34

Advanced Fracturing Fluids Improve Well Economics

The oil and gas industry h y

for hydraulic fracturing. S d

to major improvements in

Today, new additives and

providing innovative solut

more efficient and cost-ef -

Kevin ArmstrongRoger CardReinaldo NavarreteErik NelsonKen NimerickMichael SamuelsonTulsa, Oklahoma, USA

Jim CollinsCalgary, Alberta, Canada

Gilbert DumontMichael Priaro (consultant)Neal WasylyciaPetro-Canada ResourcesCalgary, Alberta, Canada

Gary SlusherEnron Oil & GasCorpus Christi, Texas, USA

For help in preparation of this article, thanJoyce, Ken Nolte and Jon Mitchell, DoweOklahoma, USA; Terry Greene, Dowell, MFrance; Richard Marcinew, Dowell, CalgaCanada; and Mike Morris, Dowell, CorpuTexas, USA.In this article, HIGHSHEAR, CleanFLOWare marks of Schlumberger.

ashis to50 atlitksre

ndarere,nd

is

e-

ngbyurng

r-6.

as witnessed a revolution in fluids technolog

tarting in the mid 1980s, focused research le

the performance of well stimulation fluids.

fluids are extending these capabilities and

ions to nagging problems. The results are

fective treatments for enhancing well produc

ks to Vicll, Tulsa,

ontrouge,ry, Alberta,s Christi,

Hydraulic fracturing is one of the oil and gindustry’s most complex operations. Ttechnique has been applied worldwideincrease well productivity for nearly years.1 Fluids are pumped into a wellpressures and flow rates high enough to spthe rock and create two opposing cracextending up to 1000 ft [305 m] or mofrom either side of the borehole (right). Saor ceramic particulates, called proppant, carried by the fluid to pack the fractukeeping it open once pumping stops apressures decline.

What defines a successful fracture? Itone that:• is created reliably and cost-effectively• provides maximum productivity enhanc

ment• is conductive and stable over time.

This article describes today’s fracturioperations and the pivotal role played the fracturing fluid. Then, it highlights fonew fluid technologies that are improvifracture success and well economics.

1. For more on hydraulic fracturing: Waters AB: “Hydraulic Fracturing—What Is It?,” Jounal of Petroleum Technology 33 (August 1981): 141Veatch RW Jr, Moschovidis ZA and Fast CR: “AnOverview of Hydraulic Fracturing,” in Gidley JL,

Oilfield Review

and PropNETHolditch SA, Nierode DE and Veatch RW Jr (eds):Recent Advances in Hydraulic Fracturing, Monograph12. Richardson, Texas, USA: Society of PetroleumEngineers (1989): 1-38.

35Autumn 1995

nFracturing operation. Modern treatments rely on a variety of process-controlled mixing, blending and pumping equipmentand computer monitoring and recording systems, which permit real-time decision making.

Continuous fluid mixer Proppant blenderPump trucks

permanent logs of job results. Engineerstracking the progress of the treatment usegraphic displays that plot actual pumpingparameters against design values to facilitatereal-time decision making. Production simu-lators compare treatment results with expec-tations, providing valuable feedback fordesign of the next job.

At the heart of this complex process is thefracturing fluid.3 The fluid, usually water-based, is thickened with high molecularweight polymers, such as guar or hydrox-ypropyl guar. It must be chemically stableand sufficiently viscous to suspend the prop-pant while it is sheared and heated in sur-face equipment, well tubulars, perforationsand the fracture. Otherwise, premature set-tling of the proppant occurs, jeopardizingthe treatment. A suite of specially designedchemical additives imparts important prop-erties to the fluid. Crosslinkers join polymerchains for greater thickening, fluid-lossagents reduce the rate of filtration into theformation and breakers act to degrade thepolymer for removal before the well isplaced on production.

The fracture is created by pumping aseries of fluid and proppant stages. The firststage, or pad, initiates and propagates thefracture but does not contain proppant. Sub-sequent stages include proppant in increas-ing concentrations to extend the fractureand ensure its adequate packing.

Fracturing fluid technology has also devel-oped in stages. Early work focused on iden-tifying which polymers worked best andwhat concentrations gave adequate prop-pant transport. Then, research on additivesto fine-tune fluid properties hit high gear.

The Rock, the Mechanics and the FluidHistorically, fracturing has been applied pri-marily to low-permeability—0.1 to 10 md—formations with the goal of producing nar-row, conductive flow paths that penetratedeep into the reservoir. These less restrictivelinear conduits replace radial flow regimesand yield a several-fold production increase(above). For large-scale treatments, as manyas 40 pieces of specialized equipment, witha crew of 50 or more, are required to mix,blend and pump the fluid at more than 50barrels per minute (bbl/min) [8 m3/min].Pumping may last eight hours with1,000,000 gal of fluid and 2,000,000 to4,000,000 lbm of proppant placed in thefracture (next page, left).

Until recently, treatments were performedalmost exclusively on poor producing wells(often to make them economically viable).In the early 1990s, industry focus shifted togood producers and wells with potential forgreater financial return. This, in turn, meantan increased emphasis on stimulating high-permeability formations.

The major constraint on production fromsuch reservoirs is formation damage, fre-quently remedied by matrix acidizing treat-ments. But acidizing has limitations, andfracturing has found an important niche.The objective in highly permeable forma-tions is to create short, wide fractures toreach beyond the damage. This is oftenaccomplished by having the proppantbridge, or screen out, at the end, or tip, ofthe fracture early in the treatment. This “tipscreenout” technique is the opposite of whatis desired in low-permeability formations

36 Oilfield Review

where the tip is ideally the last area to bepacked.

Why the different approach? The answer isfound in the relationship between fracturelength and the permeability contrastbetween the fracture and the formation.Where the contrast is large, as for low-per-meability reservoirs, longer fractures provideproportionally greater productivity. Wherethe contrast is small, as in high-permeabilityformations, greater fracture length providesminimal improvement. Fracture conductivityis, however, directly related to fracturewidth. Using short—about 100-ft [30-m]—and wide fractures can prove beneficial.2

High-permeability formation treatmentsare on a far reduced scale. Only a fewpieces of blending and pumping equipmentare required, and pumping times are typi-cally less than one hour, and often only 15minutes. Fluid is pumped at 15 to 20bbl/min [2.4 to 3.2 m3/min] with a totalvolume of 10,000 to 20,000 gal [37.9 to75.7 m3] and total proppant weight of about100,000 lbm [45,000 kg] (next page, left).This technique has been successful in theNorth Sea, Middle East, Indonesia, Canadaand Alaska, USA.

While fracturing treatments vary widely inscale, each requires the successful integra-tion of many disciplines and technologies,regardless of reservoir type. Rock mechanicsexperiments on cores, specialized injectiontesting and well logs provide data on forma-tion properties. Sophisticated computer soft-ware uses these data, along with fluid andwell parameters, to simulate fracture initia-tion and propagation. These results and eco-nomic criteria define the optimum treatmentdesign. Process-controlled mixing, blendingand high-pressure pumping units executethe treatment. Monitoring and recordingdevices ensure fluid quality and provide

n Flow regimes. Radial flow profile for an unfractured well (top left). Linear flow into and along the fracture replaces radialflow with a more conductive path for hydrocarbons (top right). Inhigh-permeability formations, short fractures are used to reachbeyond the area of matrix damage near the wellbore (bottom).

Low-permeability, linear flow

High-permeability, linear flow

Unfractured, radial flow

4 vehicles

20,000 gal 100,000 lbm

Equipment Fluid Proppant

1,000,000 gal

Propp Propp Propp

Propp Propp Propp

Propp Propp Propp

Propp Propp Propp

Propp Propp Propp

Propp Propp Propp

Propp Propp Propp

Propp Propp Propp

Propp Propp Propp

Propp Propp Propp

Propp Propp Propp

Propp Propp Propp

Propp Propp Propp

Propp Propp Propp

Propp Propp Propp

Propp Propp Propp

Propp Propp Propp

Propp Propp Propp

Propp Propp Propp

Propp Propp Propp

Equipment Fluid Proppant

Propp

Propp

40 vehicles

High-Permeability TreatmentLow-Permeability Treatment

3 million lbm

Much was learned, but what finallyemerged was a huge array of complicatedfluids—difficult to prepare and pump—andan amazing assortment of single-use addi-tives (most had to be custom manufactured)that required expensive material inventories.

