SPE

SPE 10067

Society of Petroleum EngIneers of AIME

Evaluating Constant Volume Depletion Data

by ~urtis Hays Whitson, Roga/and District College and Stein B0rre Torp Norwegian
InstItute of Technology
,

Copyr!ght 1981, Society of Petroleum Engineers of AIME

This paper.was presented at the 56th Annual Fall Technical Conference and Exhibition of the Society of Petroleum Engineers of AIME, held In
San Antonio, Texas, October 57, 1981. The material Is subject to correction by the author. Permission to copy Is restricted to an abstract of
not more than 300 words. Write: 6200 N. Central Expressway, Dallas, Texas 75206.

ABSTRACT
This paper presents results of analyzing constant
volume depletion data obtained from experimental
analyses of gas condensates and volatile oils.
Theoretical and practical developments are suppor.ted
by analyses of experimental data from two North Sea
condensate reservoirs.
The three maj or contributions of this work are:
(1) presentation of material balance equations used
to calculate fluid (particularly liquid) properties
from measured constant volume depletion data
(2) a simple method for calculating "black oh"
formation volume factors and solution gas-oil ratios
for volatile systems using material balance results
and a separator flash program, and (3) investigation
of the Peng-Robinson equation of state as a tool for
matching measured PVT data and studying vapor-liquid
equilibria phenomena during constant volume depletion.
The main example presented is a rich gas
condensate whose measured, calculated and simulated
phase behavior are fully documented in tables and
figures. Complete description of the heptanes-plus
fraction is also included so that other engineers can
check, modify and hopefully improve fluid characterization using the Peng-Robinson (or: any other)
equation of state.
INTRODUCTION
Constant volume depletion (CVD) experiments are
performed on gas condensate and volatile oil fluids
to simulate reservoir depletion performance and compositional variation. Resulting data can be used in
a variety of reservoir engineering calculations,
among the most useful being material balance calculations, generating "black oil" PVT properties and more
recently, the tuning of empirical equations of state.
All of these applications are addressed in the
present work.

References and illustrations at end of paper

Few engineers are aware of potentially-useful data

which can be derived from CVD data, some of the most
important being liquid composition (and therefrom
K-values), density, molecular mass (and specifically
C7+ molecular mass); vapon density (using two independent methods); and total ~y~tem molecular mass.
No assumptions or empirical relations are used to
calculate these data - only experimental CVD data
and simple material balance equations.
A procedure outlining these calculations was first
presented by Reudelhuber and Hinds: Their description, however, is somewhat difficult to follow and
not extensively known or used by the industry. We
therefore decided to present the material balances
in equation-form using current SPE nomenclature.
Using the material balance-derived properties,
a method is proposed for calculating "black oil"
PVT properties - Le., formation volume factors and
solution gas-oil ratios used in two-phase flow
equations and reservoir material balances. The
method is not new in principle, as it was first
suggested by Dodson, Goodwill and Mayer 2 in 1953 for
solution gas/crude oil systems. Their method,
however, requires expensive and time-consuming liquid
sample removals and experimental flash separations.
The proposed method follows the same procedure but
uses experimentally-determined vapor compositions
and material balance-derived liquid compositions
together with a routine multi-stage separator flash
program (using low pressure K-values independent of
system composition). PVT properties calculated using
this method are compared with those calculated using
the Peng-Robinson 3 equation of state.
Though more complicated, empirical equations of
state can also be used to evaluate CVD data. Several
inves tiga tors 5 ,16 have used the Peng-Robinson EOS to
simulate PVT studies of light gas condensates and
crude oils, needless to say avoiding systems operating near the critical point. Results have ranged
from excellent to poor, depending on which properties
were compared. Conrad and Gravier 16 proposed a
method to improve liquid density estimations by
adjusting properties of the heaviest plus fraction
(boiling point and Cl interaction coefficient).

SPE 10067

EVALUATING CONSTANT VOLUME DEPLETION DATA

Firoozabadi, Hekim and Katz 5 studied another lean

gas condensate and found that by only adjusting the
methane-plus fraction interaction coefficient, the
Peng-Robinson EOS overestimated liquid drop-out by
nearly 100%. [As discussed later in this work,
material balance calculations of CVD data for this
system indicate that measured liquid volumes are
approximately 100% low -. i.e. calculated liquid
densities were much too high.]
Over 30 constant volume depletion studies
performed by commercial and private laboratories
were analyzed using the material balance approach.
Three of these (North Sea fluids) were chosen to
be analyzed using the Peng-Robinson equation of state.
Their choice was based on (1) internal consistency
of measured CVD data, as indicated by material
balance calculations, and (2) availability of
extended compositional data for the heptanes-plus
fraction. All three fluids have similar paraffinnaphthene-aromatic content, withWatson 4 characterization factors ranging from 11.95 to 12.05 for the
C7+ fraction.
The first fluid (NS-l) is a rich gas condensate
and was chosen to illustrate proposed techniques for
analyzing constant volume depletion data. Extensive
data for this sample and its heptanes-plus fraction
have been included in tabular form so that other
engineers can duplicate, modify and hopefully improve
our analyses.
The second fluid (NS-2) is a lean gas condensate
similar to the systems analyzed in References 5 and 16.
Our discussion of NS-2 is limited to behavior or
observations which differ from those presented for
the rich gas condensate. The last fluid (NS-3) is a
volatile oil operating near its critical point. We
had not completed our analysis of NS-3 using the
Peng-Robinson EOS when this paper was written;
convergence problems were encountered when trying to
simulate the CVD process.. NS-3 is therefore .only
mentioned withregaud to material balance calculations
and K-value behavior. More information on any or all
of these fluids can be obtained from the authorsf
DESCRIPTION OF THE CONSTANT VOLUME DEPLETION PROCESS
A constant volume depletion experiment is
conducted at reservoir temperature and begins at
saturation pressure. Cell volume, Vcell, or the
volume contained by the fluid, is initially noted
and used asa reference volume.
Mercury is then withdrawn from the bottom of the
cell, thereby lowering the pressure as fluid expands ..
During this process, a second phase develops either retrograde liquid (for gas condensates) or
solution gas (for volatile oils). Mercury withdrawl
is ceased when a predetermined pressure is reached.
Some laboratories measure liquid volumes during the
first pressure reduction, before any vapor has been
removed; these volumes, reported relative to Vcell,
represent constant composition depletion. They
closely approximate, however, volumes which would
have been measured if the process had been constant
volume depletion. [This was checked using the PengRobinson EOS simulator for lean and rich condensates.]

* Rogaland

District College, Ullandhaug, 4000

Stavanger, Norway

Mercury is reinJected into the cell at constant

pressure while simultaneously withdrawing an equivalent volume of vapor. When initial cell volume is
reached, mercury injection is ceased. Withdrawn
vapor is analyzed using gas chromotography to determine composition, Yj' Moles of vapor produced are
calculated using the real gas law and are reported as
a cumulative percent of initial moles, n p ' Compressibility factor, Z, is also calculated by noting
produced vapor surface volume and equivilant cell
volume (at pressure and tem~erature). From measured
vapor gravity and composition, heptanes-plus
molecular mass is back-calculated. Liquid volume
is measured visually and reported as a percent or
fraction of cell volume, which in essence is a type
of hydrocarbon liquid saturation, SL.
The experimental procedure is repeated several
times (6-7) until a low pressure is reached, say
700 psig (4828 kPa). The remaining liquid is removed,
s epara ted (i. e. dis ti lled) and analyzed us ing gas
chromotography. Measured liquid composition should
check with material balance-derived composition.

[Some. majOlL taboJ1a:toJUell -6moo:th and "adjU6:t" me.a-6UJl.e.d

vapoJL c.ompoJ.lWo n-6 untU :the ma:te.JUat batane-e. ehe.ek.6.
T!UJ.> pJLoc.edUILe. 16 dMc.oUILage.d -in gene.JLaL 1:t 1.-6 good
pMc.:Uc.e. :to -inq u-LlLe whe.:the.JL a taboJ1a:toJL{1 JLe.pow
me.a-6 UlLed oJL "-6moo:thed" da..ta, and :to wha..t ex.:te.nt
ma:te.JUat batcmee-de.JUve.d da.:t.a Me. U-6e.d -in Mnat CVv
JLe.poJd6 ]
MATERIAL BALANCE EQUATIONS
Liquid Composition and K-value Calculations
Perhaps the most useful application of constant
volume depletion data is for calculating liquid ,
compositions which, together with measured vapor
compositions, yield high pressure K-values having
many important reservoir and process engineering
applications. To arrive at the final express:i:.on for
liquid composition in terms of measured CVD data, we
'first state molal and component material balances,
respec tively,

n.tk. = nLk. + nvk.

