04_Vaca_Muerta_Shale.pdf

8/18/2019 04_Vaca_Muerta_Shale.pdf

1/14

26 Oileld Review

Vaca Muerta Shale—Taming a Giant

Unconventional resource plays revolutionized oil and gas production in North

America. Because these types of plays provide many challenges to operators, they

have been slow to develop in other parts of the world. However, the Vaca Muerta

formation in the Neuquén basin of Argentina can now be added to the list of

successful producers. The triumph in the Vaca Muerta play resulted from engineers

at YPF, SA and Schlumberger collaborating on an integrated approach that includes

eld- and formation-specic workows and the application of a dynamic unconven-

tional fracture model to optimize hydraulic stimulation programs.

Matías Fernández BadessichDamián E. HrybMariano SuarezYPF, SABuenos Aires, Argentina

Laurent MosseNuncio PalermoStéphane PichonLaurence ReynoldsBuenos Aires, Argentina

Oileld Review 28, no. 1 (January 2016).Copyright © 2016 Schlumberger.Mangrove, Petrel and UFM are marks of Schlumberger.

Coaxing commercial quantities of hydrocarbonsfrom unconventional reservoirs revolutionizedthe oil and gas industry in North America. Thelessons learned there are now being applied inother parts of the world. One of the latest successstories comes from the giant Neuquén basin of Argentina, where YPF, SA is developing the VacaMuerta Shale formation for production of oil andnatural gas.

Production from shale reservoirs traces itsroots to activities in the Barnett Shale of north-central Texas, USA. In 1981, Mitchell Energy &Development Corporation drilled a well for theexpress purpose of producing natural gas fromthe Barnett Shale; the company then spent20 years guring out how to make the play com-mercially viable. Success in the Barnett was fol-lowed by the deployment of new techniques andtechnologies to nd and produce natural gas andthen oil from other basins and formations acrossNorth America. Recent success in the Neuquénbasin of South American has the potential torevolutionize the oil and gas industry in thatregion in a manner similar to that experienced in

North America. Argentina’s unconventional resource plays have

been recognized by exploration companies forsome time. The Vaca Muerta formation is consid-ered by many experts to have some of the greatestpotential for hydrocarbon production in the world(Figure 1) . One independent study compared shaleresource potential outside North America andranked the Vaca Muerta formation third globally intechnically recoverable gas and fourth in oil. 1

Figure 1. Global shale resources. Argentina hasabundant reserves in unconventional reservoirs.Globally, the country is third in technicallyrecoverable shale gas reserves ( top ) and fourthin technically recoverable shale oil ( bottom ). Themajority of Argentina’s unconventional reservesare associated with formations in the Neuquénbasin. (Adapted from the US EIA, reference 1.)

US15%

US14%

China14%

China10%

Argentina10%

Argentina8%

Algeria9%

Canada7%

Technically Recoverable Shale Gas

Canada3%

Mexico7%

Australia6%

South Africa5%

Russia4%

Pakistan

3%

Brazil3%

Others20%

Others19%

Russia22%

Technically Recoverable Shale Oil

Australia5%

Venezuela4%Mexico4%

Libya8%

1. US Energy Information Administration (EIA): “TechnicallyRecoverable Shale Oil and Shale Gas Resources: AnAssessment of 137 Shale Formations in 41 CountriesOutside the United States,” Washington, DC:US Department of Energy EIA, June 2013.


2/14

January 2016 27


3/14

28Oileld Review

Several operators are evaluating the VacaMuerta play; YPF, however, has been the mostactive and successful in initiating production to

date. The company established early on that theplay has unique characteristics that affect devel-opment programs. The fact that the Vaca Muerta

formation has reservoir-specic characteristicsthat differ from other shale plays comes as nosurprise; operators of shale resource plays gener-ally agree that, to maximize production, eachplay must be approached and evaluated based onits own intrinsic merits. The Vaca Muerta Shalepossesses features that enhance its potential todeliver nancially viable volumes of hydrocarbons.

To maximize production and reduce the

development cycle of the play, engineers at YPFare now using fracture simulators, predictivenumerical models and eld-specic workowscreated from information derived during anextensive evaluation program. The process ishelping them understand the effects of rockmechanics on fracture propagation duringhydraulic stimulations. The information gainedfrom their workows and simulations is thenused to optimize well spacing, fracture stimula-tion designs and fracture treatments. The meth-odology has helped the engineering team reducethe development time of the play compared tothe time taken by iterative approaches typicallyused in other shale developments.

The methodology YPF has taken in developingthe Vaca Muerta formation is driven in partbecause the formation has a combination of fac-tors that differentiates it from other plays: a highpore pressure, a complex stacking pattern of vary-ing lithologies and geomechanical complexityarising from the close proximity of the Neuquénbasin to the Andes Mountains. The engineers at YPF have also found fracture gradients in someparts of the basin that are abnormally high, whichgreatly impacts completion strategies. A thoroughunderstanding of the Vaca Muerta formation’sgeomechanics and reservoir heterogeneity is cru-cial for proper eld development. With thesemany factors in mind, YPF has started to movefrom a traditional, and often costly, trial-and-errormethodology of developing resource plays to amodel-and-optimize methodology.

This article provides a basic overview of the Vaca Muerta formation and YPF’s work in theplay to date. The article also reviews workowsdevised by the YPF and Schlumberger teams andthe simulation programs that are used to opti-

mize both drilling operations and stimulationtreatments and briey discusses current produc-tion and future development plans.

In the Shadow of the AndesThe Neuquén basin lies on the east ank of the Andes Mountains in west-central Argentina(Figure 2) . Oil was rst discovered there in1918.2 The basin is a major source of oil and gas

2. Hogg SL: “Geology and Hydrocarbon Potential of theNeuquén Basin,” Journal of Petroleum Geology 16, no. 4(October 1993): 383–396.

