Advancing Fluid-Property Measurements

56 Oilfield Review

Advancing Fluid-Property Measurements

Soraya BetancourtCambridge, Massachusetts, USA

Tara DaviesRay KennedyEdmonton, Alberta, Canada

Chengli DongSugar Land, Texas, USA

Hani ElshahawiShell International Exploration and ProductionHouston, Texas

Oliver C. MullinsJohn NighswanderHouston, Texas

Michael O’KeefeHobart, Tasmania, Australia

For help in preparation of this article, thanks to Gretchen Gillisand Don Williamson, Sugar Land, Texas; and Lisa Stewart,Cambridge, Massachusetts.CHDT (Cased Hole Dynamics Tester), Fluid Profiling, LFA (LiveFluid Analyzer), MDT (Modular Formation Dynamics Tester),Oilphase-DBR, PVT Express, Quicksilver Probe and RealVieware marks of SchlumbergerOLGA is a trademark of Scandpower AS.

Reservoir-fluid properties play a key role in designing and optimizing well completions

and surface production facilities to manage reservoirs efficiently. Therefore, accurate

fluid characterization is a vital part of any oil or gas production project. Advanced

fluid-analysis techniques provide the high-quality data required to develop appropriate

production strategies.

Discovery of an oil or gas accumulation imme di -ately prompts questions about its economicviability. Operators want to learn about theextent of the reservoir, the types of fluids thatwould be produced, expected production ratesand how long production might be sustained.Fluid analysis is a critical part of the process bywhich engineers perform reservoir characteri -zation, determine the reservoir architecture anddecide whether an oil or gas accumulation isworth developing. High-quality samples areessential, because erroneous data could leadengineers to misinterpret production parameterssuch as drainage volume, flow rates, reserves andfacilities design and completion. Clearly, poor ormisleading fluid data can have a severe negativefinancial impact.

If the reservoir analysis is positive, engineersbegin designing a production system that will

efficiently transport the reservoir fluids from theformation through wells, flowlines, productionfacilities and beyond. During this journey,reservoir fluids experience temperatures andpressures far different from their initial in-situconditions. These variations may inducephysical-state changes that would inhibit orinterrupt production if not understood prior todesigning tubulars and facilities. Therefore, todetermine how the fluids will respond toproduction conditions, engineers may want tocollect and analyze fluid samples from eachpotentially productive layer in the reservoir.

Traditionally, fluid samples are collected andsent to offsite laboratories for testing, a processthat delays data access and impedes an opera -tor’s ability to make time-sensitive developmentdecisions. Today, sophisticated formationsampling and testing tools allow data collection

1. Ratulowski J, Amin A, Hammami A, Muhammed M andRiding M: “Flow Assurance and Subsea Productivity:Closing the Loop with Connectivity and Measurements,”paper SPE 90244, presented at the SPE Annual Technical Conference and Exhibition, Houston,September 26–29, 2004.

2. For more on scale-removal techniques:Crabtree M, Eslinger D, Fletcher P, Miller M, Johnson Aand King G: “Fighting Scale—Removal and Prevention,”Oilfield Review 11, no. 3 (Autumn 1999): 30–45.

> Typical deepwater Gulf of Mexico oil-phase diagram. During the journey fromthe reservoir to the flowline, the oil temperature and pressure decline, and maycross phase boundaries at which asphaltenes (purple), waxes (blue) andhydrates (green) will tend to separate and form solid deposits. Gas begins toseparate from the oil as it passes through the bubblepoint boundary (red).

Wax

Hydrate

Bubblepoint

Reservoir

16,000

14,000

12,000

10,000

8,000

6,000

4,000

2,000

00 50 100 150

Temperature, °F

Pres

sure

, psi

200 250 300

Asphaltene

Flowline

60336schD7R1.qxp:Format Layout V7.2 12/4/07 7:11 PM Page 56

Autumn 2007 57

much earlier in the exploration process,including the ability to conduct real-time down -hole Fluid Profiling characterization ofreservoir-fluid properties and quantification oftheir variation. This technology gives operatorsthe ability to evaluate the reservoir while thesampling tool is in the borehole, and acquireadditional data if the reservoir is more complexthan previously thought. In addition, engineersand fluid-property specialists can betterdetermine where and when to sample, and howmany samples to collect. As a result, the qualityof fluid samples brought to surface issubstantially improved.

In the laboratory, chemists determine fluidcompositions, the temperatures and pressures atwhich phase transitions occur, and how eachphase behaves as a function of temperature andpressure. Accurate fluid characterization andknowledge of pressure-volume-temperature(PVT) behavior are critical for makingappropriate, cost-effective decisions about wellplanning, well construction, production and

monitoring. When initial PVT screening andthermodynamic modeling identify nonstandardphase behavior (such as emulsions, wax orasphaltene precipitation, hydrates and scales),specialized testing is often performed to betterunderstand reservoir-fluid behavior. All theseactivities fall under a general umbrella calledflow assurance.

When confronted with potential flow-assurance problems, engineers have several waysto mitigate or prevent difficulties.1 Thesemethods include thermal management (hot-fluidcirculation, electrical heating and insulation),pressure management (pumping, boosting andblowdown) and chemical treatments. Thesetechniques adjust the pressure-temperature paththat hydrocarbons experience during productionor, in the case of chemical treatments, alter thefluid composition to prevent phase changes ordisperse solid particles when precipitationoccurs. In addition, there are physicalremediation techniques such as pigging, jettingand cutting.2

As E&P companies venture into increasinglyremote production environments, particularlydeep water, flow assurance is criticallyimportant. Deepwater reservoir fluids follow atortuous PVT path from the formation to theproduction facility, increasing the probability offlow-assurance difficulties (previous page). Flowassurance may also be a concern in arcticenvironments where thermal differencesbetween the reservoir and surface facilities canbe extreme. Accurate knowledge of PVT behavioris vital because reservoir-fluid problems in theseremote locations could threaten the economicviability of a project.

This article describes the roles of chemistry,geology and thermodynamics during reservoir-characterization and flow-assurance studies atthe wellsite and the laboratory. Also presentedare two offshore-field case studies that demon -strate how these activities benefit well-completiondesign and operation.

Condensate

Median-weight oil

Water

Hydrocarbons

45

40

50

60

70

80

90

403530

454035302520151050

25201510

5

54

32

10

0


Reservoir-Fluid SamplingScientists and engineers of various disciplines relyon fluid-sample data when making decisions aboutreservoir development. For example, reservoirengineers use the data to determine reservoirarchitecture, estimate reserves, perform material-balance calculations and analyze fluid flow inporous media. Geologists need accurate informa -tion to perform reservoir correlations andgeochemical studies. Refining and marketingpersonnel make decisions about product yield andvalue. If erroneous data are used, unanticipatedand expensive consequences could result during production.3

A wide range of fluid behaviors can influencea sampling and analysis program. A reservoir-fluid system can be roughly categorized by itsvapor-liquid phase behavior; the classificationsrange from dry gas, wet gas and retrograde gas tovolatile oil, black oil or heavy oil (left).4 Anotherconsideration is hydrocarbon solid-phasebehavior. Wax and hydrate formation ispredominantly induced by a temperaturedecline, and pressure reductions or fluidcommingling generally cause asphaltenes to fallout of solution (next page, top).5

Engineers collect formation-water samples todetermine whether calcite, barite or halitescales will form within flowlines. Corrosive andtoxic substances such as carbon dioxide [CO2]and hydrogen sulfide [H2S] must be detected andmeasured because they influence tubular-alloyselection and the design of safety andenvironmental systems from the wellhead to thesurface production facility. Water pH is also animportant parameter governing scale andcorrosion, and may be measured downhole toavoid uncertainties.6

58 Oilfield Review

3. Nagarajan NR, Honarpour MM and Sampath K:“Reservoir-Fluid Sampling and Characterization—Key toEfficient Reservoir Management,” Journal of PetroleumTechnology 59, no. 8 (August 2007): 80–91.

4. McCain WD Jr: “The Five Reservoir Fluids,” in TheProperties of Reservoir Fluids (2nd Edition). Tulsa:PennWell Books (1990): 147–164.

