20 Oilfield Review

Bill ChaseImperial Oil Resources Ltd.Norman Wells, Northwest Territories,Canada

Walt ChmilowskiRichard MarcinewChuck MitchellCalgary, Alberta, Canada

Yen DangHouston, Texas, USA

Kevin KraussErik NelsonSugar Land, Texas

Tom LantzPhillips Petroleum CompanyLafayette, Louisiana, USA

Chuck Parham (consultant)Coastal Oil & Gas Corp.Houston, Texas

Jerry PlummerLafayette, Louisiana

For help in preparation of this article, thanks to VinceDeBonis, David Norman and Paul Price, Dowell, NewOrleans, Louisiana, USA; Chuck Ebinger, CompletionEngineers, Lafayette, Louisiana; Larry Foster and HalRiordan, Dowell, Lafayette, Louisiana; Terry Greene,Dowell, Sugar Land, Texas, USA; David Pierce, Dowell,Houston, Texas; Rob Tinis, Dowell, Red Deer, Alberta,Canada; and Mathew Samuel, Dowell, Longview, Texas.ClearFRAC, DataFRAC, PERMPAC and STIMPAC aremarks of Schlumberger.

Clear Fracturing Fluids for IncreasedWell Productivity

Hydraulic fracturing treatments can greatly

improve well productivity by decreasing wellbore

skin. Residue from conventional polymer-base

fluids, however, may clog pore spaces in the

proppant pack, reducing fracture permeability. A

new fracturing fluid offers a solution by not using

polymers. This viscoelastic surfactant fluid requires

no chemical breakers, yet cleans up better than

crosslinked polymer fluids, leading to a lower skin

and greater well productivity.

Autumn 1997 21

A successful hydraulic fracturing treatmenthas too long been defined as one that waspumped without problems. Rather, the truemeasure of a successful fracture treatment isincreased production or injectivity.1 Theobjective is to improve fluid communicationbetween the reservoir and the well. One ofthe most significant developments inhydraulic fracturing has been the realizationthat many stimulation treatments might notproduce a negative skin or decrease the skinas much as desired.2 Laboratory testing indi-cates that unbroken residue from solids-basepolymer fluids can plug the pores in theproppant pack. Some of this polymerresidue can remain in the well indefinitely,hindering production.3

The ideal fracturing fluid should showminimal pressure drop in the pipe duringplacement, have adequate viscosity tocarry proppant effectively and degradeafter the fracture closes so as not to leaveresidual material.

From a historical perspective, fracturingfluids have evolved significantly since thefirst fracture stimulation was performed in1947. Early stimulation treatments used sur-plus napalm added to gasoline to create aviscous fluid capable of initiating and propa-gating a fracture. During the 1950s, viscous

oils and gelled oils were the fluids of choice,as it was generally believed that water intro-duced into oil reservoirs would cause forma-tion damage. With the realization that waterdid not cause as much damage as originallythought, engineers began pumping watergelled with guar and guar derivatives (lineargelled water) in the 1960s.

As fracturing grew in popularity in the1970s, wells were also being drilled deeperand hotter formations were encountered.There was an increasing need for fluid vis-cosities greater than those offered by lineargels. To attain sufficient viscosity andincreased thermal stability in higher temper-ature reservoirs, linear gels were crosslinkedwith borate, zirconate or titanate ions.4

In the 1980s, foamed fracturing fluidsgrew more popular as engineers becamemore aware of the permeability damagecaused by polymeric fluids. The use offoams decreased the amount of guar intro-duced into the fracture, thereby reducingthe amount of residue and, hence, theextent of damage. In a foam fluid, the gasphase typically occupies more than half ofthe total fluid volume, so less liquid, andhence less guar, is pumped into the well.Foams also enhance cleanup after a fractur-ing treatment. The liquid volume is lower,and the entrained gas offers significantlymore energy to evacuate the fracturing fluidfrom the well.

The quest for cleaner fluids continued intothe 1990s, when advanced breaker technol-ogy and lower polymer concentrationsbecame effective tools for reducing and lim-iting damage from guar.5

The next step was the development of apolymer-free aqueous fracturing fluid. Thisfluid is unlike guar or hydroxyethyl cellulose(HEC) fluid systems; rather, it belongs to anew class of fracturing fluids—those basedon viscoelastic surfactants. This article dis-cusses the new polymer-free fracturing flu-ids and, with specific case studies, detailshow they may be used most effectively toincrease well productivity.

1. Martinez SJ, Steanson RE and Coulter AW: “FormationFracturing,” in Bradley HB (ed): Petroleum EngineeringHandbook. Richardson, Texas, USA: Society ofPetroleum Engineers (1987): 55-1–55-12.

2. Skin is a zone of reduced permeability near the well-bore. Skin damage causes an excess pressure droparound the wellbore and reduces formation fluid flowinto the wellbore.

3. Stewart BR, Mullen ME, Howard WJ and NormanWD: “Use of a Solids-Free Viscous Carrying Fluid inFracturing Applications: An Economic and Productiv-ity Comparison in Shallow Completions,” paper SPE30114, presented at the SPE European FormationDamage Control Conference, The Hague, The Netherlands, May 15-16, 1995.

4. Ely JW: Stimulation Engineering Handbook. Tulsa,Oklahoma, USA: PennWell Publishing Company(1994): 79-97.

5. Armstrong K, Card R, Navarrete R, Nelson E, Nimer-ick K, Samuelson M, Collins J, Dumont G, Priaro M,Wasylycia N and Slusher G: “Advanced FracturingFluids Improve Well Economics,” Oilfield Review 7,no. 3 (Autumn 1995): 34-51.

22 Oilfield Review

Viscoelastic Surfactant DevelopmentIn 1983, The Dow Chemical Companyintroduced fatty amine quaternary ammo-nium salts containing alkyl groups longerthan C14 as thickeners in consumer prod-ucts, such as bleach, liquid dishwashingdetergent and topical cosmetics. These vis-coelastic surfactants belong to a class ofcompounds that form micelles in an aque-ous system containing certain anions.6 Thedeformation of such systems is time depen-dent; the system acts as a solid unless a suf-ficient amount of shear has been appliedfor a certain length of time. When the sys-tem deforms, the rheological behavior isnearly Newtonian.