In the past ten years, a more productiveresearch direction has emerged. Oil compa-nies, service companies and polymer manu-facturers have concentrated on the basicphysical and chemical mechanisms underly-ing the behavior of fracturing fluids in anattempt to find improved approaches to fluiddesign and use. This initiative has led tomajor advances, including higher-performingpolymers, simpler fluids, multifunctionaladditives and continuous, instead of batch,mixing. These developments have had a sig-nificant, beneficial impact on the industry.

Recent innovations are extending the stateof the art in four areas: • controlling fluid loss to increase

fluid efficiency• extending breaker technology to

improve fracture conductivity• reducing polymer concentration to

improve fracture conductivity• eliminating proppant flowback to

stabilize fractures. Each provides new opportunities for

improving well economics, as described inthe remainder of this article.

Controlling Fluid LossA portion of the fluid pumped during a frac-turing treatment filters into the surroundingpermeable rock matrix.4 This process,referred to as fluid leakoff or fluid loss,occurs at the fracture face. The volume offluid lost does not contribute to extending orwidening the fracture. Fluid efficiency is oneparameter describing the fluid’s ability tocreate the fracture.5 As leakoff increases,efficiency decreases. Excessive fluid loss canjeopardize the treatment, increase pumpingcosts and decrease post-treatment well per-formance.

37Autumn 1995

nZones of leakoff in a fractured formation.In low-permeability reservoirs, theinvaded zone may be small or nonexis-tent, since polymer molecules are toolarge to enter the matrix. In these cases,an external filter cake controls fluid loss.In high-permeability reservoirs, signifi-cant matrix damage can occur becauseof particle penetration.

nScale of fracturing treatments. Equipment, fluid and proppant requirements to createlong fractures in low-permeability formations (left) can be 10 to 50 times more thanthose needed for short fractures in high-permeability reservoirs (right).

2. Hanna B, Ayoub J and Cooper B: “Rewriting the Rulesfor High-Permeability Stimulation,” Oilfield Review 4,no. 4 (October 1992): 18-23.

3. For background on fracturing fluids: Constien VG:“Fracturing Fluid and Proppant Characterization,” inEconomides MJ and Nolte KG (eds): Reservoir Stimu-lation, 2nd ed. Englewood Cliffs, New Jersey, USA:Prentice Hall (1989): 5-1–5-23.

4. Penny GS and Conway MW: “Fluid Leakoff,” in Gidley JL, Holditch SA, Nierode DE and Veatch RW Jr(eds): Recent Advances in Hydraulic Fracturing,Monograph 12. Richardson, Texas, USA: Society ofPetroleum Engineers (1989): 147-176.

5. Fluid efficiency, E, is defined as

in which VF is the fluid volume remaining in the fracture, VP is the total volume pumped and VL is the volume of fluid that has leaked into the formation.

External filter cakeBridging zone

Invaded zone

Uncontaminated formation

Flow

E = VF

VP = VF

VF +VL

Typically, particulates or other fluid addi-tives are used to reduce leakoff by forming afilter cake—termed an external cake—onthe surface of the fracture face. Actingtogether with the polymer chains, the fluid-loss material blocks the pore throats, effec-tively preventing invasion into the rockmatrix (below).

This approach has been applied suc-cessfully for decades to low-permeability

■■Dynamic fluid-losstest apparatus.

Fluid inlet

Fluid outlet

Fluidslot

Leakofffluid

Oil-filled annulus

Side View

Core Sleeve

To differentialpressuretransducers

End View

Core

Sleeve

Coreholder

Oil-filled annulus

■■Dynamic fluid-loss cell and core holder. A specially designed core holder simulates a slot-flow geometry, the one most representative of the actual fracturingprocess. Pressure and leakoff measurements provide the data necessary to evaluate the depth of invasion and the impact of various additives on fluid loss.

Measuring Dynamic Fluid Loss in the Laboratory

38 Oilfield Review

Dynamic fluid-loss measurements were made in

the Dowell laboratory in Tulsa, Oklahoma, USA

using special fluid-loss cells with a slot-flow

geometry and a porous test surface on one of the

slot walls (above). Cylindrical cores of the same

type and dimensions as for static tests are used

to allow direct comparison with static fluid-loss

results. Cells are constructed from stainless steel

for operation to 3500 psi and 350°F [177°C]. The

inlet design ensures fully developed flow over the

test section. A backpressure regulator and a heat

exchanger, which cools the filtrate, prevent evap-

oration of the filtrate during operations above the

ambient boiling point of the fluid. As many as

three cells can be used simultaneously for testing

cores of differing permeabilities.

A special fluid-loss simulator was designed to

prepare the fluids under dynamic conditions, sub-

ject them to shear and temperature histories and

then measure fluid loss (right).

The two components of the apparatus are a

shear history simulator and a fracture simulator.

The first uses a static mixer and 800 ft [244 m] of

small-diameter tubing to simulate preparation

and shearing of the fluid in the well tubulars.

The second subjects the fluid to shear and temper-

ature conditions of the fracture, using two large,

floating-piston accumulators and coils of tubing

immersed in a temperature-controlled bath.

A computer-controlled valving arrangement

ensures that the fluid always travels in the same

direction across the core face.

■■Forces acting on a particle. A particle,such as a fluid-loss additive, flowinginside a fracture is subjected to shear, Fxand drag, FY, where vx is the velocity inthe flow direction as a function of the dis-tance, y, from the fracture face. The ratioof shear to drag is directly proportional tothe leakoff rate and inversely propor-tional to the particle size and the shearrate at the wall. High initial leakoff andoptimized particles can help ensure thatfluid-loss additives reach and remain onthe fracture face.

■■Shear rate history of a rock segment. As the tip of the fracture passes a particularlocation, in this case a point 50 ft [15 m] from the wellbore, the maximum shear rateoccurs. As the treatment progresses and the fracture widens at this location, the shearrate falls off rapidly initially and then more slowly later. The laboratory approximationand calculated curve show good agreement.

y

vx

Fy

Fx

39Autumn 1995

(< 0.1 md) formations in which polymerand particulate sizes exceed those of thepore throats. In high-permeability reservoirs,however, fluid constituents may penetrateinto the matrix, forming a damaging internalfilter cake. This behavior has promptedmechanistic studies to determine the impacton fracturing treatment performance.6

Classic fluid-loss theory assumes a two-stage, static—or nonflowing—process.7 Asthe fracture propagates and fresh formationsurfaces are exposed, an initial loss of fluid,called spurt, occurs until an external filtercake is deposited. Once spurt ceases, pres-sure drop through the filter cake controlsfurther leakoff. For years, researchers havedeveloped fluid-loss control additives undernonflowing conditions based on this theory.

The conventional assumptions, however,neglect critical factors found under actualdynamic—or flowing—conditions presentduring fracturing, including the effects ofshear stress on both external and internal fil-ter cakes and how fluid-loss additives movetoward the fracture face. In high-permeabil-ity formations, with an internal filter cakepresent, most of the resistance to leakoffoccurs inside the rock, leaving the externalcake subject to erosion by the fluid.

Analysis of fluid loss under dynamic condi-tions relates external cake thickness to theyield stress of the cake at the fluid interfaceand the shear stress exerted on the cake bythe fluid.8 These, in turn, depend on thephysical properties of the cake and the rheo-logical properties of, and shear rate inducedin, the fluid.9 Whether an external filter cakeforms, grows, remains stable or erodesdepends on the way these parameters varyand interact over time and spatial orientation.