5. Akbarzadeh K, Hammami, A, Kharrat A, Zhang D,Allenson S, Creek J, Kabir S, Jamaluddin A, Marshall AG,Rodgers RP, Mullins OC and Solbakken T: “Asphaltenes–

Problematic but Rich in Potential,” Oilfield Review 19,no. 2 (Summer 2007): 22–43.

6. Raghuraman B, Gustavson G, Mullins OC and Rabbito P:“Spectroscopic pH Measurement for High Temperatures,Pressures and Ionic Strength,” AIChE Journal 52, no. 9(2006): 3257–3265.Xian C, Raghuraman B, Carnegie A, Goiran P-O andBerrim A: “Downhole pH as a Novel Measurement Toolin Formation Evaluation and Reservoir Monitoring,”

Transactions of the 48th SPWLA Annual LoggingSymposium, Austin, Texas, June 3–6, 2007, paper JJ.

7. Riemens WG, Schulte AM and de Jong LNG: “Birba FieldPVT Variations Along the Hydrocarbon Column andConfirmatory Field Tests,” Journal of PetroleumTechnology 40, no. 1 (January 1988): 83–88.

8. Ruiz-Morales Y, Wu X and Mullins O: “ElectronicAbsorption Edge of Crude Oils and Asphaltenes Analyzedby Molecular Orbital Calculations with OpticalSpectroscopy,” Energy & Fuels 21, no. 2 (2007): 944–952.

> Generalized pressure-temperature (PT) diagram for reservoir fluids. Thediagram contains two principal regions: single-phase (green to orange) andtwo-phase (beige). The boundary between these regions is called thesaturation envelope; it has three principal features. The bubblepoint locus isthe portion at which gas begins to separate from liquid. The dewpoint locus isthe segment at which liquid begins to condense from gas. The critical point isthe location where the loci meet. The cricondentherm is the highesttemperature on the saturation envelope, and the cricondenbar is the highestpressure on the saturation envelope. Reservoir fluids are classified accordingto their in-situ reservoir and production behavior in the PT scheme. Dry gasdoes not enter the two-phase region at any point during the production PTpath. Wet gas remains a single-phase system in the reservoir regardless ofpressure depletion; however, during production, it crosses the dewpoint locusand forms a liquid phase. Retrograde gas resides in the single-phase region attemperatures between the critical point and the cricondentherm. Duringpressure depletion at reservoir temperature, liquid forms within the reservoiritself, and persists throughout the production PT path. Volatile oil resides in thesingle-phase region just to the left of the critical point. Gas liberation occurs asthe fluid crosses the bubblepoint locus during production. Black oil exists in thesingle-phase region at reservoir temperatures far lower than the critical point.Gas evolves during production, but the relative proportion of gas is smallcompared with that of volatile oil. Heavy oil is a subset of black oil that containsvery low quantities of gas, and the liquid phase is predominantly composed ofhigh-molecular-weight components.

Oil

Gas

50%

Criticalpoint

40%

Cricondenbar

Cric

onde

nthe

rm

20%

10%

5%

Temperature

Gas Field

with Retrograde

Condensation

Gas Field with

No Retrograde

Condensation

Oil Field

with Dissolved Gas

0%

Dewp

oint

locu

s

Pres

sure

Bubblepoint lo

cus

Oil and gas

> Crude-oil samples from a single column of oil in a reservoir. The continuous color change is a vivid illustration of compositional grading. (Photograph iscourtesy of Shell.)

Increasing depth


Autumn 2007 59

Another concern is the variability ofreservoir-fluid composition within a field orformation. Petroleum reservoirs may consist ofcompartments that are isolated from oneanother. Independent flow elements may have anenormous impact on an operator’s ability to drainthe reservoir. As an analogy, consider a singlereservoir compartment to be a sponge. Like asponge with its open-cell structure, the entirecontents can be drained by a single hole or well.Carrying the analogy further, independentreservoir compartments are similar to a spool ofplastic bubble wrap—a closed-cell systemthrough which the contents of one bubble cannotflow to another. If a single hole is punchedthrough the spool, drainage occurs only fromcells that are penetrated. The bubble-wrap spoolis thus highly compartmentalized.

An additional consideration is the existenceof large compositional variations of hydrocarbonsvertically and laterally within a compartment.Compositional grading is often caused by gravity,or nonequilibrium forces of biodegradation,temperature gradients, current charging, chargehistory or incompetent sealing shales.7 Themagnitude of compositional variation can varygreatly, depending on the geological andgeochemical history of the reservoir (previouspage, bottom).8

A technically robust fluid-sampling programis vitally important when reservoir compart men -talization, compositional grading or both exist.Reservoir-formation properties influence theability to collect representative fluid samples.Sample collection requires fluid flow into theborehole, which occurs only when the wellboreflowing pressure is lower than formationpressure. However, if the flowing pressure fallsbelow the fluid saturation pressure, a gas phase(in the case of volatile or black oil) or a liquidphase (in the case of retrograde gas) will form(right). The relative mobility of each fluid phaseis different; because of unequal flow, thecomposition of the fluid exiting the formationwill not be the same as that in the reservoir. Thiseffect can be minimized or eliminated bysampling at flow rates and pressure differentialsthat are as low as feasible.

Finally, accurate reservoir-temperaturemeasure ments are vital. Errors of just a fewdegrees during PVT testing could result inmisinter pretation. For example, what is conden -sate in the formation may behave like a volatileoil at an incorrect temperature in the laboratory.This error could result in costly production-design errors.

> Common deposits that form in tubulars during hydrocarbon production.Wax and hydrate deposition mainly result from a temperature decrease,while asphaltene precipitation may be triggered by changes in pressure,temperature and composition. Inorganic scales arise from changes inpressure, temperature and composition of aqueous fluids that accompanyhydrocarbon production. (With kind permission of Springer Science andBusiness Media.)

Asphaltene Wax

Inorganic scale Hydrate

> Pressure-decline effects during reservoir-fluid sampling. If the reservoirpressure falls below the bubblepoint while sampling oil, gas separation willoccur, creating a two-phase system (top). Similarly, if the reservoir containsretrograde gas, liquid will form if the reservoir pressure falls below thedewpoint. When phase changes occur in the reservoir, the high-mobility phaseflows preferentially because of relative permeability effects, and the sample isnonrepresentative. Maintaining the reservoir pressure above the bubblepoint ordewpoint during sampling preserves single-phase behavior and ensurescollection of a representative sample (bottom).

Pres

sure

Reservoir pressure

Originalreservoirfluid

Originalreservoirfluid

BubblepointWellbore flowing pressure

Free gas

One-phaserepresentativesample fluid

Pres

sure


There are two principal hydrocarbon-sampleacquisition methods—bottomhole and surfacesampling. Bottomhole sampling involves convey -ing a sampling tool on a drillstem-test (DST)string, wireline or slickline to the producing zoneor zones. In an open hole, sampling can be

performed by the MDT Modular FormationDynamics Tester and the Quicksilver Probe tool forfocused extraction of pure reservoir fluid. Cased-hole sampling devices include the CHDT CasedHole Dynamics Tester, single-phase reservoirsampler (SRS) and SCAR sampling tool.9 Surfacesampling, which is most frequently performed atthe separator under stable flow conditions,involves collecting gas and liquid samples.Engineers may acquire surface samples duringexploration if downhole methods are unavail able,

and may continue to do so throughout a well’slifetime to monitor fluid-property evolution.10

Bottomhole fluid samples must be extractedfrom locations that will yield the most relevantinformation for making decisions. To aid in thisendeavor, today’s sampling and testing toolsinclude an array of instruments that can performdownhole fluid analysis (DFA). DFA tools providereal-time fluid-property measurements atreservoir conditions, allowing engineers toanalyze samples before they are gathered.

