Вы находитесь на странице: 1из 12

q

SPE
SPE 16959 Material Balance Calculations for Solution-Gas-Drive Reservoirs With Gravity Segregation
by A.K. Ambastha SPE Members and K. Aziz, Stanford U.

CopyrIgtN

1987. Semaly 01 Petroleum

Engineers at the 62nd Annual Techmcal Conference and ExhMion ot the SOClely of pelfoleum Engineers held m

This paper was prepared 101 preaanlatlon Dallas, TX September 27-SO. 1987

This paper was selecled for preeenlalion by an SPE Program Commmee foilowmg review of mtormaoon conlamed m an abstract subm,lted by Ihe author(s) Conlenls of the paper. as presented. have not been reviewed by the SO@y of pelrOIeum E ngmeefs and are aubfecl to correction by the autnoqsl The mstenel. es presented. does not naceaeenly reftect any POSIIIOI of the Smety ot Petroteum E ogmaers. IIS ofkers. or members Papers preeented @ SPE meetmga em subfacs to pubhcetw review by Ec!ttonat CommdSeea et the Scscmty of Peooteum Engmeare Penmesuon to copy IS reetwtad to an ebetrect ot not mcae than W words Iltustmtms may not be COPA T@ ebetmct shoutd contain conspcuoue acknWMgment ot Telex. KW6$t SPEDAL R@~rd~. TX ~ send by whom the paptM IS pmsanted Wrjte Pubkceowa Mawtaf. WE p o SOX ~. e.

AB!YflMcT

2nthiswork, areseMlrai tnu2atoria UsedtttatudyLhe ~ofww~~mviw-~= effects on reaenfoir performances. Tbs titivity study of tcsenfoi. pcrtbnnance to btockand titnc-stepai=tiwa that fortypieat Solution-gasdrive rWer@ra, large erlurain mwragereaervoirp aurcmay reauttfrom impmpercOmIol Oftime-atep inthesimukstor. --flE eunparison of aimtdatittn results with ItE Tfrmer1 method ahowathat thelatter pmdicta fsatermaervoir pmsaure andoilsatumtion decline, and thus. mderpdets the msmoir producing life and tecovery. We propose a new material balance methd for predkting the perfonnanm of thick, homogeneous, depletion-dnve reservoirs. this method accounts for the vertical pressure and saturation gradknts, and the secondary gas cap, The thickness of the srmndary gas cap cart be estimated with good accuracy using an ideslizuf saturation profile. An iterative prucedure mAues avetage reservoir prrs me to well pmsum. The pnxsure and saturation at the well am then used to calculate the producing gas-oil ratio. The idetized satutatiorr profile is also used to develop pseudo-functions to simplify he simulation of solution-gss-dnve reservoir with gravity segreg~tion,

A numb of studies-= of solution-gaadrive Ie6ewoirs under different dtions have been pubkiahed. Different col@itioM thst arise during* exploitation of @ resemtira are: 1. Mernal g8a drive =chanbL vohJrnctiG ~ n?servoim produce by Reservoirfluid expansion. fheptoduu tionisatuibuted totiquid e4Mnsion andtorOck e0tnpreSaibUity, 56 * reservoir preaaum dmpadown to the bubbkpoint P==u= Ask reaWvoir pre$Wre declitl= furtheL oil@= eontracta becauae of themtcaae ofaolution ga%andtircpnJduction is due to g~ expansion. As gas saturation rcaehea the critical value, free gas begins to now, tealdting in high gas-oil ratios and low oil recovetiea. 2. External gsa drive meehaniarm h many imxancca, reservoir pseasure in solution-gasdrive reservoir+ is maintained by gas injection, and oil is displaced by injected gas, This is mferred to as external gas drive mechanism. 3. Gravity aegrtgation (or gravity drsinage). For high relief reservoirs with good along-dip permeab~iy, favorable contk tions exist for gravity segregation of injected gas or gas released from solution. Gravity segregation is an important factor in attaining high oil twtvery from solution-gasdrive reservoirs. 011 re.cmveriesof the enter of 60 to 80% can M obtaiid with effective gravity segregation.*Z Reservoirs with significant gravity segregation also ahow low producing gas-oil ratios in structurally lower wells. Crqft and Huwkitt# describe these drive mechanisms in more detail. Tamer and Mush? proposed methods to predict the performance of depletion (solution-gas)dnve teservoira under internal gas drive mechanism, using rock and fluid properties. lM assumptions of both methods include negligible gravity segregation forces. 2hu6. these authors considered only thin, horizontal reservoirs. Both methods use the material balance principle (atadc) and a producing gas-oil ratio equation (dynamic) to Predict reservoir performance at pressutes, where gas saturation excezds the critical value. A mors detailed description of both methods appears in Crqft and HMVMIIS.M The points to note about these prediction techniques are: 1. Tme is not a factor in Ike methods because neither watel influx rmr gravity segregation are considered. Time history must be inferred horn the resenfes and well production rates.a

2NTRODUCT20N while it is usually pssible to do a detailed (and expensive!) Uueedlmensional, muhiblock reservoir simulation study to make predictions for soluiion-gssdnve reservoirs, it is otlen instructive to first do some simple material balance ealctdsticms to make ball-park projections. This is especially impott.arit because simulation of solution-gasdrive reservoirs involving gas percolation and gravity ae~gation is a numerically difficult problem that can gobble up large amounts of computer time. The evolution of solution gas and its rapid movement to the top of the reservoir is both the source of numeriesl dKticulties and the justification for assuming that gas movement is essentially hawtarw us in performance predction cslcuhtions using material balance. The performance predktion of a hydnxatbon reservoir under different drive mechanisms is of interest to any practicing restwoir Referencaa and illustrations at eod of psper.

259

MATEJUALBALANCB Calculations m SOL~ONQAS-DRIVE lU3SERVOlRSWIIM ORAVRY SEOREOATION

-$ S?81!

