Fluid Os 2012

Fluidos de Reservorio
G. Fondevila
Actividades en la
Ingeniera de Reservorios
Observaciones
Asunciones (Hiptesis)
Clculos
Decisiones
(L.P. Dake The Practice of Reservoir Engineering)
Responsabilidades Tcnicas en la
Ingeniera de Reservorios
Determinar, en conjunto con los gelogos, geofsicos y
petrofsicos, los recursos (hidrocarburos originales in situ).
Determinar la fraccin de estos hidrocarburos que puede ser
(razonablemente) recuperada.
Llevar la recuperacin a una escala temporal (pronsticos).
Efectuar el control de reservorios operativo, durante toda la
vida del proyecto.
Rol del Ingeniero/a de Reservorios
Definiciones
Reservorio: Acumulacin de hidrocarburos con
vinculacin hidrulica entre todos sus puntos.
Yacimiento: Uno o ms reservorios de hidrocarburos
agrupados y/o relacionados entre s, dentro de una
misma trampa geolgica. En un mismo yacimiento
pueden coexistir mltiples reservorios separados vertical
o lateralmente por rocas impermeables y/o barreras
geolgicas locales.
Trampa: Configuracin que impide la normal movilidad
de los hidrocarburos provocando su acumulacin. Puede
ser de origen estructural, estratigrfica o combinada.
The Need to Understand
Phase Behavior
As oil and gas are produced from the reservoir,
they are subjected to a series of pressure,
temperature, and compositional changes.
Such changes affect the volumetric and
transport behavior of these reservoir fluids and,
consequently, the produced oil and gas volumes.
(M.A. Barrufet)
The Need to Understand
Phase Behavior
Type of reservoir fluid determines depletion and
production strategies and the design of surface
facilities
Except polymer flooding, all of EOR methods
rely on the phase behavior of reservoir fluids and
fluids injected into the reservoir.
This behavior is used to classify the recovery
method (i.e., thermal, miscible, chemical, etc.),
and to design the recovery process.
(M.A. Barrufet)
Major Definitions
System: A body of matter with finite
boundaries (physical or virtual)
Closed System: Does not exchange matter
with surroundings but may exchange energy
(heat).
Open System: Does exchange matter and
energy with surroundings.
(M.A. Barrufet)
Major Definitions
Homogeneous System: Intensive
properties change continuously and uniformly
(smoothly)
Heterogeneous System: System made up
of two or more phases in which the intensive
properties change abruptly at phase-contact
surfaces
(M.A. Barrufet)
Major Definitions
Phase: A portion of the system which
has homogeneous intensive properties
and it is bounded by a physical surface.
Interface: Separates two or more
phases. These phases are solid,
liquid(s), and gas.
(M.A. Barrufet)
Major Definitions
Intensive Properties: Independent
of system mass (i.e density)
Extensive Properties: Dependent
of system mass (i.e volume)
(M.A. Barrufet)
Major Definitions
Properties: Characteristics of a system (phase) that
may be evaluated quantitatively, i.e.
Phase density (liquid, gas, solid)
Phase compositions
Isothermal compressibility
Surface tension
Viscosity
Heat capacity
Thermal conductivity
(M.A. Barrufet)
Major Definitions
Component: A molecular species,
defined or hypothetical.
Defined: C
l
, C
2
, H
2
O, etc.
Hypothetical: lumped defined (i.e. C
2
-C
6
), or
undefined C
7
+
, C
20
+
(M.A. Barrufet)
Major Definitions
State: Condition of a system at a
particular time determined when all
intensive properties are fixed
(M.A. Barrufet)
Diagrama PT: Sustancia Pura
Diagrama PT: Sustancia Pura
Diagrama PT: 2 Componentes
Diagrama PT: 2 Componentes
Diagrama PT: Mltiples Componentes
Diagrama PT: Definiciones
Tipos de Fluido de Reservorio
De qu depende el tipo de fluido de reservorio?:
Composicin de la mezcla de HC en el reservorio.
Temperatura Inicial de reservorio (Tres)
Presin Inicial de reservorio (Pi).
Ubicacin de la Temperatura de Reservorio respecto a la Temperatura
Crtica y al punto Cricondentrmico.
Pueden ser clasificados en dos tipos:
Reservorios de Petrleo: Temperatura del Reservorio (Tres), es menor
a la Temperatura Crtica (Tc). Estn clasificados en:
Subsaturados: Presin Inicial (Pi) es mayor a la Presin de burbuja (Pb).
Saturados: Presin Inicial (Pi) es igual a la Presin de burbuja (Pb).
Con Casquete/Calota de Gas: Presin Inicial (Pi) es menor a la Presin de
burbuja (Pb).
Reservorios de Gas: Temperatura del Reservorio (Tres), es mayor a la
Temperatura Crtica (Tc).
Existen 5 Tipos de Fluidos de Reservorio:
Reservorios de Petrleo:
Petrleo Negro
Petrleo Voltil
Reservorios de Gas:
Gas Retrgrado
Gas Hmedo
Gas Seco
Petrleo Negro
Petrleo Voltil
Gas Retrgrado (Gas y Condensado)
Gas Hmedo
Gas Seco
Comparacin Composicin
Componente
Petrleo
Negro
Petrleo
Voltil
Gas
Retrgrado
Gas
Hmedo
Gas Seco
C1 48.8 64.4 87.1 91.4 95.8
C2 2.8 7.5 4.4 3.5 2.7
C3 1.9 4.7 2.3 1.2 0.3
C4 1.6 4.1 1.7 1.1 0.5
C5 1.2 3.0 0.8 0.4 0.1
C6 1.6 1.4 0.6 0.3 0.1
C7+ 42.1 14.9 3.8 2.1 0.4
100.0 100.0 100.0 100.0 100.0
PM C7+ 230 180 160 140 -
GOR [scf/stb] 1,200 2,500 9,000 >15,000 >100,000
API 35 50 60 >60 -
Whitson
SPE 28214
Heavy Components Control
Reservoir Fluid Behavior
William D. McCain Jr.
Se definen 5 tipos de fluido de reservorio
(petrleo negro, voltil, gas + condensado, gas
hmedo y gas seco) debido a que cada uno
necesita diferentes tcnicas de ingeniera para
producirlo.
El mismo debe ser definido bien temprano en la
explotacin de un reservorio, ya que es un
factor crtico para la toma de decisiones sobre
cmo producir el fluido.
Introduccin (1)
El tipo de fluido de reservorio se debe definir
con una muestra representativa del mismo en
laboratorio.
Igualmente existen reglas de dedo basadas
en RGP inicial, densidad y color de lquido de
tanque, que pueden indicar el tipo de fluido.
Estos fluidos son mezclas de hidrocarburos que
van desde componentes livianos (ej: metano)
hasta molculas no voltiles ultra-pesadas.
Introduccin (2)
Los petrleos negros se caracterizan por tener
una gran cantidad de molculas pesadas. Los
petrleos voltiles poseen menos de estas
molculas, y as hasta llegar a gas seco que es
puramente metano.
Generalmente en laboratorio a las molculas
pesadas se las concentra en un pseudo-
componente llamado C
7+
.
En cuanto a RGP, los petrleos negros es la
ms baja y en gas hmedo es mucho mayor.
G
O
R

i
n
i
c
i
a
l
C7+ %molar
G
O
R

i
n
i
c
i
a
l
C7+ %molar
GASES
PETRLEOS
Petrleos Negros y Voltiles (1)
Los petrleos negros y voltiles ambos estn en
fase lquida en el reservorio.
Exhiben punto de burbuja cuando se depleta el
reservorio (baja la presin).
Ambos liberan gas en reservorio cuando la presin
est por debajo de la presin de burbuja.
El gas que liberan los petrleos negros es
usualmente gas seco. A medida que la presin
decrece el gas puede llegar a ser hmedo. Pero
esto ocurre en el final de la vida de explotacin.
El gas que proviene de petrleos voltiles es
usualmente gas retrgrado.
Este gas va a condensar una gran cantidad de
lquido en condiciones de superficie.
Por eso en las ecuaciones de balance de masa (lo
van a ver en ingeniera de reservorios...) la
hiptesis es que el gas que sale del reservorio
permanece como gas en superficie.
Por ende los petrleos voltiles no pueden tratarse
con balance de masa sino con simuladores
composicionales.
En la ecuacin de balance de masa para petrleos
negros es que es una mezcla de dos
componentes: petrleo y gas.
Sobre el punto de burbuja (slo tenemos lquido) el
balance de masa se puede realizar tanto para
petrleo negros como voltiles.
Indicadores (lmite inferior petrleos voltiles):
GOR > 300 m3/m3
Densidad > 40 API
Bo > 2
%molar C
7+
< 20%
Petrleos Voltiles y Gas Retrgrado (1)
En condiciones de reservorio los petrleos
voltiles demuestran punto de burbuja y los
gases retrgrados punto de roco, al bajar la
presin.
GOR = 600 m3/m3 y %molar C
7+
= 12.5 %
puede tomarse como un lmite razonable entre
petrleos voltiles y gases retrgrados.
Estos lmites deben ser tomados como guas,
ya que se pueden encontrar petrleos voltiles
con GOR > 600 m3/m3 y %molar C
7+
< 12.5%,
y viceversa con los gases retrgrados.
G
O
R

i
n
i
c
i
a
l
C7+ %molar
Petrleos Voltiles y Gas Retrgrado (2)
El condensado que se forma debajo de la
presin de roco es prcticamente inmvil, por
lo que se pierde para la produccin.
Esto causa una disminucin en la
permeabilidad efectiva al gas debido a que la
saturacin de lquido va aumentando.
En general se ve una disminucin marcada en
la produccin de gas cuando la presin se
encuentra por debajo del punto de roco.
Gases Retrgrados y Hmedos (1)
Se han visto gases con comportamiento
retrgrado con GOR = 27000 m3/m3.
Aparentemente, todos los gases que
demuestran condensacin en superficie, liberan
un poco de condensado en fondo.
La ecuacin de balance de masa de gases se
puede aplicar a gases hmedos:
Combinando el gas de superficie y el condensado
para determinar el gas en reservorio.
Obtener el equivalente gaseoso del condensado de
superficie a la produccin de gas en superficie.
Gases Retrgrados y Hmedos (2)
%molar C
7+
por debajo de 4% se puede
considerar un gs hmedo, aunque poco
lquido condense en reservorio.
Para ese %molar, se aplica un GOR de 2700
m3/m3 (segn la figura 1) por lo que tambin se
puede tomar como lmite entre gases
retrgrados y hmedos.
El GOR de un reservorio de gas hmedo se
mantiene constante a lo largo de la vida del
reservorio.
Gases Hmedos y Secos
Ambos permanecen en fase gaseosa al bajar la
presin en reservorio. Ninguno muestra punto de
roco y liberacin de condensado en reservorio.
La diferencia es que los gases hmedos liberan
condensado en condiciones de superficie,
mientras que un gas seco siempre permanece
como gas.
El efecto de la liberacin de condensado en
superficie es despreciable si es menor que 55 m3
condensado / 1 MM m3 de gas.
GOR = 18000 m3/m3 puede ser usado como
lmite.
Fluidos de Reservorios
Gases
G. Fondevila
Fuentes
Equations of State and PVT Analysis T. Ahmed
Fundamentals of Reservoir Engineering, L.P. Dake
SPE 75721: Simplified Correlations for Hydrocarbon
Gas Viscosity and Gas Density Validation and
Correlation of Behavior Using a Large-Scale Database,
F.E. Londono, R.A. Archer and T.A. Blasingame, Texas
A&M U.
Ecuacin de los Gases Ideales
El gas es una de las pocas substancias donde su estado
queda definido por la presin, el volumen que ocupa y la
temperatura mediante una simple relacin entre estos 3
parmetros.
Ecuacin de gases ideales:
pV = nRT
Esta ecuacin es el resultado de la combinacin de los
esfuerzos de Boyle, Charles, Avogadro y Gay Lussac,
pero es solo aplicable a presiones cercanas a la
atmosfrica.
Hiptesis de los Gases Ideales:
No existen fuerzas de interaccin (atraccin o repulsin) entre
las molculas de gas.
Choques que se producen entre las mismas son perfectamente
elsticos.
El volumen de las molculas de gas es totalmente despreciable
en comparacin con el volumen total que ocupa el mismo.
Ecuacin de los Gases Reales
Ecuacin de van der Waals (para 1 mol de gas):
(p + a/V
2
)*(V b) = RT
Esta ecuacin trata de representar los efectos
de interaccin de las molculas de los gases
(atraccin y repulsin, parmetro a) y adems
tiene el cuenta el volumen que ocupan estas
molculas (parmetro b), ambos efectos
comienzan a ser visibles al aumentar la presin.
Esta ecuacin puede ser reducida a:
pV = ZnRT
Z = Vol gas real / Vol gas ideal
Ecuacin de los Gases Reales
Determinacin del factor de compresibilidad Z:
Manera experimental: es el mtodo ms preciso,
pero consume mucho tiempo de laboratorio y dinero,
y la precisin obtenida no es determinante como para
no utilizar las correlaciones o clculos existentes.
Correlaciones:
Standing & Katz: para determinar el factor Z
Sutton & Brown: para determinar las propiedades pseudo-
crticas del gas a partir de la gravedad especfica (en el caso
que no tengamos una cromatografa del mismo).
Clculo directo del Z: Existen diferentes
metodologas de clculo hoy en da que representan
a la correlacin de Standing & Katz, podemos
mencionar:
Hall-Yarborough
Dranchuk-Abou-Kassem
Correlacin Standing & Katz
Se utiliza para determinar el factor de
compresibilidad Z.
Requiere saber la composicin del gas.
Para utilizar la correlacin, necesitamos conocer
las propiedades pseudo-crticas del gas (las
propiedades de cada componente se
encuentran tabuladas):
Ppc = (yi . Pci)
Tpc = (yi . Tci)
Luego calculamos la presin y temperatura
pseudo-reducidas:
Ppr = P/Ppc
Tpr = T/Tpc
Correlacin Sutton & Brown
Esta correlacin sirve para
determinar las propiedades
pseudocrticas del gas (Ppc y
Tpc) cuando no tenemos la
composicin del mismo, pero s
la gravedad especfica g.
Propiedades Elementos Puros
Clculo Propiedades del Gas Natural (1)
Clculo de Propiedades del Gas (2)
Factor Volumtrico de Formacin del Gas: Bg
Compresibilidad del Gas
Resumen Clculo de Propiedades
Peso Molecular:
PM_gas = (yi * PMi)
Gravedad Especfica:
SG_gas (g) = Dens_gas / Dens_aire = PM_gas / PM_aire
Dens_aire = 1.223 kg/m3
PM_aire = 28,93 gr/mol
Bg:
Bg = Vol fondo / Vol sup
Vol fondo = ZnR Tfondo/Pfondo
Vol sup = nR Tsup/Psup
Bg = Z (Tfondo/Tsup) (Psup/Pfondo)
Compresibilidad del gas real:
Cg = - 1/V dV/dP = 1/P 1/Z dZ/dP 1/P
Efecto de Impurezas en el Gas
Generalmente los gases naturales vienen acompaados
de elementos no hidrocarburos, a ser:
Nitrgeno (N
2
)
Dixido de Carbono (CO
2
)
Sulfuro de Hidrgeno (H
2
S)
A un gas se lo clasifica como dulce o amargo
dependiendo del contenido de H
2
S en el mismo.
La presencia de impurezas (generalmente por encima
de un 5% total en composicin) puede generar errores a
la hora de calcular correctamente el Z.
Correccin por Impurezas
Clculo Directo de Z (1): Dranchuk & Abu-Kassem
Clculo Directo de Z (2): Hall-Yarborough
Compresibilidad del Gas: Correlacin de Trube
cpr = cg * Ppc cg = cpr / Ppc
donde cpr la obtenemos de la correlacin de Trube
Viscosidad del Gas
Clculo de Viscosidad del Gas:
Correlacin de Carr-Kobayashi-Burrows
Paso 1
Paso 1: Obtenemos viscosidad del gas a presin armosfrica y temperatura de inters = gatm
Paso 2
Paso 2: Obtenemos los coeficientes multiplicadores para cada presin pseudo-reducida
Cromatografa Gas
Componente Fraccin
C1 91.6%
C2 3.8%
C3 1.2%
iC4 0.3%
nC4 0.4%
iC5 0.1%
nC5 0.2%
C6 0.6%
C7+ 0.2%
N2 1.2%
CO2 0.4%
PVT Gas
Pressure Z Bg cg g g
(kg/cm) ( ) (m3/stm3) (psi-1) (g/cc) (cp)
1.0 0.999 1.2384 6.81E-02 0.001 0.0127
19.5 0.977 0.0642 3.69E-03 0.012 0.0129
37.9 0.957 0.0323 1.93E-03 0.024 0.0132
56.4 0.939 0.0213 1.32E-03 0.036 0.0135
74.9 0.923 0.0158 1.00E-03 0.049 0.0139
93.3 0.910 0.0125 8.03E-04 0.062 0.0144
111.8 0.899 0.0103 6.66E-04 0.075 0.0150
130.2 0.892 0.0088 5.64E-04 0.088 0.0157
148.7 0.888 0.0077 4.83E-04 0.101 0.0164
167.1 0.888 0.0068 4.17E-04 0.113 0.0173
185.6 0.890 0.0061 3.62E-04 0.125 0.0182
204.0 0.895 0.0056 3.17E-04 0.137 0.0191
222.5 0.903 0.0052 2.78E-04 0.148 0.0201
241.0 0.913 0.0049 2.46E-04 0.159 0.0211
259.4 0.925 0.0046 2.18E-04 0.169 0.0220
277.9 0.939 0.0043 1.95E-04 0.178 0.0230
296.3 0.954 0.0041 1.74E-04 0.187 0.0240
314.8 0.970 0.0039 1.57E-04 0.195 0.0249
333.2 0.987 0.0038 1.42E-04 0.203 0.0258
351.7 1.006 0.0037 1.29E-04 0.210 0.0266
370.1 1.025 0.0035 1.18E-04 0.217 0.0275
PVT Gas - Z
0.88
0.90
0.92
0.94
0.96
0.98
1.00
1.02
1.04
0 50 100 150 200 250 300 350 400
Presin [kg/cm2]
Z
PVT Gas - Bg
0.001
0.01
0.1
1
10
0 50 100 150 200 250 300 350 400
Presin [kg/cm2]
B
g

[
m
3
/
m
3
]
PVT Gas - cg
0.0001
0.001
0.01
0.1
1
0 50 100 150 200 250 300 350 400
Presin [kg/cm2]
c
g

[
p
s
i
-
1
]
PVT Gas - g
0.00
0.05
0.10
0.15
0.20
0.25
0 50 100 150 200 250 300 350 400
Presin [kg/cm2]

g

[
g
r
/
c
c
]
PVT Gas - g
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0 50 100 150 200 250 300 350 400
Presin [kg/cm2]

g

[
c
p
]
Uso de Correlacin:
1. Obtengo Ppc y Tpc (presin y
temperatura pseudo-crticas):
Ppc = (yi * Pci)
Tpc = (yi * Tci)
2. Calculo Tpr:
Tpr = Tres/Tpc
3. Calculo Ppr:
Ppr = P/Ppc
4. Entro con Ppr y llego a la curva
de Tpr, luego leo z.
5. Ejemplo:
Tpr = 1.8
Ppr = 2
Z = 0.92
Correlacin Sutton & Brown
Uso de Correlacin:
1. Calculo GEg del gas:
GEg = PM_gas / PM_aire
PM_gas = (yi * PMi)
PM_aire 28.93 gr/mol
2. Obtengo Tpc y Ppc del grfico.
3. Ejemplo:
GEg = 0.7
Tpc = 390 R
Ppc = 670 psia
Tambin puedo usar las
ecuaciones que aparecen en el
grfico para obtener Tpc y Ppc a
partir de la GEgas.
Compresibilidad del Gas: Correlacin de Trube
cpr = cg * Ppc cg = cpr / Ppc
donde cpr la obtenemos de la correlacin de Trube
Uso de Correlacin:
1. Elijo curva de Tpr.
2. Entro con Ppr.
3. Leo el valor de cpr.
4. Ejemplo:
1. Curva Tpr = 1.8
2. Ppr = 2
3. cpr = 0.54
4. cg = cpr / Ppc
Carr-Kobayashi-Burrows
Paso 1
Paso 1: Obtenemos viscosidad del gas a presin armosfrica y temperatura de inters = gatm
Uso de Correlacin:
1. Elijo curva de Tres.
2. Entro con PM del gas.
3. Leo el valor de gas1atm.
4. Ejemplo:
1. Tres = 200 F
2. PMgas = 20
3. gas1atm = 0.0123
Paso 2
Paso 2: Obtenemos los coeficientes multiplicadores para cada presin pseudo-reducida
Uso de Correlacin Paso 2:
1. Marco lnea de Tpr.
2. Leo el multiplicador de viscosidad para
cada curva de Ppr.
3. Calculo viscosidad para cada Ppr.
4. Ejemplo (flecha celeste):
1. Tpr = 1.8
2. Curva Ppr = 2
3. Multiplicador = 1,15
4. Viscosidad @ Ppr 2 = visc_1atm *
multiplicador = 0.0123 * 1.15 =
0.0141

Copyright 2002, Society of Petroleum Engineers Inc.

This paper was prepared for presentation at the SPE Gas Technology Symposium held in
Calgary, Alberta, Canada, 30 April2 May 2002.

This paper was selected for presentation by an SPE Program Committee following review of
information contained in an abstract submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petroleum Engineers is
prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300
words; illustrations may not be copied. The abstract must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

Abstract
The focus of this work is the behavior of gas viscosity and gas
density for hydrocarbon gas mixtures. The viscosity of hydro-
carbon gases is a function of pressure, temperature, density,
and molecular weight, while the gas density is a function of
pressure, temperature, and molecular weight. This work pre-
sents new approaches for the prediction of gas viscosity and
gas density for hydrocarbon gases over practical ranges of
pressure, temperature and composition. These correlations
can be used for any hydrocarbon gas production or transport-
ation operations.

In this work we created an extensive database of measured gas
viscosity and gas density (>5000 points for gas viscosity and
>8000 points for gas density). This database was used to
evaluate existing models for gas viscosity and gas density. In
this work we provide new models for gas density and gas
viscosity, as well as optimization of existing models using this
database.

The objectives of this research are:
l To create a large-scale database of measured gas vis-
cosity and gas density data which contains all of the in-
formation required to establish the applicability of var-
ious models for gas density and gas viscosity over a
wide range of pressures and temperatures.
l To evaluate a number of existing models for gas vis-
cosity and gas density.
l To develop new models for gas viscosity and gas den-
sity using our research database these models are
proposed, validated, and presented graphically.
For this study, we created a large-scale database of gas pro-
perties using existing sources available in the literature. Our
data-base includes: composition, viscosity, density, tempera-
ture, pressure, pseudoreduced properties and the gas com-
pressibility factor. We use this database to evaluate the appli-
cability of the existing models used to estimate hydrocarbon
gas viscosity and gas density (or more specifically, the z-
factor). Finally, we provide new models and calculation pro-
cedures for estimating hydrocarbon gas viscosity and we also
provide new optimizations of the existing equations-of-state
(EOS) typically used for the calculation of the gas z-factor.

Introduction
Hydrocarbon Gas Viscosity

NIST SUPERTRAP Algorithm: The state-of-the-art mechan-
ism for the estimation of gas viscosity is most likely the com-
puter program SUPERTRAP developed at the U.S. National
Institute of Standards and Technology
1
(NIST). SUPERTRAP
was developed from pure component and mixture data, and is
stated to provide estimates within engineering accuracy from
the triple point of a given substance to temperatures up to
1,340.33 deg F and pressures up to 44,100 psia. As the
SUPERTRAP algorithm requires the composition for a parti-
cular sample, this method would not be generally suitable for
applications where only the mixture gas gravity and composi-
tions of any contaminants are known.

Carr, et al. Correlation: Carr, et al.
2
developed a two-step
procedure to estimate hydrocarbon gas viscosity. The first
step is to determine the gas viscosity at atmospheric condi-
tions (i.e., a reference condition). Once estimated, the vis-
cosity at atmospheric pressure is then adjusted to conditions at
the desired temperature and pressure using a second correla-
tion. The gas viscosity can be estimated using graphical cor-
relations or using equations derived from these figures.

Jossi, Stiel, and Thodos Correlation: Jossi, et al.
3
developed a
relationship for the viscosity of pure gases and gas mixtures
which includes pure components such as argon, nitrogen,
oxygen, carbon dioxide, sulfur dioxide, methane, ethane, pro-
pane, butane, and pentane. This "residual viscosity" relation-
ship can be used to predict gas viscosity using the "reduced"
density at a specific temperature and pressure, as well as the
molecular weight. The critical properties of the gas (specifi-

SPE 75721
Simplified Correlations for Hydrocarbon Gas Viscosity and Gas Density Validation
and Correlation of Behavior Using a Large-Scale Database
F.E. Londono, R.A. Archer, and T. A. Blasingame, Texas A&M U.
2 F.E. Londono, R.A. Archer, and T. A. Blasingame SPE 75721
cally the critical temperature, critical pressure, and critical
density) are also required.

Our presumption is that the Jossi, et al. correlation (or at least
a similar type of formulation) can be used for the prediction of
viscosity for pure hydrocarbon gases and hydrocarbon gas
mixtures. We will note that this correlation is rarely used for
hydrocarbon gases (because an estimate of the critical density
is required) however; we will consider the formulation
given by Jossi, et al. as a possible model for the correlation of
hydrocarbon gas viscosity behavior.

The "original" Jossi, et al. correlation proposed for gas vis-
cosity is given by:
) ( 10 ) (
4
1
4
JST , r g
f =
+

.........................(1)

where:

4 3
2
0.0093324 0.040758
0.058533 0.023364 0.1023 ) (
JST , r JST , r
JST , r
JST , r JST , r
f
+
+ + =
......................................................................................(2)
3
2
2
1
6
1
c w
c
p M
T
= .............................................................(3)

and,
r,JST
= /
c
, JST Reduced density, dimensionless
= Density at temperature and pressure, g/cc
c
= Density at the critical point, g/cc
T
c
= Critical temperature, deg K
p
c
= Critical pressure, atm
M
w
= Molecular weight, lb/lb-mole
g
= Gas viscosity, cp
*
= Gas viscosity at "low" pressure, cp
The Jossi, et al. correlation is shown in Figs. 1 and 2. Jossi, et
al.
3
reported approximately 4 percent average absolute error
and also stated that this correlation should only be applied for
values of reduced density (
r
) below 2.0. The behavior of the
"residual" gas viscosity function is shown in Figs. 1 and 2.

Lee, Gonzalez, and Eakin, Correlation: The Lee, et al.
5
cor-
relation evolved from existing work in the estimation of
hydrocarbon gas viscosity using temperature, gas density at a
specific temperature and pressure, and the molecular weight of
the gas. This correlation is given by:
) exp( 10
4 Y
X K
g

= ................................................(4)

where:
T M
T M
K
w
w
+ +
+
=
19.26 209.2
) 0.01607 (9.379

1.5
.................................(5)

w
M
T
.
. X 0.01009
4 986
448 3 +
+ = ...........................(6)
X Y 2224 . 0 447 . 2 = ................................................................................................
and,
M
w
= Molecular weight of gas mixture, lb/lb-mole
T = Temperature, deg R
g
= Gas viscosity at temperature and pressure, cp
Lee, et al.
4
reported 2 percent average absolute error (low
pressures) and 4 percent average absolute error (high pres-
sures) for hydrocarbon gases where the specific gravity is be-
low 1.0. For gases of specific gravity above 1.0 this relation is
purported to be "less accurate."

The range of pressures used by Lee, et al.
5
in the development
of this correlation is between 100 and 8,000 psia and the tem-
perature range is between 100 and 340 deg F. This correlation
can also be used for samples which contain carbon dioxide
(in particular for carbon dioxide concentrations up to 3.2 mole
percent). Fig. 3 shows the behavior of the Gonzalez, et al.
5
data (natural gas sample 3) compared to the Lee, et al.
4
hydro-
carbon gas viscosity correlation.

Hydrocarbon Gas Density

A practical issue pertinent to all density-based gas viscosity
models is that an estimate of gas density must be known. Al-
though there are many equation of state (EOS) correlations for
gas density (or more specifically, the gas z-factor) we found
that these EOS models do not reproduce the measured gas
densities in our database to a satisfactory accuracy. This ob-
servation led us towards an effort to "tune" the existing models
(refs. 6-8) for the z-factor using the data in our database.

For reference, the definition of gas density for real gases is
given by:
RT
w
M
z
p
.37 62
1
= ( in g/cc)...................................... (8)

where:
p = Pressure, psia
M
w
= Molecular weight, lb/lb-mole
z = z-factor, dimensionless
T = Temperature, deg R
R = Universal gas constant, 10.732 (psia cu ft)/(lb-mole
deg R)
62.37 = Conversion constant: 1 g/cc = 62.37 lbm/ft
3

The real gas z-factor is presented as an explicit function of the
pseudoreduced pressure and temperature as predicted by the
"Law of Corresponding States"
9
(see Figs. 4-6, where we use
the data of Poettmann and Carpenter
10
). It is important to note
that EOS models are implicit in terms of the z-factor, which
means that the z-factor is solved as a root of the EOS. This
must be considered in the regression process the regression
formulation must include the solution of the "model" z-factor
as a root of the EOS.

Dranchuk-Abou-Kassem,
6
Nishiumi-Saito,
7
and Nishiumi
8

provide EOS representations of the real gas z-factor. In parti-
SPE 75721 Simplified Correlations for Hydrocarbon Gas Viscosity and Gas Density 3
Validation and Correlation of Behavior Using a Large-Scale Database
cular, the Dranchuk-Abou-Kassem result is based on a Han-
Starling form of the Benedict-Webb-Rubin equation of state
(EOS) and is considered to be the current standard for the
prediction of gas density.

z-Factor Model: Dranchuk-Abou-Kassem (ref. 6)

The DAK-EOS
6
is given by:

2
11
3
2
2
11 10
5
2
8 7
9
2
8 7
6

5
5
4
4
3
3 2
1
) exp( ) A (1

2

1

r
r
r
r
r
r
r
r
r
r
r
r r r
r
A
T
A
T
A
T
A
A
T
A
T
A
A
T
A
T
A
T
A
T
A
A z
+ +
+ + +
+ + + + + =
.....(9)

where:
T
r
= Reduced temperature, dimensionless
r
=
/
c
, Reduced density, dimensionless
c
= Critical density, g/cc (z
c
=0.27)
We must note that the definition of critical density is a matter
of some debate in particular, how is critical density esti-
mated when this is also a property of the fluid? As such, we
use the "definition" of
c
as given by Dranchuk and Abou-
Kassem
6
:
0.27) (where = =
c c c
z
r
zT
r
p
z

and the "original" parameters given by Dranchuk and Abou-
Kassem (ref. 6) for hydrocarbon gases are:
A
1
= 0.3265 A
7
=-0.7361
A
2
=-1.0700

A
8
= 0.1844
A
3
=-0.5339

A
9
= 0.1056
A
4
= 0.01569

A
10
= 0.6134
A
5
=-0.05165 A
11
= 0.7210
A
6
= 0.5475.............................................................(10)
The Dranchuk-Abou-Kassem (DAK-EOS) is compared to the
data of Poettmann and Carpenter
10
in Figs. 7-9. We note that
the "original" DAK-EOS agrees quite well with the data
trends, and would, in the absence of data to the contrary, seem
to be adequate for most engineering applications. However,
we would like to extend the range of this relation as well as
provide a more statistically sound correlation of the EOS (i.e.,
add more data to the regression process).

z-Factor Model: Nishiumi-Saito (ref. 7)

The Nishiumi-Saito model (NS-EOS)
7
adds a few more terms
to the original Dranchuk-Abou-Kassem expression, and is
given by:
2
5
5
4
4
3
3 2
1
1
r
r r r
r T
A
T
A
T
A
T
A
A z
+ =

2
24
10
5
9
2
8 7
6 r
r r r
r T
A
T
A
T
A
T
A
A
+

5
24
10
5
9
2
8 7
11 r
r r r
r T
A
T
A
T
A
T
A
A
+ + + +

) exp( ) 1 (
2
15
2
15
2
18
14
9
13
3
12
r r r
r r r
A A
T
A
T
A
T
A
+
+ + +
.................................................................................... (11)

where:
T
r
r
=
/
c
, Reduced density, dimensionless
c
= Critical density, g/cc (z
c
=0.27)
Our perspective in utilizing the NS-EOS is that this relation is
purported to provide better performance in the vicinity of the
critical isotherm which is traditionally a region where the
DAK-EOS has been shown to give a weak performance. We
will compare the performance of the DAK and NS-EOS
relations in detail once these relations are regressed using our
gas density (z-factor) database.

Correlation of Hydrocarbon Gas Viscosity
In this section we provide comparisons and optimizations of
existing correlations for hydrocarbon gas viscosity (refs. 3 and
4), as well as a new correlation for hydrocarbon gas viscosity
that is implicitly defined in terms of gas density and tempera-
ture. Our approach is to use an extensive database of gas vis-
cosity and gas density data, derived from a variety of literature
sources (refs. 11-15). This database contains 2494 points ta-
ken from pure component data and 3155 points taken from
mixture data. More data were available in the literature
however, the data points chosen for this study satisfy the fol-
lowing criteria:
l Temperature is greater than 32 deg F.
l Density measurement is available for each measure-
ment of gas viscosity. (This criterion was not ap-
plied for selection of points for the correlation of
gas viscosity at one atmosphere.)
l Gas composition must be representative of a natural
gas (e.g., data for binary mixtures containing decane
were excluded (ref. 4)).
l Liquid or liquid-like (i.e., unusually high) viscosi-
ties were excluded from consideration.

Gas Viscosity: Jossi, et al. (ref. 3)

In this section we test the performance of the Jossi, et al.
model for gas viscosity against viscosity values from our
database. We then propose a "refitted" form of the Jossi, et al.
model where the coefficients of the original model were ad-
justed using regression to better match the viscosity values
provided in our database. Fig. 10 shows the results of the
original Jossi, et al. model when applied to our database we
note that there are significant departures from the 45 degree
straight line (conformance to this straight line would indicate
perfect agreement between the measured and calculated gas
viscosity values). There are 2494 pure component data points
given on this plot, where this data match has an average
absolute error of 5.26 percent in the prediction of gas viscos-
ity. We wish to note that our database does include gas vis-
cosity values measured at reduced density values greater than
2.0. We note that such points were not included in the original
study by Jossi, et al.

In an attempt to minimize the error between our data and the
Jossi, et al. model, we refitted the coefficients of the Jossi, et
al. model using non-linear regression techniques. The generic
form of the Jossi, et al. model is written in the form:
) ( 10 ) (
4
1
4
JST , r g
f =
+

.......................(12)

where:

4
3 2

5

4 3 2 1
) (
JST , r
JST , r JST , r
JST , r JST , r
f
f f f f f
+
+ + + =

....................................................................................(13)
3 2
1
e
c
e
w
e
c
p M
T
= .........................................................(14)

The "optimized" coefficients obtained from the "refitting" the
Jossi, et al. model are:

f
1
= 1.03671E-01 f
2
= 1.31243E-01
f
3
= 1.71893E-02 f
4
= -3.12987E-02
f
5
= 8.84909E-03

e
1
= -1.21699E-01 e
2
= 3.91956E-01
e
3
= -1.50857E-01
................................................................................... (15)

and the variables are defined in the same fashion as the ori-
ginal correlation proposed by Jossi, et al. where the most
important issue is that this correlation is limited to pure com-
ponent data. This means that the critical density is directly
tied to the component no alternate definition is permitted.

