously. Initially, commingled con-taminated fluid flows into both areas, but this

Winter 2006/2007 9

fluid is not collected. Fluid flow is thenseparated, or split, between the guard andsample flowlines. Fluid flow into the guard intakecan be increased, and in a short time, allcontaminated fluid is drawn into the guardflowline, allowing low-contamination formationfluid to flow into the sample flowline. Thistechnique accentuates the difference in contam -ination level between clean and contaminatedfluid, making it easier to identify a time at whicha clean sample can be collected. Case studiesfrom several environments show the samplequality that can be obtained using the newfocusing technology.

Exploring in the Gulf of MexicoIn 2004, Chevron drilled an exploration well intothe emerging Lower Tertiary play in thedeepwater Gulf of Mexico. These wells aretypically difficult to drill and complete, withwater depths down to 10,000 ft [3,000 m] andtotal well depths exceeding 25,000 ft [7,600 m].More than 20 exploration and appraisal wellshave been drilled so far in this play, and morethan half were discoveries, many with thick oil

columns. However, in such conditions, well testsusually are extremely expensive, typically costingUS$ 70 million or more. For this reason DSTs arerarely performed in this region.

Drilling in this play in Walker Ridge Block759, Chevron and partners announced discoveryof more than 350 ft [110 m] of net-pay oil sandsin Jack 1, the first well of the Jack prospect, inSeptember 2004.11 The subsalt prospect is

8. Dong C, Del Campo C, Vasques R, Hegeman P andMatsumoto N: “Formation Testing Innovations for FluidSampling,” presented at the Deep Offshore TechnologyConference and Exhibition, Vitoria, Espirito Santo, Brazil,November 8–10, 2005.

9. Del Campo C, Dong C, Vasques R, Hegeman P andYamate T: “Advances in Fluid Sampling with FormationTesters for Offshore Exploration,” paper OTC 18201,presented at the Offshore Technology Conference,Houston, May 1–4, 2006.

10. Doll HG: “The Laterolog: A New Resistivity LoggingMethod with Electrodes Using an Automatic FocusingSystem,” Petroleum Transactions of the AIME 192 (1951):305–316.

11. ChevronTexaco Announces Discovery in Deepwater Gulfof Mexico, http://www.chevron.com/news/press/2004/2004-09-07.asp (accessed September 22, 2006).

> Formation-fluid sampling with the Quicksilver Probe focused-sampling tool. The probe(left) has two intake ports, with the guard intake surrounding the sample intake. Packerssurround and separate these probes and seal against the borehole wall (right).Formation fluid is blue-gray and filtrate is light brown. When pumping begins, fluidflowing through the sample intake is highly contaminated (graph inset), but decreasesquickly with time. Soon, contamination levels are at an acceptable value.

Cont

amin

atio

n le

vel

Time

Acceptable sample

> The Quicksilver Probe toolstring. Fluids enterthe tool at the focused-sampling probe.Contaminated fluids flow downward through theguard fluid analyzer and pump. Clean fluids flowupward through the sample fluid-analyzer andpump modules to the sample-bottle module. Theconfiguration may change for different samplingjobs. For example, the pumps may be locatedupstream of the fluid analyzers for someapplications. (Adapted from Del Campo et al,reference 9.)

Power cartridge

Sample-bottlemodule

Sample pumpmodule

Sample fluidanalyzer

Hydraulic module

Focused-samplingprobe

Guard fluidanalyzer

Guard pumpmodule

10 Oilfield Review

approximately 270 miles [430 km] southwest ofNew Orleans and 175 miles [280 km] offshore (right).

To further evaluate the prospect, Chevrondrilled a second well, Jack 2, in Walker RidgeBlock 758 to a total depth of 28,175 feet[8,588 m]. Departing from typical procedures,Chevron planned a well test, which would makeJack 2 the only Lower Tertiary well ever tested inthe Gulf of Mexico. Acquiring a pure sample ofthe formation fluid prior to the production testwould aid significantly in reducing the fluiduncertainties in the test design and thereforeenhance the value of this expensive endeavor.

A unique MDT sampling toolstring config-uration allowed collection of traditionallyacquired fluid samples at two stations with anextralarge-diameter (XLD) probe, and focusedsamples at two stations with the Quicksilver Probemodule.12 Real-time analysis of flowline fluidacquired at one station with the XLD probe showsGOR increasing but not leveling off, even after8 hours of pumping (below left). Nevertheless,samples were collected at 30,000 seconds.

> The Lower Tertiary play in deepwater Gulf of Mexico, where Chevron discovered the Jack fieldin 2004. Other wells in the Lower Tertiary play are shown as dots.

S A L T

Houston New Orleans

G U L F O F M E X I C O

Jack

km

miles0

0 150

150

> Cleanup plot of flowline fluid acquired with asingle extralarge-diameter probe in the ChevronJack 2 well. The volume of fluid pumped duringsampling is shown in the top track. Real-timeanalysis of optical density measured with the LFALive Fluid Analyzer tool leads to quantification ofthe volume fraction of C6+ components,essentially liquid hydrocarbons (second track),and gas/oil ratio (GOR) (third track) as flowlinefluid becomes cleaner. GOR (blue) continues toincrease, indicating cleaner sampling, but doesnot level off, even after 8 hours of pumping.Laboratory analysis of samples collected at30,000 seconds showed the contamination levelto be greater than 10%. A data-quality flag track(bottom track) is green when data quality is high,and brown when data quality is lower.

0

Volu

me

fract

ion 1.0

0.5

0

GOR

0 5,000 10,000 15,000 20,000 25,000 30,000

Star

t pum

ping

Elapsed time, s

200,000

100,000

Pum

pout

vol

ume,

cm

3

> Pumpout volume, volume fraction and GOR plots for sample-line (left) and guard-line (right) fluidsobtained with Quicksilver Probe focused sampling. As seen in the pumpout-volume track (top), onlythe guard-line pump (red) operates from 0 to 7,340 s. Then, the sample-line pump (brown) is activatedand pumps until 11,500 s, at which time both pumps operate synchronously but at different rates.Cleanup can be seen by the increase in GOR (blue) in the guard flowline from 0 to 7,340 seconds,while the sample line is idle. Then, the guard pump stops and the sample-line pump starts. The GORseen by the sample-line LFA module increases gradually at first, and then, when flow is split at11,500 s, the sample-line GOR increases dramatically and reaches a plateau, indicating that the fluidis clean. The sample acquired at 14,000 s had a contamination level that was too small to measure.