60 Oilfield Review

9. For more on formation-fluid sampling devices:MDT tool: Colley N, Ireland T, Reignier P, Richardson Sand Joseph J: “The MDT Tool: A Wireline TestingBreakthrough,” Oilfield Review 4, no. 2 (April 1992): 58–65.Andrews RJ, Beck G, Castelijns K, Chen A, Cribbs ME,Fadnes FH, Irvine-Fortescue J, Williams S, Hashem M,Jamaluddin A, Kurkjian A, Sass B, Mullins OC, RylanderE and Van Dusen A: “Quantifying Contamination UsingColor of Crude and Condensate,” Oilfield Review 13, no.3 (Autumn 2001): 24–43.Quicksilver Probe: Akkurt, R, Bowcock M, Davies J, Del Campo C, Hill B, Joshi S, Kundu D, Kumar S, O’Keefe M, Samir M, Tarvin J, Weinheber P, Williams Sand Zeybek M: “Focusing on Downhole Fluid Sampling and Analysis,” Oilfield Review 18, no. 4 (Winter2006/2007): 4–19.CHDT tool: Burgess K, Fields T, Harrigan E, Golich GM,Reeves R, Smith S, Thornsberry K, Ritchie B, Rivero Rand Siegfried R.: “Formation Testing and SamplingThrough Casing,” Oilfield Review 14, no. 1 (Spring 2002):46–57.SRS and SCAR tools: Aghar, H, Carie M, Elshahawi H,Gomez JR, Saeedi J, Young C, Pinguet B, Swainson K,

> Schematic diagram of a MDT ModularFormation Dynamics Tester, using the QuicksilverProbe tool for focused extraction of reservoirfluid. The focused sampling probe is set againstthe borehole wall to withdraw formation fluids forFluid Profiling characterization and samplecollection. The downhole LFA Live Fluid Analyzersprovide real-time quantitative measurements ofdensity, viscosity, GOR, hydrocarbon compositionand formation-water pH.

Sampleflow

Power cartridge

Sample-bottlemodule

Pump module(sample)

LFA fluid analyzer(sample)

Hydraulic module

Focused samplingprobe

LFA fluid analyzer(guard)

Pump module(guard)

Guardflow

> Reservoir section (left) and real-time DFA GOR measurements (right) ofreservoir fluids in a deepwater exploration well. There is good agreementbetween the GOR values and the reservoir structure. The FCA analysissuggested locations for fluid sampling (blue dots, right). The reservoir fluidsvary significantly from dry gas (Fluids A and B) and condensate gases (Fluid C)at the top, to black oils with different GORs (Fluids D through J) at the bottom.At the bottom of the oil column (Fluids H, I and J), GOR variations indicate agentle fluid-composition gradient. On the other hand, a GOR inversion isevident between Fluids E and F; Fluid F is deeper than Fluid E, but has a higherGOR. A similar inversion occurs between Fluids G and J, suggesting a complexreservoir structure with a potential flow barrier at sampling station J.

Seal

Seal

Reservoir section

Water Rich gas and condensateOil

0

B

CDE

F GJH

I

500

100 1,000

Depth, ft

10,000 100,000

DFA GOR, ft3/bbl

1,000

1,500

2,000

2,500

3,000

3,500

Dry gas

A

Takla E and Theuveny B: “The Expanding Scope of WellTesting,” Oilfield Review 19, no. 1 (Spring 2007): 44–59.

10. For more on surface sampling: Aghar et al, reference 9.11. For more on optical DFA methods:

Betancourt S, Fujisawa G, Mullins OC, Carnegie A,Dong C, Kurkjian A, Eriksen KO, Haggag M, Jaramillo ARand Terabayashi H: “Analyzing Hydrocarbons in theBorehole,” Oilfield Review 15, no. 3 (Autumn 2003): 54–61.Crombie A, Halford F, Hashem M, McNeil R, Thomas EC,Melbourne G and Mullins OC: “Innovations in WirelineFluid Sampling,” Oilfield Review 10, no. 3 (Autumn 1998):26–41.Dong C, Hegeman PS, Carnegie A and Elshahawi H:“Downhole Measurement of Methane Content and GORin Formation Fluid Samples,” SPE Reservoir Evaluation &Engineering 9, no. 1 (February 2006): 7–14.

12. Betancourt SS, Fujisawa G, Mullins OC, Eriksen KO,Dong C, Pop J and Carnegie A: “Exploration Applicationsof Downhole Measurement of Crude Oil Composition andFluorescence,” paper SPE 87011, presented at the SPEAsia Pacific Technical Conference on IntegratedModeling for Asset Management, Kuala Lumpur,March 29–30, 2004.

13. Dong CM, O’Keefe M, Elshahawi H, Hashem M,Williams S, Stensland D, Hegeman P, Vasques R,Terabayashi T, Mullins O and Donzier E: “New DownholeFluid Analyzer Tool for Improved ReservoirCharacterization,” paper SPE 108566, presented at theSPE Offshore Europe Oil and Gas Conference andExhibition, Aberdeen, September 4–7, 2007.

14. Venkataramanan L, Weinheber P, Mullins OC,Andrews AB and Gustavson G: “Pressure Gradients and Fluid Analysis as an Aid to Determining ReservoirCompartmentalization,” Transactions of the 47th SPWLAAnnual Logging Symposium, Vera Cruz, Mexico, June 4–7, 2006, paper S.

15. Dong C, Elshahawi H, Mullins OC, Venkataramanan L,Hows M, McKinney D, Flannery M and Hashem M:“Improved Interpretation of Reservoir Architecture andFluid Contacts through the Integration of Downhole FluidAnalysis with Geochemical and Mud Gas Analyses,”paper SPE 109683, presented at the SPE Asia Pacific Oil and Gas Conference and Exhibition, Jakarta, October 30–November 1, 2007.

60527schD7R1.qxp:60527schD7R1 1/8/08 9:21 AM Page 60

Autumn 2007 61

DFA methods include in-situ opticalabsorption spectroscopy, optical reflectance,fluorescence and some nonoptical measure mentsincluding density, viscosity and pH. Thespectrometer operates in the visible to near-infrared range—at wavelengths between 400 and2,100 nm. Spectra are recorded in real time,revealing the proportions of methane [C1], ethaneto pentane [C2–5], hexane-plus [C6+] and CO2

fractions, as well as a gas/oil ratio (GOR) estimate.Additionally, differences between the reservoir-fluid and drilling-fluid spectra indicate the level ofsample contamination.11 Downhole fluorescencemeasurements provide fluid-phase informationthat is especially useful for retro gradecondensates and volatile oils.12 Fluorescence isalso sensitive to liquid formation in a condensategas when the flowing pressure falls below thedewpoint, allowing sampling engineers to monitorfluid-phase separation in real time, and ensurethat representative single-phase samples arecollected (previous page, top left).13

Fluid Profiling reservoir-fluid characteri -zation by DFA can diagnose compositionalgrading and help identify reservoir compart -ments. For example, abrupt fluid-composition orGOR changes between zones in a single well orbetween neighboring wells may indicate compart -mentalization. To confirm that perceivedfluid-property differences are truly significant,engineers must first consider measurementuncertainties. A recent method to evaluateuncertainties is the fluid comparison algorithm(FCA).14 FCA uses parametric models to estimateGOR and coloration uncertainties as a functionof optical-density (σε) and mud-contamination(ση) measurement variability. The algorithmcompares measurements acquired from twofluids and calculates the probability that differ -ences are statistically significant. When the FCAresult indicates that the fluids are different,sample acquisition for detailed surface analysisis justified. The following case study demon -strates how engineers employ DFA and FCA to characterize a reservoir and determinesampling locations.

Sampling and Reservoir Characterization in aDeepwater AccumulationIn a deepwater exploration well, Shell employedreal-time DFA and FCA to characterize thereservoir architecture and decide where tocollect fluid samples.15 As the MDT assemblytraveled down the well, the spectrometersmeasured the GOR at several locations (previouspage, top right). Based on FCA analysis, fluidsamples were collected at 10 different depths.

The analysis revealed wide variations inreservoir-fluid composition, ranging from dry gasand condensate gases at the top to black oils withdifferent GORs at the bottom. At the bottom ofthe oil column, the GOR varied gradually withdepth in the bottom sand, indicating a fluid-composition gradient. GOR inversions were alsodetected between the top and bottom sands,suggesting the presence of flow barriers and acomplex reservoir structure.