3. Instituto Argentino del Petróleo y del Gas, http://www.iapg.org.ar/suplemento/Agosto2015/Produccion%20por%20cuenca.html (accessed October 30, 2015.)

4. Back-arc basins are geologic features associated withisland arcs and subduction zones. Found at convergentplate boundaries, they form when one tectonic platesubducts beneath another, creating a trench. Subductionzones are frequently associated with formation ofvolcanoes and seismically active areas.

5. Hogg, reference 2.6. Garcia MN, Sorenson F, Bonapace JC, Motta F, Bajuk C

and Stockman H: “Vaca Muerta Shale ReservoirCharacterization and Description: The Starting Pointfor Development of a Shale Play with Very GoodPossibilities for a Successful Project,” paperSPE 168666/URTeC 1508336, presented at theUnconventional Resources Technology Conference,Denver, August 12–14, 2013.

7. Howell JA, Schwarz E, Spalletti LA and Veiga GD: “TheNeuquén Basin: An Overview,” in Veiga GD, Spalletti LA,Howell JA and Schwarz E (eds): The Neuquén Basin,Argentina: A Case Study in Sequence Stratigraphy and

Figure 2. Neuquén basin. Several basins in South America have potential as unconventional resourceplays. The Neuquén basin, in west-central Argentina (red box), is the major source of oil and gasproduction for the country and contains three deepwater marine shale deposits—the Agrio, VacaMuerta and Los Molles formations. These shales act as source rocks for oil and gas deposits found inconventional reservoirs in the basin.

SOUTH AMERICA

0 500 1,000 km

0 300 600 mi

Paraná basin

Chaco basin

Neuquénbasin

San Jorgebasin

Austral-Magallanesbasin

Prospective basin

CHILE ARGENTINA URUGUAY

PARAGUAY

BRAZILBOLIVIA

PERU

Basin Dynamics . London: The Geological Society,Special Publication 252 (2005): 1–14.

8. Badessich MF and Berrios V: “Integrated Dynamic FlowAnalysis To Characterize an Unconventional Reservoir inArgentina: The Loma La Lata Case,” paper SPE 156163,presented at the SPE Annual Technical Conference andExhibition, San Antonio, Texas, USA, October 8–10, 2012.

9. For more on geochemical characterization of sourcerocks: McCarthy K, Rojas K, Niemann M, Palmowski D,Peters K and Stankiewicz A: “Basic PetroleumGeochemistry for Source Rock Evaluation,”Oileld Review 23, no. 2 (Summer 2011): 32–43.

10. Stinco L and Barredo S: “Vaca Muerta Formation:An Example of Shale Heterogeneities ControllingHydrocarbon’s Accumulations,” paper URTeC 1922563,presented at the Unconventional Resources TechnologyConference, Denver, August 25–27, 2014.Badessich and Berrios, reference 8.

11. Ejofodomi EA, Cavazzoli G, Estrada JD and Peano J:“Investigating the Critical Geological and CompletionParameters that Impact Production Performance,”paper SPE 168709/URTeC 1576608, presented at theUnconventional Resources Technology Conference,Denver, August 12–14, 2013. .


4/14

January 2016 29

production from conventional reservoirs—57% ofthe natural gas and 40% of the total oil produc-tion within Argentina come from it. 3 The targetareas for shale development in the basin are LateTriassic to Early Cenozoic strata, deposited in aback-arc tectonic setting. 4 Active volcanoeslocated along the present-day Andean Cordilleraproduced pyroclastic emissions during variousgeologic periods, and the resulting ash beds cover

large areas of the basin. Ash beds are observed within the shale formations that are targets ofcurrent exploration.

The basin formed as a result of sea levelchanges and tectonic plate movement that beganin the Early Triassic period, which were followedby a succession of transgressive and regressivecycles. Restricted access to the open sea duringseveral periods led to the formation of thick evap-orite beds. 5 Reactivation of the volcanic arc dur-ing the late Jurassic also affected the currentbasin topography. 6 Present-day basin architec-ture resulted from uplift and folding during thelast Tertiary Andean movement.

The source rocks for oil and gas found in conventional reservoirs of the Neuquén basin arethe Cretaceous Agrio formation, Early Cretaceousto Late Jurassic Vaca Muerta formation and themore mature Middle Jurassic Los Molles forma-tion (Figure 3) .7 All three formations are deepwa-ter marine shales and have potential as resourceplays. The Vaca Muerta formation is the primarysource for hydrocarbons in conventional reservoirs and is in the oil window across approxi-mately 60% of the basin.

The Vaca Muerta formation was depositedbetween the Tithonian age—a Late Jurassictransgression—and the Berriasian stage of theEarly Cretaceous epoch. 8 The characteristics ofthis organic-rich formation make it a prime tar-get for exploration. 9 These include high averagetotal organic carbon (TOC) levels (1% to 8% andspikes to 12%), moderate depths of 3,150 m[10,335 ft] and overpressured conditions thathave a pressure gradient range from 13.6 to20.4 kPa/m [0.6 to 0.9 psi/ft].10

The formation varies from about 60 m [200 ft]to 520 m [1,700 ft] in thickness from the embay-

ment area to the basin center. Matrix porosity varies from 4% to 14%; the average is 9%. Matrixpermeability ranges from nanodarcies to micro-darcies; the presence of natural fractures con-tributes to the productive potential of theformation. Shale maturities based on vitrinitereectance, R o, range from less than 0.5% up to3%, which represents black oil to dry gas.11 Its

Figure 3. Lithostratigraphy of the Neuquén basin. Three major deepwatermarine shale deposits are in the Neuquén basin. The deep Los Mollesformation is Early to Middle Jurassic in age. The Vaca Muerta formationstraddles Late Jurassic to Early Cretaceous. The Cretaceous Agrioformation lies above the Vaca Muerta formation. The sediments in these formations were deposited during successive transgressive andregressive cycles. Restricted access to open seas and anoxic conditions that existed during deposition preserved their organic content. (Adaptedfrom Howell et al, reference 7.)