Thus.hepredictod t perfosmsnce ofsnintraid gsa-iivcreaerVoir is independent of Oifproduction rate. 2. Predicted frsctionaf recovery fiotn intemaf gas drive re9ervoir is independent of sixe. Thus. reservoir performance calctslstions can be based on an initisf content of one stock tank bar~.24.2J 3. Absolute rcsmoir pcrmeabdity is not a factor in USC predicted performance of an internal gas drive reservoir. Both methods assume uniform prcssum? and saturation throughout the reservoir (tank type model). That is, the whole reservoir is tr?ated as a singfe block from the resmoir simulation standpoint. We compate the mwsftsof single block simulation with those of Tarnersi methed later. To improve the mlitillity of predictions, scveraf investigators have tried to remove the assumption of uniform pnxsure and saturation throughout the reservoir in matenaf bafance calculations. Loper and Cafho@ relaxed ths aasusnption of uniform pmssum and satumtion in the reservoir for the Tarriers* method by using two-phase steady state. radiaf flow theoty. E1-Khatib4presented a modified Tamer method to predkt the performance of depletiondrive oil ~rvoirs by using the average pressure and saturation to estima~ well pmsure and saturation, assuming the rescwoir to be in pacudoateady state. He then used rock and fluid pmpetties comspmdhg to Welf pteaaure and saturation to C.slcldae * pmducii gas-oil rario. Ef-KhaiiP also pmsesttsd the effects of dndnsge srca and production rate on the msesvoir performance using his pmdktirst method. Rfilfer et al. employed tluid flow and material balance analysis to a series of succusive atedy state conditions in the resewoir. They divided the reservoir into a nurnhr of annular rings awmunding a welf. The effects of production rate, well spacing, and sock and fluid pmpties on the pcrfom2smccof solution-gasdrive reservoirs have also been considered by aeveraf invcatigstors.*M AU of these studka deal with thin horizontal reservoirs whcsc gravity effects are negligibk. Our mmMkation of the Turners method, to lx discussed later, is similar to thatf E1-Khurib/4but we account for the effect of gravity aegteo gation. Scvetaf invcsr gatomllsl-m have propmed methods for pcrformarw? prediction under extemaf gas drive mechanism with gravity segregation. Tetwilliger er aL mcrstion that recovery by gravity drainage is rate sensitive, and that rather sham dccrcasc in recovety would occur at production rates above the maximum rate of gravity drainage. Hence, this maximum rate should not be exceeded. They defined the maximum rate of gravity drainage as the production rate from a ltKs% liquid satumted system under a flow gradient equal to the gravity gradient or static pressure gradient diffcrcnsiaf bctwccn oif and gas due to dcnsi[y difference. fWs maximum rate of gravity dminagc, assuming negligible resistance to flow of gas and negligible capillary effects, is given by:n.3$ 7.83x10-6 k krOA Ap sin (a)
f?o = lb (1)

~He=-dcumpkkawgstimofmobik M Dybmi analyxed free-fail gravity dninage system where tb psoduction rstewas soblydetemtincd by gravity effects. Heemasdad Snapproximste theory of free-fall gmfky dmbssgepmesxedby Cardwefl andhrsons.mllw-m preaemdfora~presaute system, where gasiainjezre4fatt& mpofthestructmtn offset the void that occurs because of oil production, S0s22S investigators have also focuawd on pert&ttsance prediction of intemaf gas drive tcservoirs with gravity segtcgath Bzwrchueftconaidescd the effect of grsvity upas the reservoir perfonname of a high mliif pool, He outlined a pmcedttre tbr strucaxsUy weighting rock and fluid properties for u in the msterisd bslance equation. He used actuaf swervoir performance data to ptedict the future pe.rfonnsncc of a gravity five rcsctvoir. He stated that when controlled, produced gas-oil ratios wilf be low and wilf dscrcase rather than incm.se with pressure declii in reservoirs with gravity wgmgation. His statement was based on past field perfbrmsnce of high relief maervoirs. Cookts analytically studied the effkcta of depletion rate, reservoir geometry and withdrawal distribution on natural depletion. He considered distributed and segregated flow aystema. Irr the distributed flow study, thm was no vertical penneabilii and gas could only now along h dip d-on. Wlthdrawsb wem uniform SloQg the metvoirauc hthsttbmwsasmtotal thddsstigratkxt atanypczIt& Distrihlted ffow SyateS2sshowed two astWadmftonts snddscperi&rnsnce was rate sensitive. However, for segregated Ilow with vcsticd wssmuch ksarstesemdtive. CooPrn Pm@@, Wf~ thatweUs in Wcstem Vemstsdaand O@f COsthsve@muntb~ ducizsg mmtbbehsvior, andevenabw vesticalpam@sB& (CooPsused 10md)is aufEdem fhrgsssegregadm Maahswamd -~ak== Pf@=m=Umdfor de@atbm dlivese.servoir inthestripper stage using stesdysWeanalysk Inthe atripper stage, tbreservoir prcmre Isso Iowt!sst tlwgravityisthe sobdrivii fome+snd ttsus, thepr&cdon rateisqtdtebw. H*? Ouuineaa methodt opredictt hcperfomssnce of gravity drains& RSSZVOim with tklhbtg prssaum. He mpmsems ths ~tvti by a serieaof blocka, writrxmatcnal balsrsx eqwions forgsssnd oilin each block, and uses Darcys law to dtibe fluid flow in and out of blocks. He calculates rate of advance of gas-off contact using Shrew and We&h method. and describes an iterative procedure to predict the reservoir pcrfomsance. With the easy availab~ty of modem reservoir simulators, there appears to be no justification for the use of such comp~catcd and yet restrictive techniques. in tfds study, a nusnencaf simulator is used to study gas percolation and gravity segregation, and their effects on reservoir performances. Figwe 1 shows the grid arrangement for one- and two- dimensional simulations. For onedimensional verticaf simulation, the memoir is divided into NZ number of blocks in the verticaf (z) direction (Fig. la). For twodimcnsionaf simulation, the nxervoir is divided into NZ and NX number of blocks in the verricaf (z) and the horizontal (x) directions, nxpcctively, as shown in Fig, lb. The sensitivity of memoir performance to model, mervoir and operational parameters is repotted. A simple materiaf balance method for predicting the performance of thick, homogeneous depletiondsive reservoirs is proposed. A by-product of this work is a set of pseudo-functions that can be used in reswvoir simulation to study tfueedimensionaf flow problems with two-dimensional sreaf grids, Since gravity segregation or gas percolation is the dominant mechanism in the type of reservoirs bdng considered, a brief review of the literature dealing with the simulation of gas percolation along with a sensitivity study is pRsentcd next. SIMULATIONOF GAS PERCOLATION Highfy nonfincar finite-difference equations arise in a simulation of counter-cument flow due to gravity segregation. TM problem k espcciafly pronounced in thick pinnacle secfs or biohenns,s

In Eq. (l), qOis oil production rate in m. bbl/day, k is absolute permeability in md, A is cross-scctionaf area open to flow n ft2, Ap is the density difference in lb/f[3 bctw~en Ihe oil and p= phases at rcsmoir conditions, a is the dip angle in dcgrccs, and BOis the oil viscosity in cp, Afcf%rd2,and Shreve and Welch used Darcys law with gravity component and materiaf balance concept to predict h performance of solution-gas-drive rcscmoits incorporating gravity drainage and gas cap pressure maintenance. In both studies, the reservoir was divided into scvcmf blocks in the vertical direction. Martin prcsemed a method, based on the relations governing the average velocity of flow of each phase and the equations of continuity, to describe the memoir performance under prcsssuwmaintcnamx 2

SPE 16939

ANILK. AMBA!WHAANDKHALIDA22Z

Gaaprcdatitascans cvemlylimi tthatime-atepsim thatcssrbe uaedtoairn@atc suchprobkma, ifthettWsmkaRWti asetseated explicitly. Tlsia problem was recognized by Coars,w and McCreary.* fWy suggested approximate methods so that airmdatiorrcould cOntinue with large time-step size and without any numerical instability problems. Later, Send and Aziz41showed that U2ebes way to solve gas percolation problem is by implicit treatment of transrnissibiitiea. lWr Teat Pmbiem No. 2 is the same as that reported by McCfeaty.~ Azu and Setkv~2 discuss the comparison of restths from different imPUcit ~methods for this problem. In this study. wc usc fUUY ment of transmissibflitim. To understand the mechanisms to be incorporated in the material balarxx csfcrdations. wc simulate the solution-gasdrive rmwvoir deacnbcd by AfcCreay.m Saturation and pmsurc dependent properties given in Tables 1 and 2 of Ref. 40 were used along with the additional data in Table 1 of this paper. We usc a block centered grid scheme and the closed outer boundaries for both one- and two- dlmensionaf simulations. it4cCreary@ problem is referred to as Example 1 in the following.

