Copyright 1999, Society of Petroleum Engineers Inc.

This paper was prepared for presentation at the 1999 SPE Latin American and CaribbeanPetroleum Engineering Conference held in Caracas, Venezuela, 2123 April 1999.

This paper was selected for presentation by an SPE Program Committee following review ofinformation contained in an abstract submitted by the author(s). Contents of the paper, aspresented, have not been reviewed by the Society of Petroleum Engineers and are subject tocorrection by the author(s). The material, as presented, does not necessarily reflect anyposition of the Society of Petroleum Engineers, its officers, or members. Papers presented atSPE meetings are subject to publication review by Editorial Committees of the Society ofPetroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paperfor commercial purposes without the written consent of the Society of Petroleum Engineers isprohibited. Permission to reproduce in print is restricted to an abstract of not more than 300words; illustrations may not be copied. The abstract must contain conspicuousacknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

AbstractPDVSA, began its first pilot test of a steam assisted gravitydrainage (SAGD) production system in Latin America duringJune 1997. The dual well SAGD process requires thehorizontal drilling of two wells one above the other. The upperwell is dedicated to injecting steam in a continuous form andthe lower to produce the reservoir fluids.

The pilot test was carried out in the Tia Juana Field,Western Venezuelan, where reservoir fluids are highly viscous(typically 20,000 cp. under reservoir conditions). Reservoirdepth is approximately 1000 ft with a porosity of 38% and apermeability of 1-2 darcies. The pilot wells consisted of twopairs of infill wells within an original seven spot pattern with awell spacing of 231 meters.

The planning and implementation of the test was donebetween June and December 1997. Production started onDecember 15, 1997. The drilling stage of the pilot required theuse of a Magnetic Guidance Tool (MGT) to drill thehorizontal wells with a 5 meter distance between producer andinjector along the horizontal section.

The paper shows pressure and temperature data acquiredduring the SAGD process from both conventional single pointgauges and the latest generation of fibre optic distributedtemperature measurements providing real time thermalprofiles along the complete well bore.

The production results during the first year of operationshow an average production of 700 bopd, which is above theinitial production expected. This performance is promising,and it is expected that the initial objective of 60% recovery inthe pilot area will be reached approximately 3 years fromthe onset of production.

During the first year of production, a history match wasperformed using the Stars simulator. The comparison of real

and simulated production shows close agreement and a finalrecovery in the order of 60 % of STOIIP has been estimated.

IntroductionTo date the use of steam assisted gravity drainage (SAGD), forthe production of heavy oil has been concentrated in Canada.The first SAGD was attempted by Imperial Oil in 1978 by Dr.Roger Butler. The first dual well SAGD was drilled at theUnderground Test Facility (UTF) in Fort McMurray, Albertain 1987. Three pairs were drilled from an underground tunnel.In each horizontal well pair, the lower well was placed 1-2meters above the limestone with the upper well located 5meters above the producer. This project was the firstsuccessful field demonstration of the SAGD process andprovided valuable operational know-how necessary forsuccessful application to commercial production. As of 1997,some 22 SAGD pilot projects are active (1) showing the needfor optimization of the SAGD process required for economicand reliable production with a given reservoir.More recently a new concept of SAGD has been used whichinvolves only a single well. This single-well SAGD (SW-SAGD) (2) in which only one horizontal well is used for bothinjection and production has met with some success inCanada.This paper describes the planning, implementation, operationand history matching of the first pilot dual-well SAGD inLatin America.

Planning of the TestThe area selected for the pilot test was the Tia Juana Field,Western Venezuela Figure 1; it contains heavy oil with arelatively high viscosity (10,000-45,000 cp. at 110 F) whichhas historically shown low recoveries in the order of 10%under Cyclic Steam Stimulation (CCS). The CCS process hasbeen effective in parts of the Tia Juana Field, but in thenorthern part of the field where the viscosity is higher the CCSprocess has shown lower recoveries than expected. The lowrecovery obtained, together with the long steam cycle timesand the problems associated with cold production of a viscousoil, provided the stimulus for the investigation of othertechniques to improve oil recovery. SAGD is a process thatemploys long horizontal wells and can be operated at highrates with systematic sweep of the reservoir. Recoveriessignificantly greater than 50% have been achieved in the UTF(3) project.

SPE 53687

SAGD, Pilot Test in VenezuelaHumberto A. Mendoza, Jose J. Finol, PDVSA and Roger M. Butler, GravDrain Inc.