(,1)

where nL = moles of liquid with composition Xj,

nv = moles of vapor with composition Yj and nt =
total moles in the system with composition Zj, each
quantity being determined at pressure stage k.
Subscript j designates component numbers making up
each phase.
Eq. 1 s ta tes that total moles of the two phase
system equals the sum of liquid and vapor moles, while
Eq. 2 states that total moles of component j in the
two phase system equals moles of j in the liquid plus
moles of j in the vapor. The only data measured
directly and appearing in either of the equations is
vapor composition. The remaining unknowns can be
determined from reported CVD data and modified forms
of the material balance relations.

SPE 10067

C. H. WHITSON AND S. B. TORP

First we note that total moles at stage k equals

initial moles minus cumulative moles of vapor
produced. We assume a basis of one mole initial
fluid, that is nt1 = 1, yielding

All unknowns in Eq. 1 have now been defined in

terms of measured CVD data except liquid composition,
which when written in terms of the other variables,
becomes

I<.

r t..n ,

,[='2

,[='2

p.<.

(4)

j.<.

where t..npi = the incremental moles of vapor produced

from the cell during stage i, and zjl = the initial
fluid composition at stage 1 (saturated conditions).
Moles of vapor remaining in the cell can be
calculated using a volumetric balance and the real
gas law (pV=nZRT). Recalling the basis of one mole
initial fluid, cell volume can be calculated from
initial fluid properties, which for gas condensates is

Zd R.T

Pd

(Sa)

and for volatile oils (existing as a liquid at bubble

point pressure),
VeeLt = Mb/Pb

(5b)

where R = 8.3143 J/mol-K for preferred SI units, and

R '" 10.732 ~sia-ft3/mol-OR for field units. [T(K),
p(kPa), V(m ) and T(OR), p(psia),V(ft 3), respectively] 1"
At each depletion pressure, liquid volume is
measured visually and reported as a fraction of the
cell volume, SLk' Actual liquid volume VLk can then
be calculated from

vLI<.

SLI<. VeeLt

................. (9)

Equilibrium constants or K-values are defined as the

ratio of equilibrium vapor to liquid composition, or

I<.

r t..n "y

:::

I1 vk)

(I1tk

The same material balance can be applied on a

component basis, resulting in

Vc..~

ntl<.Zjk - nvk'Yjk

(3)

pA..

" . . . . . . . . . . . . . . . . . . . .. (6)

An effective means of correlating and checking the

consistency of calculated K-values (i.e. liquid
compositions) is to plot tog Kp vs the component
characterization factor Ff as suggested by Hoffman,
Crump and Hocott~ It has been our experience that
such a plot, when exhibiting linear trends approaching a common convergence point, indicates high
quality vapor composition data, and to a lesser
extent, goodness of liquid volume measurements.
A more complete discussion of this method is presented in a later section.
Physical Property Calculations
Constant volume depletion data can also be used
to calculate physical properties of equilibrium vapor
and liquid. A mass balance is employed to carry out
the necessary calculations,

mtk

mLk + mvk . . . . . . . . . . .. (11)

where mt = total mass of the system, mL = liquid mass

and my '" vapor mass at stage k. Another way of
stating the mass balance is that total mass at stage
k equals initial mass minus cumulative vapor mass
produced from the cell. Recalling the basis of one
mole initial fluid, and thereby equating initial mass
with initial (saturation) molecular mass, Ms ' gives

r t..n ,M ,

,[= '2

and from a volume balance, vapor volume Vvk is

. ,

. . . . . (12]

Vapor mass can now be calculated by noting that

mass equals molecular mass times moles, or
(8)

mvk
where pressure Pk corresponds to vapor compressibility
factor Zk'
i' Mb and Pb are buble point molecular mass and
density, respectively. Zd and Pd are dew point
compressibility and pressure, respectively.

V.<.

where Myi is the vapor phase molecular mass at

stage i.- Both Ms and My can be easily calculated
using Kay's pseudocritical mixing rule and appropriate
component molecular masses.

Using the real gas law, the corresponding moles of

vapor nvk are calculated from

P'<'

nvkMvk

* F = b(l/Tb

(13)

- liT) where b'" [tog(Pc/Pa)] I (l/Tb-l/Tc)

and: Tb = boiling point at atmospheric pressure Pa,
Tc and Pc are critical temperature and pressure,
respectively, and T is the system temperature.

SPE 10067

EVALUATING CONSTANT VOLUME DEPLETION DATA

Liquid mass is calculated as the difference between

total mass and vapor mass,

mLk

m.:tk - mvk

(14J.

Having calculated masses and volumes of

liquid and vapor, respective densities
calculated directly from the ratio of m
p ~ m/V (where volumes come from Eqs. 6

equilibrium
can be
to V. i.e.
and 7).

An independent check of vapor density can be

used to check the consistency of measured Z factors.
The relation is derived directly from the real gas
law and can be stated as

Mvk'Pk
Zk R T

'"

. .'

'"

<

1.51

We can al.so calculate molecular mass of the

equilibrium liquid, and specifically its heptanesplus fraction. Rewriting the mass balance as

m.:tk

MLk'VLCk + M0k 'VL v k

(16J.

we can solve for liquid molecular mass MLk'

Using Kay's mixing rule, the heptanes-plus molecular

mass can be back-calculated to yield

N-1

,[:/ Mj Xjl<.,[
.......... (1S)

"BLACK OIL" PVT PROPERTIES

Dodson, e;t.a~ suggested an experimental
procedure for determining so-called black oil PVT
properties used in two-phase flow equations and
solution-gas drive material balance relations.
Current laboratory procedures for estimating oil
formation volume factor Bo and solution gas-oil ratio
Rso only approximate the Dodson, e.:t.a. method without flashing the liquid phase at each stage of
the differential vaporization process. For medium
to low volatile crudes this procedure appears valid
for most engineering calculations. The "vapor"
solution gas-oil ratio Rsg is also assumed equal to
infinity - i.e. liquid condensation is neglected.
Highly volatile oils and gas condensate fluids
cannot, howeVer, be analyzed or described by the same
differential process, The basic problem posed by
these more volatile fluids is that during two phase
flow there exist both two phases aVLd two components.
That is, flowing oil contains solution gas which,
when undergoing pressure reducti0n, evolves and mixes
with the existing vapor phase. Likewise, flowing gas
contains retrograde liquid which also evolves and
mixes with the existing liquid when pressure declines.
This complex thermodynamic phenomenon is, for all
practical purposes, impossible to simulate in the
laboratory.
An alternative method is suggested which, by
making certain simplifying assumptions, approximates
the true model described above. Basically,
individual phase compositions determined from CVD
analysis (measured Or calculated) are flashed through
a multi-stage separator simulator representing field
conditions. Fig. 1 describes the process diagramatically.
Before we begin our discussion of the proposed
method, let us define the four basic PVT properties
used in two-phase flow and reservoir material
balance equations:

where Mj are molecular masses of "pure" components.

The average C7+ molecular mass of the tW0 phase
system should be calculated using the relation

VLLk'XkC7+ ,M LkC7 + + nvk'YkC7+ .MvkC7 +

R =
(19 )

.:to.:ta gal.> volume Jr.u,ulting oJr.om .:the 61Mh 06 x.

j

.60 .6.:toc.k .:tank oil volume Jr.U,ultiVLg Mom 61Mh 06 x.

j

VLLk'X kC7 + + nvk'YkC 7+

B

.:to.:ta gM volume Jr.U,ultiVLg oJr.om .:the 61Mh 06 y.

-

. j

.6.:toc.k .:tank oil volume Jr.U,ultiVLg 6Jr.om 6iMh 06 y.

j

where x' are liquid compositions determined from

material balance equations. Yj are vapor compositions
measured experimentally,

SPE 10067

C. H. WHITSON AND S. B. TORP

First, liquid composition x' is flashed using a

set of appropriate K-va1ues and ~asic vapor-liquid
equilibria equations. [Glas and Whitson 9 have
documented that Standing' s10 low pressure "black oil"
K-values are quite accurate for flash calculations of
black oils. We have since found that they are also
accurate for flash calculations of medium to highly
volatile gas condensates - i.e. systems with gas-oil
ratios less than about 50 000 SCF/STB (9000 Sm 3/Sm 3).]
The sum of
surface gas volumes divided by stock
tank oil volume is defined as the" liquid" gas-oil
ratio Rso .

