Predicting Phase Behavior of

Downloaded from http://onepetro.org/JPT/article-pdf/30/11/1649/2227059/spe-6721-pa.pdf/1 by Pontificia University Catolica Brazil user on 07 March 2025
Condensate/Crude-Oil Systeins Using Methane
Interaction Coefficients
D. L. Katz, SPE-AIME, U. ofMichigan
A. Firoozabadi, Abadan Institute of Technology

Introduction
Prediction of phase behavior of complex hydrocarbon behavior up to pressures of 64 121 kl>a (9,300 psia) is
systems has improved gradually. Recent papers using shown for several crude oil and condensate systems.
equations of state for predicting phase fugacities of con-
stituents have extended their usefulness tO higher boiling Calculations of PhaSe Behavior
components .1-5 Using extended analysis with boiling The equation of state, using the relationship of fugacity to
poirits of higher boiling hydrocarbon groups and their partial moial volumes; is used to compute the fugacity of
densities and molecular weights provides a basis for each componept .in both liquid and gaseous phases. At
improvement when. predicting phase behavior. At the equilibrium the fugacity of each component in liquid
high pressure ofres~rvoir conditions, properties ofthese equals its fugacity in the gaseous state.
C6 +groups are important. Basic parameters for each constitutent are the critical
Using the Peng-Robinson5 equation of state has the properties, normal boiling point or vapor pressure. For
advantage of improving accuracy when calculating liquid hexane arid ht(avier groups present in natural-gas and
density. When used by Bergman et al. 4 , 6 with extended crude-oil systems, average normal boiling point, liquid
gas analyses, the equation gave good prediction of dew density, and molecular weight are used 7 instead of mea-
points and liquid yield at pipeline conditions. Bergman et sured properties available for pure components.
al. used interaction coefficients but did not completely . For systemS containing various mole.cular families, the
explore their usage since phase behavior of condensates interaction coefficients between the primary constituents
at pipeline conditions appeared rather insensitive to the must be include{! to account for the degree of compatibil-
C6 + interaction coeffiCients. ity. Hete methane arid higher boiling C6 + groups are the
This paper reports the ex tension of the Peng- principal components for which interaction coefficients
Robinsori-AGA procedure 6 for phase prediction of con- appear significant.
densates and natural-gas/cnide-oil systems at reservoir Since properties of C6 + groups in crude oil and conden~
conditions. New interaction coefficients were found sates are needed here, generaiized properties are given in
necessary for methane and C6 + constituents (especially the absence of measured values ..
the highest boiling group, such as C 20 +) so that experi-
mental behavior could be matched by prediction, .Steps Properties of Hexanes and Heavier Groups
for developing interaction parameters are presented. Average boiling points, liquid densities, and molecular
Then, agreement between measured and predicted phase weights Were reported by Bergm~n et al. 6 from C6
0149-2136/78/0011-6721$00.25
through C 15 , based on analysis of 26 condensates and
@1978 Society of Petroleum Engineers o.f AIME naturally occurring liquid hydrocarbons. This correlation

This paper extends the Peng-Robinson-AGA procedure for computing phase behavior to include
gas-condensate and crude-oil systems at reservoir conditions. Interaction coefficients from
methane to the C 6 + groups were correlated with liquid density. With these coefficients, the
behavior of gas condensate systems may be predicted reliably.

NOVEMBER, 1978 1649

 sao~~~-~~-~~-~~-~,-~~-~~

r- I ! I I I I I I j//_
o Bergman, Tek a Katz //
400r--- J!
x
•
Liquid}
Vapor Hoffman, Crump a Hocott //
/--1----

~ r- i I I I ;.t -

""'.......E I /.
Y
w '

ro-
I
0
>-
1-
Ui 0.7
l I
z
1<.1
0
~ 200 c- I I ,.,Y -- c',o-+---·+------1-

'oo----;~~c:0~--1~ 1

Downloaded from http://onepetro.org/JPT/article-pdf/30/11/1649/2227059/spe-6721-pa.pdf/1 by Pontificia University Catolica Brazil user on 07 March 2025
0.5.L___-----L-_L______!__L_~__JL__--'-----'---'-----'---'--------'.-----'
-50 0 100 200 300 400 500 -100 0 I00 200 300 400 500
NORMAL BOILING POINT, °C NORMAL BOILING POINT, °C

Fig. 1-Dens ~y vs boiling points of hydrocarbon groups in crude Fig. 2-Molecular vs boiling points of hydrocarbon groups in
oil and condensate systems. crude oil and condensate systems.

