Sei sulla pagina 1di 8

Evaluation of Empirically Derived PVT Properties for Gulf of Mexico Crude Oils

Robert P.. Sutton,

I
/3/73

SPE,* and F. Farsh?zd, SPE, U. of Southwestern

Louieiana

SFz

paper evaluates several empirical PVT correlations for application in the Gulf of Mexico. Ideally, fluid properties Summary.This are determined exper@entaUy in the faboratoty; however, these data are not always available. Correlations are consequently used to determine values for bubblepoint pressure, solution GOR, FVF, and viscosity. These vafues are necewmy to compute oil reserves or for calculations involving flow though pipes m @mus medii. A review of the ~omelmions is pmvidwl, along with the results of c$dculatiom made on 31 kiividud cmde oil mmples. Itrtroductlon Tbecalculation of reserw in an oil re3ew0ir or the &termination of its performance m@res a Imowledge of the fluids physiwd properties at efevatedpresmre and temperature. Of primary importance are bubblepoint pressure, solution GOR, and FVF. In addition, viscosity tnus.t be detwmimd for calculations involving tie flow of oil tkougb pipes or porous mdla. Ideally, these pmpties are determined from laboratory studies &sign@ to duplicate the conditions of interest. Experimental data quite often are umvailable, however, because adequate samples mnnot be obtained or because the producing horizon does not warrant the expmw of an in-depth reservoir fluid study. In these cases, PVTproperdes must be determined by analogy or throngb the use of empirically derived corr,,, elations. During the last 42 years, several correlations have been propsed for determining PVT propefies. The most widely wed treat oil and gas as a two-component system with each component having a fmkd composition. Only the sf+citic gravity drelative amount of each component, the pressure, and the temperature me used to cliaracterize the oils PVT properties. Cmde oil systems from VWioti oil-prcdncing regions were used in the development of the correlations. These crude oils exbiiit regional trends in chemical composition @at categorize them as pamtfinic, naphthetic, or aronkitic. Because of the differences in composition, correlations &VeIO@ fiOIII region+ s~pl~ t&t me pR&ImiMutIy of one chimical base may not provide satisfactory results when applied to crude oifs from other regiob. This paper examines tie fluid prepeqies and correlations shown in Table 1 to determine their applicability in the offshore Gulf of b.fexico region. Correlations Fdr.years, field engineers have used empirical cofielations in lieu of laboratory data for determining the fluid piopdies ntiSSW fok calculating reserves, resemoir performance, and equipment desi~. untiI remnt&, the CXIIY guide for ~ktig a correlation ww offered by .Chierici et al., 10 who suggested the use of La.saters cotieladori for cmde oils with a gmvity >15API [<0.97 gkm3] and S@dings COfiC+kdOn crude oik with a gravity < 15API for [>0.97 g/cm3], fn 1983, Ostermann et al. 11 provided a more complete evahmtion of empirical c0rrelati0133 for determining the PvT pmpenies OfAlaskm crude oi13. Eight S3mples were anafyzed, aid it was found that the correlations pmpmed by GLW$4 (hubbkpoint pressure), SWndW1 (o1 FVF), and Seggs and Robiion8 (deti and gas-saturated oil viscosity) were the most accurate. At tbis time, $miIar analyses have not appeared in the literature for otler oil-producing regions. fhe effective use of the correlations lies in an understanding of their development and a knowledge of their limitations. The following presents a review of tbe correlations development. Tbe equations that formtkse correlations ~eprovided in the Appmdix. St&ding. Standing1,U,13 publiskf correlations for determining the bubblepoint pre$ure snd FVF of a gas-satumted oil from known Nowat Marathon 0!!Co. copyright 3X!0 1 Scdev 01P.3troleurn Engineers SPE Reservoir Engineering, FebnwY 1990

.,
values of tempemture, solution 00R, ad oil and gas gravities. In all. 105 exiwrimentallv determined data wins on 22 dfierent cmd&dh3&31-ga3 mixtures tim CdifO&a were used to arrive at the correlations. fin ga3es present in the mixtures were&e of N2 and H2S, but C02 was present in a few samples at concentmtions less than 1 mol%. AU the data were obtained in the laboratory with a hvc-stage flash separation designed to duplicate average field conditions in California. Table 2 shows the ranges of these data. Standing repmt.id an average error of 4.8% and 106 psi [731 k%] for the bubblepoint-pressure correlation and an average error of 1.17 % for the FVf!correlation. The lattez is more general than the former and will provide satisfactory .Rsult.s for a wider variety of crude oils. La$ater. ?..asate+ presented a bubblepoint-pressure correlation in 1958. A total of 15S experimentally measured bubblepoint pressum horn 137 independent cmde oil sys@x tlom Canada, western and midcondnental US., and South knerica was used in its development. The natural gases asmxiated wdth these cmdes were essemially free of nonhydrocarbons. La8ater wed Henrys law to derive a bubblepoint-pressure factor and correlated it with the mole fraction of gas in solution to obtain tie cume shown in Fig. 1. The mole fraction of gas in solution is cakufatd from values of solution GOR, oil spcitic gravity, and molecular weight. Because an oils mok.cuk weight is generally ?n uDknoWn qUantilY, I.a,Wer provided Fig. 2, which relates molecular weight to oil gmvity. l%is relationship corresponds with that observed from crude oils with a UOP characterization factor14 of 11.8. Table 2 shows tbermges data usei by Lasater in deriving tfw correlation. @sater of reported an average error of 3.8% between measured and calculat~ bubb@O~t P~ss~ms. Vazqnez and B=. fn 1976, Vazquez and Biws3 piWIlti ralationtips for deteag tie sOIutiOn GOR ~d ~ Of a g=satumti crude on. JDtO@L 6,~ ~~ PO~ts wer~u$~ ~ tbe development of these correlations. The data were separated i33t0two of v~tions fi tbe VO~flltY Of c~de Oil. fbe fit ~OUPS -use group contained oils with gravities S300API [20.88 gtcm3]. The second group contained oils with gravities > 300 API [<0.88 g/cm3]. Table 3 gives the number of data fmints and ranges of data for *h group. Vazquez and Beggs found gas gravity to be a strong correlating p=e~r iU tbe development Of tbe SObJtiOU GOR m~~~On. Bemuse g= 8w@ is *~dent ou tie COIIditiOm @r w~ch *8 gas is separated from the oil, Vazquez and Beggs &veloped a correlation to no-e g= gravity m a sV~tiOn pressme Of 100 Pig [690 ~]. ~S P~ss~ wm c~se~ ~ause it ~ ~t tO be representative of avage field sep@Or mndlti~. ~ v~ue derived from the equation for this adjustment, YgS=Tg(P)[l.0+(5 .912 x 10-5)YAPLP iog@w/H4.7)], (I) is used in all of Vazquez and Beggs correlations. A total of 124 data points from 27 differant reservoir fluids were. used to arrive at G+ 1, which was reported to have an avemge error 79 I