APPLICATION OF THE PENG-ROBINSON EQUATION OF STATE

Oil formation volume factor Bo is calculated

from the relation

The numerical solution technique used includes

a pre-iterative sucessive substitutions method
followed by Newton's method using analytical
derivatives. Convergence problems were encountered
for the NS-l fluid at temperatures approaching the
critical point - i.e. below reservoir temperature.
Several alternative numerical methods were tried
(Powell's method and a newly-developed accelerated
sucessive substitution method~ ) without success
Similar problems were noted with the volatile oil
system (NS-3) which, from all indications, lies very
near the critical point at reservoir temperature.
The lean gas condensate (NS-2) was solved problemfree OVer a wide range of temperatures.

J.>tage1>
L m + m
STO
1=1 g-<-

... . . . . . . . (20)

VSTO'P L
where VSTO is stock tank oil volume (e.g. 1 STB) and
where mg and mSTO are masses of total surface gases
and stock tank oil, respectively. Liquid density, PL'
can be determined from either material balance
calculations (ffiL from Eq. 14 and VL fromEq. 6), or
from one of several compositional density correlations
avai1ab1e 67 using material balance-derived liquid
compositions. We caution the use of PL calculated
from material balance equations since only a slight
error in retrograde liquid volume can result in a
substantial error in liquid density - and therefore
Bo ' The same error will ,not affect liquid composition
to the same degree.
At the same depletion stage k, vapor phase with
composition Yj is separated through the flash
simulator using identical K-values. The resulting
surface gas volumes divided by stock tank oil volume
defines the "vapor" solution gas-oil ratio Rsg
Gas formation volume factor, on the other hand, can
be accurately estimated from the CVD compressibility
factor Z using the real gas law,

where nSTO c moles of stock tank oil resulting from

the flash of nfeed moles of reservoir vapor.
The major assumption implicit in the proposed
method, assuming measured CVD data are" good'; is that
liquid and vapor compositions are solely dependent
on pressure. That is, the composition-pressure
relation is unique and not altered by physical flow.
Examples of PVT properties for the rich gas
condensate (NS-l) are presented in Figs. 2-5.
Here we have compared properties calculated using
material balance results with those calculated using
Peng-Robinson EOS simulated data. Identical low
pressure K-values were used for both sets of data.
"Vapor" solution gas-oil ratio is nearly the
same for both methods of calculation. "Liquid"
solution gas-oil ratio and oil formation volume
factor are both low for material balance-derived
properties. The difference is clearly a result of
the lower liquid densities estimated by the PengRobinson equation.

Measured CVD data and material balance-derived

properties were controlled using a fluid properties
package based on the Peng-Robinson equation of state
and developed by Roga1and Research Institute. A
complete description of the computer programs can be
obtained from the authors. The PVT package not only
includes general vapor-liquid equilibrium options,
but it also includes two options for characterizing
the heptanes-plus fraction - methods presented by
Whitson ll or Robinson and Peng.~

Pure component properties (critical pressure,

critical temperature, acentric factor and molecular
mass) were used for non-hydrocarbons and hydrocarbons
from methane to n-pentane. Only n-hexane was
considered for the C6 fraction.
Heptanes and
heavier properties were estimated using the procedure
and equations suggested by Whitson~1 with several
modifications given in Appendix B.
To manipulate the retrograde liquid volume curve,
the Watson charact13rization factor of the heaviest
component was adjusted, making sure that adjusted
properties were physically realistic.
Binary interaction coefficients were set equal
to zero except 13 : Nz- Nz c -0.02, COz - hydrocarbons
c
0.15, Nz - hydrocarbons c 0.12, and Cl - Cn'
n = 6,7, . , which were estimated using a linear
fit of the Katz and Firoozabadi 13 data (their Table 2),
6Cl-C~

= O. 14Y~ - 0.0668 ............ (22)

'The interaction coefficient between methane and the

heaviest component was then adjusted until a match
of the measured dew point pressure was obtained.

NS-1 is a rich gas condensate first tested at a

gas-oil ratio of 5500 SCF/STB (980 Sm 3/Sm 3) from an
initial reservoir pressure of 7300 psia (50340 kPa)
and temperature of 280 OF (138 oC). Stock tank oil
gravity was 44 API (0.8055 gm/cc). Separator samples
were taken while flowing the well at 16.3 MMSCF/D
(460.10 3 Sm 3) and a flowing bottomho1e pressure of
7260 psia (50070 kPa).

EVALUATING CONSTANT VOLUME DEPLETION DATA

Heptanes-Plus Characterization
Extended compositional data of the C7+ fraction
was not available for the NS-l fluid, only molecular
mass and specific gravity. Complete true boiling
point (TBP) data were, however, available from an
offsetting well, NS-lb. These data were adapted to
the NS-l fluid using the method presented in Ref. 11,
slightly modified as discussed in Appendix B.
Molal distribution (mole fraction vs molecular
mass) of the NS-lb fluid was fit using the gamma
distribution parameter alpha and variable upper
boundry molecular masses. The optimal alpha was 0.712
for eta (minimum molecular mass in the C7+ fraction)
of 92. Table 2 gives results of the match.
Molal distribution of the NS-l C7+ fraction was
then calculated using a ~ 0.712, n ~ 92 and MC7+ ~
184 (as compared to 177 for the NS-lb fluid). We also
chose to hold upper boundry molecular masses constant
(equivalent to paraffin values), giving the results
presented in Table 3.
Properties of the single carbon number (SCN)
groups were estimated** by defining the Kuop factors
from NS-lb SCN molecular masses and specific gravities
using the relation ll
K
~ 4 5579.MO.15178
-0.84573
uop'
.y

........ (23)

Eq. 23 was then inverted and combined with NS-l SCN

molecular masses to yield SCN specific gravities and
normal boiling points, Tb , for the NS-l fluid,
Y

6.01077.K~p1824 .MO.17947

(24)

where Tb ~ (Y'K uop )3, per definition. SCN data for

NS-l calculated using Eq. 24 are given in Table 3,
together with critical properties estimated using the
Riazi-DaubertlScorrelations (except for Tb > 850 OF,
when modified correlations were used 11 )
Tuning the Peng-Robinson Equation of State
Single carbon number groups were combined into
five multiple carbon number (MCN) groups - C7-C9,
CIO-CI3, CI4-C17, CIa-Cz4 and C2S+- as suggested in
Ref. 11. Group properties were calculated using
Kay's pseudocritical mixing rule, except for specific
gravities which used a volume-weighted mixing rule.
Methane interaction coefficients were estimated using
* The TBP analysis was performed according to the
procedure outlined in Ref. 14 and discussed by Katz,
et.al. 13 The laboratory only reported, however,
single carbon number molecular masses, mole fractions
and specific gravities. Cumulative volume percents
were then calculated by noting that incremental
volume (per mole) ~ mole fraction x molecular mass +
specific gravity. Little curvature was exhibited by
the TBP curve and normal boiling points were, therefore, merely averages of the boiling point range for
a given SCN group.

SPE 10067

MCN specific gravities and Eq. 22, as was the CI-C6

coefficient. Using these data in the Peng-Robinson
EOS yielded a dew point pressure much lower than
measured. The CI-CZS+ interaction coefficient was
then increased until dew point pressure matched.
The CVD simulator was run using the MCN
properties, as given in Table 3. The overall match
was good to excellent,except for liquid volumes
which were much too high (32% simulated maximum vs
22% measured maximum). To check if measured volumes
were low we compared material balance liquid densitieE
with Alani-Kennedy 7 densities (using material balance
compositions and molecular masses). Table 4 shows
results of the comparison, indicating that measured
volumes are consistent except for perhaps small
errors in the first two volume measurements.
Based on these results, it was decided to lower
the Peng-Robinson liquid volumes by adjusting the
characterization factor of the Czs+ fraction.
By lowering the factor from 12.42 to 11.80 resulted
in a decrease of the liquid volumes - 8% for the
maximum drop-out (from 32% to 26%). The adjustment
had little or no effect on other estimated data. To
have lowered the Kuop factor more would have created
a physically unrealistic system. Adjusted physical
properties for the C2S+ fraction are found in
Table 3, as is the methane interaction coefficient
used to adjust dew point pressure. Complete results
of the CVD simulation are presented in Table 5.
Peng-Robinson liquid densities are compared with
Alani-Kennedy estimates in Table 4.
Over twenty other adjustments of the C7+ characterization procedure were attempted for improving
liquid volume predictions. None of these were particularly helpful, though some are worth mentioning:
(1) extending the C7+ split to C40+ such that the
last component was very heavy, (2) increasing the
number of MCN groups used to nine, C2S+ inclusive,
(3) splitting the C7+ fraction into eight SCN groups
and a ClS+ fraction, (4) using TBP Kuop factors**
instead of those estimated from Eq. 23, and
(5) using the Lee-Kessler~ property correlations.***

** Actually Kuo is defined as TS/ 3 /Y and could,

therefore, have Eeen calculated dlrectly using normal
boiling points determined from TBP analysis. Using
Kuop estimated from Eq. 23 and measured molecular
masses and specific gravities, estimated normal
boiling points were calculated from Tb ~ (Y'K uop )3
and are presented in Table 2. Some of these values
were higher than upper boiling point boundries definec
for the specific SCN group. Two possible explanation!
are provided: (1) due to distillation under vacuum it
was not possible to duplicate exact boiling point
boundries as defined in Ref. 14, or (2) measured
molecular masses of the heavier fractions were in
error. It was found, however, that using estimated
Kuop factors from Eq. 23 - when used for critical
property estimations in the Peng-Robinson EOS - gave a
better match of measured constant volume depletion
data; the difference was only minor.
*** Information on these or other simulation runs can
be obtained from the authors. We would also appreciate suggestions as to how one might improve the
liquid volume prediction.