fromdm$e for bktck*of9m 4.sft. Tkrefos&8bloc& sizCof 9ftisc4maidered adequam tostudythc effecta oftime-atepaizeon theaimulation reaulta (Fig& 4x2ds). ttEpmasalm pmtUescakulatoduaing diffetcntt irste-atepsizcsazaquitediffe~ whereas the aaturation protile aarecloaei ygrouped. -fbcpmaausc pso61eain Rg.4 atmwthat thcuseofa small consmnt dn2e-stsp aiz#detaya thealtaisF mentofbubble poimpreaauminali Mockrha mlwouldrcault ina higher average memoir pmsauze at any time compared to large time-step size rum. Also, the saturation pcdiiea am mom sensitive to block size, and pressure profiles to time-step size. These results show that while the model is stable for large time steps, relhble results can onfy be obtained by careful choke of time-step and block size. The solid lii~s on Figs. 2 and 3 show idea&cd pressure and saturation protiles. tWse idealized profdes arc used to morMy the Tarnersl method in the next section. l12e idctdii satmstion profile is also used in the A-lx to develop a set of pseudo-fimciions to simplify the simulation of solution-gas-drive reae2voim with gravity segregation. lWae pseudo-functions can be used to nhrce tk dimensiotilty of the simulation problem for sohstion-gaa-dsivereservoirs by one. Figures 6 and 7 show a comparison of computed reset voir performances for several constant time-step sise nma with block aiae of 9fL Average zwservoirfmaureand oilsaturadon arocomputed aa tiSCV@ltSS2CtSiC~ Figs2re6ahow8that theaimrdation uPtothe time, whenthe bubbtepointp scaaumiacmaaed inaUb&&setarlre tmndfoz thcfhture behavior intkscconamst time-s&paizcrtms. Even though average oil saturatha arecloaely grouped fordifkztt runs @g. 7) as eFpecaed fmm Fig. 4, the average ,eoir presama (Fig. 6)azeser&ive totime-step aiae. Tlse ZCS&SWith8@O12@C time-step control arealao shovmon Fl@.6arsd7. Fortleautomdc rime-step contmlrrm, t%oh~ar xi% arcsetto lday, l,30paiarxf 0.1. reapeaively, in Eq. (3). Aa expcctul, conatam time-amp run withlOdaya time-step sizeappmc&a thezestdta from automatic time-step Contmf run. Ilmugh M aimwll graphically, chan@W ~ between o.5and2, ~bctwcCSs20and50 pai, mbetweeno.loand 0.20. and &} between 1 and 10 days does not change the simulation results. These restdts imply that the simulation should be started with a small time-step aim, and increased later, when the trends for average memoir pressure and oif saturation have been established. We have used automatic time-step control parameters within the indcated range for all subsequent oncdimensionai nsns reponed in this study.

Simulation Approach ASI ncdimensiorral simulation model was devefoped based on o the simrdtaneoua solution technique with fully impticit treatment of tzansmissibilkis. In this model, gas pressure. p,, and oil saturation. S., am the primary variables, and the noniinear equations art solved by Newtons method. An upstm.am weighting for relative pcsmeabitidea and an arithmetic averaging for pressure depmdem terms are used. The derivatives of dtfferent terms with respect to saturation or pressure are calculated numericaf.fyover saturation or pressure change of 0.0005 or 10 psi, respectively. Free gas production starts when gas saturation in tk WCU block exceeds critical gas saturation. Free gas flow rate is calculated by: (2) The variables T, and TOin Eq. (2) are the gas and oil transmissibilities. rcspec[ively, in the well block at the currwrt time. The simulation model can be run using a constant time-step size or a simplified version of automatic time-step control method used by Grabowski e! al 43 in their fully implicit generaJ purpose thermal model (lSCONl). The time-step sequence is cafculatcd by

Solid lines on Figs. 6 and 7 are the predictions from tk Tarriers mcrhod using the material balance fores proposed by Tra~ 2$ F@.f.s (j ~d 7 dso ShOW fht b Si2nUlSdon ~.dh wi(h automatic time-step control match the calculations fmm the Tarnersl (3) method to a time by which average oil saturation has dipped slightly below O-SW-SJ. Mobile 8ss. that has high mo~fiy tO P~ola@ is present in the mervoir beyond this time, causing slower avenge The variable to is an empirical paranrctcr, and 5, and q, arc the actual reservoir pressure and oil saturation declines compared to the prcdie and spcciticd changes in the primary variable i, respectively, at USC tions from the Tarnersl method. 011 saturation of approximately current lime step. Two-dimensional simulations were carried out us(I-sin-s=) in the lower portion of the reservoir @lg. 3) irsdicatesthat ing the ECLIPSE simulator.~ Results of one-dimensional runs on mobile gas travels to the top portion of the reservoir instantaneously. ECLIPSE matched the results fmm our one.dimcnsionid simulator. Thus, a high gas saturation region develops in the top portion of the rescffoir, which may be termed secondary gas cap. Figure 2 shows that the pnxsure is uniform in the secondary sss cap at a Sensitivity of Simulation Results given time. llw development of a secondary gas cap causes a slower Figures 2 and 3 show the caicufatcd pressure and saturation average rcacrvoir pmssssre decline. flsc Tamers method profiles at 900 days for runs with constant time-step sin of 10 days significantly underptilcts ttactvoir producing life and cumulative oil for Example 1. The results arc shown for different biock sizes. The recovery because the Tarnersl method does not consider gravity pnxsurc profiles am identical for block sizes of 9 and 4.5 ft. The effects. Afso, the Turners*method considers the whole reservoir as corresponding saturaticm profiles are identical, except for slight one block. FIgurcs 8 and 9 show the results of one block simulation differences for a fcw blocks close to the top of the memoir. Block for Example 1 with automatic time-step cormol. As expected, the sixes larger than 9 ft also produce comparable pressure pmtilcs. re.subs from the Tarnersl method compare favorab!y with those from However, saturation pmfrles for block sizes larger than 9 ft differ single block simulation. 261

* MA7ZRIAL BALANCS CAIXUIATIONS m REs8Rvuslts Wmi GRAVITY SmR%GAmm soLuTloNuASDRIVB s!ep I modified Tamer method are Wkightofseumduy (l-s=Zc =

. .