0 2,500 5,000 7,500 10,000 12,500

Elapsed time, s15,000

0

Pum

pout

vol

ume,

cm

3Vo

lum

e fra

ctio

n 1.0

0.5

0

GOR

Star

t pum

ping

300,000

100,000

200,000

0 2,500 5,000 7,500 10,000 12,500

Elapsed time, s15,000

300,000

100,000

0

Volu

me

fract

ion 1.0

0.5

0

Star

t pum

ping

200,000

Pum

pout

vol

ume,

cm

3GO

R

Winter 2006/2007 11

Laboratory analysis later showed that thesesamples had more than 10% OBM contamination.

Samples acquired with the Quicksilver Probemodule were less than 1% contaminated after4 hours of pumping. Pumpout volume and GORplots from the guard and sample flowlines showthe stages of fluid cleanup (previous page,bottom right). From 0 to 7,340 seconds, the lowerpump initiated fluid movement into the boreholeby pulling commingled fluids through the guardflowline. Pumping early fluids through the guard-line pump helps optimize sample acquisition.Fluids that flow early are most likely to containmud solids that might potentially plug the pump.If plugging is going to occur, field engineersprefer the guard pump to plug instead of thesample pump. Although flow will not be focusedif pumped with a single pump, samples can stillbe taken if the sample pump is functioning.Samples cannot be collected with only the guardpump operating.

In this example, the guard pump stopped at7,340 seconds and the sample-line pump started.The GOR seen by the sample-line LFA Live FluidAnalyzer module started to build. The GOR seenby the guard-line LFA tool is ignored, since nofluid is flowing through it.

At about 11,500 seconds, the flow splits. Thesample-line GOR increases rapidly, indicatingfast cleanup as the contaminated filtrate isdirected away from the sample probe.Additionally, the GOR flattens immediately,indicating that the fluid is as clean as it can be.The GOR on the guard line is still increasing; itsfluid is still cleaning up. The sample acquired at14,000 seconds was later analyzed in thelaboratory and found to have a contaminationlevel that was too small to measure.

Use of the Quicksilver Probe device allowedhigh-quality fluid samples to be compared withthose of the discovery well, and providedimproved viscosity estimation compared withsamples obtained from more conventional toolconfigurations. The lower contamination samplesalso improved the DST design by quantifyingGOR for separator and surface-facilities require -ments, predicting PVT behavior and enhancingreservoir characterization.

The Jack 2 production test, completed inSeptember 2006, was the deepest successful welltest in the Gulf of Mexico.13 During the test, thewell sustained a flow rate of more than6,000 bbl/d [950 m3/d] of crude oil from about40% of the well’s net pay. Chevron and partnersStatoil ASA and Devon Energy Corporation planto drill an additional appraisal well in 2007.

Sampling at High Pressures and TemperaturesHigh-pressure and high-temperature (HPHT)conditions present a challenging environmentfor all aspects of drilling, logging, completing andproducing a well. HPHT wells are often deep andrequire the use of the most capable rigs, whichcan be costly. Wireline logging generally requiresmultiple runs to obtain the necessary infor-mation, and often encounters tool sticking, soexpensive and lengthy fishing operations arepotential risks. Pipe-conveyed logging is oftennot the best option because it subjects tools tolonger exposure to high temperature.

Traditional wireline logging with standardcables may not provide enough pull to deploylong toolstrings in deep wells, and may lackthe extra pull necessary to get a tool out if

sticking occurs. A recently developed deploy-ment method using a high-tension cable and ahigh-tension reduction unit, or capstan, can beeffective in these cases (above).14 This allows forthe rapid deployment of the logging string andmuch higher overpull, which reduces the risk oftool sticking.

12. Weinheber P and Vasques R: “New Formation TesterProbe Design for Low-Contamination Sampling,”Transactions of the SPWLA 47th Annual LoggingSymposium, Veracruz, Mexico, June 4–7, 2006, paper Q.

13. Chevron Announces Record Setting Well Test at Jack,http://www.chevron.com/news/press/2006/2006-09-05.asp (accessed September 22, 2006).

14. Alden M, Arif F, Billingham M, Grønnerød N, Harvey S,Richards ME and West C: “Advancing DownholeConveyance,” Oilfield Review 16, no. 3 (Autumn 2004):30–43.

> High-tension logging system for logging deep wells and sticky holeconditions with heavy toolstrings. The system comprises a standardSchlumberger wireline unit, a high-strength dual-drum capstan andhigh-strength wireline cable. The capstan increases the pull that canbe exerted on the wireline, so even heavy toolstrings can be retrieved,reducing the risk of sticking. An electrical fusible weakpoint, locatedin the logging head, is normally used with high-strength cable to allowmaximum pull on the tools if needed. The weakpoint can be brokenonly electrically from surface. Extremely deep wells, such as thosedeeper than 28,000 ft [8,500 m], require a modified wireline unit.

30 to 60 ft recommended

High-strengthwireline cable

High-strength wirelinedual-drum capstan

Standard Schlumbergerwireline unit

10,000

20,000De

pth,

ft

Weakpoint

In the UK North Sea, HPHT wells are usuallydrilled with OBM. Conventional wireline fluidsampling in OBM-drilled wells normally requireslong pumping times. This extended timeincreases the chances of tool failure related tohigh temperature, and can yield unsatisfactorysamples. Shorter cleanup time and better samplequality are needed for successful fluid samplingin HPHT conditions. Some examples from HPHTwells in the North Sea show how the focusingaction of the Quicksilver Probe tool reducescleanup time, minimizing tool exposure to hightemperatures, improving sample quality andreducing the risk of tool sticking.

In one HPHT example from the central NorthSea, ConocoPhillips (UK) Ltd applied new andproven technologies to overcome difficulties in

this hostile environment. Experience had shownthat in addition to high temperatures andpressures, wells in the area were prone to stickyhole conditions; the effects of depth, filtercakeproperties, hydrostatic overbalance and wellboretortuosity combined to hinder wireline-conveyedreservoir evaluation. Obtaining any data underthese conditions required assuring the drillingteam that appropriate steps would be taken toreduce the risk of logging tools becoming stuck inthe hole. Studies were conducted to confirm thatany hydrocarbon fluids pumped into the boreholeduring fluid analysis would not destabilize themud column. Pipe-conveyed wireline logging wasruled out for safety reasons.