TABLE 1FLUID Fluid Property

PROPERTY

CORRELATIONS Correlation

55 50 E ~ 45 := 40 ~ 35 ~ 30 &

Bubblepoint pressure S.i4ution GOR FVF-. ,> .:. .. :: Jsktke:mal rnrnpresslbllity


Dead-oil viscosity G68-saiurated.6il Undersaturat6d-oil tiscosity viscosity

Standing, 1 Lash.ter,z. Va2qu62 and Seggs,3 and Glas04 St6nding, Lasater, Vazquez and Beggs, and Gla.% Standing, Vazqu~ and Beggs, and Glasd Calhoun, s Tribe, e and Vazquez and Seggs Seat, 7 Beggs and Robinson, 8 and Glas@ Chew and Connally, s and Seggs and Robinson Beal, and Vazquez and Stiggs

z ~ 25
20

?00 TABLE 2-DATA RANGES FOR STANOING AND LASATER CORRELATIONS Standing Bubblepoint pressure, psia Temperature, OF FVF, RS/BTS Solution GOR, scf/STB Tank.oil gravity, API Gas gravi~ (air= 1) Separat0rpre6sure, psia First stage Seconds tage Separat6r temperature, OF 130 to 100 to 1.024 to 20 to 16.6 to 0.59 to

200 EFFECTIVE

300 MOLECULAR OF TANK OIL

400 WEIGHT

500

Lasater
46 to 5,700 62 to 272 3 17.9 0.574 15 to to to to 2,905 51.1 1.223 605

7,000
258 2.15 1,425 63.8 0.95

lg. 2-EffecNve after Las6t6r2).

molecular weight rdat.ad to tank-cdl gravity

TABLE 3DATA FOR VAZQUEZ SEGGS CORRELATIONS Solution GOR and FVF of Gas-Saturated ,. /m S30 Number ofdita points Bubblepoint pressure, psia Average temperatwe, F IWF, RS/STB SolutionGOR, scflSTB Tank-oil gravity, API Gas gratity (air= 1) 1,141 ~ 15 to 4,572 162 1. M2tol.545 o to 881 .15.3 to 30.0 0.511 t01.351

AND

265t0465 14.7 100

Crude oil Y,4PI >30 4,883 15 to 6,055 t 80 1.02Bto2.22il 0 to 2,199 30.6 to 59.5 0.530 t01.2~9

24!0

106

Development of Eq. 1 Separator pressure, psia Temperature; F Tank-oil gravity, API Gas gravity (air= 1) 60t0 565 76to150 17t045 0.56 to 1.20 Oil ,yiscodty C6rrelati0n 141 t09,515 NIA N/A 9.3t02,199 15.3 t6 59.5 0.511 t61.351 o.l17t0 148

Under$aturated

Compressibility Correlation Pre3sure, psia Temperature, F FVF, RB/STB Solution GOR, scf/STB Tank-oil gravity, API . Gas gravity (air= 1) viscosity, Cp 141 to 9,515 NIA 1.006 to 2.226 9.3t02,199 15.3 to 59.5 0.511 tOl.351 NM

TABLE 4-DATA RANGES FOR GLAS6Y BUBBLEPOINT PRESSURE AND OIL FVF CORRELATIONS Fig. IBubblepoint pr6s6ure factor (after Lasater2): Bubblepoint pressure, psia Temperature, OF FVF, RB/STB Solution GOR, scf/STB Tank-oil gravity, API Gas gravity (air= t) Separator pressure, psia First stage .%cond stage Separator temperature, F 165t07,142 60t0280 1.025 t02.568 90 to 2,637 22,3 to 48,1 0.650 to 1.276 416 15 125

of 0.545 %. Table 3 gives the ranges of the data used in the developmentof~. 1. Average ermrsof -0.7%and 4.7wcrereprted for the solution GOR and WF correlations, respectively. Vazquaz md Beggs also investigated the physical properties of
undematurated oils, including viscosity and isotberznd compressibility. Tbelitter isrelated totbeoil FVFatpressures above the bubblepoint by co=@bP) fn(&/%b) . . . . . . . . . . . . . . . . . . . . .. . . . ...(2)

k all, 4,486 d6t6 points, encompassing tbe value ranges shown in Table 3, were used id the develqmient of the compressibility mr-

rdation. 80 3PE Reservoir Engineering, Febimiy 1990

TABLE 5DATA

RANGS8

FOR GLASO AND EISALS VISCOSITV

CORRELATIONS

Bed
Dead~l VlscOsily 98 to 260 10.0 to 52.5 0.865 to 1,550 N/A Undersaturated Viscoslly At p~ NIA NIA 0.142 to 127 140 to 4,136 12 to 1,827 Oil.