SB 16%9

.: 7

MGDIFIEDTARNERMRHGD lheaaaumptions oftbepmposed

gaacap, ?,is~by SF) - (r&_ Ssc) - (QG ~

1. An omdimenaional kar resesvoir geomemy with gas perLWhlting pwatd Ss%l il moving downward, U o 2. Uniform reservoir porosity, 3. Thermodynamic equilibrium between the oil and gas phases at all time% 4. Gss liberation meehanism in the resmoir the same as that used to determine the fluid properties, 5. Uniform absolute penneabilii everywhere in the reservoir to avoid pockets of tzapped gas, 6. A linear saturation protik in secondary gas cap ~tg. 3). 7. No water enerosehtnent and negligible water pmductiom 8. Negl@ible capillary effeets, 9. Uniform cKSSS-seetional mea open to flow everywhere in the memoir, 10. Production wells in the bottom part of the reservoir, and 11. No gas injection for pmsaurc maintenance operations. Bssedon rhesimulation result& three stages ofdepletion ofa sohttion-gasdrive memoir with gravity segregation atw resemir iaabovetrubbk point l. Freasum everywhminth pesaum. fhereisno&ee gaain*teaemoir and average Oilsaturation istksatnea stheinitiai Oilsssumtiom 2. PsWaureinto pportionofth esesenmirisbelo wbubbkpoint fxeaswe, but pressure in botzotn portion is still above bubble point pressure. Atthisstage ofde-thereismobik gssintoppostiom htnotint kbottosnp ortionoftk~voir. Seeondary gas cap h= started forming. The average reservoir preaswe can be anywhere between al@sUy lower and higher than the bubble point pressure. 3. Fressure everywhere in the memoir is below bubble point pressure. Thus, during this stage, the average rtservoir pressure is lower than bubble point pressure. There is free gas everywhere in the reservoir. i& oil saturation at the well falls below the initial oil saturation and stabilizes at (l-S#F-). The term e is a small quantity. We use c = ld. During this stage, the average oil saturation is below the oil saturation at the weU. During the first and seeond stages of depletion, the oil satutation at the weU is the same as the initial oil saturation, (l-S=). During these stages, the conventional form of Tamerst method is sdequate for performance predictions. During the third stage of depletion when the average oil saturation first qusls or falls below (l-S=-S,.), we propose the use of a modified Tamer method. The solid lines of Figs. 2 and 3 show the idealized pressure and saturation profiles at any time in the thki stage of depiction. Figurt 10 shows a schematic diagmrn illustrating the idealized p~sure and saturation profiles along with several variables to be used in the development of s modified Tamer methed. The variable u is the distsnee between the well and the reservoir bottom. The variable F represents the dwtance from the reservoir bottom to the @nt where thz reservoir ptessum. equals average reservoir pressure calculated from the Turnsrs) method, hllw hdgltt of the aeeondary gas cap is 26. The vansbles ~, pw, and p,. represent the ptesaures as the reservoir txmom, at the well, and in the seeondary gas cap, respectively. A stepby-step development of a .nodified Tamer methed is now presented. 262

(l-s=-

(4)

where (r&_ is ti average oil Saturadon calculated from the 7brnersl method. Ilw average oil asturadon in the secdary gas e.ap,(YO),C, is: (l-sw-sr)+sa (rJk = 2

(5)

Equation (5) is a direct mmqumee sumption in the aeeondsry gas cap.

of a linear saturation pdile as-

Stap II Giiaamrasion atthewellboreitc Sa= l-SW-SF-e (6) average raar-

llteeatimation of thewenpresaW& ~uaingthe voir pressure, ~involves an itemdve ~~~following (i)

cakukteoilpl wauregradk ntp./144pa i/ftatthepmasura h-. (ii) CaleulstewelI psesmuc, ~andthepresaue attbreaervoir bottom, ;, ~. F.
pw=~Tamcr+~

[1 PO

-2.

j=fim+~

(8)

For the first iteration, we use ~= hR or ~ fmm the txdeulation at ~ previous time level, if any. fiii) Stabilized pressure in the secondary gas cap is

p,=~w-ao-zc-z)

= ~Twntr

-%[-4

(9)

(iv) The csletdated average reservoir pressure based on ti idealized pressure protile is ~c+; z h

p,, Zc+ j&=

!1

(h - Z=) (lo)

SPE 16939

ANILILAMBAsfHA AND KHALIDAZZZ

(v) If&_ -&tiaksarhaa aaped&d tolermminpaLtkn a convergence has been reached Ctthewise, assume:

&M =%GU

1)

and re~ steps (ii) thlolrgh (v) Untif convergence. we We used a tolerance of 20 psi for prcssute convergence in moat cases.

19*atd20tbm!ugh araapccdvely. Tha ori@na12#lrne##mateod performc!d pooriyfor these exm@eaalao. F4puea17 through22 showm$eWoir psrformsIlos bsf&egaa@akdu@L %W . . dmlelmd (s-s) simulation weuaesgrid blocks iatba Amction AmcticQ fkcme4medor@ verdcal :2s!! Y&d& The Weu is Compkted it! * 27th block fiomthetop. Figun?a 17through 22ahowthat fbrbothcxam* 1. TIE computed performances from a twodimmsional (x-x) Simldationdonotchange aathevmllisahifted hmhntauy. 2. m computed perfmmances tlom twodimenskmal *Z) and Onedimenaioaal vertical (z) Silmdadons ate the same. 11111% the Simuladon of atdudon-gaadrive reaenfoirs with gravity segregation is easendauy an mdimenaiortal psobkm from the standpoint of average reservoir perfmmance. 3. MOditkd Tamer method dSfSCOtily fSS!diCtS lCSSSVOiS the perRnmarrxa with8mdnnane norofabottt3% inthepm-

Step III Oil sanmstion at the wellbcxe, S- and weltbwe pressure, k are used to evaluate rock and fluid properties to estimate the producing gas-oil ratio by:

=R-+[4WHW

12)

~is completes the calculations for a time step using the pmpoaed moWkd Tamr methd. Vdldakn of Modified Tarrier Method

Like the original Tarnerd method, modi6ed Tarnsr method doesnot useabsolute venical permeability andoilpmductio nrate. TIIS simulated reservoir performance wfdmk~ Ofabsolute vertical permeawty andod producdon rate(Fii.11 through 13) for Exampkl, pmvidedtheoU pmdudonratedocanot exc@ maximum rate of gravity drabge given by Eq. (1). FigurealithrOugh 13 C0r0pamt&pluUcted prlbrmanafmmtk moditkd Tenmr metbl wkh omdbsuional simulation ~ta for Example 1. ~ moditied T-r method aatiafactorUy predkta the perfonnsIIc4 of solution-gas-drive reservoir of Example 1. MtrdMed Tiuner method is mt ap@cablq if gas bmsks through into the wefl blcck and hence, mducca oif saturation in the weU block significantly below (l-S=-SW). Under rhc condWns of gas bmtcthmugh, the producing gas-oil ratio will increasedramatically as opposed to declining produced gas-oil ratio in the absence of gas breakthrough on Fig. 13. The declining produced gas-oil ratio on Fig. 13 is in agreement with Burrclmeffs statement that under controlled conditions, produced gas-oif rw:o may be low, and decline with pressure decline in ~servoirs with gravity segregation. Two more examples are considered for funher validation of modifitxl Tamer method. Example 2 uses saturation and pm.ssure dependent propemies given in Tables 1 and 2 of Ref. 39. Additional data for Example 2 is providexl in Table 2. Example 3 uses saturstion dependent properties as given in Table 2 of Ref. 39 and PVT properdes as given in Table 1 of Ref. 46, Additional data for Ecample 3 am the same as in Table 2, except the initial pmsutt which is given by gravity equilibrium with oil pressure of 4550 psia at 5 ft from the top of the column. The two sets of relative permeability cuwes used in this study am presented in Fig. 14. T?Msolution gaaoil ratio (GOR) and oil formation volume factor (FVF) of the three sets of PVT properties are presented in Figs. 15 and 16. Normalized preasrne in Figs. 15 and 16 is obmined by dividing the pressure by bubble point pressure for each PVT data set. Cwves A, B, ad C Rfer 10 the PVT data sets of McCre~,* Coars,s and A1-Khal@h et uf.,a respectively. The bubble point presatm for curves A, B, and C ~e 1275, 1644, and 4514.7 psi, respectively. I%e simulated perfonnance afong with the performance pmdictcd from modified Tamer method for the Examples 2 and 3 are presented on Figs. 17 through