The performance of our "optimized" version of the Jossi, et al.
model for gas viscosity is shown in Fig. 11. This plot shows
better conformance to the 45 degree line than the original
Jossi, et al. model shown in Fig. 10. The average absolute
error for the optimized Jossi, et al. model is 4.43 percent (as
compared to 5.26 percent for the original Jossi, et al. model).
For reference, Jossi, et al. reported an average absolute error
of 4 percent when they presented their model (fitted to a
variety of fluids including non-hydrocarbon samples). It is
relevant to note that Jossi, et al. used a relatively small
database of pure component data.
We must also note that our optimization of this model re-
quires the gas viscosity at one atmosphere (
) in this case
we used an independent correlation for
based on the rele-

vant data from our database. This correlation for
is an inde-
pendent development and is discussed in a later section.

Gas Viscosity: Lee, et al. (ref. 4)

The Lee, et al. model for gas viscosity was utilized in a similar
manner to the Jossi, et al. model i.e. the performance of the
original model was first assessed using our database, and then
the coefficients of this relation were optimized using the data-
base of gas viscosity. We note that in this work we have
utilized data for both pure components and gas mixtures.

Fig. 12 shows the performance of the Lee, et al. model on
4909 points from our database. This figure shows that gas vis-
cosity is under predicted by the Lee, et al. model at the higher
end of the gas viscosity scale. The average absolute error
associated with the comparison of this model with our
viscosity database is 3.34 percent.

The coefficients of the Lee, et al. model were then optimized
using the gas viscosity database in order to improve the per-
formance of the model. These results are shown in Fig. 13.
For the optimization the Lee, et al. relation, the correlation
model was cast in the following form:
) exp( 10
4 Y
X K
g

= ..................................................(16)

where:
T M k k
T M k k
K
w
k
w
+ +
+
=

5 4
2 1

) (

3
..................................................(17)

w
M x
T
x
x X
3
2
1
+
+ = ...............................................(18)

X y y Y
2 1
= .................................................................(19)

The optimized coefficients for this model are:
k
1
= 1.67175E+01 k
2
= 4.19188E-02
k
3
= 1.40256E+00 k
4
= 2.12209E+02
k
5
= 1.81349E+01
x
1
= 2.12574E+00 x
2
= 2.06371E+03
x
3
= 1.19260E-02
y
1
= 1.09809E+00 y
2
= -3.92851E-02
....................................................................................(20)

The average absolute error for this "optimized" model is 2.29
percent. Lee, et al. reported average absolute errors of 2 to 4
percent for their original model where we recall that the
original Lee, et al. correlations were generated using a less
comprehensive database.

Gas Viscosity: Proposed "Implicit" Model
Our correlation work based on a "non-parametric" regression
algorithm
16
shows that gas viscosity is primarily a function of
the following variables:
l Gas viscosity at 1 atm,
l Gas density, and
l Temperature.

We found that pressure and molecular weight could be dis-
carded as explicit variables for this model (these are included
implicitly in the gas density function). We then developed a
new gas viscosity model, which is simply a generic expansion
of the Jossi, et al. model using additional temperature and den-
sity dependent terms.

The relationship between the residual viscosity function (i.e.,
g
-
1atm
) and the gas density appears to be univariate, as
shown in Fig. 14. A log-log plot of the residual viscosity data
from our database shows significant scatter at low densities
where this behavior reveals a strong dependence of gas viscos-
ity on temperature at low densities (see Fig. 15).

By observation we found that the "uncorrelated" distribution
of data formed in the low-density range is directly related to
temperature. We propose a rational polynomial model in
terms of gas density with temperature-dependent coefficients
and used nonlinear regression to fit our proposed model to
temperature and gas density data. This model is given as:
) (
1
f
atm g
+ = ..................................................... (21)
3 2
3 2
) (

h g f e
d c b a
f
+ + +
+ + +
= ......................................... (22)

2
2 1 0
T a T a a a + + = ................................................... (23)
2
2 1 0
T b T b b b + + = .................................................... (24)
2
2 1 0
T c T c c c + + = ..................................................... (25)
2
2 1 0
T d T d d d + + = ................................................... (26)
2
2 1 0
T e T e e e + + = ..................................................... (27)
2
2 1 0
T f T f f f + + = ................................................... (28)
2
2 1 0
T g T g g g + + = .................................................. (29)
2
2 1 0
T h T h h h + + = .................................................... (30)

The numerical values for the parameters of our proposed "im-
plicit" model for gas viscosity (Eqs. 21 to 30) are given as fol-
lows:
a
0
= 9.53363E-01 a
1
= -1.07384E+00

a
2
= 1.31729E-03
b
0
= -9.71028E-01 b
1
= 1.12077E+01
b
2
= 9.01300E-02
c
0
= 1.01803E+00 c
1
= 4.98986E+00
c
2
= 3.02737E-01
d
0
= -9.90531E-01 d
1
= 4.17585E+00

d
2
= -6.36620E-01
e
0
= 1.00000E+00 e
1
= -3.19646E+00
e
2
= 3.90961E+00
f
0
= -1.00364E+00 f
1
= -1.81633E-01

f
2
= -7.79089E+00
g
0
= 9.98080E-01 g
1
= -1.62108E+00
g
2
= 6.34836E-04
h
0
= -1.00103E+00 h
1
= 6.76875E-01
h
2
= 4.62481E+00
................................................................................ (31)

A total of 4909 points were used in the regression calculation
of these parameters (2494 pure component data and 2415 gas
mixture data). The performance of the model is shown in Fig.
16. We note excellent agreement with the 45 degree straight-
line trend. The average absolute error for this model as com-
pared to our database is 3.05 percent. We also note there are
non-hydrocarbon components such as carbon dioxide (0.19 to
3.20 percent), nitrogen (0.04 to 15.80 percent) and helium
(0.03 to 0.80 percent) present in some of the gas mixtures used
to develop these correlations.

Gas Viscosity: Hydrocarbon Gas Viscosity at 1 Atmosphere
In order to utilize both new and existing correlations for gas
viscosity, it is imperative that we estimate the viscosity of a
hydrocarbon gas mixture at 1 atm. We propose a new correla-
tion for this purpose where this correlation is given only as
a function of the temperature (in deg R) and the gas specific
gravity (as a surrogate for molecular weight of the mixture).
The generic form of this relation is given by:
+ + +
+ + +
=
) ln( ) ln( ) ln( ) ln( 1
) ln( ) ln( ) ln( ) ln(
) ln(

3 2 1
3 2 1 0
1
T b T b b
T a T a a a
g g
g g
atm

.......(32)

In this correlation we used 261 data points for the gas viscos-
ity at 1 atm where 135 of these are pure component data and
126 are gas mixture data. This new correlation gives an aver-
age absolute error of 1.36 percent. Fig. 17 illustrates the com-
parison of the calculated gas viscosity at 1 atm and the mea-
sured gas viscosity at 1 atm.

The numerical values of the parameters obtained for the new
gas viscosity model for viscosity at 1 atm model (Eq. 32) are
given by:
a
0
= -6.39821E+00 a
1
= -6.045922E-01
a
2
= 7.49768E-01 a
3
= 1.261051E-01
b
1
= 6.97180E-02 b
2
= -1.013889E-01
b
3
= -2.15294E-02
....................................................................................(33)

Correlation of Hydrocarbon Gas Density (z-factor)
Gas Density: Dranchuk-Abou-Kassem (DAK-EOS)
In this section we present the results of our regression work
where we fitted the DAK-EOS to our gas density database.
This was a multi-step process where we first perform regress-
ion of the DAK-EOS onto the "standard" and pure component
databases. The "standard" database is a tabular rendering of
the Standing and Katz z-factor chart. These data are presumed
to accurately represent an "average" trend according to the
"Law of Corresponding States".

After the "calibration" of the EOS, we can then use the mix-
ture (and pure component) data to establish correlations for
pseudocritical properties (we must correlate pseudocritical
temperature and pressure because we will not be able to esti-
mate these parameters independently recall that we pre-
sume we have only the mixture gravity of the gas, not a full
compositional analysis).

The "first step" regression (EOS to database) is shown on Fig.
18, where we note a very strong correlation. The associated
plots for comparing the models and data for this case are
shown in Figs. 19-21 we also observe a strong correlation
(with only minor errors) near the critical isotherm. Coeffi-
cients for the DAK-EOS obtained from regression (using the
Poettmann-Carpenter
10
"standard" database) are:
A
1
= 3.024696E-01 A
7
=-1.118884E+00
A
2
= -1.046964E+00 A
8
= 3.951957E-01
A
3
= -1.078916E-01 A
9
= 9.313593E-02
A
4
= -7.694186E-01 A
10
= 8.483081E-01
A
5
= 1.965439E-01 A
11
= 7.880011E-01
A
6
= 6.527819E-01
................................................................................. (34a)

The average absolute error associated with this case is 0.412
percent (5960 data points). For reference, the original work
by Dranchuk and Abou-Kassem (ref. 6) was based on a data-
base of 1500 points and yielded an average absolute error of
0.486 percent.

Coefficients for the DAK-EOS regression using the "combin-
ed" database (Poettmann-Carpenter
10
data and pure component
data) are:
A
1
= 2.965749E-01 A
7
=-1.006653E+00
A
2
= -1.032952E+00

A
8
= 3.116857E-01
A
3
= -5.394955E-02

A
9
= 9.506539E-02
A
4
= -7.694000E-01 A
10
= 7.544825E-01
A
5
= 2.183666E-01

A
11
= 7.880000E-01
A
6
= 6.226256E-01
................................................................................. (34b)

For this case we obtained an average absolute error of 0.821
percent (8256 points). We note that this error is higher than
error we obtained using the Poettmann-Carpenter
10
"stan-
dard" database however, this error is (certainly) still accep-
table.

We now pursue the "second step" of this development by
applying the optimized DAK-EOS on our mixture and pure
component database (for the z-factor) as a mechanism to de-
velop relations for estimating the pseudocritical temperature
and pressure.

In Fig. 22 we present the calculated versus measured values of
z-factor for the "mixtures/pure component" calibration. We
note that 6032 data points were used (gas mixtures and pure
component samples), and that we achieved an average abso-
lute error of 3.06 percent for this case. In performing this
regression, we simultaneously defined the new mixture rules
for the DAK-EOS (i.e., correlations of pseudocritical tempera-
ture and pressure as quadratic polynomials as a function of the
gas gravity).

Figs. 23 and 24 present the results of the optimized DAK-EOS
for the z-factor coupled with the optimized quadratic relations
used to model the pseudocritical temperature and pseudocriti-
cal pressure (as a function of gas specific gravity). The

opti-
mized quadratic equations for the pseudocritical temperature
and pressure of a given sample are given in terms of the gas
specific gravity as follows: (DAK-EOS case only)
2
05 9 27 70 89 725
g g pc
. . . p = ..............................(35)
2
01 94 47 549 39 40
g g pc
. . . T + = ...............................(36)

where,
p
pc
= Pseudocritical pressure, psia
T
pc
= Pseudocritical temperature, deg R
g
= Gas specific gravity (air = 1.0)
Eqs. 35 and 36 were calibrated using the DAK-EOS (and the
coefficients for the DAK-EOS were taken from Eq. 34b). For
the optimized DAK-EOS based on our research database, we
note that only the combination of Eqs. 9, 34b, 35, and 36 can
be used be used to estimate the z-factor for gas mixtures.

In summary, we have recalibrated the DAK-EOS against three
databases the Poettmann-Carpenter
10
data (5960 points), an
extended database which includes the Poettmann-Carpenter
10
data and additional pure component data (8256 points), and a
database of pure component and mixture data (6032 points).

In the first two cases we provide new coefficients to replace
the original DAK-EOS (which was similarly defined by the
original authors using pure component data). The average ab-
solute errors for these cases were 0.486 percent and 0.821 per-
cent, respectively.

Lastly, we applied the optimized DAK-EOS based on the
"combined" database (Poettmann-Carpenter
10
data and pure
component data) (i.e., the combination of Eqs. 9 and 34b) for
the case of gas mixtures and developed new models for the
pseudocritical pressure and pseudocritical temperature as
functions of gas gravity. This model resulted in an overall
average absolute error of 3.06 percent for z-factors estimated
using the DAK-EOS (and the quadratic polynomials for T
pc

and p
pc
.

Gas Density: Nishiumi-Saito (NS-EOS)
This section follows a procedure similar to the previous work
which provided new forms of the DAK-EOS. The first step
was to "refit" the coefficients of the NS-EOS model using the
Poettmann-Carpenter
10
database.

The results from this regression are:

A
1
= 2.669857E-01 A
9
=-2.892824E-02
A
2
= 1.048341E+00

A
10
=-1.684037E-02
A
3
= -1.516869E+00

A
11
= 2.120655E+00
A
4
= 4.435926E+00 A
12
=-5.046405E-01
A
5
= -2.407212E+00

A
13
= 1.802678E-01
A
6
= 6.089671E-01

A
14
= 8.563869E-02
A
7
= 5.174665E-01

A
15
= 4.956134E -01
A
8
= 1.296739E+00
................................................................................. (37a)
The average absolute error achieved in this regression was
0.426 percent (5960 points), which is slightly higher than the
DAK-EOS result for the same case (0.412 percent).

The refitting procedure was also performed on the extended
database of the Poettmann-Carpenter
10
data and pure com-
ponent data (8132 points). We note that we used fewer data
for the regression as compared to the same case for the DAK-
EOS (8256 points) we found it necessary to delete certain
extreme points in this regression, particularly values near the
critical isotherm. The regression coefficients for this case are:
A
1
= 4.645095E-01 A
9
=-1.941089E-02
A
2
= 1.627089E+00

A
10
=-4.314707E-03
A
3
= -9.830729E-01

A
11
= 2.789035E-01
A
4
= 5.954591E-01 A
12
= 7.277907E-01
A
5
= 6.183499E-01

A
13
=-3.207280E-01
A
6
= 4.109793E-01

A
14
= 1.756311E -01
A
7
= 8.148481E-02

A
15
= 7.905733E -01
A
8
= 3.541591E-01
................................................................................. (37b)
The average absolute error for this model was 0.733 percent,
which is somewhat better than the DAK-EOS result for the
same case (0.821 percent).

In a similar manner to the DAK-EOS case, we also considered
gas mixtures by developing new relations for the pseudo-
critical pressure and temperature for use with the NS-EOS.
The results for this case are:
2
51 56 09 81 81 621
g g pc
. . . p + = ......................(38)
2
14 93 86 542 91 46
g g pc
. . . T + = ........................(39)

The performance for the NS-EOS (using the coefficients from
Eqs. 37a and 37b) is shown in Figs. 25 to 27 (NS-EOS case
only). The results for the "mixture" case are shown Fig. 28,
and we note that this version of the NS-EOS has an average
absolute error of 2.55 percent (with a database of 5118 points)
and is uniquely defined by Eqs. 11, 37b, 38, and 39.

We note that we again used fewer data in this regression than
the corresponding case for the DAK-EOS (5118 points for the
NS-EOS and 6032 points for the DAK-EOS). This was
necessary due to poor regression performance of the T
pc
and
p
pc
parameters and, as before, we removed extreme values
those near the critical isotherm, high pressure/high tem-
perature data, and cases of very high molecular weight. We
appreciate that this issue may cause concerns however,
based on our procedures and vigilance in the regression pro-
cess, we remain confident that the T
pc
and p
pc
correlations for
this case (i.e., NS-EOS) are both accurate and robust.

Conclusions
The following conclusions are derived from this work:
l The new correlations presented in this work for gas
viscosity, z-factor, and gas viscosity at 1 atm are ap-
propriate for applications in petroleum engineering.
l The original Jossi, et al.
3
and Lee, et al.
4
correlations
for gas viscosity appear to yield acceptable behavior
compared to our database, the average absolute errors
(AAE) for these correlations are as follows:
Jossi, et al. original:
3
AAE = 5.26 percent
Lee, et al. original:
4
AAE = 3.34 percent
However, the "refits" of these correlations (using our
research database) exhibit significantly better repre-
sentations of the data:
Jossi, et al. "refit:" AAE = 4.43 percent
Lee, et al. "refit:" AAE = 2.29 percent
For reference, the Jossi, et al. correlation was fit us-
ing pure component data only (2494 points) and
can only be applied to pure component data (this is a
requirement of the Jossi, et al. formulation). The
Lee, et al. correlation was fit using both pure compo-
nent and gas mixture data (4909 points), and should
be considered appropriate for general applications.
l Our new "implicit" viscosity correlation (given as a
function of density) works well for pure gases and for
gas mixtures over a wide range of temperatures, pres-
sures, and molecular weights. The average absolute
error for the new "implicit" viscosity correlation is
3.05 percent for our combined database of pure com-
ponent and natural gas mixture data (4909 total
points).
l Our new correlation for gas viscosity at 1 atm gave
an average absolute error of 1.36 percent based on
261 data points (135 pure component data and 126
gas mixture data).
l Although carbon dioxide, nitrogen, and helium were
present in some of the gas mixtures, the new gas
viscosity correlations match our research database
very well and, by extension, these correlations
should work well (without correction) for practical
applications where relatively small amounts of non-
hydrocarbon impurities are present.
l The original work by Dranchuk and Abou-Kassem
(DAK-EOS) for the implicit correlation of the real
gas z-factor used 1500 data points and gave an aver-
age absolute error of 0.486 percent.
6
Refitting the
DAK-EOS to our research database we considered
two cases the "standard" database given by Poett-
mann and Carpenter
10
(5960 points) and the "standard
and pure component" database (the Poettmann and
Carpenter data combined with the pure component
data) (8256 points).
The average absolute errors (AAE) for the DAK-EOS
correlations are:
DAK-EOS "standard" AAE = 0.412 percent
DAK-EOS "standard/pure" AAE = 0.821 percent
We performed a similar effort with the Nishiumi and
Saito

EOS
7
(NS-EOS) using the same databases as
for the DAK-EOS and obtained the following results:

NS-EOS "standard" AAE = 0.426 percent
NS-EOS "standard/pure" AAE = 0.733 percent

l For the case of gas mixture densities, we developed
quadratic formulations to represent the pseudocritical
temperature and pressure as functions of the gas
specific gravity. A combined database of pure com-
ponent and gas mixture data was used in this optimi-
zation.

Using the "optimized" DAK-EOS as a basis, we ob-
tained an average absolute error of 3.06 percent
(6032 data points) for the gas mixture correlation.
Proceeding in a similar fashion using the "optimized"
NS-EOS, we obtained an average absolute error of
2.55 percent (5118 data points).

Recommendations and Future Work
1. Further work should include investigations of the ex-
plicit effects of non-hydrocarbon components such as
water, nitrogen, carbon dioxide, and hydrogen sulfide
on both gas viscosity and gas density (i.e., the gas z-
factor).

2. This work could be extended to consider density and
viscosity behavior of rich gas condensate and volatile
oil fluids however, we are skeptical that any sort
of "universal" viscosity relation can be developed.

Nomenclature
AAE = Absolute error, percent
p = Pressure, psia
p
c
= Critical pressure, atm
p
pc
= Pseudocritical pressure, psia
p
pr
= Pseudoreduced pressure, dimensionless
p
r
= Reduced pressure, dimensionless
M
w
= Molecular weight, lbm/lb-mole
T = Temperature, deg F
T
c
= Critical temperature, deg K
T
pc
= Pseudocritical temperature, deg R
T
pr
= Pseudoreduced temperature, dimensionless
T
r
R = Universal gas constant, 10.732 (psia cu ft)/(lb-
mole deg R)
= Density, g/cc
r
= Reduced density, dimensionless
1atm
= Gas viscosity at 1 atm, cp
*
= Gas viscosity at low pressures used by Jossi, et
al.
4
, cp
g
= Gas viscosity, cp
g
= Gas specific gravity (air=1.0), dimensionless
yN2
,
CO2
,
H2S = Mole fraction of the non-hydrocarbon
component (fraction)

Subscripts
c = critical value
pc = pseudocritical value
r = reduced variable
pr = pseudoreduced variable

Acknowledgements
The authors wish to acknowledge the Department of Petro-
leum Engineering at Texas A&M University for the use of
computer and reference services.

References
1. Huber, M. L: Physical and Chemical Properties Division,
National Institute of Standards and Technology, Gaithers-
burg, MD.
2. Carr, N.L. Kobayashi, R., and Burrows, D.B.: "Viscosity
of Hydrocarbon Gases Under Pressure," Trans., AIME
(1954) 201, 264-272.
3. Jossi, J.A., Stiel, L.I., and Thodos G.: "The Viscosity of
Pure Substances in the Dense Gaseous and Liquid
Phases," AIChE Journal (Mar. 1962) Vol. 8, No.1; 59-62.
4. Lee, A.L., Gonzalez, M.H., and Eakin, B.E.: "The Vis-
cosity of Natural Gases," JPT (Aug. 1966) 997-1000;
Trans., AIME (1966) 234.
5. Gonzalez, M.H., Eakin, B.E., and Lee, A.L.: "Viscosity of
Natural Gases," American Petroleum Institute, Mono-
graph on API Research Project 65 (1970).
6. Dranchuk, P.M., and Abou-Kassem, J.H.: "Calculation of
z-Factors for Natural Gases Using Equations of State,"
Journal of Canadian Petroleum (Jul.-Sep. 1975) 14, 34-
36.
7. Nishiumi, H. and Saito, S.: "An Improved Generalized
BWR Equation of State Applicable to Low Reduced Tem-
peratures," Journal of Chemical Engineering of Japan,
Vol. 8, No. 5 (1975) 356-360.
8. Nishiumi, H.: "An Improved Generalized BWR Equation
of State with Three Polar Parameters Applicable to Polar
Substances," Journal of Chemical Engineering of Japan,
Vol. 13, No. 3 (1980) 178-183.
9. Standing, M.B., Katz, D.L.: "Density of Natural Gases,"
Trans., AIME (1942) 146, 140.
10. Poettmann, H.F., and Carpenter, P.G.: "The Multiphase
Flow of Gas, Oil, and Water Through Vertical Flow
String with Application to the Design of Gas-lift Instal-
lations," Drilling and Production Practice, (1952) 257-
317.
11. Lee, A.L.: "Viscosity of Light Hydrocarbons," American
Petroleum Institute, Monograph on API Research Project
65 (1965).
12. Diehl, J., Gondouin, M., Houpeurt, A., Neoschil, J.,
Thelliez, M., Verrien, J.P., and Zurawsky, R.: "Viscosity
and Density of Light Paraffins, Nitrogen and Carbon Di-
oxide," CREPS/Geopetrole (1970).
13. Golubev I.F., "Viscosity of Gases and Gas Mixtures, a
Handbook," This paper is a translation from Russian by
the NTIS (National Technical Information Service)
(1959).
14. Stephan, K., and Lucas, K.: "Viscosity of Dense Fluids,"
The Purdue Research Foundation (1979).
15. Setzmann U., and Wagner, W.: "A New Equation of State
and Tables of Thermodynamic Properties for Methane
Covering the Range from the Melting Line to 625 K at
Pressures up to 1000 MPa," J. Phys. Chem. Ref. Data
(1991) Vol. 20, No 6; 1061-1155.
16. Xue, G., Datta-Gupta, A., Valko, P., and Blasingame,
T.A.: "Optimal Transformations for Multiple Regression:
Application to Permeability Estimation from Well Logs,"
SPEFE (June 1997), 85-93.
17. McCain, W. D., Jr.: "The Properties of Petroleum Fluids,"
Second Edition, Penn Well Publishing Co., Tulsa, OK
(1990) 90-146.
18. Brill, J. P. and Beggs, H. D.: "Two-Phase Flow in Pipes,"
University of Tulsa. INTERCOMP Course, The Hague,
(1974).

Figure 1 The "residual" gas viscosity function versus reduc-
ed density for different pure substances of similar
molecular weights (Jossi, et al.
3
).

Figure 2 The "residual viscosity" function versus reduced
density for different pure components note the
effect of temperature at low reduced densities
(Jossi, et al.
3
).

Figure 3 Gas viscosity versus temperature for the Gonzalez,
et al.
5
data (natural gas sample 3) compared to the
Lee, et al.
4
hydrocarbon viscosity correlation.


Figure 4 Real gas z-factor, as attributed to Standing and
Katz,
9
plotted as a function of the pseudoreduced
pressure (data of Poettmann and Carpenter
10
).

Katz,
9
pressure divided by the pseudoreduced tempera-
ture (data of Poettmann and Carpenter
10
).

Katz,
9
density function (data of Poettmann and Carpen-
ter
10
).

Katz,
9
10
)
compared to the original DAK-EOS (coefficients
from Eq. 10).


Katz,
9
10
) compared
to the original DAK-EOS (coefficients from Eq. 10).

Katz,
9
plotted as a function of pseudoreduced den-
sity (data of Poettmann and Carpenter
10
) compared
to the original DAK-EOS (coefficients from Eq. 10).

Figure 10 Jossi, Stiel, and Thodos
4
correlation for hydrocar-
bon gas viscosity tested with our database (Car-
tesian format).

Figure 11 Optimized Jossi, Stiel, and Thodos
4
correlation for
hydrocarbon gas viscosity optimized using our
database (Cartesian format).

Figure 12 Lee, Gonzalez, and Eakin
5
correlation for hydrocar-
bon gas viscosity tested with our database (Car-
tesian format).

Figure 13 Optimized Lee, Gonzalez, and Eakin
5
correlation for
hydrocarbon gas viscosity optimized using our
database (Cartesian format).

Figure 14 Cartesian plot of the residual viscosity versus den-
sity for hydrocarbon gases.

Figure 15 Log-log plot of residual viscosity versus density for
hydrocarbon gases.


Figure 16 Cartesian plot of the calculated versus the mea-
sured viscosity for hydrocarbon gases, the vis-
cosity is calculated using the proposed implicit
model for gas viscosity (in terms of gas density and
temperature).

Figure 17 Cartesian plot of the calculated versus the mea-
sured gas viscosity at 1 atm, the gas viscosity at 1
atm is calculated using the new rational polynomial
model (Eq. 32).

Figure 18 Log-log plot of the calculated versus the measured
z-factor the z-factor is calculated using the op-
timized Dranchuk and Abou-Kassem
6
EOS (coeffi-
cients from Eqs. 34a and 34b).

Katz,
9
10
) com-
pared to the optimized DAK-EOS (coefficients from
Eqs. 34a and 34b).


Katz,
9
10
) compared
to the optimized DAK-EOS (coefficients from Eqs.
34a and 34b).

Katz,
9
10
) compared
34a and 34b).

z-factor, the z-factor is calculated using the opti-
mized DAK-EOS and the new quadratic equations
for the pseudocritical properties (coefficients from
Eqs. 34b, 35, and 36).

Figure 23 Pseudocritical pressure behavior predicted from
correlations (including pure component data) the
new correlations for ppc are derived from the opti-
mized DAK-EOS and the optimized NS-EOS.


Figure 24 Pseudocritical temperature behavior predicted from
correlations (including pure component data) the
new correlations for Tpc are derived from the opti-
mized DAK-EOS and the optimized NS-EOS.

Katz,
9
10
) com-
pared to the optimized NS-EOS (coefficients from
Eqs. 37a and 37b).

Katz,
9
10
) compared
to the optimized NS-EOS (coefficients from Eqs.
37a and 37b).

Katz,
9
10
) compared
37a and 37b).


z-factor, the z-factor is calculated using the opti-
mized NS-EOS and the new quadratic equations for
the pseudocritical properties (coefficients from
Eqs. 37b-39).

Fluidos de Reservorio
Parmetros PVT
Parmetros PVT
En los reservorios de gas tenemos una ecuacin de
estado que describe el comportamiento del fluido en
funcin de cambios en la Presin y Temperatura:
PV = ZnRT
En los reservorios de petrleo, lamentablemente no
tenemos una ecuacin de estado que lo represente,
por lo que tenemos que medir estos parmetros en
laboratorio, realizando un anlisis exhaustivo del
comportamiento de los mismos.
Petrleo Subsaturado y Saturado
Muestreo
Es muy importante tener en cuenta que el muestreo de un reservorio de debe ser
realizado al comienzo de la explotacin del mismo (una de las primeras tareas a
realizar) para que la muestra sea representativa.
El tipo de muestreo va a depender del tipo de fluido de reservorio. Tenemos dos
tipos de muestreo:
Muestra de fondo: este tipo de muestreo puede ser realizado cuando la presin de
fluencia (presin de fondo del pozo cuando se encuentra en produccin) est por
encima de la presin de saturacin del fluido. Este tipo de muestreo es utilizado para
petrleos negros o voltiles.
Muestra de superficie y recombinacin: este tipo de muestreo se utiliza para petrleos
negros y voltiles, como as para gas retrgrado y gas hmedo. Se toma una muestra de
gas y una muestra de lquido en las salidas del separador. La recombinacin se realiza
segn la RGP (GOR) correspondiente en ese momento.
Durante el muestreo es muy importante que las condiciones de produccin del
pozo se mantengan constantes a lo largo del mismo (GOR cte, Qprod cte, etc.). Se
debe producir el pozo por un periodo de tiempo para asegurar la limpieza en las
inmediaciones del pozo y luego al menor caudal posible para asegurar una presin
alta de fluencia, que est por encima de la presin de saturacin.
Muestra de Fondo
Muestra de Superficie
*Se debe corregir el GOR de campo informado, debido a que el mismo es informado sobre el lquido de tanque.
Definicin de Parmetros PVT
Rs = Relacin de gas disuelto, es el volumen de gas que va a disolver 1 m3
de petrleo en superficie a una determinada presin y la temperatura de
reservorio.
Unidades: scf/stb, Nm3/m3
Factor de conversin: 5,615 scf/stb = 1 Nm3/m3
Bo = Factor volumtrico de formacin del petrleo, es el volumen de
petrleo ocupado en fondo por 1 m3 de petrleo en superficie, junto con
su gas disuelto, a una determinada presin y temperatura de reservorio.
Unidades: rb/stb, m3/m3
Factor de conversin: 1 rb/stb = 1 m3/m3
Bg = Factor volumtrico de formacin del gas, es el volumen de gas que
ocupa 1 m3 de gas en superficie, a una determinada presin y
temperatura del reservorio.
Unidades: rb/scf, m3/Nm3
Factor de conversin: 0,1781 rb/scf = 1 m3/Nm3
Petrleo Subsaturado
Petrleo Saturado
Bt = Bo + (R Rs)*Bg
Liberacin Flash (Expansin a masa constante)
Liberacin Diferencial
Pasos de un PVT
1. Determinacin composicin de (muestreo en superficie):
Gas de separador
Gas de tnk
Lquido de tnk
2. Expansin Flash para:
Determinacin del Punto de Burbuja (Pb) o Punto de Roco (Pd).
En Petrleos se obtiene la compresibilidad del petrleo en la zona
monofsica (P > Pb): co = -1/Bo*Bo/P
3. Expansin Diferencial para determinar Bo, Rs, Bg, z.
Pasos de un PVT
4. Ensayos de separador:
Expansin Flash a travs de diferentes presiones de Separador para
determinar presin ptima de separacin y corregir los datos de la
expansin diferencial.
1er Flash: Pb, Tres -> Psep_i, Tsep
2do Flash: Psep_i, Tsep -> Pcs, Tcs
Correcciones:
Bo_corregido = Bod * Bofb_sepopt / Bodb
Rs_corregido = Rsd * Rsfb_sepopt / Rsdb
5. Composicin del efluente de las diferentes etapas de la
liberacin diferencial.
6. Medicin de viscosidad y densidad en funcin de la presin.
Correccin de GOR
Si el muestreo fue en superficie, se debe corregir el GOR medido
para realizar una correcta recombinacin:
A. Correccin por lquido de separador:
GOR medido = Gas de sep (no corregido) / Pet de tnk (STD)
GOR sep_nc = Gas de sep (no corregido) / Pet de sep (STD) =
= GOR medido * Pet de tnk (STD) / Pet de sep (STD)
B. Correccin por GE del gas:
GOR sep_nc = Gas de sep (no corregido) / Pet de sep (STD)
GE medido -> Z medido
GE lab -> Z lab
GOR sep = GOR sep_nc * Z lab / Z medido
2. Expansin Flash para la determinacin del
Punto de Burbuja (Pb) o Punto de Roco (Pd).
3. Expansin Diferencial para determinar parmetros
PVT: Bo, Rs y Bg.
4. Expansin Flash a travs de diferentes presiones de
Separador.
5. Composicin del efluente de las diferentes
etapas la liberacin diferencial.
6. Determinacin de curvas de viscosidad y densidad
en funcin de la presin.