Glas$ Temperature, F Tank.oil gravity, API Viscosity, cp Pressure, psia GOR, scflSTB 50 to 300 20.1 to 4s.1 0.616 to 3S.10 WA NIA

Above pb N/A WA 0.160 to 3151,515 to 5,515

NIA

NIA

Tbec.mektion for determbin g Ok viscosity of undersatumted cmde oils was developed from 3,593 data points with ranges also shown in Table 3. The average error for the viscosity coirelation was reported at 7.54%. Gla.w. In 1980, Gla@4 presented correlations for cahdadng bubbkpoint pressure, oil FVF (for a gas-saturated ofl), and dead-oil viscOsiLy. He prmented data from 45 oil samples, mostly from the North Sea region, and q.wd them in the development of the tirst two correlation. Table 4 gives the ranges of these data. Glas$s correlation was developed for oik with par@nicities equivalent to North Sea oils (i.e., oils with UOP characterization f@ors14 of 11.9). An adjustment to the API graviLy term in tk equatiom$ * su8gested when the correlations are used with oils of a diiYerent compositionzdmtuie. An average correction multiplier of 0.P4 was demed for offshore Gulf of Mexico crude oiis, but use of Okisfactor dedreased the accuracy of the correlation. Therefore, this modification was omitted from f!irther consideration in this study. Glas@ als6 provided a bubblepoint-press,ure correlation for voladLe oils anrfh. methcd for correcting the predicted bubblepoint pressure for the presence of C02;NZ, and H2S iq the total surface gases. These correlations were not applicable to this work. Gla?.o reported an average error of 1.28,% with a standard &viation of 6.98% for tie bubblepoint-pressure correlation. TIIeaver. age error of the oil FVF correlation was reported at 0.43 % with
a standard deviation of ,2.18%.

In addition, Bed prewmted . m-e. for etimatino the viccnsi~ of undersatumted oil. This m-ve was the resuR of cdrrelatin~ 52 viscosity observations takein tim 26 crude oif samples fm; 20 _ l~tioos, 11 of which are in Cdifomia. f%lfik M were taken &pm crude oil samples at bubblepoint pressure, while the remaining data were obtained for pressure+ above bubblepoint. These data represented the rangis shown in Table 5...Tbe resuftig ~c~on hd a reported averageerror of 2.7%. . ~ . .. Chew and Connaffy. fn 1959, Xiew &d Connallyg prisented a correlation to predict the change in oil viscosiiy as a dmcdon of solution GOR. The correlation was developed from 457 crude oil samples o.btaird tlom the major producing areas of cana&, the U. S., and South tierica. These data encommssed the ranges shown in Table 6. An analysisf the data indicated that differential oil FVF or sOo lution GO-R could be used as a corrdadng parameter. Chew and Coon8Uy selected the diffemndal solution GOR as a correlating pi+ rameter because these data are more readily obtained. Although differential &ta were used to develop the correlaQon, Chew and Connally stated that data from any Of@e usual liberation prccesses could be us@ in its application. The final correlation was.uresent. ed for solution GORs ~tig - r from O to 1,W scfNfB [0 io 288.2 M m3ktOC!+kmk-Jn3] became of a lack of c.msktmt data at higher GOR%. BWOC -----.. .~a n-b~..=T 107~: -..hl:.k-~ & - . . ,-, R.S:C ..-+ u.-IA--.8 equitioms for .sabxdadne dead-oil and ms-sabmated-oil v&mities. The equatiomi resultedI-iom a tiy ;~ 2,533 viscosity measuremeats (460 dead-oil observations ~d 2,073 gas-sa~ted&2 observation) involving .503 tierent mwie. oil systems. TabLe 7 gjves the ranges covered by these &ta.
-.,-,F, . . . . . ..... . ....-

Glas@s deackiil viscosity correlation was developd from data obtained fmm 26 crude oil samples covering the value nmges shown in Table 5. c BeaL b 194.5:Bwi7 published ~pbid CO~htiODS for deAing the viscosity of cru& oil. A total of 655 values for dead-oil vis.sosiej at lIMF [38C] wme obtained from 492 oil fields, 358 of which are located in the U.S. Viscosity data for an addltionti 98 samples were obtained for tempemtires above 10+3F F38cI. The result+gcorrelation was based on data covering the ranges shoti in Table 5. Beds dead-d viscosity correlation as presented consists of,tive curves covering the temperature range from 100 to 220F [38 to 104C], which represented the &ta witi an average error of 24.2%.

25

20

TABLE S-DATA RANGES FOR CHEW AND CONNALLY CORRELATION Bubblepoint pressure, psia Temperature, OF SO[ution GOR, scf/STB Dead.oi[ viscosity, cp 132 72 51 0.377 to to to to 5,845 292 3,544 50

?0 : r.?

15

!0

TABLS 7-DATA

RANGES FOR BEGGS AND ROBINSON CORRELATIONS 15 70 16 20

0 0.5

0.6 P.b. wkc

0.7

0.s

Pressure, psia Temperature, F Tank-oil gravity, API 80hltiOP GOR, scf/STB

to 5,265
to 295 to 58 to 2,070

Fig. 3-Correlation Calhoun ).

of isothermal

oil compressibility

(after

81

,.

*H

I..

X.,

.ee,j.,tim
-Rim,.

.-,

L,,.,.

a .ss.-,.

-.s,

..

,,

m so.

Fig. 5Appfbxlmate var!atlon of p=udostfticai pressure and psaudocdtfcal twnpefafum with specific gtsvlfy of liquid I!orrmted to 60eF (after TrubeQ).

G+ C.mm,

TEumuATum,

. . . ..

Fig. 4Variation of ps.eudocritical temperature with specific gravity and bubblepoint of liquid corfected to 60F (after Trube@).