ConocoPhillips (UK) Ltd had successfullylogged other challenging HPHT wells in theregion with a high-tension logging package. Inthese cases, wireline logging tools were loweredinto and raised from the wells using high-strengthcable and a capstan. The capstan, placed betweenthe drill floor and the wireline unit, increases thewireline pull from 9,700 to 15,500 lbm [4,400 to7,030 kg], ensuring that even long, heavy tool-strings can be retrieved.

In this near-field exploration well,ConocoPhillips (UK) Ltd planned to log the welland acquire pressures to determine mud weightsfor drilling deeper and to characterize theformation fluid. Sampling would not be requiredin the primary reservoir, because the fluids in thenearby producing fault blocks were known.Bottomhole temperatures were expected toreach 365°F [185°C] and formation pressurescould exceed 14,000 psi [97 MPa]. SlimXtremeslimhole HPHT logging tools, including the newanalog borehole seismic tool, would be run onhigh-strength cable with the capstan unit.

The well encountered an unanticipatedsecondary reservoir above the primary target,introducing uncertainty into the understanding offluid properties. The logging program wasimmediately modified to include fluid sampling inthese newly discovered zones. A Quicksilver Probemodule was readied to run in the hole and a PVTExpress onsite well fluid analysis system wasinstalled on the platform to perform surfaceanalysis of samples collected from the fournew zones.

In the first sand sampled using theQuicksilver Probe focusing device, the tooloperated smoothly, and flow through the toolbegan in commingled mode. Flow was splitbetween the sampling and guard probes after2,600 seconds of pumping, giving rise to anabrupt GOR increase from 850 ft3/bbl

[150 m3/m3] to around 1,500 ft3/bbl [270 m3/m3](left). GOR leveled off at 1,550 ft3/bbl[279 m3/m3] and remained high while sampleswere collected. Initial wellsite fluid analysisshowed high-purity oil. Final onshore laboratoryresults found contamination to be 1%.

PVT Express fluid-analysis experts on siteidentified 600 to 900 ppm of hydrogen sulfide[H2S] in the shallowest sand layer, which wouldbe incompatible with the completion design.These levels of H2S in the first sandstone led theConocoPhillips (UK) Ltd subsurface team toupgrade their scrutiny of the other unexploredlayers. Additional tool runs were scheduled forthe remaining sandstone intervals in the 8½-in.section, including many more sampling stations,assuming the Quicksilver Probe module couldwithstand the high temperatures.

At another station, Quicksilver Probe opera-tion began and remained with the tool bypassvalve in the open position. This means that theguard and sample flowlines were hydraulicallyconnected inside the tool, mimick ing traditional,single-probe sampling. Pumping continued formore than 14,000 seconds—about 4 hours—atwhich point samples were collected, because thefluid was not getting any cleaner (next page,top). Wellsite analysis determined contaminationto be 22%, which was confirmed by the onshorelaboratory result of 23% three weeks later.

Before moving away from this samplingstation, the field engineer managed to close thebypass valve and establish focused flow. Fluidflow split into the guard and sampling flowlines,GOR increased dramatically, and contaminationdecreased. PVT Express onsite analysis indicatedthat contamination levels fell from 22% to 1.5%.The fluid samples collected here showed 1%contamination in later laboratory analysis.Normal levels of H2S were found here and in theremaining layers. In all, 27 fluid samples wereacquired with good-quality results in every layer.

The well was subsequently completed,perforating only those layers that had been shownto have low H2S levels compatible with the tubingmetallurgy. Being able to acquire a suite ofuncontaminated downhole samples as part of arapidly evolving logging program was vital to thesuccess of the development of this secondaryreservoir. If high-quality samples had not beentaken and the well completion had proceededwithout this data, high concentrations of H2Swould have damaged the production tubing andentered the production facility. Mitigating thatwould have required shutting in production andperforming a costly well workover to identify andshut off the H2S-prone zone.

12 Oilfield Review

15. A safety sample is a sample that may be less than ideal,for example, one that is somewhat contaminated, but itis acquired anyway in case the tool fails and no furthersampling is possible.

16. A single-phase sample bottle maintains the single-phasenature of a fluid sample as it is brought to surface.

>Monitoring sample cleanup using the QuicksilverProbe focusing device. At 2,600 seconds, flowwas split between the sample and guard probes,with the guard probe pumping at a higher rate(brown, top track). GOR increased quickly from850 ft3/bbl to around 1,500 ft3/bbl, eventuallyreaching 1,550 ft3/bbl (third track). PVT Expressonsite well fluid analysis showed the oil samplecontains almost no OBM contamination and 600to 900 ppm of H2S. Results from onshore laboratoryanalysis found OBM-filtrate contamination to be1%. The data-quality flag (bottom track) is green,indicating high-quality data.

0

75,000

25,000

Pum

pout

vol

ume,

cm

3Vo

lum

e fra

ctio

n 1.0

0.5

0

500

0

GOR,

ft3 /

bbl

1,000 3,000 4,000 5,000 6,000 7,000

Elapsed time, s

2,000

1,000

1,500

50,000

Winter 2006/2007 13

HPHT Sampling Instead of DSTBG, drilling in another HPHT area of the centralNorth Sea, made a discovery with multiplehydrocarbon-bearing zones. This exploration wellwas designed not to have a DST, saving costs ontwo levels. First, a DST in this region would havecost US$ 10 to 20 million. Second, additionalsavings came from installing a less expensivecompletion. A DST would require heavier 97⁄8-in.casing to withstand the pressures of the well test,and a different well-test tree for supporting thewell-test equipment, totaling an additionalUS$ 4 million. Also, producing no reservoir fluidsto surface would avoid environmental risks.

Since no DST would be run, it was crucial toacquire high-purity samples by wireline. To allowreal-time analysis of formation fluids, theQuicksilver Probe tool was configured with LFAand CFA Composition Fluid Analyzer modules onthe sample flowline. A PVT Express systeminstalled on the rig analyzed contamination atthe wellsite. Shore-based experts were able toparticipate in logging and sample analysis in realtime through the InterACT real-time monitoringand data delivery system. By confirming samplepurity at the wellsite, engineers would know ifthe quality of the acquired sample was adequate,or if a new sample was required. BG fluid expertswere hoping for samples with less than 5% OBMcontamination. In addition to hydrocarbonsamples, the tool would acquire water samples ifit could sustain the high temperatures deeper inthe reservoir. Pressures were anticipated to be atleast 13,000 psi [90 MPa], and temperature wasexpected to surpass the 350°F [177°C] statedlimit of LFA and CFA operability.