Similarly, the effectiveness of additives tocontrol fluid loss depends on two factors:their ability to reach the fracture face quicklyand their ability to remain there. The formeris governed by the drag force exerted on theparticles and the latter by the shear forceexerted on them (right). The larger the ratioof drag to shear, the greater the chance thatthe particles will remain on the surface. Agreater leakoff flux to the wall, smaller parti-cle dimensions and a lower shear rate favorsticking. Promoting higher leakoff for betteradditive placement seems directly at oddswith controlling fluid loss! However, in prac-tice, higher initial leakoff can yield greateroverall fluid efficiency.

To confirm the controlling mechanisms,dynamic fluid-loss tests were conducted

using a slot-flow geometry, determined to bethe simplest representation of what occurs ina fracture. To completely describe the pro-cess, computer-controlled equipment wasconstructed to prepare and test fluids underdynamic conditions, subjecting them to thetemperature and shear histories found in afracture (see “Measuring Dynamic Fluid Lossin the Laboratory,” previous page). Cores ofvarious lengths were used in the tests to sim-ulate a fracture segment at a fixed distancefrom the wellbore.10 As the fracture tippasses a specific point, spurt occurs and theshear rate reaches a maximum (above ).Then, as the fracture widens, the shear stressdecreases. In the test apparatus, this is simu-lated by decreasing the flow rate with time.

6. Navarrete RC, Cawiezel KE and Constien VG:“Dynamic Fluid Loss in Hydraulic Fracturing UnderRealistic Shear Conditions in High-PermeabilityRocks,” paper SPE 28529, presented at the 69th SPEAnnual Technical Conference and Exhibition, NewOrleans, Louisiana, USA, September 25-28, 1994.

7. Howard GC and Fast CR: Hydraulic Fracturing,Monograph 2. Richardson, Texas, USA: Society ofPetroleum Engineers, 1989.

8. Prud’homme RK and Wang JK: “Filter-Cake Forma-tion of Fracturing Fluids,” paper SPE 25207, pre-sented at the SPE International Symposium on Oil-field Chemistry, New Orleans, Louisiana, USA,March 2-5, 1993.

9. Shear rate is proportional to shear stress, the exactrelationship depending on the rheological modelused for the fluid.

10. Navarrete RC and Mitchell JP: “Fluid-Loss Controlfor High-Permeability Rocks in Hydraulic FracturingUnder Realistic Shear Conditions,” paper SPE29504, presented at the SPE Production OperationsSymposium, Oklahoma City, Oklahoma, USA, April2-4, 1995.

300

She

ar r

ate,

sec

-1

250

200

150

100

50

0 20 60 16040 14012010080

Time, min

350

CalculatedLaboratory approximation

Fracture height = 300 ftFracture length = 540 ftPump rate = 40 bbl/minPump time = 145 min

Pressure sensors along the core monitor theprogress of the polymer front.

Laboratory tests show that, for compara-ble fluids and rocks with permeabilities ofup to 50 md, fluid loss is greater underdynamic conditions than static conditions(left). Further, examining the impact of shearstress and permeability on the magnitude offluid loss and the effectiveness of leakoff-control additives in high-permeability for-mations led to five key conclusions.11

First, high shear rates can prevent the for-mation of an external filter cake and resultin higher than expected spurt (below, left).Second, an internal filter cake controls fluidloss, especially near the fracture tip. Third,the effectiveness of fluid-loss additivesincreases with formation permeability anddecreases with shear rate and fluid viscosity.Fourth, reducing fluid loss means reducingspurt, particularly under high shear condi-tions and in high-permeability formations.Finally, at high shear rates with no externalfilter cake, efficient spurt control must beachieved by plugging the pore throats at thesurface of the rock.

The effect of shear depends on the type offluid and the formation permeability. Typi-cally, above a threshold shear level, noexternal filter cake is formed. The magni-tude of fluid loss is dependent on the type ofpolymer and whether it is crosslinked. If thepermeability is high enough and the fluidstructure degrades with shear, polymer maybe able to penetrate the rock matrix.

Dynamic tests revealed that commonlyused additives were less effective in control-ling fluid loss than static tests had previouslyindicated. Also, a direct link between fluidefficiency and shear rate was demonstrated.The higher the fraction of fluid lost underhigh shear early in the treatment, the higherthe total leakoff volume and the lower theefficiency. Spurt has a dominant effect onefficiency and the volume of fluid pumped,particularly for the short pumping timesencountered in fracturing high-permeabilityformations. If spurt is not controlled quicklyand effectively during the high-shear period,fluid efficiencies drop dramatically.

These observations prompted researchersto develop a superior additive system that:• controls spurt under high shear rates in

high-permeability formations• minimizes the influence of permeability

on leakoff12

• limits the invasion of polymer into the matrix

• reduces the overall amount of polymerpumped into the fracture.

40 Oilfield Review

6

Flui

d-lo

ss v

ol, m

l

5

4

3

2

1

0 2 4 6 16 181412108Time, min

Dynamic, 190 sec–1 shear rate

Temperature = 150°FPressure drop = 1000 psi

Dynamic, 40 sec–1 shear rate

Static

10 md

62 md

20 40 60 80 100 120 1400

2

4

6

8

10

12

14

16

She

ar r

ate,

sec

-1

0

50

100

150

250

200

300

350

Time, min

Flui

d-lo

ss v

ol, m

l

Shear rate

nDynamic versus static fluid-loss test results. For this crosslinked guar system, dynamicfluid-loss values are higher than static measurements. As shear rate increases, leakoffincreases. Here, rock permeability is 0.5 md.

nEffect of shear rate on dynamic fluid loss. Fluid loss increases with increasing perme-ability, illustrating the importance of controlling leakoff in high-permeability forma-tions. Maintaining high fluid efficiency is critical to creating a cost-effective fracturewith minimal formation damage.

Evaluation of several additive types led toa combination of particulate materials, theHIGHSHEAR system, that achieves theabove goals. One agent moves rapidly tothe fracture face during the early stage offluid loss when leakoff flux is high. This typeof particle seals a major portion of theexposed surface quickly and adheressecurely to the surface, resisting high shearforces. The second material plugs theremaining gaps in the developing filter cakeas the shear rate drops, significantly reduc-ing filtrate losses and sealing the surface sothat polymer particles cannot penetrate thematrix. Laboratory tests show a 25 to 75%reduction in spurt compared to the bestavailable products today (right). Fluid effi-ciency improves, meaning less fluid topump, less fluid to break and easiercleanup. Since this process is aided by hav-ing a less viscous fluid—one that promoteshigh initial leakoff—there is a further oppor-tunity for cost-savings and improvedcleanup by reducing the polymer concen-tration in the fluid (page 46).

One concern with fluid-loss additives hasbeen the possibility of their presence reduc-ing proppant-pack conductivity. A key ques-tion is: Does the reduction in matrix dam-age brought about by additives outweigh thepotential damage they may inflict within thefracture itself? As is well-known, the amountof conductivity damage is highly dependenton the type of fluid used.13 Flowback testswith different fluids and leakoff controladditives confirm that particulate fluid-lossadditives do limit matrix damage by mini-mizing fluid invasion. Conductivity testsshow that these agents are actually lessdamaging to the proppant pack than previ-ously thought. In most fracturing fluids, theirimpact on conductivity is minimal whenused in low concentrations.14

With these encouraging laboratory results,the next step was to test whether the HIGH-SHEAR system could reduce treatment costsand improve well productivity better thanconventional products. Extensive field trialsin Canada show an overall 15 to 20% costsavings (see “Canadian Treatments Demon-strate Cost Savings and Productivity Gains,”next page). Productivity improvements aver-aged 460% compared to 260% on offsetwells without the additive. Additional welltests are planned in areas such as the USGulf Coast where a variety of fluids are usedand larger-scale, high-permeability fractur-ing treatments are performed.