was· extended to a normal boiling point of 500°C using

added data8 • 9 for use with reservoir condensates and Development of Interaction Coefficients
crude. oils. Table 1 and Figs. 1 and 2 give average proper~ rllitial trials of predicting bubble points of crude oils with
ties for groups of compounds boiling between 0.5°C the Peng-Robinsori-AGA procedure failed to give good
above the previOUS n-paraffins 10 and 0.5°C above the agreement, even though extended analyses up to C 20 +
carbon immber used to identify the group. The higher were used. The interaction coefficient, used earlier by
boiling point properties are only best estimates and are Zudkevitch and Joffe, 11 was inserted in the equation of
used for systems for which properties and/or analyses state by Peng-Robinson. 5 Using alternative interaction
have not been performed at these high boiling points. coefficients showed promise as a way of modifying the

TABLE 1-GENERALIZED PROPERTIES OF PETROLEUM HEXANE PLUS GROUPS

Average
Hydrocarbon Boiling Range Boiling Point Dens~ Molecular
(OC) (oF)
Group ~ ~ (g/ml) Weight
c6 36.5 to 69.2 97.9to156.7 63.9 147 0.685 84
c7 69.2 to 98.9 156.7 to 210.1 91.9 197.5 0.722 96
Cs 98.9 to 126.1 210.1 to 259.1 116.7 242 0.745 107
c9 126.1 to 151.3 259.1 to 304.4 142.2 288 0.764 121
clO 151.3to 174.6 304.4 to 346.4 165.8 330.5 0.778 134
Cu 174.6 to 196.4 346.4 to 385.5 187.2 369 0.789 147
C12 196.4 to 216.8 385.5 to 422.2 208.3 407 0.800 161
c13 216.8 to 235.9 422.2 to 456.7 227.2 441 0.811 175
c14 235.9 to 253.9 456.7 to 489.2 246.4 475.5 0.822 190
c1s 253.9 to 271.1 489.2to520 266 511 0.832 206
cl6 271.1 to 287.3 520 to547 283 542 0.839 222
c17 287 to303 547 to577 300 572 0.847 237
C1s 303 to 317 577 to603 313 595 0.852 251
c19 317 to331 603 to 628 325 617 0.857 263
C2o 331 to344 62S to 652 338 640.5 0.862 275
C21 344 to357 652 to675 351 664 0.867 291
C22 357 to369 675 to696 363 686 0.872 305
c23 369 to381 696 to 717 375 707 0.877 318.
c24 381 to392 717 to737 386 727 0.881 331
c25 392 to402 737 to756 397 747 0.885 345
c26 402 to 413 756 to775 408 766 0.889 359
C21 413 to423 775 to793 419 784 0.893 374
C2s 423 to432 793 to 810 429 802 0.896 388
c29 432 to441 810 to826 438 817 0.899 402
C3o 441 to450 826 to842 446 834 0.902 416
c31 450 to459 842 to857 455 850 0.906 430
c32 459 to468 857 to874 463 866 0.909 444
c33 468 to476 874 to888 471 881 0.912 458
c34 476 to483 888 to 901 478 895 0.914 472
c35 483 to491 901 to915 486 908 0.917 486
c36 493 922 0.919 500
c37 500 934 0.922 514
Cas 508 947 0.~24 528
c39 515 959 0.926 542
c40 522 972 0.928 556
c41 528 982 0.930 570
c42 534 993 0.931 584
c4a 540 1,004 0.933 598
c44 547 1,017 0.935 612
c45 553 1,027 0.937 626

1650 JOURNAL OF PETROLEUM TECHNOLOGY

procedure. Accordingly, interaction coefficients for 0.07 I I I I I

methane binaries were compiled based on bubble points

along with descriptive properties that mightallow a corre- 0.06 "
lation related to known properties of the c6 + groups. The
1-
z
w
u 0.05
(}

/
interaction coefficientS obtained are the values that gave
minimum error for the bubble point for the range of
u:
u.
~ 0.04
/v
temperature and pressure indicated in Table 2, using data u
z
:/
z
from Refs. 12 through 20. 0
i= 0.03 0

The use of liquid density at 15.5°C (60°F) for the ~

a::
binary component in the methane binary. gave a cor- ~ 0.02 r - - - - 0

~
relation of the interaction coefficients for these systems
(Fig. 3). 0.0 I
Interaction coefficients for nonhydrocarbons used here