DFA and FCA analysis showed that fluidsabove and below the inversion had a greater than99% probability of existing in differentcompartments. A pressure discontinuity betweenthe compartments confirmed the absence ofhydraulic communication.

Shell and Schlumberger fluid specialistsexplored the oil column in detail, comparing theGOR information with petrophysical, formation-pressure and mud-gas logs, and performing FCAanalysis (below). The gamma ray and pressure

> Expanded log presentation and FCA analysis of an oil column at the bottom of a deepwaterreservoir. Correlation of the GOR log (top left) with the gamma ray log (green, top center) reveals thatFluids F and G come from one sand zone, and Fluids J, H and I reside in another. In the upper sand,Fluids F and G have the same GOR, while in the lower sand, Fluids J, H and I display a GOR decreasewith depth. The fluid densities derived from pressure gradients (blue) reveal two principal features—afluid-density inversion between Fluids G and J, and a gradual density increase from Fluid J to Fluid I—suggesting no vertical communication between the top and bottom sands. The mud-gas log (top right)offers further supporting evidence. At the depth of Fluid J, the δ13C value (red) falls abruptly andincreases gradually with depth. The relative methane concentration (blue) also fluctuates sharply atFluid J, then decreases gradually with depth—a feature consistent with the fluid-density increase.Final confirmation resulted from applying the FCA technique (bottom). The FCA algorithm generates anumber representing the probability that two fluids are statistically different. As measurementuncertainties (σ) decrease, confidence in data quality increases. Therefore, low σ values indicate ahigh probability that perceived fluid-property differences are real. The contour plot indicates that Fluid G has a 99% probability of being different from Fluid J, and a 95% probability of being differentfrom Fluid H. Discovery of reservoir compartmentalization led Shell to reevaluate the field and makesignificant strategic adjustments in reservoir development.

FG

H

I

J

Dept

h, ft

2,750

2,800

2,850

2,900

2,950

3,000

3,050

3,100

3,150

5,500 Pressure, psi 6,000

0 Gamma Ray, API 150

Relative Methane Concentration

Higher concentration ofC1 corresponds to lowerdensity of Fluid J

Increasingcontribution

of thermogeniccharge

0 2 4

–55 δ13C methane, parts per thousand –50

900 GOR, ft3/bbl 1,500

Top sand

Bottom sand

JH

I

GF

Oil-b

ase

mud

con

tam

inat

ion,

ση

Optical density, σε

0.10

0.09

0.07

0.08

0.06

0.05

0.04

0.03

0.02

0.01 0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10

Fluids G and H

P = 0.80

P = 0.99

P = 0.95

P = 0.90Fluids G and J


logs showed that the top portion of the oil residesin one sand lobe with a relatively constant GORand pressure gradient. The rest of the oil is in alower sand lobe, where the GOR decreases withdepth. The pressure log was consistent, indi -cating a significant fluid-density differencebetween the upper and lower lobes.

Recent advances in mud-gas logging havegiven engineers another tool to perform real-time detection of seals and permeability barriers,lithological variations and fluid contacts.16 Gases

gathered at the surface during drilling orseparated from fluid samples can be analyzed forisotopic content. The isotopic signature, δ13C, isthe value of the 13C/12C methane-isotope ratio ina sample relative to a standard, expressed inparts per thousand. When plotted alongside astandard mud-gas log, larger δ13C values mayindicate higher concentrations of biogenic gas inthe reservoir. Trending δ13C values may indicatenonequliibrium methane distributions, and aclear break in the methane-isotope signature in

mud gas may imply the presence of a seal.17

Engineers noted a δ13C discontinuity at around2,950 ft [899 m], suggesting yet another flowbarrier. FCA analysis provided confirmation,calculating a 95 to 99% probability that fluidsabove and below the discontinuity weredissimilar and from different compartments withno communication.

The discovery of the fluid compartments inthis deepwater field led Shell engineers to adjusttheir reservoir models and development plan -ning. Reservoir compartmentalization increasesdesign complexity and cost because engineersmust treat each zone independently. Productionforecasts, reserve calculations and enhancedrecovery schemes became proportionately more complex.

Improved regional understanding ofsubsurface architecture impacted short-termdecisions on sidetrack objectives. From thelocation of the borehole, Shell also reasoned thatcompositional grading in the lower reservoirlikely extended downward from the penetratedzone; as a result, production-facility plans werealtered to anticipate a GOR reduction with time.

Access to real-time fluid analyses allowedShell to make decisions much earlier in the field-development process, and accelerated theproject by at least six months. At today’s oilprices, the time saved was worth hundreds ofmillions of dollars.

Laboratory Fluid Preparation and Sample Chain of CustodyThe preceding case study demonstrates theconsiderable effort and care that engineers apply during the sampling process. Nevertheless,the harsh downhole environment and the nature of well operations may render DFAequipment, sampling devices and subsequentanalysis suscep tible to fouling, failure and other inaccuracies.

Schlumberger engineers addressed thisproblem by implementing a chain-of-custodyprocedure, a concept borrowed from forensicscience.18 Evidence must make the journey froma crime scene to the courtroom in a validatedand secure manner; otherwise, it may not beadmissible in court. Similarly, chemists at aremote testing laboratory should be able todetermine whether the chemical composition ofa field sample has been preserved. DFA providesa convenient way to establish a chain of custodyfor fluid samples, because chemists have theopportunity to compare analytical data acquireddownhole with those from the correspondingsamples that reach the laboratory.

62 Oilfield Review

> Chain-of-custody investigation of a valid crude-oil sample. Chemistscompare visible near-IR spectra from DFA and the laboratory sample. Bothspectra were measured at downhole conditions—15,000 psi [103 MPa] and250°F [121°C]. The spectral analysis (top) compares the downhole discretedata (red circles) with the continuous spectrum from the laboratory sample(blue). Data regression (bottom) of the LFA optical densities versus thosefrom the laboratory shows excellent agreement, evidenced by the near-perfect overlay of the regression line (red) over the X = Y line (blue). The fluidsample is well-preserved and suitable for further laboratory studies.

0.9

1.0

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0.0

-0.11,000 1,100 1,200 1,300 1,400 1,500

Wavelength, nm

Laboratory optical density

1,600 1,700 1,800 1,900 2,000

Optic

al d

ensi

tyLF

A op

tical

den

sity

1.5

1.0

0.5

0

0 0.5 1.0 1.5

Laboratory sampleLFA tool

X = YData regression

1,6001,920 1,445

1,6711,290

1,725

1,070

60527schD7R1.qxp:60527schD7R1 1/8/08 9:22 AM Page 62

Autumn 2007 63

In the field, after tool retrieval, engineersremove the reservoir-fluid samples. At this point,a PVT Express onsite well fluid analysis systemmay be available to conduct preliminarymeasurements that determine whether thecollected-sample properties agree with thosemeasured by DFA.19 If onsite analysis isunavailable or more sophisticated testing isrequired, the samples are shipped to a remotetesting laboratory in the original sample bottlesor transferred to an approved shipping container.When fluid samples arrive at the Schlumbergerlaboratory, chemists restore the fluid inside thesample container to the original reservoirtemperature and pressure, and allow the fluid toequilibrate by agitating it continuously for up tofive days. The restoration process is intended toredissolve precipitated asphaltene and waxparticles, ensure a homogeneous fluid through -out the sample cylinder and provide asingle-phase representative fluid for testing.

The visible near-infrared (IR) spectrum is aneffective hydrocarbon fingerprint.20 The DFAspectrometer performs sample analysis as thecrude-oil sample is acquired, providing a directfluid-property measurement under downholeconditions. In the laboratory, chemists performthe same measurement with a research-gradespectrometer at the downhole temperature andpressure. Differences between the DFA andlaboratory spectra may indicate that thelaboratory sample has been compromised. Forexample, if the methane concentration is lowerin the laboratory spectrum, then sample-bottleleakage or a fluid-transfer error may haveoccurred during sampling or transport to thelaboratory. The examples presented belowillustrate the chain-of-custody technique.

The first example involves a fluid sampleacquired from an offshore oil field. There isexcellent agreement between the downhole and laboratory spectra (previous page). Thisindicates that the sample is well-preserved andsuitable for further laboratory studies.