Understanding and Improving FractureConductivitySimply creating a fracture does not guaran-tee better well performance. The fracturemust provide a conductive flow path for for-mation fluids. For decades, poor treatmentresults were blamed on insufficient fracturelength or inadequate proppant transport andplacement. Studies in the late 1980srevealed that substantial fracture damage,and resulting impairment to flow, could becaused by polymer residue blocking thepore spaces between proppant particles.15

Once the proppant pack is placed andpumping stops, fluid filtrate leaks off into therock matrix and the pressure declines. Thefluid remaining in the fracture must then beflowed back to allow production of hydro-carbons. Termed cleanup, this process is crit-ical to the success of the treatment. One wayto aid cleanup is to ensure that the polymerresidue has been reduced to a minimum. Asecond is to use as little polymer as possibleto start with. The next two sections look atrecent developments in both areas.

Extending Breaker TechnologyAt the end of a treatment, the fluid left in thefracture has been partially dehydrated due tofiltrate loss. The effective polymer concentra-tion can be an order of magnitude higherthan that originally pumped, reaching 300to 600 lbm/1000 gal [36 to 72 g/cm3]. If thepolymer stays intact, an ultrahigh viscosity,gelled mass results that blocks the porespace and cannot easily be flowed back intothe well.

To prevent this, the polymer is attacked byfluid breakers—oxidizers or enzymes thatsever the polymer chain at its weakestpoints, degrading it into smaller, moremobile fragments. This reduces the viscosity

of the residual fluid, thereby allowing moreefficient cleanup. Breakers are used at reser-voir temperatures below about 325°F[163°C].16 If breaking is insufficient, con-centrated polymer remains in the proppantpack, reducing the conductivity and treat-ment effectiveness.

Historically, active chemical breakerswere dissolved in the fluid during surfacemixing. As a result, the fluid was beingattacked even as it was being pumped. Carehad to be taken to use a sufficiently lowbreaker concentration. Otherwise, viscositywould decrease too quickly, and proppantwould settle. With this low breaker concen-tration, only partial degradation of the poly-mer occurred. The result was impaired frac-ture conductivity.

Research during the past decade hasfound ways to increase breaking efficiency.

41Autumn 1995

Time, min

010 20 30 40 50

4

8

12

16

Flui

d-lo

ss v

ol, m

l

15 lbm/1000 gal HIGHSHEAR

25 lbm/1000 gal best available

No fluid-loss additive

nEffect of additives on fluid loss. The HIGHSHEAR additive system provides substantialimprovement in controlling leakoff compared to the best materials available today.

11. Navarrete et al, reference 6.12. In high-permeability formations, permeability can

vary widely. Using an average value in job designcalculations can present execution problems duringthe treatment. An additive that dampens the effect ofpermeability variation can be beneficial.

13. Brannon HD and Pulsinelli RJ: “Evaluation of theBreaker Concentrations Required to Improve the Per-meability of Proppant Packs Damaged by HydraulicFracturing Fluids,” paper SPE 19402, presented at theSPE Formation Damage Control Symposium,Lafayette, Louisiana, USA, February 22-23, 1990.

14. In instances where the impact is significant, the dam-age can be countered by lowering the polymer con-centration since fluids of lower viscosity can be usedduring the treatment.

15. Hawkins GW: “Laboratory Study of Proppant-PackPermeability Reduction Caused by Fracturing FluidsConcentrated During Closure,” paper SPE 18261,presented at the 63rd SPE Annual Technical Confer-ence and Exhibition, Houston, Texas, USA, October2-5, 1988.

16. Above this level, the fluid breaks with time due to thethermal degradation of the guar molecules.

More than 19 operational trials of the HIGHSHEAR

fluid-loss control additive have been conducted

in Alberta, Canada in clean sandstone reservoirs

(right). Here, permeabilities vary from 50 to

150 md and porosities are 20 to 24%. Well depths

typically range from 2300 to 2500 ft [700 to

760 m] with bottomhole temperatures of 85 to

95°F [30 to 35°C].

Earlier wells fractured in the field in 1993, as

part of a refracturing program, yielded proppant

placement efficiencies as low as 70%. Because of

the relatively high permeability, its variability

within the reservoir and high-concentration prop-

pant schedules, an error of 10% in the fluid-loss

design calculation could lead to 30% of the prop-

pant not being placed. In addition to verification

of laboratory test results with the additive, objec-

tives for a new refracturing campaign were:

• reduced treatment cost

• improved job design and execution

• decreased volume of polymer pumped

• reduced well cleanup times

• increased production rates.

To achieve these goals, the strategy was to

replace the 30 lbm/1000 gal borate-crosslinked

fracturing fluid normally used with a reduced-

polymer, 22 lbm/1000 gal system (page 46) to

promote higher fracture conductivity, to speed

cleanup (due to less damage from residual

polymer) and to lower wellsite costs. With its

inherently lower viscosity, the new fluid meant

higher leakoff. To counteract this, the HIGHSHEAR

additive was used to limit spurt, seal the

fracture face and smooth out the effects of

permeability variations.

Pad volumes of 2000 to 2500 gal [7.6 to 9.5m3]

were used with total slurry volumes of 4000 to

6000 gal [15.1 to 22.7 m3] in four stages. Typical

rates were 15.7 bbl/min [2.5 m3/min]. Maximum

sand concentrations ranged from 16.5 to more

than 18.5 ppa.

The field trials were highly successful. Test

wells showed consistent proppant placements of

95 to 100%. Over time, as experience with the

system grew, pad volumes were maintained and

42 Oilfield Review

A B C D E F G H I J K L M N O P Q R S

Pro

duct

ion

rate

, m3 /

day

0

5

10

15

20

25

Wells

30

Rate before refrac

Rate after refrac

Pro

duct

ion

rate

, m3 /

day

20

18

16

14

12

10

8

6

4

2

0Well

1Well

2Well

3Well

4Well

5

Rate before refrac

Rate after refrac

■■Field trial of fluid-loss additive and low-guar fluid in Alberta, Canada.

■■Comparison of pro-duction rates beforeand after treatment.Five wells were refrac-tured in 1993 usingconventional fluidstechnology, resulting in an average 260%improvement in pro-duction rate (top). Nineteen wells refrac-tured in 1995 using theHIGHSHEAR fluid-lossadditive and the low-guar fluid showed anaverage 460% improve-ment (bottom).

Canadian Treatments Demonstrate Cost Savings and Productivity Gains

higher sand concentrations were achieved as a

result of increasing placement success. Less

fluid pumped meant less fluid to break. Cleanup

times were slashed from several days to one or

two days. Post-treatment production increased

from 260% in the 1993 campaign to 460% (previ-

ous page, middle). Overall treatment costs were

15 to 20% lower.

The changeover to the lower polymer loading

was successful due to the fluid-loss additive.

In separate trials without the material present,

proppant placement efficiency was erratic.

According to Gilbert Dumont, production engi-

neer for Petro-CanadaResources, “Fluids are

coming back crystal-clear with this system,

something we haven’t seen before. The wells are

cleaning up quicker, in one to two days rather

than three to ten days. This indicates we are get-

ting faster and more complete breaking of the

system. We believe our conductivities are higher

as supported by increased well productivity.”

Michael Priaro, engineering consultant for

Petro-Canada Resources adds, “Our success rate

is greater because of more consistent placement.

We’ve been able to obtain higher proppant con-

centrations with cleaner fluid. The cleanup has

been better and quicker and well productivity

after stimulation is definitely higher, averaging

4.6-fold improvement, with some wells showing

almost 7-fold improvement. Ten-fold production

improvements may be possible with further

advances in treatment design.”

From Neal Wasylycia, production engineer for

Petro-Canada Resources, “Comparisons of

results are difficult with slant, deviated and

directional wells with varied pays being frac-

tured. We definitely know that our results on ver-

tical wells are dramatically better, and 100%

fracture placement can be repeated with confi-

dence. Our overall refracturing costs are reduced

due to lower fluid cost, quicker cleanup and

reduced fluid disposal cost. The successful

placement of fractures in deviated and slant wells

represents our next challenge, as this is where

our next large refracturing opportunity exists.”