SPE 10067

C. H. WHITSON AND S. B. TORP

RESULTS AND DISCUSSION

Most results presented in this paper are taken
from the three North Sea systems NS-1, NS-2 and NS-3.
It has been our experience, however', that some
features of CVD analysis are common to all systems.
We have tried to differentiate between observations
which are specific to a given system, and those which
are more general in nature. We note in particular
that the K-va1ue correlation developed in Appendix A
was developed from our analysis of many fluids,
ranging from volatile oils to light gas condensates.
Fig. 6 compares measured (or more correctly,
"smoothed") vapor compositions with those simulated
using the Peng-Robinson EOS. The match is excellent,
showing only slight deviation for the C7+ and C6
components. Deviation of the hexane component is
probably due to its incorrect characterization as
n-hexane.
Fig. 7 presents heptanes-p1us molecular masses of
liquid and vapor phases and the total system.
Simulated and material balance-derived values match
well. Our experience has shown that a good match of
C7+ molecular mass using the Peng-Robinson EOS is
usually difficult, and very dependent on proper
characterization of the plus fraction.
Calculated equilibrium constants were correlated
using the Hoffman, et.al. 8 method. Three main reasons
are given for this choice: (1) the log Kp vs F plot
provides a simple means of defining the approximate
pressure- and temperature-dependence of K-va1ues,
(2) material balance-derived K-va1ues can be evaluated
for consistency by checking that log Kp vs F plots are
linear and converge, more or less, to a single point
and (3) an approximate estimate of convergence
pressure can be determined by extrapolating the slope
(of lag Kp vs F plots) vs pressure curve to zero,
which can in turn be used to improve initial K-va1ue
estimates for the Peng-Robinson (or any other)
equation of state. See Appendix A.
Fig. 8 presents NS-1 K-va1ues calculated using
material balance relations. The log Kp vs F plots
are linear and appear to approach a common point.
As discussed in Appendix A, the convergence point can
give an estimate of the apparent convergence pressure.
Actually, the most accurate value is obtained by
extrapolating the slope vs pressure curve to zero,
as done in Fig. 11. The resulting estimate of PK ~
8000 psia (55170 kPa).
Fig. 9 presents NS-1 K-values calculated from
the Peng-Robinson EOS. Once again linear plots of
log Kp vs F converge to a point. From the extrapolation of slope to zero in Fig. 11, PK ~ 7500 psia
(51720 kPa). Experience with the Peng-Robinson EOS
and material balance evaluation of CVD data has shown
that rich
gas condensates and volatile oils exhibit
a more well-defined convergence point than leaner
systems.
Another interesting feature shown in Fig. 9 is
that heavy components are better correlated using the
log Kp vs F method at higher pressures. This maif
suggest that methane interaction coefficients of the
plus fractions have most influence on K-values at
low pressures.

Temperature effects on the log Kp vs F plots of

NS-l were investigated by running the Peng-Robinson
EOS simulator at 340 of (l71.l 0C), some 60 OF higher
than reservoir temperature. Fig. 10 presents the
results, indicating that temperature influence is
(1) largest for heavy components at large pressures,
(2) neglibible at low pressures (as was found in
Ref. 10), (3) relatively small compared to the
influence of pressure, and (4) not significant in
changing the apparent convergence pressure of the
system. 1hese observations are also illustrated in
Fig. 11.
For lighter systems such as NS-2, there does not
always appear such a unique convergence point for
log Kp vs F plots. We thought that this perhaps
resulted from a change in the total composition of
the system, or from alteration in the heptanes-plus
properties. We investigated these possibilities by
running a constant composition simulation of NS-l
(at 280 and 340 OF) and NS-2 (at 241 OF). Resulting
K-values were compared with CVD K-va1ues and are
presented as log K vs log P plots in Figs. 12, 13 and
14. All three systems clearly indicate that compsitional change during constant volume depletion is not
significant enough to influence K-va1ues orconvergence pressure, if in fact there exists a true
convergence of K-va1ues to unity. As seen in Fig. 14,
the lean gas condensate (NS-2) does not appear to
have a convergence pressure for components heavier
than hexane.
CONCLUSIONS
1. Measured constant volume depletion data for two
gas condensates and a volatile oil were analyzed usin
simple material balances and the Peng-Robinson
equation of state (EOS).
2. A simple method is proposed for calculating
"black oil" PV1 properties (formation volume factors
and solution gas-oil ratios) of gas condensates and
volatile oils.
3. Material balance-derived K-values can be correlated
to yield an estimate of the apparent convergence
pressure which, when used in a newly-developed K-va1u
correlation, helps calculate high pressure K-va1ues
used as initial estimates in equations of state.
4. Simulated constant composition and constant volume
depletion studies of lean and rich gas condensates
using the Peng-Robinson EOS indicate that K-va1ues
are indep~nd~t of the depletion process.
5. Temperature effects on the Hoffman, et.al. 8 K-va1ue
correlating technique (lag Kp vs F) were studied
using the Peng-Robinson EOS.
6. 1he Peng-Robinson equation of state usually overestimated liquid drop-out for gas condensates during
constant volume depletion. The problem was normally
corrected or improved by reducing the Watson 4
characterization factor of the heaviest component.

EVALUATING CONSTANT VOLUME DEPLETION DATA

NOMENCLATURE
A or A(p)

.tog Kp vs F plot

slope of the straight line connecting the

critical point and atmospheric boiling
point on a .tog vapor pressure vs1/T
plot, cycle-OR ; cycle-K

convergence

slope of

liquid phase

carbon number

= oil

produced

reduced

formation volume factor, Bb1/STB ; m /Sm

C7+

heptanes-plus component

constant volume depletion

sc

= standard

exp(x)

eX ; e = 2.71828 ..

STO

= stock

EOS

equation of state

component characterization factor, cycle.

CVD

F or F(T)
K

equilibrium constant (K-va1ue)

Watson characterization factor

.tv!
.tog

natural logarithm to base e

logarithm to base 10

f3

m
M

mass, lb m ; kg
mo1euc1ar mass, lb m/1b-mole ; kg/kg-mo1e

molecular mass of the total system

(1) moles, lb m-mo1e ; kg~mo1e

(2) exponent in K-va1ue correlation

uop

NS

North Sea sample

pressure, psia

p(x)

probability density function

Pr

kPa

=. cumulative probability function

universal gas constant, 10.732 psia-ft 3 /

mole-OR; 8.3143J/mo1e-K

"vapor" solution gas-oil ratio, SCF/STB.

Sm 3 /Sm 3

"liquid" solution gas-oil ratio (same)

saturation, fraction or percent

sg

so

absolute temperature, oR ; K

volume, ft 3 ;m 3

liquid composition, fraction or percent

vapor composition, fraction or percent.

Y or Y(p)

intercept of

.tog Kp vs F plot

total system composition

vapor compressibility factor

atmospheric

b
c

bubble point (Pb) or boiling (T )

b
critical

cell

cell, pertaining to PVT cell volume

d
g

dew point
gas

index for summation

component.identifier

depletion stage

= saturated (bubble or dew point)

condition

tank oil

total (two phase)

= vapor phase

= parameter in gamma distribution

= parameter

in gamma distribution
= gamma function

specific gravity relative to air or

water (60/60)

= incremental
interaction coefficient
parameter in gamma distribution
(minimum molecular mass)
density, 1bm/ft 3 ; kg/m 3 (gm/cc)
= acentric factor

ACKNOWLEDGMENTS
The authors wish to thank H. Norvik, H. Asheim,
D. Murphy, V. Dalen and G. Nielsen for useful coriunents
concerning this paper. We also acknowledge computer
time and facilities donated by Roga1and District
College, Norwegian Institute of Technology (NTH) and
Continental Shelf Institute (IKU). Phillips Petroleum Norway and Statoi1 should be thanked for contributing well-needed fluid data to the petroleum literature. Economical support from frying pan publications, Inc. is, as usual, appreciated.

REFERENCES
1.