In the first and shallowest hydrocarboninterval sampled, the temperature was already340°F [171°C]. Quicksilver Probe operationproceeded normally, starting with the guardprobe and sample probe connected through theinner bypass valve. The upper pump was used topump fluid through the sample line. The flow wassplit into guard and sample lines after3,050 seconds of pumping, at which time a jumpin GOR on the LFA plot indicated a significantdecrease in contamination of the fluid in thesample line (right). Less than two minutes later,when contamination reached an estimated 10%,the one-gallon sample bottle opened to collect asafety sample—a standard practice in difficultwells.15 This proved prudent, because soonafterward, the sampling-flowline pump stalled,but started again.

Contamination continued to decrease, andwhen GOR leveled off, a single-phase samplebottle was opened, filled and retrieved tosurface.16 PVT Express analysis on the rigquantified extremely low contamination,indicating that the sample was sufficiently pure,and the tool could be redeployed to the nextdeeper and hotter level. Independent onshorelaboratory testing conducted a few weeks laterdetected no contamination in this sample.

Three low-contamination samples weresuccessfully acquired at the next hydrocarbon-bearing zone, but after that, mud-check valves inthe sample flowline started to show signs ofplugging. However, the deepest and hottest zoneremained to be sampled. There, at the water-sampling station, slugs of borehole mud and OBMfiltrate were detected by the LFA module,indicating that unexpected fluid movement wasoccurring through the fluid-exit port. After sometime, this movement of fluid from the boreholehad cleared the mud-check valve, andsynchronized pumping to the guard and samplelines proceeded normally—despite the 361°F[183°C] bottomhole temperature—allowingacquisition of formation-water samples.

> Quicksilver Probe operation in a ConocoPhillips (UK) Ltd North Sea well, with the tool bypass valvein the open position (left). With the guard and sample flowlines hydraulically connected inside thetool, the effect is the same as conventional single-probe sampling. The pumpout-volume track (top)shows only the sample pump operating (blue). Cleanup is gradual, as seen by the slow increase of theGOR with time (third track). After more than 14,000 seconds, the fluid was not getting any cleaner, sosamples were collected. According to PVT Express wellsite analysis, contamination was 22%. Afterthe field engineer closed the bypass valve (right), fluid flow was split into the guard and samplingflowlines at around 15,500 seconds. Both the sample-line pump (blue) and the guard-line pump(brown) were pumping (top track), with the guard-line pump operating at a higher rate. GOR (thirdtrack) jumped to about 1,500 ft3/bbl, indicating a reduction in contamination. Onsite analysis with thePVT Express system quantified a contamination drop from 22% to 1.5%.

0

25,000

Pum

pout

vol

ume,

cm

3Vo

lum

e fra

ctio

n 1.0

0.5

0

500

0

GOR,

ft3 /

bbl 1,000

50,000

1,500

0 2,500 7,500 10,000 12,500 15,000

Elapsed time, s5,000

Star

t pum

ping

0

25,000

Pum

pout

vol

ume,

cm

3Vo

lum

e fra

ctio

n 1.0

0.5

0

500

0

GOR,

ft3 /

bbl 1,000

50,000

1,500

Elapsed time, s

Star

t pum

ping

0 2,500 7,500 10,000 12,500 15,0005,000

> Quicksilver Probe operation in BG’s HPHT NorthSea well. Before 3,050 s, fluid flowed through thesample-line pump (blue, top track). At 3,050 s, theguard-line pump (brown) and sample-line pumpoperated synchronously, with the guard-line pumpoperating at a higher rate. This split the flow,giving rise to an abrupt increase in GOR (blue) at3,050 s (second track). The sample acquired at8,700 s was found to contain no contamination.

0

40,000

20,000

Pum

pout

vol

ume,

cm

3GO

R

2,500 5,000 7,500 10,000 12,500

Star

t pum

ping

Elapsed time, s

Determining OBM contamination level ofwater samples downhole relied on interpretingdata from the color channels of the LFA module(left).17 Borehole mud, OBM filtrate andformation water each have distinctive signaturesin the visible and near-infrared frequency rangesmeasured by the tool. The two water samplescollected at this level contained some OBMfiltrate, but this was not a problem, because theOBM is immiscible in formation water.

BG estimates savings of up to US$ 24 millionby using the Quicksilver Probe focused-samplingmethod instead of a drillstem test to acquirezero-contamination samples.

Sampling Viscous OilsViscous oils can be especially difficult to sampleusing traditional sampling technology. With itsrelatively lower viscosity, OBM filtrate flowspreferentially to sampling devices, increasingsample contamination and often leaving high-viscosity formation fluids in the formation.

Cairn Energy India Pty Ltd experienced suchproblems acquiring oil samples in their Bhagyamfield in northwest India. The Bhagyam field isone of 19 fields in the Barmer basin tapping thehigh-permeability Fatehgarh sandstone. Oilreserves in the reservoir are currently estimatedat 1.5 billion bbl [240 million m3].18 Oil propertiesvary from field to field within the basin, and oilswithin the Bhagyam field exhibit compositionalgrading from crest to oil/water contact. With abetter understanding of oil properties, Cairnplans to optimize field development and surfacefacility design.

Bhagyam oils have high wax content, givingthem high pour point and high viscosity atreservoir temperature. Acquiring representative,PVT-quality samples has been a challenge.19

Before the arrival of focused-sampling tech-nology in India, most samples acquired bySchlumberger and other service companies usingtraditional openhole formation testers were toocontaminated to yield correct PVT propertiesduring laboratory analysis.20

To obtain contamination-free samples, Cairnhad resorted to collecting samples from casedwells using monophasic wireline-deployedsamplers. Samples acquired in this way may havelow levels of contamination, but can be collectedonly after the well has been completed.