CsUIoun. One of the earliest comelations for determining the isothermal compressibility of an undersatumted crude oil was presented by C?dboun5 in 1947. This correlation, shown in Fig. 3, relates a value of avefage ccmpressi%ili~ to tbeoif s@6c grsti at saturation pressure. No information, was pre*nt@ by ~0~ concerning the data used for developing this correlation.
Trube. In 1957, CrU@ presen~ a ~mel~On fOr ~~~g tbe

AII aWIZge eITOI Of 0.64% titb a mhfd deviation Of 13.s3 % was the reprted accuracy of the dead-oil visuwi!y uirrelation when

tested against the data used for its development. When tinted agti 93 cases tim the literature, the average error increased to 114.3% with a standmd deviation of 530.0%. The Sas+ahmtedti-viscosity correlation had a repotiaccmracy of 1.g3 % with a standard deviation of 27.25 %.

i.

isothermal compressibility of an m~ma~m~ ~de Off~ a ~ction of pseudoreduc.ed presswe.and tempmtnre. Methods to estimate criticsl pressure and critical temperature, shown in Figs. 4 and 5, also were presented. Trube recommended tlwtFig. 4 be used fOI dete mining critical temper-are. Bubblepoint pressure is adjusted to 60F [16W3 by use of Standings cOIIektiOn. me W$citli gravity of the oil af resewoir conditions is adjusted to 609F -. i6C] with ~. 3, which approtites the relationsbip15 recom~endexlby Tmbe in the region of intefest. ,. ,
7*=70(D+(T60)(4.6 X10-4). . . . . .. . -; ..+ . .(3) from

he pseudor.@uced compressibility

of the oil is deiermi&d

ig. 6 and is related to oil compressibtity Co=cp,lpc.

as foflowx ,. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (4)

TABLE 8-U.S.

GULF COAST DATA RANGES Gas-Saturated

oil
100 to 9,760 121 to 270 1.036 to 2.245 17.7. to 2,56S 0.55 to 7:51 0.14 to 6.44 26.0 to 54.2 0.607to 0.777 5010.1,415 ,60to120 0.589 tO.o.772

Bubblepoint pressure, psia Temperature, F WF, RBL?TB Solution GOR, scf/STB Dead:il viscosity, CP Gas-saturated-oil viscosity, CP Tank.oil gravi~, API Total gas gravity (air= 1) Separator pressure, psia Separator temperature, F Separator gas gravity (air= 1) Undersaturated .Oil Pressure, psia At bubblepoint Above bubblepoint WF, RB/STB Isothermal compressibility, psi t x 10 e Solution GOR, scf/STB Oil viscdsily, Cp At bubblepoint Above bubblepoint

727 1,000 1.059 3.o3 107

to to to to to

9,76o 12,015 2.263 28.31 2,563

lg. &Coffelation of pseudoreduced ndefsa~ rated oil, (after Tmbe ). 82

compfasdblllty

for an

0.144 to. 3.26 0.156 to 3.46

SPE Reservoir Engineering, Febmao 1590

TABLE 9-STATISTICALACCURACY

OF

uUBBLEPOINT

PRESSURE

AND SOLUTION

GOR CORRELATIONS Solution GOR

BubMepoint Pressure Pres+uri Range Stsnding -21.50 20.11 2s.96 29.40 6.03 13.03 12.10 9.22 ly4; 13.53 7.66 16.32 .6.16 16.22. 6.15 . , 14.23 7.50 14.23 7.50 10.68 17.55 16.22 12.42 0.96 27.$6 ,. S.35 6.00 6.66 5.48 18.00 11.92 6.66 11.92 -0.77 24.66 13.66 9.92 13.73 9.61 3.92 16.86 3.46 13.31 -9.27 32.66 6.60 6.96 8.98 8.42 11.65 12.61 13.17 10.76 -6.67 25.34 -21.46 14.41 21.40 14.41 -20.45 29.92 24.S4 26.00 -6.63 27.59 22.02 1s.79 . Lssater -17.52 33.86 27.51 26.29 6.62 14.24 12.17 9.25 3.92 13.59 10.91 6.80 6.06 9.76 9.84 7.67 Vszquez -26.70 41.96 41.51 37.14 -1 .5s , 16.75 13.02 10..54 348 12.28 6.96 .9.02 1254 11.11 . 12.56 11.10 Glz.% -19.31 36.25 31.29 26.47 -7.09 16.95 13.85 12.00 -2.21 12.43 9.82 7.82 7.48 9.16 9.11 7.45 Bt2nding 11.68 26.37 24.79 17.92. -13.86 25.00 18.05 2.2.12 -1 S.80 15.57 20.06 13.86 -29.06 16.63 2S.06 16.63

Lasater S.47 30.15 24.57 19.26 -11.w 27.82 16.03 26.11 -5.99 16.96 12.24 13.12 -12.82 16.07 14.96 13.63 -18.60 16.70 19.21 15.96 -.30.49 49.91 41.10 49.37 -6.00. 29.29 19.42 22.70

Vazquez 22.63 27.66 30.43 1B.65 -3.24 32.26 17.71 27.21 -6.96 20.36 12.70 17.32 -20.01 21.16 20.02 21.17-21.04 16.95 21.12 16.64 -.10.63 32.19 . 20.02 27.02 1.24 30.91 21.16. 22.54

Glas.$ 7.31 31.12 23.62 21.41 3.58 26.76 14.96 22.40 0.76 15.82 .11.60 10.68 -11.90, 15.52 13.90 .13:67 -14.14 12.59 14.65 11.97 -27.60 36.07 ~~ 30.16 36.06 0.30 27.06 17.63 20.49 ,.