Period Epoch Age

Maastrichtian

Campanian

Loncoche formation

Jaguel formationPircalaformationRoca

formation

Huitrínformation

Lohan Curaformation

La Amargaformation

Centenarioformation

Quebrada delSapo formation

Picún Leufúformation

Bajada Coloradaformation

Rio Colorado formation

Rio Neuquén formation

Rio Limay formation

Rayoso formation

Quintuco formation

Vaca Muerta formation

Tordillo formation

Lotena formation

Auquilco formationLa Manga formation

Tábanos formation

Los Molles formation

Pre-Cuyo group

Huechulafquen Formation and Piedra Santa Complex

Chachil formation

Lajas formation

Challacó formation

Lapa formation

Upper member

Lower member

Mulichinco formation

SantonianConiacianTuronian

Cenomanian

Albian

Aptian

Barremian

Hauterivian

Valanginian

Berriasian

Tithonian

Kimmeridgian

Oxfordian

Callovian

Bathonian

Bajocian

Aalenian

Toarcian

Pliensbachian

Sinemurian

Hettangian

Continental or volcanic rock

Volcanic rock

Plutonic and metamorphic rock

Evaporite rock

Offshore clastic and carbonate rock

Shallow-marine clastic and carbonate rock

Triassic

E a r l y

M i d d l e

C u y o g r o u p

M e n

d o z a g r o u p

R a y o s o g r o u p

A g r

i o

f o r m a t

i o n

N e u q u

é n g r o u p

M a

l a r g

ü e g r o u p

L o t e n a

g r o u p

L a t e

E a r l y

C r e t a c e o u s

J u r a s s

i c

L a t e

Palaeozoic

Lithostratigraphy


5/14

30Oileld Review

large gross thickness, areal extent and reservoirproperties have placed the Vaca Muerta Shale inthe ranks of a world-class resource play. 12

The deep marine sediments of the VacaMuerta Shale consist of ne laminations of blackand gray shales and limey mudstones mixed withorganic material. The anaerobic environment in which these organic materials were depositedhelped preserve their organic content. The for-mation comprises three sections—lower, middleand upper. The lower corresponds to an innercarbonate platform that consists mostly of marls,

carbonates and limestone. The middle wasformed predominantly as a slope deposit and hashigher siliciclastic content than the other twosections. The upper section returns to a predomi-nately carbonate platform. 13

The type of hydrocarbon expected in the VacaMuerta Shale generally depends on the location within the basin. The eastern and southernregions have potential for oil production, the western region is predominately dry gas and thearea between the oil and dry gas zones is wet gasand condensate prone (Figure 4) .

Effects from the early Cenozoic Andes orog-eny introduced structural complexities in theNeuquén basin, especially along the westernfront.14 The close proximity of the western VacaMuerta to the Andes Mountains created stressesand natural fractures that are less prevalentalong the eastern edge of the basin. The inu-ence of the Andes on the stress state of the reservoir rocks helps distinguish it from other shale

plays, especially those in North America.

Unlocking PotentialMost unconventional plays have unique featuresthat differentiate them from other plays. In thenascent period of shale play development, opera-tors were looking for the next Barnett Shale. Although lessons learned from the Barnett Shaleand other unconventional plays were crucial forhastening commerciality and acceptability ofshales as viable prospects, operators realizedearly on that each new development required aspecic and often unique approach. Because ofits many distinct characteristics, this is espe-cially true of the Vaca Muerta Shale. Distinctionsfor the formation include the prevalence of ashbeds, complex facies stacking, abnormally highpore pressure and geomechanical stresses result-ing from the proximity of the Andes Mountains.

To optimize the development program for the Vaca Muerta play, YPF collaborated with otheroperators and service companies that had exper-tise in shale developments. Working withSchlumberger geologists and engineers, YPFengineers devised workows and simulations toefciently exploit the Vaca Muerta formation(Figure 5) .15 The objective was to avoid the costlyand time-consuming trial-and-error approachcommonly employed in the development of manyshale plays. Creating and using a predictivemodel rather than relying on a statisticalapproach was deemed crucial for nancial viability.

To meet the objective of creating a useful pre-dictive model, the team members rst denedspecic known problems for developing the play.They established data requirements for eachstage of the process that included seismic, petro-

physical and geophysical data (Figure 6) . Theexistence of 3D seismic data over much of thearea was benecial in creating these models.Hundreds of legacy wells have also been drilledfor exploring and developing conventional reservoirs, although the data acquired in those wellsmay not always provide the type of informationneeded to dene parameters for drilling optimi-zation, well spacing and fracture design in theunconventional sections. Additional data were

Figure 4. Hydrocarbon type in the Vaca Muerta Shale. Hydrocarbon types in the Vaca Muerta Shalevary throughout the Neuquén basin (blue outline). The eastern and southern regions contain mostlyoil (green), the west-central section is predominately dry gas (pink) and the region between the twoyields wet gas and condensate (yellow). (Adapted from the US EIA, reference 1.)

S O U T H A M E R I C A

CHILE

ARGENTINA

CHILE

ARGENTINA

0 125 250 km

0 75 150 mi

Neuquén basinOilWet gas and condensateDry gas

Figure 6. Data requirements for predictive model.

