By Tom Bieltz,Alex Obvintsev,Paul Riegler,Aviral Sharma,Jeffrey Kok,and Shim Yen “Yenny” Han

HOUSTON–High water production isa major issue in horizontal oil wells, es-pecially in longer laterals, commonly be-cause of the high drawdown from theheel. In addition, the presence of hetero-geneity along the lateral section can leadto uneven sweep of hydrocarbons, whichcan result in poor recovery. To control thewater production and achieve better sweepefficiency, inflow control devices (ICDs)have been introduced that balance fluidflux along the producing horizontal well. This article presents a case study of a

Gulf of Mexico Shelf horizontal wellthat was completed with prepackedscreens, and assesses the results of using

integrated technologies such as ICDsalong with rotary steerable drilling systemsand azimuthal logging-while-drilling wellplacement technology to achieve higherreservoir sweep efficiency. The well wassanded in and plugged after five years ofhigh water production. A saturation logwas run in an offset well, which grantedthe viability to revisit this reservoir. Con-sequently, a new offset horizontal wellwas proposed and drilled next to the ex-isting well to sweep remaining reserves.For the new horizontal offset, full field

dynamic simulations were performed toevaluate attic placement methodology andthe optimization of ICDs into the integrateddesign. The results illustrate that attichorizontal well placement is feasible usingintegrated drilling with ICD technologyto maximize the sweep efficiency.In a mature field development, opera-

tors are challenged to place wells as highas possible to minimize attic traps and

produce from the thin oil rims. Never-theless, early water breakthrough is stilla problem with such horizontal wells be-cause of the imbalanced drawdown alongthe horizontal and the close proximity tothe oil/water contact (OWC). In practice,to increase hydrocarbon recovery fromthese wells, flow rates are increased. Thisresults in drawing higher localized waterproduction that can cause “hot spots” inthe completion string. Mechanical failureof the sand control media often follows,leading to well plugging.To effectively exploit such reservoirs,

it is necessary to geosteer horizontal wellsas far as possible away from the OWCwith the means to balance fluid influxalong the horizontal. Developments inLWD technology have enabled structureor bed boundary mapping in real time toaccurately place wells below the reservoirtrap or within thin sand columns. Com-pletion technology incorporating stand-

Modeling Optimizes Completion Design

Structural Map of the Reservoir and Well LocationsFIGURE 1A

North-to-South Reservoir Cross-SectionFIGURE 1B

-5,370

-5,380

-5,390

-5,400

The “Better Business” Publication Serving the Exploration / Drilling / Production Industry

JUNE 2012

Reproduced for Schlumberger with permission from The American Oil & Gas Reporter www.aogr.com

alone screens combined with customizeddownhole ICDs can be introduced todelay or regulate water breakthrough.While these technologies individuallyadd value to horizontal well production,in combination, they provide the neededsolution to optimize recovery and prolongthe lifespan of a producing lateral well.

Anticlinal Reservoir

The reservoir is located on the OuterContinental Shelf near the Louisiana Coast.It is a fluvial/deltaic sand deposited inshallow water. The reservoir is an anticlinewith a sealing fault to the east, as shown inFigure 1A. The north-to-south cross-sectionof the sand in Figure 1B shows the anticlinalstructure of the reservoir. Sand thicknessvaries from 48 to 77 feet, and is thicker atthe center of the reservoir. The sand ishighly unconsolidated, making sand controlintegral to the completion strategy.The unconsolidated sand is highly

porous with an average porosity of 32percent. The log data were upscaled tocalculate a variogram, which showed majorcontinuity in the northeast direction. Themodeled variogram was then used as inputto create the porosity model. The reservoirmodel is based on a cell size of 100 x 100x 2 feet. Sequential Gaussian simulationwas used to create the porosity model. The permeability of the reservoir was

modeled based on sidewall core analysis,which was used to find the correlationbetween porosity and permeability tomodel the permeability of the reservoir.