A viscoelastic surfactant fluid providesexcellent particle suspension. Dowellapplied this fluid technology first as agravel-pack fluid, PERMPAC fluid. The sur-factant is added to common completionbrines—potassium chloride, ammoniumchloride, calcium chloride or calcium bro-mide—to suspend gravel effectively. The sur-factant concentration varies from 2.5 to 6%by volume, depending on the anticipatedtemperature in the well. The main advantageof this fluid, unlike polymer-base systemssuch as guar or HEC, is that little residue isleft upon breaking. This type of treatmentresults in a gravel pack with significantlyhigher conductivity.7

The principal advantages of viscoelastic-surfactant fluids are ease of preparation,minimal formation damage and highretained permeability in the proppant pack.The fluids are typically prepared by continu-ous mixing of the surfactant into the brinebefore it passes through a high-shearblender. The blender provides sufficientshear for complete dispersion of the surfac-tant and fluid viscosification. Viscoelasticsurfactant fluids can also be used in fracpack and in conventional high-permeabilityfracturing treatments.8 The originalPERMPAC surfactant works well in theseapplications, but cost and temperature limi-tations—less than 140°F [60°C]—preventedwidespread use in hydraulic fracturing appli-cations. Recent modifications to the chemi-cal structure of the surfactant have reducedthe fluid cost and increased the temperaturelimit to 200°F [93°C], opening the door tohydraulic fracturing applications.9

The ClearFRAC surfactant is a blend of aquaternary ammonium salt, erucyl bis (2-hydroxyethyl) methyl ammonium chloride(derived from rapeseed oil), with iso-propanol. ClearFRAC fluid is a mixture ofthis surfactant in brine. The preferred brinecompositions include 3% by weight of

ammonium chloride and 4% by weight ofpotassium chloride solutions. At tempera-tures greater than 150°F [66°C], sodiumsalicylate is added as a stabilizer. The sur-factant concentration varies from 0.7 to 4%by volume.

The surfactant molecule consists of ahydrophobic tail that is 22 carbon atomslong. The head group is hydrophilic, and iswhere the quaternary ammonium group islocated. When the surfactant is added to thebrine, the surfactant molecules aggregateinto structures in which the hydrophobic

■■Micelle structure.The micellar struc-ture of the vis-coelastic surfactantin brine appearsrod-shaped orworm-like as themicelles becomeentangled. Theworm-like micellesexceed 100 micronsin length and con-tain several thou-sand molecules.When the viscoelas-tic surfactant fluidbreaks, the micellesbecome spherical,with a diameterroughly equal tothe width of therod-shapedmicelles. Thisimage was takenwith a cryo-trans-mission electronmicroscope at theUniversity of Minnesota.

Hydrocarboncore

Surfactantmolecule

ActivatorSurfactantmolecules

Activator

Brine environment Oil environment

■■Breaking ClearFRAC fluids. No additional chemicals are usually needed to break ClearFRAC fluids. Dilution by formation water or contact with hydrocarbons will disruptthe rod-shaped micelles, breaking the fluid.

1µm

Scale

Autumn 1997 23

tail groups are on the inside, and thehydrophilic groups are on the outside. Suchstructures are called micelles. In the case ofClearFRAC fluid, the micelles are rod-shaped or worm-like (previous page, top). Ifthe surfactant concentration is above a criti-cal concentration, the micelles entangle andhinder fluid movement. Such interactionsproduce the fluid’s viscosity.

The viscosity of ClearFRAC fluids is brokenby two mechanisms: contact with hydrocar-bons or dilution by formation water (previ-ous page, bottom). Because one or both ofthese conditions occur in fractured wells, noadditional breaker chemicals are required;however, there are some common additiveswhich can contribute to the break mecha-nism. Produced oil, condensate or dry gasaffects the electrical environment in thefluid, disrupting the micelles. The micelleschange shape from rods to spheres, andfluid viscosity is lost because the micellescan no longer become entangled. In thecase of formation water, dilution of ClearFRAC fluid reduces the surfactant con-centration, and the rod-shaped micelles nolonger entangle with one another.

The field application of these fluids hasbeen successful. Present applicationsinclude wells in which fracture conductiv-ity or fracture length is important, mobi-lization of complex mixing equipment isdifficult, or situations where cleanup is anoverriding concern.

Operationally, the preparation of Clear-FRAC fluids is simple (left). Because nopolymer hydration is required, the surfactantconcentrate can be metered continuouslyinto the brine for easy mixing. No crosslink-ers, breakers or other chemical additives arenecessary. The mixing of the fracturing fluidis simplified by elimination of variances dueto polymer hydration and breaker effectsand the need for extensive metering andpumping systems. Moreover, there is lesswaste due to elimination of tank bottoms,the unpumpable residual fluid remaining inthe bottom of the containers used in batch-mixed jobs.

Equipment and Materials for ClearFRAC Fluids

Frac tank

Brine

Proppant

ClearFRAC surfactant

Pump truckBlender

Well

Proppant

Guar,breakers andsurfactants

Fracturing Equipment and Materials

Continuousmixer

Frac tank

WaterBactericide

KCI orsubstitute

Pump truck

Blender

Breakers andfluid-loss additives

Well

Crosslinker,buffers and

breakers

■■Mixing procedures. A typical crosslinked guar-base fracturing fluid may use up to 13different additives to provide the desired rheological properties. Complex metering andpumping equipment is needed to ensure accurate delivery of the chemical concentra-tions (top). In the simplified ClearFRAC mixing system, the surfactant is added to a brinesolution, then mixed with proppant and pumped downhole (bottom). Fewer chemicalsand less equipment reduce the likelihood of problems during a job.

6. A micelle is a molecular aggregate.7. Parlar M, Nelson EB, Walton, IC, Park E and DeBonis

VM: “An Experimental Study on Fluid-Loss Behaviorof Fracturing Fluids and Formation Damage in High-Permeability Porous Media,” paper SPE 30458, pre-sented at the SPE Annual Technical Conference andExhibition, Dallas, Texas, USA, October 22-25, 1995.

8. Stewart et al, reference 3.9. Samuel M, Card RJ, Nelson EB, Brown JE, Vinod PS,

Temple HL, Qu Q and Fu DK: “Polymer-Free Fluid forHydraulic Fracturing,” paper SPE 38622, presented atthe SPE Annual Technical Conference and Exhibition,San Antonio, Texas, USA, October 5-8, 1997.

24 Oilfield Review

Fluid-Loss ControlUnlike polymer-base fracturing fluids, ClearFRAC fluid does not form a filter cakeas a result of leakoff into the formation.Consequently, the fluid-loss rate is essen-tially constant with time (left). Also, unlikepolymer-base fluids where a lower viscosityaqueous phase enters the formation matrix,leaving most of the solids behind, wholeClearFRAC fluid with full viscosity enters thematrix. At formation permeabilities less thanabout 5 mD, it is difficult for an elastic, vis-cous fluid such as ClearFRAC fluid to enterthe pore throats.

As a result, the leakoff rate of ClearFRACfluid, with no fluid-loss additives, is less thanthat of a 20-lbm/1000 gal crosslinked-boratefluid. In high-permeability formations,ClearFRAC fluid is compatible with fluid-loss additives, and significant improvementsin fracturing fluid efficiency are observed.

Proppant TransportThe conventional guideline for proppanttransport is that the viscosity of a fracturingfluid should be at least 100 cp at a shearrate of 100 sec-1 or 50 cp at a shear rate of170 sec-1 (below left). This guideline wasderived from experience with conventionalpolymer-base fracturing fluids whose rheo-logical behavior generally follows thepower-law rheological model. This rule-of-thumb may not apply to ClearFRAC fluid.

Viscoelastic surfactant fluids behave morelike Newtonian fluids, with a flatter appar-ent-viscosity profile across the shear-ratespectrum. This response is different fromthat of most aqueous-base polymer sys-tems. The viscoelastic surfactant fluid isshear thinning, but its rheology is com-pletely reversible and has no permanentdegradation of viscosity when exposed tohigh-shear conditions. Viscoelastic surfac-tant fluids provide ample viscosity forproppant transport in the fracture.