<1,000 psla Averageermr,$i Standard deviation Average absolute error, % Standard deviation 1,001 to Average Btandsrd Average Standard 2,000 psln error, Oh deviation absolute error, % deviation

2,001 t03,000Psia Average error, % Stsiidsrd deviation Average absolute error, % Standard detiation 3,001 to 4,000 pzla Average error,% Siandsrd. deviation Average absolute errw, % Standard deviation 4,001 t05,000fXIa Average error, % Stsndzrd deviation Average absolute error, % stsndzrd deviation >5,000 pda Averafje error,% standard deviation. Average absolute erior, % Standard deviation Totial Average, error; % 8tandsrd. deviation...

Average absolute 6rror, %


Standard deviation

16.66
19.56

16.66
18.20

21.16
26.4\

17.79
19.29

Trub-s did not specify the data used to develop the correlation, nor did he alh@ to its accuracy, z@x@ the examples presented in bis p.zper showed an aveI28e absolute errnr of 7.9%. between Cahdatd ziil me33ured vzfues. lLs. auncOastPvTDati systems were 8ath~ from PVT san@es taken along the J..oui@na and Texas !@ mask. This. information consists of 285Ma point.$ gzsfor sa@rated oil ad 134 data points fnrunderzaurated oil. These samples,were typical of the crude oil syskrns fonnd in the Gulf of Mexico. The mzjoiity of the dztz came from oifs of low to mderate vola~, however, two of the oifs had a bighfy volahle nature. The gaws witirdatively free of impurities, but C02 and N2 were present at ~ncentmtiom averaging 0.4 and 0.3%, rqectively. Tabfe 3 shows the ranges of values covered by tkse data. The data were, obtained Llnm conventional ~ reports that derive the various fluid prnperti~ through a duTerenti@ liberation process. Fnr comparison wi!b the vnriouz PVT cnrrekidons, the SOhldql GOR @d FVF dztz Wsr3 a@ted to represent a 53h liiadm$rncess by the technique proposed by AMyX et nf. 16 and used by Vazquez &d Be8Ss3 in tbe devefopmnt of their correlations
Bo=&@ofi/B~) ,, Ad ... .. .. . . . . . . . . . . . . . . . .. (6) . . . . . . . . . . . . . . . . . . . . . . .. . . .. . . ...(5) Daia represendng 31 dWerent cmde-dlmturzl-gzs

Remilts of v2ri0u2PVT pmperdes wersmzde with 2Uthe zfoti mentioned correlztiom. The equations or chartz of the3e mrrilaIn zrriving zt the tiom were wed in the folfowing evzluzdon. czlctdzted result3, measmred quzntitiw were used for input into the c011ek3ti01w r3mphs of czkufat.ed vs. mea.mred properdes Ye provided k Ref. 17 for each of the cmrelatiom exaniined. A refem ence line was placed on each 8r2ph so that the correkition,s accuricy cm readily be ascertained. fn addition, the statistical error exhibited by each correlation is provided.
C!aMad0n2

Bubblepoint Presure. F@ULtSof bubblepin-pressure &kub@m are shown in Figs. 7A through 7D of Ref. 17. ,~e statistics accuracies of the correlations are provided in Table 9. Ov@l, the results were less than desirable, and Glasos correlation provided the best resuks for the entire data set. Because of the large errors, Table 9 provides the statistical results for several pressure ranges: All the m-relations performed poorly at pressurez less tb.m l,CKKl psia [6.9 MPa]. The resuks are somewhat more aiceptzble for higher pressure ranges. Snlntion GOR. Resufts of the solution GOR caladadom are shown in F@. 8A through SD of Ref. 17. Tbe statistical ncczracy is @cNided in Table 9. fn this imtznce, GIZWYScorrelation provided ~e best rewd~ however, its accuracy declines for solution GORS in excess of 1,4CNIscf/STB [252 std m3/stcc&tznk m3]. Table 9 provides the stadstical re+uh.s for the various pressure ranges, because of tbe genemf inaccuracy of the correlations, and their relationship to the bubblepoint ,pressuie correlations. titers correlation is limited in its application above solution GORS of 1,000 scf/STB 83

R$=R&,-[(Zfl)J(Bofd/Bd).

The solutinn GOR was tiuther adjusted to reflect a.presiure base of 14:7 psia [101 IcP2] to bs cxmzistent with the correlztiom. ,,, :

8PE Reservoir Engineering, Febmery 1%0

TABLE 1O-STATISTICAL

AC6UiACYOFCOI?RELATIONS
Isothermal Compre3$ibilii Undemturated Oils Calhoun -8.45 Trube .. -5.20 for , Under88turated.Gil FVF Calculations Trube -0.31 va3@z & B31ggs -0.49

Bubblepoint Pressure/Gil FVF Standing -0.90 VB2quez & Beggs -2.67 --Glaw3 1.69

Average error, %

Vazquez & Beggs 5.00

Calhoun . .0.21

St8ndard deviation Average abaolute error, % Pd8ndard deviation

2.52 1.59 2.16

257 2.77 2.46

2.28 2.28 1.55

18.17 13.07 15.17

37.84 27.99 20.57

30.78 23.20 20.74

0.54 0.25 0.52

0.78 .0.39 0.74

0.98 0.52 0.96

Undmaatumted-. Dead.Oil v@odty


Beggs

Gas-Saturated~I
Chew

@x@
Begg6

Oil Viic08ily

Average error, % Standard deviation


Aversge sbwlute error, % Standard deviation

Besl G 82.4a
30.75 18.01

il Robinson -12.27 25.07


28.01 23.95

GIBs.# 17.48 23.23


25.36 13.89

&Connally -18.84 21.71


22.14 17.80

& Robinson 11.05 19.21


17.31 13.81

Beal . ...2.91 4.64


3.23 4.24

V82quez & Beggs -0.32 5.50


3.47 -4.28.