Figure 2A shows the core analysis withcorrelation and Figure 2B shows the mod-eled permeability that was eventuallyused in the flow simulations. The highpermeability of the sand brings great pro-duction challenges for the wells in thisreservoir as a result of the strong aquiferat the bottom. This is a matter of significantconcern with horizontal wells because ofthe desired high production rates in areservoir with high vertical-to-horizontal

permeability ratios.The oil-water capillary zone was neg-

lected because of the reservoir’s highvertical permeability. Because of a lack ofpressure-volume-temperature data, fluidcorrelations were used to create PVTtables. Oil gravity of 29 degrees API, gasspecific gravity of 0.7, bubble point pressure,and initial fluid contacts were some of thekey inputs for the fluid model. A black oilmodel was used for the flow simulations.

Sidewall Core AnalysisFIGURE 2A

Modeled Permeability MapFIGURE 2B

2,500

2,750

3,000

3,250

3,750

4,000

4,250

3,500

Permeability

0.3375 0.34 0.3425 0.345 .0350.3475

Porosity

Symbol legendCore.txt PhiVsK

Gamma Ray and Resistivity Log ResponseFIGURE 3

SpecialReport: Offshore & Subsea Technology

History Matching

The reservoir has three vertical wellsand one horizontal. The first two verticalwells, V1 and V2, depleted most of thefree gas cap. The horizontal well, H1,started producing more than 5,000 barrelsof water and less than 200 barrels of oil aday after two months of production. Thelateral section of well H1 was completedwith prepacked screens that faced me-chanical failures after four years becauseof possible hot spots in the screens causedby the high-rate water production.Recent logging in an offset well re-

vealed unswept hydrocarbons in place(Figure 3). Since the hydrocarbon intervalis only 12 feet, the most viable approachis to place a horizontal well high withinthe pay. A new horizontal well, H2, wasplanned to be placed close to the H1 hor-izontal well. Two main challenges had tobe addressed for the H2 horizontal well:placing the horizontal well in a 12 foot-thick pay zone and steering it away fromthe OWC, and delaying early water break-through and balancing fluid influx.Even after placing the H2 well at the

top of the structure, the recent resistivitylog indicated that early water breakthroughwas inevitable, given that the originalOWC had moved up by 12 feet becauseof the high permeability and strong bot-tom-drive aquifer. To avoid early waterbreakthrough and high water cuts, anICD completion design was considered.To model the ICDs and their behaviorover time, it was important to understandthe dynamics of the reservoir. Dynamicsimulation incorporating pressure, satu-ration and reservoir heterogeneity wasperformed to optimize the ICD designfor the H2 well.Before modeling the ICDs, full-field

history matching was performed to cali-brate the fluid saturations, pressure andpetrophysical properties. The black oilsimulator was used for the history match-ing. The historic oil rate control was usedas a development strategy along with aminimum bottom-hole pressure limit.Figure 4 shows the results of historymatching analysis for the field. Thesimulated historic oil, water and gas rateswere matched with the resepectiveobserved data. From Figure 4, the firsttwo vertical wells (V1 and V2) producedmost of the gas cap. Well V3 had highwater production. It started producing inthe late 1970s and was shut in after adecade of high water production. It cameon line again in late 2007 and produced

for a couple years.The H1 horizontal well was completed

with prepacked screens in early 1995.The well began producing at very highwater rates, but it also had higher oilproduction compared with the verticalwells. Higher water production has beenan issue in every well drilled in this sandbecause of the high vertical permeability,and this issue became severe in the caseof the H1 wells. After more than fouryears of production, H1 was pluggedbecause of sanding from a mechnicalfailure of the completion string. The history match was also done on

individual wells to assure the consistencyof the simulation model. After matchingthe rates, observed pressure data wereplotted against the simulated pressure.