The second example involves another samplefrom an offshore oil field (right). The downholeand laboratory spectra have good overallagreement, but there are subtle differences inthe wavelength region above about 1,600 nm.Chemists investigated the spectral differencesfurther by employing algorithms to calculateGORs and probabilities of similarity between thetwo samples.21 The calculations showed that theGOR from the downhole spectrum, 580 ft3/bbl[103.3 m3/m3], was significantly higher than thatfrom the laboratory spectrum, 320 ft3/bbl[57.0 m3/m3]. This difference corresponded to a

16. Venkataramanan L, Elshahawi H, McKinney D,Flannery M, Hashem M and Mullins OC: “Downhole Fluid Analysis and Fluid Composition Algorithm as an Aid to Reservoir Characterization,” paper SPE 100937,presented at the SPE Asia Pacific Oil and GasConference and Exhibition, Adelaide, Australia,September 11–13, 2006.

17. Berkman T, Ellis L and Grass D: “Integration of Mud GasIsotope Data with Field Appraisal at Horn MountainField, Deepwater Gulf of Mexico,” AAPG Bulletin 86,no. 13 (2002): supplement.

18. Betancourt SS, Bracey J, Gustavson G, Mathews SG andMullins O: “Chain of Custody for Samples of Live CrudeOil Using Visible Near-Infrared Spectroscopy,” AppliedSpectroscopy 60, no. 12 (December 2006): 1482–1487.

19. Aghar et al, reference 9.20. For more on visible and near-IR spectroscopy:

Crombie et al, reference 11.21. Mullins OC, Beck G, Cribbs MY, Terabayashi T and

Kagasawa K: “Downhole Determination of GOR onSingle-phase Fluids by Optical Spectroscopy,”Transactions of the 42nd SPWLA Annual LoggingSymposium, Houston, June 17–20, 2001, paper M.Venkataramanan L, Fujisawa G, Mullins OC, Vasques RRand Valero H-P: “Uncertainty Analysis of Near-InfraredData of Hydrocarbons,” Applied Spectroscopy 60, no. 6(June 2006): 653–662.

> Chain-of-custody investigation of a compromised crude-oil sample. Bothspectra were measured at downhole conditions—20,000 psi [138 MPa] and 200°F[93°C]. The visible near-IR spectral analysis (top) shows subtle OD differences inthe region above about 1,600 nm. At 1,671 nm, indicating methane, thelaboratory-sample OD is lower than that measured by the LFA tool. At 1,725 nm,the methylene “oil peak,” the laboratory sample OD is higher than that measuredby the LFA tool. These differences are also visible in the linear regression plot(bottom). The ratio between the methane and oil peaks can be used to calculatethe GOR. In this case, the algorithms indicated that the laboratory-sample GORwas significantly lower than that of the field sample. Therefore, further analysiswould be necessary before this sample could be trusted.

1,000 1,200

0 0.5 1.0 1.5

1,400 1,600 1,800 2,000

Wavelength, nm

Laboratory optical density

1.6

1.4

1.2

1.0

0.8

0.6

0.4

0.2

1.5

0.5

0

1.0

0

Optic

al d

ensi

tyLF

A op

tical

den

sity

X = YData regression

Laboratory sampleLFA tool

1,600

1,671

1,820

1,280

1,445

1,725

1,070

60336schD7R1.qxp:60527schD7R1 1/14/08 5:06 PM Page 63

93% likelihood that either one (or both) of thespectra were in error, or the laboratory samplewas compromised. Further investigation ofsurface procedures and DFA-tool performancewould be required before this sample could beused with confidence in the laboratory.

Laboratory Techniques for Flow AssuranceIn the laboratory, chemists determine fluidcompositions and measure fluid propertiesrelated to flow assurance. For compositionalanalysis, an accurately measured volume of fluidis isobarically and isothermally transferred to apycnometer to measure mass and density. Then,the pycnometer is connected to an apparatus inwhich the oil sample is cooled to ambienttemperature and decompressed. The volume ofgas liberated by this procedure allowscalculation of the GOR.

Chemists employ gas chromatography todetermine the vapor-phase composition up toC15+ and the liquid-phase composition up to C36+.The bulk crude-oil composition is calculated bysumming the individual contributions from eachphase (above left). This procedure ensures that aconsistent reservoir-fluid composition is avail ablefor subsequent fluid-property characterizationand reservoir-production simulation.

The bubblepoint pressure is determined byperforming a constant composition expansion(CCE) test. Technicians place a known volume ofequilibrated fluid in a PVT cell at reservoirtemperature and pressure (left). The fluid isinitially single phase, and testing begins byreducing the pressure isothermally andmonitoring the fluid-volume change. Eventually,the fluid separates into two phases. Pressurereductions continue in increments ranging from100 to 500 psi [0.69 to 3.45 MPa], the vapor andliquid phases are equilibrated at each step, andphase volumes are measured and plotted as afunction of pressure. For a black oil, theintersection of the single- and two-phase lines inthe PV plot defines the bubblepoint.

The flow-assurance testing protocol dependson the nature of the crude oil. For example, whenasphaltenes are of concern, then saturate,aromatic, resin and asphaltene (SARA) analysisand paraffinic solvent titration with dead oil areprincipal screening techniques.22 It is alsocommon to measure the asphaltene-precipitationpressure on a live-oil sample. If testing identifiesan asphaltene-precipitation problem, additionalstudies are conducted to map out the asphaltenephase diagram and evaluate the effectiveness ofchemicals or coatings as prevention strategies.23

64 Oilfield Review

> Typical crude-oil compositional analysis, determined by gas chromatography. The plot shows thehydrocarbon contribution from C1 to C30+, as well as the CO2 concentration.

100.00

10.00

1.00

Components

CO2

H2S N2 C1 C2 C3

0.10

0.01

I-C4

N-C

4I-C

5N

-C5 C6

MCY

C-C5

Benz

ene

CYCL

-C6 C7

MCY

CL-C

6To

luen

e C8C2

-Ben

zene

M&

P-Xy

lene

O-Xy

lene C9 C1

0C1

1C1

2C1

3C1

4C1

5C1

6C1

7C1

8C1

9C2

0C2

1C2

2C2

3C2

4C2

5C2

6C2

7C2

8C2

9C3

0+

Wei

ght p

erce

nt

> Pressure-volume-temperature (PVT) cell and bubblepoint determination. The fully visual PVT cellallows direct confirmation of bubblepoints at various temperatures and pressures (top). A magneticmixer provides vigorous agitation to maintain phase equilibrium. A video-based cathetometermeasures fluid levels in the cell for phase-volume calculations. The oven heats the PVT cell to thetest temperature, and technicians monitor and plot phase volumes as a function of pressure. Theinflection point in the curves as pressure decreases defines the bubblepoint. In the case depictedhere, the bubblepoint is approximately 5,000 psi [34.5 MPa] (bottom).

High-pressure pumps

Oil

Solv

ent

Side view

Magnetic mixer

Cathetometer

16,030

30

2,030

4,030

6,030

10,030

12,030

14,030

8,030

Pres

sure

, psi

Volume, cm3

25 30 35 40 45 50 55 60 65 70 75

Tres = 176°FT = 120°FT = 75°F


Autumn 2007 65

Waxy crude oils pose different production andtransportation challenges. Wax deposition insidetubulars and pipelines reduces the effective flowarea, increasing the pressure drop and potentiallycausing complete blockage. Therefore, it isimportant to fully understand the oil’s behaviorthroughout the pressure and temperature pathfrom the formation to the production facilities.Temperature is the dominant parameter affectingwaxy crude-oil viscosity, gel strength, pour point,wax crystallization and deposition. Although deadoils such as stock-tank oil (STO) can be used togenerate preliminary data, it is important toinclude live oils in the testing program becausepressure and dissolved gases may stronglyinfluence wax solubility.

The first characterization step is to measurethe amount of wax that can precipitate anddeposit on a solid surface. Live-oil filtration andhigh-temperature gas chromatography (HTGC)are common methods to measure the waxcontent. HTGC is more valuable because itprovides the n-paraffin composition at highcarbon numbers (from C60 to C100)—informationchemists enter into thermodynamic models topredict wax behavior.