Reude1huber, F.O. and Hinds, R.F.: "A Compositional Material Balance Method for Prediction of
Recovery from Volatile Oil Depletion Drive Reservoirs," T!l.an.o. ,AIME(1957) 21 0,19-26

2.

Dodson, C.R., Goodwill, D. and Mayer, E.H.:

"Application of Laboratory PVT Data to Reservoir
Engineering Problems," TMn.o. ,AIME(1953) 79 g,
287-298

3.

Peng, D.-Y. and Robinson, D.B.: "A New TwoConstant Equation of State," IV!d. EYIfJ.Chem. FuV!d.
(1976)15,No.l,59-64

Subscripts
a

SPE 10067

SPE 10067
4.

5.

6.

C. H. WHITSON AND S. B. TORP

Watson, K.M., Nelson, E.F. and Murphy, G.B.:

"Characterization of Petroleum Fractions,t' Ind.
Eng.Chem.(1935)27,1460-l464
Firoozabadi, A., Hekim, Y. and Katz, D.L.:
"Reservoir Depletion Calculations for Gas Condensates Using Extended Analyses in the PengRobinson Equation of State," Can.J.Chem.Eng.
(Oct.,1978)56,610-6l5
Standing, M.B. and Katz, D.L.: "Vapor-Liquid
Equilibria of Natural Gas-Crude Oil Systems,"

20.

On'

Standing, M.B.: Votumwue and Phcu,e Behav-tOlt

Oil F-tud HyMoeMbon Sy.6teJnJ.>, 8th Printing,
Society of Petroleum Engineers of AlME, Dallas
(1977)

21.

Robinson, D.B. and Peng, D.-Y.: "The Characterization of the Heptanes and Heavier Fractions,"
Research Report 28, GPA Tulsa, Oklahoma (1978)

22.

Kessler, M.G. and Lee, B.1.: "Improve Prediction

of Entha1apy of Fractions," HydM. PMe. (March,
1978)153-158

Tnano.,AIME(1944)155,232
7.

A1ani, G.H. and Kennedy, H.T.: "Volumes of Liquid

Hydrocarbons at High Temperatures and Pressures,"

APPENDIX A

Tnano."AIME(1960)219,288-292
IMPROVED K-VALUE ESTIMATION AT HIGH PRESSURES
8.

Hoffman, A.E., Crump, J.S. and Hocott, C.R.:

"Equilibrium Constants for a Gas-Condensate
System," Tltano., AIME( 1953) 198,1-10

9.

Glas, C/J. and Whitson, C. H.: "The Accuracy of PVTParameters Calculated from Computer Flash Separation at Pressures Less Than 1000 ps ia," SPE Paper
8033 (1979)

10.

Standing, M.B.: "A Set of Equations for Computing

Equilibrium Ratios of a Crude Oil/Natural Gas
System at Pressures Below 1,000 psia," J.Pd.
Teeh. (Sept. ,1979)1193-1195

11.

Whitson, C.H.: "Characterizing Hydrocarbon Plus

Fractions," EUR Paper 183 Presented at the EUROPEC
Meeting held in London, England, Oct. 21-24, 1980

12.

13.

14.

Risnes, R., Dalen, V. and Jensen, J.1.: "Phase

Equilibrium Calculations in the Near-Critical
Region," Paper Presented at the 1981 European
Symposium on Enhanced Oil Recovery, Sept. 21-23,
1981
Katz, D.L. and Firoozabadi, A.: "Predicting Phase
Behavior of Condensate/Crude-Oil Systems Using
Methane Interaction Coefficients," TMno.,AlME
(1978)265,1649-1655
"Selected Values of Properties of Hydrocarbons and
Related Compounds," API Project 44, Texas A&M
Univ., College Station (1969)

15.

Riazi, M.R. and Daubert, T.E.: "Simplify Property

Predictions," Hydno. PMe. (March, 1980)115-116

16.

Conrad, P.G. and Gravier, J.F.: "Peng-Robinson

Equation of State Checks Validity of PVT Experiments," OilWcu, J.(April 21,1980)77-86

17.

Yarborough, L.: "Application of a Generalized

Equation of State to Petroleum Reservoir Fluids,"
from Equationo 06 State by Chao and Robinson

18.

Wilson, G.M.: "A Modified Redlich-Kwong Equation

of State, Application to General Physical Data
Calcua1ations," paper presented at the 65th
National AIChE Meeting, Cleveland, 1969

Solution of the Peng-Robinson (or any other)

cubic equation of state requires initial estimates
of K-values. At higher pressures (> 500 psia) and
particularly near phase boundries or the critical
point, these estimates are very important for determining the "correct" solution to the equation.
Accurate K-va1ue estimates can also reduce numerical
divergence when searching for the solution.
Wi1son w proposed the following thermodynamic
relation for estimating K-values which should, in
practice, only be used at low pressures,

K.
j

where

TIT,
ej

plPej
1 + W,
j

Tb/T ej

7 1 - Tb/T ej

5.37 = '?"tn(70)
3
'T and p define the system's temperature and pressure,
Tb is the atmospheric boiling point at Pa' and Tc and
Pc are critical temperature and pressure, respectively.
Actually, Eq. A-l appeared in the petroleum literature
some 10 years before Wilson proposed his relation.
Hoffman, d.at. 8 presented Eq. A-1 graphically (their
Fig. 4) and suggested the following generalization
(albeit graphically),
A(p)' Fj(T) + YIp)

19.

Brinkman, F.H. and Sickling, J.N.: "Equilibrium

Ratios for Reservoir Studies," Tnano.,AlME(1960)
219,313-319

(A-2l

where

11T b , - liT
J
tog (Pe/Pal
11Tb, - 1IT ,
j

ej

....... (A-3l

10

EVALUATING CONSTANT VOLUME DEPLETION DATA

A(p)

pJr.eJ.,!.>U!te-depertdent !.>,tope

Y(p)

~e!.>!.>U!te-depertdent

SPE 10067

lnteJr.eept
and that exponent n = 0.6 for all systems.

It is easily shown that Eqs. A-l & A-2 are identical

for A(P=Pa) = 1 and Y(P=Pa) = ,tog Pa - i.e. at
atmospheric pressure.
Engineers familiar with Eq. A-l are aware that
its accuracy is usually limited to low pressures.
An investigation of the pressure-dependent slope in
Eq. A-2 showed, however, that Eq. A-l could be
reformulated to yield equally-accurate results at
higher pressures.

We can now reformulate Eq. 1 (the Wilson-type

equation) to give K-values at all pressures and
temperatures,

K,
j

.. ..... (A-7)

The quasi-thermodynamical model used to extend

Eq. A-l is based on the suggestion by Brinkman and
Sicklingl~ that plots of ,tog Kp vs F(T) at several
pressures intersect at a common point defining the
apparent convergence pressure of the system. Their
relation, after dropping pressure- and temperaturedependent notation, is

K,

I .

Note that A=l for pcp , reducing E.q. A-7 to Eq. A-I.
For P=PK' K.=l, thus ~aking Eq. A-7 consistent at
both pressute boundries. Using Eqs. A-5 (n~0.6),
A-6 and A-7, K-values can be estimated directly from
input data used by equations of state .

(A-3)

APPENDIX B
where PK = convergence pressure and FK = the F-va1ue
corresponding to PK at the common point of intersection. Eq. A-3 is easily derived by noting that Kj=l
at convergence pressure PK' We can also show from
Eq. A-l that FK = ,tog(PK/Pa), or in terms of the
intercept in Eq. A-2,

Methods presented in Ref. 11 were used to

determine the molal distribution of the C7+ fraction
with the following modifications:
(1) Minimum molecular mass (eta) was set equal to 92

y
The pressure-dependence of slope A was investigated and found to have the following two properties:
(1) it ranged from 0 ~ A ~ 1 forPK < p < Pa and
(2) its form varied from sligt1y concave up to
linear (see Fig. 11). From these observations, a
general equation-form was found to fit A(p),

MODIFIED C7+ SPLITTING METHOD

A(p)

where exponent n varied from 0.5 to 0.8 for various

condensate and volatile oil systems. We found,
however, that n = 0.6 gave reasonable K-va1ue
estimates for all fluid systems at most pressures and
temperatures. In practice, exponent nand PK for a
particular system can be determined exactly by
plotting ,tog(l-A) vs ,tog(P-Pa) where n is the slope
and the intercept equals n.,tog(PK-Pa)' Slope A is
usually determined by drawing a best straight line or
using linear regression through data for components
ethane through hexane, where our experience has shown
that methane and carbon dioxide also lie on the line.
For more general use of Eq. A-5 (for example,
initial K-value estimates for an equation of state),
we have found that a good estimate of apparent
convergence pressure can be obtained from the
heptanes-plus molecular mass~

(2) The. error function used to determine the optimal

alpha (~) was

and was minimized using the half-interval method.

