.

3 *

I Sodetyof Pstrobum Erl@wsrs SPE 28749 Production Performance of a Retrograde Gas Reservoir: A Case Study of the Arun Field
. . .. . .. . Deddy Afidick, N.J. Kaczorowski, and Srinivas Bette, Mobil UII maonesia inc. SPEMembers

Cqyright 1994, Sooiefyof PefrolaumEngineers,Inc. mh paper wss ~pared for prasenfekn at the SPEAsiaPsctfIc al 6 Geemnfefenos~~ in ~l~me, Auef@sI 7-$0 N~mbef l~.

ABSTRACT The Arun field is one of the worlds giant retrograde gas reservoirs.Approximately10 years after productionbegan, a significant 10ssin well productivityoccurred in some of the wells. The study showsthat thii productivity losswas due to naar wellbore condensate accumulation,and documentsits effects on productionand pressuretransientresponse. A radial, single well, compositionalmodel was used to study this effect and confirm that the productively loss was due to ~~ ~mmuia~on. The - IIluuul --A- I ----la- ,,a Lw plwrw .warlin+ +hm Wua Umu W+ed *A Uc=
future

apply for two phase conditionsof a gas condensate well, several semi-analytical and numerical methods were developed. Here, ourintension is notto present an exhaustive Iiirature review. However,the reader is referred to Chopra3 forsome referenceto priorwork. In ths paper we presentthe applicationof compositionalmodellingto pressuretransient response of wells affected by condensate dropout, and to predkt futurewell performance. The Arun field is one of the worlds giant retrograde gas reservoirs. Well test analyses indicated possible Iiiuid accumulation effects. Ths was confirmed with weii
pidtitiv+t~ @&S. A Uul -*A n Iw+nual, k-la haunr t-nmnndtinnal slnl~lw KaJv, -, , by. ..-..

performance of the well. The model matches well production date and the pressure transient response of affected wells.

Ths work identifiesnear wellbore condensate accumulation as an extremelyimpoftant factor to considerwhen predicting future well performance es some of the produdivitiesare reducedby 50%. The work also details how productiondata and well test analysis can be used to quantifythe effects of nearwellbore mndensete accumulationon well productivity.

model was used to verify that liquid accumulation would cause the same type of behavior obsenmd in the field. Subsequently, a multilayer compositionalmodel was usedto model a specificwell.

BACKGROUND The Arun field ~ located on the northern coast of Aceh Province in North Sumatra, Indonesia (Fiiure 1). Mobil operates the field, which began production in 1977. The averageresenfoir pressureand temperature were 7,100 psia and 352F at a datumelevationof 10,050 ft-ss. The reservoir isa thicklimestone formationwith a thicknessof over 1,000 ft in local areas and covers a productivearea of over 23,000 acres. The initial condensate to gas ratio (CGR) was 65 Bbl/MMscf at separator conditionsof 1,250 psia and 68eF. The field currentlyproduces 3.4 Bscf/day of separator gas from a total of 78 producers with an average resetvoir

INTRODUCTION The engineering aspectsof gas condensatewell performance have been a subject of research and developmentfor many years, Recognizing that classicalanalytical methods(such as A1-Husseiny, et. al.f and Govie#) for dry gas wells do not References at end ofpaper 73

PRODUCTION PERFORMANCEOF A RETROGRADE GAS RESERVOIR A CASE STUDY OF THE ARUN FIELD

SPE 28749

. ...................... ................................
s
:!,

WOlwolw2w025wm f%WaUre, @g

3500

40004W

Fiiure 2- Constantcompos#ionexpansbn
.A

Figure 1- Locat/on map pressureof 2,250 psia.

Akr inihl aas .is swmt . -. ..... . . sa~rstinn --~ .--- ... ------- via . .. oioeli~e ~-r --- -- to .- PT Arun. -... an LNG plant. Unstabilizedcondensate is also sent to the LNG

!I!m
2000

-----! ------ .--:-.---.--. --.:-.-----... --~-... -.--.------.--. --.--.:--------------...* ............~ ............ E.........+ ............1 .......................