Seismic and geologic data

Petrophysical data

Geomechanics

Microseismic

Discrete fracture networks (DFN)

Completion design

Fracture design

Horizons, faults and stratigraphy

Water saturation, porosity andpermeability

Pore pressure, stresses and elasticproperties

Event type and location

Azimuth, spacing and length

Staging perforation cluster andperforation properties

Pumping schedule, proppant, fluidvolumes and pressures

Property

Geologic model

Openhole logs and core data

Mud data, openhole logs andformation tests

Geophysical interpretations

Image logs, image interpretaionand model estimates

Job data

Job execution data

Source


6/14

January 2016 3

12. US EIA, reference 1.13. Ejofodomi et al, reference 11.14. The Andes Mountains were formed by the Cenozoic

tectonic shortening of the South American plate as itoverrode the subducting Nazca plate, an oceanic platein the Eastern Pacic Ocean off the west coast ofSouth America.

Figure 5. Unconventional reservoir integrated workow. In developing the Vaca Muerta resource play, engineers at YPF and Schlumbergercollaborated on a workow that reduced the degree of trial and error.Geologists rst characterized the reservoir using 3D seismic data,petrophysical data, a 3D mechanical earth model and 3D discrete fracturenetwork models (1). Using these models, they can develop the hydraulicstimulation program based on automatic fracture gridding (2). A complexhydraulic fracture model can be compared to stimulated volumes computed

during treatments (3). During the stimulation treatment, uid rates andpressures are recorded, and microseismic events may be captured (4). Asensitivity analysis helps engineers understand the effectiveness of thechanged parameters (5). If the outcome of the stimulation does not match the model, the empirical data are used to update the simulations. If thestimulation results match the model, additional information, including coreand uid properties, is used to predict production, which can eventually bevalidated using production logs. (Adapted from Hryb et al, reference 16.)

Automatic Fracture GriddingPumping Schedule

Reservoir Characterization

Sensitivity Analysis

Production Evaluation

Complex Hydraulic Fracture Model

1 2

3

Pressure

Production

3D seismic

3D DFN

Petrophysics

3D MEM

Rates

4

5

acquired that were then used to model the forma-tion and create the simulations.

The engineering teams formulated and

adopted integrated workows to help them ana-lyze the old and new data. A 3D geomechanicalmodel was constructed from well log and 3D seis-mic data. Petrophysical data were acquired frompilot wells and used to ne-tune the models andcompute geomechanical properties such asin situ stresses, Young’s modulus, Poisson’s ratioand pore pressure. To calibrate the modeledresults, fracture injection tests were performedin some pilot wells.

As part of their process, the YPF andSchlumberger engineers used history matching to validate predicted responses from the simulations

and then compared the model responses with real- world well results. They were able to determinethe plausible geometry of existing and created

fracture networks in near real time from thesemodels, which allowed them to make well-to-weoptimizations in addition to stage-to-stage modi-

cations during hydraulic fracture operations.The methodology adopted for the complex

hydraulic fracture modeling takes geologic models,

15. Lacentre P, Pichon S, Suarez M and Badessich MF:“Simulación dinámica integrada de fracturamientohidráulico y reservorio para pozos horizontales enVaca Muerta Shale Oil & Gas,” presented at theSegundo Jornadas de Simulación, Buenos Aires,July 7–8, 2015.


7/14

32Oileld Review

3D seismic models and other data sources to cre-ate the 3D geomechanical model. 16 The geome-chanical model is combined with discretefracture network (DFN) models to simulate frac-ture treatments. Natural fracture models arebased in part on image log data that honormechanical ow units, fracture azimuth, dipangle distribution and fracture density.

A fracture design based on simulator resultscould then be executed with real-time microseis-mic monitoring.17 By using actual productiondata, the engineers are able to evaluate theeffectiveness of the methodology. If the resultsindicate that stimulation pressures, fracturepropagation and stimulated volumes are in line

with predictions, then future production fore-casts can be made. When the results are not inline with predictions, the models are updated ormodied to reect the results.

Reservoir Heterogeneity The reservoirs of some early shale developmentprojects were treated as if the rocks were isotro-pic. In the case of organic shales, large variationsin formation properties and characteristics canexist both laterally and vertically; the extent ofthese variations is often much greater than what isencountered in developing conventional reservoirs. Because hydraulic fracturing and horizontaldrilling are basic requirements in the majority of

unconventional plays, reservoir heterogeneity,especially regarding geomechanical formationproperties, must be understood and accounted forin stimulation and drilling programs.

In shale reservoirs, geologic discontinuities, which include faults, natural fractures, beddinggeometry and changes in mechanical properties,inuence stimulation results. Evaluation tech-niques generally treat mechanical properties as

affecting the completion quality (CQ) of the res-ervoir rocks (Figure 7) . The CQ is a predictiveattribute of the reservoir rocks that dependslargely on Young’s modulus, Poisson’s ratio, bulkmodulus and rock hardness. Natural fracturedensity and orientation, rock anisotropy and pre- vailing in situ stresses are also incorporated indetermining the CQ of the reservoir rocks. 18 TheCQ values incorporate information about claytype and volume, which are obtained from spec-troscopy measurements as well as mechanicalproperties. The mechanical properties includePoisson’s ratio, Young’s modulus and stress pro-les computed from acoustic and other petro-physical data.

Along with CQ, the reservoir quality (RQ) ofthe rocks—another predictive property—deter-mines the economic viability of shale plays. Fororganic-rich shales, RQ is a measure of the capac-ity to produce hydrocarbons after hydraulic frac-ture stimulation. The RQ property is a product ofporosity, hydrocarbon saturation, mineralogy,organic content and thermal maturity. Rocks thathave a combination of superior RQ and CQrespond to stimulation treatments better than dorocks with poor RQ and CQ properties.

To accurately characterize unconventionalresources such as organic shales, the reservoirevaluation often relies on advanced petrophysicsbased on spectroscopy, nuclear magnetic reso-nance and dielectric dispersion measurements.Once the formation properties are calibratedusing advanced petrophysical data, a minimaldataset may be used to approximate RQ. Thisminimal set of log measurements includes resis-tivity, bulk density, porosity, water saturation andpore pressure.