Completion Design

After calibrating the reservoir modelfor pressure and saturations, productionwas forecast for the new H2 horizontal

well. Based on the reservoir heterogeneity,the production challenges for a horizontalwell can be described as friction pressureloss from toe to heel, and coning at theheel (large drawdown for long wells) inthe case of homogeneous reservoirs, andhigh permeability layers, faults and frac-tures, and early water or gas breakthroughsin the case of heterogeneous reservoirs.In this Gulf of Mexico Shelf applica-

tion, the reservoir has relatively homo-geneous sand, making water coning morelikely to occur at the heel of the new hor-izontal well. Because of the higher verticalpermeability, high water production wasinevitable, even though the H2 well wasplanned to be landed on top of the struc-ture. To delay/control high water cutsand sand production, ICDs along withgravel packing were planned as part ofthe completion design. For the forecast,four completion designs were comparedto show the value of ICDs.The most common ICD has a nozzle-

Full-Field History Match of Oil, Water and Gas Rates FIGURE 4

1964 1972 1976 19801968 1984 1988 1992 1996 2000 2004 20080.0001

1,000

Liquid Flowrate [STB/d], [STB/d]

Water production rate History_Oil ControlOil production rate History_Oil Control

Gas production rate History_Oil ControlOil production rate Observed

Gas production rate ObservedWater production rate Observed

Gas F

lowrate [M

cf/d], [Mcf/d]

0.001

0.01

0.1

10

0.00011

10010,000

0.0010.01

0.110

100

1

1,000

SpecialReport: Offshore & Subsea Technology

Cell Permeability along the Lateral Section and ICD Joints Separated by Packers

FIGURE 5

based design, with a nozzle in the frontand a screen at the back to protect thenozzle from sand and debris. The nozzlediameter can be designed to control thepressure drop across it, and therefore,the fluid rate.ICDs create a pressure differential

across the screen such that the annuluspressure is greater than the tubing pressure.This lowers the drawdown pressure onthe reservoir. By designing the ICDs cor-rectly, the drawdown pressure balanceacross the length of the well can beachieved. To achieve this balance, suffi-cient pressure is needed at the heel toarrive at the same flow rate as the ICDs.This pressure can be delivered using apump or gas lift.In heterogeneous reservoirs where per-

meability varies significantly across thelateral, especially in fractured carbonateswhere fluid flow along the annulus betweenthe completion tubing and well bore willtake effect, packers can be used for seg-mentation to isolate zones. Unless ICDsare employed, the flow rate in each sectionwill be roughly proportional to the sectionpermeability. By adjusting the flow re-striction in each section, the fluid fluxcan be balanced within the section andalong the lateral from heel to toe.Four completion designs were simu-

lated: open-hole, gravel pack (GP), GPand ICD (GICD), and optimum ICD(OICD). Although open-hole completionis not practical in unconsolidated sand,it was used only as a base case for com-

parison. The open-hole lateral section is8.5 inches in diameter and 750 feet long.The gravel pack completion has a 4.5-inch screen in a 750-foot lateral packedwith gravel in the annulus. In the case ofthe GICD, the ICD joints are installedinside the gravel pack screen, with eightICD joints geometrically distributed andeach joint (12 meters long) separated byswellable packers for annulus isolation.The nozzle size of each ICD is 2 x 3 mil-limeters (two nozzles three millimetersin diameter).The OICD completion design is similar

to GICD, but the number of ICD jointsand their configuration is designed basedon permeability and saturations. In thiscase, 12 ICD joints were installed, separated

by swellable packers (Figure 5). The firstseven ICDs from the heel had the smallestnozzle sizes of 1 x 1.6 millimeters to pre -vent early water coning because of highpermeability. The remaining five ICDs,installed closer to the toe, had nozzle sizesof 2 x 1.6 millimeters to allow the water topush oil at the toe section for higher sweepefficiency. Different nozzle sizes and packerconfigurations were simulated along thehorizontal to compare the performances,and OICD configuration yielded the highestrecovery.The four completion designs were

simulated dynamically for one year ofproduction using the same reservoir con-trols, including reservoir volume controlsand BHP limits. The GP and GICD pro-duced the same volume of oil and waterafter one year of production because theICD nozzles in the GICD completionscenario were all the same size and toobig to control water coning.In the case of OICD, the cumulative

oil production was increased by 15 percentin one year compared with the GICD/GPbecause of the strategically selected nozzlesizes based on the permeability and satu-ration responses. Also, there was no increasein the cumulative water production afterone year, indicating that OICD allowedthe same water production, but it alsohelped push the oil. The key is to understandthe geology and petrophysics that eventuallyhelp to optimize the ICD configuration.An average pressure drop of 170 psi

was achieved across the nozzles for theGICD completion. The pressure dropacross each nozzle is proportional to thesquare of the fluid rates. In the OICDdesign, for the nozzles close to the heel,a very high pressure drop of more than300 psi was achieved initially, whichhelped choke back the high rate movingfluids into the production tubing fromthe heel section, promoting water to dis-place oil through the nozzles closer to

Structure Map from Offset Well ControlFIGURE 6

FIGURE 7Undulated Profile of Well No. 1 Trajectory along Modeled Structure

the toe for better sweep.