plantforfurtharseperetion,Aside stream of separatorgas is sentto a fieldNGL plantwhere exkactionof LPG components is removed and sent to the LNG plant. The residue gas supplii field fuel, domestic sales, and injection. Gas injectionwas implemented as soon as fmld production began to accelerate liquid recovery. Currently25% of the produced gas is injected. The lean gas is injected on the pdphery of the resenfoir to sweep condensate rich gas towardsthe producers. This makes the Arun resewoir a composilionallydynamic system where retrograde condensation,water vaporization, and lean gas injectionaffect resenmirbehavior. A fluid sample was taken priorto production. Experimental data revealed the dew point pressure to be 4,400 psi. Retrograde behaviorwas determined es shown in F~ure 2. Gas began condensinget the dew pointand increasedwith lowerpreasuretoa maximumliquiddrqooutfillingabout 1.1% of the pore volume. Further reductionin pressure caused vaporizationof a small portionof the Iiiuid. Asdelkerability became more cdlicalto meet LNG contracts, deWrebility estimates became more important. To improve these @mates an intensive pressuretransientwell testing programbegan in 1989, et whiih time the resammir pressure had fallen below the initialdew point. By 1993 all walls were tested et least once.
A. tu@~sl . .,~ --Arm . . . .. -II .. . .. t~ ---@SKK)nW ~ ShOWI in F@!rQ 3. The --r-

n Woo -----

1900 .---l ..----

..l ............i ............{ ............y.........~ ............".......~;~;~;;

-----.
;o~

Fqure 3- Typh#Arun ptwssumtransienf response buildup period. Atypical log-log derivativecurve of the build up period is shown in Figure 4. The curve exhibits two different stabilization regbns which represent zones of

+ s

107

Ida tabiuzation

2ndsb~~ 106 10 I I 111111 I lo~ &;::

J:

: 0 TO

test consisted of three one-hour flow periodsfollowed by a

Figure4- T~&g-@d#va#vepW

74

SPE28749

D. AFIDICK, N. KACZOROWSKI, S. BEITE

differenteffectivegas permeabilii-tilckness products&h). In this case, ~h in the inner zone is lower than that of the outer zone. The most common explanationsfor thii type of behaviorare spherical flow or multi layer effects. With the resewoir pressure below the dew point, another possibilitywas the effects of ~quidaccumulationaround the well bore. Liquidaccumulation occumbecause produdng a well creates a relatMy lam pressuredrop in the vicinityof the wII. Gas migmting to the well originatesaway from the wII where the pressure is hiiher. This gas is in va~- liquide@*iti,rn at the higherpmasura. Asthe gas m&atea to the wII, pressure dec#ases andaarnafl fmctionoftbegascondanses closeto thewell. Thiscondensateis belovvthecrilicalliquidsaturation (SJ and does not flow. As more gas is produced,the small amount of gas which condenses beginsto accumulate until the critical ~quid saturation is reached. Condensate then flowsintothe wll as a liquidphase. The bank of condensate which accumulates aroundthe wII bore can be envisionedconceptually as shown in F~ure 5. Initially a small bank forms and is entirelybelow the crilkal saturation. tier, tie a~~a iiii~Wdi=A@y =~tind the W!! reachesthe ciUicalsaturationfollowed by a transitionzone of decreasing liquidsaturation. Eventually,when the resatir pressure reaches the point of maximum ~quiddropout,the transitionzone terminates at the maximum Iiiuid saturation (8A in the resewoir.

1.1% affects the gas relative permeabilii very little. Even whenthe maximumliquiddrop out is reached the gas relative permeabilii is 0.99. The flow of fluids in the reaavoir is affected VeIY Iiie with condensate dropout. However, the liquid accumulation aroundthe well severely restrictsthe flow of gas in the near wII region.

1
0.9

0.8 0.7 0.6 0.5 0.3

k 0.4

i 0.2 a 0.1 0

CMml Lkyid

I 1

1 1

f
0 0.2

0.4

0.6

0.8

Tow LiquidS8timbii

F~ure 6- Expenmenfalmlafiva panneabllliy dafa

CdUalLiquid ~ \ \ 8 ~ . . . J tl * \ *
G

affeting the wall tests, ProductivityIndex (PI) plots ware generated. PI is defined as the total wII stream (TWS) rate divided by the drawdown pressure. Psuedo-pressums (rri~~) ~i~ iised h calculating P1. Flowing bottom hole pressures (PJ were estimated from measured flowingwell head pressureswith composNonal tubinghyhzwtics. Interpolationof atatk bottom hole pressureswas used to estimate the resmoir pressure (p,). The PI plotfor an Arun well is shqvn in FUure 7.
!~ o~~~,
,

.~w A --

. i

Q Iid swmqnulaa&m
W-

----

-----

---

Y13S
-

tl<a<t2 \ \ . 0.38

..