By encapsulating breakers so they do notinteract with the fluid until released by rup-ture of the protective coating with time orstress, higher concentrations can be used,degrading more of the polymer. This hasbeen a major innovation for improvingcleanup and proppant-pack conductivityand is a common industry practice today.17

The mechanisms of proppant-pack dam-age have been the subject of much studyand much controversy. Recent findings haveled to a more consistent and reproducibleapproach to testing proppant-pack damage.By understanding the controlling processes,scientists have been able to develop new-generation additives to improve fluidcleanup and fracture conductivity.

A keystone of the investigation was devel-oping a relationship between retained per-meability and porosity reduction in prop-pant packs. A graph of this correlationshows that, for example, a 10% change inporosity (from 30 to 27%) can reduce per-meability by 35%. Laboratory tests verify thetheoretical relationship (above, right). Small,random blockages by residual polymer or athin, concentrated polymer filter cake at the

fracture face lead to large reductions inretained permeability (below).

Another part of the study involved identi-fying the parameters and test procedureswith the most significant effects on mea-sured conductivity. Tests were performed atthe Dowell laboratory in Tulsa, Oklahoma,USA and two independent laboratories. The

43Autumn 1995

■■Retained permeability as a function ofporosity. The theoretical correlation relat-ing the two properties of the proppantpack has been confirmed in laboratorytesting. Small decreases in porosity canlead to significant reductions in retainedpermeability, resulting in decreasedhydrocarbon production.

■■Pore blockages.Retained perme-ability can bereduced by anexternal filter cake (upper left). The larger the filtercake, the lower the permeability.Retained perme-ability can also bereduced by random,undegraded poly-mer fragments fill-ing the void spacesbetween proppantparticles. As theresidue increases,the permeabilitydecreases (upperright, lower left, lowerright). Even smallincreases in residuecan have a dra-matic effect and in both cases, the permeabilitydecrease can dra-matically impactproduction rates.

Fractional reduction in porosity

Frac

tiona

l ret

aine

d pe

rmea

bilit

y

0.2 0.4 0.6 0.8 10

0.2

0.4

0.6

0.8

1

Theoretical model1 Layer2 Layer, test 12 Layer, test 2

50% retained permeability 25% retained permeability

75% by filter cake

75% retained permeability50%25%

17. Gulbis J, King MT, Hawkins GW and Brannon HD:“Encapsulated Breaker for Aqueous Polymeric Flu-ids,” paper SPE 19433, presented at the SPE Forma-tion Damage Control Symposium, Lafayette,Louisiana, USA, February 22-23, 1990.

results confirm that polymer concentration,for example, is critical. Its influence is non-linear (below).

The relationships of other parameters totest results and reproducibility were alsoquantified, leading to a refined testing pro-cedure. Testing by these laboratories usingthe new guidelines demonstrates that reli-able results can be achieved, provided testconditions are closely controlled. Testingwithin the same laboratory is expected tovary by 15 to 20% while lab-to-lab varia-tions have a 25 to 30% relative deviation(right ) . Before uniform procedures, testresults often deviated by 100% or more.

The thorough investigation of conductivitytesting provided a road map of how toimprove breaker effectiveness. The problemwas attacked on two fronts: selection of anoptimal breaker for a given temperaturerange and development of an additive toassist in removal of the degraded polymer.

Each type and composition of breaker areeffective over a certain limited temperaturerange, based on performance and cost crite-ria. This requires tailoring of chemical prop-erties and encapsulation technology so thatthe breaker remains active in the proppantpack long enough to do its job. A suite ofmaterials is needed to cover the range oftemperatures that are encountered in frac-turing operations.

But a breaker by itself is often not enough.Conductivity testing shows that residualfragments left after the primary polymerstructure has been degraded can still causesignificant pore blockage (below). Worse,under certain conditions, the fragments cancoagulate—or bind together into a viscousmass—and reduce conductivity further.Sometimes, adding more breaker can aggra-vate coagulation.

To avoid this, a special blend of anticoag-ulants was developed. This blend, theCleanFLOW additive, works synergisticallywith the breaker to reduce the size of thepolymer fragments and prevent their ten-dency to coagulate. The dispersive action ofthe additive increases mobility and available

flow paths. Pore blockage is reduced andproppant-pack permeability increases.

Laboratory testing in standard conductiv-ity cells shows that the breaker plus addi-tive system outperforms breaker alone,yielding almost 40% higher proppant-packpermeabilities at low-end additive concen-trations (next page, top). Retained perme-abilities improve to as much as 90% atincreased concentrations.

In field testing, treatments with theCleanFLOW additive showed higher levelsof polymer returns during flowback thanprevious treatments on offset wells. Higherreturns are a direct measure of improvedproppant-pack cleanup. Over a given flow-back period in a well in western Wyoming,USA, 62% of the polymer was returnedwhen the additive was used with breaker,compared to 31% for an offset well with thebreaker alone. At the end of the test phase,the well treated with the CleanFLOW addi-tive was still returning polymer, while theoffset well was not (next page, bottom).

Net present value (NPV) calculations onvarious formation types show that conduc-tivity-related production increases realizedfrom the use of the additive can significantlyimprove well economics.

44 Oilfield Review

0 100 200 300Polymer concentration, lbm/1000 gal

Ret

aine

d pe

rmea

bilit

y, %

100

80

60

40

20

0

Borate crosslinked guarTemperature = 160°F

Lab 1 Lab 2 Lab 3

Test 1Test 2

Ret

aine

d pe

rmea

bilit

y, %

35

30

25

20

15

10

5

0

+ breaker

+ CleanFLOWand breaker

nEffect of polymer concentration onretained permeability. As the residualpolymer concentration remaining in theproppant pack increases, the permeabilitydecreases, emphasizing the need for com-plete polymer degradation and removal.

nEnhanced proppant-pack cleanup. Use of a breaker alone (top)results in polymer fragments that may be difficult to removefrom the fracture. The CleanFLOW additive system works withthe breaker (bottom) to prevent fragment coagulation, facilitatecleanup and improve fracture conductivity.

nInterlaboratory comparison of conduc-tivity results. Major improvements inreproducibility stem from consistency intesting methods and close control overtest conditions, removing a major road-block in correlating data between labs.

Reducing Polymer ConcentrationToday, over 70% of the fracturing treatmentsconducted use guar- or hydroxypropyl guar-based fluids.18 The rheology of such fluidshas been studied for years. When added towater, guar molecules hydrate and swell,increasing in diameter and length. Thehydrated strands overlap and hinder motion,giving rise to an increased viscosity of thesolution. For adequate proppant transportand placement, a viscosity of 100 centipoise(cp) at a shear rate of 100 sec-1 is generallyaccepted as a minimum guideline.

To minimize fluid leakoff and counteractthe inherent thinning of guar systems at ele-

vated temperatures, however, higher viscosi-ties have historically been used. In non-crosslinked (linear) systems, this meansadding more polymer, resulting in polymerconcentrations of 40 lbm/1000 gal [4.8g/cm3] or higher. This approach is expensiveand often results in fluid mixing and han-dling difficulties. Fracture conductivity canbe impaired, since more polymer must bebroken and produced back. Crosslinkinghas become a common means of enhancingviscosity at lower polymer levels. Whileeffective in building viscosity, this practicecan lead to complicated, hard-to-breakstructures. Using crosslinkers, polymer lev-

els can be reduced to about 25 lbm/1000gal [3 g/cm3]. The goal of recent researchhas been to formulate a reliable fluid usingeven less polymer to reduce cost andimprove fracture conductivity.