At a given surfactant concentration, the vis-cosity of a ClearFRAC fluid will decreasewith temperature (next page, top). This tem-perature-related thinning can be reduced byincreasing the surfactant concentration oradjusting the salt concentration. Unlike con-ventional polymer systems, however, the vis-cosity does not degrade with time at a giventemperature. Until the fluid is contaminatedwith hydrocarbons or diluted with formationwater, the viscosity will remain stable.

100,000

10,000

1000

100

10

Vis

cosi

ty, c

p

0.01

Shear rate, sec-1

Formation and fracture Tubulars and perforations

0.1 1 10 100 1000

2.5% viscoelastic surfactant

40-lbm/1000 gal hydroxyethyl cellulose

Viscosity profileat 75°F [24°C]

■■Viscosity profile. At high shear rates—above 170 sec-1—the viscosity of the viscoelasticsurfactant fluids is less than that of HEC fluids, giving the viscoelastic surfactant fluidslower friction pressures during pumping.

4% ClearFRAC fluid through brine core20-lbm/1000 gal borate-crosslinked guar through brine core4% ClearFRAC fluid through hydrocarbon core20-lbm/1000 gal borate-crosslinked guar throughhydrocarbon core

■■Leakoff-rate. ClearFRAC fluids have a more predictable leakoff rate than guar-base flu-ids. In these tests, 1-in. Berea sandstone cores were saturated with either brine or hydro-carbons. Guar-base and ClearFRAC fracturing fluids were pumped through the cores.The collected fluid volume was measured relative to the core pore volume to determinethe rate at which each type of fluid leaks off into a formation. The leakoff for ClearFRACfluids is not drastically different in both hydrocarbon-saturated and brine-saturated cores.In contrast, the leakoff of guar fluids is affected by the type of fluid in the formation. Thus,fracturing design with ClearFRAC fluids is less complex because the leakoff is more pre-dictable and less affected by the nature of the formation fluids.

Clear Fracturing 2/21/98 4:53 PM Page 24

Autumn 1997 25

The behavior of surfactant-base fluids dif-fers substantially from that of guar-base flu-ids, and laboratory observation and fieldexperience have suggested that viscoelasticsurfactant fluids with viscosities below theconventional guideline are efficient and arecapable of placing proppant as per design.These observations led to a series of prop-pant-transport tests at STIM-LAB, an inde-pendent laboratory in Duncan, Oklahoma,USA. ClearFRAC fluids were tested at vari-ous flow rates, proppant concentrations andtemperatures to correlate fluid propertieswith proppant transport capability (see“Laboratory Testing,” page 27).

Effective proppant transport was demon-strated in a large-scale fracture simulator atfluid viscosities as low as 30 cp at 100 sec-1.This result was due to the fluid’s elasticityand high viscosity at low-shear rates.

The proppant transport tests proved thatviscoelastic surfactant fluids provide ade-quate proppant transport throughout the 75to 175°F [24 to 79°C] fluid temperaturerange. Even when small amounts of prop-pant settling occurred, the perforation areawas kept clear at all times, and more than90% of the proppant remained in suspen-sion throughout the fluid volume. It isimportant to mention that when a nonvis-cosified brine/sand slurry is pumpedthrough the slot, the sand immediatelydrops and plugs the perforation.

The tests demonstrated that the conven-tional viscosity versus proppant transportguideline derived for polymer-base fluidsmay not apply to ClearFRAC fluids. At ambi-ent temperature, a 42-cp fluid providedexcellent proppant transport (no apparentsand settling), yet at 175°F, a 100-cp fluidallowed some sand to settle. A better guide-line might be to use the minimum concen-trations that were empirically estimatedduring this experimental program (left).

Fluid viscosity, as calculated by the power-law model, does not adequately predict theproppant transport capability of ClearFRACfluids. Further work is needed to determinerheological parameters that better describethe proppant transport behavior of ClearFRAC fluids.

500

400

300

200

100

075

Fluid temperature, °F

3% ammoniumchloride base fluid

100 125 150 175 200

1.0% 1.5% 2.0% 3.0% 4.0%

Viscoelasticsurfactantconcentration

Vis

cosi

ty a

t 100

sec

,-1cp

■■Typical rheological performance. Viscoelastic surfactant fluids generate high viscosityat low temperatures but thin with increasing temperature, making them excellent frac-turing fluids at temperatures below 200°F and especially in low-temperature wells.

Sur

fact

ant c

once

ntra

tion,

%

Fluid temperature, °F

50 75 100 125 150 175 200

5

4

3

2

1

0

■■Proppant transport. With increasing bottomhole temperature, a larger percentage ofClearFRAC surfactant is required for adequate proppant transport.

26 Oilfield Review

Foam Stability and RheologyTo help reduce fluid cost, improve fluid effi-ciency and accelerate cleanup, ClearFRACfluids have been mixed with nitrogen to pro-duce foamed fracturing fluids. Experimentsto quantify the stability and rheologicalbehavior of these systems at elevated tem-peratures and pressures were performed atSchlumberger Cambridge Research (SCR),Cambridge, England, and at STIM-LAB.

Initial experimental work at ambient tem-perature and pressure showed that stablefoams with a half-life—the time at whichhalf of the liquid phase has separated—exceeding 12 hr could be obtained withClearFRAC fluids containing only the surfac-tant itself as the foaming agent.

To determine the suitability of ClearFRACfoams for use in the field, it was necessaryto conduct experiments to quantify their sta-bility and rheological behavior at elevatedtemperature and pressure. The initial testswere conducted with the SCR test appara-tus, which uses a heated syringe pump tofeed the liquid phase into a foam generator.A nitrogen bottle with a digital flow con-troller feeds the gas into the foam generator.The foam passes by a sight glass, allowingobservation of foam texture and bubble sizedistribution. A collection vessel with a glasswall is then filled. The cell is heated by awater bath to control temperature. The foamgenerator and collection cell can be pres-sured up to 1200 psi [8273 kPa] (right).

Four base fluids were tested, all preparedfrom 3% ammonium chloride brine andvarying concentrations of ClearFRAC surfac-tant. The foams were generated at about1100 psi [7583 kPa] and ambient tempera-ture, then pumped into the preheated collec-tion vessel. The foams were evaluated atthree test temperatures: 110°F [43°C], 150°F[66°C] and 190°F [88°C]. Each experimentwas videotaped to show the texture of thefoam, foam during filling of the collectionvessel and the condition of the foam afteraging in the collection vessel. The half-livesof foam prepared from ClearFRAC fluidsrange from greater than 12 hr at fluid tem-peratures less than 150°F to 40 min at 190°F.

At STIM-LAB, the rheological behavior ofClearFRAC foams was tested with base fluidscontaining 0.2 to 3.0% ClearFRAC surfac-tant. At a fluid temperature of 75°F, the rheo-logical behavior of ClearFRAC foams wasevaluated at various concentrations andfoam qualities (next page). In most cases, themaximum viscosity was achieved at a foamquality of approximately 80%. The finest andmost uniform texture was observed at a 70%foam quality. There is little apparent morpho-

logical difference between the foams pre-pared from different brine compositions.

