Reservoir-Fluid Property Correlations—State of the Art W.D. McCain Jr.,* SPE, Cawley, Gillespie & Assocs. Inc. ‘Sunemary. This paper presents correlations to determine reservoir-luid properties from field data. The best available correlations were selected by comparison with a data base of hundreds of reservoir-flud stdies of samples representing all areas ofthe free world involved in active petroleum exploitation from 1980 to 1986. Also, correlations of formation-water properties are given Introduction ‘Values of reservoir liquid and gas properties are often needed when laboratory PVT data are not availabe. This paper shows how to use normally available feld dala to estimate fluid properties. ‘While at Texas A&M U., Thad access to a data base of hundreds of reservoir-fuid studies provided by Core Laboratories Inc. The ‘geographical and geological origins of the reservoir samples had ‘been carefully removed from the data but the samples were known, to represent all areas ofthe free world in which petroleum exploi- {ation was active during the first 6 years of the 1980's. All eservoir-fluid property correlations available inthe petrole- ‘um engineering literature were compared with this data base. This paper gives the best correlations. Identification of Reservolr-Fluld Type ‘Surprisingly accurate ‘‘roles of thumb" are available! to identify ‘eservoir-fid type from field data. When the inital producing GOR is <3,300 scf/STB, the fluid isa liquid at reservoir conditions. Possibie exceptions occur if the stock-tank liquid is colorless of has a gravity higher than about 5O° API Reservoir liquids are either black oils or volatile oils; the gener- al material-balance equations? work only for black oils, The be- Iavior of volatile oils does not fit the assumptions inherent inthe derivation of the materal-balance equations. Black ols are identi- fied as having initial producing GOR’s below 2,000 scf/STB and deeply colored stock-tank oll with gravities below 45° APT. Reservoir gases are clasified as retrograde gases (often called condensate gases or gas condensate), wet gases, and dry gases. Retrograde gases have inital producing GOR's >3,300 sef/STB. ‘The few exceptions of ols that ave ratios higher than this are iden tified as having deeply colored stock-tank liquids with gravities < 40° API. Retrograde behavior occurs for gases with inital produc ing GOR's of 150,000 scf/STB or higher; however, as a practical ‘alter, gases with initial producing GOR's 2 50,000 scf/STB can, be treated as wet gases. ‘The term wet gas is used fora gas that does not release conden- sate inthe reservoir but does form hydrocarbon liquid atthe sur face. The term dry gas is used for a gas that does not form any Inydrocarbon liquid atthe surface. In this context the tems “wet” and "dry" do not refe to water or Water vapor, which is always present to some extent. Properties of Reservoir Liquids ‘The physical properties discussed next apply onl to black ols. En ‘ineering a volatile-ol reservoir requires a special laboratory study ‘ot discussed here. Solution GOR at Bubblepoint, Ryy- The initial producing GOR ‘provides a good estimate of solution GOR for use at pressures equal, to and above bubblepoint pressure. This will not be tue if fee gas froma gas cap or another formation is produced with te oil. Field data ofen exhibit a great deal of scatter; however, a trend of con- stant GOR usually can be discemed before reservoir pressure drops, ‘below the bubblepoint. ‘Often the reported values of producing GOR do not include stock- tank vent gas. In this case, the use of intial producing GOR for Cony 90 Sot of Poseum Engrs 26 solution GOR results in values that are low by 10% or more. The Socktank GOR ean be estimated wit Jog Rsp=Ay +Ap log 79 +A3 log ygsp +a log Psp +43 log Top, vesseneensee where 4;=0.3818, 4,=—5.506, 4,=2.902, A As= ~0.7355. Bq. 1 should not be used if the separator tempera- ture is > 40°F, ‘Addition of the estimate of stock-tank GOR from Eq. 110 the separator GOR resul in an estimate of solution GOR accurate 1 within 3 Bubblepoint Pressure, py. The bubblepoint pressure of the oil at reservoir conditions can be estimated with* Po=18.Cpp-1.4), where Cy) 100031700124) to an accuracy of 15%. The specific gravity ofthe separator gas can be used for 7g; however, R, should include stock-tank vent 22s. The equations are valid to 325°F. 