Various metal ions, including titaniumand zirconium, have been used for decadesas crosslinking agents. In recent years,boron [as B(OH)4

-] has grown in popularityand is by far the most common elementtoday. Various boron salts and compounds

45Autumn 1995

0200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400

40

35

30

25

20

15

10

5

Cumulative flowback volume, bbl

Pol

ymer

con

cent

ratio

n, lb

m/1

000

gal

CleanFLOW system and breaker

Breaker only

62% polymer return

31% polymer return

CleanFLOWsystem

2 gal/1000 gal

No breaker

Ret

aine

d pe

rmea

bilit

y, %

Breaker only0

10

20

30

40

CleanFLOWsystem

4 gal/1000 gal

Temperature = 225°F

nPolymer flow-back comparisons.For a test well inwestern Wyoming,USA, 62% of thepolymer wasreturned duringflowback. For anoffset well withbreaker alone, dur-ing a similar flow-back time, only31% of the polymerwas returned.

nConductivitycomparison. Thenew additive sys-tem provides sub-stantially greaterretained perme-abilities comparedwith no breaker orwith use of abreaker alone, con-firming that theadditive enhancesremoval of residualpolymer and limitspore blockages.

18. Gulbis J: “Fracturing Fluid Chemistry,” in Econo-mides MJ and Nolte KG (eds): Reservoir Stimulation,2nd ed. Englewood Cliffs, New Jersey, USA: PrenticeHall (1989): 4-1–4-14.

can be used (below). If the pH19 of the fluidis above 8, crosslinking occurs almostinstantaneously when borate ions are addedto hydrated guar (right). A unique feature ofthe resulting fluid is that the crosslinking isreversible. As temperature increases, the pHfalls and the solution thins because there isless borate ion available in solution. Viscos-ity recovers as temperature decreases andthe pH rises.

If crosslinking is too rapid, high frictionpressure in wellbore tubulars and shearthinning may occur during pumping. Thismay not be a severe problem in shallow,low-temperature wells where pumpingtimes are short, but is of major concern fordeep, higher temperature treatments requir-

ing extended pumping times. For theseapplications, it is necessary to delay thecrosslinking process until the fluid has trav-eled through most, or all, of the well tubu-lars. This can be accomplished by varyingthe chemistry and additives in the fluid or byencapsulating the crosslinker, which permitstimed release of the active material. Delayedcrosslinking reduces friction pressure andhorsepower requirements and provides forhigher injection rates.

For a stable and reliable borate crosslinkedfluid to be prepared at a given temperature,sufficient polymer chains must be present forentanglement to occur. For most guars, solu-tions containing less than 20 lbm/1000 galcannot effectively be crosslinked. In addition

to having sufficient polymer in solution, twoother criteria must be satisfied: the chainsneed to have enough active crosslink sitesand the proper number of borate ions mustbe present to build a network structure.Careful balance between the two has to bemaintained to produce a stable system.

In the past, borate crosslinked fluids at thelower threshold of 20 lbm/1000 gal guarconcentration have been formulated in thelaboratory and even tested in the field.Operational experience, however, showedthem to be unreliable and overly sensitive tosmall variations in fluid chemistry.

How can the threshold polymer level nec-essary for crosslinking be reduced success-fully? Based on earlier work with zirconatesystems, scientists evaluated what types ofmaterials associate with guar molecules toincrease their solubility. After studying thebehavior of a variety of materials, researchersidentified a combination of chemicals thatpermits stable borate crosslinking at polymerconcentrations as low as 15 lbm/1000 galand reliable field formulations to be mixed inthe 15- to 20 lbm/1000 gal range (next page,bottom).

These systems, referred to as low-guar flu-ids, can be used at fluid temperatures up to175°F [80°C]. They exhibit viscosities nor-mally measured for borate crosslinked fluidswith polymer concentrations 5 to 10lbm/1000 gal higher.

Proppant suspension properties are excel-lent, allowing use of proppant concentra-tions as high as 17 to 20 pounds of proppantadded (ppa). The fluid can be continuouslymixed and then crosslinked. It can be effec-tively degraded using available breaker sys-

46 Oilfield Review

Borate Crosslinking of Guar

Structure of Guar

B(OH)3 + OH- B(OH)4-

B(OH)4- +

B(OH)4- +

R

2R

OH

OH

OH

OHR

O

O

BOH

OH+ 2H2O

RO

OB

O

O

R + 4H2O

H

CH2

H H

OH OH

H H

O

H

HO

H

H

H

H

HO O O

H

O

CH2OHCH

2

H

OH OH

H H

OH OH

H H

OH OH

H H

O

CH2

OH

Galactosesubstituents

Mannosebackbone

R= Mannose backbone

H

CH2OH

HHO

H

OH H

OH

O

H

CH2OH

HOH

OH

OH

H

OHO O

nStructure of guarand crosslinkingby boron. Guar iscomposed of amannose back-bone with intermit-tent galactose sub-stituents (top).Boron, in the formof B(OH)4-, reactswith the hydroxyl(OH) groups on thepolymer in a two-step process to linktwo polymerstrands together(bottom).

nBorate ion fraction as a function of tem-perature and pH. The amount of borate ionin solution directly impacts the crosslinkingprocess. As the pH increases past 8, theamount of borate available for crosslinkingincreases dramatically. The solubilityeffect gives rise to the reversible crosslink-ing exhibited by borate fluid systems.

75°F100°F150°F200°F250°F300°F

1

0.8

0.6

0.4

0.2

05 6 7 8 9 10 11 12 13 14

pH

Bor

ate

ion

fract

ion

tems for rapid, efficient cleanup followingthe treatment. Conductivity tests on proppantpacks placed with low-guar fluids show 150to 200% higher pack permeabilities thanthose obtained with conventional polymerconcentrations (right).

Over 500 treatments have been pumpedsince the introduction of the fluid in 1994.Because of the success of these treatments,the low-guar system is rapidly becoming thefluid of choice for applications to 175°F.

Reduced polymer loading can lead toincreased fluid leakoff. This can be avertedby the adding fluid-loss agents, particularlyin higher-permeability formations. Field stud-ies, however, have shown that this may notalways be necessary, depending on the typeof formation and its permeability. In morethan 130 jobs performed in Kansas, USA in5- to 50-md formations, no fluid-loss addi-tives were used. Fluid efficiency was notadversely affected, and conductive fractureswere obtained. Cleanup times were reducedby 50% compared to previous treatments.Proppant placement efficiencies—the per-centage of proppant effectively placed in thefracture—met or exceeded 97%.

In even higher-permeability formations,as tests in Alberta, Canada (page 42) con-clusively demonstrate, low-guar systemscan be used effectively with the newHIGHSHEAR fluid-loss additive to improvefracture conductivity, well cleanup and wellperformance.

Controlling Proppant FlowbackFlow of proppant into the wellbore follow-ing a fracturing treatment is of major con-cern. This phenomenon may occur duringinitial cleanup or sometime after the well isput back on full production. Termed prop-pant flowback or proppant backproduction,it can lead to expensive, time-consumingremedial operations and safety concerns. Inlow-rate wells, proppant may settle in thecasing, requiring periodic wellborecleanouts. Loss of near-wellbore fractureconductivity can result, and production maycease entirely if the productive zone is fullycovered. In high-rate wells, erosion occursto tubulars, control valves and wellheadequipment. Disposal costs for producedproppant may be substantial.

The frequency of this problem hasincreased markedly as greater fracture

widths and higher proppant concentrationshave become the norm. In areas such asAlaska and the North Sea, up to 20% of theproppant may be produced back, whilesome instances of up to 50% have beenreported.20 This can translate into anywherefrom 1000 to 100,000 lbm [454 to 45,400kg] of proppant per treatment. Although theflowback may stop with time, many wellsproduce proppant throughout their life-times. Wells often must be placed onrestricted chokes to limit pressure dropsand stabilize the proppant pack.21

While changes in fracture design and exe-cution can sometimes alleviate the problem,a typical solution has been the use of resin-coated proppant (RCP). RCPs are pumpedinto the fracture near the end of the treat-ment, referred to as tailing in. The well may

Autumn 1995

nViscosity of low-guar fluids overtime. Now, reliablefluids can be for-mulated with poly-mer concentrationsof between 15lbm/1000 gal (left)and 20 lbm/1000gal (right). Even atthe lower polymerconcentration, vis-cosities are abovethe 100-cp levelrequired for ade-quate proppantsuspension andtransport.

nRetained conduc-tivity of low-guarsystems. By reduc-ing the polymerconcentration,cleanup is easierand the retainedconductivity of theproppant pack isgreater than forconventional fluidsthat require higherpolymer concen-trations.