The results of these experiments showedthat ClearFRAC surfactant is an excellentfoaming agent in its own right, and no addi-tional surfactants are necessary to producestable foamed fracturing fluids from 50 to

90% quality up to a temperature of at least175°F. In addition, in the 70 to 80% qualityrange, viscosities far greater than those ofthe base fluids can be attained. This effectreduces the surfactant concentrationrequired to prepare useful ClearFRAC fluids.

These findings have been borne out by asuccessful series of more than 100 experi-mental foamed fracturing treatments in

Liquid outlet

Collectionvessel

Lightsource

Lightsource

Videocamera

Visual cell: 1-in. diameter

Pressure taps, 10 cm apart

Bubble size analyzer

Visual cell: 1-in. diameter

Digital processindicator

Precision pressuretransmitter

Backpressure

Gas outlet

Gas inlet

Transparent sand pack cell:5.00-cm OD4.12-cm ID100-cm length30- to 50-mD permeability

Check valve

Check valveValve

Thermal massflow controller

Antisurge valve

Microprocessorand control equipment

Pump

N2 BottleFoam generator

Precision pressuretransmitter

Safe container

Relief valve

■■SCR foam test equipment.

Autumn 1997 27

Canada and Kansas, USA. Use of nitrogenor carbon dioxide to prepare energized—less than 52% foam quality—or true foamedClearFrac fluids with foam quality greaterthan 52% has been established.

To date, nearly all of the foamed ClearFRACjobs have used ammonium chloride as thebrine salt, partly due to the low foam viscosi-ties at low temperatures observed in the labo-ratory experiments.

In Canada, ammonium nitrate is now usedas the brine component in many wells. Mostof the spent fracturing fluid is disposed of bylandfarming, so brines with chlorides mustbe carefully monitored to remain withinenvironmental limits. Because ammoniumnitrate is a fertilizer, landfarming thesebrines is advantageous to the environment.

In proppant transport tests conducted at STIM-LAB

in Duncan, Oklahoma, the laboratory equipment

consisted of two 50-gal [0.2-m3] mixing tanks and

a heated section of tubing. The fluid was agitated

with mixing blades and then pumped through the

1-in. (0.89-in. ID) tubing, which was heated by gas

burners. Shear was simulated by pumping the

fluid through the tubing. Once the fluid left the tub-

ing, it traveled through a series of four sizes of

pipe ranging from 1/2 in. to 1 in. Pressure drops

were measured over each section of pipe and a

rheogram of shear stress versus shear rate was

generated to calculate the flow behavior index, n,

and consistency coefficient, K, values.1 Changes

in pressure were measured in duplicate on all

pipes and slots to allow a backup measurement in

the event of transducer line plugging by proppant.

Flow rate and pressure data were recorded contin-

uously throughout the test runs, and the data were

processed to plot shear stress versus shear rate.

The capability of a fluid to transport proppant

was assessed visually when the fluid system was

pumped through a slot 1 ft [0.3 m] high by 8 ft

[2.4 m] long with a 5/16-in. [0.8-cm] gap width.

One panel of the slot was transparent to allow

visual inspection. The diameter of the perforation

in the center of the slot entrance was 5/16 in.

Dynamic transport of proppant in the slot was

recorded on videotape. A grid system was placed

on the slot to observe proppant-settling velocities

as the slurry traveled across the 8-ft length. The

fluid was stopped at the end of a run to record the

static settling time. The pumping rate of fluid

through the slot was varied from 1 to 3 gal/min [4

to 11 L/min]. The shear rate at the slot varied from

20 to 60 sec-1, and the shear rate at the perforation

varied from 1300 to 3900 sec-1.

The proppant used in all tests was 20/40 Ottawa

sand, and tests were performed at 4 and 8 ppa.2

The goal of the tests was to determine the mini-

mum concentration of surfactant necessary to pro-

vide adequate proppant transport. The tests were

run at ambient temperature, 150°F and 175°F.

The first test fluid consisted of 0.5% ClearFRAC

surfactant in 3% ammonium chloride at 75°F

[24°C] with a viscosity of 22 cp at 100 sec-1. Before

the proppant was added, the fluid was allowed to

circulate through the system. At 3 gal/min

[11 L/min], significant turbulence was observed

around the perforation opening. The fluid exhibited

a chunky texture upon exiting the perforation,

showing a quick recovery after experiencing the

high-shear environment.

When red dye was added to the fluid to observe

flow behavior, it was immediately apparent that

the fluid in the center of the slot was moving more

quickly than at the edges. Such laminar flow

behavior may explain why lower friction pressures

are observed in the fluid compared to polymer-

base systems when ClearFRAC surfactant is used.

Laboratory Testing

1. In a power-law fluid, viscosity is a function of shear rate:µ=Kγn-1.

2. The abbreviation ppa indicates pounds of proppant addedto 1 gal of fluid.

When sand was added at 4 ppa, the perforation

area was kept clear at all times; however, some

minor amounts of sand settled at the bottom of the

slot. Nevertheless, the settled proppant continued

to be dragged along the bottom of the slot to the

exit perforation. An in-line densitometer showed

that greater than 90% of the proppant was in circu-

lation. Since the area around the perforation open-

ing never plugged, this can be considered as ade-

quate proppant transport. This finding is supported

by field results.

The test was repeated with a doubling of the

ClearFRAC concentration to 1%; the fluid viscosity

increased to 42 cp at 100 sec-1. When dye was

shot into the fluid prior to the addition of proppant,

laminar flow behavior was again apparent; how-

ever, the fluid displayed a much tighter structure.

After exiting the perforation, the fluid instantly

recoiled to form a gelatin-like mass. When sand

was added at 4 ppa, there was no evidence of set-

tling. The fluid appeared to grab the sand grains at

the perforation exit and carry them across the slot.

The same effect was seen at higher concentra-

tions. Pumping was stopped, and sand settling

was not evident after 30 min. Further tests were

conducted with varying concentrations of ClearFRAC

surfactant and at varying temperatures.

Viscoelastic surfactant

concentration

1.5% 0.5% 0.4% 0.2%

550

500

450

400

350

300

250

200

150

100

50

0

Vis

cosi

ty a

t 100

sec

,-1cp

Foam quality, %

50 60 70 80 90

3% ammoniumchloride base fluid

■■Foam rheology. Mixed in a 3% ammonium chloride brine at 75°F, ClearFRAC foamreached maximum viscosity at 80% foam quality.