'A more accurate estimate of bubblepoint pressure can be obtained it weservoir pressure is measured regularly. Plot reservoir pressure and producing GOR vs. cumulative production. For a volumetric solution-gas-drive reservoir, pressure will decline rapidly initial- ly, then flaten when reservoir pressure drops below the oil bub- blepoint pressure (the pressure at which the line changes slope) ‘The producing GOR will begin o increase shorly after bubblepoint pressure is reached. Solution GOR, R,. Eqs. 2 and 3 can be used to estimate solution GOR for pressures below the bubblepoint. Enter any pressure be- low bubblepoint in place of py in Eq. 2 and calculate the corre- sponding value of solution GOR with Eq. 3. The results should be within 15% of measured values. fa field-derived bubblepoint pressure has been obtained from pressure measurements as described above, the accuracy ofthe es- timates of solution GOR can be improved.} Start by creating a ta- ble of pressures and solution GOR's. Subtract the field-derived ‘bubblepoint pressure from the bubblepoint pressure calculated with Eqs. 2 and 3 to obtain a“ delta pressure.” Subtract tis “delta pres- sure’ from all pressures in the R, vs. p table. This procedure ‘works very well for pressures near the bubblepoint. I i less a curate at low pressures. OW FVE, B,. The oil FVF for use at pressures equal to or below ‘ubblepoint‘can be estimated with* Byy=0.9759+12(10-5)Cpa'2, sees) where Cans Rsig/t=)°5 41.257. : 6 “The eto canbe ned fray presi el 10 or blow the ‘bubblepoint by inserting the corresponding value of solution GOR ‘timated as discussed above. The resuling FVF vale wl be within 5% of laboratory-measured values if accurate values of solution GOR are used. If solution GOR’s are obtained with Eqs. 2 and 3, the accuracy ofthe resulting FV values willbe some unlmown combination ofthe 15% accuracy of Eqs. 2 and 3 and the 5% ac- curacy of Eqs. 4 and 5. Do not use at temperatures above 325°F. SPE Reservoir Engincering, May 1991‘At pressures above bubblepoint pressure, the oil FVF is caleu- lated with Bo=Bey expleo(Po-PIl- © ‘where FVF atthe bubblepoint is estimated as discussed above, Es- timation ofthe coefficient of isothermal compressibility of ol, cy, is discussed Iter. Oil Density at Reservoir Conditions, pop. Eq. 7 may be used to calculate’ the density of the oil in the reservoir at bubblepontpres- ‘sure (and below) from estimated values of B, and R,. Pan sr0 +0.01357R,74)/Bos - o ‘where pag isthe density of the reservoir liquid atthe pressure and temperatize at which B, and R, were estimated. A weighted aver- ‘age of separator and stock-tank-gas specific gravites should be sed {or yg; however, the use of separator gas gravity gives adequate results, The accuracy of this calculation should be some unknown, ‘combination ofthe accuracy of the estimates of B, and R,; how- ever, in practice, the calculated density is within 5% of laboratory ‘measured values. Density of reservoir ol at pressures above the bubblepoint can be calculated with Po=Pob €XPICa( P~Ps)}: Coefficient of Isothermal Compressibility of Oil, cy. The coffi- cient of isothermal compressibility of ol, often called oil compress- ibility, is defined for pressures above the bubblepoint as. @) co=—CUVY@VIap)r, - (9a) om ~(1/B.)(@B,)/ pr. ) oF ¢4=(Np_NAP0/9P)r- (Se) {At pressures below te tablet, i compres is dened 1//aB, aR, —| -a,() |... Bap /, Nap /, ‘At pressures sbove the bubblepoat,” omy tag, tAsT Agr tAsrapd/AsPs ro) ‘where Ay = 1,433.0, Ap=5.0, Ay=17.2, Ag=—1,180.0, As 1261, and 4g~108: Values of ol compressibility calculated with Eq. 11 are general- Iy ow, by as much as 5O% at high pressures. Accuracy i improved at pressures ner the bubblepoint, The data set used to develop this ‘equation included pressures as high as 9,500 psig: however, the author did not give a temperature range. * "AL pressures below the bubblepoint,” In(¢,) = ~7.633~1.497 In(p)+1.115 In(P)+0.533 Intyae) ++0.184 In(). (2) ‘The results are accurate to within 10% at pressures above 500 pris. Below 500 psi, the accuracy is within 20%. 