-%-

Mmulnum Li@ompwt

Di8tanoa from

Well

0.14 33005100-47W -41OOWO37W

FUure 5- Typical cOt?cknsafa accumulaffonas a iimcfbn of

tfme
In May 1990 some experimental work was performed to etimata the mikal liquid saturation and the gas relative permeability (IQ Figure 6 shows the resultsobtainedfor a coresamplefromthe Arun field. The criticalliquidaaturaikm was 51% while the gas relative perrneabilii at the critical liquid saturation was 0.18. Notice that for the Arun fluid systemthe small amount of liquiddropoutin the reservoirof

Reservoir Prwwre, p8i8 FMure 7- PI of a t@cal Arun wall as a WcfiOn of fesendr pssure

A significant drop in well productivityoccurred when the Ilov#ngMom hda pressurewent balowthe dew point. Thii was considered strong evidence that the wII tests wre

75

affected by liquid accumulation. Sphedcel flow and multi layereffects are not affected by the dewpoint so no change in productivityshould occur. Only liquidaccumulationcan account for both the well test effects and the significantloss of productivity below the dewpoint.

1-

0.0. #

SINGLE WELL MODEL To confirmthat a well undergoingliquidaccumulationwould behavein the same manner observedin Arun wells, a single well, compositional,2dlmeneional (r-z coordinate system) simulation model was used. The effects of liquid water vapodzing intothe vapor phase because of the highreservoir temperature is included in the model (Sette and Heinemann?. The modelconeMadof a single76Wt layer of homogeneous prope- with11 radialcellsof varyingwidths. The innercell radiuswas 10 ft with subsequent celtegettinglarger. Figure 8 illustrates the cell dimensions along with the resewoir umwtiea. The well was mmdetad over the entire intervalto elitinate pwtial penetrationiffects.

0+ o

0.2

0.4

0.6

0.8

Total Llquld Saturation

Figure9- hbdelgas relative penneabiiUycurve

mat Ca8 . .. wK-,

-..2..-

i
. g

a
n

e230
7400

1 2 3 4 5

10 25 30 30 l!io

: * ~ well . i a G Homoganatxm a

The appliibility of this model for generating well test data was confirmedby generating a pressuretransienttest while the memoir pressurein the model was at 5,250 psi. Care was taken to ensure that the flowing bottom hole pressure remained above the dew point pressurethroughoutthe test. The pressure response was analyzed using a well test analysis SOflWareo Excellent agreement was obtained bekmenthe parametersfrom the analysisversusthose used inthe simulation model. A comparisonof the resultsis shown inF&ure 10. This confirmedthe applicabilityof the simulator to model pressure transient behavior. Notice that the dedvathre curvein Fuure 10 does not exhibitthe hump during the early#me asehwn bythe fieldtest (Figure 4). The hump .--AA -Ilk.. k. .H.d U#hiahI U99LSnn+ Qimi da+d cUpl 9UUI ILU Wullw! u *WI WV UBIWWL,WI FIWO Wau oowc-9 EIurakw
G

inthemdei. tirmatbn

8350

*m .,--1,700

a :

m0h088755n

#
9 @

l%r0aky15%
krmmhilh99 . - . ...--... --md ...-

10

og108 10

Figure 8- Sk@ layarmo# Initialwork with this model indicatedthat the experimentally &d d+. hilmu m, -Pa nd raa~ n4 -m -ea p0310may uu~wuu wwcu OIVU Iur-iiWuwu Wthe reaenmirand the fluid phases. Therefore, to accountfor the intetfacial properliea of the accumulated I&@ near the wellbore, the gas relattve permeabilitycurve was modlfiad slightly as shown in Fvure 9. The experimentallydefined critical liquid saturation was honored and the gas relative permeability at the crilical liquidsaturationwas increasedto 0.435. This forced the gas curveto be a straightline.