In a campaign designed to improve samplequality, the Quicksilver Probe device collectedsamples in two Bhagyam wells.21 Of the18 samples acquired, 15 were of PVT quality. Sixof these showed no contamination. One such

14 Oilfield Review

> Detecting OBM contamination in water samples using the LFA module in the BG HPHT North Sea well.The top track shows quartz-gauge pressure and strain-gauge pressure along with unscaled resistivity(pink) and pump strokes (blue and green). The second track shows volume fraction of C6+ components,indicating OBM-filtrate contamination (green), OBM and solids (red) and water (blue). At 2,600 s, soonafter the guard and sample pumps start to pump synchronously, the sample line receives formationwater (blue). The third track contains the 10 LFA optical channels. Channel 0 (black) detects methane.Channels 6 and 9 (darker blues) detect water. The volume-fraction track detects water (blue) when LFAchannels 6 and 9 reach high amplitudes.

06

9

LFA

Optic

al C

hann

els

01.

0

Volu

me

Frac

tion

Pres

sure

Time, s1,000 2,000 3,000 4,000

Strain-gauge pressure

Quartz-gauge pressure

> Composition from laboratory analysis of viscous-oil samples acquired for Cairn Energy India inthe Bhagyam field in the Barmer basin. Samples, all from the same location, were collected usinga monophasic wireline (WL) sampler in cased hole (green), a conventional wireline formation testerin open hole (blue) and Quicksilver Probe focused sampling in open hole (red). The conventionalwireline formation tester yielded a contaminated sample, while the openhole sample acquired usingthe focusing method compared favorably with the one obtained in cased hole.

Mol

e co

mpo

sitio

ns

100

10

1.0

0.10

0.01

Contaminationweight %, 39.5

Uncontaminated sample, WLContaminated sample, 1035 Well AWell 1, Quicksilver Probe sample

Hydrog

en

Carbon

dioxi

de

Methan

e

Propa

ne

n-Buta

ne

i-Pen

tane

Hexane

s

Benzen

e

Heptan

es

Tolue

ne

Ethyl-

benze

ne

Ortho-x

ylene

Tri-M

e-ben

zene

Undeca

nes

Tride

canes

Penta

decan

es

Heptad

ecane

s

Nonad

ecane

s

Heneic

osane

s

Tricos

anes

Penta

cosan

es

Nonaco

sanes

Heptac

osane

s

Hentria

contan

es

Tritria

contan

es

Penta

triacon

tanes

I N D I A

RajasthanP A K I S

T AN

NEPAL

C H I N A

BANGLADESH

SRI LANKA

Barmer Basin

km

miles0

0 500

500

Winter 2006/2007 15

sample was analyzed for chemical compositionand compared with a sample acquired at thesame location by a traditional openhole form a tiontester and one obtained in cased hole (previouspage, bottom). The traditional openhole methodproduced a sample that was clearly contami natedwith several OBM components. The sampleacquired in cased hole showed an overallcomposition that was similar to the uncontam-inated Quicksilver Probe sample, with smallconcentration variations in a few components.

Following the successful acquisition ofcontamination-free samples and the availabilityof detailed fluid PVT data, Cairn has better fluid-property data for carrying out field-developmentstudies that involve reserves estimation,facilities design and flow assurance. Thesestudies will make a significant contribution toproduction at Bhagyam and demonstrate thepotential for improved field-developmentstudies worldwide.

Sampling in Mature FieldsThe Quicksilver Probe device is also a valuabletool for evaluating the efficiency of hydrocarbonrecovery in mature fields. Examples fromcomplex, mature fields in the Middle East showhow focused sampling acquired pure samples inhigh- and low-permeability formations andhelped assess gas-sweep efficiency.

The first well is in a reservoir that is undergas-cap expansion drive and water drive.Recently, several evaluation wells were drilled tomonitor sweep efficiency. The wells were drilledwith OBM and logged with the Platform Expressintegrated wireline logging tool and nuclearmagnetic resonance tools for openhole analysis.The Quicksilver Probe module of the MDT toolwas used to collect fluid samples.

The objective was to evaluate the efficiency ofthe gas-cap expansion in the main reservoir—aheterogeneous sandstone formation with perme -ability exceeding 1 darcy. Although low oilsaturations from MRX Magnetic ResonanceeXpert measurements indicated highly efficientsweep of the oil by expansion of gas, theformation tester was run to confirm that therewas no mobile oil in the swept zone. Theidentification of any remaining mobile oil would

indicate incomplete sweep. The QuicksilverProbe module identified and sampled fluids atfour stations in the gas zone and one station inthe oil column (below). All zones show thecharacteristic increase in GOR when flowthrough the guard line is split from the sampleline. Several gas samples were captured with no

OBM-filtrate contamination and with no mobileoil, indicating highly efficient recovery.

In the oil zone, the GOR measured by the LFAmodule was within 1% of the GOR already knownfor the field. The pumping time required for aclean oil sample in this zone was about 1,600seconds, roughly one-third of the time normally

17. Betancourt S, Fujisawa G, Mullins OC, Carnegie A,Dong D, Kurkjian A, Eriksen KO, Haggag M, Jaramillo ARand Terabayashi H: “Analyzing Hydrocarbons in theBorehole,” Oilfield Review 15, no. 3 (Autumn 2003):54–61.

18. “Wireline Sampling Technology Enables Fluid SamplingWithout Contamination,” JPT 5, no. 9 (September 2006):32, 34.

19. PVT-quality samples are those that have sufficiently lowcontamination, such that PVT properties measured in thelaboratory correspond to those of an uncontaminatedsample. The maximum allowable contamination variesby company and laboratory. A general rule is 7%contamination for this basin.

20. Alboudwarej H, Felix J, Taylor S, Badry R, Bremner C,Brough B, Skeates C, Baker A, Palmer D, Pattison K,Beshry M, Krawchuk P, Brown G, Calvo R,Cañas Triana JA, Hathcock R, Koerner K, Hughes T,

Kundu D, López de Cárdenas J and West C:“Highlighting Heavy Oil,” Oilfield Review 18, no. 2(Summer 2006): 34–53.

21. Kumar S and Kundu D: “Fluid Sampling in Oil BasedMud Environment: Quicksilver Probe Aids AcquiringContamination Free Samples in a ChallengingEnvironment,” to be presented at Petrotech 2007, the7th International Oil & Gas Exhibition, New Delhi, India,January 15–19, 2007.