The wax-appearance temperature (WAT) isone of the most important flow-assurancemeasure ments, indicating the temperature atwhich wax crystals begin to form in a crude-oilsample. This measurement provides a preliminaryassessment of the likelihood of wax-relateddeposition problems. Laboratory workers placedead oil on the stage of a cross-polar microscope(CPM) and block light transmission by adjustingpolarized prisms at opposite ends of the sample.When illuminated by polarized light, crystallinematerials disturb the polarization plane;therefore, as the fluid sample cools, wax-crystalformation is clearly visible as bright spots appearagainst the black background (above right). Somelaboratories have high-pressure CPM instrumentsthat can measure the WAT in live oils. The samplecell operates at pressures up to 20,000 psi[138 MPa] and temperatures up to 392°F [200°C].

Formation and growth of wax crystals mayaffect the rheological behavior of the crude oil.Above the WAT, most hydrocarbon systemsbehave as Newtonian fluids; however, non-Newtonian behavior such as shear thinning maycommence as the fluids cool and phase changesoccur. This behavior must be properly quantifiedto allow engineers to design a suitableproduction system.24 Rheological properties ofinterest in the context of waxy crude are pourpoint, apparent viscosity and gel strength.

The pour point is the temperature belowwhich a fluid is no longer pourable because ofviscosification, gelation or solids formation.Pour-point testing with dead oils conforms toASTM Standard D97.25 For live oils, engineers usean apparatus containing a visual sapphire cellmounted on an automatic, vibration-free pivotingbracket that resides in a programmableconvection oven. The oil sample is heated to thereservoir temperature, and slowly cooled untilfluid movement in the cell ceases. Most waxycrude oils begin to display non-Newtonianbehavior near the pour point. If the pour pointexceeds anticipated oil-production temperatures,engineers may add pour-point-depressantchemicals to maintain flow.

The apparent viscosity of waxy crude oilsgenerally increases dramatically as thetemperature and shear rate decrease,particularly at temperatures near the pour point.Schlumberger engineers measure viscosity with arheometer that can operate at 6,000 psi[41.4 MPa] and 302°F [150°C], allowing workwith live oils.

Waxy crude oils tend to form gels attemperatures below the pour point. In the eventof a production shutdown, high pumping pres -sures may be required to break the gel andrestore flow. Therefore, yield-stress data fromwaxy crude oils are necessary to properly designflowlines and avoid production problems. The

22. Live oil contains the dissolved gases present in thereservoir. If the gases are released, or “flashed,” at thesurface or in the laboratory, the residual liquid is calleddead oil. Stock-tank oil (STO) is also a dead oil.

23. For a detailed discussion of asphaltene flow assurance:Akbarzadeh et al, reference 5.

24. At constant temperature and pressure, Newtonian fluidsdisplay constant viscosity at all shear rates. Theviscosity of non-Newtonian fluids is not constant at allshear rates. Apparent viscosity is the viscosity of a fluidat a given shear rate and temperature.

25. ASTM D97-06 Standard Test Method for Pour Point ofPetroleum Products. West Conshohocken, Pennsylvania,USA: ASTM International, 2006.

> Determination of wax-appearance temperature (WAT) by cross-polar microscopy (CPM). Themicroscope is equipped with a heated stage between two polarized prisms (top). At the beginning ofthe test, the prisms are adjusted to block light transmission. As the stage cools, wax-crystal formationchanges the light polarity, and the video camera sees the appearance of bright spots. In this example,no light is transmitted at 111°F [44°C] (bottom, right panel) because the fluid temperature is above theWAT. Spots begin to appear when the fluid cools to the WAT (center panel), and the image becomesbrighter as the fluid cools to 0°C (left panel). This method is accurate to ± 2°F [1.1°C].

0°C 44°CWAT = 42°C

10 50Temperature 48.6

IR filter

Polarizer

Analyzer

20

360°rotatable

stage

Coolinggas

Hot stagetop view

Chargedcoupleddevice

Hot stage


yield stress of live fluids can be determined by amodel pipeline test (MPT) (left).26

Another important measurement in thecontext of production-system design and flowassurance is the wax-deposition rate. Theprincipal controlling parameters are fluidtemperature, heat loss through the pipeline wall,wax content, shear rate and fluid viscosity. Thewax-deposition rate can be determined undersimulated pipe-flow conditions in a wax-deposition flow loop (WDFL) (below left).

The following case study illustrates howoperators use laboratory measurements to developoperational strategies that prevent, mitigate orremediate wax deposition and gel formation.

Optimizing Subsea System Design in West AfricaA West African field, located at a water depth lessthan 1,000 ft [300 m], has a reservoirtemperature and pressure of 170°F [76.7°C] and3,180 psi [21.9 MPa]. The seabed temperature is55°F [12.8°C]. The operator planned to produceoil through a looped 6-in. ID subsea tieback to ahost facility 2 miles [3.2 km] from the reservoir.With the large temperature difference betweenreservoir and seabed, and the long distance thatfluids would flow along the cold seafloor, theoperator needed assurance that solids depositionwould not impede flow. Schlumberger collecteddownhole fluid samples and sent them to theOilphase-DBR fluid sampling and analysislaboratory in Edmonton, Alberta, Canada for aflow-assurance study.27

Compositional analysis of the reservoir fluidrevealed a black oil with a GOR of 230 ft3/bbl[41.0 m3/m3] and an API gravity of 36.3.Compositional analysis indicated that thefraction with a carbon number above C30 was35.8%. Further characterization of the C30+

fraction and n-paraffin distribution revealed thatthe crude oil contained about 13.1 wt% C17+

n-paraffins.28 The relatively high n-paraffinconcentration was cause for concern that the

66 Oilfield Review

26. The yield stress, τy, is calculated by the following force-

balance equation: τy = where Py is the hydraulic pressure necessary to cause fluid movement, D is theinner diameter of the coil and L is the coil length.

27. Alboudwarej H, Huo Z and Kempton E: “Flow-AssuranceAspects of Subsea Systems Design for Production ofWaxy Crude Oils,” paper SPE 103242, presented at theSPE Annual Technical Conference and Exhibition, San Antonio, Texas, September 24–26, 2006.

28. Paraffin is a common name for a group of alkanehydrocarbons with the general formula CnH2n+2, where n is the number of carbon atoms. The simplest paraffinmolecule is methane, CH4, a gas at room temperature.Octane, C8H18, is liquid at room temperature. The solidforms of paraffin are heavier molecules from C20 to C40.Linear members of the series (those with no branches orcyclic structures) are called n-paraffins.

>Model pipeline test (MPT) apparatus. Stock-tank oil circulates through acoil of tubing immersed in a temperature-controlled bath. Flow through thecoil stops when the bath reaches the test temperature, allowing the fluid toage and form a gel structure. After the aging period, laboratory workersmeasure the nitrogen pressure necessary to initiate flow in the coil andcalculate the gel strength from a simple force-balance equation.

N2 pressure to break gel

System charging pump

Conv

ectio

n ov

en

High-pressurecirculation pump

Fluidsamplecylinder

Backpressureregulator

Heated lines

Temperature-controlled

bath

>Wax deposition flow loop (WDFL). The WDFL is a miniature flow loop that exposes stock-tank oils(STO) to a range of heat fluxes and shear rates that would be expected in the flowline during actualproduction. The deposition section of the flow loop is a 39-in. [1-m], 0.375-in. OD stainless steel tube. A 0.53-galUS [2.0-L] oil reservoir maintains the oil above the WAT throughout the test. The oil exits thereservoir and flows at a controllable rate through 50 ft [15 m] of coiled copper tubing in a temperingbath for temperature adjustment. The deposition loop is immersed in a cooling bath to simulate heatloss in the pipeline. As the oil flows through the deposition loop, a data-acquisition system monitors the wall temperature, oil and water temperatures, the pressure drop between the inlet and outlet of thecoil and the flow rate. Wax deposits constrict flow inside the deposition loop, increasing the pressurerequired to maintain flow. The increase in pressure allows calculation of the amount of deposited wax.