L=L==
...
G

El

Figure 10- Test hnsacclsnuidOn

kyar-i,

pdwto condensate

76

SPE28749

D. AFIDICK, N. KACZOROWSKI, S. BETTE

Usingthe model, a second well test was generated after a 3monthshutin periodat a resenroirpressureof 3,880 psi,well belowthadewpoint pressure. The generatadpressureprofile was analyzed analytically. The resultsare shown in Figure 11. The derivativecurve exhibitsthe same characteras that Qbaenmd from our field tests (Fiiure 4). Stabilizedregions developed depictingtwo regions;f dflerent ~h. 109

Both stabilized regionsof the derivativecurve matched with the transitionperiod between these regions matched fairly well. Howver, in some of the field teats, the transitionzone dd not match very well. Thm is a result of the simplified assumptionsused in the radml mmposite model. Fuure 12 illustratesgas relative permeabilityas a functionof distance from the well from the simulation model and that assumed in the radial composite model. The analytical model, whkh consistsof onlytwo regions,does not account forlhe transition fmrnthe inner zone with Sk to the outer zone with connate water saturation (Q or S, ~. Thus, the analyticallydetermined dedvativacurve reaches the second stabilizedregionsooner than the simulationmodel.

. loa F !gy

1
s c

1. >

--

106

L
.

2M8mumM

1
0.8 n

l?~1

~- -Eia --
I

. I I 111111. I [ 111111 I I 111111 . I I 111111 .

Q1 8
{

0.8

i 1=
Oietence from

0.4 +

Figure 11- Test hm sin@ layer model, al?arconchsata accumulation An analytical radial mmposite model was used to interpret this test. The interpretation results are very close to the Vaiuaa used in me simulation modei. Tine ratio ~M3 of me innerk to the outer k is .445 which is va?yclose to the relative gas permeabilii at the critical liquidsaturationused in the model. It is importantto note that gas relativeparmeabilii at critical liquid (~ at S~ saturation can be determined from the two stabilized regions of the derivative curve for the Awn fluid system. Ths isthe most importantfactor in determiningwell productivity10ss. Ths is so for the Arun system where ~ awayfromthe well is essentiallyunaffected by ~quiddropout. However, other flukt systems which have h~her maximum ~quki dropout can impact ~ away from the well. In these systems,the ratio of inner k and outer k representsthe ralio If M is available from core (M) of ~ at Se and ~ at S, ~ data,the effectof condensateaccumulationcan be estimated from the inflowequation for me radii composharnodei:

Well, It Wall

Figure 12 -km as a tincfion of dstanca hm h

SatMiedthatliquidaccumulationcan cause the characteristic behavior seen on Arun wII tests, .wII productivitywas generated as a function of reservok pressure. Figure 13 W@ratasthe reautk of the simulationmodel. When the wll pressure passed through the dew point productivitywas quickly and severely affected by liquid accumulation. PI

* 0.26 -.--.-..-..~-..-.--.-}..------*--.-{------..--{.-----------~-.-----.-.+ ?~;: a k 0.24 ......................... ............ ............ ....................... ............ ...........
i _. ~~~::i ............~.........~T...7 0.2. ............~ 0.16 G -&051004eoo ......~ ............+...........+............

P:-PM2=1422QZT(M( ln(r#)-3/4) + ln(r#~+Se+DQ)

.............. (1)

t
470046004200 41m am

I 3700

Rswmclr Preeeurs, PO& Figure 13 -PI of d?asingk Iayerrno&/

The analytical solution to the radwl composite model is supenimposadon the results of the simulationin Fwure 11.

77

PRODUCTION PERFORMANCE OF A RETROGRADE GAS RESERVOIR A CASE STUDY OF THE ARUN FIELD

SPE 28749

dropped from 0.341 to 0.166, a reductionof about 45%, as soon as condensate accumulated in just the firstcell. The first cell was filled up to Sk within a shott time due to its relatively small volume compared to the gas throughput. Productivitycontinuesto decline, reatilng a 50% reduction, as liquidaccumulates but at a much lower rate. To fwther &westigatethii rapid decline in Pl, the single layer modelwas run with the firstcell refinedto five l-ft cells. The resuttisshownin Fuure 14. PI draatkally dropswhen the limt l-ft cell was filled to ~. At the time, mndenaate had not started accumulating beyond the l-ft radhia. The declhing liquid saturation priortothe rapid accumulationof mndenaate is due to the water vaporizingintothe vapor phase.

immediately surroundhg the well reaches vapor ~quid aqwm withthe condensateat the highershut in pressure but does not change significantlywith time. Shuttingin the well doea not improvethe well productivity (FusseI~. To investigate if ~quid will revaporize at lower resmmir pressure,this model was depleted to a reservoirpressureof 500 peia. F~ure 15 shows liquidsaturationin the firstthree innercellsas a functionof reservoirpressure and the effect it The reductkn in oil saturation due to has on PI, ravapobtion occurredincell#3 long before cells #2 and #l. PI was notsfgnikantlyimproveduntilthe oil saturatkn in cell #1 was reduced. Again, this confirms that condensate accumulation immediately around the wellbore significantly affectswell productivity.