Downloaded from http://onepetro.org/JPT/article-pdf/30/11/1649/2227059/spe-6721-pa.pdf/1 by Pontificia University Catolica Brazil user on 07 March 2025
I I I J I
are 0.12 for nitrogenhydrocarbons, 20 • 21 and 0.15 for C0 2 0
o.5 0.60 0.70 0.80 0.90 1.0
DENSITY AT 15.5°C, g/m!
hydrocarbons. 6 Ethane and propane interaction
coefficients were considered with other hydrocarbons. Fig. 3-lnteraction coefficient from methane to other
hydrocarbons as a function of the density of the heavier
Because of the variation fm.ind with· temperature and hydrocarbon.
pressure, and the relatively small influence on the predic-
tions, the value of0.01 used in the Peng-Robinson-AGA
procedure was continued. The interaction coefficients
between methane and ethane through n-pentane were Hoffman et al. 8 Lean-Gas Condensate
zero. Later, the significance of the interaction coeffi- Hoffman et al. 's data on the gas-cap material at the
cients for methane and the last composite group, such as reservoir temperature of 93. 9°C (201 °F) are the rriost
c2{) +' will be discussed. complete because true boiling-point distillations with
densities and molecular weights were determined on cuts
Natural-Gas Condensate Systems corresponding to n-paraffins up to C22 • The analysis
The natural-gas condensate system used when comparing given and the densities and molecular weights for each
experimental data with calculated values is listed in Table cutwere used to predict (1) dew-point pressure, (2) liquid
3. The measured dew-point pressure at stated tempera- yield as a function of pressure down to 3551 kPa (515
tures, the liquid yield when flashed at given temperatures psia), and (3) vapor-liquid equilibrium ratios all at
and pressures, and the corresponding equilibrium ratios 93. 9°C. Table 4 compares experimental data, predicted
are used to give three methods for comparison. For each values in this study, and published predictions by
system, the extended analysis was available. The predic- Starling .1 Fig. 4 compares equiW>rium ratios. The dew-
tions for the condensates were based on the interaction point pressure and liquid-yieHlpf.~~ictions are excellent,
coefficients of Fig. 3, which is correlated on liquid den- and predicted equilibrium ratios for methane have an
sity of the hydrocarbon groups as developed here. For average absolute error of0.76% with experimental data.
natural-gas condensates, none of the experimental data
being compared were used to generate the calculation Starling 1 Rich-Gas Condensate (Fluid B)
procedure. Each comparison will be discussed in tum. Starli?g reported the analysis on a rich-gas condensate
TABLE 2-INTERACTION COEFFICIENTS FOR BINARY METHANE/HYDROCARBON SYSTEMS
Boiling Density Pressure Temperature
Heavier Interaction Point (g/ml, Molecular Range Range
Component Coefficient ~ 15:5"C) Weight .(kPa) ec) --
Ref.
nC4 0.02 -0.50 0.584 58.12 41 oo to 12 400 71.1 ,87.8, 104.4, 121.1 12
nCs 0.02 36.07 0.6312 72.15 4100·to 16 100 71.1 '1 04.4, 137.8 12
nCs 0.025 68.74 0.6640 86.18 4050 to 16 200 75,100,150 13
nC1 0.025 98.43 0.6882 100.20 4100 to 20 600 71.1 '104.4, 137.8 14
nCs 0.035 126.66 0.7068 114.23 4050to7090 75,100,150 15
nC9 0.035 15,0.80 0.7217 128.26 4050 to 30 400 75,100,150 16
nC10 0.035 174.12 0;7341 142.28 4100 to 32 750 71.1 '1 04.4, 137.8 12
nC:w 0.054 343.8 0.795 282.57 500to 6080 40 17
Benzene 0.06 80.09 0.8844 78.11 9970 to 24 200 37.8, 71.1' 104.4 18,19
Cyclohexane 0.03 80.72 0.7834 84.16 41 00 to 24 100 71.1 '1 04;4, 137.8 20