[180 std m3/stwk-t8nk m3]. Above this value, some of the talc.lsted vahes feachd t@ uppef lhrit of the ccmefation and any extfafmhtion gave unrs8E3dc remfts. Oil FVF. Resufts of PVF cahwfstiom for cfude oil at bubblspoiat pressure are shown irIFgs. 9Atbr0ughPC ofRef. 17. Thestadstical accuracies are pmvidd. in Table 10. Acmrding to the re3~ts Of G1athe caIcufatiofm, Gksds.correlation offem the best a~w. typically undefpmdicts FVF. Standings correlaSO:Scorrel*On tion teds to ove redict FVFs greater tbm 1.2 RLVSTB [1.2 res ~3/stwk.W *! I. 31w vazquez and Beygs correlation Overpredicted FVF for the majority of c+iaes examined.
Lsothermaf Compfrx+bffity. Tbs resuls of isotfmfmal compfemibifi~ cabxdadom for undermoded cmde oifs are sbovm in Fkg.s. 10A through 10C of Ref. 17. Stsdsdcal reaultr ue provided in TabIe 10. Calbounrs correlation yiefds a singfe value of mmpressibility based on the speciric gravity of the bubble@nt .1 and provides Oie best accuracy. Although tbe errors for all Ore cOqefatiom W6 father high, the sWk-tics for the calculated FVF for undematufated oils are excslfent. This qmndty is related to ~~saiity by the fol10win8 eqnadcnr, which dampens the errm

Furthermore, Ciiew and Comtally found that rhe f8bomtorYmeasured dats can be in emorby 20%, which is comparable to the St3,tiStiC8found km. Undef@umted-6if Vf8&ity. Rmdts of the undersaturated-oif visrosity cdculadons am shown in Figs. 13A and 13B of Wf. 17. The stati$ticaf accuracies of the correlations are 8iven in Table 10. Both cmrelations gave vefy sadsfactoty remdk% hotievsf, tie msthcd of Vazquez snd Ekggs is recomm ended LW8use of tie sifnpficity of it3 fofmufation and sfigbtfy better stsdadcs. Conclrmlons
Tfd3 papers goal was to review the vsrious PVT ccmelsdons ad to compsre @eir accuracy for application in the Gulf of Mmco.

Bo>Bdexp[co(;b-p)].

. . . . . . . . . . . . . . . . . . . .. . . . ...(7)

Glaw$s correlation yields the best reatdt.?for calcuhtkxm of bubblepoint pfessme, solution GOR, snd oilFVF. ,At solution GORS above 1,402 scflSTB [252 srd m3/stcck-tauk m3] and Lwbblepint pressures gmarer tbm 7,02+3psi. [48.3 MPaI, Gla.w>s cmrelatioos for these ties show a !iecreas& in accuracy. For these cams, Vazcornquez and Beggscorrelation is more accmate. Tbe@frermsl prrxsibiliw of sn und ermhuated oil k LXst detcrmindwitb f&lboims cotilation. Dead-oil vis~sity is +stmm@j moat accurately with GfmOs correlation and the metbrxl ,of lkggs +d Robinson. Vazquez and Ikg@ correlation shbufd bk LM~ fOr .&tmmining the vkasities of gawatumted and undema~kd Ofls. Recommendations .IMSworkcompared the.-cy of sevemf PVT correfadom using seclusively mea.swed &ta br the cmrefations. Further work should be undertaken to determine the interaction betweem the ,dif6erent correlations. Thiswork w&ld entsil, for instsnce, using calculated values of solutiou GOR in the oil v correlation and rafculated vslur+ of viscosity (both for drad and gas-saturated oil) in the correlations for gas-saturated and undefsahuated-oil viscosities. flus, the accuracy of tlwe, carrekadom cOuf@be .3wmmined for institis, where meawd *P are completely factig.:. Nomsmchture , ~

These stadstks sfs also given in Table 10. Trubes coffelation, descfibed by.~gs. 4 through 6, is not adequate for volatile oifs t@t exceed tie rnethcfls mnges of speciric gravity. The w+ufadons perdorm.e+for this paper were cmnatr@ed by the rang~ :f v~u~ shown in Figs. 4 through 6. Wbife tie results of the calculations for volatile oils did not result in any degea.re in accuracy, the limbs of the correlation are evident.
Dead-Oil Vf.wasity. Rewfts of the viscosity calcuhiti~s for dead cmde oik are shown in Figs. 11A through llC of Ref. 17. The st@istfcal accuracies of the ermektions are provided in Tabfe 10. For Ods evduadon, Beds corrsfation war emapolated to temperatures up to 270F [132C] without sdversely affecting the ccmrefstions festits. Gk@s correlation showed the best accuracy of the thfee correlations examined.

G8s-Sahusted4if Vffty. Results of the viscosiw calcufadons ftir gar+atwated crude oifs are shown k Figs. 12A and 12B of Ref. 17. The stsiistiwd accuwcirs of the cOrreltiom are provided in Table 10. Chew and Connalfjs coffelation tsnded to overpredkt oil visccsity, while Beggs and Robmi wrfelation showed an opposite trend. Begg8 snd Robinsons cmrelation was found to be more accurate. AIIY improvement in sccuracj should not be expected, according to a cbmpmismof the restdt.$ fmmdby this work and the resufts determined during the ccmelations development.

a,b .= .meKlcimrts in cm-relation equations B. = oil FVF, RBISTB [res m3/stock-@,m3J Bob = oiLFVF at bubblepoi@ PS3.W% RB/STB

[res m3/stock-tank

m3]