(3) The upper molecular mass boundry was allowed to
vary between 14i - 12 and l4i + 12, except for C7
which had a range of 93 to l4i + 12. Measured mole
fraction was matched by varying the upper botindry.
Only two iterations are required since the relation
between boundry molecular masses and mole fractions
~s nearly linear.
That is ,. if upper b.oundry Ml yields
Zl, and M2 yields 22 , and Z = measured mole fraction,
then the optimal tipper molecular mass boundry is very
nearly

M' .
opt

M2 ~.Ml
A

2 -

. A

......... (B-2)

In some cases, Eq. 2 was used twice in order to

increase the accuracy of the estimated mole fraction.

SPE 10067

C. H. WHITSON AND S. B. TORI'

(4) In the original method presented in Ref. 11,

carbon number molecular masses corresponding
to z with molecular mass boundries Mi and Ml+ l was
mereh the arithmetic average. Dale Embry (Phillips
Petroleum Company, Bartlesville) noted in a
personal comnunication that this approximation was
not necessary and that the exact analytical expression is given by
sin~le

M.{,',

n + a~

P~(M~;+I,a+l,~,n) - P~(M<M~,a+l,~,n)
--~"--::,:",,:,_-------~.{,:::,-,,_--

P~(M~Ml+l,a,~,n)

where

P~(M~l,a,~,n)

... (B-3)

'

00

P~(X~x,a"b,c.)

e,-Y

t'

(B . . 4)

[ya+ j / r(a,+j+l)] (B-5l

j=O
y

= (x-c.)

/b

(B-6)

and, as previously mentioned, n = 92. Concerning

Eq. B-4, ~ is the same whether a or a+l is used in
the cumulative.probability function Pro

11

TABLE 1 - MEASURED CONSTAID VOLUME DEPLETION DATA FOR THE NS-1 FLUID AT 280 of
Compositions

Equilibrium
Liquid

F:quilibriuni
Vapo,r.
Pressure - psia

Exp.

Calc.

714.7

714.7

Cor~ponent

6764.7

5514.7

4314.7

3114.7

2114.7'

1214.7

Carbon Dioxide
Nitrogen

2.37
0.31
73.19
7.80
3.55
0.71
1.45
0.64
0.68
1. 09
8.21

2.40
0.32
75.56
7.83
3.47
0.67
1. 37
0.59
0.62
0.97
6.20

2.45
0.33
77 .89
7.87
3.40
0.65
1.31
0.55
0.58
0.88
4.09

2.50
0.34
79.33
7.92
3.41
0.64
1.30
0.53
0.56
0.83
2.64

2.53
0.34
79.62
':8.04
3.53
0.66
1.33
0.54
0.57
0.82
2.02

2.57
0.34
78.90
8.40
3.74
0.72
1.44
0.59
0 . 61
0.85
1.84

2.60
0.33
77 .80
8.70
3.91
0.78
1.56
0.64
0.66
0.90
2.12

0.59
0.02
12.42
3.36
2.92
0.91
2.09
1.40
1.60
3.68
71.01

0.535
0.017
10.704
3.220
2.896
0.916
2.103
1.417
1.624
3.755
72 .815

Totals

100.00

100.00

100.00

100.00

100.00

100.00

100.00

100.00

100.000,

MC7 +

184.0

160.0

142.0

127.0

119.0

115.0

114.0

213.0

207.9

Methane

Ethane
Propane
iso-Butane
normal-Butane
iso-Pentane
normal-Pentane
Hexanes

Heptanes-plus

YC7+
Z
n

0.816

0.799

0.783

0.770

0.762

0.758

0.757

1. 238

1.089

0.972

0.913

0.914

0.937

0.960

9.024

21. 744

38.674

55.686

72.146

81. 301

19.7

21.6

21.3

20.2

19.3

0.000

-. %

SL - %

714.7

14.1

0.0

0.833

TABLE 2 - COMPOSITIONAL AND PROPERTIES DATA OF FLUID Ns-IB SAMPLED FROM

A WELL OFFSETTING NS-1 COMPARED WITH CALCULATED DATA
GENERATED USING THE METl10DPRESENTED IN REF. 11
Calculatedi'

Measured

Single
Carbon

Number
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25+

Mole

Molal

Factor

Boiling
Point
(Pa)

0.7158
0.7365
0.7757
0.7639
0.7723
0.7814
0.7939
0.8053
0.8096
0.8152
0.8255
0.8303
0.8341
0.8400
0.8477
0.8531
0.8577
0.8666
0.8826

12.05
11. 93
11.90
11.99
12.04
12.04
12.03
12.02
12.13
12.30
12.19
12.20
12.26
12.28
12.29
12.32
12.36
12.38
12.87

641. 7
678.3
727.3
768.4
804.0
832.7
871.2
907.0
947.1
1008.1
1019.0
1039.4
1069.4
1097.6
1130.8
1161.0
1191.4
1234 .9
1465.6

0.8061

12.25
(12.02)*

Mass

Specific
Gravity

0.94
0.84
0.74
0.60
0.41
0.34
0.31
0.26
0.22
0.19
0.17
0.15
0.13
0.11
0.08
0.07
0.06
0.06
0.51

95
104
118
132
144
154
167
180
197
212
226
234
250
262
277
292
308
329
471

6.19

177

Percent

uop

Upper

Mole

Molal

Molal

Percent

Mass

Mass

0.935
0.838
0.739
0.600
0.410
0.340
0.310
0.260
0.220
0.190
0.170
0.150
0.130
0.110
0.080
0.070
0.060
0.070
0.508

95.1
105.4
118.9
134.2
148.7
161. 7
175.0
188.7
202.1
215.4
228.9
242.8
256.8
270.6
283.0
294.1
304.8
316.7
439.1

99.6
111.9
126.8
142.4
155.5
i68.4
182.2
195.7
209.1
222./,
236.1
250.1
2M .1
277 .8
289.0
299.9
310.3
324.0

6.190

177.0

* The higher average Kuop value was calculated using a weight-average

mixing rule, whereas the lower value was estimated using the Whitson
correlation.
i" The gamma distribution (Ref. 11) was used where an optimal alpha of
0.712 was fou.nd for eta (minimum molecular mass in the C7+ fraction)
of 92. Upper molecular masses were found by fitting the measured
compositions.

0.843

TABLE 3. - PHYSICAL PROPERTIES OF THE C7+ SINGLE AND MULTIPLE CARBON NUMBER
GROUPS USED IN THE PENG-ROBINSON EQUATIllN OF STATE TO DESCRIBE
RESERVOIR FLUID BEHAVIOR OF THE NS-1 FLUID

Single

Carbon
Number

7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25+

Mole
Percent

Molal
Mass

Specific
Gravity

1. 2136
1.1730
0.8600
0.6872
0.5681
0.4783
0.4074
0.3498
0.3021
0.2621
0.2282
0.1992
0,1744
0.1529
0.1343
0.1182
0.1041
0.0918
0.7054

95.3
106.5
120.7
134.7
148.7
162.7
176.7
190.7
204.7
218.7
232.7
246.6
260.6
274.6
288.6
302.6
316.5
330.5
462.3

0.7177
0.7409
0.7599
0.7682
0.7781
0.7908
0.8034
0.8153
0.8168
0.8210
0.8310
0.8389
0.8432
0.8486
0.8554
0.8602
0.8639
0.8690
0.9192

8.2100

184.0

0.8160

Boiling
Point
(OR)
646.8
690.6
739.4
781.4
822.2
863.1
902.8
941.2
972.6
1029.8
1039.5
1072 .0
1104.7
1131.6
1161.9
1190.2
1217.4
1245.1
1488.0

Critical
Temp.
(OR)
971.6
1021.4
1073.2
1112.9
1152.1
1192.3
1231. 2
1268.3
1294.2
1321.5
1354.0
1383.6
1408.7
1434.4
1461.0
1484.9
1507.0
1530.2
1734.6

Methane

Press.