Well Placement

For this study, the well placementchallenges were described in a feasibilityphase that determined a fit-for-purposebottom-hole assembly for drilling andsteering measurements for both the landingand lateral.The reservoir was evaluated for well

placement or geosteering off a structuremap (Figure 6). The target sand packagewas evaluated using a cross-section cor-relation of the offset well log data. Thereservoir pay interval was defined as aseven- to 12-foot zone across these wells,which was further refined to a five-footdrilling target window. The property ofthe reservoir derived from the log datawas used to generate a measurement re-sponse for populating a preconstructedstructure to provide log response simula-tions along the proposed horizontal well.In addition to generating a response

log model for the planned lateral, theoffset lateral (well No. 1) provided themeans to study the structure model andproperties using a pre-existing lateral togain insights into the well’s structuralprofile and production history (Figure7). The reservoir top boundary depictedby offset resistivity measurements is rep-resented by the black line with the reservoirbeneath it.Based on this model, it is obvious

that the well’s trajectory (drilled with apositive displacement mud motor) undu-lates to weave in and out of the top of thereservoir. This trajectory undulation may

not be accurately represented since thereservoir top boundary was modeled withundulations to match the log with thestructure property response. This can beexplained by trajectory divergence fromthe actual when employing positive dis-placement motors, taking stationary surveymeasurements at fixed distances.Based on this understanding, the tra-

jectory could be more tortuous. Althoughwell No. 1 was located high in the struc-ture, the entire lateral was not placedwithin the zone. The sumps along thelateral may have caused a higher influxat the heel, consequently damaging themedia that could have resulted in sandparticle breach into the well that prema-turely ended the well’s production life.This assessment clearly justifies using

a rotary steerable system to provide asmoother, less undulating well. Besidesdrilling and steering needs, near-bit meas-urements such as gamma ray and contin-uous inclination along with the closed-loop steering capacity to hold or maintainthe angle of the well is instrumental toprovide a smooth lateral profile for pro-duction and completion success.Prior to placing the production hole, ac-

curate landing within the five-foot targetcolumn was the initial obstacle. While uti-lizing RSS with gamma ray and continuousinclination measurements at the bit wouldbe the ideal BHA configuration, positivedisplacement motors with measurements30 feet or more behind the bit could be em-ployed with strategic well placement method-ologies. Continuous LWD and inclinationmeasurements in combination with geosteer-

ing response modeling is critical to overcomestructural and survey uncertainties whilethe well builds through the curve to land ata 90-degree inclination.

Higher Landing Precision

Based on offset well evaluations, theoverburden can be distinguished and cor-related with gamma ray while the five-foot target can be defined only using re-sistivity measurements. Therefore, LWDgamma ray and resistivity measurementswere concluded as the minimum require-ment for the landing phase, while az-imuthal bed boundary mapping meas-urements would be the ideal measurementsfor engaging higher landing precision.Since lateral placement high within

the target for a maximum displacementfrom the OWC was critical to the successof the well, azimuthal resistivity meas-urement or a bed boundary mapper is es-sential. To validate this recommendation,the feasibility for engaging the bed bound-ary mapping measurement was evaluatedby generating synthetically modeled re-sponses from structure model properties.This was achieved by selecting fit-for-purpose azimuthal LWD measurementsto simulate the responses for the horizontalwell across the targeted reservoir. Forthis reservoir, the response simulationsillustrated that azimuthal resistivity meas-urements from a proprietary LWD tooldesigned with tilted receiver coils couldhelp achieve the primary objective ofplacing the well in the optimal distancefrom the OWC.Figure 8 demonstrates that the bed