IX3!2

::: , ::: , ........... ........., 0.34 ......+..... 0.23

0.24
!! :: :: . . . . . . . . . . . . . . . . .. . .. . . ..!...... . . . . . . .
~, ;; :: ~~ . . . . . . . .. . . . . . . .. . . . . . .
-L:.

0.45
0.3 II

0.271 .....................

0.56

i .......

0.15

. .
503050004@34m042504m04730 Reaervolr 4700 -4ao

0
Preeeure, pie
0.15 i

-.~~m

.. .. .. . .... ..

. ... .................................. .
Coil#3-

-*= am m
l!!QO
1000

--+ 0.23

Reeervdr Preeeumpsie Figure14- EWofcondsnsate Wellbore accumulation14?anxmdthe Fiiure 15- EM of Iiquidnwapohdion on PI

This example illustrates a very small zone with liquid saturation at Sk significantlyaffects well P1. For a hUh rate, high CGR well, this small zone is filled to SEas soon as the flowingbottom hole pressurefelts below the dew point. The model with the orighal radml cells was atso used to invedig~e if the accumulated condensate revaporizeswhen the wetlisshutinfora longperiod. The model simulateda 2yearshutin petiodafterthe reaewoir pressurewas well below the dew pointwith the followingresulls on Satu ah Cell h .15 .15 .15

MULTI LAYER MODEL

To investigate the effect of multiple layers on both ~quid accumulationand well teat response, a * layer model was constructed. The radial dimensions are identical with the singlelayermodal. Figure16 illustratesthe thicknessof each layer along with the reeewoir prqwrties assumed. These _ were obtainedfrom a detailed geologk description in the region of an individualArun well. In general, POand permeabilitydecrease from top to bottom. The same relativepermeabilii curvesas in the single layer model was used for all layers. The modelshowd that the rate of condensate accumulation differed from layer to layer. The amount of ~qu~ the gas throughput. accumulation is influenced by hin&w narmea~ity aCCUmUlakd -...=, ,.* l=-- ~... . uu!rquucwJ, ,,. ... .. . ~-condensate and developed the inner zone of reduced ~ Mar than the low permeabilitylayers. Figure 17 showsthii

Year

0
1 2

Cell #l .512 .512 .512

Cell #l 1 .142 .142 .142

The resultsindicatethat ~quidsaturationin each cell remains OS* UUCCSIM u. , . =hlfi . ... . in ... =* ~-----~ There are several comtan itti* 9-*ar explanations for thii but the primary reason is that there is very little gas migration at the shut in well . The gas

78

SPE 28749

D. AFIDICK, N. KACZOROWSKI, S. BETTE

r
\_

-.--_:~-
-y :Well
h,ft 30.5 55 95 52 162 370 k,md Porosity ~ / ..---49.0 .214 :-- - --28.0 .22 :_ . . . . . ..--- 11.8 .209 ~ 17.1 .219 : 2.6 .127 ~ 1.5 .12 :

..
.-

D* K
o
G

108
3 0

..

u-.

.#ield
. . . . .

test

Leyer

1 2 3 4 5 6

107-

...

jo-t
Fgure 16-

~o-3

1-2 dm(P)(psi2/cp)

10-1 vs dt (hr)

100

101

Six-/ayer mo&/ Figure 18- Actual vs 64ayer mocH tests

clearly. The proportionof condensate accumulationin each layerisalmostidentical to the proportionof kh. The h~her kh layers accumulated condensate sliihtly lower than their kh proportion as these layers were more severely affected by condensate accumulation. Consequently,the proportionof gasthroughput inthese h~her kh layerswas curtailed. Thus, liquid accumulation has a normalizing affect on layered systemsaffecting high kh layers more than low kh layers.
30
25

_. n
5K1051004K)04700450 4X@ 41OO3XIO37OO

;20

Resewoir Pressure, psia

I!

15

10 5

Fiiure 19- PI of the 6-layer mod# vs. the actual Arun well the modeltothat generated from field data is shown if F~ure 19. An excellentmatch was obtained.