FGz=?z:: lute vertical penneabilky and OUproduction rate., if the oil prdcdon raredoeamc excecdmaxitnm~tyu rate given by Eq. (l). Sii Od~ratsa uaedinthe Exampka 2amt3 are below "maximum gravity drabageraw givatby Eq. (lJ&aandon 4iaeaauyjuatilkd obaerv* !and2aIc expMnedby Pig. 23. Figum23showatheofJ aawadon atsfttblth smaervdrtopfbr Exmnple 3atcJamwveoit mcOvcryof14. 72%. ou~ ploftka an!shown fordifb@ wall bc#iwla for HabmMbU, Figurc23alsD ahowadIeoil aaura&mrn5 fkfFosntbe rueNoirsop tbronedimedod venicaI (z)afmuiadlxL Eventhtmgh thsoil aaturatkm pmfueaam difkanL avemgeo$l Utumdon &sq@w&wdy thLssme lkwaucaau. Though notahDwn gr@UUy, tMs~ tionistrtle forthepmalra protuaaaiao. Ills satwdonmdprcaawa pm61eaat diffelerElo@orla andtfmea ahowaimuarbelwior. nlat is why, * reservoir performance from x-z and z aimuhtiom? are ~ same befors gas bakthmugh. Thiadiscussion alsoappuesto Example 2. Implications of Idealised Saturation Protile F@we 24 shows how oif saturation at 5 ft from the Icservoir top varies with cumulative oil recovety (or time) for the thee exampies. For Examples 2 and 3, residual oil saturation is 0.15. For Example 1 residual oil saturation is 0,3895. OU saturation at tfw top portions of the reservoir approaches residual oil aatmtion faster for Example 1 than Examples 2 and 3 because of b steeper oif relative permeability curve mar residual oil saturation for Example 1 (Hg. 14). 011 saturation at 5 ft from the top does not reach residuaf oil satumtion at any time for Examples 2 and 3, But moditkd Tamer method satisfactorily predkts the reservoir perfonnana assundng a linear saturation profUe with S. at the top and (l-SW+J at Z. in the secondary gas cap. Since modilled Tamer method satisfactorily predicts the mservoir prfonnance with an inaccurate saturation profUe, it may also be possible to predict the resefvoir performance just as accurstety, aasuming complete segregation with S- from the resewoir top to 2. and (1-S6-$=-s) fmm ZCto the reservoir bottom. ModMed Tar&r method was used for the three examples with 13Jq = S- lhe mstdts from assuming ~,),, = Se wete identical with UMmaulta from aaauming ~,),, given by Eq. (S), except for aligtu ovew%ietations of producing gas+il ratio late in reservoir tife by assuming CJ,C. Sm lltus, the assumption of incomplete or complete gravity segregation does not affect calculated average reservoir performance

263

MATERIAL BALANCECAlzuLAf10N8 Rx SOLUTIO?WASDRIWRESERVOlR8 WITH GRAVHY SEGREGATR3N NOMENCLATURE A= B. = Bt = c, = 8 . gc = h= k= kti = kmt = K* = m. ~.


p.

SPB 169s9

Signfflcsmtly.vertical hrekky-Levereu theory can approximate %etual aaturadon profile most closely. But a signitlcant isnpaet on calculated reservoir performance is not expeued. However. the time to is ti~ by th; UIOVCSIHli Of SCCOndSryas g gSS ~Sk-U@ cap with time, and may depend snore ssmngly on the shape of mrmmedaaturadon profile. In the simulation suns, gas breaks through etiler for a=z simulation for a welf in the end block (127) or (527) compared to a welf in the center block (327), as expected. G= breakthrough for onedimensional vertical (z) simulation occurs after gas breakthmugfr for a wefl in the center block for two-&mensiorsafsimulation This is because at any time, gas is more Unifomy distributed in the top portions of the reservoir for an one-dimensionaf vertical simrdstion than a two-dimensional simulation. Producing gas-oil ratio increases dramatically, and a high oil production rate can not be sustaiswd after gas bmkthrough. Thus, even though overalf reservoir performances are the same from twodlmensionsl and one-dimensional simsdations, there art differences in the time at which gas breakthrough occurs. McxMied Tamer method, using @JQ given by Q. (5) and Q = SW predicts gas breakthkxrgb (2<. 275 R = center of welf block) after a cumrdative oil smovery of about 35% and 75%, rcqtectively, for Example 2. But twodimensiod (*z) simulation shows ~w_~@kamtiveofl-~oftit~% for the well in k center block (327). This example cakadadon showa that moditied Tarmr method using lii aaturadon pro61e pmdicrs tbetimetogas Wakthmugh Snotc accutaely than rnodi6ed Tamer method using the Cotnple&segsBgationasstur@on. lmproving nsoditied Tamer method for mom realistic mpmrmtm 0ssof h saturadon pmtlle in the aemdary gas cap wifl improve Uw pmdktion of @ time to gas bmskrbrougts, Further impsovemeats to the modified Tamer method as presented in this study carsbe either based on empirical data or on one-dimensional disptacesnent theory,

csoss-s4ctional ams opal to flow Gil formation volume factor Gas formation volume f-r Rock compmsaibility Gravitational asxeleraion, 32.2 ft/see? Conversion eonstanL 32.2 lbm-fk/lbf-sa? Reswoir thickness Pemleabfity Relative permesbtiy of phase I Relative permesbtii to oil at critical gas asturadon Pseudo relative ~rmeabiiity to phase 1 Number of blocks in z direction Number of blocks in x dirwticxs
Pmsrm

~= F= -= p; . R= S= $= f. = = ;=
x=

2= Zc =

Pressure at the resemoir bottom Average reservoir pressure Pseudo Capillarypreastrse Production rate, volJDay C&oil ratio saturation Depth averaged oil saturation Average Oti SSWration Time TmaaMbWty ofphaaef=&&Jkr, Horiaorad disunm Vertical distance ~sas@Pw-

Gmk a= Ar= 4= s . = ~= y= p. co= +. e. = p.

symbofa M @e Time-step size Density diffemce betweezsoil and gas, W@ Change in the Mrinsic variable during a time-step Arbitrarily small quantity Arbitrssily specified SKIrm Dens@ in terms of pstssure/distance = Pglge Viscosity Empirical parameter Porosity porosity at pressure # Fluid density, lbnt/ ft3

CONCLUSIONS 1. Sensitivity study of simulation results to time-step and block size shows that even though simultaneous solution technique is rmconditionally stable, it is not sufficient to ensure reliable answers with large time-step and block sizes. The simulation should be stwted with a small time-step size, which can he increased later, when average reservoir pressure and saturation trends have beers established. 2. The Twrcrsi method significantly underpnxhcts the nxervoir producing life and cumulative oil recovery. The Tarnersl method has been modified empirically to predict the performance of a solution-gasdrive tesmoir with gravity segregation, flris modified method satisfactorily matches the simulated performsnces for three examples, and is applicable before gas breakthrough. 3. Simulated examples show that befote gas hduhrough, the perfonnsnce of a solution-gas-drive reservoir with gravity ae-gregation is independent of absolute vestical Penrseatility and oil productiors rate. if the oil production rate is below the maximum gravity drainage rate. 4. New pseudo cspilla~ and relative permeability functions for a solution-gas-drive resemir with gravity segregation have been developed using the idealizd aaturstiors pmtile from this study and the vertical equilibrium approach proposed by Coats et uf!s pseudo-functions should improve perfotmaru predictions These from threedlmensionsf simrdations using the two-dimensional amaf grids.