PVTBlackOil
Compaa : XXXXXXXXXXXXXX
Yacimiento : XXXXXXXXXXXXXX
Pozo : XXXXXX
pgina-6-
DATOS OBTENIDOS DE LAS PLANILLAS DE TOMA DE MUESTRA
Y CONFIRMADOS POR PERSONAL DE XXXXXXXXXXXXX
Formacin Quintuco
2471.5 - 2475.5 mbbp
Presin Esttica del Reservorio 2334 psi @ 2474 mbbp
Temperatura de la capa productiva 91.5 C @ 2474 mbbp
2.7/8" a 2449 mbbp
Estado del Pozo Cerrado
Profundidad del Packer 2449 mbbp
Profundidad del Tapn 2539 mbbp
Fecha de toma de muestra 04/04/01
Profundidad del muestreo 2430 mbbp
Presin a prof de muestreo 2273 psi
Temperatura a prof del muestreo 90.4 C
Presin esttica de boca 20 psi
Produccin de petroleo diaria
122.4 m
3
/d
Nivel de petroleo 710 mbbp
Porcentaje de agua 0%
Contacto gas-petrleo 710 mbbp
Muestra 1 2
N de Botelln 14308 28312
Fecha de toma de muestra 04/04/01 04/04/01
Tiempo de cierre del Pozo 12 hs. 14 hs.
Tiempo de bajada del tomamuestras 60 min. 60 min.
Presin de apertura del tomamuestras 960 psi 940 psi
Temp de transferencia de la muestra 13 C 11 C
Temperatura ambiente 14 C 11 C
cm
3
para elevar la presin a 2400 psi
35.7 36
Presin de transferencia de la muestra 2800 psi 2800 psi
Presin de embarque de la muestra 2500 psi 2500 psi
cm
3
retirados del botelln de lquido
30 30
Presin de burbuja de campo 1140 psi @ 13 C 1135 psi @ 10 C
Presin de apertura del botelln 2735 psia @19.2C
Presin de burbuja del botelln 1180.1psia @ 19.5 C
Mediciones efectuadas en el laboratorio
Caractersticas del pozo
Intervalo productor
Dimetro de Tubing
Condiciones de la toma de muestra (muestra de fondo)

Compaa : XXXXXXXXXXXXX
Yacimiento : XXXXXXXXXXXXX
Pozo : XXXXXX
pgina-7-
[Kg/cm
2
]
abs
[psia] Experim. Ajustada [Kg/cm
2
]
abs
[psia] Experim. Ajustada
212.75 3026.0 821.75 821.74 80.57 1146.0 810.00 810.00
179.00 2546.0 819.16 819.18 78.81 1121.0 807.00 807.00
144.90 2061.0 816.68 816.66 77.76 1106.0 804.00 804.00
107.29 1526.0 813.94 813.95
Pb 82.97 1180.1 812.23 812.23
Presin Lectura de bomba Presin Lectura de bomba
EXPANSIN A MASA CONSTANTE Y 19.5C
RELACIN PRESIN-VOLUMEN DEL FLUIDO DE RESERVORIO
Botelln: XXXXX
Estado bifsico Estado monofsico
802
804
806
808
810
812
814
816
818
820
822
824
50 70 90 110 130 150 170 190 210 230
PRESIN [Kg/cm
2
]
L
e
c
t
u
r
a

d
e

B
o
m
b
a

[
c
m
3
]
P. de Burbuja

Pozo : XXXXXX
pgina-8-
Composicin de la muestra

Pozo : XXXXXX
pgina-9-
COMPONENTE Densidad Peso
[g/cm
3
] Molecul ar
Metano 0.009 0.114 0.344 16.04
Etano 0.029 0.205 0.484 30.07
Propano 0.214 1.025 0.508 44.09
i-Butano 0.166 0.604 0.563 58.12
n-Butano 0.672 2.440 0.584 58.12
i-Pentano 0.703 2.058 0.625 72.15
n-Pentano 1.252 3.662 0.631 72.15
Hexanos 2.285 5.742 0.691 84.00
Heptanos 4.000 8.797 0.729 96.00
Octanos 4.799 9.469 0.752 107.00
Nonanos 4.188 7.307 0.771 121.00
Decanos 4.174 6.577 0.785 134.00
Undecanos 3.809 5.471 0.796 147.00
Dodecanos 3.204 4.202 0.807 161.00
Tridecanos 3.205 3.867 0.819 175.00
Tetradecanos 3.102 3.447 0.830 190.00
Pentadecanos 3.640 3.730 0.840 206.00
Hexadecanos 2.586 2.459 0.847 222.00
Heptadecanos 2.401 2.139 0.855 237.00
Octadecanos 2.785 2.343 0.860 251.00
Nonadecanos 2.497 2.005 0.865 263.00
Eicosanos y Sup 50.279 22.338 0.930 475.20
100.000 100.000
PROPIEDADES MEDIDAS EXPERIMENTALMENTE
Densidad media @ 15.5C [g/cm
3
] 0.8491
Gravedad (API) 35.1
Peso Molecular Medio 215.6
PROPIEDADES CALCULADAS
3
] 0.8491
Gravedad (API) 35.1
PROPIEDADES CALCULADAS DE LA FRACCION C7+
Porcentaj e Molar 84.149
Densidad @ 15.5C [g/cm
3
] 0.865
COMPOSICION MOLECULAR DEL PETROLEO DE TANQUE OBTENIDO A PARTIR DE
UNA SEPARACION FLASH @ P. ATM. Y 20.6 C
[% en peso] [% Mol ar]
Val ores Asi gnados Botell n: 14308
CROM ATOGRAM A

Pozo : XXXXXX
pgina-10-
Gas de Petrleo Valores Asignados
Flash de Flash
[% Molar] [% Molar]
Nitrogeno 1.411 0.622 0.804 28.013
Dixido de Carbono 0.093 0.041 0.809 44.010
Metano 59.548 0.114 26.330 0.344 16.04
Etano 13.136 0.205 5.909 0.484 30.07
Propano 13.244 1.025 6.415 0.508 44.09
i-Butano 2.428 0.604 1.409 0.563 58.12
n-Butano 5.467 2.440 3.775 0.584 58.12
i-Pentano 1.505 2.058 1.814 0.625 72.15
n-Pentano 1.601 3.662 2.753 0.631 72.15
Hexanos 0.961 5.742 3.633 0.691 84.00
Heptanos 0.455 8.797 5.118 0.729 96.00
Octanos 0.109 9.469 5.341 0.752 107.00
Nonanos 0.032 7.307 4.098 0.771 121.00
Decanos 0.010 6.577 3.680 0.785 134.00
Undecanos 5.471 3.058 0.796 147.00
Dodecanos 4.202 2.348 0.807 161.00
Tridecanos 3.867 2.161 0.819 175.00
Tetradecanos 3.447 1.926 0.830 190.00
Pentadecanos 3.730 2.085 0.840 206.00
Hexadecanos 2.459 1.374 0.847 222.00
Heptadecanos 2.139 1.195 0.855 237.00
Octadecanos 2.343 1.309 0.860 251.00
Nonadecanos 2.005 1.120 0.865 263.00
Eicosanos y Sup 22.338 12.485 0.930 475.20
100.000 100.000 100.000
PROPIEDADES MEDIAS
G. Esp. (aire=1) 0.966
3
] 0.8491 0.7850
Peso Molecular Medio 28.0 215.6 130.4
PROP. CALCULADAS DE LA FRACCIN C7+
Porcentaje Molar 0.606 84.149 47.299
3
] 0.865 0.865
Peso Molecular Medio 99.9 237.5 236.7
PARMETROS GLOBALES DE ESTE ENSAYO
Relacin GAS-PETRLEO: 73.4 m
3
/m
3
en condiciones STD
Fraccin molar de gas 0.4411
Fraccin molar de lquido 0.5589 Condiciones del FLASH (no equil.)
Factor de Volumen del Flash 1.169 De 212.9 Kg/cm
abs
F. de Volumen Calculado y 20.6 C
@Pb & T Reserv 1.275 hasta 1.03 Kg/cm
abs
y 20.6 C
COMPOSICION MOLECULAR DEL FLUIDO DE RESERVORIO, A PARTIR DE UNA
SEPARACION FLASH @ P. ATM. Y 20.6 C
COMPONENTE
Botelln: 14308
Fluido de
reservorio
recombinado
[% Molar]
Densidad
[g/cm]
Peso
Molecular

Pozo : XXXXXX
pgina-11-

Expansin a masa constante

Pozo : XXXXXX
pgina-12-
Estado monofsico Estado bifsico
Vol umen relati vo Vol umen relati vo
Experim. Aj ustado Experim. Aj ustado
212.89 0.9853 0.9853 1.323E-04 110.57 1.0099 1.0099
176.97 0.9902 0.9902 1.448E-04 98.87 1.0561 1.0561
Pr 164.10 0.9920 0.9921 1.492E-04 82.67 1.1491 1.1493
159.97 0.9927 0.9927 1.506E-04 71.77 1.2424 1.2422
141.97 0.9954 0.9954 1.567E-04
128.97 0.9975 0.9975 1.611E-04
Pb 113.60 1.0000 1.0000 1.663E-04
Vr = 1.02122 + -2.0727E-04 x P + 1.8055E-07 x P para P>Pb
Vr = 1 + ( 113.60 - P ) / ( 9.1311E-03 x P + 1.7508E+00 x P ) para P<Pb
Donde P = Presin y Vr = Volumen Relativo
EXPANSIN A MASA CONSTANTE Y 91.5 C
RELACIN PRESIN-VOLUMEN DEL FLUIDO DE RESERVORIO
Presin
[Kg/cm
2
]
abs
Presin
[Kg/cm
2
]
abs
Coef. de
Compresib
1/[Kg/cm]
0.950
1.000
1.050
1.100
1.150
1.200
1.250
1.300
50 100 150 200 250
PRESIN [Kg/cm
2
]
abs
V
o
l
.

R
e
l
a
t
i
v
o
P. de Burbuja

Pozo : XXXXXX
pgina-13-

Liberacin Diferencial

Pozo : XXXXXX
pgina-14-
Presin
[Kg/cm
2
]
abs
Rd
[m
3
/m
3
]
Rl
[m
3
/m
3
]
Pb 113.60 75.22 0.00
97.37 66.62 8.60
81.47 58.32 16.91
65.57 49.97 25.26
49.67 41.48 33.74
33.77 32.64 42.59
17.87 22.48 52.74
1.03 0.00 75.22
LIBERACION DIFERENCIAL A 91.5 C
RELACIN GAS-PETRLEO
(Disuelto [Rd] y Liberado [Rl])
0
20
40
60
80
0 20 40 60 80 100 120
PRESIN [Kg/cm
2
]
abs
R
G
P

[
m
3
/
m
3
]
Gas Disuelto
Gas Liberado

Pozo : XXXXXX
pgina-15-
Presin
[Kg/cm
2
]
abs
Factor de Vol umen
(Bo)
Presin
[Kg/cm
2
]
abs
Factor de Vol umen
(Bo)
212.89 1.2664 97.37 1.2637
176.97 1.2727 81.47 1.2434
Pr 164.10 1.2751 65.57 1.2226
159.97 1.2759 49.67 1.2010
141.97 1.2794 33.77 1.1776
128.97 1.2821 17.87 1.1488
Pb 113.60 1.2853 1.03 1.0629
FACTOR DE VOLUMEN DE PETRLEO (Bo)
1.00
1.05
1.10
1.15
1.20
1.25
1.30
0 50 100 150 200 250
PRESIN [Kg/cm
2
]
abs
B
o
P. de Burbuja

Pozo : XXXXXX
pgina-16-
Estado monofsi co Estado bifsi co
Presi n
[Kg/cm
2
]
abs
Densidad
[g/cm
3
]
Presi n
[Kg/cm
2
]
abs
Densidad
[g/cm
3
]
212.89 0.7435 97.37 0.7390
176.97 0.7398 81.47 0.7451
Pr 164.10 0.7384 65.57 0.7516
159.97 0.7379 49.67 0.7586
141.97 0.7359 33.77 0.7662
128.97 0.7344 17.87 0.7754
Pb 113.60 0.7325 1.03 0.7996
DENSIDAD DEL PETRLEO
0.720
0.730
0.740
0.750
0.760
0.770
0.780
0.790
0.800
0.810
0 50 100 150 200 250
PRESIN [Kg/cm
2
]
abs
D
e
n
s
i
d
a
d

[
g
/
c
m
3
]
P. de Burbuja

Pozo : XXXXXX
pgina-17-
Presin
[Kg/cm
2
]
abs
Viscosidad
[cP]
Presin
[Kg/cm
2
]
abs
Viscosidad
[cP]
212.89 1.005 97.37 0.874
176.97 0.939 81.47 0.933
Pr 164.10 0.919 65.57 1.013
159.97 0.910 49.67 1.123
141.97 0.878 33.77 1.252
128.97 0.850 17.87 1.440
Pb 113.60 0.823 1.03 2.088
VISCOSIDAD DEL PETRLEO
0.60
0.80
1.00
1.20
1.40
1.60
1.80
2.00
2.20
0 50 100 150 200 250
PRESIN [Kg/cm
2
]
abs
V
i
s
c
o
s
i
d
a
d

[
c
p
]
P. de Burbuja

Pozo : XXXXXX
pgina-18-
Presi n [Kg/cm]abs
COMPONENTE 97.37 81.47 65.57 49.67 33.77 17.87 1.03
Nitrogeno 3.633 3.031 2.373 1.678 1.063 0.465 0.159
Dixido de Carbono 0.057 0.071 0.079 0.087 0.112 0.138 0.097
Metano 79.841 79.476 78.489 76.451 71.771 60.127 20.580
Etano 7.828 8.384 9.212 10.461 12.686 16.970 18.579
Propano 4.815 5.130 5.669 6.574 8.400 12.847 27.756
i-Butano 0.704 0.739 0.808 0.932 1.200 1.915 5.994
n-Butano 1.476 1.539 1.675 1.929 2.475 3.984 13.951
i-Pentano 0.417 0.424 0.452 0.513 0.645 1.037 4.232
n-Pentano 0.473 0.477 0.505 0.571 0.713 1.141 4.505
Hexanos 0.379 0.367 0.383 0.422 0.509 0.785 2.624
Heptanos 0.232 0.220 0.215 0.246 0.285 0.406 0.949
Octanos 0.088 0.088 0.088 0.087 0.092 0.121 0.378
Nonanos 0.038 0.037 0.036 0.034 0.036 0.045 0.139
Decanos 0.018 0.017 0.016 0.015 0.015 0.019 0.056
100.000 100.000 100.000 100.000 100.000 100.000 100.000
PROPIEDADES MEDIAS
G. Esp. (aire=1) 0.722 0.727 0.737 0.759 0.805 0.926 1.479
Factor de Desviacin ('Z') 0.884 0.895 0.907 0.921 0.936 0.952 0.993
Peso Molecular Medio 20.9 21.1 21.4 22.0 23.4 26.9 42.9
P. Molec de la Fraccin C7+ 102.9 103.0 103.0 102.3 101.8 101.4 102.4
Viscosidad [cp] 0.0150 0.0143 0.0137 0.0129 0.0120 0.0114 0.0099
Factor de Volumen (Bg) 0.01189 0.01438 0.01810 0.02426 0.03627 0.06974 1.25859
P. Calorf. Inf. [Kcal/m
3
] 9,790 9,924 10,150 10,522 11,203 12,845 20,045
P. Calorf. Sup. [Kcal/m
3
] 10,796 10,941 11,186 11,587 12,318 14,080 21,801
COMPOSICIN MOLECULAR DEL EFLUENTE
0.01
0.1
1
10
100
0 20 40 60 80 100
PRESIN [Kg/cm]abs
C
o
m
p
o
s
i
c
i
n
N2
CO2
C1
C2
C3
i-C4
n-C4
i-C5
n-C5
C6
C7
C8
C9
C10

Pozo : XXXXXX
pgina-19-
Valores Asignados
COMPONENTE [% en peso] [% Molar] Densi dad Peso
[g/cm
3
] Molecular
Metano 0.001 0.010 0.344 16.04
Etano 0.010 0.070 0.484 30.07
Propano 0.112 0.522 0.507 44.09
i-Butano 0.108 0.380 0.563 58.12
n-Butano 0.488 1.719 0.584 58.12
i-Pentano 0.595 1.687 0.624 72.15
n-Pentano 1.146 3.251 0.631 72.15
Hexanos 2.308 5.622 0.691 84.00
Heptanos 4.342 9.255 0.729 96.00
Octanos 5.362 10.253 0.752 107.00
Nonanos 4.710 7.964 0.771 121.00
Decanos 4.713 7.198 0.785 134.00
Undecanos 4.338 6.038 0.796 147.00
Dodecanos 3.658 4.649 0.807 161.00
Tridecanos 4.194 4.904 0.818 175.00
Tetradecanos 2.777 2.991 0.830 190.00
Pentadecanos 3.607 3.583 0.840 206.00
Hexadecanos 3.533 3.257 0.847 222.00
Heptadecanos 2.927 2.527 0.855 237.00
Octadecanos 2.892 2.358 0.860 251.00
Nonadecanos 2.927 2.277 0.865 263.00
Eicosanos y Sup 45.251 19.485 0.930 475.20
100.000 100.000
PROPIEDADES MEDIDAS EXPERIMENTALMENTE
3
] 0.8499
Gravedad (API) 35.0
PROP. CALCULADAS A PARTIR DE LOS VALORES ASIGNADOS
3
] 0.8451
Gravedad (API) 35.9
PROPIEDADES CALCULADAS DE LA FRACCION C7+
Porcentaje Molar 86.739
3
] 0.858
COMPOSICIN MOLECULAR DEL PETRLEO DE TANQUE OBTENIDO AL FINAL DE
LA LIBERACIN DIFERENCIAL
CROM ATOG RAM A

Pozo : XXXXXX
pgina-20-
Presin
[Kg/cm
2
]
abs
Factor de desviacin del
gas
1 97.37 0.884
2 81.47 0.895
3 65.57 0.907
4 49.67 0.921
5 33.77 0.936
6 17.87 0.952
7 1.03 0.993
FACTOR DE DESVIACIN DEL GAS ('Z')
0.86
0.88
0.90
0.92
0.94
0.96
0.98
1.00
0 20 40 60 80 100 120
PRESIN [Kg/cm
2
]
abs
F
a
c
t
o
r

d
e

D
e
s
v
i
a
c
i
n

[
Z
]

Pozo : XXXXXX
pgina-21-
Presi n
[Kg/cm
2
]
abs
Factor de vol umen del
gas (Bg)
1 97.37 0.0119
2 81.47 0.0144
3 65.57 0.0181
4 49.67 0.0243
5 33.77 0.0363
6 17.87 0.0697
7 1.033 1.2586
FACTOR DE VOLUMEN DEL GAS
0.000
0.010
0.020
0.030
0.040
0.050
0.060
0.070
0.080
0 20 40 60 80 100 120
PRESIN [Kg/cm
2
]
abs
B
g

Pozo : XXXXXX
pgina-22-
Presi n
[Kg/cm
2
]
abs
Vi scosi dad del gas
(Bg)
1 97.37 0.0150
2 81.47 0.0143
3 65.57 0.0137
4 49.67 0.0129
5 33.77 0.0120
6 17.87 0.0114
7 1.03 0.0099
VISCOSIDAD DEL GAS
0.009
0.010
0.011
0.012
0.013
0.014
0.015
0.016
0 20 40 60 80 100 120
PRESIN [Kg/cm
2
]
abs
V
i
s
c
o
s
i
d
a
d

[
c
p
]

Pozo : XXXXXX
pgina-23-
Valores Asignados
PRODUCTOS CONDENSABLES (L/1000m
3
) Densidad Peso
Presin [Kg/cm
2
] [g/cm
3
] Mol ecul ar
COMPONENTE 97.37 81.47 65.57 49.67 33.77 17.87 1.03
Nitrogeno - - - - - - - 0.803 28.02
D. de Carbono - - - - - - - 0.809 44.01
Metano - - - - - - - 0.300 16.04
Etano - - - - - - - 0.356 30.07
Propano 177.3 188.9 208.8 242.1 309.3 473.1 1022.1 0.507 44.09
i-Butano 30.8 32.3 35.3 40.8 52.4 83.7 262.0 0.563 58.12
n-Butano 62.2 64.8 70.6 81.3 104.3 167.9 587.9 0.584 58.12
i-Pentano 20.4 20.8 22.1 25.1 31.6 50.8 207.2 0.624 72.15
n-Pentano 22.9 23.1 24.5 27.7 34.5 55.2 218.1 0.631 72.15
Hexanos 20.3 19.7 20.5 22.6 27.3 42.1 140.6 0.664 84.00
Heptanos 13.0 12.3 12.0 13.7 15.9 22.7 53.1 0.727 96.00
Octanos 5.3 5.3 5.3 5.3 5.6 7.3 22.9 0.749 107.00
Nonanos 2.5 2.5 2.4 2.3 2.4 3.0 9.3 0.768 121.00
Decanos 1.3 1.2 1.2 1.1 1.1 1.4 4.0 0.782 134.00
356.1 370.9 402.7 461.9 584.4 907.1 2527.3
PRODUCTOS CONDENSABLES ACUMULATIVOS (L/1000m
3
)
Presin [Kg/cm
2
]
COMPONENTE 97.37 81.47 65.57 49.67 33.77 17.87 1.03
Nitrogeno+ - - - - - - -
D. de Carbono+ - - - - - - -
Metano+ - - - - - - -
Etano+ - - - - - - -
Propano+ 356.1 370.9 402.7 461.9 584.4 907.1 2527.3
i-Butano+ 178.7 182.0 193.9 219.8 275.1 434.1 1505.2
n-Butano+ 148.0 149.7 158.6 179.1 222.7 350.3 1243.1
i-Pentano+ 85.7 84.9 88.0 97.8 118.4 182.5 655.2
n-Pentano+ 65.3 64.1 65.8 72.7 86.8 131.7 448.0
Hexanos+ 42.4 41.0 41.4 45.0 52.3 76.5 229.8
Heptanos+ 22.1 21.3 20.9 22.4 25.0 34.4 89.2
Octanos+ 9.1 9.0 8.9 8.7 9.1 11.7 36.2
Nonanos+ 3.8 3.7 3.5 3.4 3.5 4.4 13.3
Decanos+ 1.3 1.2 1.2 1.1 1.1 1.4 4.0
CONTENIDO DE PRODUCTOS CONDENSABLES DEL EFLUENTE

Pozo : XXXXXX
pgina-24-
Presin
[Kg/cm
2
]
abs
Coef. de
Compresib.
1/[Kg/cm
2
]
Gas Disuelto
[m
3
/m
3
]
Factor de
Volumen del
Petrleo
(Bo)
Factor de
Volumen del
Gas
(Bg)
Densidad
[g/cm
3
]
Viscosidad
del Petrleo
[cP]
Viscosidad
del Gas
[cP]
Factor de
Desviacin
del Gas
'Z'
212.89 1.323E-04 1.2664 0.7435 1.005
176.97 1.448E-04 1.2727 0.7398 0.939
Pr 164.10 1.492E-04 1.2751 0.7384 0.919
159.97 1.506E-04 1.2759 0.7379 0.910
141.97 1.567E-04 1.2794 0.7359 0.878
128.97 1.611E-04 1.2821 0.7344 0.850
Pb 113.60 1.663E-04 75.22 1.2853 0.7325 0.823
97.37 66.62 1.2637 0.01189 0.7390 0.874 0.0150 0.884
81.47 58.32 1.2434 0.01438 0.7451 0.933 0.0143 0.895
65.57 49.97 1.2226 0.01810 0.7516 1.013 0.0137 0.907
49.67 41.48 1.2010 0.02426 0.7586 1.123 0.0129 0.921
33.77 32.64 1.1776 0.03627 0.7662 1.252 0.0120 0.936
17.87 22.48 1.1488 0.06974 0.7754 1.440 0.0114 0.952
1.03 0.00 1.0629 1.25859 0.7996 2.088 0.0099 0.993
RESUMEN DE LOS VALORES OBTENIDOS

Pozo : XXXXXXX
pgina-25-
Presin
[Kg/cm
2
]
abs
Gas liberado
(cond reservorio)
[m
3
/m
3
]
Pr 164.10 0.00
Pb 113.60 0.00
97.37 0.08
81.47 0.19
65.57 0.36
49.67 0.64
33.77 1.20
17.87 2.86
VOLUMEN DE GAS LIBERADO EN CONDICIONES DE RESERVORIO
(Relativo al volumen a presin de burbuja)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
0 50 100 150 200 250
Presin [Kg/cm
2
]
abs
G
a
s

l
i
b
e
r
a
d
o

(
c
o
n
d
.

r
e
s
.
)

[
m
3
/
m
3
]
P. de Reservorio

Pozo : XXXXXX
pgina-26-
Estado Monofsico Estado Bifsico
Presi n
[Kg/cm
2
]
abs
Factor de
Reduccin del
Volumen
Presi n
[Kg/cm
2
]
abs
Factor de
Reduccin del
Volumen
212.89 0.9853 97.37 0.9832
176.97 0.9902 81.47 0.9674
Pr 164.10 0.9921 65.57 0.9512
159.97 0.9927 49.67 0.9344
141.97 0.9954 33.77 0.9162
128.97 0.9975 17.87 0.8938
Pb 113.60 1.0000 1.03 0.8270
* Volumen relativo al Volumen a la Presin de Burbuja
FACTOR DE REDUCCIN DEL VOLUMEN DE PETRLEO (Shrinkage factor)*
0.800
0.820
0.840
0.860
0.880
0.900
0.920
0.940
0.960
0.980
1.000
0 50 100 150 200 250
Presin [Kg/cm
2
]
abs
R
e
d
u
c
c
i
n

d
e
l

p
e
t
r
l
e
o
P. de Burbuja

Pozo : XXXXXX
pgina-27-
Rsfb: Gas disuelto del ensayo flash
Rl: Gas liberado en el ensayo diferencial
Presi n Rs=Rsfb - Rl * Bofb / Bodb
[Kg/cm
2
]
abs
16 [Kg/cm]abs 12 [Kg/cm]abs 8 [Kg/cm]abs 4 [Kg/cm]abs
113.60 67.92 67.02 66.25 66.82
97.37 59.55 58.65 57.88 58.45
81.47 51.47 50.58 49.81 50.37
65.57 43.35 42.45 41.68 42.25
49.67 35.09 34.20 33.43 34.00
33.77 26.49 25.59 24.82 25.39
17.87 16.61 15.72 14.95 15.52
1.03 0.00 0.00 0.00 0.00
AJUSTE DE LA RELACIN GAS PETRLEO A LAS CONDICIONES DE
SEPARADOR
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
0 50 100 150
Presin [Kg/cm
2
]
abs
R
G
P

Pozo : XXXXXX
pgina-28-
Bod: Factor de volumen del ensayo diferencial
Bofb: Factor de volumen del ensayo flash
Bodb: Factor de volumen del punto de burbuja del ensayo diferencial
Presin Bo=Bod * Bofb/Bodb
[Kg/cm
2
]
abs
113.60 1.2505 1.2468 1.2435 1.2459
97.37 1.2293 1.2256 1.2224 1.2248
81.47 1.2096 1.2060 1.2028 1.2051
65.57 1.1894 1.1858 1.1827 1.1850
49.67 1.1683 1.1648 1.1618 1.1640
33.77 1.1456 1.1421 1.1392 1.1414
17.87 1.1175 1.1142 1.1113 1.1134
1.03 1.0629 1.0629 1.0629 1.0629
AJUSTE DEL FACTOR DE VOLUMEN A LAS CONDICIONES DE SEPARADOR
1.05
1.10
1.15
1.20
1.25
1.30
0 50 100 150
Presin [Kg/cm
2
]
abs
B
o

Pozo : XXXXXX
pgina-29-

Ensayos de separacin
Pozo : XXXXXX
pgina-30-
Presin [Kg/cm
2
]
abs
COMPONENTE 16.00 12.00 8.00 4.00
Nitrogeno 2.186 2.023 1.850 1.646
Dixido de Carbono 0.097 0.101 0.103 0.101
Metano 78.926 76.565 73.178 67.699
Etano 10.443 11.351 12.344 13.195
Propano 5.877 6.917 8.435 10.747
i-Butano 0.605 0.741 0.976 1.469
n-Butano 1.173 1.447 1.941 3.073
i-Pentano 0.243 0.301 0.414 0.710
n-Pentano 0.281 0.347 0.476 0.831
Hexanos 0.100 0.122 0.168 0.308
Heptanos 0.049 0.060 0.082 0.154
Octanos 0.015 0.018 0.025 0.048
Nonanos 0.004 0.005 0.007 0.013
Decanos 0.001 0.002 0.002 0.004
100.000 100.000 100.000 100.000
PROPIEDADES MEDIAS
G. Esp. (aire=1) 0.711 0.734 0.770 0.838
Factor de Desviacin ('Z') Calc. 0.952 0.962 0.972 0.984
Peso Molecular Medio 20.6 21.3 22.3 24.3
P. Molecular de la Fraccin C7+ 100.8 100.7 100.7 100.7
P. Calorf. Inf. [Kcal/m
3
] 9,823 10,144 10,633 11,543
P. Calorf. Sup. [Kcal/m
3
] 10,835 11,180 11,705 12,680
ENSAYOS DE SEPARACIN FLASH A 25.0 C
ANLISIS DE LOS GASES DE SEPARADOR
0.01
0.1
1
10
100
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Presin [Kg/cm]abs
C
o
m
p
o
s
i
c
i
n
N2
CO2
C1
C2
C3
i-C4
n-C4
i-C5
n-C5
C6
C7
C8
C9
C10

Pozo : XXXXXX
pgina-31-
RELACIN GAS-PETRLEO
(De Separador, Tanque y Total)
PRESION DE RELACION GAS-PETRLEO [m
3
/m
3
]
SEPARADOR
[Kg/cm
2
]
abs
SEPARADOR TANQUE TOTAL
16.00 45.05 22.87 67.92
12.00 49.31 17.71 67.02
8.00 54.53 11.72 66.25
4.00 62.18 4.64 66.82
0
10
20
30
40
50
60
70
80
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
PRESIN [Kg/cm
2
]
abs
R
G
P

[
m
3
/
m
3
]
TOTAL
SEPARADOR
TANQUE

Pozo : XXXXXX
pgina-32-
FACTOR DE VOLUMEN DE PETRLEO
PRESION
[Kg/cm
2
]
abs
FACTOR DE VOLUMEN
Bo [m
3
/m
3
]
1 16.00 1.2505
2 12.00 1.2468
3 8.00 1.2435
4 4.00 1.2459
5
Bo = (Vol de Petrl eo a Pb y T de Reservorio) / (Vol Petrl eo de TK STD)
1.20
1.22
1.24
1.26
1.28
1.30
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
PRESIN [Kg/cm
2
]
abs
B
o

[

m
3
/
m
3
]

Pozo : XXXXXX
pgina-33-
PRESION
[Kg/cm
2
]
abs
DENSIDAD
[g/cm
3
]
GRAVEDAD API
1 16.00 0.8426 36.43
2 12.00 0.8419 36.58
3 8.00 0.8412 36.71
4 4.00 0.8417 36.61
5
DENSIDAD DEL PETRLEO DE TANQUE
33.0
34.0
35.0
36.0
37.0
38.0
39.0
40.0
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
PRESIN [Kg/cm
2
]
abs
A
P
I

Pozo : XXXXXX
pgina-34-
Presin de Separador
[Kg/cm
2
]
abs
16.0 12.0 8.0 4.0
Presin de Separador
[psi g]
213 156 99 42
Vol umen l qui do @ Pb &
T Reserv
100.00 100.00 100.00 100.00
Vol umen de Lquido de
Separador (Cond Sep)
85.88 85.02 83.86 81.98
Vol umen de Gas de
Separador STD
3602 3955 4385 4991
Vol umen de Lquido de
Tanque STD
79.97 80.21 80.42 80.26
Vol umen de Gas de
Tanque STD
1829 1421 943 372
G. Esp. del Gas de
Separador (Ai re=1)
0.711 0.734 0.770 0.838
G. Esp. del Gas de
Tanque (Ai re=1)
1.221 1.249 1.273 1.258
Densi dad del Lqui do de
Tanque STD
0.8426 0.8419 0.8412 0.8417
NOTAS
Todos l os val ores se normal i zan con respecto a 100 cc de fl ui do de reservori o
en condi ci ones de Presin de Burbuj a y Temperatura de Reservori o.
A parti r de la presente tabl a, es posi bl e deri var todos l os Factores de Vol umen
y Relaci ones Gas-Petrleo de i nters.
RESUMEN DE DATOS VOLUMTRICOS

Pozo : XXXXXX
pgina-35-

Caracterizacin del petrleo

Pozo : XXXXXX
pgina-36-
VISCOSIDAD DEL PETRLEO DE TANQUE
EN FUNCIN DE LA TEMPERATURA
Temperatura
[C]
Viscosidad
[cp]
1 69.8 3.06
2 89.5 2.21
Tr 91.5 2.14
3 110.5 1.66
120 110 100 -10 0 10 20 30 40 50 60 70 80 90
1
2
3
4
5
6
7
8
9
10
Temperatura [C]
V
i
s
c
o
s
i
d
a
d

[
c
p
]