28 Oilfield Review

Fracturing Fluid FlowbackTo maximize well productivity, it is essentialto maximize fracture cleanup. Polymerresidues that stay in the fracture contributesignificantly to a lowered proppant-packpermeability, leading to a loss in treatmenteffectiveness. Even a small amount of poros-ity loss can cause major loss in retained per-meability.10 Parameters such as types andconcentrations of gelling agent, crosslinker,breaker, reservoir temperature, flowbackrate and shutdown time can affect thedegree of permeability damage. To under-stand the relationship of these parameters tofracture cleanup, quantification of the poly-mer in the flowback fluid is crucial.11

A basic assumption is that a cleaner frac-ture will produce reservoir fluids at a higherrate. But how is fracture cleanup related toproduction? A reasonable analysis is that agiven mass of returned polymer produces agiven volume of pore space available forflow in the proppant pack. Therefore, underequivalent reservoir conditions, a directrelationship should exist between returnedpolymer and production. The conventionalmethod of quantifying cleanup from ahydraulic fracture has been to report load-water recovery. This amount may be affectedby produced formation water, and hencemay be inaccurate. Instead, a colorimetricmethod that involves a phenol-sulfuric acidreaction is used to accurately test thereturned fluids for guar or HEC.12

Analysis of the fracturing fluid returned tothe surface after hydraulic fracturing indi-cates that only 35 to 45% of the guar-basepolymer that is pumped during the treat-ment flows back out of the well during the1- to 10-day flowback period (below). Theremaining polymer stays in the fracture anddecreases well productivity. Thus, there is aneed for a fracturing fluid that can bebrought back to the surface more efficiently.

One method of assessing damage involvescore-flow tests. Leakoff tests were con-ducted on 12-in. [30-cm] Berea sandstonecores at a differential pressure of 1000 psi[6890 kPa]. Two core permeabilities weretested: 230 mD and 1000 mD. The coreswere flooded with 40-lbm/1000 gal borate-crosslinked guar, an 80-lbm/1000 gal HECpolymer and 4% ClearFRAC solution,respectively (next page, top). After theleakoff test, the cores were left in a fluid-losscell, and brine was injected in the oppositedirection. Steady flow rates were reached todetermine the retained permeabilities. In thecore flooded with ClearFRAC fluid, flowbegan immediately as the brine diluted andbroke the fluid. The guar- and HEC-floodedcores had significantly lower permeabilities,even after 24 hr of cleanup. These core-flood tests clearly indicate that polymerresidues can decrease the core permeability.

When wells treated with ClearFRAC fluidare initially flowed back, the tail of theslurry may still have significant viscosity ifit has not yet contacted hydrocarbons orformation water. To help clean up the frac-turing fluid during the initial flowback,especially in underpressured reservoirs,small amounts of some polar organic com-pounds can be added to the tail slurry toaccelerate breakdown. Used in conjunc-tion with proper flowback techniques, thisprocedure can minimize or eliminate earlysand production.13

Shallow Gas Foam FracturesIn September 1996, PanCanadian PetroleumLtd. conducted a field test of the ClearFRACsystem in 10 wells in a shallow gas field inthe Princess East field in Brooks, Alberta,Canada. Like many stimulated wells, thesewells were in poor-quality reservoir areas.Five wells were fractured with this fluid, andfive were fractured with a low-guar systemas a control. Each well had four formationstreated. These low-pressure gas wells werefractured with foamed ClearFRAC fluid; thenitrogen in the foam provided the energy toassist the wells back to production andimprove cleanup. The wells treated withClearFRAC fluid flowed back a greater vol-ume of fracturing fluid—based on fluid vol-umes and colorimetric analysis—than didthe control wells. This improved initial

Pol

ymer

retu

rned

, %

Well

70

60

50

30

10

0

40

20

East Texas Colorado South Texas Central Texas

ETX

04E

TX01

ETX

02E

TX03

ETX

05E

TX10

ETX

08E

TX06

ETX

07E

TX09

CO

-HC

O-A

CO

-FC

O-I

CO

-KC

O-E

CO

-LC

O-C

CO

-JC

O-B

CO

-GC

O-D

STX

01S

TX02

STX

03S

TX04

STX

05S

TX06

STX

07C

TX07

CTX

08C

TX01

CTX

02C

TX03

CTX

04

■■Flowback analysis. An analysis of 150 hydraulic fracturing treatments in the USA showsan average return of 35 to 45% of the polymer pumped. The large volume of polymerresidue remaining in the fracture could impede fluid flow into the well.

10. Barree R and Mukherjee H: “Engineering Criteria forFracture Flowback Procedures,” paper SPE 29600,presented at the SPE Rocky Mountain Regional/Low-Permeability Reservoirs Symposium, Denver, Col-orado, USA, March 19-22, 1995.

11. Armstrong et al, reference 5.12. Pope D, Britt L, Constien V, Anderson A and Leung L:

“Field Study of Guar Removal from Hydraulic Frac-tures,” paper SPE 31094, presented at the SPE For-mation Damage Control Symposium, Lafayette,Louisiana, USA, February 14-15, 1996.

13. Anderson AJ, Ashton PJN, Lang J and Samuelson ML:“Production Enhancement Through Aggressive Flow-back Procedures in the Codell Formation,” paper SPE36468, presented at the SPE Annual Technical Con-ference and Exhibition, Denver, Colorado, USA,October 6-9, 1996.Card RJ, Howard PR and Féraud J-P: “A Novel Tech-nology to Control Proppant Backproduction,” SPEProduction & Facilities 10, no. 4 (November 1995):271-276.Willberg DM, Card RJ, Britt LK, Samuel M, EnglandKW, Cawiezel KE and Krus H: “Determination of theEffect of Formation Water on Fracture Fluid CleanupThrough Field Testing in the East Texas Cotton Val-ley,” paper SPE 38620, presented at the SPE AnnualTechnical Conference and Exhibition, San Antonio,Texas, USA, October 5-8, 1997.

Autumn 1997 29

cleanup and flowback of the fracturing fluidare believed to have contributed to higherproduction rates.

An initial comparison of well productionduring the first few months following thefracture stimulations was impossiblebecause fluid from all zones in each welland from all the wells were commingledinto a common pipeline and not monitoredindividually. From June through August1997, PanCanadian conducted flow-provertests on these wells to quantify the long-term differences in production attributableto the different fracturing fluids.

Before the wells were tested, each wascleaned out with coiled tubing to removeany water and sand fill. The formations inthis part of the Princess East gas field are ofpoor quality; hence, overall production perwell is low. In addition, the wells tend toload with water over time if water is pre-sent, as was the case here. The gas rates foreach group of five wells were plotted dur-ing the test period (below right).

The results were encouraging. A year later,the wells treated with ClearFRAC fluid stillaveraged about 9 to 10% more production.The curves converged slightly after onemonth as certain wells had water loadingand slugging. It is possible that the quality ofthe reservoirs in the specific wells mayaccount for the incremental production;however, statistical averaging of the 20zones in each group of wells in the samearea indicates this cause is unlikely. Assum-ing an average rate difference of 500 scm/d[17 Mscf/D] for the five wells over time—not integrated for specific decline curves buttaken from the cumulative difference duringthe first 51 days of the test shown in the fig-ure—ClearFRAC wells produced a cumula-tive volume of 150,000 m3 [5300 Mscf]more than the control wells. The incremen-tal revenue may be small from these low-producing wells, but the favorable trend hascaused PanCanadian to analyze the poten-tial long-term production improvement inother wells in Brooks and Drumheller,Alberta, with 500 to 600 wells slated forClearFRAC treatments next year.