B& = oif FVF reaufdqg from differemdal liiraha,. B@ = oil FVF at bubblepoint

B*

,., RB/STB [res m3/stock-tank m3] pressure re@dng from diKermdaI lifwadon, Rf2/ST8 [ES m3/stock@@m3] = ofi, FVF at bubblqoint plWSllIe resufting %rn M likration, RS/S~ IRS m3@ock-t8nk m.3j

ti

SPE Reservoir Engineering, February 1P9+I

co = isothemal mqmessibtity of ojl, psi-l @a-l] CP, = pssudoreduced compfessibfity c,,c~, C3 = coefficients in V~quez and Beggs correlation Ei = error (measured value -cafdated vzlue X 102/mea3ured value), %
~ = avetige eyor, EE<ln, % IE[ = average absolute error, Z@i[/n, % FP = bubble@nt:po&sure factor m = @z-efficient in Vazquez and Beggs correlation M. = effwdve moleaik weight of stcck-tank ofl n = n~~r of Ob~~atiOps ND = correlating number for calculating FVF NPb. = correlating number for cahdating bubblepoint presmre p = resxvoiz pressure, psia Na] Pb = bu,bblevzint pressure, psia ~a] PC = cmcaf pressure, ptia &Pa] pp, = ~eudoreduced pressure P.p = 3oparat0r pressure, psia ma] R = ~Iution GOR, Sti/S~ [std m3/stwk-tank m3] s R,p = gas Iibefated as a resolt of flashing a bubblepoint oil, scffSTB ~std m3/stock-tank m3] (Rl)w = gas Iibefated as a resuft of differentially lowering the pressure below bubblepoint, scf/STB [std m31stmf@aok m3] T = reservoir temperature, F ~C] T, = critical temperamre, R ~ TP, = pseudoreduced tempefamre T.p = sqnrator temperature, ,F X, Y,Z = coefficients ,Jn Beggs and Robinson correlation Yc = T.mI = 78 = ?g(p) = 78s = 70 = _f06 = 70m 7&o I% kob Ad mole ff~tion of g% ti Solution st~k-ti C@gravity, API tg/cm3] total gas, specific gravity (air= 1) @ spe.~c gmvit? al pressure p (air= 1) s~P~at~r-gas spafic gmvity (air= I) Ofi s~fic gravity (water= 1) ofi SPWfIC gravity at bubblepoint pressure (Water=l) ,,

8. 23%2$,H.D. and Robinson, J. R.: Tsdmadng tie Viscosity of CN6e Oil Systems: JPT (Sept. 1975) 1140-41. ( fly, C.A. Jr.: A Vkcmity Correlation for Gas9. Chew, 1. and M Salur?ted Cmde Oifs, Trans., MhfB (1959) 216,22-25. 10. Chierici, G.L., Cicci. G.M., and ScIcdd, M,: l%m-Pbaw VerdWI Flow in OJ Wdfs-Predicdon of pressure Dmm, , JPT(Au% 1974) . ...,. , 927-3% Trm., .4fb MS, 237. Ii. Ostermann, R. D., E3dig-Emnomides, C. A., and Owakdd, 0.0.: I congelations for the Reservoir Fluid Pro@ies of Alaskan Crudes,,, ? paper SPE 11703pmsented at fhe 1983 SPE CaOf..mia Reqional Meet. ills, Velmu-a, Mad i 23-23. 12. Standing, M.B.: ( )il-system Correlations,>, Pefrokrun Produc!im Handbook, .C. Frk :k (d.), SPE, Ricbald.wn, TK (1%2) 2, Chap. i9. T 13. Standing, M. B.: VM-C mufPba.w Bduniorqf Oil FiebfJJ@ocar. bon Sy~en!& SPE, Richardson, TK (1977). 14. Watson, K.M., Nelson, E.F., and Murphy,G.B.: Charactedzation of Petroleum Fractions, Ind & E.g.. C&n. (1935) 1460-64. 15. Brown. G.G. et af. : Nutumf Gasoline mid Zk VokuileH@roc.wbons, Naturai Gas Asn. of ~iri;: ti-m-: OK194~;4R; f ... . ... 16. Amyx, J.W., Bass, D.M. Jr., and whiting,, .R.L.: Pezrolewn Reservoir i?ngineerin~: Physical Prqxrtks, McGraw-Hill Book Co. fnc., New York City (1960) 395-97. 17. Suffon, R.P. and Farshad, F.: Supplement m SPE 13172, Evahadon of Fmpiricany Deriv4 PVT Properdes for Gulf of Mexico Cm& Oil?,,$, paper SPE 20277available fmm SPE Bwk Order Dept., Richardson, TK. 1S. .&@ K., Govier, G.W., and Fogarasi, M.: pressure Drop in WdIs Prcducing Oil and G88, 3. @n. Pet Tech. (July-Sept. 1972) 38-48. AppendixCorrsistlon Bubblepoint Pmssnre Equations

Cafcufations.

Standing. 13 Pb 18:21(Rs/7g)0.83 mtiog(O.C0091T-O.0125TApI) 1.4].


Lu.wfer. ! Y8=(&/379.3)/[(Rs/379 .3)+(35070/Me)]

and Pb = [@f)(T+459.67)] /yl, Wherepf and .UO are determined from Figs. 1 and 2, resjwcdvely. Vazquez.and Beggs. 3
Pb = {( CIRs/7~s)~tibzg[ C3YAPIKT+459.67JI} l/cZ >

= sP@fic ~avity of tirsatumtd ~fi at ~eSWoU .@mperalure (water= 1) = sFCifiC gfavity of undersaturated oi3 adjusted to 643F (water= 1) = V@qsity of undematurated oif, cp [ml%.s] = viscOsity of gas-sahmted 01, w [mPa. s] = viscosity of gas-free 01, cp [mPa. S]

where C, =27.64 for ?.z.. s30 and 56.06 --- wiw>30. for ,=. . --C2 = 1.09~7 for yfiS3~and 1.1870 for yNI >30, and C3=11. 172 for .ym s30 and 10.393 for Yml >30. Gfasc6.4
Pb ~tfiOg[l.7669+1.744710g Npb 0.30~18(Iog .989)

Npb)21,

Acknowledgments We thank the management of Mafathon Oil Co. for p-emission to publish tlds article and the U. of Southwestern Louisiana, where this research was undertaken. References 1.Wdi&, M.B. A Pmss_Volume-Tempemtu ComdationFor hfixtwes of California Oils and Gases,,, D,-iIl. & Prod P,..., API (1947) 275-87. 2. Lasater, J.A.: ,Bbble point Pressure Correlation,Ss Tmrs., A2ME (1958)213, 379-81. M.E. and L@@, H.D,: ConeJadors for FIuidPhysicalpmp 3. VW, erty Prediction,,, JPT (km 1980) 968-70. Correlation, 4. Glas+?, Yieneralkd PmweVolume-TemFrature 0.:

where NPb = (R,l@o Solution

81$ 0.lWTm10 (T

GOR Correfaliom.