CHAPTER 2
PVT ANALYSIS FOR OIL
2.1 INTRODUCTION
In Chapter 1, the importance of PVT analysis was stressed for relating observed
volumes of gas production at the surface to the corresponding underground withdrawal.
For gas this relationship could be obtained merely by determining the single or two
phase Zfactor, and using it in the equation of state. The basic PVT analysis required
to relate surface production to underground withdrawal for an oil reservoir is
necessarily more complex due to the presence, below the bubble point pressure, of
both a liquid oil and free gas phase in the reservoir.
This chapter concentrates on defining the three main parameters required to relate
surface to reservoir volumes, for an oil reservoir, and then proceeds to describe how
these parameters can be determined in the laboratory by controlled experiments
performed on samples of the crude oil.
The subject is approached from a mechanistic point of view in merely recognising that
PVT parameters can be determined as functions of pressure by routine laboratory
analysis. No attempt is made to describe the complex thermodynamic processes
implicit in the determination of these parameters. For a more exhaustive treatment of
the entire subject the reader is referred to the text of Amyx, Bass and Whiting
1
.
Finally, a great deal of attention is paid to the conversion of PVT data, as presented by
the laboratory, to the form required in the field. The former being an absolute set of
measurements while the latter depend upon the manner of surface separation of the
gas and oil.
2.2 DEFINITION OF THE BASIC PVT PARAMETERS
The PressureVolumeTemperature relation for a real gas can be uniquely defined by
the simple equation of state
pV = ZnRT (1.15)
in which the Zfactor, which accounts for the departure from ideal gas behaviour, can
be determined as described in Chapter 1, sec. 5. Using this equation, it is a relatively
simple matter to determine the relationship between surface volumes of gas and
volumes in the reservoir as
sc
sc
T p 1 p
E 35.37 (scf / rcf )
p T Z ZT
= = (1.25)
Unfortunately, no such simple equation of state exists which will describe the PVT
properties of oil. Instead, several, so-called, PVT parameters must be measured by
laboratory analysis of crude oil samples. The parameters can then be used to express
PVT ANALYSIS FOR OIL 44
the relationship between surface and reservoir hydrocarbon volumes, equivalent to
equ. (1.25).
The complexity in relating surface volumes of hydrocarbon production to their
equivalent volumes in the reservoir can be appreciated by considering fig. 2.1.
oil
stock tank
oil
(a)
solution gas
gas
oil
stock tank
oil
(b)
free gas
+
solution gas
SURFACE
RESERVOIR
Fig. 2.1 Production of reservoir hydrocarbons (a) above bubble point pressure,
(b) below bubble point pressure
Above the bubble point only one phase exists in the reservoir the liquid oil. If a
quantity of this undersaturated oil is produced to the surface, gas will separate from the
oil as shown in fig. 2.1(a), the volume of the gas being dependent on the conditions at
which the surface separation is effected. In this case, it is relatively easy to relate the
surface volumes of oil and gas to volumes at reservoir conditions since it is known that
all the produced gas must have been dissolved in the oil in the reservoir.
If the reservoir is below bubble point pressure, as depicted in fig. 2.1(b), the situation is
more complicated. Now there are two hydrocarbon phases in the reservoir, gas
saturated oil and liberated solution gas. During production to the surface, solution gas
will be evolved from the oil phase and the total surface gas production will have two
components; the gas which was free in the reservoir and the gas liberated from the oil
during production. These separate components are indistinguishable at the surface and
the problem is, therefore, how to divide the observed surface gas production into
liberated and dissolved gas volumes in the reservoir.
Below bubble point pressure there is an additional complication in that the liberated
solution gas in the reservoir travels at a different velocity than the liquid oil, when both
are subjected to the same pressure differential. As will be shown in Chapter 4, sec. 2,
the flow velocity of a fluid in a porous medium is inversely proportional to the fluid
viscosity. Typically, gas viscosity in the reservoir is about fifty times smaller than for
liquid oil and consequently, the gas flow velocity is much greater. As a result, it is
normal, when producing from a reservoir in which there is a free gas saturation, that
gas will be produced in disproportionate amounts in comparison to the oil. That is, one
barrel of oil can be produced together with a volume of gas that greatly exceeds the
volume originally dissolved per barrel of oil above bubble point pressure.
Control in relating surface volumes of production to underground withdrawal is gained
by defining the following three PVT parameters which can all be measured by
laboratory experiments performed on samples of the reservoir oil, plus its originally
dissolved gas.
R
s
The solution (or dissolved) gas oil ratio, which is the number of standard
cubic feet of gas which will dissolve in one stock tank barrel of oil when
both are taken down to the reservoir at the prevailing reservoir pressure
and temperature (units scf. gas/stb oil).
B
o
The oil formation volume factor, is the volume in barrels occupied in the
reservoir, at the prevailing pressure and temperature, by one stock tank
barrel of oil plus its dissolved gas (units rb (oil + dissolved gas)/stb oil).
B
g
The gas formation volume factor, which is the volume in barrels that one
standard cubic foot of gas will occupy as free gas in the reservoir at the
prevailing reservoir pressure and temperature (units rb free gas/ssf gas).
Both the standard cubic foot (scf) and the stock tank barrel (stb) referred to in the
above definitions are defined at standard conditions, which in this text are taken as
60F and one atmosphere (14.7 psia). It should also be noted that R
s
and B
o
are both
measured relative to one stock tank barrel of oil, which is the basic unit of volume used
in the field. All three parameters are strictly functions of pressure, as shown in fig. 2.5,
assuming that the reservoir temperature remains constant during depletion.
Precisely how these parameters can be used in relating measured surface volumes to
reservoir volumes is illustrated in figs. 2.2 and 2.3.
pi
p
T
Phase diagram
P
Bo rb ( oil + dissolved gas) / stb
1 stb oil
+
R scf / stb si
solution gas
Fig. 2.2 Application of PVT parameters to relate surface to reservoir hydrocarbon
volumes; above bubble point pressure.
Fig. 2.2 depicts the situation when the reservoir pressure has fallen from its initial value
p
i
to some lower value p, which is still above the bubble point. As shown in the PT
diagram (inset) the only fluid in the reservoir is undersaturated liquid oil. When this oil is
produced to the surface each stock tank barrel will yield, upon gas oil separation, R
si
standard cubic feet of gas. Since the oil is undersaturated with gas, which implies that it
could dissolve more if the latter were available, then the initial value of the solution gas
oil ratio must remain constant at R
si
(scf/stb) until the pressure drops to the bubble
point, when the oil becomes saturated, as shown in fig. 2.5(b).
Figure 2.2 also shows, in accordance with the definitions of B
o
and R
s
, that if R
si
scf of
gas are taken down to the reservoir with one stb of oil, then the gas will totally dissolve
in the oil at the reservoir pressure and temperature to give a volume of B
o
rb of oil plus
dissolved gas. Figure 2.5(a) shows that B
o
increases slightly as the pressure is reduced
from initial to the bubble point pressure. This effect is simply due to liquid expansion
and, since the compressibility of the undersaturated oil in the reservoir is low, the
expansion is relatively small.
Typical values of B
o
and R
s
above the bubble point are indicated in fig. 2.5, these are
the plotted results of the laboratory analysis presented in table 2.4. The initial value of
the oil formation volume factor B
oi
is 1.2417 which increases to 1.2511 at the bubble
point. Thus initially, 1.2417 reservoir barrels of oil plus its dissolved gas will produce
one stb of oil. This is a rather favourable ratio indicating an oil of moderate volatility
and, as would be expected in this case, the initial solution gas oil ratio is also relatively
low at 510 scf/stb. Under less favourable circumstances, for more volatile oils, B
oi
can
have much higher values. For instance, in the Statfjord field in the North Sea, B
oi
is
2.7 rb/stb while the value of R
si
is approximately 3000 scf/stb. Obviously the most
favourable value of B
oi
is as close to unity as possible indicating that the oil contains
hardly any dissolved gas and reservoir volumes are approximately equal to surface
volumes. The small oil fields of Beykan and Kayaky in the east of Turkey provide good
examples of this latter condition having values of B
oi
and R
si
of 1.05 and 20 scf/stb
respectively.
Below the bubble point the situation is more complicated as shown in fig. 2.3.
p
i
p
T
p
rb ( oil + dissolved gas) / stb
1 stb oil
B
o
(R - R ) B
s g rb (free gas) / stb
R = R +
s
(R - R ) scf / stb
s
Fig. 2.3 Application of PVT parameters to relate surface to reservoir hydrocarbon
volumes; below bubble point pressure
In this case each stock tank barrel of oil is produced in conjunction with R scf of gas,
where R (scf/stb) is called the instantaneous or producing gas oil ratio and is measured
daily. As already noted, some of this gas is dissolved in the oil in the reservoir and is
released during production through the separator, while the remainder consists of gas
which is already free in the reservoir. Furthermore, the value of R can greatly exceed
R
si
, the original solution gas oil ratio, since, due to the high velocity of gas flow in
comparison to oil, it is quite normal to produce a disproportionate amount of gas. This
results from an effective stealing of liberated gas from all over the reservoir and its
production through the relatively isolated offtake points, the wells. A typical plot of R as
a function of reservoir pressure is shown as fig. 2.4.
R
scf / stb
R = R
si
510 scf / stb
p
b Reservoir pressure
4000 scf / sfb
Fig. 2.4 Producing gas oil ratio as a function of the average reservoir pressure
for a typical solution gas drive reservoir
The producing gas oil ratio can be split into two components as shown in fig. 2.3, i.e.
R = R
s
+(RR
s
)
The first of these, R
s
scf/stb, when taken down to the reservoir with the one stb of oil,
will dissolve in the oil at the prevailing reservoir pressure to give B
o
rb of oil plus
dissolved gas. The remainder, (R R
s
) scf/stb, when taken down to the reservoir will
occupy a volume
s g s g
scf rb
(R R ) B (R R ) B (rb. free gas / stb)
stb scf

=

(2.1)
and therefore, the total underground withdrawal of hydrocarbons associated with the
production of one stb of oil is
(Underground withdrawal)/stb = B
o
+ (R R
s
) B
g
(rb/stb) (2.2)
The above relationship shows why the gas formation volume factor has the rather
unfortunate units of rb/scf. It is simply to convert gas oil ratios, measured in scf/stb,
directly to rb/stb to be compatible with the units of B
o
. While B
g
is used almost
exclusively in oil reservoir engineering its equivalent in gas reservoir engineering is E,
the gas expansion factor, which was introduced in the previous chapter and has the
units scf/rcf. The relation between B
g
and E is therefore,
g
rb 1
B
scf 5.615E

=

(2.3)
thus B
g
has always very small values; for a typical value of E of, say, 150 scf/rcf the
value of B
g
would be .00119 rb/scf.
B
(rb / stb)
o
1.3
1.2
1.1
1.0
1000 2000 3000 4000
PRESSURE (psia)
p = 3330 psia
b
a
b
c
600
400
200
R
(scf / stb)
s
1000 2000 3000 4000
1000 2000 3000 4000
B
(rb / scf)
g
.010
.008
.006
.004
.002
E
(scf / rcf)
- 200
- 100
- 0
PRESSURE (psia)
PRESSURE (psia)
Fig. 2.5 PVT parameters (B
o
, R
s
and B
g
), as functions of pressure, for the analysis
presented in table 2.4; (p
b
= 3330 psia).
The shapes of the B
o
and R
s
curves below the bubble point, shown in fig. 2.5(a) and
(b), are easily explained. As the pressure declines below p
b
, more and more gas is
liberated from the saturated oil and thus R
s
, which represents the amount of gas
dissolved in a stb at the current reservoir pressure, continually decreases. Similarly,
since each reservoir volume of oil contains a smaller amount of dissolved gas as the
pressure declines, one stb of oil will be obtained from progressively smaller volumes of
reservoir oil and B
o
steadily declines with the pressure.
EXERCISE 2.1 UNDERGROUND WITHDRAWAL
The oil and gas rates, measured at a particular time during the producing life of a
reservoir are, x stb oil/day and y scf gas/day.
1) What is the corresponding underground withdrawal rate in reservoir barrels/day.
2) If the average reservoir pressure at the time the above measurements are made
is 2400 psia, calculate the daily underground withdrawal corresponding to an oil
production of 2500 stb/day and a gas rate of 2.125 MMscf/day. Use the PVT
relationships shown in figs. 2.5(a) (c), which are also listed in table 2.4.
3) If the density of the oil at standard conditions is 52.8 lb/cu.ft and the gas gravity is
0.67 (air = 1) calculate the oil pressure gradient in the reservoir at 2400 psia.
EXERCISE 2.1 SOLUTION
1) The instantaneous or producing gas oil ratio is R = y/x scf/stb. If, at the time the
surface rates are measured, the average reservoir pressure is known, then B
o
, R
s
and B
g
can be determined from the PVT relationships at that particular pressure.
The daily volume of oil plus dissolved gas produced from the reservoir is then
xB
o
rb, and the liberated gas volume removed daily is
s
y
x( R )
x
B
g
rb. Thus the
total underground withdrawal is
o s g
y
x(B ( R )B ) rb/ day
x
+ (2.4)
2) At a reservoir pressure of 2400 psia, the PVT parameters obtained from table 2.4
are:
B
o
= 1.1822 rb/stb; R
s
= 352 scf/stb and B
g
= .0012 rb/ scf
Therefore, evaluating equ. (2.4) for x = 2500 stb/d and y = 2.125 MMscf/d gives a
total underground withdrawal rate of
2500 (1.1822 + (850 352) .0012) = 4450 rb/d
3) The liquid oil gradient in the reservoir can be calculated by applying mass
conservation, as demonstrated in exercise 1.1 for the calculation of the gas
gradient. In the present case the mass balance is
Mass of 1 stb of oil Mass of B
o
rb of oil
+ = +
R
s
scf dissolved gas
at standard conditions
dissolved gas in the
reservoir
or
sc sc
r
o s g
o o
lb lb
1(stb) 5.615 R (scf )
cu.ft cu.ft
lb
B (rb) 5.615
cu.ft

+

=

in which the subscripts sc and r refer to standard conditions and reservoir
conditions, respectively.
The gas density at standard conditions is
sc
=
g
0.0763 (refer equ. (1.30))
= 0.0511 lb/cu ft
Therefore,
sc sc
r
o s g
o
o
( 5.615) (R )
B 5.615
(52.8 5.615) (352 0.0511)
47.37 lb/ cu ft
1.1822 5.615

+
=
+
= =
and the liquid oil gradient is 47.37/144 = 0.329 psi/ft.

2.3 COLLECTION OF FLUID SAMPLES
Samples of the reservoir fluid are usually collected at an early stage in the reservoir's
producing life and dispatched to a laboratory for the full PVT analysis. There are
basically two ways of collecting such samples, either by direct subsurface sampling or
by surface recombination of the oil and gas phases. Whichever technique is used the
same basic problem exists, and that is, to ensure that the proportion of gas to oil in the
composite sample is the same as that existing in the reservoir. Thus, sampling a
reservoir under initial conditions, each stock tank barrel of oil in the sample should be
combined with R
si
standard cubic feet of gas.
a) Subsurface sampling
This is the more direct method of sampling and is illustrated schematically in fig. 2.6.
sample chamber
p
wf
p
i
p
b
pressure
r
Fig. 2.6 Subsurface collection of PVT sample
A special sampling bomb is run in the hole, on wireline, to the reservoir depth and the
sample collected from the subsurface well stream at the prevailing bottom hole
pressure. Either electrically or mechanically operated valves can be closed to trap a
volume of the borehole fluids in the sampling chamber. This method will obviously yield
a representative combined fluid sample providing that the oil is undersaturated with gas
to such a degree that the bottom hole flowing pressure p
wf
at which the sample is
collected, is above the bubble point pressure. In this case a single phase fluid, oil plus
its dissolved gas, is flowing in the wellbore and therefore, a sample of the fluid is bound
to have the oil and gas combined in the correct proportion. Many reservoirs, however,
are initially at bubble point pressure and under these circumstances, irrespective of
how low the producing rate is maintained during sampling, the bottom hole flowing
pressure p
wf
will be less than the bubble point pressure p
b
as depicted in fig. 2.6. In this
case, there will be saturated oil and a free gas phase flowing in the immediate vicinity
of the wellbore, and in the wellbore itself, and consequently, there is no guarantee that
the oil and gas will be collected in the correct volume proportion in the chamber.
In sampling a gas saturated reservoir, two situations can arise depending on the time at
which the sample is collected. If the sample is taken very early in the producing life it is
possible that the fluid flowing into the wellbore is deficient in gas. This is because the
initially liberated gas must build up a certain minimum gas saturation in the reservoir
pores before it will start flowing under an imposed pressure differential. This, socalled,
critical saturation is a phenomenon common to any fluid deposited in the reservoir, not
just gas. The effect on the producing gas oil ratio, immediately below bubble point
pressure, is shown in fig. 2.4 as the small dip in the value of R for a short period after
the pressure has dropped below bubble point. As a result of this mechanism there will
be a period during which the liberated gas remains in the reservoir and the gas oil ratio
measured from a subsurface sample will be too low. Conversely, once the liberated gas
saturation exceeds the critical value, then as shown in fig. 2.4 and discussed
previously, the producing well will effectively steal gas from more remote parts of the
reservoir and the sample is likely to have a disproportionately high gas oil ratio.
The problems associated with sampling an initially saturated oil reservoir, or an
undersaturated reservoir in which the bottom hole flowing pressure has been allowed to
fall below bubble point pressure, can be largely overcome by correct well conditioning
prior to sampling. If the well has already been flowing, it should be produced at a low
stabilized rate for several hours to increase the bottom hole flowing pressure and
thereby re-dissolve some, if not all, of the free gas saturation in the vicinity of the well.
Following this the well is closed in for a reasonable period of time during which the oil
flowing into the wellbore, under an ever increasing average pressure, will hopefully re-
dissolve any of the remaining free gas. If the reservoir was initially at bubble point
pressure, or suspected of being so, the subsurface sample should then be collected
with the well still closed in. If the reservoir is known to be initially undersaturated the
sample can be collected with the well flowing at a very low rate so that the bottom hole
flowing pressure is still above the bubble point. With proper well conditioning a
representative combined sample can usually be obtained.
One of the main drawbacks in the method is that only a small sample of the wellbore
fluids is obtained, the typical sampler having a volume of only a few litres. Therefore,
one of the only ways of checking whether the gas oil ratio is correct is to take several
downhole samples and compare their saturation pressures at ambient temperature on
the well site. This can be done using a mercury injection pump and accurate pressure
gauge connected to the sampler. The chamber normally contains both oil and a free
gas phase, due to the reduction in temperature between wellbore and surface. Injecting
mercury increases the pressure within the chamber until at a saturation pressure
corresponding to the ambient surface temperature all the gas will dissolve. This
saturation pressure can be quite easily detected since there is a distinct change in
compressibility between the two phase and single phase fluids. If it is experimentally
determined, on the well site, that successive samples have markedly different
saturation pressures, then either the tool has been malfunctioning or the well has not
been conditioned properly.
In addition, it is necessary to determine the static reservoir pressure and temperature
by well testing, prior to collecting the samples. Further details on bottom hole sampling
techniques are given in references 2 and 3 listed at the end of this chapter.
b) Surface recombination sampling
In collecting fluid samples at the surface, separate volumes of oil and gas are taken at
separator conditions and recombined to give a composite fluid sample. The surface
equipment is shown schematically in fig. 2.7.
gas meter
gas
sample
oil
sample
p
T
st
st
stock tank oil
p
T
sep
sep
well
separator
Fig. 2.7 Collection of a PVT sample by surface recombination
The well is produced at a steady rate for a period of several hours and the gas oil ratio
is measured in scf of separator gas per stock tank barrel of oil. If this ratio is steady
during the period of measurement then one can feel confident that recombining the oil
and gas in the same ratio will yield a representative composite sample of the reservoir
fluid. In fact, a slight adjustment must be made to determine the actual ratio in which
the samples should be recombined. This is because, as shown in fig. 2.7, the oil
sample is collected at separator pressure and temperature whereas the gas oil ratio is
measured relative to the stock tank barrel, thus the required recombination ratio is
REQUIRED MEASURED SHRINKAGE
sep
scf scf stb
R R S
sep.bbl stb sep.bbl

=

Dimensionally, the measured gas oil ratio must be multiplied by the shrinkage factor
from separator to stock tank conditions. This factor is usually determined in the
laboratory as the first stage of a PVT analysis of a surface recombination sample by
placing a small volume of the oil sample in a cell at the appropriate separator
conditions and discharging it (flash expansion) to a second cell maintained at the field
stock tank conditions. During this process some gas will be liberated from the separator
sample, due to the reduction in pressure and temperature, and the diminished stock
tank oil volume is measured, thus allowing the direct calculation of S. In order to be
able to perform such an experiment it is important that the engineer should accurately
measure the pressure and temperature prevailing at both separator and stock tank
during sampling and provide the laboratory with these data.
One of the attractive features of surface recombination sampling is that statistically it
gives a reliable value of the producing gas oil ratio measured over a period of hours;
furthermore, it enables the collection of large fluid samples. Of course, just as for
subsurface sampling, the surface recombination method will only provide the correct
gas oil ratio if the pressure in the vicinity of the well is at or above bubble point
pressure. If not, the surface gas oil ratio will be too low or too high, depending upon
whether the free gas saturation in the reservoir is below or above the critical saturation
at which gas will start to flow. In this respect it should be emphasized that PVT samples
should be taken as early as possible in the producing life of the field to facilitate the
collection of samples in which the oil and gas are combined in the correct ratio.
2.4 DETERMINATION OF THE BASIC PVT PARAMETERS IN THE LABORATORY AND
CONVERSION FOR FIELD OPERATING CONDITIONS
Quite apart from the determination of the three primary PVT parameters B
o
, R
s
and B
g
,
the full laboratory analysis usually consists of the measurement or calculation of fluid
densities, viscosities, composition, etc. These additional measurements will be briefly
discussed in section 2.6. For the moment, the essential experiments required to
determine the three basic parameters will be detailed, together with the way in which
the results of a PVT analysis must be modified to match the field operating conditions.
The analysis consists of three parts:
flash expansion of the fluid sample to determine the bubble point pressure;
differential expansion of the fluid sample to determine the basic parameters B
o
,
R
s
and B
g
;
flash expansion of fluid samples through various separator combinations to
enable the modification of laboratory derived PVT data to match field separator
conditions.
The apparatus used to perform the above experiments is the PV cell, as shown in
fig. 2.8. After recombining the oil and gas in the correct proportions, the fluid is charged
to the PV cell which is maintained at constant temperature, the measured reservoir
temperature, throughout the experiments. The cell pressure is controlled by a positive
displacement mercury pump and recorded on an accurate pressure gauge. The
plunger movement is calibrated in terms of volume of mercury injected or withdrawn
from the PV cell so that volume changes in the cell can be measured directly.
The flash and differential expansion experiments are presented schematically in
figs. 2.9(a) and 2.9(b). In the flash experiment the pressure in the PV cell is initially
raised to a value far in excess of the bubble point. The pressure is subsequently
reduced in stages, and on each occasion the total volume v
t
of the cell contents is
recorded. As soon as the bubble point pressure is reached, gas is liberated from the oil
and the overall compressibility of the system increases significantly. Thereafter, small
changes in pressure will result in large changes in the total fluid volume contained in
the PV cell. In this manner, the flash expansion experiment can be used to "feel" the
bubble point. Since the cell used is usually opaque the separate volumes of oil and
gas, below bubble point pressure, cannot be measured in the experiment and
therefore, only total fluid volumes are recorded. In the laboratory analysis the basic unit
of volume, against which all others are compared, is the volume of saturated oil at the
bubble point, irrespective of its magnitude. In this chapter it will be assumed, for
consistency, that this unit volume is one reservoir barrel of bubble point oil (1rb
b
).
PV
cell
thermal
jacket
Heise pressure
gauge
mercury
reservoir
mercury pump
Fig. 2.8 Schematic of PV cell and associated equipment
p
i
oil v
t
= v
o
Hg
p
b
oil v
t
= 1
Hg
p < p
b
oil
v
t
Hg
gas
(a)
p
b
oil
Hg
oil
Hg
gas
v
o
v
g
p < p
b
v
o
Hg
oil
(b)
gas
v
o
= 1
Fig. 2.9 Illustrating the difference between (a) flash expansion, and (b) differential
liberation
Table 2.1 lists the results of a flash expansion for an oil sample obtained by the
subsurface sampling of a reservoir with an initial pressure of 4000 psia and
temperature of 200F; the experiment was conducted at this same fixed temperature.
Pressure
psia
Relative Total
Volume
v
t
= v/v
b
= (rb/rb
b
)
5000 0.9810
4500 0.9850
4000 (p
i
) 0.9925
3500 0.9975
3330 (p
b
) 1.0000
3290 1.0025
3000 1.0270
2700 1.0603
2400 1.1060
2100 1.1680
TABLE 2.1
Results of isothermal flash expansion at 200F
The bubble point pressure for this sample is determined from the flash expansion as
3330 psia, for which the saturated oil is assigned the unit volume. The relative total fluid
volumes listed are volumes measured in relation to this bubble point volume. The flash
expansion can be continued to much lower pressures although this is not usually done
since the basic PVT parameters are normally obtained from the differential liberation
experiment. Furthermore, the maximum volume to which the cell can expand is often a
limiting factor in continuing the experiment to low pressures.
The essential data obtained from the differential liberation experiment, performed on
the same oil sample, are listed in table 2.2. The experiment starts at bubble point
pressure since above this pressure the flash and differential experiments are identical.
Pressure
psia
Relative Gas
Vol. (at p and T)
vg
Relative
Gas Vol. (sc)
Vg
Cumulative
Relative
Gas Vol. (sc)
F
Gas expansion
Factor
E
Zfactor
Z
Relative Oil
Vol. (at p and T)
vo
3330 (pb ) 1.0000
3000 .0460 8.5211 8.5211 185.24 .868 .9769
2700 .0417 6.9731 15.4942 167.22 .865 .9609
2400 .0466 6.9457 22.4399 149.05 .863 .9449
2100 .0535 6.9457 29.3856 129.83 .867 .9298
1800 .0597 6.5859 35.9715 110.32 .874 .9152
1500 .0687 6.2333 42.2048 90.73 .886 .9022
1200 .0923 6.5895 48.7943 71.39 .901 .8884
900 .1220 6.4114 55.2057 52.55 .918 .8744
600 .1818 6.2369 61.4426 34.31 .937 .8603
300 .3728 6.2297 67.6723 16.71 .962 .8459
14.7 (200F) 74.9557 .8296
14.7 ( 60F) 74.9557 .7794
All volumes are measured relative to the unit volume of oil at the bubble point pressure of 3330 psi
TABLE 2.2
Results of isothermal differential liberation at 200 F
In contrast to the flash expansion, after each stage of the differential liberation, the total
amount of gas liberated during the latest pressure drop is removed from the PV cell by
injecting mercury at constant pressure, fig. 2.3. Thus, after the pressure drop from
2700 to 2400 psia, table 2.2, column 2, indicates that 0.0466 volumes of gas are
withdrawn from the cell at the lower pressure and at 200F. These gas volumes v
g
are
measured relative to the unit volume of bubble point oil, as are all the relative volumes
listed in table 2.2. After each stage the incremental volume of liberated gas is
expanded to standard conditions and remeasured as V
g
relative volumes. Column 4 is
simply the cumulative amount of gas liberated below the bubble point expressed at
standard conditions, in relative volumes, and is denoted by F = V
g
. Dividing values in
column 3 by those in column 2 (V
g
/v
g
) gives the gas expansion factor E defined in
Chapter 1, sec. 6. Thus the .0466 relative volumes liberated at 2400 psia will expand to
give 6.9457 relative volumes at standard conditions and the gas expansion factor is
therefore 6.9457/.0466 = 149.05. Knowing E, the Zfactor of the liberated gas can be
determined by explicitly solving equ . (1.25) for Z as
sc
sc
T p 1 p
Z 35.37
p T E ET
= =
and for the gas sample withdrawn at 2400 psia
2400
Z 35.37 0.863
149.05 660
= =
These values are listed in column 6 of table 2.2.

Finally, the relative oil volumes, v
o
, are measured at each stage of depletion after
removal of the liberated gas, as listed in column 7.
Before considering how the laboratory derived data presented in table 2.2 are
converted to the required field parameters, B
o
, R
s
and B
g
, it is first necessary to
compare the physical difference between the flash and differential liberation
experiments and decide which, if either, is suitable for describing the separation of oil
and gas in the reservoir and the production of these volumes through the surface
separators to the stock tank.
The main difference between the two types of experiment shown in fig. 2.9(a) and (b) is
that in the flash expansion no gas is removed from the PV cell but instead remains in
equilibrium with the oil. As a result, the overall hydrocarbon composition in the cell
remains unchanged. In the differential liberation experiment, however, at each stage of
depletion the liberated gas is physically removed from contact with the oil and
therefore, there is a continual compositional change in the PV cell, the remaining
hydrocarbons becoming progressively richer in the heavier components, and the
average molecular weight thus increasing.
If both experiments are performed isothermally, in stages, through the same total
pressure drop, then the resulting volumes of liquid oil remaining at the lowest pressure
will, in general, be slightly different. For low volatility oils, in which the dissolved gas
consists mainly of methane and ethane, the resulting oil volumes from either
experiment are practically the same. For higher volatility oils, containing a relatively
high proportion of the intermediate hydrocarbons such as butane and pentane, the
volumes can be significantly different. Generally, in this latter case, more gas escapes
from solution in the flash expansion than in the differential liberation, resulting in a
smaller final oil volume after the flash process. This may be explained by the fact that
in the flash expansion the intermediate hydrocarbon molecules find it somewhat easier
to escape into the large gas volume in contact with the oil than in the case of the
differential liberation, in which the volume of liberated gas in equilibrium with the oil, at
any stage in the depletion, is significantly smaller.
The above description is a considerable simplification of the complex processes
involved in the separation of oil and gas; also, it is not always true that the flash
separation yields smaller oil volumes. What must be appreciated, however, is that the
flash and differential processes will yield different oil volumes and this difference can
be physically measured by experiment. The problem is, of course, which type of
experiment will provide the most realistic values of B
o
, R
s
and B
g
, required for relating
measured surface volumes to volumes withdrawn from the reservoir at the current
reservoir pressure and fixed temperature.
The answer is that a combination of both flash and differential liberation is required for
an adequate description of the overall volume changes. It is considered that the
differential liberation experiment provides the better description of how the oil and gas
separate in the reservoir since, because of their different flow velocities, they will not
remain together in equilibrium once gas is liberated from the oil, thus corresponding to
the process shown in fig. 2.9(b). The one exception to this is during the brief period
after the bubble point has been reached, when the liberated gas is fairly uniformly
distributed throughout the reservoir and remains immobile until the critical gas
saturation is exceeded.
The nature of the volume change occurring between the reservoir and stock tank is
more difficult to categorise but generally, the overall effect is usually likened to a non-
isothermal flash expansion. One aspect in this expansion during production is worth
considering in more detail and that is, what occurs during the passage of the reservoir
fluids through the surface separator or separators.
Within any single separator the liberation of gas from the oil may be considered as a
flash expansion in which, for a time, the gas stays in equilibrium with the oil. If two or
more separators are used then the gas is physically removed from the oil leaving the
first separator and the oil is again flashed in the second separator. This physical
isolation of the fluids after each stage of separation corresponds to differential
liberation. In fact, the overall effect of multi-stage separation corresponds to the
process shown in fig. 2.9(b), which is differential liberation, only in this case it is not
conducted at constant temperature. It is for this reason that multi-stage separation is
commonly used in the field because, as already mentioned, differential liberation will
normally yield a larger final volume of equilibrium oil than the corresponding flash
expansion.
The conclusion reached, from the foregoing description of the effects of surface
separation, is somewhat disturbing since it implies that the volume of equilibrium oil
collected in the stock tank is dependent on the manner in which the oil and gas are
separated. This in turn means that the basic PVT parameters B
o
and R
s
which are
measured in terms of volume "per stock tank barrel" must also be dependent on the
manner of surface separation and cannot be assigned absolute values.
The only way to account for the effects of surface separation is to perform a series of
separator tests on oil samples as part of the basic PVT analysis, and combine the
results of these tests with differential liberation data. Samples of oil are put in the PV
cell, fig. 2.8, and raised to reservoir temperature and bubble point pressure. The cell is
connected to a single or multi-stage model separator system, with each separator at a
fixed pressure and temperature. The bubble point oil is then flashed through the
separator system to stock tank conditions and the resulting volumes of oil and gas are
measured. The results of such a series of tests, using a single separator at a series of
different pressures and at a fixed temperature, are listed in table 2.3 for the same oil as
described previously (tables 2.1 and 2.2).
Separator Stock tank Shrinkage factor GOR
p T p T f
si
R
(psia) ( F) (psia) ( F) f
b
c (stb/rb
b
)
(scf/stb)
200 80 14.7 60 .7983 512
150 80 14.7 60 .7993 510
100 80 14.7 60 .7932 515
50 80 14.7 60 .7834 526
TABLE 2.3
Separator flash expansion experiments performed on the oil sample
whose properties are listed in tables 2.1 and 2.2
The shrinkage factor
f
b
c , listed in table 2.3, is the volume of oil collected in the stock
tank, relative to unit volume of oil at the bubble point (stb/rb
b
), which is the reason for
the subscript b (bubble point). The subscript f refers to the fact that these experiments
are conducted under flash conditions. All such separator tests, irrespective of the
number of separator stages, are described as flash although, as already mentioned,
multi-stage separation is closer to a differential liberation. In any case, precisely what
the overall separation process is called does not really matter since the resulting
volumes of oil and gas are experimentally determined, irrespective of the title.
f
si
R is
the initial solution gas oil ratio corresponding to the separators used and is measured
in the experiments in scf/stb.
Using the experimental separator flash data, for a given set of separator conditions, in
conjunction with the differential liberation data in table 2.2, will provide a means of
obtaining the PVT parameters required for field use. It is considered that the differential
liberation data can be used to describe the separation in the reservoir while the
separator flash data account for the volume changes between reservoir and stock tank.
What is required for field use is B
o
expressed in rb/stb. In the differential liberation data
the corresponding parameter is v
o
(rb/rb
b
), that is, reservoir barrels of oil per unit barrel
at the bubble point. But from the flash data it is known that one reservoir barrel of oil, at
the bubble point, when flashed through the separators yields
f
b
c stb. Therefore, the
conversion from the differential data to give the required field parameter B
o
is
f
o b
o
b b
v rb rb rb
B
stb c stb rb

=

Similarly, the solution gas oil ratio required in the field is R
s
(scf/stb). The parameter in
the differential liberation data from which this can be obtained is F (cumulative gas vol
at sc/oil vol at p
b
= stb/rb
b
). In fact, F, the cumulative gas liberated from the oil, must be
proportional to
f
si s
R R (scf/stb), which is the initial solution gas oil ratio, as determined
in the flash experiment, minus the current solution gas oil ratio at some lower pressure.
The exact relationship is
f
f
b
si s
b b
rb scf stb scf 1
(R R ) F 5.615
stb rb stb c stb

=

Finally, the determination of the third parameter B
g
can be obtained directly from the
differential parameter E as
g
rb 1 rcf 1 rb
B
scf E scf 5.615 rcf

=

Thus the laboratory differential data can be transformed to give the required field PVT
parameters using the following conversions
Laboratory
Differential
Parameter
Required
Field
Parameter
Conversion
v
o
(rb/ rb
b
) B
o
f
o
o
b
v rb
B
c stb

=

(2.5)
F (stb/rb
b
) R
s
f
s si
b
f
5.615 F rb
R R
c stb

=

(2.6)
E (scf/rcf) B
g g
1 rb
B
5.612 E scf

=

(2.7)
EXERCISE 2.2 CONVERSION OF DIFFERENTIAL LIBERATION DATA TO GIVE
THE FIELD PVT PARAMETERS B
o
, R
s
AND B
g
Convert the laboratory differential liberation data presented in table 2.2 to the required
PVT parameters, for field use, for the optimum separator conditions listed in table 2.3.
EXERCISE 2.2 SOLUTION
The optimum separator pressure in table 2.3 is 150 psia since this gives the largest
value of the flash shrinkage factor
f
b
c as 0.7993 (stb/rb
b
) and correspondingly, the
lowest flash solution gas oil ratio
f
si
R of 510 scf/stb. Using these two figures the
laboratory differential data in table 2.2 can be converted to give the field parameters B
o
,
R
s
and B
g
using equs. (2.5) (2.7), as follows
Pressure
(psia)
f
o
o
b
v
B
c
=
(rb/stb)
f
f
s si
b
5.615 F
R R
c
=

(scf/stb)
g
1
B
5.615 E
=
(rb/scf)
4000 (p
i
)
1.2417
f
oi
(B ) 510
f
si
(R )
3500 1.2480 510
3330 (p
b
)
1.2511
f
f
ob
b
1
(B )
c
=
510 .00087
3000 1.2222 450 .00096
2700 1.2022 401 .00107
2400 1.1822 352 .00119
2100 1.1633 304 .00137
1800 1.1450 257 .00161
1500 1.1287 214 .00196
1200 1.1115 167 .00249
900 1.0940 122 .00339
600 1.0763 78 .00519
300 1.0583 35 .01066
TABLE 2.4
Field PVT parameters adjusted for single stage, surface separation
at 150 psia and 80F;
f
b
c = .7993 (Data for pressures above 3330 psi
are taken from the flash experiment, table 2.1)
The data in table 2.4 are plotted in fig. 2.5(a) (c).
In summary of this section, it can be stated that the laboratory differential liberation
experiment, which is regarded as best simulating phase separation in the reservoir,
provides an absolute set of PVT data in which all volumes are expressed relative to the
unit oil volume at the bubble point, the latter being a unique volume. The PVT
parameters conventionally used in the field, however, are dependent on the nature of
the surface separation. The basic differential data can be modified in accordance with
the surface separators employed using equs. (2.5) (2.7) in which
f
b
c and
f
b
R are
determined by flashing unit volume of reservoir oil through the separator system. The
modified PVT parameters thus obtained approximate the process of differential
liberation in the reservoir and flash expansion to stock tank conditions. Therefore if,
during the producing life of the reservoir, the separator conditions are changed, then
the fixed differential liberation data will have to be converted to give new tables of B
o
and R
s
using values of
f
b
c and
f
si
R appropriate for the altered separator conditions.
This combination of differential liberation in the reservoir and flash expansion to the
surface is generally regarded as a reasonable approximation to Dodson's PVT analysis
technique
4
. In this form of experiment a differential liberation is performed but after
each pressure stage the volume of the oil remaining in the PV cell is flashed to stock
tank conditions through a chosen separator combination. The ratio of stock tank oil
volume to original oil volume in the PV cell prior to flashing gives a direct measure of
B
o
, while the gas evolved in the flash can be used directly to obtain R
s
. The process is
repeated taking a new oil sample for each pressure step, since the remaining oil in the
PV cell is always flashed to surface conditions. This type of analysis, while more
accurately representing the complex reservoir-production phase separation, is more
time consuming and therefore more costly, furthermore, it requires the availability of
large samples of the reservoir fluid. For low and moderately volatile crudes, the manner
of deriving the PVT parameters described in this section usually provides a very good
approximation to the results obtained from the Dodson analysis. For more volatile
crudes, however, the more elaborate experimental technique may be justified.
2.5 ALTERNATIVE MANNER OF EXPRESSING PVT LABORATORY ANALYSIS
RESULTS
The results of the differential liberation experiment, as listed in table 2.2, provide an
absolute set of data which can be modified, according to the surface separators used,
to give the values of the PVT parameters required for field use. In table 2.2 all volumes
are measured relative to the unit oil volume at the bubble point. There is, however, a
more common way of representing the results of the differential liberation in which
volumes are measured relative to the volume of residual oil at stock tank conditions.
This volume is obtained as the final step in the differential liberation experiment by
flashing the volume of oil measured at atmospheric pressure and reservoir
temperature, to atmospheric pressure and 60F. This final step is shown in table 2.2 in
which 0.8296 relative oil volumes at 14.7 psia and 200F yield 0.7794 relative oil
volumes at 14.7 psia and 60F. This value of 0.7794 is the shrinkage factor for a unit
volume of bubble point oil during differential liberation to stock tank conditions and is
denoted by
d
b
c . The value of
d
b
c ,is not dependent on any separator conditions and
therefore, relating all volumes in the differential liberation to this value of
d
b
c , which is
normally referred to as the "residual oil volume", will provide an alternative means of
expressing the differential liberation results.
It should be noted, however, that the magnitude of
d
b
c is dependent on the number of
pressure steps taken in the differential experiment. Therefore, the differential liberation
results, in which all volumes are measured relative to
d
b
c do not provide an absolute
set of data such as that obtained by relating all volumes to the unit volume of oil at the
bubble point.
In the presentation of differential data, in which volumes are measured relative to
d
b
c ,
the values of v
o
and F in table 2.2 are replace by
d
o
B and
d
s
R where
d
o
B = Differential oil formation volume factor
(rb/stb-residual oil)
and
d
s
R = Differential solution gas oil ratio
(scf/stb-residual oil)
Alternatively, by replacing
f
b
c in equs. (2.5) and (2.6) by
d
b
c , these parameters can be
expressed as
d
d
o b
o
b b
v rb rb
B
c stb residual rb