Acid FracturesSeveral cases involving acid fractures illus-trate the importance of using a nondamag-ing fracturing fluid. Imperial Oil ResourcesLimited’s Norman Wells field is an oil fieldnow under waterflood in the Northwest Ter-ritories, Canada. The field has 325 wells—166 producers and 159 injectors arrangedin a five-spot pattern on six man-made and

three natural islands in the MackenzieRiver. The field’s economic life is expectedto run through the year 2010. At present,some areas of the field are marginally prof-itable, and a solution is needed to improverecovery rate.

The formation is the Kee Scarp, a micriticlimestone with natural fractures, vertical andhorizontal, induced by the tectonics of themountain ranges on both sides of the river.

Because the overburden stress is the leastprincipal stress, any induced fracture will behorizontal, not vertical. Thus, to increaseproduction, each low-rate layer in each wellhas to be treated. The frequency of naturalfractures varies throughout the field.

The formation has up to 18 distinguishablelayers in four separate reef or shoalsequences, with an average reservoir heightof 80 m [262 ft]. The respective layer heights

40-lbmborate-

crosslinkedguar

Retainedpermeability

HEC 80

Retainedpermeability

Greater than 24-hr

cleanup

Immediatecleanup

1000-mD permeability

1 in. 2 in. 4 in. 5 in.

12 in.

12 in.

52% 82% 95%

47%

73%

20%

ClearFRACfluid

Retainedpermeability

1000-psi differential pressure

Greater than 24-hr

cleanup

■■Core-flow test results. Berea sandstone cores treated with ClearFRAC fluid cleaned upmore quickly and regained more permeability than similar cores treated with borate-crosslinked guar or HEC.

1400

1200

1000

800

600

400

200

300

250

200

0

50

100

150

Pro

duct

ion

rate

, m3 /d

Tota

l pro

duct

ion,

thou

sand

m3 /d

June

ClearFRAC fluidControl fluid

0July

ClearFRAC fluidControl fluid

■■ClearFRAC test wells. PanCanadian Petroleum Ltd. tested ten wells in the Princess Eastgas field in Alberta, Canada. Five wells were fractured with ClearFRAC foam fluid andfive with a crosslinked low-guar fluid. After 11 months of commingled production, thewells were tested individually over a three-month period. Analysis of each group of fivewells shows the wells treated with ClearFRAC fluid produce about 10% more than thecontrol wells.

30 Oilfield Review

vary from 2 to 10 m [7 to 33 ft], and not alllayers are in every well. The average truevertical depth is 450 m [1476 ft], and theaverage bottomhole temperature is 25°C[77°F]. The permeability is typically greaterat the top of the reservoir and decreaseswith depth. Average formation permeabilityis 4 mD, but there are few data on verticalpermeability. The higher permeability layersat the top of the Kee Scarp do not requirestimulation because many are already pro-ducing with increasing water cuts withinexpected production rates.

Historically, stimulation of the producersin the Norman Wells field has involvedselective acid treatments, and more recently,foam-diversion acidizing treatments. Neithertechnique has yielded consistent results onthe producers, although the foam-diversionacidizing treatments have had excellentresults on the injectors. A side effect of theearly acid treatments has been a weakeningor loss of wellbore isolation between layers.Thus, to induce a fracture in lower layers, aremedial isolation strategy was needed.

Acid-fracturing treatments were done inthe past with no attempt to direct theinduced fracture into the lower permeabilitylayers. Thus, the fractures went into the topof the zone, with no increase in production.The treatments pumped 50 to 100 m3

[13,000 to 26,000 gal] of fluid and hadexpected fracture lengths of about 50 m[164 ft]. The only net result directlyattributable to these treatments, however,was accelerated water breakthrough. At thispoint, it was determined that the only wayto increase production in the poorer per-forming wells was with a small inducedfracture, 10 to 15 m [33 to 49 ft], directed atthe tighter permeability zones.

The isolation problem was solved, and aninitial fracturing program was proposed withalternating stages of a low-pH guar gel, forfluid leakoff control, with 25 m3 [6600 gal]of 15% hydrochloric acid. This gel systemwas selected because of its compatibilitywith acid at low temperature and its ease ofpreparation, with minimal manpower andequipment. A crosslinked guar was deemedbetter for leakoff control than gelled acidbecause the bottomhole temperature inthese wells was below or at the minimumrecommended for these fluids. One of themain concerns in using a guar-base fluidwas suspected loss of whole polymer intosecondary fissures and fractures and theinability to recover this fluid. Unrecoveredpolymer could remain in these pathways,thereby limiting production.

ClearFRAC fluid became available at thebeginning of the stimulation program in thefield. The fracturing fluid system waschanged because the viscoelastic surfactantfluid is easy to mix, fracturing fluid viscositycan be altered simply by changing surfactantconcentration, crosslinking times are not anissue, and no breakers or extra chemicalsare needed. A viscoelastic surfactant fluidwas most desirable because it would notplug the secondary fissures and fractures.

The first treatment was on a well in thesoutheast end of the field where the wellsintersect the lower, tighter layers of the KeeScarp. The ClearFRAC fluid was mixed in a4% potassium chloride brine. The treatmentcalled for alternating pads of ClearFRACfluid and acid, followed by an overdisplace-ment with brine. A DataFRAC analysis indi-cated the pad volumes of ClearFRAC fluidcould be reduced by 60% from the originaldesign because of a low leakoff rate. It is notknown whether the low leakoff rate was a

local formation phenomenon or a propertyof the viscoelastic surfactant fluid, becausethe original leakoff rates were determinedfrom matrix treatments using gel-base fluidsin other parts of the reservoir.

In total, seven fracturing treatments withstaged ClearFRAC fluid and acid were doneon three wells. The production results so farare encouraging (left). In all three wells, oilflowed to surface within 30% of the recov-ered injected volume with no signs of emul-sion or other fluid incompatibility problems.If a polymer fluid had been used, with littlefluid returned, there would be the potentialfor a large amount of unrecovered polymerto plug the secondary fractures and fissuresand impede production.

Frac PackingHydraulic fracturing has long been consid-ered a stimulation treatment for low-perme-ability formations in hard-rock areas. Inthese treatments, the goal is to create a long,thin fracture with a large surface area. Frac-ture half-lengths, or wings, can reach 500 to1000 ft [150 to 300 m] and have widths intenths of an inch or less.

In contrast, frac packs typically have shortwing lengths of 10 to 50 ft [3 to 15 m] andfracture widths of 1 to 2 in. [3 to 5 cm]. Afrac pack, or STIMPAC treatment, is a frac-ture created using high-viscosity fluid,pumped above fracture pressure, to placesand outside the annulus between the cas-ing and downhole screen and a short dis-tance into the formation. The aim is to cre-ate a high-conductivity sand pack extendinga sufficient distance from the wellbore,beyond any wellbore damage, to create aconduit for the flow of reservoir fluids atlower pressure differentials.14

In late 1991, Dowell and BP Explorationperformed the first true frac pack in the Gulfof Mexico.15 Since then, the incidence offrac packing has skyrocketed, with some600 frac-pack treatments conducted lastyear out of roughly 1200 sand-control com-pletions in the Gulf. The trend is continuingas more operators are considering frac packsin areas that were once gravel packed.