Svbacripts g= gc = i= 1= n= *= of = p. s=
w= Wc =

Gss Ciiticsf gas ith specified norm or change in ith inthnsic variable Phase, oil or gas nth time interval Oii Residuaf oil Producing Solution water or well water Connate

ACKNOWLEDGMENTS Finznciaf support was provided by the Depsnment of Energy Contract No. AC03-glSFl 1564, SUPRI A Industsiaf AffWea, --264

SF%169S9

ANILK. AMBASTHAAND KHAL3DAZEZ 22.

WPfU B Irrdtumial Affiliates, and Stanford Univcraity. We thank 5xpioration Constdttts Lmited for the permission to use U@ re$erIoir simulator ECIJPSE for this project. ?EPERENCES 1. Tamer, J.: How DifferentSize Gas Caps and PressureMaintenance ProgramsAffect &nount of RecoverableOILOil Week&. (JW 12, 1944) 144, No. 2,32-34. Buckley, S.E. and Leverett. M.C.: Mechanism of ~uid Displacement in %ndsy Tram., AIME (1942) 146, 107. Welge, H. J.:Dkplacement of Od from Porous Me@a by Water or Gas, Trons., AIME (1949) 179, 133. Welge, H.J.: A Simplified Method for Computing 011 Recoveries by Gas or Water Drive, Trans., MIME(1952) 19S A, Y1. Muskat, M.: The Production Hktories of 011 Producing Gasdrive Reservoirs, Journol of Applied Physics, (1945) 16, 147. Kern,, L.R.: Displacement Mechanism in Multi-Well $KWIS. Tram.. AIME (1952) 195,39. Weve, D.R. and Welch, L.W., Jr.: Gas Drive and Gravity Drainage Analysis for Pressure Maintenance Gpmdons. Trans., AIME (1956) 207,136. Smith, R.H. reprted by Klot% LA.: l%e Gravity Dminagc MechankamJournaf c# Per. Tech.. Vol. V, No. 4 (April 1953) 19. Woody, L.D., Jr. and Moscrip HI. R.: performance Cabtlations for Combination Drive Reservoir% Trans.. AfME (1956) 207, 129. Burtehaell, E.P.: -Weaemoir Performance of a High Relief Pool, Trons.. AfME (1949) 171. Terwilliger, P.L., W~y, L.E., HalL H.N., Btidge$. P.M. m Morse, R.A.: An Experimental ard Tkoriticai hlVeSti@tiOtl of Gravity Drainage Performance Trotrs.,AME (1951) 19% 285. McCord, D.R.: performance Pmdkxions incorporating Gravity Drainage and Gas Cap Pressure Maintenance - LL-370 Are% Bolivar Coastal Field; Truns., AIME (1953) 198,231. Martin, J.C.: Reservoir Analysis for Pressure Maintenance Operations Based on Complete segregation of Mobde Fluids, Trons., AIME (1958) 213,220.

HagoortoJ.: Oil Remvery by Gravity Dminage; Sot. Per. Eng. J. (June 1980) 139-150. 23. YpmL J.G.J.: Analytical and Numerical Modeliig of Immiaeible Gravity-Stable @rs Injection Into Stmti6ed Reamoira, Sot. Pet. Eng. J. (Aug. 1985) 5S4-564. 24. Crafti B.C. and Hawkins, M. F.: Applkd PerrofeunrReservoir En@eering, mntice Hall IrE., Ertglewood Cliffs, NJ. (1%9) 375. Wilson, W.W.: Engineering Study of the Cook Ranch Field, Shackelford Counry, Texas: Trurrs., AIME (1952) 195.77. Katz, D.L.: Possibilities of Secondary Recovery for the Oklahoma County Wilcox Sand. huts., AIME (1942)

) ..

25, 26. 27.

). L

;.

5. 1.

Lewis, LO.: Gravity Drainage in 011 Fmlds, Tram., AIME (1942) 28. Tracy, G.W.: Simplified Form of the Material Balance Eqtration; Trons., AIME (1955) X)4, 243. 29. Loper, R.G. and Calhoun, J.C,, Jr.: flte Effect of Well Spacing and Draw&wn on Recovery from fntemal Gas Drive Reservoirs, Truns., AIME (1949) 1* 211. 30. E1-Khatib, N.A.F.: Whe Effeet of Drainage Area & PmducRrhman@ of Dcplction Dtiveoil Re.sertionllarconrhe Voirs, PapersPE llo19pmae@ed at the 57th Amttlal Mtg. of SFE of AIME. New Gtleam, LA (SepL 26-2!%1982).

B.

31.

9,

10. 11.

Miller, C.C., Btownacom~ E.R. and Kie@miclL W.F., Jr.: A Cat@ation oftlKEffect of Producdon Rate uptm Ukimate Recovery by solution Gaa Drive, Trum., AIME (1949) 1= 235. 32. West WJ,, Garvim W.W. ad SheldorL J.W.: solution of the Equations of Unsteady state -fwo-phaae now in oil hvoir% Tram, AIME (1954) 201,217. ~ JJ. and Roberts, T.G.: The Effect of the Relative Permeability Ratio, the Oil Gravity, and the soltrtkm Gaa-Oil Ratio on the f%mary Recovery ticun a Depletion Type Reservoir. Trum., AME (1955) 204, 120. Heuer, G.J., Jr., Clark, G.C., and Dew, J.N.: The Influence of Production Rate, Permeability Variation and Well Spacing on Solution-Gas-Drive Perfonnanc% J. Pet. Tech. (May 1%1) 469-74.

33.

12.

34.

13.