=

(2.8)
and
d d
d
s si
b
5.615 F scf
R R
c stb residual

=

(2.9)
where
d
si
R is the initial dissolved gas relative to the residual barrel of oil at 60F, and is
proportional to the total gas liberated in the differential experiment, thus
d
d
si
b
(Maximumvalueof F) scf
R 5.615
c stb residual

=

(2.10)
and for the differential data presented in table 2.2
d
si
74.9557 5.615
R 540 scf stb residual oil
.7794
= =
The majority of commercial laboratories serving the industry would normally present
the essential data in the differential liberation experiment (table 2.2) as shown in
table 2.5.
There is a danger in presenting the results of the differential liberation experiment in
this way since a great many engineers are tempted to use the
d
o
B and
d
s
R values
directly in reservoir calculations, without making the necessary corrections to allow for
the surface separator conditions. In many cases, the error in directly using the data in
table 2.5 is negligible, however, for moderate and high volatility oils the error can be
quite significant and therefore, the reader should always make the necessary
correction to the data in table 2.5 to allow for the field separator conditions, as a matter
of course.
Pressure
(psia)
Formation Vol. Factor
d d
o o b
B v / c =
Solution GOR
d d d
s si b
R R 5.615 F/ c =
4000 1.2734 540
3500 1.2798 540
3300
1.2830
d
ob
(B ) 540
d
si
(R )
3000 1.2534 479
2700 1.2329 428
2400 1.2123 378
2100 1.1930 328
1800 1.1742 281
1500 1.1576 236
1200 1.1399 118
900 1.1219 142
600 1.1038 97
300 1.0853 52
14.7 (200F) 1.0644 0
14.7 ( 60F) 1.0000 0
TABLE 2.5
Differential PVT parameters as conventionally presented by laboratories, in which
B
o
and R
s
are measured relative to the residual oil volume at 60F
The conversion can be made by expressing and R
s
d
, in table 2.5, in their equivalent,
absolute forms of v
o
and F, in table 2.2, using equs. (2.8) and (2.9) and thereafter,
using equs. (2.5) and (2.6) to allow for the surface separators. This will result in the
required expressions for B
o
and R
s
. Alternatively, the required field parameters can be
calculated directly as
d f
d
f d f f
b ob
o o
o o
b b b ob
c B
v v
B B
c c c B

= = =

(2.11)
where
d
o b
v / c =
d
o
B the differential oil formation volume factor measured relative to the
residual oil volume as listed in table 2.5 (rb/stb-residual);
f
ob
B =
f
b
1/ c is the oil formation volume factor of the bubble point oil (rb
b
/stb)
determined by flashing the oil through the appropriate surface separators
and is measured relative to the stock tank oil volume (refer tables 2.3 and
2.4); and
d
ob
B =
d
b
1/ c is the oil formation volume factor of the bubble point oil determined
during the differential liberation experiment and is measured relative to
the residual oil volume (refer table 2.5) (rb
b
/stb-residual).
Similarly, the required solution gas oil ratio for use under field operating conditions is,
equ. (2.6)
d
f f
f d f
b
s si si
b b b
c
5.615 F 5.615 F
R R R
c c c
=

=

which, using equ. (2.9), can be expressed as
f
f d d
f
ob
s si si s
ob
B
R R (R R )
B

=

(2.12)
where
f
si
R = solution gas oil ratio of the bubble point oil, determined by flashing the oil
through the appropriate surface separators, and is measured relative to the oil volume
at 60F and 14.7 psia (refer tables 2.3 and 2.4) (scf/stb).
d
si
R = solution gas oil ratio of the bubble point oil determined during the
differential experiment and measured relative to the residual oil volume at
60F and 14.7 psia (refer table 2.5 and equ. (2.10)) (scf/stb-residual).
The differential data, as presented in table 2.5, can be directly converted to the
required form, table 2.4, using the above relations. For instance, using the following
data from table 2.5, at a pressure of 2400 psi
d
o
B = 1.2123 (rb /barrel of residual oil at 60F and 14.7 psia)
d
s
R = 378 (scf/ " )
d
ob
B = 1.2830 (rb / " )
d
si
R = 540 (scf/ " )
while from the separator flash tests (table 2.3), for the optimum separator conditions of
150 psia and 80F
f f
ob b
B (1/ c ) 1.2511(rb/ stb) = =
f
si
R 510 (scf / stb) =
Therefore, using equ. (2.11)
o
1.2511
B 1.2123 1.1822 rb stb
1.2830
= =
and equ. (2.12)
s
1.2511
R 510 (540 378) 352 scf / stb
1.2830
= =
2.6 COMPLETE PVT ANALYSIS
The complete PVT analysis for oil, provided by most laboratories, usually consists of
the following experiments and calculations.
a) Compositional analysis of the separator oil and gas, for samples collected at the
surface, together with physical recombination, refer sec. 2.3(b), or; compositional
analysis of the reservoir fluid collected in a subsurface sample.
Such analyses usually give the mole fractions of each component up to the
hexanes. The hexanes and heavier components are grouped together, and the
average molecular weight and density of the latter are determined.
b) Flash expansion, as described in sec. 2.4 (table 2.1), conducted at reservoir
temperature. This experiment determines
the bubble point pressure
the compressibility of the undersaturated oil as
o o
o
o o
dv dB 1 1
c
v dp B dp
= = (2.13)
the total volume v
t
of the oil and gas at each stage of depletion.
c) Differential liberation experiment as described in sec. 2.4 to determine
E, Z, F and v
o
(as listed in table 2.2), with F and v
o
measured relative to the
unit volume of bubble point oil.
Alternatively, by measuring
d
b
c during the last stage of the differential liberation,
the above data can be presented as
E, Z,
d
si
R
d
s
R (or just
d
s
R ) and
d
o
B (as listed in table 2.5), with
d
s
R and
d
o
B measured relative to residual oil volume. In addition, the gas gravity is
measured at each stage of depletion.
d) Measurement of the oil viscosity at reservoir temperature (generally using the
rolling ball viscometer
1,3
), over the entire range of pressure steps from above
bubble point to atmospheric pressure. Gas viscosities are normally calculated at
reservoir temperature, from a knowledge of the gas gravity, using standard
correlations
5
.
e) Separator tests to determine the shrinkage,
f
b
c , and solution gas oil ratio,
f
si
R , of
unit volume of bubble point oil (1 barrel) when flashed through various separator
combinations (refer table 2.3). Instead of actually performing these tests, in many
cases the results are obtained using the phase equilibrium calculation
technique
1
.
f) Composition and gravity of the separator gas in the above separator tests.
Copyright 2002, Society of Petroleum Engineers Inc.

This paper was prepared for presentation at the SPE Annual Technical Conference and
Exhibition held inSan Antonio, Texas, 29 September 2 October 2002.

This paper was selected for presentation by an SPE Program Committee following review of
information contained in an abstract submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petrol eum Engineers is
prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300
words; illustrations may not be copied. The abstract must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

Abstract
Combinations of data from two laboratory procedures,
differential liberation and separator test, are used to determine
values of oil formation volume factors and solution gas-oil
ratios for pressures below the bubblepoint pressure of the
reservoir oil. The equation commonly used to calculate the
solution gas -oil ratio is incorrect: the correct equation is
derived. The equation used to calculate the oil formation
volume factors is correct; however, a derivation illustrating the
underlying assumptions has not been published. This
derivation is also examined.

Introduction
The bad news is that the equation usually used in the
petroleum industry to calculate solution gas -oil ratios from
black oil PVT reports,

( )
oSb
s sSb sDb sD
oDb
B
R R R R ,
B
= . . . . . . . . . . . . . . . . . . (1)

is incorrectly formulated.
The good news is that the errors at high pressures, where
solution gas -oil ratios are more often used, are not severe. A
correct formulation of the equation will improve estimates of
solution gas -oil ratio throughout the full range of
depletion pressures.
Further good news is that the equations used to calculate
oil formation volume factors from black oil PVT reports,

( ) ( )
o oE oSb b
B B B at p>p = . . . . . . . . . . . . . . . . . . . . . (2a)

and

( )
oSb
o oD b
oDb
B
B B at p<p ,
B
= . . . . . . . . . . . . . . . . . . . . . (2b)

are correct.
This paper presents a brief discussion of the laboratory
procedures, which will lead to the derivations of Eq. 2b and a
correct replacement for Eq. 1.

Laboratory Procedures
The two laboratory procedures that provide the necessary
data are the differential liberation (sometimes called
differential vaporization or differential distillation) and the
separator test (sometimes, incorrectly, called flash test). Each
of these laboratory procedures will be described briefly. What
will be seen is that each procedure starts with a quantity of
reservoir oil in the laboratory cell at its bubblepoint pressure at
the reservoir temperature. In each procedure, gas is removed
in a sequence of flash vaporizations, with the resulting liquid
ending up at atmospheric pressure and 60F
(standard conditions).

Differential Liberation. Theoretically, the differential
liberation starts with a liquid at some pressure and
temperature, then the liquid is partially vaporized, and each
small increment of vapor is at once removed from the contact
with the liquid.
1
Thus, the liquid is in equilibrium at any
instant with a small amount of vapor. This procedure is
somewhat tedious, so the petroleum industry emulates the
differential liberation with a series of flash vaporizations in
which a definite fraction of a batch of liquid is vaporized and
kept in intimate contact with the liquid until the gas is
withdrawn at the end of each step in the series.
In practice, the process starts with a sample of reservoir
oil at its bubblepoint pressure in the laboratory cell with the
temperature controlled at reservoir temperature. A dozen or
so steps (the exact number depends on the starting bubblepoint
pressure) consisting of flash vaporizations are carried out.
Each step starts with a pressure reduction at constant reservoir
temperature. This causes gas to be vaporized. The gas is

SPE 77386
Analysis of Black Oil PVT Reports Revisited
William D. McCain, Jr., Texas A&M University
2 WILLIAM D. MCCAIN, JR. SPE 77386
allowed to come to equilibrium with the liquid at the lower
pressure and reservoir temperature, and then the gas is
removed, and its quantity and specific gravity are measured.
The volumes of the liquid remaining at the end of each step
are determined. The last flash vaporization step ends at
atmospheric pressure. Then the temperature of the remaining
liquid is reduced to 60F and the liquid volume is adjusted to
maintain the pressure at 0 psig. This final liquid is called the
residual liquid from the differential liberation or, simply,
residual oil.
The volumes of gas removed in all steps are added; this is
the amount of gas in solution at bubblepoint pressure. The gas
volumes are decremented to determine the gas remaining in
solution at the pressure at the end of each step.
Finally, the volumes of gas in solution (in standard cubic
feet, scf) and the volumes of the oil in the cell (in reservoir
barrels, res. bbl) at the end of each step are divided by the
volume of the residual oil (at atmospheric pressure and 60F).
These are presented in the laboratory report as gas in solution,
scf/residual barrel, and relative oil volume, reservoir
barrel/residual barrel. Other properties, such as reservoir oil
density, incremental gas specific gravity, and gas
compressibility factor, are also measured and reported.

Separator Test. The separator test starts with a sample of the
reservoir oil at its bubblepoint pressure in the laboratory cell at
reservoir temperature (same as the start of the differential
liberation). A measured volume of the reservoir oil is expelled
through a sequence of two (usually, though sometimes three)
flash vaporizations. In the first, at separator temperature and
pressure, the gas that was vaporized is removed, and the liquid
goes to the second flash at stock-tank temperature and
atmospheric pressure. The volume of resulting liquid is
determined at atmospheric pressure and 60F; this liquid is
usually called stock-tank oil and the volume is reported in
stock-tank barrels, STB, which also could be interpreted as
standard barrels.
The volumes and specific gravities of the gases from the
two flash vaporizations are measured. The two gas volumes
are added, and the sum is reported as solution gas-oil ratio at
the bubblepoint, scf/STB. The volume of reservoir oil that
was expelled at the start of the test is divided by the volume of
stock-tank oil (at standard conditions) and reported as oil
formation volume factor at the bubblepoint, res. bbl/STB. The
density of the stock-tank oil is measured and usually reported
in API.

Example Laboratory Data
Table 1 shows selected data from a differential liberation
and a separator test for a black oil. The bubblepoint pressure
is 3043 psig at a reservoir temperature of 262F.
The quantity and properties of oil and gas produced by the
two procedures are different. However, this is difficult to see
in the formats in which the data are presented. A comparison
of the results of the two procedures may be made by placing
the data on a basis of one barrel of reservoir oil at bubblepoint
pressure and reservoir temperature. The oil in place at the
bubblepoint is the starting point of both processes and is
independent of the process. Table 2 shows this comparison
for the data of Table 1. The quantity of oil resulting from the
separator test is eighteen percent higher than the oil resulting
from the differential liberation. The quantity of gas resulting
from the separator test is fifteen percent lower than the gas
from the differential liberation. And compositions and
properties of the resulting oils and gases differ significantly, as
represented by the different API gravities of the oils and
specific gravities of the gases.

Calculation of Oil Formation Volume Factors
The ratio

oDb b oSb
oDb
oSb b
1 residual bbl
B res. bbl @ p B residual bbl
1 STB
B STB
B res. bbl @ p
= . . . . . . . . . . . (3)

provides a convenient relationship between residual barrels
from the differential liberation and stock-tank barrels from the
separator test. The term reservoir barrel at bubblepoint
pressure (at reservoir temperature), which appears twice on the
left-hand side of Eq. 3, can be eliminated since both laboratory
procedures start with the reservoir oil in the laboratory cell at
its bubblepoint pressure at reservoir temperature.
The change in volume of oil in the reservoir during
pressure depletion can be written as

( )
oSb
oSb o oDb oD
oDb
B
B B B B
B
= . . . . . . . . . . . . . . . . . . . (4)

The units of this equation are

change in res bbl of oil change in res bbl of oil residual bbl
STB residual bbl STB

=

Under the assumption that the change in volume (res. bbl) of
oil in the reservoir is the result of a differential liberation
process, the change units in the numerator of each side of
the equation are identical. Thus, the equation is correct.
Eq. 4 can be rearranged as

oSb
o oD
oDb
B
B B ,
B
= . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2b)

which is the equation generally used in the petroleum industry
to calculate oil formation volume factors from differential
liberation and separator test data.
Note that the use of Eq. 2b requires that the ratio of the
volumes of residual oil to stock-tank oil remains constant (for
a particular oil sample), regardless of the starting pressure.
The limited experimental data that are available show that this
is true.
2

SPE 77386 ANALYSIS OF BLACK OIL PVT REPORTS REVISITED 3
There are very limited data to test the results of Eqs. 2a
and 2b against measured oil formation volume factors.
However, Dodson, et al.,
2
provided one set of data. Dodson,
et al., proposed a laboratory procedure for determining oil
formation volume factors (and solution gas -oil ratios) that is
generally considered more accurate than the differential
liberation/separator test procedure usually used and discussed
in this paper. Unfortunately their composite liberation,
although considered a superior method,
3
requires a large
sample of reservoir fluid and is very time consuming in the
laboratory. Thus, it is not used in routine laboratory analysis.
But Dodson, et al., did provide one example of a routine
laboratory report and the results of a composite liberation for
the same black oil. Fig. 1 shows values of oil formation
volume factor calculated with Eqs. 2a and 2b compared with
the more accurate results of the composite liberation.
The small differences between the results of Eq. 2b and
the data from the composite liberation are approximately one
percent, well within experimental accuracy. The composite
liberation is an entirely different laboratory procedure; the
differential liberation/separator test discussed here is the
industrys (less expensive) method of approximating the
composite liberation.

Calculation of Solution Gas-oil Ratios
Eq. 1 is used commonly in the petroleum industry to combine
data from differential liberation and separator tests to calculate
solution gas -oil ratios. The validity of this equation can be
examined easily by rearrangement:

( )
oSb
sSb s sDb sD
oDb
B
R R R R
B
= . . . . . . . . . . . . . . . . . . . (5)

The units are

gas liberated in sep test, scf gas libera ted in diff lib, scf residual bbl
STB residual bbl STB

=

Eq. 5 shows that the gas volume liberated during the
separator test has been set equal to the gas volume liberated
during the differential liberation. If the sources of the data are
not taken into account, the units, scf/STB, appear to be
correct. However, Table 2 shows that the gas liberated during
a separator test is significantly different in quantity and quality
from the gas liberated during a differential liberation. The
ratio B
oSb
/B
oDb
in Eq. 5 takes into account the differences in
the oils from the separator test and differential liberation, but
the differences in the gases are ignored.
Thus, the material balance expressed in Eq. 5 must be
incorrect. It follows that values of solution gas-oil ratio
calculated with Eq. 1 must be in error! In fact, this is
illustrated every time the equation is used because calculated
values of solution gas -oil ratio are generally negative at
low pressures.
The correct formulation is as follows. The equation must
calculate the gas remaining in solution in the reservoir oil at a
pressure after pressure depletion from p
b
to some p, R
s
,
scf/STB. Further, R
s
should be the amount of gas to be
liberated through a separator/stock-tank sequence.
R
sSb
is the gas originally in solution in the reservoir oil at
its bubblepoint pressure as measured in a separator test, scf of
gas from sep. test/STB. R
sDb
R
sD
is the volume of gas
liberated in the reservoir during a differential liberation from
p
b
to p, scf of diff. lib. gas/residual bbl.
The ratio

sSb
sDb
scf of gas from sep. test
R
STB
scf of gas from diff. lib.
R
residual bbl

takes into account both the difference in the two oils, residual
bbl/STB, and the difference in the two gases, scf of gas from
sep. test/scf of gas from diff. lib.
Thus,

( )
sSb
sDb sD
sDb
R
R R
R
scf of gas lib. by diff. lib. scf of gas from sep. test, residual bbl
residual bbl scf of gas from diff. lib., STB

is the gas differentially liberated converted to scf of sep.
gas/STB. The difference between the gas originally in
solution and the gas liberated during depletion from p
b
to p is
the gas remaining is solution at p.

( )
sSb
s sSb sDb sD
sDb
R
R R R R
R
= . . . . . . . . . . . . . . . . . . (6)

Again, there is the assumption that the differential liberation
mimics the depletion in an oil reservoir, i.e., the gas remaining
in solution is that left after gas has been removed by
differential liberation.
Eq. 6 can be rearranged into a simpler form

sSb
s sD
sDb
R
R R
R
= . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (7)

The use of Eq. 7 implies that the ratio of the gas liberated
by separator test to the gas liberated by differential liberation
be constant (for a particular oil sample), regardless of the
starting pressure. The limited available data from Dodson, et
al.
2
show this to be true.
Again, the Dodson, et al.
2
data are the only data available
to use in comparing the results of Eq. 1 and Eq. 7 with
experimental solution gas -oil ratios measured at pressures
below bubblepoint pressure of the oil.
4 WILLIAM D. MCCAIN, JR. SPE 77386
Fig. 2 shows the results of Eqs. 1 and 7 using the data
from the routine laboratory report compared with the more
accurate results of the composite liberation. Eq. 7 gives
values of R
s
, which fit the composite liberation more closely
than the results of Eq. 1. Further, the values of R
s
from Eq. 7
converge to zero at 0 psig (as required by the definition of R
s
),
while the values from Eq. 1 are negative at low pressures.
The small differences between the results of Eq. 7 and the
composite liberation data in Fig. 2, as well as the good
comparison of the results of Eqs. 2a and 2b in Fig. 1, show
that the simpler, less costly, routinely used PVT procedures
give results adequate for reservoir engineering work.
The difference between the results of Eq. 7 and Eq. 1 in
Fig. 2 are not very dramatic. However, notice that the
Dodson, et al.
2
oil does not have much dissolved gas, with R
sb

of about 600 scf/STB. Fig. 3 shows that the oil of Table 1,
which has about 1000 scf/STB originally in solution, shows a
much larger difference between the results of the two
equations, about thirty percent at a reservoir pressure of
1100 psig.

Conclusions
Eqs. 2a, 2b, and 7 are the correct equations for calculating oil
formation volume factors and solution gas-oil ratios at
reservoir pressures below the bubblepoint pressure when
combining the data from differential liberation with separator
test data.
The use of differential liberation data for these
calculations is nearly as accurate as the more costly composite
liberation data and is certainly adequate for reservoir
engineering calculations.

Nomenclature
B
o =
Oil formation volume factors at pressure,
res. bbl/STB
B
oSb =
Oil formation volume factor at bubblepoint
pressure measured in a separator test, res.
bbl @ p
b
/STB
B
oD =
Oil relative volumes at pressures less than
bubblepoint pressure measured in a
differential liberation, res. bbl/residual bbl
B
oDb =
Oil relative volume at bubblepoint pressure
measured in a differential liberation, res. bbl
@ p
b
/residual bbl
B
oE =
Oil relative volume at pressures above
bubblepoint pressure measured in a constant
mass expansion, res. bbl/res. bbl @ p
b

R
s =
Solution gas-oil ratios at pressures less than
bubblepoint pressure, scf/STB
R
sSb =
Solution gas-oil ratio at bubblepoint
pressure measured in a separator test, scf
from separator test /STB
R
sD =
Solution gas-oil ratios at pressures less than
bubblepoint pressure measured in a
differential liberation, scf from differential
liberation/residual bbl
R
sDb =
Solution gas-oil ratio at bubblepoint
pressure measured in a differential
liberation, scf from differential
liberation/residual bbl

References
1. Dodge, B.F.: Chemical Engineering Thermodynamics,
McGraw-Hill, New York (1944), 592.
2. Dodson, C.R., Goodwill, D., and Mayer, E.H.:
Application of Laboratory PVT Data to Reservoir
Engineering Problems, Trans., AIME (1953) 198, 287
298.
3. Moses, P.L.: Engineering Applications of Phase
Behavior of Crude Oil and Condensate Systems, J. Pet.
Tech. (July 1986) 38, 715723.

TABLE 1. SELECTED DATA FROM DIFFERENTIAL LIBERATION
AND SEPARATOR TEST OF A BLACK OIL
Differential Liberation at 262F
Pressure,
psig
Gas in
Solution,
scf/residual bbl
Relative Oil
Volume,
res. bbl/residual bbl
Incre-
mental
Gas
Specific
Gravity
3043 1401 2.022
2900 1307 1.965 0.822
2600 1140 1.866 0.867
2300 998 1.783 0.858
2000 868 1.709 0.853
1700 751 1.644 0.856
1400 642 1.584 0.861
1100 537 1.526 0.881
800 436 1.468 0.922
500 331 1.407 1.008
263 235 1.339 1.197
143 176 1.290 1.432
0 0 1.111 2.194
0 at 60F = 1.000
Gravity of Residual Oil = 43.2API @ 60F

SEPARATOR TEST
Flash
Vapori -
zation
Pres-
sure,
psig
Temper-
ature,
F
Gas-oil
Ratio,
scf/STB
Oil
Forma-
tion
Volume
Factor,
res.
bbl/
STB
Stock-
tank Oil
Gravity,
API @
60F
Gas
Specific
Gravity
Separator 100 70 935 0.775
Stock
Tank 0 72 102 1.714** 48.0 1.309
1037*
*Solution gas-oil ratio at bubblepoint pressure
**Oil formation volume factor at bubblepoint pressure

SPE 77386 ANALYSIS OF BLACK OIL PVT REPORTS REVISITED 5

Fig. 1 Comparison showing excellent fit of oil formation
volume factors calculated from routine laboratory PVT
data using Eqs. 2a and 2b with data from composite
liberation.

1
1.05
1.1
1.15
1.2
1.25
1.3
0 500 1000 1500 2000 2500 3000 3500 4000
Reservoir pressure, psig
O
i
l

f
o
r
m
a
t
i
o
n

v
o
l
u
m
e

f
a
c
t
o
r
,

r
e
s
.

b
b
l
/
S
T
B
data from composite liberation calculated with Eqs. 2a and 2b

Fig. 2 Comparison of solution gas-oil ratios calculated
from routine laboratory PVT data using Eqs. 1 and 7 with
data from composite liberation showing that Eq. 7 has a
better fit.

-100
0
100
200
300
400
500
600
700
0 500 1000 1500 2000 2500 3000 3500 4000
S
o
l
u
t
i
o
n

g
a
s
-
o
i
l

r
a
t
i
o
,

s
c
f
/
S
T
B
data from composite liberation calculated with Eq. 7 calculated with Eq. 1

Fig. 3 Comparison of solution gas-oil ratios calculated
from routine laboratory PVT data using Eqs. 1 and 7,
showing that the deviation between the results is greater
when initial solution gas-oil ratio is large.

-200
0
200
400
600
800
1000
1200
0 500 1000 1500 2000 2500 3000 3500
S
o
l
u
t
i
o
n

g
a
s
-
o
i
l
r
a
t
i
o
,
s
c
f
/
S
T
B
Eq. 7 Eq. 1
TABLE 2. THE VOLUMES AND PROPERTIES OF THE OIL AND
GAS RESULTING FROM THE TWO LABORATO RY PROCEDURES
ARE QUITE DIFFERENT (DATA FROM TABLE 1)

Differential
Liberation Separator Test
Volume of oil at standard
conditions at end of
process
0.495
b
residual bbl
res. bbl @ P
0.583
b
STB
res. bbl @ P

Gravity of oil at standard
conditions at end of
process
43.2API 48.0API
Volume of total gas at
standard conditions
liberated during process
692.9
b
scf
res. bbl @ P
605.0
b
scf
res. bbl @ P