2 HUMBERTO MENDOZA, JOSE FINOL, ROGER BUTLER. SPE 53687

The proposed dual well SAGD involved the use twohorizontal wells. The horizontal injector well, located close toand vertically above a horizontal producer, was used to injectsteam along the horizontal portion of the well. The injectedsteam tends to rise and forms the underground steamchamber .The heated oil drains downwards to the horizontalproducer well located vertically below the injector well.It was decided to drill two parallel pairs as shown in figure 2perpendicular to the dipping of the reservoir. This decisionwas based on the good production performance of the LSE4703 very close to the pilot area.

Geology of Pilot AreaHeavy Oil deposits in the Tia Juana Field occur in the LowerLagunillas Member of the Lagunillas Formation. The reservoiris composed of channel deposits, with the top of the reservoirvarying from 450 to 1000 ft (TVD). The reservoir structure isa monoclynal 3 dipping to the southwest with no faults in oraround the pilot area. The total thickness of the lowerlagunillas formation averages 280 ft and is divided into twomain layers. The top layer 40-85ft. was selected for the pilottest because of the homogenous condition of the rock (99%quartz).

Drilling and CompletionsThe SAGD drilling process has become possible with theadvent of technology to drill a horizontal well for steaminjection above and parallel to another horizontal forproduction. The magnetic Guidance Tool (MGT), guides thesecond well (injector) and keep it a measured distance fromthe first. In this drilling method, the first well is drilled usingconventional directional steering. The second well is landeddirectly above the first, tracking the magnetic ranging datafrom the MGT in the first well. Having drilled certain distance(30 ft.) in the second well the MGT is moved farther into thelower well where it is tracked by the directional tool in theupper well and so on, until the complete measured distance isdrilled. Vertical distance bettwen wells was chosen as 5meters based on viscosity and past experience from otherpilots in Canada.All wells were drilled vertically to approximately 200 ft. with13 3/8 casing, follows by 9 5/8 production casing up to 1300ft. and 7 0.015x 1.5 slotted liner to final depth of 2350 ft. inan 10 hole( see figure 3)The first pair (LSE 5085-5088) was completed with 6conventional thermocouples at the heel, middle and toe of thehorizontal sections. Capillary tubing was installed to the heeland toe of the wells to monitor pressure at the same time. Thesecond pair was completed additionally with a fiber opticsystem to measure a complete well temperature profile in realtime. This technology was chosen in part to evaluate thetechnology itself but also to evaluate enable monitoring ofsteam hanneling into the producer and progressive growth ofthe steam chamber with time.

Production HistoryFirst pair LSE 5085-5088 has been on production sinceDecember 15, 1997. Figure 4 shows one year production withan average oil production of 700 BOPD, 50 % water. Theoriginal expectation was in the order of 300 BOPD and a peakproduction of 700 BOPD after one year of operation.Cumulative production after one year is 230,000 bls, thecumulative steam injection is 270,000 Bls which indicate acumulative Oil Steam Ratio of 0.8 B/B which is consider anexcelent initial performance when is compared to othersSAGD projects.Pressure measurements from capillary tubing installed at theheel and toe are shown in figure 5. Several stages are observedfrom preheating up to normal operation after december 1997.Under normal operation we measured a 40 psi pressure dropbettwen heel and toe sections operating under Steam Trapconditions with 30 F of subcooling (according to temperaturereading from thermocouples).Typical Fiber Optic measurements from the second well pairare shown in Figure 6. Fiber optically acquired thermalprofiles are indicative of the the energy state of the completewell. This is extremely helpful to detect potentail problemssuch as steam channeling and should enable qualitativelypreferential flow and flow patterns along the horizontalsection in the producer to be determined.

Reservoir Fluid PropertiesTypically in the Tia Juana Field there are two layers separatedby shale. The in situ oil viscosity exhibits wide variation,(10,000 to 20,000 cp.), porosity is around 38% andpermeability 2.0 Darcies. Initial oil saturation is 85% and hasa gravity of 9-11API. The reservoir thickness for the firstpair is 40 ft at the heel section of the well. The reservoirthickness increases to 85 ft at the toe section for a horizontaldistance of 1390 ft. The gas oil ratio is low (60 scf/bl) and wasmeasured in neighboring wells.