TABLE3-GAS-CONDENSATE/CRUDE-OtLSYSTEMSUSEDINCOMPARISONSOFEXPERiMENTALWITH
PREDICTED BEHAVIOR
Temperature Range Pressure Range
Authors System (OC) (of) (kPa x 1o-3 ) {psia)
Hoffman, eta/. 8 Condensate, crude
oil 83.89 201 26.49 to 3.55 3,842·to 515
Starling1 Condensate 101.11 ' 121.11 214,250 26.16 to 3.55 3,795 to 515
Roland 23 Crude o~ 48.89, 93.33 120,200 7.22 to 64.63 1,047 to 9;374
Standing and
Katz2" Crude oil 48.89, 121.11 120,250 6.96 to38.13 ~ ,010 to 5,530
Katz and
HachmutljS Crude oil 93.33 200 0.813 to 16.54 118to 2,399
Evans and Harris9 Crude oil 87.78 190 7.65 to 41.37 1,110 to 6,000

NOVEMBER, 1978 1651

 TABLE4-COMPARISONOFPREDICTEDWITHEXPERIMENTALDATAFORGAS-CAPCOMPOSITIONOF
HOFFMAN eta/. 8 AT 98.9°C
Liquid Yield*
, Pressure Experimental This Work Starling 1
(kPa) (psia.) {m3 /m 3 x1 0 6 ) (bbi/MMscf) {m 3 /m 3 x1 0 6 ) (bbi/MMscf) (m3 /m 3 x1 06 ) {bbi/MMscf)
20098 2,915 51.02 9.07 54.50 9.69 63.56 11.3
17340 2,515 69.97 12.44 70.25 12.49 77.62 13.8
13893 2,015 87.52 15.56 84.49 15.02 90.56 16.1
10445 1,515 95.51 16.98 92.87 16.51 98.44 17.5.
6998 1,015 95.29 16.94 94.50 16.80 97.31 17.3
3551 515 84.82 15.08 86.00 15.29 83.81 14.9
Experimental dew,.po int pressure is 36 490 kPa (3,842 psia). .
Predicted dew point for this study is 26 682 kPa (3,870 psia), Starling 1 > 35 267 kPa (5, 115 psia).
Average absolute deviation between experimental and predicted yields forth is study is 2.6%, and 8.2% for Starling.

Downloaded from http://onepetro.org/JPT/article-pdf/30/11/1649/2227059/spe-6721-pa.pdf/1 by Pontificia University Catolica Brazil user on 07 March 2025
*Liquid volume at reported pressure and temperature; gas standard conditions for metric units are 15°C and 101.325 kPa.

with extended analysis to C 20 , but did not include liquid gression analysis on measured equilibrium ratios to de-
densities and molecular weights for the groups. His termine parameters for the common Bendict-Webb-
analysis with generalized liquid densities and molecular Rubin equation of state and then predictions were made
weights of Table 1 were used with th~ interaction coeffi- on the same systems.
cients in Fig. 3. Dew-point measurements at tempera-
tures of 101.1°C and 121.1°C (214°F and 250°F) are Crude-Oil/Natural-Gas Systems
compared with predictions along with liquid yields at Data available for crude-oil/natural-gas systems are listed
these temperatures in Table 5. The method in this study in Table 3. Before making phase calculations, extended
showed no improvement wheri predicting dew-point analyses had to be prepared for some systems by convert-
pressure, but a modest improvement when predicting ing distillations to individual hydrocarbon groups us-
liquid yield. Note that the Starling method. involved re- ing the generalized properties in Table 1. Density and
molecular weight of the last fraction were adjusted
PRESSURE • psia to provide an over-all match of reported density and
500 1000 2000 4000 molecular weight of the c7 + fraction.
Preliminary trials when predicting the behavior of
crude-oil/natural-gas systems indicated that the interac-
tion coefficient relationship with density did not give
satisfactory results when extended to the last fraction. We
knew that asphaltic and high-bOiling constituents occur-
ring in the ''last fraction,'' such as C20 +, were not likely
to have a continuous volatility behavior when measured
in terms of liquid density. In the work of Roland23 and
Standing and Katz,24 the downward curvature of the
vapor-liquid equilibrium ratios with increased pressure
was unexpected. At that time, the explanation was that
concentration of the high-boiling and asphaltic fractions
~ in the l~quid state by retrograde vaporization of the inter-
0
- mediate constituents was responsible for the behavior;
~
0::
equilibrium cell contents were approaching dew points at
~ 0.1 increased pressure. One might, therefor~, expect to have
:::> a discontinuity in the interaction-coefficient/density
a:CD function as the last constituent is reached. Accordingly,
:J the interaction coefficient for the last group was deter-
:::>
8 mined from even a single reliable point (a bubble point,
for example) because its use for predicting equilibria for
other conditions is shown to be reliable.