Weighted average specific
gravity at standard
conditions of total gas
liberated during process
1.0931 0.8252
88 PHASE BEHAVIOR
Chapter 6
Conventional PVT Measurements
6.1 Introduction
This chapter reviews the standard experiments performed by pres-
sure/volume/temperature (PVT) laboratories on reservoir fluid
samples: compositional analysis, multistage surface separation,
constant composition expansion (CCE), differential liberation ex-
pansion (DLE), and constant volume depletion (CVD). We present
data from actual laboratory reports and give methods for checking
the consistency of reported data for each experiment. Chaps. 5 and
8 discuss special laboratory studies, including true-boiling-point
(TBP) distillation and multicontact gas-injection tests, respectively.
Table 6.1 summarizes experiments typically performed on oils
and gas condensates. From this table, we see that the DLE experi-
ment is the only test never performed on gas-condensate systems.
We begin by discussing standard analyses performed on oil and gas-
condensate samples.
6.1.1 General Information Sheet. Most commercial laboratories
report general information on a cover sheet of the laboratory report,
including formation and well characteristics and sampling condi-
tions. Tables 6.2 and 6.3
1,2
show this information, which may be
important for correct application and interpretation of the fluid anal-
yses. This is particularly true for wells where separator samples
must be recombined to give a representative wellstream composi-
tion. Most of these data are supplied by the contractor of the fluid
study and are recorded during sampling. Therefore, the representa-
tive for the company contracting the fluid study is responsible for
the correctness and completeness of reported data.
We strongly recommend that the following data always be reported
in a general information sheet: (1) separator gas/oil ratio (GOR) in
standard cubic feet/separator barrel, (2) separator conditions at sam-
pling, (3) field shrinkage factor used ( +B
osp
), (4) flowing bottom-
hole pressure (FBHP) at sampling, (5) static reservoir pressure, (6)
minimum FBHP before and during sampling, (7) time and date of
sampling, (8) production rates during sampling, (9) dimensions of
sample container, (10) total number and types of samples collected
during the drillstem test, and (11) perforation intervals.
6.1.2 Oil PVT Analyses. Standard PVT analyses performed on res-
ervoir oils usually include (1) bottomhole wellstream compositional
analysis through C
7)
, (2) CCE, (3) DLE, and (4) multistage-separa-
tor tests. The CCE experiment determines the bubblepoint pressure
and volumetric properties of the undersaturated oil. It also gives
two-phase volumetric behavior below the bubblepoint; however,
these data are rarely used. The DLE experiment and separator test
are used together to calculate traditional black-oil properties, B
o
and R
s
, for reservoir-engineering calculations. Occasionally,
instead of a DLE study, a CVD experiment is run on a volatile oil.
Also, the C
7)
fraction may be separated into single-carbon-number
cuts from C
7
through approximately C
20)
by TBP analysis or simu-
lated distillation (see Chap. 5).
6.1.3 Gas-Condensate PVT Analyses. The standard experimental
program for a gas-condensate fluid includes (1) recombined well-
stream compositional analysis through C
7)
, (2) CCE, and (3) CVD.
The CCE and CVD data are measured in a high-pressure visual cell
where the dewpoint pressure is determined visually. Total volume/
pressure and liquid-dropout behavior is measured in the CCE ex-
periment. Phase volumes defining retrograde behavior are mea-
sured in the CVD experiment together with Z factors and
produced-gas compositions through C
7)
. Optionally, a multistage-
separator test can be performed as well as TBP analysis or simulated
distillation of the C
7)
into single-carbon-number cuts from C
7
to
about C
20)
(see Chap. 5).
6.2 Wellstream Compositions
PVT studies usually are based on one or more samples taken during
a production test. Bottomhole samples can be obtained by wireline
with a high-pressure container during either production testing or a
shut-in period. Alternatively, separator samples can be taken during
a production test. Bottomhole sampling is the preferred method for
most oil reservoirs, while recombined samples are traditionally used
for gas-condensate reservoirs.
3-8
Taking both bottomhole and sepa-
rator samples in oil wells is not uncommon. The advantage of sepa-
rator samples is that they can be recombined in varying proportions
to achieve a desired bubblepoint pressure (e.g., initial reservoir
pressure); these larger samples are needed for special PVT tests
(e.g., TBP and slim tube among others).
6.2.1 Bottomhole Sample. Table 6.4 shows the reported wellstream
composition of a reservoir oil where C
7)
specific gravity and molec-
ular weight are also reported. In the example report, composition is
given both as mole and weight percent although many laboratories re-
port only molar composition. Experimentally, the composition of a
bottomhole sample is determined by the following (Fig. 6.1).
1. Flashing the sample to atmospheric conditions.
2. Measuring the volumes of surface gas, V
g
, and surface oil, V
o
.
3. Determining the normalized weight fractions, w
gi
and w
oi
, of
surface samples by gas chromatography.
4. Measuring surface-oil molecular weight, M
o
, and specific
gravity, g
o
.
CONVENTIONAL PVT MEASUREMENTS 89
TABLE 6.1LABORATORY ANALYSES PERFORMED ON
RESERVOIR-OIL AND GAS-CONDENSATE SYSTEMS
Laboratory Analysis Oils Gas Condensates
Bottomhole sample D d
Recombined composition d D
C
7+
TBP distillation d d
C
7+
simulated distillation d d
Constant-composition expansion D D
Multistage surface separation D d
Differential liberation D N
CVD d D
Multicontact gas injection d d
D+standard, d+can be performed, and N+not performed.
5. Converting w
gi
weight fractions to normalized mole fractions
y
i
and x
i
.
6. Recombining mathematically to the wellstream composition, z
i
.
Eqs. 6.1 through 6.5 give Steps 1 through 6 mathematically.
z
i
+F
g
y
i
)(1 *F
g
) x
i
; (6.1) . . . . . . . . . . . . . . . . . . . . . . . .
F
g
+
1
1 )
133, 300
gM
o
R
s
, (6.2) . . . . . . . . . . . . . . . . . .
where R
s
+GOR V
g
V
o
in scf/STB from the single-stage flash;
y
i
+
w
g i
M
i
j0C
7)
w
g j
M
j
w
g C
7)
M
g C
7)
; (6.3) . . . . . . . . .
x
i
+
w
o i
M
i
j0C
7)
w
o j
M
j
w
o C
7)
M
o C
7)
; (6.4) . . . . . . . . . .
and M
o C
7)
+
w
o C
7)
1M
o
*
j0C
7)
w
oj
M
j
. (6.5) . . . . . . . . . . . . .
Surface gas usually contains less than 1 mol% C
7)
material con-
sisting mainly of heptanes and octanes; M
g C
7)
+105 is usually a
good assumption. Surface oil contains less than 1 mol% of the light
constituents C
1
, C
2
, and nonhydrocarbons. Low-temperature dis-
tillation can be used to improve the accuracy of reported weight
fractions for intermediate components in the surface oil ( C
3
through
C
6
); however, gas chromatography is more widely used.
The most probable source of error in wellstream composition of a
bottomhole sample is the surface-oil molecular weight, M
o
, which
appears in Eq. 6.2 for F
g
and Eq. 6.4 for x
i
. M
o
is usually accurate
within "4 to 10%. In Chap. 5, we showed that the Watson character-
ization factor, K
w
, of surface oil (Eq. 5.35) should be constant (to
within "0.03 of the determined value) for a given reservoir. Once an
average has been established for a reservoir (usually requiring three
separate measurements), potential errors in M
o
can be checked. A
calculated K
w
that deviates from the field-average K
w
by more than
"0.03 may indicate an erroneous molecular-weight measurement.
Eqs. 6.1 through 6.4 show that all component compositions are
affected by M
o C
7)
, which is backcalculated from M
o
with Eq.
6.5. Fortunately, the amount of lighter components (particularly C
1
)
in the surface oil are small, so the real effect on conversion from
weight to mole fractions of the surface oil usually is not significant.
6.2.2 Recombined Samples. Tables 6.5 and 6.6 present the separa-
tor-oil and -gas compositional analyses of a gas-condensate fluid
and recombined wellstream composition. The separator-oil com-
position is obtained by use of the same procedure as that used for
bottomhole oil samples (Eqs. 6.1 through 6.5). This involves bring-
ing the separator oil to standard conditions, measuring properties
TABLE 6.2EXAMPLE GENERAL INFORMATION SHEET
FOR GOOD OIL CO. WELL 4 OIL SAMPLE
Formation Characteristics
Name Cretaceous
First well completed / /19 (m/d/y)
Original reservoir pressure at 8,692 ft, psig 4,100
Original produced GOR, scf/bbl 600
Production rate, B/D 300
Separator temperature, F 75
Separator pressure, psig 200
Oil gravity at 60F, API
Datum 8,000
Original gas cap No
Well Characteristics
Elevation, ft 610
Total depth, ft 8,943
Producing interval, ft 8,684 to 8,700
Tubing size, in. 2
7
/
8
Tubing depth, ft 8,600
PI at 300 B/D, B-D/psi 1.1
Last reservoir pressure at 8,500 ft, psig 3,954*
Date / /19 (m/d/y)
Reservoir temperature at 8,500 ft, F 217*
Well status Shut in 72 hours
Pressure gauge Amerada
Normal production rate, B/D 300
GOR, scf/bbl 600
Separator pressure, psig 200
Separator temperature, F 75
Base pressure, psia 14.65
Well making water, % water cut 0
Sampling Conditions
Sample depth, ft 8,500
Well status Shut in 72 hours
GOR
Separator pressure, psig
Separator temperature, F
Tubing pressure, psig 1,400
Casing pressure, psig
Sampled by
Sampler type Wofford
*Pressure and temperature extrapolated to the midpoint of the producing
interval+4,010 psig and 220F.
and compositions of the resulting surface oil and gas, and recombin-
ing these compositions to give the separator-oil composition; Tables
6.5 and 6.6 report the results.
Separator gas is introduced directly into a gas chromatograph,
which yields weight fractions, w
g
. These weight fractions are con-
verted to mole fractions, y
i
, by use of appropriate molecular
weights; Tables 6.5 and 6.6 show the results. C
7)
molecular weight
is backcalculated with measured separator-gas specific gravity, g
g
.
M
g
C
7)
+w
g
C
7)
1
28.97g
g
*
i0C
7)
w
g i
M
i
*1
. (6.6) . . . . . . .
90 PHASE BEHAVIOR
TABLE 6.3EXAMPLE GENERAL INFORMATION SHEET
FOR GOOD OIL CO. WELL 7 GAS CONDENSATE
Formation Characteristics
Formation name Pay sand
Date first well completed / /19 (m/d/y)
Original reservoir pressure at 10,148 ft, psig 5,713
Original produced-gas/liquid ratio, scf/bbl
Production rate, B/D
Separator pressure, psig
Separator temperature, F
Liquid gravity at 60F, API
Datum, ft subsea 8,000
Well Characteristics
Elevation, ft KB 2,214
Total depth, ft 10,348
Producing interval, ft 10,124 to 10,176
Tubing size, in. 2
Tubing depth, ft 10,100
Open-flow potential, MMscf/D
Last reservoir pressure at 10,148 ft, psig 5,713
Date / /19 (m/d/y)
Reservoir temperature at 10,148 ft, F 186
Status of well status Shut in
Pressure gauge Amerada
Sampling Conditions
Flowing tubing pressure, psig 3,375
FBHP, psig 5,500
Primary-separator pressure, psig 300
Primary-separator temperature, F 62
Secondary-separator pressure, psig 20
Secondary-separator temperature, F 60
Field stock-tank-liquid gravity at 60F, API 58.5
Primary-separator-gas production rate, Mscf/D 762.14
Pressure base, psia 14.696
Temperature base, F 60
Compressibility factor, F
pv
1.043
Gas gravity (laboratory) 0.737
Gas-gravity factor, F
g
0.902
Stock-tank-liquid production rate at 60F, B/D 127.3
Primary-separator-gas/stock-tank-liquid ratio
In scf/bbl
In bbl/MMscf
5,987
167.0
Sampled by
For the example PVT report (Tables 6.5 and 6.6), the separator
gas/oil ratio, R
sp
, during sampling is reported as standard gas vol-
ume per separator-oil volume (4,428 scf/bbl). In this report, the units
are incorrectly labeled scf/bbl at 60F, where in fact the separator-oil
volume is measured at separator pressure (300 psig) and tempera-
ture (62F). The separator-oil formation volume factor (FVF), B
osp
,
is 1.352 bbl/STB and represents the volume of separator oil required
to yield 1 STB of oil (i.e., condensate).
The equation used to calculate wellstream composition, z
i
, is
z
i
+F
gsp
y
i
)(1 *F
gsp
) x
i
, (6.7) . . . . . . . . . . . . . . . . . . . . .
where F
gsp
+mole fraction of wellstream mixture that becomes
separator gas. In the laboratory report, F
gsp
is reported as primary-
separator gas/wellstream ratio (801.66 Mscf/MMscf), which is
equivalent to mole per mole ( F
gsp
+0.80166 mol/mol). The re-
ported value of F
gsp
can be checked with
F
gsp
+1 )
2, 130
osp
M
osp
R
sp
*1
, (6.8) . . . . . . . . . . . . . . . . . . . . .
where M
osp
+
N
i+1
x
i
M
i
. (6.9) . . . . . . . . . . . . . . . . . . . . . . . . . . .
osp
in lbm/ft
3
is calculated with a correlation (e.g., Standing-Katz
9
)
or with the relation (62.4g
o
)0.0136g
g
R
s
)B
o
, where R
s
and
B
o
+separator-oil values in scf/STB and bbl/STB, respectively;
TABLE 6.4WELLSTREAM (RESERVOIR-FLUID)
COMPOSITION FOR GOOD OIL CO. WELL 4
BOTTOMHOLE OIL SAMPLE
Component mol% wt%
Density*
(g/cm
3
) API*
Molecular
Weight
H
2
S Nil Nil
CO
2
0.91 0.43
N
2
0.16 0.05
Methane 36.47 6.24
Ethane 9.67 3.10
Propane 6.95 3.27
i-butane 1.44 0.89
n-butane 3.93 2.44
i-pentane 1.44 1.11
n-pentane 1.41 1.09
Hexanes 4.33 3.97
Heptanes plus 33.29 77.41 0.8515 34.5 218
Total 100.00 100.00
*At 60F.
g
o
+stock-tank-oil density; and g
g
+gravity of gas released from
the separator oil.
Finally, the value of stock-tank-liquid/wellstream ratio in bbl/MMscf
represents the separator barrels produced per 1 MMscf of wellstream.
In terms of F
gsp
and separator properties, this value equals
bbl
MMscf
+
470(1*F
gsp
)
M
osp
osp
B
osp
, (6.10) . . . . . . . . . . . . . .
where 470+(1 million scf/MMscf)/[(379 scf/lbm mol)(5.615 ft
3
/bbl)].
The separator-oil and -gas compositions can be checked for con-
sistency with the Hoffman et al.
10
K-value method and Standings
11
low-pressure K-value equations.
6.3 MultistageSeparator Test
The multistage-separator test is performed on an oil sample primari-
ly to provide a basis for converting differential-liberation data from
a residual-oil to a stock-tank-oil basis. Occasionally, several separa-
tor tests are run to help choose separator conditions that maximize
stock-tank-oil production. Usually, two or three stages of separation
are used, with the last stage at atmospheric pressure and near-ambi-
ent temperature (60 to 80F). The multistage-separator test can also
be conducted for high-liquid-yield gas-condensate fluids.
Fig. 6.2 illustrates schematically how the separator test is per-
formed. Initially, the reservoir sample is at saturation conditions and
the volume is measured ( V
ob
or V
gd
). The sample is then brought to
the pressure and temperature of the first-stage separator. All the gas
is removed, and the oil volume at the separator stage, V
osp
, is noted
together with the volume of removed gas, DV
g
; number of moles of
removed gas, n
g
; and specific gravity of removed gas, g
g
. If re-
quested, the gas samples can be analyzed chromatographically to
give molar composition, y.
The oil remaining after gas removal is brought to the conditions
of the next separator stage. The gas is removed again and quantified
by moles and specific gravity. Oil volume is noted, and the process
is repeated until stock-tank conditions are reached. Final oil volume,
V
o
, and specific gravity, g
o
, are measured at 60F.
Table 6.7 gives results from four separator tests, each consisting
of two stages of separation. The first-stage-separator pressure is var-
ied from 50 to 300 psig, and stock-tank conditions are held constant
at 0 psig and 75F. GORs are reported as standard gas volume per
separator-oil volume, R
sp
, and as standard gas volume per stock-
tank-oil volume, R
s
, respectively.
DR
sp
+
DV
g
V
osp
(6.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
and DR
s
+
DV
g
V
o
. (6.12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Total GOR is calculated by adding the stock-tank-oil-based GORs
from each separator stage.
R
s
+
N
sp
k+1
DR
s
k
. (6.13) . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Separator-oil FVFs, B
osp
, are reported as the ratio of separator-oil
volume to stock-tank-oil volume.
B
osp
+
V
osp
V
o
. (6.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Accordingly, the relation between separator gas/oil ratio and stock-
tank gas/oil ratio at a given stage is
DR
sp
+
DR
s
B
osp
. (6.15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Because B
osp
u1, it follows that R
sp
tR
s
.
Bubblepoint-oil FVF, B
ob
, is the ratio of bubblepoint-oil volume
to stock-tank-oil volume.
B
ob
+
V
ob
V
o
. (6.16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The average gas gravity, g
g
, is used in oil PVT correlations and
to calculate reservoir densities on the basis of black-oil properties.
The average gas gravity is calculated from
g
g
+
N
sp
k+1
gg
DR
s
N
sp
k+1
DR
s
k
, (6.17) . . . . . . . . . . . . . . . . . . . . . . . . . .
Fig. 6.1Procedure for recombining single-stage separator samples to obtain wellstream
composition of a bottomhole sample; BHS+bottomhole sampler, GC+gas chromatograph,
FDP+freezing-point depression, and DM+densitometer.
92 PHASE BEHAVIOR
TABLE 6.5SEPARATOR AND RECOMBINED WELLSTREAM COMPOSITIONS
FOR GOOD OIL CO. WELL 7 GAS CONDENSATE
Separator Products Hydrocarbon Analysis
Separator Liquid Separator Gas Wellstream
Component (mol%) (mol%) (gal/Mscf) (mol%) (gal/Mscf)
CO
2
Trace 0.22 0.18
N
2
Trace 0.16 0.13
Methane 7.78 75.31 61.92
Ethane 10.02 15.08 14.08
Propane 15.08 6.68 1.832 8.35 2.290
iso-butane 2.77 0.52 0.170 0.97 0.317
n-butane 11.39 1.44 0.453 3.41 1.073
iso-pentane 3.52 0.18 0.066 0.84 0.306
n-pentane 6.50 0.24 0.087 1.48 0.535
Hexanes 8.61 0.11 0.045 1.79 0.734
Heptanes plus 34.33 0.06 0.028 6.85 3.904
Total 100.00 100.00 2.681 100.00 9.159
Heptanes-Plus Properties
Oil gravity, API 46.6
Specific gravity
at 60/60F
0.7946 0.795
Molecular weight 143 103 143
Parameters
Calculated separator gas gravity (air+1.000) 0.735
Calculated gross heating value for separator gas at 14.696 psia and
60F, BTU/ft
3
dry gas
1,295
Primary-separator-gas*/-separator-liquid* ratio, scf/bbl at 60F 4,428
Primary-separator-gas/stock-tank-liquid ratio at 60F, bbl at 60F/bbl 1.352
Primary-separator-gas/wellstream ratio, Mscf/MMscf 801.66
Stock-tank-liquid/wellstream ratio, bbl/MMscf 133.9
*Primary separator gas and liquid collected at 300 psig and 62F.
TABLE 6.6MATERIAL-BALANCE CALCULATIONS FOR
GOOD OIL CO. WELL 7 GAS-CONDENSATE SAMPLE
Liquid Composition at Specified Pressures
(mol%)
Component At 3,500 psig At 2,900 psig At 2,100 psig At 1,300 psig At 605 psig
CO
2
0.18 0.18 0.18 0.15 0.08
N
2
0.13 0.08 0.06 0.03 0.01
C
1
13.18 45.04 32.22 19.69 11.77
C
2
8.12 14.05 13.99 12.32 7.44
C
3
12.59 9.67 11.25 11.66 9.31
i-C
4
3.44 1.14 1.59 1.85 1.64
n-C
4
5.21 4.82 6.12 7.35 7.17
i-C
5
2.67 1.25 1.77 2.43 2.79
n-C
5
5.74 2.16 3.48 4.62 5.50
C
6
8.47 3.11 4.55 6.40 8.37
C
7+
40.27 18.51 24.79 33.50 45.91
Total 100.00 100.00 100.00 100.00 100.00
M
o
, gmol
96.6 54.1 64.3 78.2 95.6
M
oC7)
, gmol 168.8 160.1 152.1 149.9 150.3
o
, gcm
3 0.3235 0.2642 0.1625 0.0892 0.0398
Fig. 6.2Schematic of a multistage-separator test.
p
st
+14.7 psia
T
st
+60F
where
g
g
k
+separator-gas gravity at Stage k. This relation is based
on the ideal gas law at standard conditions, where moles of gas are di-
rectly proportional with standard gas volume ( v
g
+379 scf/lbm mol).
Table 6.8 gives the composition of the first-stage-separator gas
at 50 psig and 75F. The gross heating value, H
g
, of this gas is calcu-
lated by Kays
12
mixing rule and component heating values, H
i
,
given in Table A-1.
H
g
+
N
i+1
y
i
H
i
. (6.18) . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Component liquid yields, L
i
, represent the liquid volumes of a
component or group of components that can theoretically be pro-
cessed from 1 Mscf of separator gas (gallons per million standard
cubic feet). L
i
can be calculated from
L
i
+19.73y
i
M
i
i
, (6.19) . . . . . . . . . . . . . . . . . . . . . . . . . . .
where M
i
+molecular weight and
i
+ component liquid density
in lbm/ft
3
at standard conditions (Table A-1). The C
7)
material in
separator gases is usually treated as normal heptane.
6.4 Constant Composition Expansion
6.4.1 Oil Samples. For an oil sample, the CCE experiment is used
to determine bubblepoint pressure, undersaturated-oil density, iso-
thermal oil compressibility, and two-phase volumetric behavior at
pressures below the bubblepoint. Table 6.9 presents data from an
example CCE experiment for a reservoir oil.
Fig. 6.3 illustrates the procedure for the CCE experiment. A blind
cell (i.e., a cell without a window) is filled with a known mass of reser-
voir fluid. Reservoir temperature is held constant during the experi-
ment. The sample initially is brought to a condition somewhat above
initial reservoir pressure, ensuring that the fluid is single phase. As the
pressure is lowered, oil volume expands and is recorded.
The fluid is agitated at each pressure by rotating the cell. This
avoids the phenomenon of supersaturation, or metastable equilibri-
um, where a mixture remains as a single phase even though it should
exist as two phases.
13-15
Sometimes supersaturation occurs 50 to
100 psi below actual bubblepoint pressure. By agitating the mixture
at each new pressure, the condition of supersaturation is avoided, al-
lowing more accurate determination of the bubblepoint.
Just below the bubblepoint, the measured volume will increase
more rapidly because gas evolves from the oil, yielding a higher sys-
tem compressibility. The total volume, V
t
, is recorded after the two-
phase mixture is brought to equilibrium. Pressure is lowered in steps
of 5 to 200 psi, where equilibrium is obtained at each pressure.
When the lowest pressure is reached, total volume is three to five
times larger than the original bubblepoint volume.
The recorded cell volumes are plotted vs. pressure, and the result-
ing curve should be similar to one of the curves in Fig. 6.4.
16
For a
black oil (far from its critical temperature), the discontinuity in vol-
ume at the bubblepoint is sharp and the bubblepoint pressure and
volume are easily read from the intersection of the p-V trends in the
single- and two-phase regions.
Volatile oils do not exhibit the same clear discontinuity in volu-
metric behavior at the bubblepoint pressure. Instead, the p-V curve
is practically continuous in the region of the bubblepoint because
the undersaturated-oil compressibility is similar to the effective
two-phase compressibility. This makes determining the bubble-
point of volatile oils in a blind cell difficult. Instead, a windowed cell
TABLE 6.7SEPARATOR TESTS (RESERVOIR-FLUID) OF
GOOD OIL CO. WELL 4 OIL SAMPLE
Separator
Pressure
(psia)
Separator
Temperature
(F)
GOR
b
(ft
3
/bbl)
GOR
c
(ft
3
/bbl)
Stock-Tank
Gravity
(API)
FVF
d
(bbl/bbl)
Separator
Volume
Factor
e
(bbl/bbl)
Flashed-Gas
Specific
Gravity
50
to
0
75
75
715
41
737
41 40.5 1.481
1.031
1.007
0.840
1.338
100
to
0
75
75
637
91
676
92 40.7 1.474
1.062
1.007
0.786
1.363
200
to
0
75
75
542
177
602
178 40.4 1.483
1.112
1.007
0.732
1.329
300
to
0
75
75
478
245
549
246 40.1 1.495
1.148
1.007
0.704
1.286
a
Gauge.
b
In cubic feet of gas at 60F and 14.65 psi absolute per barrel of oil at indicated pressure and temperature.
c
In cubic feet of gas at 60F and 14.65 psi absolute per barrel of stock-tank oil at 60F.
d
In barrels of saturated oil at 2,620 psi gauge and 220F per barrel of stock-tank oil at 60F.
e
In barrels of oil at indicated pressure and temperature per barrel of stock-tank oil at 60F.
94 PHASE BEHAVIOR
TABLE 6.8FIRST-STAGE SEPARATOR-GAS
COMPOSITION AND GROSS HEATING VALUE FOR
GOOD OIL CO. WELL 4 OIL SAMPLE*
Component mol% gal/Mscf
H
2
S Nil
CO
2
1.62
N
2
0.30
C
1
67.00
C
2
16.04 4.265
C
3
8.95 2.449
i-C
4
1.29 0.420
n-C
4
2.91 0.912
i-C
5
0.53 0.193
n-C
5
0.41 0.155
C
6
0.44 0.178
C
7+
0.49 0.221
Total 100.00 8.793
Heating Value
Calculated gas gravity (air+1.000) 0.840
Calculated gross heating value, BTU/ft
3
dry gas at 14.65 psia and 60F
1,405
*Collected at 50 psig and 75F in the laboratory.
is used to observe visually the first bubble of gas and the liquid vol-
umes below the bubblepoint.
Reported data from commercial laboratories usually include bub-
blepoint pressure, p
b
; bubblepoint density,
ob
, or specific volume,
v
ob
(v +1); and isothermal compressibility of the undersaturated
oil, c
o
, at pressures above the bubblepoint (Table 6.9). The table also
shows the oils thermal expansion, indicated by a ratio of undersatu-
rated-oil volume at a specific pressure and reservoir temperature to
the oil volume at the same pressure and a lower temperature.
Total volumes are reported relative to the bubblepoint volume.
V
rt
+
V
t
V
ob
. (6.20) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Traditionally, isothermal compressibility data are reported for pres-
sure intervals above the bubblepoint. In fact, the undersaturated-oil
compressibility varies continuously with pressure, and, because
V
t
+V
o
(V
rt
+V
ro
) for p up
b
, oil compressibility can be ex-
pressed as
c +
1
V
rt
V
rt
p
T
+
1
V
ro
V
ro
p
T
; p up
b
. (6.21) . . . . . . . . .
Fig. 6.3Schematic of a CCE experiment for an oil and a gas
condensate.
TABLE 6.9CCE DATA (RESERVOIR-FLUID)
FOR GOOD OIL CO. WELL 4 OIL SAMPLE
Saturation (bubblepoint) pressure*, psig 2,620
Specific volume at saturation
pressure*, ft
3
/lbm
0.02441
Thermal expansion of undersaturated
oil at 5,000 psi+V at 220F/V at 76F
1.08790
Compressibility of saturated oil at
reservoir temperature
From 5,000 to 4,000 psi, vol/vol-psi
13.48x10
6
15.88x10
6
18.75x10
6
Pressure/Volume Relations*
Pressure
(psig)
Relative volume
(L)
Y function
5,000 0.9639
4,500 0.9703
4,000 0.9771
3,500 0.9846
3,000 0.9929
2,900 0.9946
2,800 0.9964
2,700 0.9983
2,620** 1.0000
2,605 1.0022 2.574
2,591 1.0041 2.688
2,516 1.0154 2.673
2,401 1.0350 2.593
2,253 1.0645 2.510
2,090 1.1040 2.422
1,897 1.1633 2.316
1,698 1.2426 2.219
1,477 1.3618 2.118
1,292 1.5012 2.028
1,040 1.7802 1.920
830 2.1623 1.823
640 2.7513 1.727
472 3.7226 1.621
*
At 220F.
**
Saturation pressure.
1
Relative volume+V/V
sat
in barrels at indicated pressure per barrel at saturation
pressure.
Y function+( p
sat
*p)/(p
abs
)(V/V
sat
*1).
The V
rt
function at undersaturated conditions may be fit with a se-
conddegree polynomial, resulting in an explicit relation for under-
saturated-oil compressibility (see Chap. 3).
Total volumes below the bubblepoint can be correlated by the Y
function,
16,17
defined as
Y +
p
b
*p
p(V
rt
*1)
+
p
b
*p
p
V
t
V
b
*1
, (6.22) . . . . . . . . . . . . . .
where p and p
b
are given in absolute pressure units. As Fig. 6.5
shows, Y vs. pressure should plot as a straight line and the linear
trend can be used to smooth V
rt
data at pressures below the bubble-
point. Standing
16
and Clark
17
discuss other smoothing techniques
and corrections that may be necessary when reservoir conditions
and laboratory PVT conditions are not the same.
6.4.2 Gas-Condensate Samples. The CCE data for a gas condensate
usually include total relative volume, V
rt
, defined as the volume of
gas or of gas-plus-oil mixture divided by the dewpoint volume. Z fac-
Fig. 6.4Volume vs. pressure for an oil during a DLE test (after Standing
16
).
a
t

2
9
0

p
s
i
a
tors are reported at pressures greater than and equal to the dewpoint
pressure. Table 6.10 gives these data for a gas-condensate example.
Reciprocal wet-gas FVF, b
gw
, is reported at dewpoint and initial
reservoir pressures, where these values represent the gas equivalent
or wet-gas volume at standard conditions produced from 1 bbl of
reservoir gas volume.
b
gw
+
5.615 10
*3
T
sc
p
sc
p
ZT
+0.198
p
ZT
, (6.23) . . . . . . . .
with b
gw
in Mscf/bbl, p in psia, and T in R.
Most CCE experiments are conducted in a visual cell for gas con-
densates, and relative oil (condensate) volumes, V
ro
, are reported at
pressures below the dewpoint. V
ro
normally is defined as the oil vol-
ume divided by the total volume of gas and oil, although some re-
ports define it as the oil volume divided by the dewpoint volume.
6.5 Differential Liberation Expansion
The DLE experiment is designed to approximate the depletion pro-
cess of an oil reservoir
18
and thereby provide suitable PVT data to
calculate reservoir performance.
16,19-21
Fig. 6.6 illustrates the labo-
ratory procedure of a DLE experiment. Figs. 6.7A through 6.7C
and Table 6.11 give DLE data for an oil sample.
A blind cell is filled with an oil sample, which is brought to a
single phase at reservoir temperature. Pressure is decreased until the
fluid reaches its bubblepoint, where the oil volume, V
ob
, is recorded.
Because the initial mass of the sample is known, bubblepoint densi-
ty,
ob
, can be calculated.
The pressure is decreased below the bubblepoint, and the cell is
agitated until equilibrium is reached. All gas is removed at constant
pressure. Then, the volume, DV
g
; moles, Dn
g
; and specific gravity,
g
g
, of the removed gas are measured. The remaining oil volume, V
o
,
is also recorded. This procedure is repeated 10 to 15 times at de-
creasing pressures and finally at atmospheric pressure. Residual-oil
volume, V
or
, and specific gravity, g
or
, are measured at 60F.
Other properties are calculated on the basis of measured data
( DV
g
, V
o
, Dn
g
, g
g
, V
or
, and g
or
), including differential solution
gas/oil ratio, R
sd
; differential oil FVF, B
od
; oil density,
o
; and gas
Z factor, Z. For Stage k, these properties can be determined from
96 PHASE BEHAVIOR
Fig. 6.5PVT relation and plot of Y function for an oil sample at pressures below the bubblepoint.
Bubblepoint
Temperature
5F
80
163
185
205
Pressure
psia
1,970
2,437
2,520
2,615
Volume
cm
3
82.30
86.88
87.92
89.05
R
sd
k
+
k
j+1
379
Dn
g
j
V
or
, (6.24) . . . . . . . . . . . . . . . . . . . . . . . .
B
od
k
+
V
o
k
V
or
, (6.25) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
k
+
V
or
(62.4g
or
) )
k
j+1
28.975.615
Dn
g
g
g
V
o
k
+
350g
or
)
k
j+1
0.0764
DR
sd
g
g
j
5.615
B
od
k
,
(6.26) . . . . . . . . . . . . . . . . . .
and (Z)
k
+
1RT
pDV
g
Dn
g
k
, (6.27) . . . . . . . . . . . . . . . . . .
with V
or
and Vo in bbl, R
sd
in scf/bbl, B
od
in bbl/bbl, DV
g
in ft
3
, p
in psia, Dn
g
in lbm mol,
o
in lbm/ft
3
, and T in R. Note that the sub-
script j+1 indicates the final DLE stage at atmospheric pressure and
reservoir temperature. Reported oil densities are actually calculated
by material balance, not measured directly.
6.5.1 Converting From Differential to Stock-Tank Basis. Perhaps
the most important step in the application of oil PVT data for reservoir
calculations is conversion of the differential solution gas/oil ratio,
R
sd
, and oil FVF, B
od
, to a stock-tank-oil basis.
16,20
For engineering
calculations, volume factors, R
s
and B
o
, are used to relate reservoir-
oil volumes, V
o
, to produced surface volumes, V
g
and V
o
; i.e.,
R
s
+
V
g
V
o
(6.28) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
and B
o
+
V
o
V
o
. (6.29) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Differential properties R
sd
and B
od
reported in the DLE report are
relative to residual-oil volume (i.e., the oil volume at the end of the
DLE experiment, corrected from reservoir to standard temperature).
R
sd
+
V
g
V
or
(6.30) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
and B
od
+
V
o
V
or
. (6.31) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The equations commonly used to convert differential volume fac-
tors to a stock-tank basis are
R
s
+R
sb
*
R
sdb
*R
sd
B
ob
B
odb
(6.32) . . . . . . . . . . . . . . . . . .
and B
o
+B
od
B
ob
B
odb
, (6.33) . . . . . . . . . . . . . . . . . . . . . . . . . . .
where B
ob
+bubblepoint-oil FVF, R
sb
+solution gas/oil ratio
from a multistage-separator flash, and R
sdb
and B
odb
+differential
volume factors at the bubblepoint pressure. The term ( B
ob
B
odb
),
TABLE 6.10CCE DATA FOR GOOD OIL CO.
WELL 7 GAS-CONDENSATE SAMPLE
Pressure
(psig) Relative volume
Deviation Factor
Z
6,000 0.8808 1.144
5,713* 0.8948 1.107**
5,300 0.9158 1.051
5,000 0.9317 1.009
4,800 0.9434 0.981
4,600 0.9559 0.953
4,400 0.9690 0.924
4,300 0.9758 0.909
4,200 0.9832 0.895
4,100 0.9914 0.881
4,000
1.0000 0.867
3,905 1.0089
3,800 1.0194
3,710 1.0299
3,500 1.0559
3,300 1.0878
3,000 1.1496
2,705 1.2430
2,205 1.5246
1,605 2.1035
1,010 3.5665
Pressure/volume relations of reservoir fluid at 186F.
*
Reservoir pressure.
**
Gas FVF+1.591 Mscf/bbl.
Dewpoint pressure.
Gas FVF+1.424 Mscf/bbl.

representing the volume ratio, V
or
V
o
, is used to eliminate the resid-
ual-oil volume, V
or
, from the R
sd
and B
od
data. Note that the conver-
sion from differential to flash data depends on the separator
conditions because B
ob
and R
sb
depend on separator conditions.
Although, the conversions given by Eqs. 6.32 and 6.33 typically
are used, they are only approximate. The preferred method, as origi-
nally suggested by Dodson et al.,
22
requires that some equilibrium
oil be taken at each stage of the DLE experiment and flashed through
a multistage separator to give the volume ratios, R
s
and B
o
. This lab-
oratory procedure is costly and time-consuming and is seldom used.
However, the method is readily incorporated into an equation-of-
state (EOS) -based PVT program.
6.6 Constant Volume Depletion
The CVD experiment is designed to provide volumetric and com-
positional data for gas-condensate and volatile-oil reservoirs pro-
ducing by pressure depletion. Fig. 6.8 shows the stepwise procedure
of a CVD experiment schematically, and Figs. 6.9A through 6.9D
and Table 6.12 give CVD data for an example gas-condensate fluid.
The CVD experiment provides data that can be used directly by
the reservoir engineer, including (1) a reservoir material balance
that gives average reservoir pressure vs. recovery of total well-
stream (wet-gas recovery), sales gas, condensate, and natural gas
liquids; (2) produced-wellstream composition and surface products
vs. reservoir pressure; and (3) average oil saturation in the reservoir
(liquid dropout and revaporization) that occurs during pressure
depletion. For many gas-condensate reservoirs, the recoveries and
oil saturation vs. pressure data from the CVD analysis closely
approximate actual field performance for reservoirs producing by
pressure depletion. When other recovery mechanisms, such as wa-
terdrive and gas cycling, are considered, the basic data required for
reservoir engineering are still taken mainly from a CVD report. This
section provides a description of the data provided in a standard
Fig. 6.6Schematic of DLE experiment.
CVD analysis, ways to check the data for consistency,
23-25
and how
to extract reservoir-engineering quantities from the data.
23,26
Initially, the dewpoint, p
d
, or bubblepoint pressure, p
b
, of the res-
ervoir sample is established visually and the cell volume, V
cell
, at
saturated conditions is recorded. The pressure is then reduced by
300 to 800 psi and usually by smaller amounts (50 to 250 psi) just
below the saturation pressure of more-volatile systems. The cell is
agitated until equilibrium is achieved, and volumes V
o
and V
g
are
measured. At constant pressure, sufficient gas, DV
g
, is removed to
return the cell volume to the original saturated volume.
In the laboratory, the removed gas (wellstream) is brought to at-
mospheric conditions, where the amount of surface gas and conden-
sate are measured. Surface compositions y
g
and x
o
of the produced
surface volumes from the reservoir gas are measured, as are the vol-
umes DV
o
and DV
g
, densities
o
and
g
and oil molecular weight
M
o
. From these quantities, we can calculate the moles of gas re-
moved, Dn
g
.
Dn
g
+
DV
o
o
M
o
)
DV
g
379
. (6.34) . . . . . . . . . . . . . . . . . . . . . . . .
These data are reported as cumulative wellstream produced, n
p
, rel-
ative to the initial moles n.
n
p
n
k
+
1
n
k
j+1
(Dn
g
)
j
, (6.35) . . . . . . . . . . . . . . . . . . . . . . . . .
where j+1 corresponds to saturation pressure and (Dn
g
)
1
+0. The
initial amount (in moles) of the saturated fluid is known when the cell
is charged. The quantity n
p
n is usually reported as cumulative wet
gas produced in MMscf/MMscf, which is equivalent to mol/mol.
Surface compositions y
g
and x
o
of the removed reservoir gas and
properties of the removed gas are not reported directly in the labora-
tory report but are recombined to yield the equilibrium gas (well-
stream) composition, y
i
, which also represents the equilibrium gas
remaining in the cell. The C
7)
molecular weight of the wellstream,
MgC
7)
, is backcalculated from measured specific gravity
( g
w
+g
g
) and reservoir-gas composition, y. C
7)
specific gravity of
the produced gas, ggC
7)
, is also reported, but this value is calculated
from a correlation.
Knowing the cumulative moles removed and its volume occupied
as a single-phase gas at the removal pressure, we can calculate the
equilibrium gas Z factor from
Z +
pDV
g
Dn
g
RT
. (6.36) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
A two-phase Z factor is also reported that is calculated assum-
ing that the gas-condensate reservoir depletes according to the ma-
terial balance for a dry gas and that the initial condition of the reser-
voir is at dewpoint pressure.
98 PHASE BEHAVIOR
Fig. 6.7ADLE data for an oil sample from Good Oil Co. Well 4; differential solution gas/oil
ratio, R
sd
.
p
Z
2
+
p
d
Z
d
1 *
G
pw
G
w
, (6.37) . . . . . . . . . . . . . . . . . . . . . . . .
where G
pw
+cumulative wellstream (wet gas) produced and
G
w
+initial wet gas in place. As defined in Eq. 6.37, the term G
pw
G
w
equals n
p
n reported in the CVD report. From Eq. 6.37, the only un-
known at a given pressure is Z
2
, and the two-phase Z factor is then giv-
en by
Z
2
+
p
p
d
Z
d
1 *
n
p
n
. (6.38) . . . . . . . . . . . . . . . . . . . . .
Theoretical liquid yields, L
i
, are also reported for C
3)
through
C
5)
groups in the produced wellstreams at each pressure-depletion
stage. These values are calculated with
L
i
+19.73y
i
M
i
i
(6.39) . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
and by summing the yields of components in the particular plus
group. For example, the liquid yield of C
5)
material at CVD Stage
k is given by
L
C
5)
k
+
C
7)
j+i C
5
L
j
k
+19.73
C
7)
j+i C
5
y
j
M
j
. (6.40) . . . . .
Table 6.13 gives various calculated cumulative recoveries based
on the reservoir initially being at its dewpoint. The basis for the cal-
culations is 1 MMscf of dewpoint wet gas in place, G
w
; the corre-
sponding initial moles in place at dewpoint pressure is given by
n +
G
w
v
g
+
1 10
6
scf
379 scflbm mol
+2, 638 lbm mol. (6.41) . . . . . . . . .
The first row of recoveries (wellstream) simply represents the
cumulative moles produced, n
p
n, expressed as wet-gas volumes,
G
pw
, in Mscf.
G
pw
+nv
g
n
p
n
+(2, 638 lbm mol)
379 scflbm mol
1 10
3
Mscfscf
n
p
n
+ 1 10
3
n
p
n
. (6.42) . . . . . . . . . . . . . . . . . . . . . . . . .
Recoveries in Rows 2 through 4 (Normal Temperature Separa-
tion, Total Plant Products in Primary-Separator Gas, and Total Plant
Products in Second-Stage-Separator Gas) refer to production when
the reservoir is produced through a three-stage separator. Fig. 6.10
Fig. 6.7BDLE data for an oil sample from Good Oil Co. Well 4; differential oil FVF (relative
volume), B
od
.
illustrates the process schematically. The calculated recoveries are
based on multistage-separator calculations that use low-pressure K
values and a set of separator conditions chosen arbitrarily or speci-
fied when the PVT study is requested.
6.6.1 Recoveries: Normal Temperature Separation. Column
1: Initial in Place. In Column 1, Row 2a the stock-tank oil in solu-
tion in the initial dewpoint fluid (N+135.7 STB) is calculated by
flashing 1 MMscf of the original dewpoint fluid, G
w
, through a
multistage separator.
Rows 2b through 2d give the volumes of separator gas at each
stage of a three-stage flash of the initial dewpoint fluid: 757.87,
96.68, and 24.23 Mscf, respectively. The mole fraction of well-
stream resulting as a surface gas F
gg
is given by
F
gg
+
G
d
G
w
+
757.87 )96.68 )24.23 Mscflbm mol
1 10
3
scfMscf
379 scflbm mol
+0.8788 lbm mollbm mol, (6.43) . . . . . . . . . . . . . . .