> Pressures and sampling results from a high-permeability zone in a complex Middle East field.Formation pressures appear in Track 1, with gas identified as orange circles and oil as green circles.Gamma ray and caliper appear in the depth track. Track 3 contains resistivity curves and drawdownmobility (circles). Track 4 plots density and neutron porosity. Focused sampling stations with theQuicksilver Probe module are indicated with small probe insets. At the top sampling station, gas wassampled from a high-permeability zone. Downhole fluid analysis at this station (top right) showsvolume fraction and GOR values from the CFA Composition Fluid Analyzer. In the volume-fraction plot,yellow, red and green indicate C1, C2 to C5, and C6+, respectively. GOR values (magenta) point tosampling-line cleanup. At the third station, optical densities from the LFA Live Fluid Analyzer in thesample line (dark blue) are greater than those in the guard line (light blue), showing cleaner fluid inthe sample line. Analysis of CFA results at the same station displays similar results. At the fourthstation, in a high-permeability oil zone, LFA measurements of GOR detect cleanup in the sampleflowline (bottom right).

10,0005,0000

0.5

1

1.5

2

2.5

3

3.5

4

4.55

Sample line

Guard line

Formation Pressure, psi

Depth, ft

Gamma Ray

0 200gAPI

Caliper

4 12in.

Drawdown Mobility

0.2 20,000 0.45 -0.15vol/vol

Neutron Porosity1.95 2.95g/cm3

Density

mD/cP

100

50

0

30,00020,00010,000

05,000 10,000

100

50

040,000

20,000

05,000 10,000

750

500

250

0

1,000 2,000

0.2 20,000ohm.m

Resistivity

X,X90

X,X00

X,X10

X,X20

X,X30

X,X40

X,X50

X,X60

X,X70

X,X80

X,X90

X,X00

X,X10

X,X20

X,X30

X,X40

required for this zone in other wells (above).Laboratory analysis determined that contami-nation in the oil sample acquired by theQuicksilver Probe tool was too small to measure.

In a second Middle East example, an evalu-ation well penetrated six reservoirs, including adiscovery. The objectives were to obtain pressureprofiles, identify fluids downhole and acquireclean fluid samples. In addition, the operatorwanted to establish flow from low-porosity zonesthat previously had not been directly tested fortheir potential producibility. The operatorselected the Quicksilver Probe device because itslarge probe area was better equipped thanconventional large-diameter probes to establishflow and obtain samples from the low-porosity,low-permeability formations.

The well was drilled with OBM, and caliperdata showed good hole condition. Fluid samplingpoints were selected using free-fluid porosityreadings from the nuclear magnetic resonancelog. On the first pressure and sampling run, theMDT tool, deployed on wireline, acquiredpressure profiles using the large-diameter probe.On the second descent, the Quicksilver Probemodule and sampling units were run on drillpipe,and sampled at five stations (next page, top).

At the second sampling station, downholefluid analysis identified oil in a previouslyuntapped interval. Fluid mobility was so low in

this tight zone that a pressure measurementcould not be acquired. However, the LFA modulein the Quicksilver Probe tool successfullymonitored formation-fluid cleanup and founda low GOR for the liquid, leading to additionaloil reserves.

Downhole Fluid AnalysisIn most reservoirs, fluid composition varies withlocation in the reservoir. Fluids may exhibitgradations caused by gravity or biodegradation,or they may be segregated by structural orstratigraphic compartmentalization. One way tocharacterize these variations is to collectsamples for surface analysis. Another way is toanalyze fluids downhole, without bringing themto surface. Downhole fluid analysis (DFA) isemerging as a powerful technique to charac-terize fluids downhole. DFA helps determine thebest intervals for sample collection, if necessary.Analyzing fluid composition while the tool is stillin the hole also allows more detailed fluidcharacterization, because interpreters canmodify the fluid-scanning program in real time toinvestigate unexpected results.22

The ability of the Quicksilver Probe module tosupply uncontaminated fluids ensures optimalDFA results, and the faster cleanup time allowsseveral DFA fluid-scanning stations to beconducted efficiently without the long stationtimes associated with conventional sampling. Acombination of DFA and sample collection

helped a Norwegian operator understandfluids in a well drilled on the NorwegianContinental Shelf.23

The well was drilled as a final appraisalbefore development of an oil field. Because ofenvironmental restrictions, a production test wasnot planned, so it was critical to obtain uncon-taminated samples and fully characterize fluidvariations within the reservoir. The fluid analysiswould be used in the material selection of subseapipeline and surface facilities, process designand production planning. Because of the highpriority to capture representative hydrocarbonsamples without miscible contamination, thewell was drilled with water-base mud (WBM).

The Quicksilver Probe tool was run in the121⁄4-in. and 81⁄2-in. sections to collect samples ofgas condensate, oil and formation water, andfilled 19 sample chambers from many levels. Anexample from one of the more challenging zones,sampling oil in a relatively tight zone withmobility of 17 mD/cP, shows how the focusingtechnology results in an uncontaminated sample.

Fluid cleanup began with commingled flowfirst through the guard flowline, then through thesample flowline. After 1,300 seconds, flow is splitand focusing is achieved by increasing the flowrate in the guard probe (next page, bottom). Thereal-time GOR detected by the CFA modulestabilized at around 2,300 seconds, indicatingthat the fluid was clean. However, pumpingcontinued, and a sample was acquired at 2,800seconds. The spikes in the GOR curve indicatethe presence of produced fines from theformation, confirmed later when the sample wasanalyzed at surface. Wellsite analysis showedsome sand in the samples, but no detectablelevel of WBM filtrate.

In the same well, the focusing method createdoptimal conditions for DFA. The spectroscopicanalyzers that indicate when fluid in the flowlineis pure enough to sample also characterized thefluid composition in terms of three componentgroups: methane (C1), ethane to pentane (C2

to C5), and hexane and heavier (C6+). This allowsin-situ compositional analysis without collectinga sample and retrieving it to surface.

16 Oilfield Review

22. Fujisawa et al, reference 2.Mullins OC, Fujisawa G, Elshahawi H and Hashem M:“Coarse and Ultra-Fine Scale Compartmentalization byDownhole Fluid Analysis,” paper IPTC 10034, presentedat the International Petroleum Technology Conference,Doha, Qatar, November 21–23, 2005.