Frac packs are ideal completion tech-niques to bypass formation damage. Manymechanisms can cause wellbore damage—crushing due to stress, fluid and solids inva-sion during drilling, perforating damage,fines migration or precipitation of paraffinor scale. Some of this damage is stressinduced and, therefore, may not be uniform

Well A

Well B

Well C

4

12

22

5

22

34

3

1

2

4

3

5

Oil

Before fracturing

Flow rates, m3/d

After fracturing

Water Oil Water

■■Norman Wells field production data. In the Norman Wells field in the Northwest Territories, Canada, the first wells fractured with ClearFRAC fluid immediately began producing more oil.

Autumn 1997 31

around the wellbore. Only a few of thesedamage mechanisms can be corrected withacid treatments. Because the damage isoften nonuniform, flow into the wellboremay not be equal in all directions. Even aperfect gravel pack will produce throughless than 100% of the available flow area insuch cases.

Radial flow toward the wellbore causeshigher velocities and pressure drops adja-cent to the well (above). These high ratesand pressure drops can cause formationminerals to become mobile and bridge nearthe wellbore, in the perforation tunnels, orin the annular gravel-pack region. This dam-age, along with declining reservoir pressure,causes production from gravel-packed wellsto decline over time. As wells age, operatorsoften open the wells up, increasing draw-down to maintain production; however,such actions also increase fines migrationand damage.

A frac pack addresses these problems bycreating a conduit perpendicular to the mini-mum principal stress, extending beyond anynear-wellbore damage. The flow area intothe wellbore increases, reducing pressuredrop and fluid velocities in the formation,thereby eliminating the causes of finesmigration. Formation fluids establish a some-

what bilinear flow at lower velocities (right).The key property of the fracture is that itmust be highly conductive.

Perhaps the most critical factor to ensurefuture production in a frac pack is keepingthe proppant pack clean. With polymer frac-turing fluids, polymer residue can be signifi-cant on the small fracture face in a frac-packcompletion. At the end of a treatment, asthe formation begins to close, a polymerfluid will have no place to go but into theformation. The fluid leaks off into the forma-tion, and the polymer residue can becomehighly concentrated as a filter cake. Part ofthis problem can be mitigated by usinggreater concentrations of breakers or encap-sulated breakers, which are crushed by theclosing formation and release the breaker inthe appropriate location. These mechanismswork well, but are still imperfect.16 Someestimates put the efficiency of these break-ers at less than 50%. The breaker itselfcould leak off into the formation, bypassingdehydrated polymer.

HEC has often been used as a frac-packfluid because of its small particle size. Thedisadvantage, however, is that too muchpolymer is needed to create the viscosity tofracture high-permeability zones; a viscouspolymer membrane may form and requirehigh volumes of breakers. Borate-crosslinked low-guar systems are used inmany frac packs worldwide, especiallywhere temperatures exceed 200°F. The useof crosslinked fluids cut frac-pack costs andexpanded use of the technique. The fluids,however, can still produce a positive skinfactor in the completion.

Viscoelastic surfactant fluids are particlefree and behave like linear fluids, makingthem ideal for frac-pack operations. Vis-coelastic fluids typically have lower fluidefficiency than normal fracturing fluids,but, in this application, that is a desirablefeature.17 Viscoelastic surfactant fluids havea nearly constant leakoff response to pres-sure. In addition to being solids free andnondamaging, viscoelastic surfactant fluidshave advantages over crosslinked polymersin the way they propagate fractures.

As the viscoelastic surfactant fluid ispumped, much of it leaks off into the forma-tion. The viscosity of the remaining fluid pro-duces drag forces on the rock, initiating thefracture. With a fluid that does not build a fil-ter cake, the rate required to create a fracturecan be calculated from Darcy’s law. Themain factors that control this fracturing rateare the kh product, injected fluid viscosity,fracturing pressure and reservoir pressure.18

Of these, only the injected fluid viscosity can

Fluidflow

Fluidflow

Fluid, solids andperforating damage

Borehole

Radial Flow

■■Stress and radial flow. Damage fromdrilling fluids, drilling-induced stress, perfo-rating damage and fines migration all con-tribute to nonuniform damage in the near-wellbore region. The damaged area shownabove is an area of increased stress. Higherstress on an unconsolidated formation willreduce permeability. Fluid no longer flowsequally in all directions, increasing fluidvelocities and fines migration, which fur-ther damage the near-wellbore region.

Bilinear Flow

Borehole

Fracture wings extending past damage

Fluidflow

Fluidflow

Stress-induceddamage

Fluid, solids andperforating damage

■■Fracture flow. With a frac pack extendingpast the near-wellbore damage, flowbecomes bilinear. Fluid flows almost per-pendicular to the fracture and then downthe clean proppant pack toward the well.Fluid velocities are then lower in the forma-tion for the same or even higher flow rates.The result is a lower drawdown and lessmigration of fines into the wellbore area.

14. DeBonis VM, Rudolph DA and Kennedy RD: “Expe-riences Gained in the Use of Frac-Packs in UltralowBHP Wells, U.S. Gulf of Mexico,” paper SPE 27379,presented at the SPE International Symposium onFormation Damage Control, Lafayette, Louisiana,USA, February 7-10, 1994.Fan Y and Economides MJ: “Fracture Dimensions inFrac&pack Stimulation,” paper SPE 30469, presentedat the SPE Annual Technical Conference and Exhibi-tion, Dallas, Texas, USA, October 22-25, 1995.Ebinger CD: “Frac Pack Technology Still Evolving,”Oil & Gas Journal 93, no. 43 (October 23, 1995): 60-70.

15. McLarty JM and DeBonis V: “Gulf Coast Section SPEProduction Operations Study Group—TechnicalHighlights from a Series of Frac Pack Treatment Sym-posiums,” paper SPE 30471, presented at the SPEAnnual Technical Conference and Exhibition, Dallas,Texas, USA, October 22-25, 1995.

16. Brannon HD and Pulsinelli RJ: “Breaker Concentra-tions Required to Improve the Permeability of Prop-pant Packs Damaged by Concentrated Linear andBorate-Crosslinked Fracturing Fluids,” SPE Produc-tion Engineering 7 no. 4 (November 1992): 338-342.Samuelson ML and Constien VG: “Effects of HighTemperature on Polymer Degradation and Cleanup,”paper SPE 36495, presented at the SPE Annual Tech-nical Conference and Exhibition, Denver, Colorado,USA, October 6-9, 1996.

17. Efficiency is the volume of the fracture divided bythe total volume pumped. With very permeable rock,more fracturing fluid can leak off to the formation,decreasing fluid efficiency.

18. The kh product is the permeability times the heightof the fracture.

32 Oilfield Review

be controlled, unless a fluid-loss agent isused—thereby reducing the kh product bydecreasing the permeability. The fluidleakoff limits the length of fracture that canbe created, but it affects the job cost if largevolumes of fluid leak off. The fracture is cre-ated by a volume of fracturing fluid thatdoes not contain proppant. Once the prop-

pant-laden slurry reaches the tip of the frac-ture, it bridges off and the fracture can nolonger increase in length. Proppant at thetop and bottom similarly prevent heightgrowth. This situation is called a tip screen-out. Further pumping causes the fracture toballoon or widen. In soft rock, the fracturewidth can increase four to six times withthis technique.