Sweet spots may be determined from a com-

posite quality index (CQI) that accounts for boththe RQ and CQ properties. For the completiondesign, perforation clusters and fracture stagesare assigned to the best CQI intervals. TheMangrove engineered stimulation design used inthe Petrel platform automates the process ofdetermining the best CQI intervals, and engi-neers are using completion programs devisedfrom the Mangrove software recommendations(Figure 8) .

Figure 8. Quality indicators for well completion and design from Mangrove software. Log data are used to generate reservoir quality (RQ) and completion quality (CQ) indicators. These data are processedusing Mangrove stimulation design software that automatically highlights optimal intervals based onpredetermined cutoff valves. By focusing on intervals that have the best RQ and CQ, and eliminating those with poor RQ and CQ, better stimulation results can be obtained compared to those obtainedusing a geometrical approach in which intervals are completed and stimulated uniformly withoutregard to the formation properties.

Engineered Placement of Fracture Stages and Perforation Clusters

Rock quality

Stress gradient

HighLow

Stress gradient

Rock QualityGood RQ and good CQ

Bad RQ and good CQGood RQ and bad CQ

Bad RQ and bad CQ

Perforation cluster

Fracture stage

Figure 7. Data requirements for fracture modeling.

Organic content

Thermal maturity

Effective porosity

Intrinsic permeability

Fluid saturations—oil, gas,condensate and water

Organic shale thickness

Hydrocarbons in place

Mineralogy—mainly clay, carbonate and silica

Mechanical properties—Young’s modulus, Poisson’sratio and tensile strength

Natural fractures—presence, density, orientation andcondition (open, closed or cemented)

In situ stress—variations between intervals accountingfor mechanical properties anisotropy

Reservoir Quality (RQ)Completion Quality (CQ)


8/14

January 2016 33

The CQI is a continuous index based on RQand CQ at each depth, although the completionengineers must account for the interactions thatoccur in the formation and around the perfora-tion clusters during the stimulation process. Asingle value of CQ at a given depth can be com-pared with neighboring CQ values, whereas theRQ values must be integrated over the fracture’sestimated vertical propagation length and vol-ume to represent the capacity of the stimulatedformation to produce hydrocarbons.

Data acquisition in a vertical pilot well isessential for dening the predicted CQ response,especially in plays such as the Vaca Muerta, where vertical heterogeneity is common. Thedenition of RQ and CQ for Vaca Muerta play con-tinues to evolve as the formation characteristicsare better understood.

Unconventional Fracturing ModelTo effectively produce from ultralow permeability

shale formations, operators try to generateclosely spaced fractures during stimulations toincrease the reservoir contact and increasedrainage. However, stress effects may oppose thecreation of new fractures in the presence of prop-agating fractures. To design and properly executean effective hydraulic stimulation, engineersmust understand the interactions of fractures incomplex fracture networks typical of those in the Vaca Muerta play.

One distinctive aspect of the YPF workow isthe UFM unconventional fracture model softwareused to model hydraulic fracture creation andpropagation. The UFM modeling is part of theMangrove software.19

Observations from actual jobs show that pre-dicting fracture propagation in multistage hydrau-lic fracture operations has not always yieldedaccurate results. Most simulation models do notproperly handle complex fracture propagation inthe presence of natural and newly induced frac-tures. The UFM simulation, rst introduced in2011, tackles this weakness in other simulationsby determining whether a hydraulic fracture willpropagate through an existing fracture, stimulat-ing fresh formation, or if it will be subjugated byexisting fractures and simply continue along anexisting fracture plane (Figure 9) .20 The crossingmodel used in the UFM simulation is an essentialcomponent in the predictions.

The UFM software also accounts for stress

shadow effects—the stress exerted on surround-ing rocks and adjacent fractures by propagatingfractures—in simulating the propagation of mul-tiple or complex fractures. Stress shadow causesrestrictions in fracture width, which can signi-cantly reduce the effectiveness of proppantplacement to the point of screenout. 21

When multiple fractures are propagatingsimultaneously in close proximity, the transportof uid and proppant into the newly formed

fractures may be adversely affected. The stressshadow results in narrower fracture widths thanstatic propagation models suggest. The predic-tions from the UFM simulations more closelyreplicate the dynamics of actual treatments

16. Hryb D, Archimio A, Badessich M, Ejofodomi E, Diaz A,Cavazzoli G, Zalazar F, Lagarrigue E and Pichon S:“Unlocking the True Potential of the Vaca Muerta ShaleVia an Integrated Completion Optimization Approach,”

paper SPE 170580, presented at the SPE AnnualTechnical Conference and Exhibition, Amsterdam,October 27–29, 2014.

17. For more on real-time microseismic monitoring:Burch DN, Daniels J, Gillard M, Underhill W, Exler VA,Favoretti L, Le Calvez J, Lecerf B, Potapenko D,Maschio L, Morales JA, Samuelson M and Weimann MI:“Live Hydraulic Fracture Monitoring and Diversion,”Oileld Review 21, no. 3 (Autumn 2009): 18–31.

18. Glaser KS, Miller CK, Johnson GM, Toelle B,Kleinberg RL, Miller P and Pennington WD: “Seeking the Sweet Spot: Reservoir and Completion Qualityin Organic Shales,” Oileld Review 25, no. 4(Winter 2013/2014): 16–29.

19. For more on UFM software processing: Wu R, Kresse O,Weng X, Cohen C and Gu H: “Modeling of Interaction ofHydraulic Fractures in Complex Fracture Networks,”paper SPE 152052, presented at the SPE HydraulicFracturing Technology Conference, The Woodlands,Texas, February 6–8, 2012.

20. Cipolla C, Weng X, Mack M, Ganguly U, Gu H, Kresse Oand Cohen C: “Integrating Microseismic Mapping andComplex Fracture Modeling to Characterize HydraulicFracture Complexity,” paper SPE 140185, presented at the SPE Hydraulic Fracturing Technology Conferenceand Exhibition, The Woodlands, Texas, January 24–26,2011.