Hoffman et al. 8 Crude-Oil System

Data are composed of composition of the equilibrium
liquid and gas phases at 93. 9°C (201 °F) and 26 455 kPa
(3 ,837 psia) at reservoir conditions. The interaction coef-
ficient of the last group, c22 +' is. computed using the
bubble-pomt pressure of the liquid of 0.17. Using this
value, the gas-condensate composition calculated as part
0.001 [.___ _L___L_L_L..j_LL.-Ul------'-~
of the bubble-point computation is compared with the
1000 10,000 40,000 measured composition of the gas-cap fluid in Table 6.
PRESSURE. k Po
Roland 23 Natural-Gas/Crude-Oil System
Fig. 4-Comparison of predicted with experimental K values of
Hoffman eta/. s condensate at 93.9°C (201 °F). Based on the distillation curve, density and molecular

1652 JOURNAL OF PETROLEUM TECHNOLOGY

 . TABLE 5-COMPARISON OF PREDICTED WITH EXPERIMENTAL DATA FOR STARLING 1 CONDENSATE AT
TEMPERATURES OF 101.1 and 121.1oC (214° and 250"C)
Liquid Yield*
Pressure Temperature Experimental This Work Starling
(kPa) (psia) (oC) (m3 /m 3 x10 6 ) (bbi/MMscf) ( m3 /m 3 x1 O") (bbi/MMscf) (m3 /m 3 x1 06 ) (bbi/MMscf)
13893 2,015 101.1 880.2 156.5 823.50 146.4 1056.9 187.9
10445 1,515 101.1 877.5 156.0 780.7 138.8 990.0 176.0
6998 1,015 101.1 822.9 146.3 706.5 125.6 872.4 155.1
3551 515 101.1 648.8 115.3 576.0 102.4 872.4 119.7
13893 2,015 121.1 672.2 119.5 690.2 122.7 830.8 147.7
8722 1,265 121.1 711.0 126.4 633.9 112.7 755.4 134.3
Dew-Point Temperature Measured (kPa) This Study (kPa) Starling (kPa)
101.10C (214°F) 26166 (3,795 psia) 26641 24 235 to 25 614
121.10C (250°F) 24 235 (3,515 psia) 25786 23 546 to 24 235

Downloaded from http://onepetro.org/JPT/article-pdf/30/11/1649/2227059/spe-6721-pa.pdf/1 by Pontificia University Catolica Brazil user on 07 March 2025
*Liquid volume at reported pressure and temperature, and gas standard cooditions for metric units are 15°C and 101.325 kPa. Ave.rage absolute deviation for this study is 9.4%
and 12.1% for Starling.

weight of the C7 + residue, and the generalized properties (7 ,965 psia), which seems to be in some error, is ex-
in Table_1, an extended analysis up to the C 28 +group was cluded, the absolute deviation will improve to 1. 68%.
conducted for the crude oil of constant composite com- Furthermore, for 13 runs of 48.9°C (120°F) and 93.3°C
position ofRoland. 23 (200°F) and below 46 471 kPa (6,740 psia), absolute
The interaction coefficient for methane through c28 + deviation for methane equilibrium ratios is only 0.83%.
was computed by adjusting methane K values for the Methane K values and components through C 6 + (Table 7)
24 587 -kPa (3,566 psia) and 38 480-kPa (5,581 psia) indicate excellent agreement between predicted and
runs at 120°F. The equilibrium ratios of all other runs of measured values .
constant composition then were computed. Predicted
equilibrium ratios at 93.3°C (200°F) are compared with Standing and Katz 24 Natural-Gas/Crude-Oil
experimental data in Fig. 5. The average absolute devia- System
tion of 14 runs at 93.3 oc (200°F) for methane K values is Six runs were selected from Standing and Katz data.
2.30%. However, if the K value of the run at 54 91 7 kPa These runs belong to 49.9°API crude oil with a GOR of
650 m 3 /m3 (3,660 scf/bbl) at 48.9°C (120°F) and
121.1 oc (250°F). Standing26 gives extended analysis for
crude oil with cut intervals of 37. 8°C (100°F), and also
PRESSURE. psia
data for molecular weight and· density of high-boiling
200 500 1000 3000
constituents vs boiling point.
The interaction coefficient between methane and the
last group, with boiling points starting at 371.1°C
(700°F), was obtained by using the methane K values for
oo Experimental two runs at 121.1 ac (250°F) with pressures of 21 960
-Calculated
kPa (3,185 psia) and 29 715 kPa (4,310 psia). Fig. 6
compares predicted with experimental K values for three
I.Of-,------ runs at 48.goc (120°F). The proposed procedure is capa-
. ble of predicting downward curvature of ethane through
heptane plus K values at high pressures. Deviation of the
K values is given in Table 7.
~
c)
Katz and Hachmuth 25
~
0::
~
~
0::
IIl
::::i
0.1
--::1- 10
I
Crude-Oil/Natural-Gas System
Although the composite composition is not provided for
this set of early. data on the equilibrium ratios of the
5 i
aw
TABLE 6-COMPARISON OF PREDICTED WITH MEASURED
I
GAS-CAP PHASE COMPOSITION WITH HOFFMAN et at. •