14. E1-Khatib, N.A.F.: A Modified Methcd for Performance Predictionof Depletion Drive Oil Reservoirs: Paper No. 82-33-04 presented at the 33rd Annual Mtg. of the Pctmleurn Society of CIM held in Calgary (June 6-9, 1982). 15. Cook, R.E.: Analysis of Gravity Segregation Performance During Natural Depletion; Sot. Pet. Eng. J. (Sept. 1962) 261-274. 16. Matthews, C.S. and Lefkovits, H.C.: Gravity Drainage Pcrformsnce of Depletion-type Reservoirs in the Stripper Stage, Tronr., AIME (1956) 207, 265. 17. Hall, H.N.: Analysis of Gravity Drainage, J. Pet. Tech. (Sept. 1961) 927-936. 18. Dykstra, H.: The Prediction of Od Recovery by Gravity Drtinage, J. Pet. Tech. (May 1978) 818-830. 19. Templeton, E.E., Nielsen, R.F., and Stahl, C.D.: A Study of Gravity Counterblow Segregation. Sot. Pet. Eng. J. (June 1%2) 185-193. 20, Ridings, R,L., Dalton, R,L., Greene, H,W,, Kyte, J.R., and Naumann, V.O.: Experimental and Calculated Behavior of Dissolved-Gas-Drive Systems. Sot. Per. Eng. J. (March 1%3) 4148. 21. Joslii, W.J.: Applying the Frontal Advance Equation to Venical Segregation Reservoirs, J. Pet. Tech. (Jan. 1964) 87-94.
1 .-.

of Levine, J.S. and Prats, M.: lhe Calculated Performance Solution-Gas-Drive Reservoim~ Sot. Pet. Eng. J. (Sept. 1%1) 142. 36. Morse, R.A. and Whking, R.L.: A Numerical Model Study of Gravitational Effects and Production Rate on Solution Gas Drive Performanceof Oil Reservoirs: J. Pet. Tech. (May 1970) 625. 37. Csrdwell, W.T., Jr., and Parsons, R.L.: Gravity Drainage Theory: Trons., AIME (1949) 179, 199-215. 35. 38. Richardson, J.G. and Blackwell, R.J.: Use of Simple Mathematical Models for Predicting Resemoir Behavior: J. Pet. Tech. (Sept. 1971) 1145-54. 39. coats, K.H.: A Treatment of the Gas Rrcolation Problem in simulation of ?luec-DmenaiorraL Three-Phase Flow in Reservoir. Sot. Per. Eng. J. (Dec. 1%8) 413-19. 40. McCreary, J.G.: A Siiple Method for Controlling gas Per@ation in Numerical Slmtdadon of Solution Gas Drive Reservoir Sot. Pet. Eng. J. (March 1971) 85-91. 41. Settari, A. and Aziz, K.: Treatment of Nonlinear Terms in UE Numerical Solution of Partial Differential Equations for Mtdtiphase Flow in Porous Meda? ht. J. Afbdtf@we Ffow, 1, 1975, 817-44.

8 42.

MAIS?UAL BALAWB CAUXJLAfKWS K1R SOLUIION-GAS-DRIVBRESERVOIRSWmi c2MvrrY sBGREGAnO!! Ifweaelect areferemeplsm fortheaswlcabdadm Oftheresenfoir (z.Oplane), the31atz=(k Pg-PO=FC=ATZC

SE 169S9 aadsetop

.-$

.. .

Aziz, K. and settari, A.: Petrofenm Reservoir SbmdUI@ & #Ied science Publii Ltd., London (1979) 169. 43. Grabowskio J.W., Vie, P.K., L@ RC,, Behie, A. atd Rubin, B.: A F@y Ir@cit @ner@ purpose f%lite ~lffelwlce Thennrd Model for In-situ Combustion and Steam, Paper SPE 83% presented at the S4th hnual Mtg. of the SPE of AIME, LiISVegSS(Sept. 23-26, 1979). 44. ECLIPSE Refenmce Manual, Version S4/11, Exploration ConSUkSIUS Ltd. (NOV. 1984). 45. Coats, K.H., Dempsey, J.R. and Henderson, J.H.: The Use of Vertical Equilibrium in Two-Dmensional simulation of ThreeDimensional Reservoir Perfonnanee: Sot. Pet. Eng. J. (Mach 1971) 63-71, 46 A1-Khalifah, A.A., MIZ, K., and Home, R.N.: A New Approach to Multi-phase Well Test Analysis, paper SPE 16743 pzwented at 62nd Annual Meeting, Dallas, TX (Sept. 27-30, 1987).

(A-5)

PIwure diffemmceat z = O is a pseudo capillary pressure deaoted by p.. Substitution of Zcfrom Eq. (A-5) in Eq. (A-1) givess relation between $ and PCax

s=

(Q,e -(1

-S* h AT

- S8C) PC 1

+(1 - SW - ~gc) (A-d)

Pseudo Relative Permeability Pseudo relative perrneabilities to oil and gas are given by I ~.= krom.wZ. + kmcg (h - z.) h
(A-7)

APPENDIX - DEVELOPMENT OF PSEUDO-PUNCTIONS

IrI this Appendix, we develop pseudo-funetions for solutiongas-drive reservoirs with gravity segregation. The development is based on the verdctd quilibrium concept proposed by Coau et uf.a and idealizd satwation profile ahown in Fig. 3. Pseudo
capillary

Pressure I .. Where k.and+ale


~CtiVdy,

(A-8)

Depth averaged oil saturation at some position and time is given by ~= (h thetelative penneabilitics tooil and gas, Of (~~k given by ~. (5). =G between Eqa. (A-7) and (A-1) reatdta itx
M Oil WtU@OII

(s0)2.

% +

(1

&

~r)

2=)

(A-1)

where (~o)z,is given by Eq. (S). Equation (A-1) is beed on the SSsum@on of idealized saturation pmfi.leof Fig. 3. Equation (A-1) Rlales the saturation F to the secondary gas cap-oil mntsct position, ZC. At the pliMe z = z<: Pm= Pgc (A-2)

g,=

(k,om - k,ag) [~-(1 (Qc

-SW - S-)]

+ ~~

(A-9)

- (1 -SW - Sgc)

A similar elimination between Eqs, (A-8) srd (A-1) results in:

because of insignificant rock capiUarypressure. At any position other

than z = z., oil and gas pressures are given by: Po (z)= Pm + y. (z - 2.) P8 (z) = D& + 78 (Z Zc) (A-3) (AA)

68 =

k,g,aw [~-(1 (&)zc -(1

- $w - Sgc)l - Sw - Sgc)

(A-1O)

The expressions developed for pseudo capillary pressure and selative permeability are useful only before g~ b~~~ugh, which means that Z, c h or quivdently, ~> (s),..

1 ---

1
TABLE L ADDITIONALDATA FOR EXAMPLE 1
Rsxavohaiac Oifpmduaimruc

Nx

GaadmaiIyalaldaldcoXKmOM
bxilixloila&l#knl coXmalCwJtaaawatim Initial psaxoc

Tonf aimulxkm*C

dmmwnal column2327 X 2327 X 135 ft tafl ~ ~~ay k the ]OWSI bfock mmd 6olwlF 0.0s ibm 0.99 0.01 Givenby gravityquitibnum wiUx pmssumof oil 17S0paiaatthc top of the column 13s0 Daya

~..

NX+l

NX+2

2 NX

FwosiIyfouOWSlbc -

O=++CJ( P+)]

NZ

NX NZ

Whuc

d= 0.04 pa= 17s0 @a


x K@ psi-

a
Fig.

b
la

c,= 3.0

umqunau for one-

qu!

twdbnendonai

aimulxtlons.

TABLE 2. ADDITIONALDATA FOR EXAMPLE 2


RcsuvOuaiae Oifpmduaimmc VCnicaf Ftmdmly

2500x 2500x 1000 SrBmay 0.0s OYfP 1.00

3mfl

1050 AZ 11 1030 ~ ,*~~ q

OilIlauityz!atmdax domiitiGzsfkoaiv alamdmlcmlwms


lxitialoil~

2??