0 1
2 3 4 5 6

Layer

F~ure 17- Condensate accumuhtion in d..erent /ayers

FUTURE WORK wlm mesa results,we are confidentthat the model used in thii studyrepresentsthe actual Arun well. Some future work to be done based on hisstudyinclude : - Effect of lean gas injectionbreakthroughon condensate revaporization - Methodsof removingthe zone of condensate accumulation immediatelyaround the wellbore to improve productivity by injectingmisciblefluids
. . .... ..

A welltestwas generatedat a resewoir pressureof 3,660 psi. The calculated derivative curve is shown in F~ure 18 superimposed on the derivativecurve generated from field data. An excellent match to the two stabilized regionswas obtained. Eventhough,@x Iaym were used in this model, the generatedtransition betweenstabilizedregionsand the radius of the zone with~ dd not match the actual data. This is due to the limitationof usinga finite number of cells. A comparisonof the well productivity profilesgenerated with

79

PRODUCTlON ~ERF~RiifAhiCE OF A i?~~RoG~E ~AS RESEi?\~!R A CASE STUDY OF THE ARUN FIELD = average reservoirpressure, psia = flowingbottom hole pressure, psia = flowrate, Mscflday 2 r, = drainage rdus, ft ri = radiusof inner zone, ft = wellbore radius,ft = ctitical~quidsaturation 2, sInml = maximum liquiddropout = Skin factor at Q=O = connate water saturation > = compressibility factor v = w, CP P,

. .-.-

- Effect of condensate accumulationon a much leaner gas reservoirs

CONCLUSIONS I@uidaccumulationhas occurred in the Arun resewoir. Thw was identified through well teat interpretationand PI plots. This conclusionwas verifmdwith compositionalsimulation. Other conclusionsfrom thii study are: - Even with a fairly lean gas, liquidaccumulationreduced indtiual well productivities by about 50% - The mostdominantfactorwhichdetermines productivity 10ss is&at Sk - Forthe Arun fluid system, ~ at Sk can be determinedfrom well teat analysis. Sk cannot be determined. - The mostcriticalregion affecting productivity is immediately around the wellbore.

ACKNOWLEDGMENTS

The authomwouldlike to thank Pertamina-BPPKA and M Oil Indonesiafor their permissionand supportto publis paper. Also, Mobil Exploration and Producing Techn Center (MEPTEC) for their reviewand valuable comme

REFERENCES - The amountof accumulationis controlledpredominantlyby gas throughput. Consequently,zones of h~her kh contah the most liquidaccumulation. - The accumulated liquiddoes not re-vaporizeif the well is shut in for an extended period. - A radiil compositemodel can be used to analyze well tests. ~ of the innerand outer regionscan be determined butthe transitionregion cannot be modelled. - Condensate revaporizationbegins in zones away from the well. Productivity does not significantly improveuntil revaporizationbegins immediatelyaroundthe wellbore.

1.A1-Husaainy, R., Ramey, H.J.,Jr., and Crawford, P.B.: Flow of Real Gases through Porous Media: JPT (Ma 1966) 624-636; A1-Huaaainy, R. and Ramey, J.H., Jr Applicationof Real Gas Flow Theory to Well Testing Deliverability Forecasting: JPT (May 1966) 637-642

2. Govier,G.W. : Theory and Practice of the Testingof Wd/s, Energy Resources ConservationBoard,Albert (1975)

3. Chopra,A.K.: Transientand Steady-State Aspects of Condensate Well Performance, paper No. 88-3942 presentedat the 39th Annual Technical Meeting of th .- .-.-a Petroleum Soaety of cihi, Caigary, June 1z-lo, 1s85

NOMENCLATURES CGR E
h. .

= Condensate to Gas Ratio - = non-i3amy coeffderltt Ciay?vkfaf

= fQrma~Qn ticcknes

4. Bette,S., Heinemann, R.F.: CompositionalModellin HighTemperatureGas-condensate ReservoirswithW Vaporization; paper SPE 16422 presented at SPE Fahrlm ---:.. - UI ~mn--a~,-;=-, il=timm bloli~tmq symlpUalulll I Ix=-l Vull Wll I Bulauvl 1, I c U-*U! , . .
6-8

k * k ~~G LPG M m(P) NGL PI

= permeability = effective gas permeabilii =&at Sk = r~!div~ gas *rmeabiiity = Liqu~ed Natural Gas = Lquified Petroleum Gas khlllefZO#OUbr. = gas pseudo pressure = Natural Gas Liquid = Productivity Index

l 1989 ----

5. Fuaaell, D.D.: Single-Well Performance Predictions Gas Condensate Reservoirav JPT (July 1973) 660-87