-~---------0---
c7
0
0 l-
1
Measured
Composition
of Gas-Cap
'Bubble Point
Calculation on
Well Effluent Crude Oil
Component (mol%) (mol%)
ci 91.35 92.12
c2 4.03 3.504
c3 1.53 1.52
0.00 I L__L__l__J____J__LLL...l..L---'------L..---'--L..LJ...L.LI nC5 0.19 0.16
1000 10,000 100,000 Cw 0.158 0.116
PRESSURE. kPa c15 0.050 0.046
Fig. 5-Comparison of predicted with experimental K values for C:n 0.004 0.0036
Roland 23 crude oil at 93.9°C (200°F). C2z+ 0.002 0.0019

NOVEMBER, 1978 1653

 TABLE 7-AVERAGE ABSOLUTE PERCENTAGE DEVIATION FOR METHANE THROUGH
C1+K VALUES FOR CRUDE-OIL SYSTEMS
Component Roland 23 Standing and Katz 24 Katz and Hachmuth 25 Evans and Harris9
c1 2.4* 2.2 2.3 2.8
c2 3.0 4.8 2.3 9.6
c3 3.4 8.1 3.6 6.8
c4 6.0 4.7 1.7 10.3
Cs 4.3 15.0 7.3 13.1
Cs 8.7 13.6 10.1 13.9
c7+ 19.8 18.9 33.5 37.2
Equilibria
compared 20 6 7 10
*Average absolute deviations as percent of experimental values.

Evans and Harris 9

Downloaded from http://onepetro.org/JPT/article-pdf/30/11/1649/2227059/spe-6721-pa.pdf/1 by Pontificia University Catolica Brazil user on 07 March 2025
crude-oil/natural-gas system, the smoothness of the
measured K values indicates that the composite composi- Crude-Oil/Natural-Gas System
tion change is negligible. Based on this assumption, one Extended analysis for the crude oil was conducted for up
can compute the. composite composition in the cell. to c28 +;the interaction coefficient between this last frac-
Vapor- and liquid-phase .compositions of the run at tion and methane was obtained by adjusting methane K
12 024 kPa ( 1,7 44 psia) and 93.3 oc (200°F) were se- values for runs ofMixture B. Fig. 7 compares predicted
lected to compute the composite composition. Liquid and with experimental K values for constituents of Mixture
vapor phases of this run were combined as 2 mol gas and A, and Table 7 shows percentage deviations.
1 mol liquid. Extended analysis was conducted the same
way as for Roland crude oil. The interaction coefficient General Comparison
for methane and the last fraction (i.e., C28 +)was obtained Table 8 assembles the interaction coefficient between
by matching pre4icted vaporization to equal 66.67%. methane and the last group for crude oils, accompanied
With this value of interaction coefficient, K values were by properties of the last fraction. All extended analysis,·
predicted for other runs at 93.3°C (200°F). Table 7 indi- except for Hoffman et al. crude oil, did not have true
cates that predicted values agree well with experimental boiling-point distillation analyses. Extended analyses
measurements .. were improvised using judgment and generalized ex-

PRESSURE, psio
PRESSURE, psio I 000 2000 10000
1,000 10,000 10.0 ' '
10.0
I
I I ~ I

c,N o oExperimental
-Predicted

~
l~
1.0

r-~ u
I I
C2 ~
:X:: c3o I I 1.0
Q :X::

~ I 6
r-3~

~·:a
0::: c4 -~
:::;;!

t
0:::
::J ~

~~ bo1o
0:: 0.1 ::J
CD
0:::
::::i CD
::J ::::i
0
w ::J
0
w 0 Ao
o
I
0. I
,___ cs~ I

~~
- --
0

o o Experimental
-Predicted
A
c+7 0
0.001 ,___.__...__._._J...J..L...L...l..-__;____J..._._._...J.....J'--L-J...J..I..J
1,000 10,000 100,000 I I
0.0 I
PRESSURE, kPo 10,000 100;000
Fig.·s-Comparison of predicted with experimental K values for PRESSURE, kPa
650-m 3 /m 3 (3,660 scf/bbl) gaJ)-to-oil ratio of Standing and Katz 24 Fig. 7-Comparison of predicted with experimental K values for
crude oil at 48.9°C (t20°F). Mixture A crude oil of Evans and Harris9 at 87.8°C (190°F) ..