Fluid Temperature (°F)Permeability

(darcies)Percent Retained

Conductivity

Proppant Type: Sand

Proppant Concentration: 2 lbm/gal

2% KCI20 lbm30 lbm

175175

125

32

17561

-4926

2% KCI 125 216 -15 lbm 125

125130106

604920 lbm

125 63 2930 lbm

2% KCI 100 244 -15 lbm 100 148 61

100 103 4220 lbm

Vis

cosi

ty a

t 100

sec

-1, c

p

Visc

osity

at 1

00 s

ec-1

, cp

180

160

140

120

100

80

60

40

20

0

400

350

300

250

200

150

100

50

01 2 30 0 1 2 3Time, hr Time, hr

100°F125°F150°F

15 lbm/1000 gal 20 lbm/1000 gal 19. The measure of acid intensity equal to the logarithmof the reciprocal of the hydrogen ion concentrationof the solution. pH 7 is neutral; below 7 is acidic,above alkaline.

20. Martins JP, Abel JC, Dyke CG, Michel CM and Stew-art G: “Deviated Well Fracturing and Proppant Pro-duction Control in the Prudhoe Bay Field,” paperSPE 24858, presented at the 67th SPE Annual Tech-nical Conference and Exhibition, Washington, DC,USA, October 4-7, 1992.

21. Vreeburg R-J, Roodhart LP, Davies DR and PennyGS: “Proppant Backproduction DuringHydraulic Fracturing—A New Failure Mechanismfor Resin-Coated Proppants,” Journal of PetroleumTechnology 46 (October 1994): 884-889.

47

be shut in for a period of time to allow theresin to cure as its temperature rises, bindingthe proppant particles together at theirpoints of contact. Ideally, a consolidated,high-conductivity matrix is formed.

RCPs, however, are not universally appli-cable and have certain severe limitations.Performance is sensitive to shear, tempera-ture, closure pressure and shut-in time. Con-ductivities are frequently lower thanexpected. In low-temperature wells, anexpensive activator must be added to theRCP at a 0.5 to 2.0% concentration for theresin to cure. Shut-in times may be as longas 24 hours. Resin coatings can interactchemically with fracturing fluid additives.Cyclic loadings, caused by the well beingshut in and put back on production overtime, can cause the pack to fail. In extremecases, gelled masses of resins can be pro-duced back.22

Because of these drawbacks, there is amajor incentive to introduce a more consis-tent means of controlling proppant flowbackthat also improves well cleanup efficiencyand maximizes well productivity.

Recent research has helped define themechanisms underlying proppant-packdegradation and has led to the invention ofa physical, rather than a chemical, solutionto the problem. This innovation, calledPropNET technology, uses fibers to hold theproppant in place. Pumped together withthe proppant in the fracturing fluid, the

material forms a web, or network, whichstabilizes the proppant-fiber pack andallows high production rates of oil or gas(above). The technology is based on theprinciples of fiber reinforcement commonlyused in a variety of industrial and commer-cial applications as a strengthening method.For example, natural and synthetic fibers areused to protect dams and other concreteand soil structures from erosion. The inher-ent ability of fibers to stabilize highlyporous, particulate-containing materialsprovided a basis for these investigations.

A comprehensive set of tests has beenapplied to determine the applicability offiber reinforcement and to characterize theproperties and performance of fiber-contain-ing packs. Conductivity test cells measuredproppant-pack permeability. Special pack-mobility tests simulated conditions before,during and after well cleanup. Three config-urations of test cells were developed andconstructed to evaluate the key parame-ter—pack resistance to proppant flow-back—as measured by the maximum flowrate that the pack can withstand or the max-imum pressure drop across the pack beforeproppant is produced.

A variety of fiber types were investigated,including polymer, glass, ceramic, metaland carbon. Based on several evaluation cri-teria, a special, flexible glass fiber was cho-sen for its performance, cost and availability.The material has a 2.55 g/cm3 [21.3lbm/gal] bulk density, similar to that of mostproppants. Selection of the particular glassfiber was based on its stability to fluids anddownhole conditions. Fibers must meetminimum size criteria to be effective. Short,small diameter fibers are less effective instrengthening the proppant pack. Glass fiberdiameters of 10 to 20 microns and lengthsof 10 mm or more provide optimum packstability and ease of handling.

How the fibers work is open to debate,but it is thought that they interweave among

the proppant particles, providing increasedstrength, or that they stabilize and distributestress, aiding bridging, within a significantarea of the pack. The fiber structure is moreflexible than cured RCPs, allowing the prop-pant-fiber pack to shift without failing.

Laboratory tests show that the ability ofthe pack to resist proppant flowback is afunction of fiber concentration. Stabilityincreases with fiber content until a plateauis reached (below). While laboratory datashow that use of 1.5% fibers by weight canreduce permeability by up to 30% com-pared to packs without fibers,23 field resultsshow less reduction. Conductivity values forpacks with fibers are superior to those mea-sured for postcured RCPs.24

Despite the low concentration of 1.5%,the fiber level is about 30% by number ofparticulates, or about one fiber for every twoproppant particles. Fiber length is an orderof magnitude more than proppant diameter,so, for example with sand proppants, eachfiber touches approximately five particles(next page, top).

Single-phase flow tests using water andtwo-phase tests with water and gas were

48 Oilfield Review

Crit

ical

flow

rate

, lite

r/m

in

Weight fraction fiber, %

9

8

7

6

5

4

3

2

1

0 0.5 1 1.5 2 2.5

nPropNET proppantflowback control.Random glass fibers pumped inthe fracturing fluidform a net-likestructure that stabi-lizes the proppantpack while allowinghigh productionrates of formationhydrocarbons.

nEffect of fiber concentration on criticalflow rate. As fiber is added, up to about1.5% by weight, the flow rate necessary tofail the pack increases. Beyond 1.5%,additional fibers do not improve packstrength significantly.

conducted without a confining stress andwith a stress of 1000 psi. Results show thatpack stability increases with pack width to acertain limit and that packs with morespherical and uniform ceramic proppant aregenerally less stable than irregular grainsand packs. Packs can withstand pressure

gradients in excess of 100 psi/ft, but typicallevels are 45 psi/ft (right). Worst-case flowconditions (80%/20% gas-to-water ratio)were used to establish maximum flow-rateguidelines. Under these conditions, packstability reaches a minimum at about 40%of the maximum pack strength in single-phase fluid flow. Based on laboratory data,maximum cleanup rates of 30 bbl/day/per-foration [4.8 m3/day/perforation] for sandand 20 bbl/day/perforation [3.2 m3/day/per-foration] for ceramic proppants are beingused in the field.

Tests results show that no minimum con-fining stress, shut-in time or reservoir tem-perature are required for the fibers to beeffective, overcoming some of the severestlimitations of RCPs.

Packs were cycled under stress to simulateshut-in and production periods, with packssubjected to alternating stress levels of 1000psi and 4000 psi. No pack failures wereobserved, even after more than 30 cycles,for both sand and ceramic proppant packs.Aging studies were also conducted, sinceglass fibers can be dissolved by formationwaters. The solubility rate depends on sev-eral factors, including temperature, pH and

the type of minerals present in the water (sil-ica being the most important). Results showthat fibers are expected to retain 50% oftheir effectiveness for at least two years at300°F [149°C] when in contact with a sil-ica-saturated formation brine. Lifeexpectancy could be greater, depending ondownhole conditions.