23. O’Keefe M, Eriksen KO, Williams S, Stensland D andVasques R: “Focused Sampling of Reservoir FluidsAchieves Undetectable Levels of Contamination,” paperSPE 101084, presented at the SPE Asia Pacific Oil andGas Conference and Exhibition, Adelaide, Australia,September 11–13, 2006.

> Comparison of pumping times to acquire clean oil samples in a Middle East field usingthe Quicksilver Probe module and a traditional probe. In this high-permeabilitysandstone, it took the Quicksilver Probe tool (red) only about 1,600 seconds of pumpingtime to draw low-contamination fluid into the sampling line, while the traditional probe(blue) pumped about three times as long to obtain a low-contamination sample.

GOR,

ft3 /b

bl

Time, s

0 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000 4,500 5,000600

650

700

750

800

Quicksilver Probe samplingSplit flowlines

Traditional probe sampling

Winter 2006/2007 17

> Sampling a new discovery in a mature Middle East field. This evaluation well penetrated six reservoirs, including a discovery. In additionto obtaining pressure samples, the operating company performed tests in bypassed low-porosity zones. Formation pressures appear inTrack 1, with oil identified as green circles and water as blue circles. Open circles indicate pressure measurements that do not fall on anygradient. Stars are pressures measured with the Quicksilver Probe tool. Track 2 contains drawdown mobility. Track 3 plots porosity andpore-fluid content with red for oil and blue for water. The second sampling station, at X,300 ft, was a discovery. LFA volume fraction andGOR results are plotted to the right of the porosity track (top right). Pump rates and GOR values for the third station are also shown to theright of the porosity track (middle right). A low-contamination sample was also acquired at the fifth station, at Y,100 ft—the first time oil hadflowed from this low-porosity formation. The GOR from this interval (bottom right) was found to be 250 ft3/bbl.

X,300

Y,100

1.0

0.5

0.00

–250

–500

0 5,000 10,000 15,000

40

30

20

10

0150

125

100

75

50

25

04,000 6,000 8,000 10,000

0 2,500 5,000 7,500 10,000 12,500

1.0

0.5

0.0300

200

100

0

Formation Pressure, psi

Dept

h, ft Drawdown Mobility

0.01 10,000

vol/volPorosity

mD/cP Water

Oil0.5 0

< Fluid cleanup in an oil well offshore Norway. QuicksilverProbe tool operation began with commingled flow throughthe guard flowline, as seen by the increase in guard-flow-line pumpout flow rate (light green, top track), then throughthe sample flowline (dark green, top track). After 1,300 s,flow is split and focusing is achieved by increasing thepumping rate in the guard probe. The GOR (bottom track)responds by stabilizing at around 2,300 s, indicating thatlow-contamination fluid is flowing through the sampleflowline. Sample flowline GOR is red, and guard flowlineGOR is blue. A sample was acquired at 2,800 s, and wasfound to contain no detectable WBM. The spikes in the GORcurve indicate the presence of produced fines from the formation sand. (Adapted from O’Keefe et al, reference 23.)

500

500

0

5

10

15

1,000

GOR,

ft3 /

bbl

Flow

rate

, cm

3 /s

1,000 1,500 2,000 2,500 3,0000

Elapsed time, s

Guard pump

Sample pump

Guard flowlineSample flowline

Two DFA fluid-scanning stations straddledthe gas/oil contact, only 0.5 m [1.6 ft] apart(above). The lower station at X,X46.0 m indicateda black oil with apparent density of 0.868 g/cm3

and a low GOR of 133 m3/m3 [740 ft3/bbl]. Half ameter higher, at X,X45.5 m, the fluid compositionshows a dry gas with apparent density of0.126 g/cm3 and C1 content in excess of 91 wt%.

The DFA measurements defined the gas/oilcontact within 0.5 m, which was a higheraccuracy than could be achieved using pretestpressure gradients in this well. The focusingcapability of the Quicksilver Probe tool ensuredthat the fluids being analyzed were repre -sentative of reservoir fluids, adding confidence tothe DFA results. Similar measurements at 15additional DFA stations helped quantify reservoir-fluid composition and delineate fluid contacts.

Collecting Formation Water Free of WBM ContaminationIn another example from offshore Norway,focused sampling helped acquire formation-water samples in an exploration well drilled withWBM. In this well, water characterization was akey factor in reservoir description and economicevaluation. The operator, taking special care tostudy water composition at different times in thelife of the well, collected water samples using theQuicksilver Probe tool, then also with a dual-packer probe in casing, and finally with adrillstem test.24

To facilitate quantification of WBM filtratecontamination, tritium, a naturally occurringisotope of hydrogen, was added to the WBM as atracer. Formation waters do not contain tritiumin measurable amounts, so tritium levelsdetected by laboratory testing could be easilyconverted to contamination levels.

The first sample was taken with theQuicksilver Probe focused-flow tool. Flow wassplit after 18,700 seconds, and the sample wascollected after 24,960 seconds of pumping.Laboratory analysis of tritium content showedthe sample to have 0% contamination.

The well was then cased with a 7-in. liner andperforated over the zone of interest. A wirelineformation tester was run inside the liner, alongwith inflatable dual packers to isolate the flowinterval. The increased flow area provided by the packers would minimize the drawdownrequired to extract samples and so reduce therisk of tool sticking.

After a long cleanup time—24 hours—duringwhich 1,700 liters [450 galUS] of fluid werepumped from the formation into the borehole, theformation tester collected two samples. Laboratoryanalysis indicated that the samples containedelevated concentrations of tritium, potassium,calcium and bromide, indicative of contaminationby completion brine and mud filtrate.

The operating company then performed a DST to test a gas zone above the water zone.Water flowing with the gas was collected foranalysis, but was found to be heavily contaminatedwith completion brine and also contained 46%hydrate inhibitor.

The Quicksilver Probe samples proved to bethe purest water samples ever collected from thefield, surpassing the quality obtainable fromconventional-probe sampling, cased-hole samplingor a DST. Analysis of the samples revealedunexpected compositional characteristics thatwere difficult to believe at first, but furtheranalysis of core and logs corroborated the newwater-composition results.25

In another water-sampling example fromoffshore Norway, both focused and conventionalmethods applied to the same formation helpedcompare cleanup performance. This explorationwell was drilled with water-base potassiumchloride [KCl] drilling fluid, adding difficulty tothe water-sampling program. Because theformation water and the WBM had similar opticalproperties, real-time qualification of contami -nation relied not on spectroscopic measure mentsbut on resistivity differences. For quantitativedetermination of contamination levels, theconcentration of potassium in the sample,corrected by subtracting the level assumed presentin the formation water, was divided by the knownconcentration in the WBM filtrate at each depth.