1654 JOURNAL OF PETROLEUM TECHNOLOGY

 TABLE &-INTERACTION COEFFICIENTS BETWEEN METHANE AND LAST
FRACTION OF CRUDE-OIL SYSTEMS
Last Fraction
Last Interaction Density, at 15.5°C Molecular
Crude Oil Fraction Coefficient (g/ml) Weight
Hoffman et 8 a/. C22+ 0.17 0.902 361
Roland 23 C2s+ 0.17 0.928 492
Standing and Katz24 · c23+ 0.25 1.008 622
Katz and Hachmuth25 C2s+ 0.20 0.939 508
Evans and Harris9 c2"8+ 0.25 0.914 342

tended hexane plus its properties. Therefore, any correla- 8. Hoffmann, A. E., Crump, J. S., and Hoco.tt, C. R.: ''Equilibrium
tion of methane with the last group's interaction coeffi- Constants for a Gas-Condensate System,'' Trans., AIME (1953)
198, 1-10.
cient should wait for more data with systems, including

Downloaded from http://onepetro.org/JPT/article-pdf/30/11/1649/2227059/spe-6721-pa.pdf/1 by Pontificia University Catolica Brazil user on 07 March 2025
9. Evans, R. B. and Harris, D.: "Equilibrium Vaporization Ratios-
true boiling-point analyses. Hydrocarbon Mixture Containing Two Concentrations of Hep~
tanes and Heavier Fractions," Ind. Eng. Chem. (1956) 1, No. 1,
Recommendation 45-50.
10. "Selected Values of Properties of l:Iydrocarbons and Related
When obtaining composition data for crude-oil reser- Compounds," API Project 44, Texas A&M U., College Station
voirs, bubble-point pressure customarily is obtained at (1969).
the reservoir temperature on a bottom-hole or recom- 11. Zudkevitch, D. and Jaffe, J.: "Correlation and Prediction of
bined sample. The composition with extended analyses Vapor-Uquid Equilibria with the Redlich-Kwong Equation of
State, "AIChE Jour. (1970) 16, No. 1, 112.
to C20 + or more, using a true boiling distillation along 12. Sage, B. H. and Lacey, W. N.: Monograph on API Research
with bubble point, Peng-Robinson-AGA procedure, and Project 37 - Thermodynamic Properties of the Light Paraffin
our proposed technique (including properties of the Hydrocarbons and Nitrogen, API, Dallas (1950).
groups, if necessary) can be used to assess the interaction 13. Shim, J. and Kohn, J.P.: "Multiphase and VDlumetric Equilibria
coefficient for methane and the last fraction. The interac- of Methane-n-Hexane Binary System at Temperatures Between
-110° and 150"C, "J. Chern. Eng. Data (Jan. 1962) 7, No.1, 3-8.
tion coefficient then will give the information required 14. Reamer, H. H., Sage, B. H., andl.;acey, W. N.: ''PhaseEquilibria
for predicting phase behavior at other conditions. in Hydrocarbon Systems-Volumetric and Phase Behavior of the
Methane-n-Heptane System," Ind. Eng. Chern. (1956) 1, No. 1,
Conclusions 29-42.
15. Kohn, J. P. and Bradish, W. F.: "Multiphase and Volumetric
Proper estimation of methane K values in both crude-oil Equilibria of the Methane-n-Octane System at Temperatures
and condensate systems is essential for an accurate esti- Between -110° and 150°C," J. Chern. Eng. Data (Jan. 1964) 9,
mate of bubble point, dew point, and flashing of these No. 1, 5-8.
systems. Using our proposed procedure, predicted meth- 16. Shipman, L. M. and Kohn, P. J.: "Heterogeneous Phase and
aneK values agree well with measured values. Volumetric Equilibrium in the Methane-n-Non:ane System,"
J; Chem. Eng. Data (April1966) 11, No.2, 176-180.
For gas-condensate systems, the Peng-Robinson-AGA 17. PUri,S. andKohn,J. P.: "Solid-Liquid-VaporEquilibriuminthe