21. Screenout occurs when the proppant carried in a treatment uid creates a bridge across the perforationsor in fracture networks. The restriction to uid owcauses a rapid rise in pump pressure and ends theplacement of additional proppant.

Figure 9. Unconventional fracture model. The engineers at YPF adopted aUFM simulator to characterize Vaca Muerta Shale stimulations. Inputs to the simulation ( left ) begin with recorded microseismic (MS) events thatare compared to a microseismic monitoring (MSM) model. The results areused to calibrate the UFM simulation and the distributed fracture network(DFN). The DFN model is used to predict fracture geometry created duringhydraulic stimulations. The UFM simulations rely on material balance,geomechanical properties, DFNs and wellbore geometry to predict fracture

propagation properties ( right ). The cell-based outputs of the simulationspredict fracture geometry—whether the induced fractures will propagate through existing fractures, rebranch after crossing fractures or follow the track of existing fractures ( top right ). The software computes fractureheights based on these interactions of natural and induced fractures.Proppant transport and settling ( lower right ), which directly impactproduction and the stimulated volume predictions are generated with theUFM simulation software. (Adapted from Cipolla et al, reference 20.)

DFN realizationMicroseismic events

Compare to MSM

Prediction fracture geometry

C a l i b r a

t e U F M

a n d D F N o

u t p u t s

Fracture treatmentdataEarth model

Natural fracture

Rebranching

SlurryFluid

Well

Hydraulicfracture Crossing


9/14

34Oileld Review

because it accounts for fracture geometry, which ultimately controls uid transport and

proppant placement.The UFM software simulates fracture propa-

gation, rock deformation and uid transport inthe complex environment of stimulation treat-ments by using material balance, geomechanics,DFN models and well geometry. The programdelivers fracture height estimates, predicts inter-actions of natural and induced fractures andmodels proppant transport.

Typical Vaca Muerta WellThe workow and modeling that YPF uses in the

Vaca Muerta development offer considerable ex-ibility; however, the general process can be out-lined for a typical well. A subject well, drilledparallel to another well from the same pad and vertically separated by approximately 260 ft[80 m], was drilled laterally for several thousandfeet.22 Engineers used Mangrove software todetermine staging, perforation cluster locationsand completion strategy (Figure 10) .

The RQ and CQ outputs from the Mangroveprocessing were limited using predetermined

cutoff values initially determined from the fullycharacterized vertical pilot well. The process forstimulation design comprised the following steps:Identify intervals that have favorable RQ and CQ values based on log measurements along the lat-eral, determine sweet spots based on CQI, cross-reference CQI with formation properties andinformation from the vertical pilot well, andstimulate the zones highlighted from theMangrove processing. The intervals and perforation

Figure 10. Vaca Muerta Shale completion design. The Mangrove softwareautomatically determines optimal stimulation intervals. The typical inputsof gamma ray (Track 1), resistivity (Track 2), bulk density (Track 3) andwater saturation (Track 4) are derived from basic well log data. Effectiveporosity (Track 4) and clay volumes (Track 5) are computed using datafrom advanced logging tools, as are mechanical properties such as pore

pressure (Track 6), Poisson’s ratio (Track 7), Young’s modulus (Track 8) andstresses (Track 9). The RQ and CQ indices (Tracks 11 and 12) are combined to produce a composite quality index (Track 10) from which the stages andperforation clusters are determined (Track 13). Perforation clusters aregrouped by similar properties to ensure maximal outcomes. (Adapted fromHryb et al, reference 16.)

3 , 3

0 0

3 , 4

0 0

3 , 5

0 0

3 , 6

0 0

3 , 7

0 0

3 , 8

0 0

3 , 9

0 0

4 , 0

0 0

4 , 1

0 0

4 , 2

0 0

G G

G o o

d

G o o

d

G o o

d

G o o

d

G o o

d

G o o

d

G o o

d

G o o

d

G o o

d

G o o

d

G o o

d

G o o

d

G o o

d

G o o

d

G o o

d

B a

d

B a

d

B a

d

B a

d

B a

d

B a

d

B a

d

G G

G G

B G

B G

B G

B G

B G

G G

G G

G G

G G

G G

G G

G G

G G

G G Composite Quality Index

Completion Quality Index

Bulk Density

Resistivity

Gamma Ray

Reservoir Quality Index

Young’s Modulus

Poisson’s Ratio

Pore Pressure

Clay Volume

Minimum HorizontalStress Gradient

Effective PorosityWater Saturation

Missingdata

S t a g e

2

S t a g e

3

S t a g e

4

S t a g e

5

S t a g e

6

S t a g e

7

S t a g e

8

S t a g e

9

S t a g e

1 0

S t a g e

1 1

DesignedStages


10/14

January 2016 35

clusters were also grouped based on similarstress proles (one of the CQ indicators) to helpensure that the treatment breakdown pressures would be similar for individual stages. Otherwise,the treatment would have a tendency to follow apath of least resistance and ow through a lim-ited number of perforations, which would notstimulate the rock effectively.

The workow also allows the design team to

simulate scenarios to determine an optimal stim-ulation schedule. For example, for another well inthe project, the predicted results of a hybrid treat-ment were compared to that of a slickwater-onlytreatment (Figure 11) . The hybrid treatment—acombination of slickwater, gel and crosslink gel—should produce a higher gas rate and maintainthat output compared to the simulated resultsobtained using a slickwater-only treatment.

The predicted normalized ow rate per clus-ter for the fracture treatment was used to groupsimilar clusters. The treatment consisted of acombination of slickwater, 20 lbm [9 kg] cross-linked gel and 20 lbm crosslinked gel carrying100, 40/80, 30/60 and 20/40 ceramic proppant(Figure 12) . Each stage consisted of 500,000 lbm[227,000 kg] of proppant pumped at a rate of70 bpm [11 m3 /min].