1 the bubblepoint pres- sure is known, the accuracy of estimates of oil compressibility at pressures below bubblepoint can be improved by using? Ja(cq) = ~7.573~1.450 In(p)~0.383 In(pp)+ 1.402 I(T) £0.256 In(y ap) +0.449 Inf). 3) Eqs. 12 and 13 ate valid to 330°F and 5,300 psia. «10 Oi Viscosity, uy. Estimation of ol viscosity at pressures below the bubblepotnt a two-step procedure. First, the viscosity ofthe cil without dissolved gas (dead ol), ap, is estimated at reser temperature og log(40+1) £8653 0.025086 ,p1—0.5644 log T. . a4) Bg, His based on data with ages of $10 58°APTand 6010 175°F. ‘SPE Reservoir Engineering, May 1991 ‘Then, the effect of solution GOR corresponding to the pressure of interest is taken into account with! Hom Awan? where A=10.715(R, +100) 0515 a6) and Bes.44(R, +150)-033, a Eqs. 15 through 17 were derived with data to 295°F and 5,250 psig. Oil viscosity at pressures above the bubblepoint is estimated by first calculating viscosity atthe bubblepoint with Eqs. 14 through 17 from the solution GOR atthe bubblepoint and then adjusting Viseosity to higher pressures with? Hom Heal DiPs)®, where B=Cyp% exp(Cs+Cyp), - a9) and C) =2.6, Cy=1.187, Cy=~11.513, and Cy=-8.98%10-5. Eqs. 18 and 19 were developed from a data base with pressures to 9,500 psig; the applicable temperatures were not given. * It appears thatthe relationships of ol viscosity to other proper- ties ofthe oil are too complicated to be explained by the limited field data normally available. Thus, the values of ol viscosity cal- culated with Eqs. 14 through 19 should be considered tobe “order ‘of magnitude" estimates only Properties of Reservolr Gases Properties of dry gases will be considered first. Then, the calcul tions necessary for estimating properties of reservoir wet gases will be discussed. Retrograde gases will not be considered because a special laboratory report is requited for these gases, Compressibilty Equation of State. The equation of state most often used by petroleum engineers is pV=2nkT. = (20) ‘The Standing Katz? correlation of « factors has stood the test of time. Theit graphical correlation may be represented by 5 +A tal Tyr +As Typ? 4a! Tor’ +ASlTo Vopr FAG + Aa Tyr Aa! Tye? Doge? ~A9(A TIT ¥ ABT bp’ $A 01+ 1109770 pr2ITpy 3 XR(—A pr?) and ppp = 0.21 Ppl Ie : . 3265, Ay ~1,0700, 430.5339, 44 =0.01569, 0.05165, 44=0.5875, Ay=~0.7361, dy=0-1844, y= 011056, 490.6136, and Ay) "0.7210. Bq. 21 represents the Standing-Katz correlation to within 1% for 0.2 1.801 175 (she 285) SPERE Author’s Reply to Discussion of Reservoir-Fluid Property Correlations—State of the Art William D. McCain Jr., SPE, S.A. Holditch & Assocs. Inc. Hayworth may have found an example reservoir Nuid that doesnot fit the ‘rues of thumb" for determining reservoir fluid type.12 However, he does not give the color ofthe stock-tank liquid. Stock tank oll colors the third rule of thumb and is useful in confirming fluid type. ‘Composition was not considered because the paper? was written to show how to use field data to estimate reservoir fuid properties (see the first sentence ofthe Summary or the second sentence of ‘the Introduction). Extensive information in the use of field and laboratory data to distinguish between black and volatile oils is available elsewhere,? 510 References 1. Moses, P.L.: "Engineering Application of Phase Behavior of Crude Oil and Condensate Systems," JPT (ly 1986) 15-23 2. MoCain, WD. Jr: "Reseroit-Fuid Property Corelations—State of the Art." SPERE (May 1991) 266-72 3, MeCain, W.D. Ie The Properties of Petroleum Fluids, second edition, PeanWeil Books, Tulsa (1989) 149-53 (sre 200) SPERE SPE Reservoir Enginering, November 1991