boundary mapping tool is capable of de-tecting the upper boundary of the targetsand five and a half feet away and theOWC eight feet TVD away from the wellbore. The actual distances are a functionof a resistivity profile that is computed inreal time, where no variables or constantsare needed to manipulate the bed-to-tooldistances. The ability to map the forma-tion’s bed boundary facilitates the meansto proactively steer the well. This reduceswell path tortuosity to facilitate the smoothrunning of the completion string andallows the production hole trajectory po-sitioning objectives to be achieved.The planned trajectory is placed in a

geometrically updip structure to providesensitivity from the conductive bed forthe response analysis. The resistivityphase and directional curves in the firsttwo tracks from the top in Figure 8 aresimulated raw measurements from thebed boundary mapper, while resistivityand gamma ray are simulated standardlog measurements. The dots in the az-imuthal resistivity detection view represent

Azimuthal Resistivity Mapping Distance to the Bed BoundaryFIGURE 8

Azimuthal resistivity detection view showing top and bottom boundaries

Azimuthal resistivity canvas view showing 2 boundary mapping model.

~5.5 ft below the top; ~8ft from OWC

direction and distance of the conductivebeds that are mapped, with blue aboveand red below the tool. The azimuthalresistivity canvas view shows these bedsmapped for two boundaries, and the insertillustrates the distances from the tool tothe bed.

To overcome the identified well place-ment and geosteering challenges, RSSwith azimuthal resistivity bed boundarymapping measurements is the optimalchoice for drilling and steering the curveand landing sections of the well. Azimuthalresistivity with bed boundary mapping

enables precise lateral trajectory placementhigh within the target and maximum dis-placement from the OWC for optimumrecovery using ICD completions. RSSwith azimuthal resistivity bed boundarymapping measurements is essential tothe success of this project. r

TOMBIELTZ

Tom Bieltz is a reservoir engineerworking the East Bay Field for EnergyPartners Ltd. He has 30 years of expe-rience in both reservoir and reserveevaluations, and spent most of his careerwith Amoco/BP prior to joining EnergyPartners as a consultant. Bieltz earneda degree in petroleum engineering fromPennsylvania State University.

ALEXOBVINTSEV

Alex Obvintsev manages drilling op-erations for Energy Partners Ltd. Inhis 15-year career, he has performedwell design, planning, management, su-pervision and post-analysis for onshore,shelf and deepwater international anddomestic projects. Obvintsev joined En-ergy Partners in November 2004, fol-lowing eight years with Schlumberger’sintegrated project management division.He holds an M.S. in petroleum engi-neering from Freiberg University.

PAULRIEGLER

Paul Riegler is a geologist with En-ergy Partners Ltd., working offshoreLouisiana. He started his career withExxon in Kingsville, Tx. Riegler holdsa B.S. in geology from the New YorkState College at Oneonta, an M.S. ingeology from the University of RhodeIsland, and an M.S. in management,computers and systems from HoustonBaptist University.

AVIRALSHARMA

Aviral Sharma is as a productionengineer for Schlumberger North Amer-ica, and has been involved in multipleproduction optimization and reservoirstudies for both deepwater and uncon-ventional plays. He joined Schlumbergerin 2008 as a reservoir engineer, beforemoving to his current position. Sharmaholds an M.S. in petroleum engineeringfrom the University of Texas at Austin.

JEFFREYKOK

Jeffrey Kok is the well placement in-structor for the Schlumberger HoustonTraining Center. He began his career asa wireline field engineer in France,Trinidad & Tobago, and Argentina. Hethen worked with Baker Hughes INTEQin geosteering in the Asia Pacific region.Kok holds a bachelor’s in mechanicalengineering from the University of Sal-ford.

SHIM YEN “YENNY” HAN

Shim Yen “Yenny” Han is the petro-physics instructor for the SchlumbergerHouston Training Center. In 15 yearswith Schlumberger, she has worked asan LWD and geosteering engineer, aswell as a petrophysics domain championfocused on marketing, introducing newtechnology, technical support, and de-veloping new products. Han holds apetroleum engineering degree from theUniversity Technology of Malaysia.

SpecialReport: Offshore & Subsea Technology