22. Hryb et al, reference 16.Figure 11. Stimulation simulation comparisons. Engineers modeled and compared treatment scenarios to determine which was more effective. Both designs were based on 80-m [260-ft] stage lengths. Thehybrid design ( top left ) is based on four clusters per stage, and the slickwater design ( top right ) isbased on eight clusters per stage. The hybrid design included slickwater, gel and crosslink gel stagesalong with a mixture of proppant sizes. The slickwater-only treatment was predominately 100 meshsand with some 40/70 mesh sand. The cumulative gas production (red line) of the hybrid design clearlyexceeds that of the slickwater-only treatment (blue line). The predicted hybrid ow rates are higher than those of the skickwater-only treatment ow rates, and the measured data validates the hybridsystem prediction. (Adapted from Hryb et al, reference 16.)

D i m e n s

i o n

l e s s g a s r a t e

1.00

SlickwaterGelCrosslink

100 mesh sand40/70 mesh sand30/50 mesh sand

SlickwaterGelCrosslink

100 mesh sand40/70 mesh sand30/50 mesh sand

0.75

0.50

0.25

022015

100%

81%

19%20%

Hybrid Treatment80-m stage length, 4 clusters per stage

Slickwater Treatment80-m stage length, 8 clusters per stage

Hybrid treatmentCumulative production

Predicted rate

Measured rate

Slickwater treatment

38%

14%

36%50%

42%

Days1050

Figure 12. Simulated designs. The hydraulic simulation was optimized for theparameters of the specic well. Each of the ten stages used a specic uid type and volume ( left ). Each stage consisted of a mixture of four proppant

mesh sizes ( right ), pumped as a percentage of the total volume. (Adaptedfrom Hryb et al, reference 16.)

%

100

90

80

70

60

50

40

30

20

10

0

Simulation DesignHybrid completion design Slickwater

Stage1 Average2 3 4 5 6 7 8 9 10

Crosslink gel

%

100

90

80

70

60

50

40

30

20

10

0

Simulation Design4 proppant mesh sizes500,000 lbm proppant at 70 bpm 30/60 mesh proppant

Stage1 Averag2 3 4 5 6 7 8 9 10

20/40 mesh proppan100 mesh proppant 40/80 mesh proppan

Linear gel


11/14


12/14

January 2016 37

production uses small grid blocks in the near wellbore and in the fractures. In this simulation,engineers used 445,000 cells and 19 layers, which provided high-resolution results and nedetails. The team used the simulation to deter-mine that the proppant had a tendency to settleout upon fracture closure.

To conduct the history matching, engineerscontrolled the total liquid ow rates—both oiland water—to match the observed uid phasesand owing pressures from well tests. From theseresults, they determined that the most importanthistory matching component for the matrix (as itexisted prior to fracture stimulation) was pres-

sure-dependent permeability. For the newlyinduced fractures, it was both porosity- and pres-sure-dependent permeability; however, matchingthe owing pressures using xed permeabilitiesand porosities was not possible. For the model-ing, a pressure-dependent permeability/porosityratio was developed (Figure 14) .

Workow UpdatesNow that there are more than 400 wells in theeld, the impact of history matching and job mod-ications becomes more evident. The UFM software outputs, production models and continuousfeedback provide an understanding of important variables for future wells and treatments. The

data are used to determine optimal spacingbetween lateral wells for maximizing productionbut minimizing negative crosswell impacts.

Engineers can use the simulation outputs todetermine the optimal fracturing uids, prop-pant type and volumes, pumping schedules, per-foration cluster locations and placement,longitudinal staging along the wellbore, horizon-tal staging related to other wells, stress shadoweffects and the impact of changes made from jobto job and from stage to stage. They use thesedata along with sensitivity analyses to determine which parameters have the greatest impact on well performance (Figure 15) . Future hydraulic

Figure 14. Production predictions from stimulation simulations. Theestimated relative permeability ( left ) is a key component in predictingproduction from fractured reservoirs. The simulator can produce plots for

increases in permeability, detect unstimulated matrix, determine unproppedfractures and quantify the expected production from the stimulated volumeand propped fractures ( right ).

N o r m a l

i z e

d f r a c t u r e

d i s t r

i b u t i o n

16

0

2

4

6

8

10

12

14

Grid cell fracture permeablilty, mD-ft

Propped Fracture Permeability Histogram

0.12 0.40 0 .80 1.20 4.00 8 .00

P e r m e a

b i l i t y

i n c r e a s e

1.0

0

0.2

0.4

0.6

0.8

Pressure, psi0 1,000 2,000 3,000 4,000 5,000

Unpropped

Propped

Matrix

6,000 7,000 8,000 9,0

Figure 15. Postfracture calibration of models. Input parameters havesome degree of uncertainty because of the nature of the shale formation’sheterogeneity and fracture complexity. Uncertainty analyses can beperformed after the stimulation program to quantify these uncertainties anddetermine model sensitivity ( left ). Similarly, evaluation of production results

can provide levels of uncertainty for the inputs used to predict production(right ). The degree to which this uncertainty affects the simulation can beused to evaluate the quality of the predictions. (Adapted from Hryb et al,reference 16.)

DFN

Vertical stress contrast

Horizontal stress anisotropy

Injected fluid properties

Maximum stress orientation (angle)

Fracture extension

Fracture compaction

Propped fracture

Matrix permeability

Fluid properties

Fracture porosity

Matrix porosity

Matrix compressibility

Uncertainty Analysis, Hydraulic Fracture Model Uncertainty Analysis, Production Prediction


13/14

38Oileld Review

fracture stimulation designs are modied from

these simulations to improve efciency and maxi-mize the stimulated volume.