where G
d
+total separator dry gas and the corresponding mole
fraction of stock-tank oil is 0.1212 mol/mol. F
gg
is used to calculate
dry-gas FVF (see Eq. 3.41). For the dewpoint pressure, this gives
B
gd
+
B
gw
F
gg
+
p
sc
T
sc
ZTp
F
gg
+
14.7520
[0.867(186 )460)]4015
0.8788
+4.487 10
*3
ft
3
scf. (6.44) . . . . . . . . . . . . . . . . . . .
The producing GOR of the dewpoint mixture for the specified
separator conditions can be calculated as
R
p
+
G
N
+
757.87 )96.68 )24.23 Mscflbm mol
1 10
3
scfMscf
135.7 STBlbm mol

+6, 476 scfSTB. (6.45) . . . . . . . . . . . . . . . . . . . . . . . .
The dewpoint solution oil/gas ratio, r
sd
, is simply the inverse of R
p
.
r
sd
+r
p
+
1
R
p
+1.544 10
*4
STBscf +154.4 STBMMscf.
(6.46) . . . . . . . . . . . . . . .
Note that specific gravities of stock-tank oil and separator gases are
not reported for the separator calculations.
100 PHASE BEHAVIOR
Fig. 6.7CDLE data for an oil sample from Good Oil Co. Well 4; oil viscosity, m
o
.
Column 2 and Higher. On the basis of 1 MMscf of initial dew-
point fluid, Rows 2a through 2d give cumulative volumes of separa-
tor products at each depletion pressure ( N
p
, G
p1
, G
p2
, and G
p3
).
The producing GOR of the wellstream produced during a depletion
stage is given by
R
p
k
+
G
p1
)G
p2
)G
p3
k
*
G
p1
)G
p2
)G
p3
k*1
N
p
k
*
N
p
k*1
.
(6.47) . . . . . . . . . . . . . . . . . .
For 2,100 psig, this gives
R
p
+
[(301.57 )20.75 )5.61) *(124.78 )12.09 )3.16)]
1 10
3
(24.0 *15.4)
+ 21, 850 scfSTB. (6.48) . . . . . . . . . . . . . . . . . . . . . . . .
In terms of the solution oil/gas ratio,
r
s
+r
p
+
1
R
p
+
1
21, 580 scfSTB
+4.58 10
*5
STBscf
+45.8 STBscf . (6.49) . . . . . . . . . . . . . . . . . . . . . . . . . .
At a given pressure, the mole fraction of the removed CVD gas
wellstream that becomes dry separator gas is given by
F
gg
k
+
G
p1
)G
p2
)G
p3
k
*
G
p1
)G
p2
)G
p3
k*1
G
w
n
p
n
k
*
n
p
n
k*1
.
(6.50) . . . . . . . . . . . . . . . . . .
For p+2,100 psig, this gives
F
gg
+[(301.57 )20.75 )5.61)*(124.78 )12.09
)3.16)]
1 10
3
1 10
6
(0.35096 *0.15438)
+0.9558 . (6.51) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The dry-gas FVF at 2,100 psig is
B
gd
+
14.7520
0.762(186 )460)2, 115
0.9558
+6.884 10
*3
ft
3
scf . (6.52) . . . . . . . . . . . . . . . . . . .
In summary, the information provided in the rows labeled Normal
Temperature Separation gives estimates of the condensate and
sales-gas recoveries assuming a multistage surface separation. For
example, at an abandonment pressure of 605 psig, the condensate
recovery is 35.1 STB of the 135.7 STB initially in place (in solution
in the dewpoint mixture), or 26% condensate recovery. Dry-gas re-
covery is (685.02)37.79)10.40)+733.21 Mscf of the 878.78
TABLE 6.11DLE DATA FOR GOOD OIL CO. WELL 4 OIL SAMPLE
Differential Vaporization
Pressure
(psig)
Solution
GOR
(scf/bbl
*
)
Relative
Oil Volume
(RB/bbl
*
)
Relative
Total Volume
(RB/bbl
*
)
Oil
Density
(g/cm
3
)
Deviation
Factor
Z
Gas FVF
(RB/bbl
*
)
Incremental
Gas Gravity
2,620 854 1.600 1.600 0.6562
2,350 763 1.554 1.665 0.6655 0.846 0.00685 0.825
2,100 684 1.515 1.748 0.6731 0.851 0.00771 0.818
1.850 612 1.479 1.859 0.6808 0.859 0.00882 0.797
1,600 544 1.445 2.016 0.6889 0.872 0.01034 0.791
1,350 479 1.412 2.244 0.6969 0.887 0.01245 0.794
1,110 416 1.382 2.593 0.7044 0.903 0.01552 0.809
850 354 1.351 3.169 0.7121 0.922 0.02042 0.831
600 292 1.320 4.254 0.7198 0.941 0.02931 0.881
350 223 1.283 6.975 0.7291 0.965 0.05065 0.988
159 157 1.244 14.693 0.7382 0.984 0.10834 1.213
0 0 1.075 0.7892 2.039
1.000
**
DLE Viscosity Data at 220F
Pressure
(psig)
Oil Viscosity
(cp)
Calculated Gas
Viscosity
(cp)
5,000 0.450
4,500 0.434
4,000 0.418
3,500 0.401
3,000 0.385
2,800 0.379
2,620 0.373
2,350 0.396 0.0191
2,100 0.417 0.0180
1,850 0.442 0.0169
1,600 0.469 0.0160
1,350 0.502 0.0151
1,100 0.542 0.0143
850 0.592 0.0135
600 0.654 0.0126
350 0.783 0.0121
159 0.855 0.0114
0 1.286 0.0093
Gravity of residual oil+35.1API at 60F.
*
Barrels of residual oil.
**
At 60F.
Mscf dry gas originally in place, or 83.4%. These recoveries can be
compared with the reported wet-gas (or molar) recovery of 76.787%
at 605 psig. In addition to recoveries, the calculated results in this
section can be used to calculate solution oil/gas ratio, r
s
, and dry-gas
FVF, B
gd
, for modified black-oil applications.
6.6.2 Recovery: Plant Products. Rows 3 through 5 consider
theoretical liquid recoveries for propane, butanes, and pentanes-
plus assuming 100% plant efficiency. Recoveries in Rows 3 and 4
are for the calculated separator gases from Stages 1 and 2 of the
three-stage surface separation. Recoveries in Row 5 are for the pro-
duced wellstreams from the CVD experiment and represent the ab-
solute maximum liquid recoveries that can be expected if the reser-
voir is produced by pressure depletion. Fig. 6.10 illustrates the
recovery calculations schematically. Liquid volumes (in gal/MMscf
of initial dewpoint fluid) at CVD Stage k are calculated from
(L
i
)
k
+ 19, 730
M
i
k
j+1
Dn
g
n
y
i
j
, (6.53) . . . . . . . . .
Fig. 6.8Schematic of CVD experiment.
102 PHASE BEHAVIOR
Fig. 6.9ACVD data for gas-condensate sample from Good Oil Co. Well 7; liquid-dropout curve, V
ro
.
where j +1 represents the dewpoint, y
i
+compositions of well-
stream entering the gas plant at various stages of depletion,
M
i
+component molecular weights, and
i
+ liquid component
densities in lbm/ft
3
at standard conditions (Table A-1).
Calculated liquid recoveries below the dewpoint use the moles of
wellstream produced ( Dn
g
n) and the compositions y
i
from the sep-
arator gas (Rows 3 and 4) or wellstream (Row 5) entering the plant.
Column 1 (Initial in Place) gives the total recoveries assuming that
the entire initial dewpoint fluid is taken to the surface and processed
[i.e., k +1 and (Dn
g
n)
1
+1 in Eq. 6.53].
Note that cumulative recovery of propanes from the first-stage
separator during depletion (1,276 gal) is larger than the liquid pro-
pane produced in the first-stage-separator gas of the original dew-
point mixture (1,198 gal). This means that the stock-tank oil from
the separation of original dewpoint mixture contains more propane
than the cumulative stock-tank-oil volumes produced by depletion
and three-stage separation.
The results given in Rows 3 and 4 cannot be calculated from re-
ported data because surface separator compositions from the three-
stage separation are not provided in the report. The results in Row
5 can be checked. As an example, consider the C
3
recoveries for the
initial-in-place fluid at 2,100 psig.
L
C
3
p
d
+19, 730
44.0931.66
(1)(0.0837)
+2, 299 galMMscf (6.54a) . . . . . . . . . . . . . . . . . . . .

and
L
C
3
2100
+19, 730
44.0931.66
[0.0825(0.05374)
)0.0810(0.15438 *0.05374)
)0.0757(0.35096 *0.15438)]
+754 galMMscf. (6.54b) . . . . . . . . . . . . . . . .
For the C
5)
recoveries at the dewpoint,
L
C
5)
p
d
+19, 730[(72.1538.96) (0.0091)
)(72.1539.36) (0.0152)
)(86.1741.43) (0.0179) )(14349.6) (0.0685)]
+5, 513 galMMscf . (6.55) . . . . . . . . . . . . . . . . .
6.6.3 Correcting Recoveries for Initial Pressure Greater Than
Dewpoint Pressure. All recoveries given in Table 6.13 assume that
the reservoir pressure is initially at dewpoint. This assumption is
made because initial reservoir pressure is not always known with
certainty when PVT calculations are made. However, adjusting re-
ported recoveries is straightforward when initial pressure is greater
than dewpoint pressure. With Q
Table
as recoveries given in Columns
2 and higher in Table 6.13, Q
d
as hydrocarbons in place in Column
Fig. 6.9BCVD data for gas-condensate sample from Good Oil Co. Well 7; equilibrium gas
compositions, y
i
.
D
e
w
p
o
i
n
t
P
r
e
s
s
u
r
e
1 at dewpoint pressure, and Q as actual cumulative recoveries based
on hydrocarbons in place at the initial pressure,
Q +Q
d
pZ
pZ
d
*
pZ
pZ
d
; p yp
d
, (6.56) . . . . . . . . . . . .
Q +Q
Table
)DQ
d
; p tp
d
, (6.57) . . . . . . . . . . . . . . . . . . .
and DQ
d
+Q
d
( pZ)
i
( pZ)
d
*1, (6.58) . . . . . . . . . . . . . . . . . . . .
where DQ
d
+additional recovery from initial to dewpoint pres-
sure.
For the example report,
DQ
d
+
5, 7281.107
4, 0150.867
*1Q
d
+0.1173Q
d
, (6.59) . . . . . . . . . . . . . . . . . . . . . . . . . . .
recalling that moles of material at dewpoint is 2,638 lbm mol, moles
of material at initial pressure of 5,728 psig is n +2, 638(1 )0.1173)
+2, 947 lbm mol, and the basis of calculations is G
w
+1.173
MMscf of wet gas in place at initial pressure of 5,728 psia.
The cumulative wellstream produced at the dewpoint pressure of
4,000 psig is 0.1173(1, 000) +117.3 Mscf. Recovery at 3,500 psig
is 117.3 )53.74 +171.0 Mscf. Likewise, wet-gas recovery
should be increased by 117.3 Mscf for all depletion pressures in the
CVD table.
For stock-tank-oil recovery, Q
d
+135.7 STB, so DQ
d
+15.9
STB. Stock-tank-oil recovery at 4,000 psig is 15.9 )0 +15.9
STB; at 3,500 psig the recovery should be 15.9 )6.4 +22.3 STB,
and so on.
On the basis of 1 MMscf wet gas at the dewpoint or 1.1173 MMscf
at initial reservoir pressure, the laboratory hydrocarbon pore vol-
ume (HCPV), V
pHClab
, is the same.
V
pHClab
+
G
w
B
gw
d
+
1 10
6
14.7
520
0.867(186 )460)
4, 015

+3, 943 ft
3
+
G
w
B
gw
i
+1.1173 10
6
14.7
520
1.107(186 )460)
5728

+3, 943 ft
3
. (6.60) . . . . . . . . . . . . . . . . . . . . . . . . . .
The actual HCPV of a reservoir is much larger than V
pHClab
, and the
conversion to obtain recoveries for the actual HCPV is simply
104 PHASE BEHAVIOR
Fig. 6.9CCVD data for gas-condensate sample from Good Oil Co. Well 7; equilibrium gas Z
factor, Z
g
.
Q
actual
+Q
lab
V
pHCactual
V
pHClab
, (6.61) . . . . . . . . . . . . . . . . . . . . . . .
where Q
lab
+laboratory value given by Eqs. 6.55 and 6.57. As an ex-
ample, suppose geological data indicate a HCPV of 625,000 bbl
(82.45 acre-ft), or 3.50910
6
ft
3
. Then, original wet gas in place is
G
w
+1.1173 10
6 3.509 10
6
3, 943
+994.3 MMscf (6.62) . . . . . . . . . . . . . . . . . . . . . . . . . . .
and condensate in solution at initial pressure is given by
N +135.7(1.1173)
3.509 10
6
3, 943
+134, 900 STB. (6.63) . . . . . . . . . . . . . . . . . . . . . . . . . .
6.6.4 Liquid-Dropout Curve. Table 6.11 and Figs. 6.9A through
6.9D show relative oil volumes, V
ro
, measured in the example CVD
experiment. V
ro
is defined as the volume of oil, V
o
, at a given pres-
sure divided by the original saturation volume, V
s
. This relative vol-
ume is an excellent measure of the average reservoir-oil saturation
(normalized) that will develop during depletion of a gas-condensate
reservoir. Correcting for water saturation, S
w
, the reservoir-oil satu-
ration can be calculated from V
ro
with
S
o
+(1 *S
w
)V
ro
. (6.64) . . . . . . . . . . . . . . . . . . . . . . . . . . .
For most gas condensates, V
ro
shows a maximum near 2,000 to
2,500 psia. Cho et al.
27
give a correlation for maximum liquid drop-
out as a function of temperature and C
7)
mole percent in the dew-
point mixture.
V
ro
max
+93.404 )4.799z
C
7)
*19.73 ln T, (6.65) . . . . . .
with zC
7)
in mole percent and T in F. The correlation predicts
(V
ro
)
max
+23.2% for the example condensate fluid compared with
24% measured experimentally (at 2,100 psig). Fig. 6.11 shows val-
ues of (V
ro
)
max
vs. T and zC
7)
from Eq. 6.65.
Considerable attention usually is given to matching the liquid-
dropout curve when an EOS is used. Some gas condensates have-
what is referred to as a tail, where liquid drops out very slowly
(sometimes for several thousand psi below the dewpoint) before fi-
nally increasing toward a maximum. Matching this behavior with
an EOS can prove difficult, and the question is whether matching the
tail is really necessary (see Appendix C).
What really matters for reservoir calculations of a gas-condensate
fluid is how much original stock-tank condensate is lost because
of retrograde condensation in the reservoir. The shape and magni-
Fig. 6.9DCVD data for gas-condensate sample from Good Oil Co. Well 7; wet-gas material
balance.
tude of liquid dropout reflects the change in producing oil/gas ratio,
r
p
[r
s
. A tail on a liquid-dropout curve implies that the producing
wellstream is becoming only slightly leaner (i.e., r
s
is decreasing
only slightly). The cumulative condensate recovery is given by
N
p
+
G
p
0
r
s
dG
p
, (6.66) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
where G
p
+cumulative dry gas produced. Cumulative condensate
production is readily evaluated from a plot of r
s
vs. G
p
.
One of the most important checks of an EOS characterization for
any gas condensate, particularly one with a tail, is N
p
calculated
from CVD data vs. N
p
calculated from the EOS characterization. It
is alarming how much the surface condensate recovery can be un-
derestimated if the tail is not matched properly. We do not recom-
mend matching the dewpoint exactly with a liquid-dropout curve
that is severely overpredicted in the region where measured results
indicate little dropout. If the EOS characterization cannot be modi-
fied to honor the tail of liquid-dropout curve, it is preferable to
underpredict the measured dewpoint pressure and match only the
higher liquid-dropout volumes.
In summary, oil relative volume, V
ro
, is not important per se; how-
ever, the effect of liquid dropout on surface condensate production
should be emphasized. In fact, the effect of shape and magnitude of
liquid dropout on fluid flow in the reservoir is negligible, and any
EOS match will probably have the same effect on fluid flow from the
reservoir into the wellbore (i.e., inflow performance).
6.6.5 Consistency Check of CVD Data. Reudelhuber and Hinds
24
give a detailed procedure for checking CVD data consistency that
involves a material-balance check on components and phases and
yields oil compositions, density, molecular weight, and MC
7)
. To-
gether with reported data, these calculated properties allow K values
to be calculated and checked for consistency with the Hoffman et
al.
10
method.
11,28
Whitson and Torps
23
material-balance equations
are summarized later. Similar equations can also be derived for a
DLE experiment when equilibrium gas compositions and oil rela-
tive volumes are reported. Reported CVD data include temperature,
T; dewpoint pressure, p
d
, or bubblepoint pressure, p
b
; dewpoint Z
factor, Z
d
, or bubblepoint-oil density,
ob
. Additional data at each
Depletion Stage k include oil relative volume, V
ro
; initial fraction
of cumulative moles produced, n
p
n; gas Z factor (not the two-
phase Z factor), Z; equilibrium gas composition, y
i
; and equilibrium
gas (wellstream) C
7)
molecular weight, Mg C
7)
.
The equilibrium gas density,
g
; molecular weight, M
g
; and well-
stream gravity, g
w
+M
g
M
air
, are readily calculated at each
106 PHASE BEHAVIOR
TABLE 6.12CVD DATA FOR GOOD OIL CO. WELL 7 GAS-CONDENSATE SAMPLE 2*
Reservoir Pressure, psig
Component, mol% 5,713** 4,000
3,500 2,900 2,100 1,300 605 0
CO
2
0.18 0.18 0.18 0.18 0.18 0.19 0.21
N
2
0.13 0.13 0.13 0.14 0.15 0.15 0.14
C
1
61.72 61.72 63.10 65.21 69.79 70.77 66.59
C
2
14.10 14.10 14.27 14.10 14.12 14.63 16.06
C
3
8.37 8.37 8.26 8.10 7.57 7.73 9.11
i-C
4
0.98 0.98 0.91 0.95 0.81 0.79 1.01
n-C
4
3.45 3.45 3.40 3.16 2.71 2.59 3.31
i-C
5
0.91 0.91 0.86 0.84 0.67 0.55 0.68
n-C
5
1.52 1.52 1.40 1.39 0.97 0.81 1.02
C
7
1.79 1.79 1.60 1.52 1.03 0.73 0.80
C
7+
6.85 6.85 5.90 4.41 2.00 1.06 1.07
Total 100.00 100.00 100.00 100.00 100.00 100.00 100.00
Properties
C
7+
molecular weight 143 143 138 128 116 111 110
C
7+
specific gravity 0.795 0.795 0.790 0.780 0.767 0.762 0.761
Equilibrium gas deviation factor, Z 1.107 0.867 0.799 0.748 0.762 0.819 0.902
Two-phase deviation factor, Z 1.107 0.867 0.802 0.744 0.704 0.671 0.576
Wellstream produced, cumulative
% of initial
0.000 5.374 15.438 35.096 57.695 76.787 93.515
From smooth compositions
C
3+
, gal/Mscf 9.218 9.218 8.476 7.174 5.171 4.490 5.307
C
4+
, gal/Mscf 6.922 6.922 6.224 4.980 3.095 2.370 2.808
C
5+
, gal/Mscf 5.519 5.519 4.876 3.692 1.978 1.294 1.437
Retrograde Condensation During Gas Depletion
Retrograde liquid volume,
% hydrocarbon pore space
0.0 3.3 19.4 23.9 22.5 18.1 12.6
*Study conducted at 186F.
** Original reservoir pressure.
Dewpoint pressure.
0-psig residual-liquid properties: 47.5API oil gravity at 60; 0.7897 specific gravity at 60/60F; and molecular weight of 140.
Depletion Stage k [and at the dewpoint ( k +1) for a gas-conden-
sate sample] from
M
g
k
+
N
i+1
(y
i
)
k
M
i
, (6.67) . . . . . . . . . . . . . . . . . . . . . . . . . .
k
+
p
M
g
k
(Z)
k
RT
, (6.68) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
and

g
g
k
+
g
w
k
+
M
g
k
28.97
. (6.69) . . . . . . . . . . . . . . . . . . . .
On a basis of 1 mol initial dewpoint fluid ( n +1), the cell vol-
ume is
V
cell
+
Z
d
RT
p
d
(6.70) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
for a gas condensate and
V
cell
+
M
ob
ob
(6.71) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
for a volatile oil. Oil and gas volumes, respectively, at Stage k are
(V
o
)
k
+V
cell
(V
ro
)
k
and

V
g
k
+V
cell
1 *(V
ro
)
k
. (6.72) . . . . . . . . . . . . . . . . . . . .
Moles and mass of the total material remaining in the cell at Stage k
are given by
(n
t
)
k
+1 *
n
p
n
k
,
n
g
k
+
V
g
k
(Z)
k
RT
,
and (n
o
)
k
+(n
t
)
k
*
n
g
k
, (6.73) . . . . . . . . . . . . . . . . . . . . . . . .
and moles and mass of the individual phases remaining in the cell at
Stage k are given by
(m
t
)
k
+M
s
*
k
j +2
Dn
g
n
M
g
j
,
m
g
k
+
n
g
M
g
k
,
and (m
o
)
k
+(m
t
)
k
*
m
g
k
. (6.74) . . . . . . . . . . . . . . . . . . . . . .
In Eqs. 6.73 and 6.74,
Dn
g
n
j
+
n
p
n
j
*
n
p
n
j*1
, (6.75) . . . . . . . . . . . . . . . . . . . .
M
s
+saturated-fluid molecular weight, and (n
p
n)
1
+0.
Densities and molecular weights of the oil phase are calculated from
TABLE 6.13CALCULATED RECOVERIES* FROM CVD REPORT
FOR GOOD OIL CO. WELL 7 GAS-CONDENSATE SAMPLE
Reservoir Pressure (psig)
Initial in Place 4,000** 3,500 2,900 2,100 1,300 605 0
Wellstream, Mscf 1,000 0 53.74 154.38 350.96 576.95 767.87 935.15
Normal temperature separation
Stock-tank liquid, bbl 135.7 0 6.4 15.4 24.0 29.7 35.1

Primary-separator gas, Mscf 757.87 0 41.95 124.78 301.57 512.32 658.02
Second-stage gas, Mscf 96.68 0 4.74 12.09 20.75 27.95 37.79
Stock-tank gas, Mscf 24.23 0 1.21 3.16 5.61 7.71 10.4
Total plant products in primary separator
Propane, gal 1,198 0 67 204 513 910 1,276

Butanes, gal 410 0 23 72 190 346 491
Pentanes, gal 180 0 10 31 81 144 192
Total plant products in second-stage
separator
Propane, gal 669 0 33 86 149 205 286

Butanes, gal 308 0 15 41 76 108 159
Pentanes, gal 138 0 7 19 35 49 69
Total plant products in wellstream
Propane, gal 2,296 0 121 342 750 1,229 1,706

Butanes, gal 1,403 0 73 202 422 665 927
Pentanes, gal 5,519 0 262 634 1,022 1,315 1,589
*
Cumulative recovery per MMscf of original fluid calculated during depletion.
**
Dewpoint pressure.
Recovery basis: primary separation at 500 psia and 70F, second-stage separation at 50 psia and 70F, and stock tank at 14.7 psia and 70F.
Recovery assumes 100% plant efficiency.
k
+
(m
o
)
k
(V
o
)
k
(6.76) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
and (M
o
)
k
+
(m
o
)
k
(n
o
)
k
, (6.77) . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
and the oil composition is given by
(x
i
)
k
+
(n
t
)
k
(z
i
)
k
*
n
g
y
i
k
(n
t
)
k
*
n
g
k
. (6.78) . . . . . . . . . . . . . . . . . . .
K values can be calculated from K
i
+y
i
x
i
, and z
i
+overall com-
position of the mixture remaining in the cell at Stage k.
(z
i
)
k
+
1
(n
t
)
k
(z
i
)
1
*
k
j+2
Dn
g
n
y
i
j
. (6.79) . . . . . . . . . . .
C
7)
molecular weight of the oil phase can be calculated from
M
o C
7)
k
+
(M
o
)
k
*
i0C
7)
(x
i
)
k
M
i
x
C
7)
k
. (6.80) . . . . . . . . . . . . . .
Table 6.6 summarizes these calculations for the sample gas-conden-
sate mixture.
Fig. 6.10Schematic of method of calculating plant recoveries in a CVD report for a gas
condensate.
(Separator Gas 1)
(Separator Gas 2)
108 PHASE BEHAVIOR
Fig. 6.11Calculated maximum retrograde oil relative volumes from the Cho et al.
27
correlation.
Heptanes Plus, mol%
Nonphysical
The oil composition at the last depletion state (605 psig for the ex-
ample condensate) can be measured, but it must be requested specif-
ically. Also, the residual-oil molecular weight, Mor, and specific
gravity
,
gor
,
remaining after depletion at atmospheric pressure are
typically measured and reported as shown in Table 6.12. These val-
ues can be compared with calculated values by use of the material-
balance equations shown earlier.
The material-balance calculations are more accurate for rich gas
condensates and volatile oils. In fact, obtaining reasonable material-
balance oil properties for lean gas condensates is difficult. Some-
times it is useful to modify the reported oil relative volumes (partic-
ularly those close to the dewpoint) to monitor the effect on
calculated oil properties.
An alternative material-balance check that may be even more
useful for determining data consistency (particularly for leaner gas
condensates) involves starting with reported final-stage condensate
composition, (x
i
)
k+N
, and adding back the removed gases, (y
i
)
k
, for
each stage from k +N to k +1. This results in the original gas
composition, (z
i
)
k+1
, which can be compared quantitatively with
the laboratory-reported composition.
References
1. Core Laboratories Good Oil Company Oil Well No. 4 PVT Study, Core
Laboratories, Houston.
2. Core Laboratories Good Oil Company Condensate Well No. 7 PVT
Study, Core Laboratories, Houston.
3. Flaitz, J.M. and Parks, A.S.: Sampling Gas-Condensate Wells, Trans.,
AIME (1942) 146, 13.
4. Katz, D.L., Brown, G.G., and Parks, A.S.: NGAA Report on Sampling
Two-Phase Gas Streams from High Pressure Condensate Wells, (Sep-
tember 1945).
5. Reudelhuber, F.O.: Sampling Procedures for Oil Reservoir Fluids, JPT
(December 1957) 15.
6. Clark, N.J.: Sampling and Testing Oil Reservoir Samples, JPT (Jan.
1962) 12.
7. Clark, N.J.: Sampling and Testing Gas Reservoir Samples, JPT
(March 1962) 266.
8. Recommended Practice for Sampling Petroleum Reservoir Fluids, API,
Dallas (1966) 44.
9. Standing, M.B. and Katz, D.L.: Density of Natural Gases, Trans.,
AIME, (1942) 146, 140.
10. Hoffmann, A.E., Crump, J.S., and Hocott, C.R.: Equilibrium Constants
for a Gas-Condensate System, Trans., AIME (1953) 198, 1.
11. Standing, M.B.: A Set of Equations for Computing Equilibrium Ratios
of a Crude Oil/Natural Gas System at Pressures Below 1,000 psia, JPT
(September 1979) 1193.
12. Kay, W.B.: The Ethane-Heptane System, Ind. & Eng. Chem. (1938)
30, 459.
13. Kennedy, H.T. and Olson, C.R.: Bubble Formation in Supersaturated
Hydrocarbon Mixtures, Oil & Gas J. (October 1952) 271.
14. Silvey, F.C., Reamer, H.H., and Sage, B.H.: Supersaturation in Hydrocar-
bon Systems: Methane-n-Decane, Ind. Eng. Chem. (1958) 3, No. 2, 181.
15. Tindy, R. and Raynal, M.: Are Test-Cell Saturation Pressures Accurate
Enough?, Oil & Gas J. (December 1966) 126.
16. Standing, M.B.: Volumetric and Phase Behavior of Oil Field Hydrocar-
bon Systems, eighth edition, SPE, Richardson, Texas (1977).
17. Clark, N.J.: Adjusting Oil Sample Data for Reservoir Studies, JPT
(February 1962) 143.
18. Moses, P.L.: Engineering Applications of Phase Behavior of Crude-Oil
and Condensate Systems, JPT (July 1986) 715.
19. Amyx, J.W., Bass, D.M. Jr., and Whiting, R.L.: Petroleum Reservoir En-
gineering, McGraw-Hill Book Co. Inc., New York City (1960).
20. Craft, B.C. and Hawkins, M.: Applied Petroleum Reservoir Engineering,
first edition, Prentice-Hall Inc., Englewood Cliffs, New Jersey (1959).
21. Dake, L.P.: Fundamentals of Reservoir Engineering, Elsevier Scientific
Publishing Co., Amsterdam (1978).
22. Dodson, C.R., Goodwill, D., and Mayer, E.H.: Application of Labora-
tory PVT Data to Reservoir Engineering Problems, Trans., AIME
(1953) 198, 287.
23. Whitson, C.H. and Torp, S.B.: Evaluating Constant-Volume-Depletion
Data, JPT (March 1983) 610; Trans., AIME, 275.
24. Drohm, J.K., Goldthorpe, W.H., and Trengove, R.: Enhancing the Eval-
uation of PVT Data, paper OSEA 88174 presented at the 1988 Offshore
Southeast Asia Conference, Singapore, 25 February.
25. Drohm, J.K., Trengove, R., and Goldthorpe, W.H.: On the Quality of
Data From Standard Gas-Condensate PVT Experiments, paper SPE
17768 presented at the 1988 Gas Technology Symposium, Dallas,
1315 June.
26. Reudelhuber, F.O. and Hinds, R.F.: Compositional Material Balance
Method for Prediction of Recovery From Volatile-Oil Depletion-Drive
Reservoirs, JPT (January 1957) 19; Trans., AIME, 210.
27. Cho, S.J., Civan, F., and Starling, K.E.: A Correlation To Predict Maxi-
mum Condensation for Retrograde Condensation Fluids and Its Use in
Pressure-Depletion Calculations, paper SPE 14268 presented at the
1985 SPE Annual Technical Conference and Exhibition, Las Vegas, Ne-
vada, 2225 September.
28. Clark, N.J.: Theoretical Aspects of Oil and Gas Equilibrium Calcula-
tions, JPT (April 1962) 373.
SI Metric Conversion Factors
API 141.5/(131.5)API) +g/cm
3
bbl 1.589 873 E*01+m
3
Btu 1.055 056 E)00+kJ
cp 1.0* E*03+Pa@s
ft 3.048* E*01+m
ft
3
2.831 685 E*02+m
3
F (F*32)/1.8 +C
gal 3.785 412 E*03+m
3
in. 2.54* E)00+cm
lbm mol 4.535 924 E*01+kmol
psi 6.894 757 E)00+kPa
*Conversion factor is exact.
Petrleos Voltiles
Fluidos de Reservorios
G. Fondevila
Diagrama de Fases
Caractersticas
Temperatura Reservorio < T
C
GOR: 350 600 m3/m3 (2000 3200 scf/stb)
Densidad > 40 API (generalmente 45 55API)
mol% C
7+
: 12.5 20 %
Bob > 2
Gran cantidad de volumen de gas liberado una vez
alcanzada la presin de saturacin.
Gas liberado por Petrleos Voltiles es de naturaleza
Retrgrada: no puedo hacer un balance de masa
convencional (ya que debo tener en cuenta el
condensado aportado por el gas)
PVT Petrleo Voltil
Mismos Pasos que un PVT Normal, pero adems:
Liberacin a Volumen Constante (CVD - Constant Volume Depletion):
Similar a la liberacin diferencial, pero solo libero gas hasta llegar al volumen
inicial.
Mejor manera de representar a lo que ocurre en el reservorio (variacin
composicional del fluido).
El camino termodinmico afecta a las propiedades del fluido.
Se recomienda para este tipo de Petrleos realizar la tcnica de Dodson para
PVT: Para cada etapa diferencial, realizar el ensayo flash a condiciones de
separador. Esta tcnica requiere de mucho volumen de muestra inicial, es
costosa y lleva mucho tiempo. Esto es debido a que la presin ptima de
separador vara durante el tiempo de explotacin.
En los Ensayos de Separador, se prueban adems mltiples etapas de
separacin (2 o ms).
En petrleos voltiles, para maximizar la produccin de lquido se emplea el
uso de mltiples etapas de separacin (2+).
Curvas PVT Tpicas
Bo Rs [m
3
/m
3
]
P [kg/cm
2
] P [kg/cm
2
]
Liberacin Volumen Constante
Liq
Liq
Gas
Liq
Gas
P
Pi P1 >
Bajo la presin hasta P1.
Hay liberacin de gas, por lo que P1 < Pb.
Saco gas hasta llegar al Volumen Inicial
P1
Vol Cte
Diagrama de Fases
Bajo P, Alta liberacin de gas
Pi
Temp Res
Temperatura Reservorio
muy cercana a Tc
50% Vol liq !!
Variacin Bo/Bob segn el tipo de Petrleo
Tipo A o B Tipo C Tipo D
Gas y Condensado
Diagrama de Fases
Zona de condensacin retrgrada
Caractersticas
T
CRIT
< T
RES
< T
CT
GOR: 600 2700 m3/m3
Densidad lquido > 40 API (generalmente > 50API)
mol% C
7+
: 4 12.5 %
Cuando la presin del reservorio se encuentra por debajo del punto
de roco, se produce una condensacin de lquido en fondo.
El condensado liberado en fondo debe alcanzar una saturacin
crtica mnima para poder fluir, por lo que queda inmvil en fondo.
Esta condensacin produce una disminucin de la permeabilidad
relativa al gas por lo que la productividad del reservorio disminuye.
Se necesitan altas saturaciones de lquido en fondo para que el
mismo se vuelva mvil, por lo que la mayora del condensado
queda en fondo.
PVT
Recombinacin: Recordemos que por la naturaleza de los Gases
Retrgrados, los mismos deben ser muestreados en superficie y
luego las muestras son recombinadas. El pozo debe ser producido
a un mnimo caudal para mantener la presin de fluencia (si es
posible) por encima del Pd, pero este caudal debe ser mayor al
crtico, que nos permita desplazar el condensado generado dentro
del pozo (resultado: ahogamos el pozo, muestra no representativa
de gas menos rico).
Relacin PV: Para obtener el Punto de Roco. Celda visual PVT.
Lquido Retrgrado: Se obtiene la curva de produccin de lquido
en funcin de la deplecin. Esta curva se obtiene en liberacin a
masa constante (flash) y liberacin a volumen constante.
CVD (Deplecin a Volumen Constante): Tipo de expansin que
representa el tipo de deplecin sufrida en el reservorio. Esto es para
tener en cuenta el condensado inmvil que queda en el reservorio
en contacto con el gas.
Expansin Flash Curva PV: Determinacin de Pr (visual)
PVT: Gas y Condensado
Composicin del efluente a distintas P
Curva de produccin de lquido en una LVC
Comparacin de Recupero de Lquido: CCE vs CVD
Comparacin de Recupero de Lquido: CCE vs CVD
Reciclaje de Gas en Gas y Condensado (1)
Reciclaje de Gas en Gas y Condensado (2)
Z bifsico
- Se calcula a partir de una
liberacin CVD:
Z2f(P) = P / [Pd/Zd * (1 Gp/GOIS)]
- Donde:
-Z2f: Z bifsico
-P: Presin
-Pd: Punto de Roco
-Zd: Factor de desviacin en Pd
-Gp: Gas producido acumulado a CS
-GOIS: Gas Original In Situ a CS
Bloqueo por lquido en Gas y Condensado
Client: Lucky Offshore Field: Mississippi Canyon 999
Well: OCS-G-999 No. BBB