An interesting application of ClearFRACfluids is in the treatment of selective com-pletions. In the spring of 1996, PhillipsPetroleum Company performed two fracpacks in a well in the High Island area inthe Gulf of Mexico. The lower gas sand wascompleted with a crosslinked-guar fluid sys-tem, and the selective zone above was com-pleted with ClearFRAC fluid because thefrac pack was to be left behind pipe untilthe lower zone was depleted. Phillips didnot plan to flow back the fracture immedi-ately for cleanup; hence, a polymer fluidsystem might undergo severe dehydrationand damage the formation over time. TheClearFRAC fluid remains in place andbreaks over time as it contacts the formationgas. When the selective zone is produced atsome point in the future, it should flow backas if it had just been fractured.

Fracturing for Water ControlFrac packing with ClearFRAC fluid hasanother advantage—the capability to frac-ture near but not into water zones. Theleakoff rate of the viscoelastic surfactantfluid helps prevent the fracture from grow-ing up or down, possibly contacting nearbywater zones. In certain applications, a typi-cal guar-fluid frac pack might break downthe formation being treated, allowing thefracture to propagate downward directlyinto an underlying water zone, acceleratingwater production. The predictable leakoffcharacteristics of viscoelastic surfactant flu-ids make them the fluids of choice in theseoperations. Once a tip screenout isachieved, the frac will widen rather thanincrease vertically into the water zone.

Coastal Oil & Gas Corp., like many oper-ators in the Gulf of Mexico, has changed itscompletion practices from gravel packing tofrac packing. Frac packing is the comple-tion of choice for high-permeability wellsthat may have potential sand-control prob-lems, drilling-induced formation damage,fines migration problems, perforation dam-age, or other borehole damage. A typicalfrac pack accesses the formation beyondthese damaged zones. Coastal’s frac-packcompletions have held up over time, requir-ing less remedial work than offset gravel-pack completions.

Coastal has been completing wells withfrac packs at a rate of one or two per monthduring 1997 and has had six wells stimu-lated with ClearFRAC fluid. The reasons forusing ClearFRAC fluid instead of a guar fluidinclude the capability of fracturing withouthitting water zones, a shortened cleanuptime and a cleaner fracture.

Density porosity, p.u.

Neutron porosity, p.u.

Density correction, g/cm

0

0

60

60

-0.25 0.253

XX100

Perforated interval XX075 to XX097 ft

Gas-water contact XX110 ft

Gas zones

XX200

■■Fracturing above a water zone. This gas well in the High Island area in the Gulf of Mex-ico was a perfect candidate for a ClearFRAC stimulation. The gas zone from XX072 toXX110 ft sat directly atop a water zone from XX110 to XX160 ft. A frac pack was carefullypumped through the perforated interval from XX075 to XX097 ft to bypass wellbore dam-age yet stay above the gas-water contact.

Autumn 1997 33

In a High Island area well located offshoreLouisiana, Coastal encountered a 36-ft [11-m] gas zone overlying a 50-ft [15-m] waterzone (previous page). Many operatorswould have perforated the upper section ofthe gas zone, performed a gravel-pack com-pletion and flowed the well at reduceddrawdown to stave off water coning.

The well was perforated underbalancedwith tubing-conveyed perforating guns, 13 ft[4 m] above the gas-water contact. Thescreen and packer assembly were thenpicked up and run in the well. The packer

was set, and the pipe was pickled. ADataFRAC analysis was performed to deter-mine the optimum pumping scheduledesign based on closure pressure andleakoff coefficients. Pumping time for thetreatment was about half an hour (top).

The pumping operation was successful(above). Based on bottomhole pressuredata and computer simulation results, thefinal frac pack was estimated to have a 30-ft[9-m] half-length and 1.5-in. [4-cm] width.The well was flowed back at moderate rates

that were increased hourly until reachinganticipated production rates about 12 hrlater. The well flowed back normal amountsof sand and load water and appeared toclean up properly as the gas contacted andbroke down the ClearFRAC fluid. The wellwas then shut in as the rig was skidded tothe next slot, and production resumed sev-eral days later. Initial production rates aver-aged 11 MMscf/D [300,000 scm/d] of gasand 300 BOPD [48 m3/d] with only traceamounts of water.

The frac pack successfully avoided thewater zone. A comparison of the productionto offset wells is difficult and potentiallyinaccurate due to the extremely faultednature of the producing reservoir, however.On this well, the frac-pack treatment wasslightly more costly than for a guar fluid. Thetotal cost to drill and complete the well wasmore than $3 million. Coastal felt the costdifference for the fluid was minor, given theability of the ClearFRAC fluid to ensure aclean frac pack above the water zone.

Looking AheadClearFRAC fluid was commercialized onMay 12, 1997. As of November 1997, morethan 400 ClearFRAC jobs had been pumpedin the USA and Canada. At present, use ofClearFRAC fluid is highest in Canada, fol-lowed by the Gulf of Mexico and then theeastern USA.

The two largest hurdles for ClearFRAC fluidto overcome are its upper temperature limi-tation and cost relative to guar fluids. Theupper temperature limit for ClearFRAC fluidsis 200°F, but research is under way toincrease the temperature range.

The higher chemical cost of ClearFRACsurfactant increases the total cost of a frac-ture treatment by some 5 to 20% on smalltreatments, such as frac packs and jobs withless than a few hundred thousand pounds ofproppant placed. Reducing the cost of thesurfactant will increase the population ofwells where the application of ClearFRACtechnology is economically feasible. Toaddress this cost issue, research is also inprogress to increase the efficiency of thecurrent surfactant and to develop lessexpensive second-generation surfactants.

—KR

9000

6000

15

10

5

Pre

ssur

e, p

si

Pum

p ra

te, b

bl/m

in

Time, min

3000

0 600 610 620 630 640 650

Beginning of tip screenout

Treating pressureBottomhole pressurePump rate

■■Pumping pressure and rate chart. Following the tip screenout in this frac pack operation,the bottomhole pressure began to increase slightly as the treating pressure decreased,indicating the fracture width was expanding. The final fracture had a half-length of 30 ftand a width of 1 in.

Pad

Fluidvolume, gal

FluidPump rate,

bbl/minStage Proppant,

lbProppant

sizeProppant

concentration, ppa

12 4% surfactant 3500 0 20/40 0

1 12 4% surfactant 500 250 20/40 0.5

2 12 4% surfactant 750 1500 20/40 2

3 12 4% surfactant 1000 4000 20/40 4

4 12 4% surfactant 1000 6000 20/40 6

5 12 4% surfactant 250 2000 20/40 8

Flush 12 2% KCI water 3340 0 20/40 0

■■Proppant schedule. This STIMPAC treatment placed 13,750 lbm of 20/40 proppant with7000 gal of 4% ClearFRAC solution followed by 3340 gal of 2% potassium chloride.