4.5 :5 225

c~

aaaaahl Ixilialprualw

O.m of Givenby gmvitycquihbriumwith oil prcasurc


1700paia x5 ftliomtkt0Pof*mI5400

: +

Toialaimlmkmtht RYxoaily Rock ~v

0.10

bays -/ .~
o

3.0 x IF pat

+ Idcdized Profile

30

60
Di2tanwthmt

90
hetop,fi

lm

150

Fig.2-anDat

othJoefcsia90xl

P18-uNP-0(Ex-x@o

1. At=? OdWS,

UmO=900dWS).

..

Lo

a9

!
[ 80

&Metftod AutoowicAt Cuatrol

0.93 a7 0.9

22.5

10

0.3 1 1

ldeahacd Refile I I

Os$ -:

z +90
I I 0.1 I 0.1s Fractkon
nmwmd 10I E,.

30

a w 120 Distancerom f thetop,fl


PIOflm IE..mw 1. v. w day.. th..

I 1s0
9w tiy.)

<
0.8 o

O.OJ

0.2

Cumulative Oil Recovery.


Fig 74* *1 smlmim tram s. Unlilbuon (w.

3-EWU1 01 Mock q,z* FIE

on wt.raoon

* It) MIJ TM!U*

1100

At.Days q 10
q 30 A 45 +90

~ ,W
7 ;
c!

Tamer

Metkl

Single Block Simulation

q **9

.*

***

;.

..9
w,
q

:::: . ...+ Wlo 30

1::: q .* ,d en

::::::2 +*

q .+*
Im

km 6 $ m

90

150

003

0.1
0)1

0.1s

02

Diatanccfmn thetcp.
F* 4.m9e10fmu.,18@0no ruWwOwlnmta,m Qlol.

n
81.9 R. Hmo.9006wy8,

Cumulative

Raovay, Fraction

I .0

L q *

d .ma6&.9bb

09 -

At. Days
u. g s ~ .6 ~ 09 ,095 -

Tamer

Method

Single Block Simulation

07

10 o 30 A 45 +90
q

05

-*
q

oaf

I
o,~ o 3 60 w 120

I
1$0

E 4 OH o 005 Cumulative FIQ 9-Av.rw. I 01 Oil Recovery. 01$ Fmc[ion


1

02

Disumccfrom the top, ft


n~ s. tw,ct et t#u,..*t.* 0 ,.1,,110 p.ttle
(En*rnpl. I ,r .9 Il. tom,. 900 6,s,

.311 uralm. Iron!q w qlc.aiorb ofwlallon and lame, s In.thou lot i!.mnpl.

-5-!!+
P:.

U201 o
~Ia 8-hwmq9r.wrvw
W.*UW*

l\

I
02

00s
Cumulative 0:1
IrOm *.D smwl.tmn

0.1
I \t.

01$
? Ill Md TWO

Raovery,Fraction
mothcd 10+E**rn*b !

l-sgc -&J
208

PA

.. .

. .... . .=. ., ,. . ..

....

WE 1695g
E
!! ,i I2(Q

m Mcdfd
Itm
q

Tamer Methcd

500 STB/D&20md

6
8 ~:: i 3

1.2 ~, . o .. ,..>+
I

c;
. ,

A . Ref. 40 B - Ref. 39 C - Ref. 46

& i!! $
..,

;/0 ---;---I I

m,

005
~,,u,. I,OM ,.0 gmnJIS,W

01
,na nwu,,ud

015
law IW,hod

02

0.5

1.5

Cumulative Oil Recovery. Fraction


P19 I I .
h,,,+ ,,,vM 10I

Normalized Presawc
E ,m@, 1. Fig Is.aahltlall ~oll ,!0 rot trwaoPVT m ml,.

1.8 !? ~ 16

i
8

14

12

..

0 0s
Cumulative

01
Oil Recovery.

015
Fracoon

02

I@ I
. . B - Ref. 39
#

c:.

A - Ref. 40

C - Ref. 46

,.

,.f

-d----

B __

. ...

~-

A
I

. ,..

0.s

1.s

Nonnaliad Preaaum

F*

*$-A-

@ h8ml-o

and

Wnw-ractaalmnl

a I

2-D
q

I (X-Z) !hd42i041 Wefl in

bfeck (327)

+ 4

Well in block (L27)Well in block (S.27)

~ 1200 -0 t 4
km

A o

1000 STWDay 2(KI0s?8~y Modified Tarrier Method

0.0s

0.I

0.15

01

Cmnul.10\c 011 P.ccovcr>.


Fq I 3- P,.XI.
<89 g,, , , ,,,,0 ,,mn ,.0 ,,,n,.,,.n ,M

I.r.h.la,m
.l.,h~ ,., E,*In*,, ,

CumulativeOil Recovery,Frwion
F@ 17-A.crqo ,,WIVOI, LU,W, 101 E,9w18 2

,nga,,,..,J ,,,,,

1.0
1

hk(rcwy

Ilht. 4(II
V)) 011

2.D (x.z) Simulation


. W41

_= $ ~

,.

on
(Ml

\
\

( ;J> \

CmII\ ( Rut

in block (3,27)

+ +
09 \ \ \

Well in block (1,27)Well in blak (5,27)

\ y fj u ~ [1..l

i
8

0.8s -0

().2
11 , .,

! q

I-D (z) Simulation loflommay A 2W0 ~Day Modlfkj Tuner Method 0.05 0.1

0.2
Fq Ia-nel,,,,,

0..1

01 ..

0.(3

().X

0.1s

0.2

Od Saturahon
pg,maabu,,
c,,

CumulaIIweOil Recovery. Fraction

well in blosk (327) Well knblock(1,27) ~ 1200 -

.Aw
i -

+ t

well knW (3#7~ v$llinblock(}~)

+ Wellinblock(547)

WxilInblock(S&L

!!! 3s - 1-D (a) Sirnuixth

!
q+

lMO 900

1.D (z) Skmuluion

Km STwDey

E ~~
o
O.m Cumulative

A 2ooo-y Modified TunerMethod 0.1 0.1s oil Recove$y, Frxction

4N
1

0 3omd A 10nrd
ModkfWlkIWMdKd
&w al 0.1$ 0.2

0.2

7s0

w
c Uso +
q

Wetl in W

(3,27)

Wxuinblock(lm)

~ 0.30s -----m

.-.

.-

-----

well M h

(s.27)

mOO IOD (~S~u2edoo MM -

0.306m--0.304

2-2)(xc) Simulxtia wou b200k(3!27)

10md ModilbdTanmMedld
A

i
8

0.302 :

--w8ublock(l,27) ........ WX21 ak (s27) W -. _ . \.D(*)skm~~

@w

al

als

al

Cumuldve oil koveey, Prmion

}.0
2.D (X.2) Slrnulmkrll s Well in blak (3,27) i
0.9 + Well in

A
4

I!D (z) Skmudion

A
q

Bxxmple 1

block (1,27)-

Sxxmple 2
Eaxmple 3
j

Q Well In block (S,27) 00


I -

4 -*
4

3
0.7

s
0.6o

1OM(I
Modi!led Txmer MeUmd
I 0.03 0.1 31$ 0.2

0,,0

0.15
Cumuhfhe Oil Recovety, Fmction

0.2

CumulativeOil Recovcv, Frection

Вам также может понравиться