Thermocouples

Pressure transducer

Reservoir

Remelting bath Tempering bath

Pump

Flowmeter

Thermocouples

Deposition section

Cooling water

Cooling bath Thermocouples

PyD____4L


Autumn 2007 67

fluid might exhibit flow-assurance problemsrelated to wax deposition. Therefore, theOilphase-DBR team objective was to generatemeasurements that would provide guidanceconcerning ways to mitigate and remediate waxdeposition during steady-state and transientevents during crude-oil production.

Constant composition expansion (CCE)testing measured a bubblepoint pressure of 700 psi [4.8 MPa] at the reservoir temperature.WAT and pour-point tests were performed withboth live- and dead-oil samples (right). As thefluid pressure fell below the bubblepoint toambient conditions, dissolved gas escaped, theaverage sample composition shifted towardheavier hydrocarbons, and both the WAT andpour point increased approximately 20°F [11.1°C].

Rheological testing revealed gel-structureformation in live crude oil at low shear rateswhen the fluid temperature fell below the pourpoint and approached the seabed temperature(below). As the rheometer pressure decreased,the shear stress required to break the gelsincreased—behavior consistent with the loss oflighter hydrocarbons. Around the pour point,

>Wax-appearance temperature (WAT) and pour-point (PP) data measuredwith live and dead West African crude oil. The dead-oil measurements areshown at atmospheric pressure. As the live-oil fluid pressure decreased from700 psi to atmospheric pressure, both the WAT (red) and pour point (blue)increased approximately 20°F. The increases resulted from the loss ofdissolved gases in the fluid.

160

40

70

100

130

Tem

pera

ture

, °F

Pressure, psi0 200 400 600 800

WATPour point

1.0E+11

1.0E+10

1.0E+09

1.0E+08

1.0E+07

1.0E+06

1.0E+05

1.0E+04

1.0E+03

1.0E+02

1.0E+01

1.0E+00

Visc

osity

, mPa

-s

Shear stress, Pa

300 psi

10 100

1.0E+11

1.0E+10

1.0E+09

1.0E+08

1.0E+07

1.0E+06

1.0E+05

1.0E+04

1.0E+03

1.0E+02

1.0E+01

1.0E+00

Visc

osity

, mPa

-s

Shear stress, Pa

1.0E+11

1.0E+10

1.0E+09

1.0E+08

1.0E+07

1.0E+06

1.0E+05

1.0E+04

1.0E+03

1.0E+02

1.0E+01

1.0E+00

Visc

osity

, mPa

-s

Shear stress, Pa

100 psi

10 100

700 psi

10 100

Viscosity at 55° FViscosity at 65° FViscosity at 80° F

> Rheological behavior of a live West African crude oil. Shear-stressversus viscosity measurements were performed near and below the pourpoint at 100, 300 and 700 psi [0.69, 2.07 and 4.83 MPa]. At 700 psi (top left),elevated low-shear-rate plateau viscosities at 55° and 65°F [12.8° and18.3°C] indicated the presence of gel structures. Little gelation occurrednear the pour point at 80°F [26.7°C]. Similar behavior occurred at 300 psi(top right); however, higher shear stresses were required to break the gelsat 55° and 65°F—behavior consistent with the loss of lighter hydrocarbonsin the sample. At 100 psi (bottom left), the 80°F sample exhibited strongshear thinning behavior.


shear-thinning behavior with no gelationoccurred at all pressures. Further investigationof gel strength involved aging live- and dead-oilsamples for 12 hours at the seabed temperature.The live-oil gel strength between 100 and 700 psivaried from 38 to 42 Pa, and the dead-oil gelstrength at ambient pressure was more thanthree times greater—142 Pa.

Dead-oil wax-deposition tests wereconducted in the WDFL at temperaturesbetween 81° and 122°F [27.2°C and 50.0°C](top). The results revealed a deposition-ratediscontinuity at the higher end of the shear-raterange. For each shear rate, chemists entered themeasured deposition rates, the n-paraffindistribution, C30+ composition and viscosity

profile into a standard equation-of-state modelthat calculates an n-paraffin diffusioncoefficient. The model assumes that moleculardiffusion of wax molecules is the principaldriving force governing wax deposition.29

The next step involved entering the flow-assurance measurements, wax-deposition dataand diffusion coefficients into the OLGAsimulator—a commercial multiphase-flow fluid-transport model—to assess and predict crude-oilbehavior in various production scenarios.Schlumberger engineers applied OLGA simula -tions to two cases: producing at a steady stateand restarting production after a shutdown.

Because the WAT of the West African crude oilwas about 50°F [27.7°C] higher than the pourpoint, wax deposition and gel formation were theprincipal flow-assurance risks during steady-state production. The operator’s design goal wasto prevent wax deposition at production ratesabove 5,000 bbl [795 m3] of oil per day. Methodsto prevent or slow wax deposition in a pipelineinclude using insulated pipe, installing heatersand injecting paraffin inhibitors into the crude-oil stream. OLGA simulations determined thatpipe insulation was the most suitable flow-assurance method (next page). Having both live-and dead-oil data proved important. Knowingonly the dead-oil WAT would have led theoperator to believe that expensive pipe-in-pipeinsulation was necessary. Calculations using thelower live-oil WAT showed that more economicalwet insulation would be sufficient to prevent wax deposition.

Because the oil cools as it flows through thepipeline to the collection facility, it was essentialto predict the likelihood of gel formation.Assuming a production rate of 5,000 bbl/dthrough wet insulated pipe, the OLGA simulatorshowed that live oil could flow through thepipeline for about 20 hours before reaching its pour point—sufficient time to reach the stock tank.

For flow rates less than 5,000 bbl/d, the OLGAsimulator predicted rates at which wax deposi -tion would occur in the pipeline. Withoutexperimental WDFL wax-deposition data,engineers would have to use the standard modelto estimate the n-paraffin diffusion coefficient,predict a deposition rate, and schedule wax-removal operations. For this West African crudeoil, the standard model predicted that remedi -ation would be necessary every two weeks. WithWDFL data, the simulator predicted a muchlower deposition rate, increasing the timebetween remedial jobs to six weeks. Methods toremove wax deposits include pipeline pigging,

68 Oilfield Review

>Wax-deposition behavior of a dead West African crude oil. Depositionmeasurements at two shear rates, 170 and 511 s–1, revealed unusual behavior.At the lower shear rate (red), deposition slowed steadily with increasingtemperature. However, when the fluid temperature exceeded about 100°F[37.8°C], the deposition rates at the higher shear rate (blue) suddenly surpassedthose at the lower shear rate. Chemists repeated the measurements to verifythe reproducibility of this behavior. The deposition rates were entered intoequation-of-state and fluid-transport models that help engineers predict howthe crude oil will behave in various production scenarios.

40

10

15

20

25

30

35

Depo

sitio

n ra

te, m

g/m

2 s

Temperature, °F70 80 90 100 110 120 130

170 s-1

511 s-1

> OLGA calculation of flowline-restart pressures in 6-in. pipe. If the oil gels inthe flowline during a shutdown, pressure must be applied to overcome the gelstrength and initiate flow. The maximum pressure that can safely be applied is500 psi. The simulation shows that maintaining an internal-flowline pressuregreater than 100 psi would allow engineers to safely restart the flowline.

2,000

0

250

500

750

1,000

1,250

1,500

1,750

Rest

art p

ress

ure,

psi

Fluid pressure, psi

0 100 200 300 400 500 600 700 800

Maximum restart pressure


Autumn 2007 69

hot-oil circulation and solvent treatments withcoiled tubing.

A vital flow-assurance question concerns theability of the system to restart after a productionshutdown. According to the operator, 500 psi wasthe highest pressure that could be safely appliedto overcome gel strength and initiate flow(previous page, bottom). Using live-oil gel-strength data and fluid-transport parameters,the OLGA simulator calculated that fluid flowcould be restored as long as the pressure in thepipeline remained above about 100 psi.

This case study shows that reliance on dead-oil experimental data and standard wax-deposition calculations could lead operators tomake unduly conservative decisions whendesigning production systems. In this case, flow-assurance predictions based on live-oil dataallowed the operator to save millions of dollars in flowline costs and less frequent wax-removal operations.