Acentric

(psia)

Factor

457.5
423.4
383.2
345.9
316.8
294.0
274.9
258.4
240.4
225.7
214.6
204.3
194.0
185.1
177 .4
170.0
162.9
156.8
91.4

0.2742
0.3056
0.3454
0.3861
0.4251
0.4622
0.4984
0.5342
0.5740
0.6127
0.6486
0.6855
0.7246
0.7634
0.8016
0.8416
0.8830
0.9241
1.0590

Interaction
Coefficient

MULTIPLE CARBON NUMBER PROPERTIES USED IN THE FINAL CVD SIMULATION

7- 9
10-13
14-17
18-24
25+

3.2466
2,1410
1.1421
0.9749
0.7054

106.1
152.7
209.2
281.5
462.3

0.7385
0.7837
0.8205
0.8524
0.9192

8.2100

184.0

0.8160

688.3
835.0
984.4
1146.9
1276.1*

1016.5
1163.6
1304.5
1446.0
1584.2*

425.5
313.1
237.4
182.7
168.8*

0.3044
0.4348
0.5856
0.7832
0.8819*

0.03659***
0.04292***
0.04807***
0.05254***
0.18400**

* Adjusted values representing a Kuop factor of 11.80. The original values

correspond to a Kuop factor of 12.42 and are given above (e.g. 1488.0).

** Adjusted value used to match the measured dew point pressure.

*** Calculated using the Katz and Firoozabadi correlation, curve-fit

to yield

the following relation 6 -G == O.14Yn - 0.0668. Though not shown in this

C
table, the metane-hexane 1 n interaction coefficient was also calculated using
this relation.

TABLE 4 - CALCULATED LIQUID DENSITIES AS A FUNCTION OF PRESSURE FOR NS-1

Calculated Liquid Densities (gm/cc)
Simulated CVD Data

Measured CVD Data

( psia)

Material
Balance
Density

Alani-Kennedy
Density
(M.B. Liquid
Properties)

5514.7

0.670

(l.608

0.541

4314.7

0.680

0.649

0.554

0.596

3114.7

0.688

0.670

0.580

0.632

Pressure

PengRobinson
Density**

Alani -Kennedy
Density
(P~R

Liquid

Properties)

0.570

2114.7

0.700

0.682

0.608

0.664

1214.7

0.711

0.700

0.636

0.692

714.7

0.722

0.711

0.653

0.707

The Alani-Kennedy density correlation requires liquid compositions,

total liquid molecular mass, heptanes-plus molecular mass and specific
gravity (as well as pressure and temperature). These data were available
from either material balance calculations or P-R simulation results.

**

The Peng-Robinson simulation used properties given in Table 3 with an

adjusted Kuop = 11.8 for the C2S+ fraction. Using the original Kuo ]? factor
of 12.42 gave even lower liquid densities than those given above, wlth a
larger deviation from the Alani-Kenedy values.

TABLE 5 -

SllIDLATED CONSTANT VOLUME DEPLETION DATA FOR THE NS-1 FLUID AT 280 of USING THE PENG-ROBINSON
EQUATION OF STATE
Compositions
Equilibrium
Vapor

Equilibrium
Liquid

Pressure - psia
1214.7
2114.7

714.7

Exp.
714.7

Calc.
714.7

2.583
0.321
77.772
8.711
3.989
0.778
1.567
0.638
0.659
0.877
2.105

0.590
0.020
12.420
3.360
2.920
0.910
2.090
1.400
1.600
3.680
71.010

0.595
0.029
11. 939
3.623
3.133
0 .. 967
2.314
1.509
1.770
4.223
69.897

100.000

100.000 100.000

100.600

100.000

121.2

116.4

213.0

209.1

Component

6764.7

5514.7

4314.7

3114.7

Carbon Dioxide
Nitrogen
Methane
Ethane
Propane
iso-Butane.
normal-Butane
iso-Pentane
normal-Pentane
Hexanes
Heptanes-P1us

2.370
0.310
73.190
7.800
3.550
0.710
1.450
0.640
0.680
1.090
8.210

2.403
0.323
75.549
7.779
3.474
0.686
1.390
0.604
0.639
0.996
6.157

2.447
0.336
77.644
7.793
3.405
0.660
1.326
0.564
0.592
0.889
4.343

2.497
0.344
79.135
7.880
3.383
0.644
1.281
0.530
0.550
0.789
2.969

2.541
0.343
79.712
8.057
3.444
0.647
1.282
0.516
0.532
0.727
2.198

Totals

100.000

100.000

100.000

100.000

184.0

161.0

142.7

129.1

C7 +

0.816

C7+

p
SL -

0.783

0.770

0.762

0.758

114.5
0.757

1.203

1.037

0.937

0.890

0.886

0.911

0.936

0.000

9.637

22.581

39.492

56.196

72 .413

81. 535

0.00

26.11

26.65

25.11

23.00

21.58

0.799

----

2.576
0.334
79.242
8.372
3.660
0.691
1.375
0.548
0.563
0.744
1.895

19.55

0.833

0.843

IDENTICAL MULTI-STAGE SEPARATION

B
9

Vg
"-

VSTO

P ZT

"_~_c._

hc.p

VAPOR

Yj

PVT CELL

Fig. 1 - Schematic description of the procedure for calculating "black

011" PVT properties.

u..

0.030

NS-1
0
280 F

(/)

.......
"'.j..J

'I-

co"

OJ

---0----

0.025

---6-

PENG-ROBINSON Z-FACTORS
MEASURED Z-FACTORS

c:::

0
I-

e(

u..

0.020

uJ

::E
::J
...J

>
z

0.015

l-

e(

::E

c:::

0
u..

0.010

(/)

e(
I.!)

0.005

o. 000

1-.L...L......L....l.-I-.L...J-1.....l.-L....l-..I-l..-'---'-l-.l-I.....L......L-l-.l-L....L......L-L-J.....J-..I-l.......L-J-l-.l-J

1000

2000

3000

4000

5000
6000
7000
PRESSURE, psia

Fig. 2 - Gas formation volume factor vs pressure for NS1 at 280 'F.

50000

In
l-

,........-,-,-,--.--r-r--r-r-,-,........-,-,-,--.--r-r--r-r-,-,r-r-...,..--r-r-T""""T-'-1'"""""T'"-,-,---r--t
-0-

(/)

.......

-6-

I.L.
U
(/)

PENG-ROBINSON RESULTS
MATERIAL BALANCE RESULTS
USING MEASURED DATA

..

~ 40000

a:::

..

l-

e:(

a:::

30000

...J
H

o
I

(/)

e:(
(,.!)

o
H

20000

I-

::>
...J

o(/)
a:::

10000

SEPARATOR
PRESSURE
(psia)
1014.7
264.7
14.7

CONDITIONS *
TEMPERATURE

(0 F)

155

80

60

STANDING LOW PRESSURE

"BLACK OIL" K-VALUE
CORRELATION USED

5000
6000
7000
PRESSURE, psia
Fig. 3 - "Vapor" solution gas-oil ratio vs pressure for NS1 at 280 'F.

In
l-

2 0

.......
-J

..c
OJ

.
..
a:::

In

1.8

I-

U
c:(
I.L.

,........-,-,r-r-..,-,r-r-..,-,-,..,-,-,.,--,..-,..,..-,..,--r-,..,.,-o--T--r-T1""-o--T..."
~PENG-ROBINSON RESULTS
-6-MATERIAL BALANCE RESULTS
-o-MATERIAL BALANCE RESULTS
USING P-R LIQUID VOLUMES (SL)

(/)

1.6

::>
...J

o
>

SEPARATOR
PRESSURE
(psia)
1014.7
264.7
14.7

CONDITIONS*
TEMPERATURE
(0 F)

155
80

60

::;;:

1.4

a:::
o

I.L.

...J
H

1.2

*STANDING LOW PRESSURE,

"BLACK OIL" K-VALUE
CORRELATION USED

1000

2000

3000

4000

5000
6000
7000
PRESSURE, psia

Fig. 4-011 formation volume factor vs pressure for NS-1 at 280 of.

2000 rr-'--~--,-~-'--r-r-'--~-'--rr-,--,r-r--nr-r---,-;-r--,-;--r-,-,--r-r-r-,-,

(/)

--0-

.......

u..

-6-

U
(/)

-0-

a"

PENG-ROBINSON RESULTS
MATERIAL BALANCE RESULTS
MATERIAL BALANCE RESULTS
USING P-R LIQUID VOLUMES (SL)

NS-1
280 of

(f)
~

o" 1500

le:(
~

--l

SEPARATOR CONDITIONS*
PRESSURE TEMPERATURE
(psi a)

( oF)

1014.7
264.7
14.7

155
80
60

o
I

(/)

~ 1000
z

o
H

I-

::::>

--l

o(/)

e
::::>

STANDING LOW PRESSURE

"BLACK OIL" K-VALUE
CORRELATION USED

500

C!J
H

--l

1000

2000

3000

4000

5000

6000

7000

PRESSURE, psia
Fig. 5 - "Liquid" solution gas-oil ratio vs pressure for NS1 at 280 0 F.

(/) 240 r--r--,....-T""""""T---r-..,--,---.--.---r-r-r-~--,-~--,-r-r-,....--r-r---,-...-T--r,...,,-,-..---,-,-,......,

UJ
(/)
(/)

-0-

-t:.-

e:(

::E

PENG-ROBINSON MATCH
MATERIAL BALANCE RESULTS
USING MEASURED DATA

220

--l

::::>
u
UJ

--l

~ 200
(/)

::::>
--l

e..