Analytical SimulationThe main parameters used in the analytical recoverycalculations are listed in the following table:Bottom Hole Quality 70%Residual Oil Saturation 10%Reservoir Ther. Cond. Btu/ft.day.F 22.4Initial Res. Temp. 100FSteam Chamber Pres. 500 psiWell spacing 328 ft.Soi 85%Swi 15%Effective Permea. 1.5 DRatio of Vert.To Hori. Perm. 0.43Thickness above producer 30 ft.

Table1: Reservoir Data for Analytical Simulation

The results of the calculation for the selected area using theButler (4) equations are shown in figure 4. The peak oil rate isin the order of 700 BOPD after one year operation. The

SPE 53687 SAGD, Pilot Test in Venezuela 3

cumulative Oil Steam Ratio was 0.22 B/B for a period of 3.5years. For the same period of time the estimated CumulativeOil recovery was 60% of STOOIP.

Numerical SimulationA coupled reservoir and wellbore model was constructedto simulate the two-well SAGD process using anadvance grid option technique called the DiscretizedWellbore Option. This technique treats the horizontalwells as a second porosity as in the dual porosityapproach for a naturally fractured reservoir. TheDiscretized Wellbore Option is implemented in theadvanced CMGs thermal simulator STARS. It shouldbe noted that the model is in a strip amongst otherinterfering and producing wells which have not beenconsidered. The objective of the numerical simulationwas to gain a better understanding of the fluids flowinside the strip and to assess several scenarios forimproving oil recovery.

Model DescriptionThe reservoir model consists of a strip according to thegeological section along the horizontal length of SAGD wellpair. This section indicates an increase in the contributorysand thickness from the heel towards the toe of the well. TheGamma Ray Log for well LSE-5018 indicates isotropy in thevertical direction for the principal sand. This has beenrespected in the simulation model. Figure 7. depicts thereservoir/well grid model. It is an 8 25 17 Cartesian gridrepresenting half of the two-well pattern.The average reservoir thickness is 74 feet and the seventeencells in a vertical direction vary in size from 4 to 8 feet. In theJ direction, the symmetric half model extends 218 feet and has25 cells of varying width. The reservoir top has been assigneda value of 785 feet. A porosity value of 0.05 and permeabilityvalue of 1 md. have been arbitrarily assigned to the existingupper shaly sand. The horizontal permeability of the cleansand has a base value of 1500 md. However, this permeabilityvalue has been altered during several sensitivity analyses. Fig.8 depicts sand-shale variance and sand divergence along thehorizontal section of the wells.Table 2 gives some of the initial reservoir data used in thesimulation.

Thickness,ft. 85Horizontal Permeability,md 1500Vertical Permeability,md 1000Porosity 0.37Rock heat capacity, Btu/cu.ft-F 35Thermal Oil Conductivity Btu/ft.day.F 1.8Thermal Water Conductivity 8.6Thermal Gas Conductivity 0.025 Initial Pressure, psi 350 Initial Temperature,F 100 Soi 0.85 Swi 0.15

Table 2. Reservoir Data for Simulation

The characterisation of reservoir fluids, particularlythe viscosity vs. temperature table was obtained byusing PVT Builder (CMG Software) which allowsusers to interactively define PVT tables. The PVTdata for simulation was generated from well-knowncorrelations and matched to several real dependentpressure data points; GOR, Bo, Bg and mo, so thatthe resulting PVT data was more representative ofactual conditions.

Discretized Wellbore ModelIn order to model well flow more accurately the horizontalinjector and producer were defined as a discretized wellbore.In this case, every wellbore section is treated as a grid blockand the completion to the reservoir is handled as an interblockconnection.The fluid flow equations and the energy equations were solvedfor each discretized wellbore block the same way as equationsdescribing flow in the reservoir. Fluid properties and relativepermeability tables were assigned to the wellbore whilewellbore porosity was automatically set to one. Initialpermeability was calculated from the Hagen-Poiseuilleequation for laminar flow in a pipe as .8/2wr The initialpermeability and relative permeability curves were adjusted toyield the pressure drop and phase slip from multiphase flowcorrelations. Equations for wellbore flow and fluid transferbetween wellbore and reservoir are described in [5].

Well ConfigurationThe horizontal injector was completed in layer 11 and thehorizontal producer in layer 15, 3 meters above the bottom ofthe reservoir. The separation between both wells was exactly 5meters. Table 2 shows the well data employed in this study.