Coming Advances in Waxy-Crude Flow AssuranceSignificant work is underway to continueimproving flow-assurance testing and fluid-property surveillance during a field’s productivelife. The West Africa case study demonstrated the benefits of performing flow-assurance experi ments with live crude oils. However, wax-deposition testing in devices like the WDFL hasbeen confined to dead oils. The WDFL could bemodified to perform high-pressure tests, butconsuming two liters of live oil would beprohibitively expensive.

Oilphase-DBR scientists overcame thisproblem by designing and introducing theRealView live solids deposition cell.30 Requiringonly 0.04 galUS [150 mL] of oil, the cell canoperate at pressures up to 15,000 psi [103.4 MPa],temperatures up to 392°F and Reynolds numbersup to 500,000.31 Unlike the WDFL, the oil resides in a cylindrical vessel. A rotating spindleat the center induces fluid movement. Thedevice can simulate production conditions oftemperature, pressure, composition, pipe-surface roughness and both laminar andturbulent flow. Turbulent-flow testing is useful

29. Hayduk W and Minhas BS: “Correlations for Predictionof Molecular Diffusivities in Liquids,” Canadian Journalof Chemical Engineering 60, no. 2 (April 1982): 295–299.

30. Zougari M, Hammami A, Broze G and Fuex N: “Live OilsNovel Organic Solid Deposition and Control Device: Wax Deposition Validation,” paper SPE 93558, presentedat the 14th SPE Middle East Oil and Gas Show andConference, Bahrain, March 12–15, 2005.

31. In fluid mechanics, the Reynolds number is adimensionless ratio of inertial forces to viscous forces.Turbulent-flow conditions exist when the Reynoldsnumber exceeds 3,000.

> OLGA model simulations of West African crude-oil behavior in a flowlineduring steady-state production. Engineers used the simulations as a designtool to choose pipe insulation and avoid wax deposition during production. The temperatures at which oil would arrive at the collection station are plottedagainst production rate (top). Wet insulation (red) is more economical than apipe-in-pipe (PIP) enclosure (blue), but it is four times less efficient. Having thelive-oil WAT was fortunate because it showed that wet insulation would besufficient at production rates greater than 5,000 bbl/d. The next graph (center)reveals the cooldown profile as oil travels through the flowline. Using wetinsulation (red), the cooldown time to gel formation would either be 12 hours to the dead-oil pour point or 20 hours to the live-oil pour point. The latter timeperiod is sufficient to displace oil from the wellhead to the collection station.The OLGA simulator also predicted two-week wax-deposition profiles alongthe flowline (bottom). The wax-deposit thickness predicted by the standardmodel (blue) is nearly three times higher than the value calculated fromlaboratory wax-deposition tests in the WDFL (red). As a result, the frequency of wax-removal treatments could be reduced significantly.

160

140

120

80

60

40

100

160

150

140

120

110

100

130

Tem

pera

ture

, °F

Time, h

0 5 10 15 20 25

Arriv

al te

mpe

ratu

re, °

F

Oil production rate, bbl/d

0 2,500 5,000 7,500 10,000 12,500 15,000 17,500 20,000

STO WAT

STO pour point

Live-oil pour point

Live-oil WAT

1.2

1.0

0.8

0.4

0.2

0

0.6

Depo

sit t

hick

ness

, mm

Flowline distance, mi

0 0.5

Wellhead Collectionfacility

1.0 1.5 2.0

Recommended wax-removal thickness

PIP insulationWet insulation

PIP insulationWet insulation

Standard n-paraffindiffusion coefficient

WDFL n-paraffindiffusion coefficient


because it simulates the shear environment atthe flowline wall (above).

Thanks to live-oil testing in the RealView cell,significant advancements in asphaltene flow

assurance have already been realized.32

Preliminary laboratory results show that live-oiltesting will lead to similar improvements for waxycrude oils. A recent laboratory study tested thebehavior of a black waxy crude oil in a hypothet -ical flow-assurance situation. Oilphase-DBRscientists assumed the oil entered a 6-in. ID, 2-mi long pipeline at a temperature and pressureof 170°F and 3,170 psi [21.9 MPa]. The simulatedoil-production rate was 5,000 bbl/d, and theseabed temperature was 65°F [18.3°C]. Withinthis scenario, engineers performed deposition-rate tests that compared the performance of STO and live oils in laminar and turbulent flow(above right). The results showed that waxdeposition is slower in turbulent flow, and live-oilwax-deposition rates are far lower than thoseobserved with STO.

The low wax-deposition rates observed withlive oils would significantly impact the hypo -thetical flowline design and the frequency ofwax-removal operations. OLGA simulationsdemon strated that the 2-mi flowline could beconstructed from foam-insulated pipe instead ofpipe-in-pipe, potentially saving US $4 million. Inthe context of flowline remediation, traditionalWDFL measurements with STO indicated thatwax removal would be necessary every two weeks.Simulations with live-oil data predicted that wax-removal would be necessary about once per year,providing significant operational savings.

As a field produces, the crude-oil propertiesoften change. For example, as gas condensatefalls below the saturation pressure duringdepletion, the condensate yield and the WAT mayfall. In a compositionally graded accumulation,

composition may change as fluids are producedfrom regions that were originally distant from theinitial sampling point. The impact on flowassurance may be significant.

Traditionally, engineers monitor fluid-property evolution by periodically sampling fromseparators or extracting live oil near perfora -tions, and performing flow-assurance tests in thelaboratory. In offshore and deepwater fields, thisapproach is costly.33 With the advent ofintelligent completions, equipped with sensorsthat transmit downhole temperature, pressureand flow rates in real time, production moni -toring can be performed remotely. Thesecompletions also incorporate remote-controlpumps and valves that engineers can use tomitigate flow-assurance problems.

Chemical sensors are being developed thatcan detect fluid-composition changes. Wheninstalled at strategic locations in the wellcompletion and along a pipeline, the sensors willprovide real-time data for monitoring of solidsdeposition, corrosion rates and rheologicalproperties. As a result, interventions forsampling or remediation will be performed onlywhen necessary.

Today, fluid sampling and analysis areprogressing to a point at which consistentstandards are applied along the continuum fromDFA and reservoir characterization, to samplingand laboratory analysis, and on to productionsurveillance. This integrated approach will beincreasingly valuable to operators makingreservoir exploration, development andproduction decisions, particularly in high-risk,remote locations. —EBN

70 Oilfield Review

32. Akbarzadeh et al, reference 5.33. Amin A, Smedstad E and Riding M: “Role of Surveillance

in Improving Subsea Productivity,” paper SPE 90209,presented at the SPE Annual Technical Conference andExhibition, Houston, September 26–29, 2004.

> Illustration and schematic diagram of theRealView live solids deposition cell. The cell canachieve turbulent flow and wall-shear conditionsthat reflect those found in flowlines (top). Theinner spindle inside the cell rotates to createfluid movement (middle). Wall temperature andsystem pressure can be independentlycontrolled. The deposition surface androughness can be changed by inserting specialsleeves. The wax deposit appears on thestationary-cylinder surface (bottom).

>Wax-deposition behavior of waxy crude oil in the WDFL and RealViewapparatus. Engineers performed tests with dead and live oils. In laminar flow,dead-oil wax-deposition rates were essentially the same in the WDFL andthe RealView cell. Turbulent flow in the RealView cell reduced the dead-oildeposition rate substantially. Live-oil deposition rates in the RealViewapparatus were even lower for both flow regimes.

30

25

20

15

10

5

0

Tota

l wax

-dep

ositi

on ra

te, m

g/m

2 s

STO,laminar

STO,laminar

STO,turbulent

Live oil,laminar

Live oil,turbulent

24.5 23.7

15.5

9.2

3.3

WDFL RealView cellMultipoint

thermocouples

Shear cell

Variable-speedDC motor

High-pressureflow-controlvalves

Mechanicalmountingstand

Innerrotatingcylinder

Outerstationarycylinder

Deposit

Coolant

Coolantout

Charge

Discharge

Coolantin

Oil

Electricalheatingcartridge


Documents

Advancing Fluid-Property Measurements