Glass fibers do not interact with commonfracturing fluid systems or additives, a keyconcern with RCPs. Their presence in thefluid slurry also reduces proppant settling,aiding proppant transport and placement.Glass fibers have certain limitations to beconsidered during treatment design. Theyare not effective at temperatures above300°F and under certain conditions: if for-

49Autumn 1995

Max

imum

pre

ssur

e dr

op b

efor

e fa

ilure

, psi

/ft

140

120

100

80

60

40

20

00 20 40 60 80 100

Nitrogen content, % total flowing

nPhotomicrographof fiber-reinforcedproppant pack.The glass fibers areabout an order ofmagnitude longerthan the diameterof a typical prop-pant particle,allowing each fiberto contact aboutfive particles. Thefibers stabilize thepack by inter-weaving amongparticles, provid-ing increasedstrength and sta-bility. They mayalso promotebridging and stressdistribution withinthe pack.

nResistance of sand-fiber packs to prop-pant flowback. At a fiber content of 1.5%,the pressure drop the pack can withstandis a maximum when no nitrogen is pre-sent, exceeding 100 psi/ft. As the nitro-gen content increases, pack strengthdecreases until a minimum is reached atabout an 80% gas volume.

22. Almond SW, Penny GS and Conway MW: “FactorsAffecting Proppant Flowback with Resin CoatedProppants,” paper SPE 30096, presented at the Euro-pean Formation Damage Conference, The Hague,The Netherlands, May 15-16, 1995.

23. This observation is consistent with calculations inwhich fibers represent 5 to 6% of the pore volume.

24. Card RJ, Howard PR and Feraud J-P: “A Novel Tech-nology to Control Proppant Backproduction,” paperSPE 31007, SPE Production & Facilities (in press).

Texas, USA:

A 10,000-ft [3080-m] gas well in south Texas,

USA with a bottomhole temperature of 275°F

[135°C] was fractured using a borate-crosslinked

fluid. A fiber-reinforced pack was placed with

1.5% fiber by weight of proppant. Fibers were

added during the entire proppant stage. Pumping

pressure levels were similar to treatments with-

out fibers present. Of the nearly 16,000 lbm

[7300 kg] of proppant placed, less than 0.07%

was produced back. The well was later cycled

through four shut-in and production periods, with

a shut-in closure stress of 1900 psi and a flowing

tubing pressure of 4200 psi. Production was

essentially proppant free and remains so today

well over a year later. Productivity from the well

exceeded that from offsets.

In another south Texas gas well, having two

sandstone layers separated by a shale layer,

226,000 lbm [102,500 kg] of proppant were

placed using a borate-crosslinked guar fluid.

Here, fibers were tailed in with the last 15% of

the proppant. Flowback started as soon as the

treatment was completed. The initial flowback

rate of 500 bbl/day [80 m3/day] was increased

to 1000 bbl/day [160 m3/day]. Less than 0.05%

of the proppant was produced back over a

four-day period.

How fiber reinforcement compares with RCPs

was evaluated in an offset well. An upper produc-

tive zone was fractured with a 15% fiber tail-in

treatment, while a lower zone used a 23% RCP

tail-in. The PropNET zone had a much higher ini-

tial flowback rate, earlier gas breakthrough and

more rapid fluid returns (middle). Cleanup costs

were reduced and fracturing fluid recovery was

maximized, allowing the well to be placed on

production sooner (bottom). Conductivities were

higher with PropNET treatments than for RCPs.

50 Oilfield Review

Economic Benefits of Fiber Reinforcement: Field Results

More than 150 hydraulic fracturing treatments using PropNET fiber-reinforced packs have been conducted in

the USA (top) and Venezuela. Here is a look at three typical examples.

■■Fluid returns in asouth Texas well.Faster flowback ratesand larger cumulativeflowback volumeswere found with thePropNET system thanwith RCPs. Gas break-through was muchearlier and productsales were realizedsooner.

■■Decrease in flowbackcosts for a south Texaswell. With a muchshortened flowbacktime, well monitoringand associated costsare substantiallyreduced, providing anincreased return oninvestment for thetreatment.

■■Fracture treatment in south Texas.

400

300

200

100

0

Cum

ulat

ive

fluid

retu

rned

, bbl

7010 20 30 40 50 60Flowback time, hr

Gas

RCP

PropNET

Gas

10,000

0

Flow

back

cos

ts, $

142 4 6 8 10 12

Flowback time, days16

20,000

30,000

PropNET flowback

RCP flowbackFlowback costsdecreased by$12,600

According to Gary Slusher, project production

engineer for Enron Oil & Gas, in Corpus Christi,

Texas, “We have had to live with RCPs. They can

be effective, but they are expensive and require

long cure times, and once you start pumping

them you’re committed. When we bring a well

back, we have to do it gingerly to avoid prob-

lems. With PropNET, we can start and stop pump-

ing as necessary and aggressively flow back the

wells. The system is more effective in stopping

proppant production than RCPs.

“We’ve seen both direct and indirect benefits

with PropNET. We’re getting much higher fluid

recoveries and more rapid cleanup. This reduces

our total job costs substantially, by 12 to 15% or

more. We believe our effective frac lengths are

longer, too. But, most importantly, we get the gas

to market sooner by turning the wells around

faster, sometimes in about 24 hours instead of

seven to ten days. We’ve also been able to use

simpler fluids with less additives. That’s saving

us additional money and reducing friction pres-

sures and horsepower requirements. When the

job has been executed according to design and

the material has been placed where it’s needed,

the results have been excellent.”

Indiana, USA:

In low-temperature gas wells in Indiana, USA,

RCPs are commonly used when fracturing multi-

ple zones. This requires an activator and an

extended shut-in period for curing of each zone

after treatment, or a total time of four to five days

per well. When fiber technology was used

instead, the initial zone could be flowed back

only 10 min after the job was completed. Follow-

ing a limited flowback period, the next zone could

be fractured immediately, completing operations

in one day, a 75 to 80% savings in rig labor, rig

time and equipment costs.

mation water is not silica-saturated, if con-fining stress exceeds the crush strength ofthe proppant, or if the fracture will betreated later with certain types of acid.25

While fibers could be distributed through-out the entire length of the proppant pack,field experience supports laboratory testsshowing that only the tail-in portion of thetreatment, typically the last 15 to 20% of theproppant pumped, normally requires fiberaddition. If, however, there are concernsthat the entire height of the producing zonecannot be covered or that sand control willbe a problem, it is prudent to use fibersthroughout the proppant pack.

Extensive field experience in the USA(see “Economic Benefits of Fiber Reinforce-ment: Field Results,” previous page) hasdemonstrated the superior performance offiber-reinforced packs, including reducedshut-in times, faster cleanup rates andgreater execution efficiency. This translatesdirectly into rapid well turnaround anddecreased wellsite costs.

On average, less than 0.2% of the prop-pant pumped was produced back in over150 treatments. Cleanup rates as high 4000bbl/day [636 m3/day] water and gas rates ofup to 10 MMscf/day were observed, 10times those used by local operators forRCPs. Since cleanup can represent 10 to60% of the total cost of a fracturing treat-ment, this time savings has a dramatic eco-nomic impact.

With the concept proven both in the labo-ratory and field, this technology will rapidlybe applied beyond the Western Hemi-sphere. To support this expansion, alterna-tive fibers are being tested for applicationabove 300°F in areas such as the North Sea.

New Solutions with Novel TechnologyThe impressive developments in fracturingfluids technology from 1985 to 1993 havebeen reinforced by innovations during thepast two years. Advanced fluid-loss addi-tives are improving fluid efficiency at lowerfluid viscosities. Combined with new low-guar systems, this means reduced costs andreduced damage to the fracture. Enhancedbreaker and additive systems are speedingwell cleanup, giving more complete fluiddegradation and cleaner, more conductivefractures. And, the fracture is now more sta-ble, thanks to innovations in proppant flow-back control.

Individually and collectively, these newtechnologies are benefiting oil and gas oper-ators by reducing treatment and wellcleanup costs, increasing well productivityand speeding product sales to market. In thefuture, the increased emphasis on more effi-cient and effective fluids and the synergisticapplication of low-cost innovations willyield further economic gains. —DEO

51Autumn 1995

25. Fibers are resistant to hydrochloric acid [HCl], theacid commonly used, but not to hydrofluoric acid[HF] which is known to dissolve glass.