4 HUMBERTO MENDOZA, JOSE FINOL, ROGER BUTLER. SPE 53687

Wellbore:

length, ft 1390inner diameter of tubing, ft 0.249outer diameter of tubing, ft 0.291inner diameter of casing, ft 0.515outer diameter of casing, ft 0.583

Table 3. Wellbore Data for Simulation

Steam was initially injected in both wells at 200 psi. throughthe tubing, and fluids were produced from the annulus. Theinjection temperature and steam qualities were specified at thetubing inlet, and a minimum bottom hole pressure wasimposed as the main constraint at the annulus outlet.

History MatchThe average daily historic oil production and steam injectionwere specified for each well throughout the history matchingperiod and were updated monthly. Modifications were madeprimarily to absolute vertical and horizontal permeabilities.The rock water/oil and gas/oil relative permeabilities wereused with no adjustments. The model was adjusted so thatsimulated flowing BHP for the horizontal producer matchedthat measured in the field by bottom hole sensors. Figs. 9,10and 11 show the well history match for the horizontalproducing well of oil rate, water cut and flowing bottom holepressure from start-up through July 1998.

Simulation ResultsPrediction base case and two sensitivity runs were performedfor gross steam injection rates of 160, 120 and 200 T/Drespectively from July 1998 to Dec. 2000. Fig. 12 shows a plotof cumulative oil production for the existing producinghorizontal well. The plot shows an ultimate recovery of6.4 105 STB for the base case, which represents 52% of theOOIP of the drainage area. Thus far, the incremental oilrecovery under the higher steam injection rate (200 T/D), isestimated at 6.7 105 STB; it is nearly the same cumulativegross production rate for the base case (160 T/D), while theultimate recovery under the lower steam injection rate (120T/D) is estimated at 6.1 105 STB.

Fig. 13 exhibits a plot of the cumulative oil-steam ratio for thehorizontal producing well, being the higher cumulative oil-steam ratio at lower steam injection rate.Fig. 14 shows the steam saturation at the end of the steamcirculatin period and figure 15 illustrates the growth of thesteam chamber, with time, through a vertical plane parallel tothe horizontal wells. It shows the first steam breakthroughoccuring at the beginning of the horizontal section, causing thesteam to heat the surrounding formation by thermalconduction making the oil less viscous. This effect allowsmore steam to be injected, causing the steam to sweep thereservoir preferentially at the heel of the well pair. Figures 16

to 19 shows how the steam chamber moves from left to rightwith time .

Conclusions

1. - A SAGD pilot test has been successfully implemented inthe Tia Juana Field. After one year, field experience has beengained in planning, drilling and operating the process.

2. - Initial production rate was better than expected (700 vs-300 BOPD) but in the longer term real production has thetendency to match the analytical and numerical simulations.

3.- Final recovery are in the order of 52 and 60% for samesimulations which means an increase in ultimate recovery bymore than 40% over the conventional steam soak recovery inthis field.

4.- Steam Injection needed to keep the producer underNatural Flow, Steam Trap and producing at low water cut(50%) was in the order of 120 140 T/D. This is close to70% less steam than expected which provides an OSR of 0.8B/B

5.- From 3 years simulation, optimal injection rate is about120 to 140 T/D and oil production was not greatly affected byby higher steam injection rates.

ACKNOWLEDGEMENTThe authors wish to thank PDVSA for supporting this workand for permission to publish this paper. Planning andimplementation of the pilot were posible due to clearinstructions given by Dr. R Butler, Mr. C. Bhome andperformed by a high quality multidisciplinary team inPDVSA.

REFERENCES

1. Corbett, B., Steam Assisted Gravity Drainage (SAGD)Technologies and the Oil Sands, Alberta Chamber ofResources Directory 1997, pp-30-36.

2. Nzekwu, N.,Designing Thermal Improved Oil Recoverywith Horizontal Wells, Inernational Heavy Oil Symposium,Calgary ,Alberta June 1995.

3. Mukherjee,N.J., et al, A Comparison of Field vs ForecastPerformance for Phase B of the UTF SAGD Project in theAthabasca Oild Sands, 6 ta UNITAR Conference, Feb, 1995

4. Butler,R.M., Baratha, S and Yee,C. T., SAGD Study EastTia Juana Field, C-7 Project, Internal Report August 15,1996.

5. Oballa, V., Coombe, D.A and Buchanan .L., Aspects ofDiscretized Wellbore Modelling Coupled toCompositional/Thermal Simulation,JCPT, VOL. 36 N 4,April 1997.