procedure with extended analysis and interaction coeffi- Methane-n-Eicosane and Ethane-n-Eicosane Binary Systems,"
cientS fot methane and C 6 + constituents correlated with J. Chern. Eng. Data (1970) 15; No.3, 372-374.
density seems to give reliable results. For crude-oil/ 18. Sage, B. H., Webster, D. C., and Lacey, W: N.: "Phase Equi-
libria in Hydrocarbon Systems," Ind. Eng. Chern. (1936) 28,
mitural-gassystems, the interaction coefficient for me- 1045-1047.
thane and the last fraction should be determined with 19. Schoch, E. P., Hoffmann, A. E., kasperick, A. S., Lightfoot,
a bubble point, or selected from Table 8. Gathering J. H., and MayfielQ, F. D.: "Solubility of Methane in Benzene,"
data with true boiiing-point analyses of etude oils should Ind. Eng. Chem. (Jun:e 1940)32, 788-791.
20. Reamer, H. H., Sage, B. H., and Lacey, W. N.: ''Pha8eEquilibria
help correlate the interaction coefficients for the last
in Hydrocarbon Systems, VDlumetric and Phase Behavior of the
fraction containing the high boiling-point material and Methane-Cyclohexane System," Ind. Eng. Chem. (1958) 3, No.
asphalt. 2, 240-245.
21. Poston, R. S. and McKetta, J. J.: "Vapor-Liquid Equilibrium in
References then-Hexane-Nitrogen System,'' J. Chem. Eng. Data (July 1966)
1. Starling, K. E. : ' 'A New Approach for Determining Equation-of- 11, No.3, 364-365.
State Parameters Using Phase Equilibria Data,'' Soc. Pet. Eng. J. 22. Azarnoosh, A. and McKetta, J. J.: "Nitrogen-n-Decane System
(Dec. 1966) 363-371; Trans., AIME, 237. in the Two-Phase Region," J. Chem. Eng. Data (Oct. 1963) 8,
2. Soave, G.: "Equilibrium Constants From. a Modified Redlich- No.4, 494-496.
Kwong Equation of State," Chern. Eng. Sci. (1972) 27, 23. Roland, C. H.: "Vapor-Liquid Equilibria for Natural Gas-Crude
1197-1203. Oil Mixtures," Ind. Eng. Chern. (Oct. 1945)37,No. 10,930-936.
3. West, E. H. andErbar, J. H.:' 'An Evaluation of Four Methods of 24. Standing, M. B. and Katz, D. L.: "Vapor-Liquid Equilibria of
Predicting the Thermodynamic Properties of Light Hydrocarbon Natural Gas-Crude Oil Systems," Trans., AIME (1942) 155,
System," Proc., NGPA 52nd Annual Convention, Dallas (1973) 232-242.
50-61. 25. Katz, D. L. and Hachmuth, D. H.: "Vaporization Equilibrium
4. Bergman, D. F.: "PredictingthePhaseBehaviorofNaturalGas in Constants in a Crude Oil-Natural Gas System," Ind. Eng. Chern.
Pipelines," PhD dissertation, U. of Michigan, Ann Arbor (1976). (Sept. 1937)29, 1072-1077.
26. Standing, M. B.: "Vapor-Uquid Equilibria of Natural Gas-Crude
5. Peng, D. Y. and Robinson, D. B'.: "A New Two-Constant Equa-
Oil Systems," PhD dissertation, U. of Michigan, Ann Arbor
tion of State," Ind. Eng. Chern. Fund. (1976) 15, No. 1, 59-64. (1941). . JPT
6. Bergman, D. F., Tek, M. R., and Katz, D. L.: Retrograde Con-
densation in Natural Gas Pipelines, Monograph Series, American
Original manuscript received in Society of Petroleum Engineers office Sept. 15, 1977.
Gas Association, New York City (1975). Paper accepted for publication Jan. 26, 1978. Revised manuscript recejved July 19,
7. Cavett, R. H.: "Physical Data for Distillation Calculations 1978. Paper (SPE 6721) first presented at the SPE-AIME 52nd Annual Fall Technical
-Vapor-Liquid Equilibria,'' Proc., API 27th Mid-Year Meeting Conference and Exhibition, held in Denver, Oct. 9·12, 1977.
(1962) 52, 351-366. This paper will be included in the 1978 Transactions volume.

NOVEMBER, 197"8 1655