The workows used by YPF and Schlumbergerput all the analysis tools in a single platform, which simplies the modeling. The results can bequickly analyzed, and analysts are able to visual-ize the effects of parameter changes. The resultsinclude the uncertainties associated with each variable, which further help in determining which

parameters have the greatest impact on the out-

come and the optimal course of action to take.Engineers compared simulations results that

were based on previous generation fracture mod-eling software with simulations that used theUFM software and found the results more closelymatched the empirical data. The engineersattributed the more accurate results to the UFMsoftware’s proper treatment of the developmentand propagation of complex fractures. The level

of uncertainty from these simulation models can

be high due to nonunique solutions; however, byusing data from individual wells to calibrate thesimulations, the analysts can improve the modelsfor future drilling and stimulation programs(Figure 16) .

24. Collins PW, Badessich MF and Ilk D: “AddressingForecasting Non-Uniqueness and Uncertainty inUnconventional Reservoir Systems Using ExperimentalDesign,” paper SPE 175139, presented at the SPE AnnualTechnical Conference and Exhibition, Houston,September 28–30, 2015.

Figure 16. Optimization strategies as part of the standard workow. After the calibrated and corrected models and simulations become available,engineers can develop optimization strategies. Using the simulations, the team can vary landing location, proppant characteristics, uid deliverymethods, perforation strategies and contributions from existing conditions

to optimize well locations and spacings and eld development. Simulationscan be presented graphically, and then engineers can compare the results,which helps them choose the best course of action. (Adapted from Hryb etal, reference 16.)

• Horizontal landing location• Various proppant systems• Various fluid systems• Fibers• Perforation strategy• Impact of natural fractures

Optimization Strategy


14/14

Not only do the workows help users under-stand well dynamics and help them focus on themost critical parameters for adjusting programs,they also benet the operations. Field personnelare aware of the need for quality measurementssuch as permeability, uid saturations, effectiveporosities and pressure data that will be used inthe simulations and modeling. When followingthese well-dened programs, eld personnel canfocus on what is required to ensure successful jobexecution and efcient job operations.

Current StatusCompared to its North American counterparts,the Vaca Muerta Shale development is a relativelynew play. The rst shale gas well in the play thatinvolved hydraulic fracture stimulation was com-pleted in mid-2010 and was brought online inDecember of that year. Ramped-up developmentdrilling did not start until 2013, when the rstcluster of 100 wells operated by YPF was drilled inthe Loma Campana block. Spaced at one well per20 to 40 acres [0.08 to 0.16 km2], the majority ofthe early wells were drilled vertically. Productionforecasting is included in the normal workows,and specic datasets have been dened to ensurethe necessary inputs are available for the predic-tive software (gure 17) . Because the develop-ment is new, forecasting medium-to-long-term well performance is still a challenge. 24

Today, several wells that were drilled horizon-

tally and completed using multistage stimulationprograms are producing from the Vaca Muertaformation. These lateral wells were initially com-pleted with four to six stages per well; however,more stages, with as many as fteen stages per well, are being attempted. Most of the wells weredrilled vertically, which is in direct contrast tohow wells are drilled in the majority of the uncon-

ventional plays in North America, where horizon-tal well drilling is standard practice. YPF initiallychose vertical well congurations for targetingthe Vaca Muerta play because of the signicant vertical heterogeneity of the formation and sub-stantial pay thickness.

The vertical well methodology employed by YPF engineers is beginning to shift. Althoughlarge-scale drilling of horizontal wells that werecompleted using multistage fracturing did notbegin until late 2014, by the end of 2015, approxi-mately 10% of the producing wells were horizon-tal. This increase in horizontal drilling as aproportion of the total wells drilled has beenadopted because engineers at YPF now feel con-dent that they have identied sweet spots forlanding horizontal targets within the vertical por-tion of the reservoir.

The LLL-992h horizontal well that YPFrecently drilled and brought online is a goodexample of this shift in methodology. The well isregarded as a major milestone for the project as itheralds a transition from drilling primarily verti-cal wells to drilling a high percentage of horizon-tal wells. This well, drilled to a lateral length ofapproximately 6,600 ft [2,000 m], achieved apeak production of 1,630 bbl/d [260 m3 /d], whichfar exceeded initial expectations. These resultsare encouraging the operator to migrate morerapidly from a vertical well architecture to a hori-zontal well development scheme for wells drilled

in both the oil and the gas windows.In spite of reservoir complexity and hetero-

geneity, the integrated approach YPF has takenis helping the company develop the play. Byunderstanding the discrete fracture networksthrough use of the UFM software, engineers andgeologists are able to optimize stimulationdesigns and develop realistic modeling of

proppant placement and fracture growth.Microseismic measurements are used on some wells in real time to adjust programs during jobexecution and then are used to update and ne-tune existing mechanical models.

The Vaca Muerta Shale is massive, and eventhough YPF has drilled numerous wells, thisoperator and others have only just begun thedevelopment and commercialization of the play.

As more wells come online, and eld complexities become better understood, this giant has thepotential to revolutionize the oil and gas industryin Argentina as well as encourage operators toexplore the possibilities for shale resources inother South American basins. —TS

Figure 17. Data requirements for generating production predictions.

Production history

PVT

Changes in rock properties

Relative permeabilities

Oil, gas and water volumes

Bottomhole pressure

Oil and gas properties

Changes in the matrix

Changes in unstimulated rock

Changes in conductivity ofthe propped fractures

Oil, gas and water permeabilities

Property

Production data

Laboratory measurements

Calculated from wellhead pressure

Laboratory measurements on core

Adjustments in parameters using empirical data

Data from service provider

Adjustments in parameters using empirical data

Source

04_Vaca_Muerta_Shale.pdf

Documents

Transcript of 04_Vaca_Muerta_Shale.pdf