R.=-/g L7.wer.

[(
%

+1.4

1.2C48 -0.CG29iT)

antilog(0.0125yMr

Pj=[(Pb)(Yg)l/(T+459.67)

and R.= 132,755yJg/MO(l

yg),

of Resewoir Engineering, U. of Ok lahonm Press. Nmuwi. OK 0947) 35. 6. Tmbe, A-3.: ;Compre&Mi6 of LfndemamratedHydmarbm RewrVOUFbiidi, Trans., A2&fE (1957) 210,341-44, 7. Bed, C.: The Viscosity of Air, Water, Namrd Gas, Crude Oil and Its Awxiated Gases at Oil Field Temperatures and Pressures,,, OiI Evaluation and Rewve Estimates,Reprint Series, and Gus Propq SPE, Ricbardsmt, TK, (1970)3, 114-27.

JPT (May 1980) 785-95. 5. Cdhoqn, J.C,. Jr.: Funabmnfs

where yg and MO are ,dewmzined from Figs. 1 and 2, respectively.

Vazquez andBeggs. R,=(Tg@bcz/C1)mtilog[C37mr/(T+459.67)1 >

where Cl =27.64 for 7mIs30 and 56.06 for ym >30, C2 = 1.0937 for T*l s30 and 1.1S70 for yeI >30, and C3 = 11.172 for -ymI <30 and 10,393 for YUI >30.

SPE Reservoir Bn@eerimg, February 19S0

85

(%3!4
&
=@b(~M~0.939/~0.

Begijs and Robinson. 8

ln)]l.fis ,
.i811-3.ti3 IOg Pb).s].

p~=lw-1, wheie X= YT-l.lfi, G@. #&=(3.141


X 101a) T-3.WI.Iog

Whm NPb =mtiog~.88@-(14 033 FVF c0iTetat30ils.

Y=l#,

and z=3.0324o.02023iAPI.

smdirlg.~
B&=0.972+ 1.47X 10-4@,(Y@0.5
+

-yml)[10.3:1300E m-36.4471,
-

1.25T11.17S.

Ga3&tmnt&koif

Viscdy C0r3@b3tiomi

?iiqwz and Beggs.


chew Landcullnalb.9Js

B&=l.o+clR.+c*(T-@)
whsre Cl =4.677x10-4

( )
~ ?$.
foi 7MS30

+cJf$(T-60)

()
% ?*. ,

JLob=a(lkd)b,

and 4.670x10+ for 7M>30, C2=1.751X10-S for~M1s30 and 1.100 x10-5 for 7AP3>30, and Cs= I.8106x10-8 for ~m1S30 and 1.337X 10-9 for y~ >30.

where a= O.20+0.80 antilog(-0.00081 0.57 and.10g(-o.oOonRJ.

R,)

and

b= O.43+

Beggs and Robinson. A+ =WJ,


where a=10.715(R, + 160) -0.5i5 and b=5.44 (R, +150) 0.338.
Viity iorrdsths.

Gk@. Bd =1.0+.md16g[-6
O.27683@S ~~m NB)2], .58511 +2.91329 10g NB

u331iers3tnkited-oil

NB =R$(~8/yo)0.5M +0.968Z
C0mpmwi33ifity C0rrek3ti0313.

Bed.
po=p~+O.OO1@-pb) (0.@4p&16+0.W8pob0 56).

Unde3vstmWed-Oif
c7fh014n5 (see F@

3).

k?quez Beggs. Und


/% f%b@lpb)m. whers m=2.6p1.1rnandlog[( -3.9 x 10-5) (P-5.0)].

Y* =(70 +2.18X 10-4@J/Bo.

Tiubd (We Figs. 4 tbiougb 6). TP,$(T+459.67)ITC, Pp, =P~P.7


and CoCP,/PC. vazqu& k

si Motrlc
API bbl m ~~ psi R scfhbl
Tawe.lon

COnvOrslon Factors 141.51(131 .5+API) X 1.589873 x 1.0* (oF-32)/l.8


X

~_ol

:y3

E-03

= i%
= o~ =K = std rn31m3

Beggs. 12.60

O. O=[(- 1433.0 +5.0) (R, +17.2)( T-1 18W(7W+

6.894757

E+OU = kpa
E01

x7ml/(pxlo5). D@d-oi3 Vi3c03ityCorratiom%


Bed. 13 360/(T+2WJ)], Pmi=[0.32 +(1.8 X 107)/yAF.1453][
where a=andlog[O.43

Rfl.8 X 1.801175 exact.

fmxor b

WERE sept. 6,334. raw 1 !


acqted

-WfIal SPE .muwpt I!C+I w, mm. Rw


praented at tha W34 SE

rwni+ torpiw
Ann!Jal Ta3@.A

nmmampt maim JUT 9.


C@t?renca

forPublh.

+(8.33 /yMI)].

7$s9. FW W+ 7SW fl~ InHouston, m4 EMbnim IwJd

344X, 13-19.

S6

SPE R&rmir

E@ndng,

FebrwY

1990

Potrebbero piacerti anche