SPE 53687 SAGD, Pilot Test in Venezuela 5

6. Edmunds, N.R., and Suggett, J.C., Design a CommercialSAGD Heavy Oil Project. SPE Paper# 30277.

7. Edmunds, N.R., and Gittins, S.D., Effective SteamAssisted Gravity Drainage to Long Horizontal Well PairsPaper# CIM/AOSTRA 91-65.

ZON

A EN

REC

LAM

AC

ION

MARACAIBOCABIMAS

TIA JUANA

LAGUNILLAS

BACHAQUERO

LAGODE

MARACAIBO C-7D-7

D-8

C-8

TIA JUANA FIELD

SAGD PILOT TEST

FIG. 1 LOCATION SAGD PILOT TEST

1264

1591

3303

4207

4211

4213

4252

4258

4270

4274

4277

4301

4307

4398

4399

4400

4445

4491

4510

4511

4512

4517

4540

4543

4544

4545

4546

4554

4560

4564

4618

4620

4625

4629

4646

4648

4692

4693

4696

4650

4652

4654

4657

4556

4567

S 4762

S 4763

4758

S 4765

4768

4622

4628

4695

4561

4541

4626

4662

4772 C7

-643'

-751'

- 765'

-766'

-834'

-728'

-687'

-648'

-795'

-685'

-645'

-812'

-835'

-744'

-725'

-646'

-649' -692'

-704'

-744'

-590'

-751'

-621'

-616'

-537'

-707'

-678'

-693'

-604'

-550'

-564'

-554'

-583'

-498'

-513'

-518'

-513'

-520'

-692'

L D4798

4799

-466'

-483'

4844

-475'

4703

S

F

-565'

-427'

-552'-692'

-634'

-634' -570'

4902 4921

E

4944

4991

-803'

4993

-789'4994

-791'

4995

4996

-826'

4998

-830'

4999

-833'

5002

5004

-824'5005

5006 -717'5008

-739'5009

-744'

5010 -719'

5011

5018 -701'

5027

-757'5029

-776'

5030 -840'

5079

E

F

5083

E

F

5085E

F

5091E

F

5092

5088

SS

SS

SS

E

E

F

F

5177

5175

4692 UAUS-2

UAUS-1

F

FF

E

E

UAS-3

UASN-2

FIG. 2 PROJECT AREA OF SAGD WELLS

050

100

150

200

250

300

350

400

450

MEASURED DEPH FT.

13-3

/8 C

AS

. @21

4

9-5/

8 C

AS

. @ 1

281

7 LINER 0.015 x 292 SLOTS/FT @ 2936 FT

1100

- 2

-7/8

TU

BIN

G

10 HOLE @ 2936 FT

LIN

ER

@ 1

176

3-1/

2 T

UB

ING

@ 2

706

TE

MP

ER

AT

UR

E

F

0 500 1000 1500 2000 2500 3000

FIG. 6 TYPICAL FIBER OPTIC TEMPERATURE PROFILE Vs. MEASURED DEPTH, PRODUCER LSE-5091

ji

FIG. 7 GRID MODEL (8 x 25 x 17), PAIR LSE-5085 / 5088

FIG. 8. SAND-SHALE (POROSITY) DIVERGENCE ALONG THE HORIZONTAL WELLS

J F M A

1998J

NUM. SIM. ACTUAL DATA ANAL. SIMU.

J F M A M J J

1998

FIG. 9. COMPARISON OF SIMULATED OL PRODUCTION RATE AGAINST TIME

0

200

400

600800

1000

1200

1400

D

1997

FIG. 10. COMPARISON OF SIMULATED WATER CUT AGAINST TIME

20

30405060

7080

90100

Wat

er C

ut

(%)

D J F M A

1997 1998

FIG. 11. COMPARISON OF SIMULATED BOTTOM HOLE PRESSURE AGAINSTTIME

Simulated BHP Toe Press Heel P ress

100

150

200

250

300

350

400

BH

P (

PS

I)

D

1997

FIG. 16. STEAM SATURATION 17/06/98 FIG. 17. STEAM SATURATION 18/12/98

FIG. 18 STEAM SATURATION 18/12/99 FIG. 19 STEAM SATURATION 18/12/2000