~
180
UJ

e:(

l-

e..
UJ

:c 160

140

120

* MEASURED

NS-1
(-6-)

280 0 F
7000
psia

Fig. 6 - Caloulated and measured vapor oomposltiQns vs pressure for NS1 at 280 0 F.

o
......
o
...

. . . - - _ ) - - - -....0 - -_ _--<

(/)

I-!

lI-!

(/)

0:E

>------< > - - - - _ 0

PENG-ROBINSON MATCH
MEASURED (ACTUALLY THE DATA HAVE BEEN
"SMOOTHED" AND ADJUSTED ACCORDING
TO CORE LABORATORIES PROCEDURE)

10

...J

-0-_-=.........:::8

C2 < > - - - - 0 - - - 0 - - - - - 0 - - _ - 0 -

<C

...J

:E

C3

0<C

C02 o-~---<:::;:;_.-<:::::...--e-----o---__<>------o
o
C7 +
n-C 4 ~-_ _....OJ___ __<O>------<O>-----O>-----O

>

N20---<>-----00----0>------0>------0>-

NS-1
280F

10 -1 L...J..-l-J...,....!.-I-.J......J.-l.....J.......J--'-...J......JL......l..-l-.l-l......l-J....J..-l-J...,....!.-I.-.J......J.-l.....J.......J--'--'--L......l.......L.....J

1000

2000

3000

4000

5000

6000

7000

PRESSURE, psia
Fig. 7 - Calculated heptanesplus molecular masses vs pressure for NS1 at 280 of.

C\l

'r-

10 5 ~r-..-..-,---,---.-..,--r--r-r--.- r-r-"-'-'---'-'-7r-r--rr-r-r-r--;

Vl

Q.

...

~ 10

...

I-

INSERT

25o

10 3

0W

~ 10 2
=>
Vl
Vl
W
~

0I

10

IZ

<C

IVl

EXTRAPOLATION OF SLOPES
TO ZERO YIELDS
APPARENT CONVERGENCE
PRESSURE'" 7500 psia

:E

=>
H
~

CO

I-!

...J
H

=>

1 0'L-c-~~--'-'--'-~--'---'~~---"----'----'-~L....L~~-L-J
2.0
2.5
3.0
3.5
4.0

OJ
W

10-3

*ADJUSTED C7+ F VALUES

I....-.l....-.l....-.l....-.l....-.l....-.l....-.i--.i--.i--.i--.i--.i--.i--.i--.l....-.i--.i--.i--.i--.i--........................" ' - - '

-4

-2

COMPONENT CHARACTERIZATION FACTOR, F=b(1/Tb - 1/T)

Fig. 8- NS1 Kvalues at 280 of calculated using the material balance approach.

ro

10 5 r-r--r-r--,-..,-.-r-,-r--,--'--.-r-,-,--,--,--,....,-,-,--,--,--,....,-,-,--,-.,..--,

'r-

PRESSURE

MATERIAL BALANCE RESULTS

USING MEASURED DATA
1- 5514.7 _ LEAST SQUARES LINEAR
2-4314.7
REGRESSION
3-3114.7
(N 2 AND C7+ DATA EXCLUDED)
4-2114.7
5-1214.7
6- 714.7

(fj

Q.

,
I-

25

o
0::
a..

(psia)

10

UJ

0::
:::J
V)
V)

UJ

0::

0...

103

V)

C3

a..
o

IZ

UJ

....J

Iii -C 4
n- C4
Iii -C 5

c:(

lV)

V)

n-C5

~ 10

t-l

0::

CO
t-l

....J
t-l

:::J

C!J

UJ

C7+
10 1 L......J.---l..--'-:..L...-l-.J.---l..--'--'--l-.J.---l..--'-...L-l-.J.---l..--'-...L-l-.J.---l..--'--'--l-.J.---l..--'--l-l
-1
0
1
2
3
4
5

COMPONENT CHARACTERIZATION FAtTOR, F=b(1fTb - 1fT)

Fig. 9 - NS1 Kvalues at 280 'F oaloulated using the Peng-Robinson EOS.

ru

280 of
--- 340 of

'r(fj

Q.

Q.

1-'

10 4

__

~":<"';'ifr'

:::J
Cl

o
0::
a..

APPARENT
CONVERGENCE
PRESSURE

UJ

0::
:::J
V)
V)

10 3

7500 psia

UJ

0::

a..
I

IZ
c:(

10 2

lV)

:E
:::J

CO

10

t-l

....J
t-l

:::J

NS-1

C!J
UJ

280 of

1L......L.--'-.L.......L.--'-.L.......L.--'-.L.......L.....l-l-.L......l-l-.L......l-L-l.....J.....J-J-..J-J-J--'-'-'--'-'-'-..l-.J.--'-l
-2

-1

COMPONENT CHARACTERIZATION FACTOR, F=b (1fTb - 1fT)

Fig. 10 - Temperature effeot on the Hoffman, et aI, Kvalue oorrelatlon for NS1.

t-

4.0 r-----,--,----,--,-----,--r-----,--r-----,--,----,--,----,--,----,----..

--- PENG-ROBINSON MATCH (2800F)

r::s

--- PENG-ROBINSON MATCH (340F)

!Z

log

MATERIAL BALANCE RESULTS

Cl

Z
<C

3.0

UJ

C-

-l

en

u..

en

>

{f 2.0

EXTRAPOLATED

1.0

'"w

'">
z

NS-1
280 of 1340 of
0.00
2000

6000
PRESSURE,

8000
ps~a

Fig. 11 - Slopes and Intercepts of log Kp vs F plots vs pressure for NS-1.

- COSTANT CO POSITION
----CONSTANT VOLUME DEPLETION

NS-1
280 of

N2

C1
C0

..

C~

t-

Z
<C

t-

en
z

;-0,

10-1

~=~~;;~~~~;;.~~~~.~~.it
n-C6~

C7 -C 9

a::

co

C -C

-l
H

::::>

lO

10-2

13

ro

I .~
I .[ ~

5514.7"'; ~

:1; ~
~ ~

C!J
UJ

II

'":::>

'">'"

U)
U)

'"0.
c;';

2114.7

g:
z

uJ

o.

c-

3
W

uJ

'"

0.
0.

10- 5

C25 +

10-6

1214.7

PRESSURE
(ps i a) 714.7
'----:---'--'--'-.L.....J...-'--_ _---'_ _'---L......L.....L....l-.l-L..J

3x10

103
PRESSURE,

104
psi a

Fig. 12 - Peng-Roblnson K-values for NS1 at 280 of representing two

depletion processes.

CONSTANT VOLUME DEPLETION

CONSTANT COMPOSITION
N2
C1
CO 2
C2

10

...

IZ

ex:

. C3
,-C 4

l(/)

n-C4

:E
:::J
H

'"
'"

D.

IJ)
H

oo

-l

C 1 0 - C 1 3 1 .[ ~

:::J

r:JJ

UJ

llJ

'":::>
U)
U)

UJ

'"
0-

UJ
U

UJ

'":::>

Z
UJ

U)

:z '"
'" z>
OJ

UJ

0-

f--

I
2114.7

'"~

f--

0-

25+~~~""""~1

UJ

'"
00-

PRESSURE I
(psia)714.7

1214.7

10 -5 L---L-l-....l...-.!-l.....l-I.-::-_ _..l.-_..l.---L.---l..-l-...l--I.....J...J

103

3 X 10 2

104
PRESSURE, psia

Fig. 13- Peng-Robinson K-values for NS-1 at 340 'F representing two
depletion processes.

>< 10 2 ~---,----r--,---.-TT..,----~--r----,----r--,---,--rT""'I

--..>..

CONSTANT COMPOSITION
CONSTANT VOLUME DEPLETION

II

NS-Z
241F

...

I-

ro

10

(f)

l-

Q.

(/)

C2
C3
i -C4 ---------

:E

:::J
H

n - C4 =::::::::::::::

IJ)
H

-==

:::::::==----:::::;;

1\
llJ
~

:::J

-l

(/)

(/)

:::J

o
o
o

llJ

10- 1

l:l..
llJ
U

Z
llJ

ClO
C11
C12
C13
C14

l!:l
~

llJ

>
Z
o
u

10 -4 L--'---'----.l..--L-LLL-_ _--'--_--'--_.L.-...L.-.L.....l---.l..-.L.

3x10

10 3

10 4
PRESSURE, psia

Fig. 14-Peng-Roblnson K-values for NS-2 at 241

depletion processes.

'F representing two