In the first sampling sequence, samples werecollected at three times during focused flow, at1,050 seconds, 7,050 seconds and 7,800 seconds,and showed 8.35%, 0.02% and 0% contamination,respectively. Temporarily switching off the guard pump shows the corresponding effect on contamination (next page, top). An addi-tional sample collected at 1,550 seconds, afterthe guard pump had been stopped, yielded 33.4% contamination.

18 Oilfield Review

24. O’Keefe et al, reference 23.25. O’Keefe et al, reference 23.

> Downhole fluid analysis to identify the gas/oil contact between two fluid-scanning stations separated by 0.5 m. Measurements with the CFA modulefound the pure fluid at X,X46.0 m (left) to be a black oil. The composition track (top) indicates 91 wt% C6+ (green) and 8.9 wt% C1 (yellow), with a GOR of 133 m3/m3. At the station 0.5 m higher (right), the compositional analysis yields C1 content (yellow) of 91 wt%, and C6+ (green) less than 5 wt%, indicating a dry gas with a GOR of 56,602 m3/m3. (Adapted from O’Keefe et al, reference 23.)

Elapsed time, s Elapsed time, s1,100 1,200 1,300 1,400 1,500 1,600 1,700 1,800 700

00.25

0.50

0.75

2,500

5,000

0

50

100GO

R, m

3 /m

3W

eigh

t%W

ater

fract

ion

00.25

0.50

0.75

2,500,000

5,000,000

0

50

100

GOR,

m3 /

m3

Wei

ght%

Wat

erfra

ctio

n

800 900 1,000 1,100 1,200 1,300 1,400 1,500 1,600

Data-qualityflag

Data-qualityflag

Depth = X,X46.0 m Depth = X,X45.5 m

C6+ weight%C2 to C5 weight%C1 weight%

C6+C2 to C5C1 GOR

8.9 weight% 0.1 weight% 90.9 weight% 133 m3/m3

C6+ weight%C2 to C5 weight%C1 weight%

C6+C2 to C5C1 GOR

91.0 weight% 2.7 weight% 4.6 weight% 56,602 m3/m3

Winter 2006/2007 19

For the next run, the field engineer shiftedthe toolstring 3.5 m [11.5 ft] higher in the sameformation and used a conventional single-probetool with a large-diameter packer to sampleformation fluid without focusing (above).Sampling at this station was designed to allow

direct comparison between the performance ofthe focused probe and the conventional probe.Three sample bottles were filled at this stationafter the same cleanup times as the threesamples in the focused-sampling sequence. Theconventionally acquired samples showed con -

tam ination levels of 26.2%, 8.6% and 8.2%. Notonly were the focused samples cleaner at everytime comparison, but the focused sampleacquired at 1,050 seconds (17.5 minutes) wascleaner than the conventional sample after7,050 seconds (2 hours).

Focusing on the FutureThe focused-sampling capability of the QuicksilverProbe tool provides higher purity fluid samples inmuch less time than traditional sampling tools.Benefits include higher quality fluid-propertydata, access to accurate fluid information earlierin the reservoir-charac terization process, reducedrisk of tool sticking, enhanced capabilities fordownhole fluid analysis and all the savingsassociated with getting the fluid characterizationright the first time.

For some E&P operators, especially thoseinvolved in deepwater activities, the technologyrepresents the best existing substitute forprohibitively expensive or environmentallyunfeasible DSTs. The tool can be run without theextensive upfront planning required for DSTs.Other operators have used high-qualityQuicksilver Probe results to optimize DST plans.In either case, the focused-sampling approachincreases efficiency, quality and safety in thesedemanding environments.

In several cases, Quicksilver Probe samplinghas yielded surprising results. The technology isencouraging some operators to review currentplans and resample zones where other technologieshave given unsatisfactory answers. Operatingcompanies that have had to drill with WBM for thepurpose of obtaining clean oil samples can nowconfidently drill with OBM, safe in the knowledgethat pure samples can be obtained downhole withno miscible contamination.

As a result of the high purity of new, incomingsamples, some laboratories have had to create anew classification for such low contamination,called “too small to measure,” or TSTM. Now thatsuch pure samples are available for laboratoryanalysis, researchers and experimentalists maybe able to perform additional analyses and devise new measurements to better understandfluid behavior.

An important consequence of the ability toobtain zero-contamination samples downhole isthe improvement in accuracy of real-timedownhole measurements. This will encouragecompanies to perform downhole fluid analysis fora more complete mapping of reservoir fluids thanis done today, and will also promote the additionof new DFA measurements. —LS

> Focused sampling of formation water in a WBM-drilled well offshoreNorway. Because the formation water and the WBM had similar opticalproperties, contamination was determined by changes in resistivity (blue),which leveled with time. This portion of the sampling log highlights theincrease in fluid resistivity when the guard-line pump (black) is temporarilyswitched off, operating only the sample-line pump (green). Sample-flowlinepressure is shown in red. (Adapted from O’Keefe et al, reference 23.)

X12

Pres

sure

, bar

0.030

Resi

stiv

ity, o

hm.m

X08

X10

X14

X16

X06

X04

X02

X00400 600 800 1,000 1,200 1,400 1,600 1,800 2,000

0.029

0.028

0.031

0.032

Time, s

Quartz-gaugepressure

Resistivity

Sample-pumpmotor speed

Guard-pumpmotor speed

> Conventional sampling of formation water in a WBM-drilled well offshoreNorway, for comparison with the results of the previous figure (above).The resistivity sensor was coated with mud initially, but began to respond(blue) partway through the sampling program. Resistivity never leveled off,indicating that water samples were still contaminated. Samples acquiredat this station after the same cleanup times as the three samples in thefocused sampling sequence showed contamination levels of 26.2%, 8.6%and 8.2%. (Adapted from O’Keefe et al, reference 23.)

Pres

sure

, bar

X10

X15

X20

X25

X30

X35

X40

X45

X50

0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000

0.060

Resi

stiv

ity, o

hm.m

0.055

0.050

0.045

0.040

Time, s

Sample-pump motor speed