Table 3: Summary of Results of the Selected Sample
(Sample 1.01)

Reservoir Conditions
Pressure (p
i
) 10,000 Psia
Temperature (T
i
) 204 F

Properties
OBM Contamination Wt%

Dew Point Pressure (p
d
)
At T
i
5,975 psia
150F 5,917 psia
100F 5,699 psia

Gas-Oil Ratio
Single-stage Flash: 18,640 scf/stb

Properties at 60F STO API Gas Gravity (Average)
Single-stage Flash: 43.2 0.784

Properties at Reservoir Conditions
Viscosity: cP
Compressibillity (C
o
): 28.8 10
-6
/psi
Density: 0.417 g/cm
3

Z Factor: 1.484

Properties at Saturation Conditions
Viscosity: cP
Compressibillity (C
o
): 69.6 10
-6
/psi
Density: 0.345 g/cm
3

Z Factor: 1.072

Total Depletion Recovery
Abandonment Pressure: 1000 psia
Wellstream Recovery 77.62%

Oilphase - DBR Report # NAM 1112
8

Table 4: C30+ Composition, GOR, API of Selected Sample (Sample 1.01)
Component Flashed Gas Flashed Liquid Monophasic Fluid
MW WT % MOLE % WT % MOLE % WT % MOLE %

N2 28.01 0.62 0.50 0.00 0.00 0.50 0.49
CO2 44.01 12.75 6.57 0.02 0.09 10.17 6.36
H2S 34.08 0.00 0.00 0.00 0.00 0.00 0.00
C1 16.04 53.22 75.30 0.04 0.40 42.46 72.79
C2 30.07 12.70 9.59 0.06 0.32 10.14 9.28
C3 44.10 7.63 3.93 0.13 0.49 6.12 3.81
I C4 58.12 2.05 0.80 0.09 0.26 1.65 0.78
N C4 58.12 3.03 1.18 0.20 0.56 2.45 1.16
I C5 72.15 1.40 0.44 0.23 0.53 1.16 0.44
N C5 72.15 1.14 0.36 0.26 0.60 0.96 0.37
C6 84.00 3.28 0.86 2.51 4.84 3.13 1.00
C7 96.00 1.37 0.31 6.33 10.50 2.37 0.65
C8 107.00 0.54 0.11 7.48 10.88 1.94 0.47
C9 121.00 0.12 0.02 4.90 6.35 1.09 0.23
C10 134.00 0.09 0.02 5.87 7.28 1.26 0.26
C11 147.00 0.04 0.01 6.19 7.00 1.28 0.24
C12 161.00 0.02 0.00 7.06 7.29 1.44 0.25
C13 175.00 0.01 0.00 8.12 7.71 1.65 0.26
C14 190.00 0.00 0.00 8.40 7.35 1.70 0.25
C15 206.00 0.00 0.00 10.05 8.11 2.03 0.27
C16 222.00 0.00 0.00 5.29 3.96 1.07 0.13
C17 237.00 0.00 0.00 4.84 3.39 0.98 0.11
C18 251.00 0.00 0.00 6.83 4.53 1.38 0.15
C19 263.00 0.00 0.00 3.88 2.45 0.79 0.08
C20 275.00 0.00 0.00 1.87 1.13 0.38 0.04
C21 291.00 0.00 0.00 1.32 0.75 0.27 0.03
C22 300.00 0.00 0.00 0.69 0.38 0.14 0.01
C23 312.00 0.00 0.00 0.72 0.38 0.15 0.01
C24 324.00 0.00 0.00 0.75 0.38 0.15 0.01
C25 337.00 0.00 0.00 0.39 0.19 0.08 0.01
C26 349.00 0.00 0.00 0.41 0.19 0.08 0.01
C27 360.00 0.00 0.00 0.42 0.19 0.09 0.01
C28 372.00 0.00 0.00 0.44 0.19 0.09 0.01
C29 382.00 0.00 0.00 0.91 0.38 0.18 0.01
C30+ 580.00 0.00 0.00 3.29 0.94 0.67 0.03

Calculated MW 22.7 166 27.5
Mole % 3.35 96.65

OBM Contamination Level (wt%) STO Basis
Live Oil Basis

Stock Tank Oil Properties at Standard Conditions: C30+ Properties
Measured Calculated
MW 166 166 580
Density (g/cm
3
) 0.810 0.810 0.950

Single Stage Flash Data
Original STO Corrected
GOR (scf/stb) 18640
STO Density (g/cm
3
) 0.810
STO API Gravity 43.2

OBM Density (g/cm
3
) @60F

9

Table 5: Calculated Group Properties for Selected Sample (Sample 1.01)

Properties Flashed Gas Flashed Liquid Monophasic Fluid

Mole %
C7+ 0.46 91.91 3.52
C12+ 0.00 49.89 1.673
C20+ 0.00 5.09 0.170
C30+ 0.00 0.94 0.032

Mass %
C7+ 2.18 96.46 21.258
C12+ 0.03 65.70 13.314
C20+ 0.00 11.23 2.271
C30+ 0.00 3.29 0.666

Molar Mass
C7+ 102.23 172.39 163.48
C12+ 168.60 218.44 218.34
C20+ - 361.78 361.78
C30+ - 580.00 580.00

Density
C7+ 0.739 0.816 0.809
C12+ 0.810 0.844 0.844
C20+ - 0.900 0.900
C30+ - 0.950 0.950
STO at 60
o
F 0.810

Gas Gravity (Air = 1) 0.784

Dry Gross Heat Content (BTU/scf) 3659
Wet Gross Heat Content (BTU/scf) 3595

10

Table 6: Constant Composition Expansion at T
i

(Sample 1.01)

Pressure Relative Vol % Liquid % Liquid Bulk Compressibilty Z
(psia) (V
r
=V
t
/V
s
) (V
L
/V
t
) (V
L
/V
b
) Density (g/cm
3
) (10
-6
/psia) Factor

p
i
10000 0.828 0 0.417 28.78 1.484
9500 0.841 0 0.410 32.35 1.433
9000 0.856 0 0.403 36.32 1.382
8000 0.891 0 0.387 45.53 1.279
7000 0.937 0 0.368 56.54 1.176
p
d
5975 1.000 0 0 0.345 69.58 1.072
5500 1.039 0.22 0.23 0.332
5450 1.044 0.24 0.25 0.330
5000 1.090 0.53 0.58 0.316
4500 1.156 1.83 2.12 0.298
4000 1.251 4.68 5.85 0.276
3500 1.384 6.53 9.04 0.249
3000 1.576 6.92 10.91 0.219
2500 1.866 6.33 11.81 0.185
2000 2.329 5.18 12.06 0.148
1800 2.597 4.62 12.00 0.133

11

Figure 1: Constant Composition Expansion at T
i
Relative Volume
(Sample 1.01)

0.6
0.8
1
1.2
1.4
1.6
1.8
2
2.2
2.4
2.6
2.8
0 2000 4000 6000 8000 10000 12000
Pressure (psia)
R
e
l
a
t
i
v
e

V
o
l
u
m
e

(
V
t
/
V
t

@

P
b
)
Figure 2: Constant Composition Expansion at T
i
Liquid Volume %
(SAMPLE 1.01)
0
2
4
6
8
10
12
14
0 1000 2000 3000 4000 5000 6000 7000 8000
Pressure (psia)
L
i
q
u
i
d

V
o
l
u
m
e

%
Vl/Vt
Vl/Vd

12

Table 7: Constant Composition Expansion at 150F
(Sample 1.01)

(psia) (V
r
=V
t
/V
s
) (V
L
/V
t
) (V
L
/V
b
) Density (g/cm
3
) (10
-6
/psia) Factor

p
i
10000 0.852 0 0.441 24.14 1.526
9500 0.864 0 0.435 27.03 1.469
9000 0.877 0 0.429 30.24 1.413
8000 0.907 0 0.415 37.67 1.299
7000 0.945 0 0.398 46.58 1.184
p
d
5917 1.000 0 0 0.376 57.88 1.060
5500 1.028 0.21 0.22 0.366
5450 1.032 0.23 0.24 0.364
5000 1.069 0.56 0.60 0.352
4500 1.123 2.59 2.91 0.335
4000 1.202 7.22 8.68 0.313
3500 1.317 9.76 12.85 0.286
3000 1.486 10.08 14.97 0.253
2500 1.747 9.06 15.83 0.215
2000 2.181 7.29 15.90 0.172
1800 2.437 6.46 15.74 0.154

13

Figure 3: Constant Composition Expansion at 150F Relative Volume
(Sample 1.01)

0.6
0.8
1
1.2
1.4
1.6
1.8
2
2.2
2.4
2.6
0 2000 4000 6000 8000 10000 12000
Pressure (psia)
R
e
l
a
t
i
v
e

V
o
l
u
m
e

(
V
t
/
V
t

@

P
b
)
Figure 4: Constant Composition Expansion at 150F Liquid Volume %
(SAMPLE 1.01)
0
2
4
6
8
10
12
14
16
18
20
0 1000 2000 3000 4000 5000 6000 7000 8000
Pressure (psia)
L
i
q
u
i
d

V
o
l
u
m
e

%
Vl/Vt
Vl/Vd

14

Table 8: Constant Composition Expansion at 100F
(Sample 1.01)

(psia) (V
r
=V
t
/V
s
) (V
L
/V
t
) (V
L
/V
b
) Density (g/cm
3
) (10
-6
/psia) Factor

p
i
10000 0.869 0 0.465 19.99 1.578
9500 0.879 0 0.460 22.26 1.516
9000 0.890 0 0.454 24.77 1.454
8000 0.915 0 0.442 30.57 1.328
7000 0.946 0 0.427 37.51 1.201
p
d
5699 1.000 0 0 0.404 48.33 1.034
5500 1.011 0.10 0.10 0.400
5450 1.014 0.13 0.13 0.399
5000 1.041 0.38 0.40 0.388
4500 1.080 1.59 1.72 0.374
4000 1.139 8.30 9.46 0.355
3500 1.229 13.02 16.00 0.329
3000 1.364 13.96 19.04 0.296
2500 1.583 12.77 20.21 0.255
2000 1.964 10.32 20.27 0.206
1800 2.197 9.12 20.04 0.184

15

Figure 5: Constant Composition Expansion at 100F Relative Volume
(Sample 1.01)

0.6
0.8
1
1.2
1.4
1.6
1.8
2
2.2
2.4
0 2000 4000 6000 8000 10000 12000
Pressure (psia)
R
e
l
a
t
i
v
e

V
o
l
u
m
e

(
V
t
/
V
t

@

P
b
)
Figure 6: Constant Composition Expansion at 100F Liquid Volume %
(SAMPLE 1.01)
0
2
4
6
8
10
12
14
16
18
20
22
24
0 1000 2000 3000 4000 5000 6000 7000 8000
Pressure (psia)
L
i
q
u
i
d

V
o
l
u
m
e

%
Vl/Vt
Vl/Vd

16

Table 9: Constant Volume Depletion at T
i

(Sample 1.01)

Pressure Total Produced Vapor Properties Calculated Liquid Properties Two Phase
Recovery Z Factor MW Density Volume % Density Z Factor*
psia (%) (g/gmol) (g/cm
3
) (g/cm
3
)

5975 0.00 1.067 27.39 0.345 0.00 1.067
5200 6.34 0.991 27.21 0.321 0.40 0.784 0.992
4500 13.66 0.926 26.72 0.292 2.03 0.615 0.931
3800 23.11 0.878 25.29 0.246 6.48 0.590 0.882
3200 32.96 0.854 24.17 0.204 8.69 0.603 0.853
2700 42.28 0.845 23.47 0.169 9.38 0.619 0.837
2100 54.25 0.848 22.89 0.127 9.37 0.651 0.826
1500 66.94 0.869 22.57 0.088 8.90 0.681 0.827
1000 77.63 0.899 22.55 0.056 8.29 0.704 0.831

*Z = (Z
g
*V + Z
L
*L), where V and L are mole fractions of vapor and liquid at each stage (i,e., after displacing the
excess gas to achieve constant volume)

17

Figure 7: Constant Volume Depletion Total Recovery %
(Sample 1.01)

0
10
20
30
40
50
60
70
80
90
0 1000 2000 3000 4000 5000 6000 7000
Pressure (psia)
T
o
t
a
l

R
e
c
o
v
e
r
y

(
%
)
Figure 8: Constant Volume Depletion Retrograde Liquid Deposit
(Sample 1.01)

0
1
2
3
4
5
6
7
8
9
10
0 1000 2000 3000 4000 5000 6000 7000
Pressure (psia)
R
e
t
r
o
g
r
a
d
e

L
i
q
u
i
d

D
e
p
o
s
i
t

(
%
)

18

Figure 9: Constant Volume Depletion Vapor Deviation Factor Z
(Sample 1.01)

0.8
0.85
0.9
0.95
1
1.05
1.1
0 1000 2000 3000 4000 5000 6000 7000
Pressure (psia)
G
a
s

Z

F
a
c
t
o
r

Figure 10: Constant Volume Depletion Vapor Molecular Mass
(Sample 1.01)
20
21
22
23
24
25
26
27
28
0 1000 2000 3000 4000 5000 6000 7000
Pressure (psia)
G
a
s

M
W

(
g
/
g
m
o
l
)

19

Table 10: Constant Volume Depletion Vapor Composition
(Sample 1.01)

Component p-psia
5200 4500 3800 3200 2700 2100
MW Mol%

N2 28.01 0.49 0.49 0.50 0.51 0.51 0.51
CO2 44.01 6.36 6.37 6.42 6.47 6.50 6.55
H2S 34.08 0.00 0.00 0.00 0.00 0.00 0.00
C1 16.04 72.85 73.12 74.05 74.86 75.39 75.82
C2 30.07 9.28 9.28 9.27 9.27 9.29 9.34
C3 44.10 3.81 3.80 3.75 3.70 3.68 3.68
I C4 58.12 0.78 0.78 0.76 0.74 0.73 0.72
N C4 58.12 1.16 1.15 1.12 1.09 1.06 1.04
I C5 72.15 0.44 0.44 0.42 0.40 0.39 0.37
N C5 72.15 0.37 0.36 0.34 0.33 0.31 0.30
C6 84.00 0.99 0.98 0.91 0.84 0.78 0.71
C7 96.00 0.65 0.63 0.57 0.49 0.43 0.36
C8 107.00 0.47 0.45 0.39 0.33 0.27 0.21
C9 121.00 0.23 0.22 0.19 0.15 0.12 0.08
C10 134.00 0.26 0.25 0.21 0.16 0.12 0.08
C11 147.00 0.24 0.23 0.18 0.14 0.10 0.06
C12 161.00 0.24 0.23 0.18 0.12 0.08 0.05
C13 175.00 0.26 0.24 0.18 0.12 0.07 0.04
C14 190.00 0.24 0.23 0.16 0.10 0.06 0.03
C15 206.00 0.27 0.25 0.16 0.09 0.05 0.02
C16 222.00 0.13 0.12 0.07 0.04 0.02 0.01
C17 237.00 0.11 0.10 0.05 0.03 0.01 0.00
C18 251.00 0.15 0.13 0.07 0.03 0.01 0.00
C19 263.00 0.08 0.07 0.03 0.01 0.01 0.00
C20 275.00 0.04 0.03 0.01 0.00 0.00 0.00
C21 291.00 0.02 0.02 0.01 0.00 0.00 0.00
C22 300.00 0.01 0.01 0.00 0.00 0.00 0.00
C23 312.00 0.01 0.01 0.00 0.00 0.00 0.00
C24 324.00 0.01 0.01 0.00 0.00 0.00 0.00
C25 337.00 0.01 0.00 0.00 0.00 0.00 0.00
C26 349.00 0.01 0.00 0.00 0.00 0.00 0.00
C27 360.00 0.01 0.00 0.00 0.00 0.00 0.00
C28 372.00 0.01 0.00 0.00 0.00 0.00 0.00
C29 382.00 0.01 0.01 0.00 0.00 0.00 0.00
C30+ 580.00 0.01 0.00 0.00 0.00 0.00 0.00

Total 100.00 100.00 100.00 100.00 100.00 100.00
Calculated MW 27.21 26.72 25.29 24.17 23.47 22.89

Viscosity (cP) 0.030 0.027 0.024 0.022 0.020 0.017
Heat Content (BTU/scf) - Dry 1450 1423 1344 1282 1243 1210

20


Table 11: Constant Volume Depletion Vapor Composition Table 11: Constant Volume Depletion Vapor Composition
(Sample 1.01) (Sample 1.01)

Component Component p-psia p-psia
1500 1500 1000 1000
MW MW Mol% Mol%

N2 N2 28.01 28.01 0.51 0.51 0.50 0.50
CO2 CO2 44.01 44.01 6.59 6.59 6.61 6.61
H2S H2S 34.08 34.08 0.00 0.00 0.00 0.00
C1 C1 16.04 16.04 75.95 75.95 75.66 75.66
C2 C2 30.07 30.07 9.45 9.45 9.60 9.60
C3 C3 44.10 44.10 3.73 3.73 3.85 3.85
I C4 I C4 58.12 58.12 0.73 0.73 0.76 0.76
N C4 N C4 58.12 58.12 1.05 1.05 1.09 1.09
I C5 I C5 72.15 72.15 0.37 0.37 0.38 0.38
N C5 N C5 72.15 72.15 0.29 0.29 0.30 0.30
C6 C6 84.00 84.00 0.66 0.66 0.66 0.66
C7 C7 96.00 96.00 0.30 0.30 0.28 0.28
C8 C8 107.00 107.00 0.16 0.16 0.14 0.14
C9 C9 121.00 121.00 0.06 0.06 0.05 0.05
C10 C10 134.00 134.00 0.06 0.06 0.04 0.04
C11 C11 147.00 147.00 0.04 0.04 0.03 0.03
C12 C12 161.00 161.00 0.03 0.03 0.02 0.02
C13 C13 175.00 175.00 0.02 0.02 0.01 0.01
C14 C14 190.00 190.00 0.01 0.01 0.01 0.01
C15 C15 206.00 206.00 0.01 0.01 0.00 0.00
C16 C16 222.00 222.00 0.00 0.00 0.00 0.00
C17 C17 237.00 237.00 0.00 0.00 0.00 0.00
C18 C18 251.00 251.00 0.00 0.00 0.00 0.00
C19 C19 263.00 263.00 0.00 0.00 0.00 0.00
C20 C20 275.00 275.00 0.00 0.00 0.00 0.00
C21 C21 291.00 291.00 0.00 0.00 0.00 0.00
C22 C22 300.00 300.00 0.00 0.00 0.00 0.00
C23 C23 312.00 312.00 0.00 0.00 0.00 0.00
C24 C24 324.00 324.00 0.00 0.00 0.00 0.00
C25 C25 337.00 337.00 0.00 0.00 0.00 0.00
C26 C26 349.00 349.00 0.00 0.00 0.00 0.00
C27 C27 360.00 360.00 0.00 0.00 0.00 0.00
C28 C28 372.00 372.00 0.00 0.00 0.00 0.00
C29 C29 382.00 382.00 0.00 0.00 0.00 0.00
C30+ C30+ 580.00 580.00 0.00 0.00 0.00 0.00

Total Total 100.00 100.00 100.00 100.00
Calculated MW Calculated MW 22.57 22.57 22.55 22.55

Viscosity (cP) Viscosity (cP) 0.016 0.016 0.014 0.014
Heat Content (BTU/scf) - Dry Heat Content (BTU/scf) - Dry 1191 1191 1189 1189

21

21

Table 12: Constant Volume Depletion Liquid Composition at Abandonment Pressure
(Sample 1.01)
Component MW Wt% Mol%

N2 28.01 0.00 0.05
CO2 44.01 0.09 2.38
H2S 34.08 0.00 0.00
C1 16.04 0.24 17.50
C2 30.07 0.16 6.11
C3 44.10 0.19 5.02
I C4 58.12 0.08 1.64
N C4 58.12 0.15 2.91
I C5 72.15 0.10 1.67
N C5 72.15 0.10 1.58
C6 84.00 0.48 6.60
C7 96.00 0.54 6.50
C8 107.00 0.52 5.64
C9 121.00 0.34 3.22
C10 134.00 0.43 3.75
C11 147.00 0.48 3.75
C12 161.00 0.57 4.11
C13 175.00 0.68 4.53
C14 190.00 0.74 4.50
C15 206.00 0.91 5.14
C16 222.00 0.49 2.58
C17 237.00 0.46 2.27
C18 251.00 0.67 3.08
C19 263.00 0.38 1.69
C20 275.00 0.19 0.79
C21 291.00 0.13 0.54
C22 300.00 0.07 0.27
C23 312.00 0.07 0.27
C24 324.00 0.08 0.28
C25 337.00 0.04 0.14
C26 349.00 0.04 0.14
C27 360.00 0.04 0.14
C28 372.00 0.05 0.14
C29 382.00 0.09 0.29
C30+ 580.00 0.39 0.77

Total 100.00
MW 116.03

22

Table 13: Calculated Cumulative Recovery at Surface Separators
(Sample 1.01)

Cumulative Recovery per Initial Reservoir Pressure (psia)
MMSCF of Original Fluid in Place 5975 5200 4500 3800 3200 2700 2100 1500 1000

Wellstream Produced*, MSCF 1000.00 0.00 63.45 136.63 231.13 329.66 422.85 542.53 669.45 776.29

Stock Tank Liquid - bbls 69.81 0.00 4.33 8.94 13.44 16.86 19.32 21.56 23.31 24.60

Primary Sep. Gas - MSCF 913.04 0.00 57.99 125.20 213.50 307.01 396.43 512.47 636.40 740.95
Stage 2 Gas - MSCF 12.78 0.00 0.81 1.70 2.63 3.40 3.98 4.55 5.02 5.39
Stock Tank Gas - MSCF 20.14 0.00 1.27 2.67 4.16 5.39 6.35 7.29 8.10 8.74

Total Plant Products in
Primary Separator Gas - Gallons
Ethane Plus 3655.00 0.00 232.27 502.18 860.12 1243.50 1614.21 2102.98 2636.18 3184.38
Propane Plus 1376.63 0.00 87.58 189.92 328.34 479.50 627.87 826.69 1046.99 1277.29
Butanes Plus 544.37 0.00 34.69 75.53 132.31 196.07 259.98 347.45 445.91 549.92
Pentanes Plus 175.26 0.00 11.19 24.49 43.62 65.90 88.90 121.31 158.48 197.86

Stage 2 Separator Gas - Gallons
Ethane Plus 85.13 0.00 5.38 11.31 17.62 22.85 26.87 30.86 34.25 37.74
Propane Plus 33.34 0.00 2.11 4.45 6.98 9.12 10.79 12.47 13.93 15.45
Butanes Plus 12.54 0.00 0.79 1.68 2.67 3.53 4.21 4.92 5.54 6.19
Pentanes Plus 3.65 0.00 0.23 0.49 0.79 1.07 1.29 1.53 1.74 1.96

Stock Tank Gas - Gallons
Ethane Plus 425.10 0.00 26.81 56.43 88.36 115.28 136.24 157.38 157.42 176.03
Propane Plus 294.55 0.00 18.59 39.23 61.81 81.14 96.37 111.93 111.94 126.03
Butanes Plus 156.14 0.00 9.88 20.92 33.36 44.31 53.13 62.36 62.37 71.06
Pentanes Plus 52.60 0.00 3.34 7.11 11.52 15.57 18.95 22.62 22.62 26.23
*Wellstream production is from total recovery data in the CVD test.

Table 14: Calculated Instantaneous Yields at Surface Separators
(Sample 1.01)

Pressure psia 5975 5200 4500 3800 3200 2700 2100 1500 1000
SCF 1st Stage Gas/bbl STO Liquid 13079 13403 14570 19630 27298 36463 51715 70709 81424
bbl STO Liquid/MMSCF 1st Stage Gas 76.46 74.61 68.64 50.94 36.63 27.43 19.34 14.14 12.28
bbl STO Liquid/MMSCF Wellstream 69.81 68.18 63.04 47.60 34.77 26.32 18.75 13.81 12.02

Surface Separator Conditions Used:
Primary Separator at 575 psia and 74F
Stage 2 Separator at 250 psia and 74F
Stock Tank at 15.025 psia and 60F

23

APPENDIX A: DETAILED SINGLE STAGE FLASH DATA

24

Table 15: C30+ Composition, GOR, API of Sample 1.02
Component Flashed Gas Flashed Liquid Monophasic Fluid
MW WT % MOLE % WT % MOLE % WT % MOLE %

N2 28.01 0.62 0.50 0.00 0.00 0.50 0.49
CO2 44.01 12.75 6.57 0.02 0.09 10.17 6.36
H2S 34.08 0.00 0.00 0.00 0.00 0.00 0.00
C1 16.04 53.22 75.30 0.04 0.40 42.46 72.79
C2 30.07 12.70 9.59 0.06 0.32 10.14 9.28
C3 44.10 7.63 3.93 0.13 0.49 6.12 3.81
I C4 58.12 2.05 0.80 0.09 0.26 1.65 0.78
N C4 58.12 3.03 1.18 0.20 0.56 2.45 1.16
I C5 72.15 1.40 0.44 0.23 0.53 1.16 0.44
N C5 72.15 1.14 0.36 0.26 0.60 0.96 0.37
C6 84.00 3.28 0.86 2.51 4.84 3.13 1.00
C7 96.00 1.37 0.31 6.33 10.50 2.37 0.65
C8 107.00 0.54 0.11 7.48 10.88 1.94 0.47
C9 121.00 0.12 0.02 4.90 6.35 1.09 0.23
C10 134.00 0.09 0.02 5.87 7.28 1.26 0.26
C11 147.00 0.04 0.01 6.19 7.00 1.28 0.24
C12 161.00 0.02 0.00 7.06 7.29 1.44 0.25
C13 175.00 0.01 0.00 8.12 7.71 1.65 0.26
C14 190.00 0.00 0.00 8.40 7.35 1.70 0.25
C15 206.00 0.00 0.00 10.05 8.11 2.03 0.27
C16 222.00 0.00 0.00 5.29 3.96 1.07 0.13
C17 237.00 0.00 0.00 4.84 3.39 0.98 0.11
C18 251.00 0.00 0.00 6.83 4.53 1.38 0.15
C19 263.00 0.00 0.00 3.88 2.45 0.79 0.08
C20 275.00 0.00 0.00 1.87 1.13 0.38 0.04
C21 291.00 0.00 0.00 1.32 0.75 0.27 0.03
C22 300.00 0.00 0.00 0.69 0.38 0.14 0.01
C23 312.00 0.00 0.00 0.72 0.38 0.15 0.01
C24 324.00 0.00 0.00 0.75 0.38 0.15 0.01
C25 337.00 0.00 0.00 0.39 0.19 0.08 0.01
C26 349.00 0.00 0.00 0.41 0.19 0.08 0.01
C27 360.00 0.00 0.00 0.42 0.19 0.09 0.01
C28 372.00 0.00 0.00 0.44 0.19 0.09 0.01
C29 382.00 0.00 0.00 0.91 0.38 0.18 0.01
C30+ 580.00 0.00 0.00 3.29 0.94 0.67 0.03

Calculated MW 22.7 166 27.5
Mole % 3.35 96.65

OBM Contamination Level (wt%) STO Basis
Live Oil Basis

Stock Tank Oil Properties at Standard Conditions: C30+ Properties
Measured Calculated
MW 166 166 580
Density (g/cm
3
) 0.810 0.810 0.950

Single Stage Flash Data
Original STO Corrected
GOR (scf/stb) 18640
STO Density (g/cm
3
) 0.810
STO API Gravity 43.2

OBM Density (g/cm
3
) @60F

25
Inyeccin de Gas
Procesos de Inyeccin de Gas
Objetivos:
- Se aplican tanto en reservorio de petrleo como gas y
condensado.
- Ests diseados para mejorar la recuperacin de
petrleo (lquidos).
- La aplicacin principal es el mantenimiento de presin
para mantener los caudales de produccin.
- en Gas y condensado se utiliza para prevenir la baja
recuperacin de lquidos por el fenmeno de
condensacin retrgrada.
- La inyeccin de gas pobre (mayormente CH
4
o N
2
)
logra recuperar una cantidad significativa de HC
intermedios (C5 a C12)
Inyeccin de Gas Pobre en
Yac. de Gas y Condensado
- La inyeccin de gas pobre en Gas y
Condensado puede ser:
- Miscible: Si la presin de reservorio se encuentra por
encima del Punto de Roco.
- Si la presin de reservorio est por debajo del Punto
de Roco, este gas pobre puede re-vaporizar
lquidos que fueron liberados en el reservorio.
- Miscibilidad: se define como la condicin donde dos
fluidos se encuentran mezclados en alguna
proporcin en la cual la mezcla resultante se
encuentra en estado monofsico.
Reciclado de Gas en
Yac. de Gas y Condensado (1)
- El reciclado de gas en reservorios de Gas y
Condensado es utilizada para minimizar las prdidas en
la recuperacin de lquidos.
- Cuando la presin cae por debajo del Pd (punto de
roco), lquidos condensan en fondo y permanecen como
fase inmvil.
- El gas producido se vuelve menos rico y la
recuperacin final de lquidos puede llegar a bajar hasta
15/20%.
- Para maximizar la recuperacin de lquidos la presin de
reservorio se debe mantener ms alta que el Pd para no
permitir la condensacin retrgrada.
- Esto se logra re-inyectando el gas producido luego de
ser separado y procesado para extraerle todo el
condensado.
Reciclado de Gas en
Yac. de Gas y Condensado (2)
- El gas producido no es suficiente para reemplazar el
vaciado del reservorio causado por la produccin, por lo
que se debe agregar gas para lograr un mantenimiento
de presin completo.
- Si la presin del reservorio est muy por encima del Pd
se puede utilizar solamente el gas producido para re-
inyeccin y luego al ir acercndose la presin al Pd se
agrega ms gas.
- El resultado econmico de retrasar las ventas de gas
(por su re-inyeccin) para incrementar el recupero del
condensado puede ser prohibitivo.
- Alternativas para hacerlo econmico puede ser la re-
inyeccin de parte del gas, comprar gas pobre para
inyeccin o utilizar N2.
Inyeccin de N2 en
Yac. de Gas y Condensado
- La utilizacin de N2 generado en locacin para re-
inyeccin ha sido uno de los mtodos ms usados para
proyectos de reciclado de gas.
- En los 80s algunos estudios demostraron que el N2
causaba una sustancial condensacin de lquido cuando
era mezclado junto con gas y condensado. Este
comportamiento caus la preocupacin de que el N2
podra hasta empeorar el problema de la condensacin
retrgrada.
- Estudios posteriores de desplazamiento demostraron
que prcticamente todo el lquido condensado por el
contacto inicial con el N2 fue re-vaporizado con el
contacto continuo con N2.
- Ensayos de slim-tube demuestran resultados similares
en inyeccin de N2 con respecto a inyeccin de gas
pobre.
Efectos de la Inyeccin de N2
Inyeccin de N2 y su influencia en
la condensacin retrgrada
14 Oileld Review
Understanding Gas-Condensate Reservoirs
Li Fan
College Station, Texas, USA
Billy W. Harris
Wagner & Brown, Ltd.
Midland, Texas
A. (Jamal) Jamaluddin
Rosharon, Texas
Jairam Kamath
Chevron Energy Technology Company
San Ramon, California, USA
Robert Mott
Independent Consultant
Dorchester, UK
Gary A. Pope
University of Texas
Austin, Texas
Alexander Shandrygin
Moscow, Russia
Curtis Hays Whitson
Norwegian University of Science and
Technology and PERA, A/S
Trondheim, Norway
For help in preparation of this article, thanks to Syed Ali,
Chevron, Houston; and Jerome Maniere, Moscow.
ECLIPSE 300, LFA (Live Fluid Analyzer for MDT tool), MDT
(Modular Formation Dynamics Tester) and PVT Express are
marks of Schlumberger. CHEARS is a mark of Chevron.
Teon is a mark of E.I. du Pont de Nemours and Company.
How does a company optimize development of a gas-condensate eld, when
depletion leaves valuable condensate uids in a reservoir and condensate blockage
can cause a loss of well productivity? Gas-condensate elds present this puzzle.
The rst step must be to understand the uids and how they ow in the reservoir.
A gas-condensate reservoir can choke on its
most valuable components. Condensate liquid
saturation can build up near a well because of
drawdown below the dewpoint pressure,
ultimately restricting the ow of gas. The near-
well choking can reduce the productivity of a
well by a factor of two or more.
This phenomenon, called condensate
blockage or condensate banking, results from a
combination of factors, including fluid phase
properties, formation flow characteristics and
pressures in the formation and in the wellbore.
If these factors are not understood at the
beginning of eld development, sooner or later
production performance can suffer.
For example, well productivity in the Arun
field, in North Sumatra, Indonesia, declined
significantly about 10 years after production
began. This was a serious problem, since well
deliverability was critical to meet contractual
obligations for gas delivery. Well studies,
including pressure transient testing, indicated
the loss was caused by accumulation of
condensate near the wellbore.
1
Arun is one of several huge gas-condensate
reservoirs that together contain a significant
global resource. Other large gas-condensate
resources include Shtokmanovskoye eld in the
Russian Barents Sea, Karachaganak field in
Kazakhstan, the North field in Qatar that
becomes the South Pars field in Iran, and the
Cupiagua eld in Colombia.
2
This article reviews the combination of uid
thermodynamics and rock physics that results in
condensate dropout and condensate blockage.
We examine implications for production and
methods for managing the effects of condensate
dropout, including reservoir modeling to predict
eld performance. Case studies from Russia, the
USA and the North Sea describe eld practices
and results.
Forming Dewdrops
A gas condensate is a single-phase fluid at
original reservoir conditions. It consists
predominantly of methane [C
1
] and other short-
chain hydrocarbons, but it also contains long-
chain hydrocarbons, termed heavy ends. Under
1. Adick D, Kaczorowski NJ and Bette S: Production
Performance of a Retrograde Gas Reservoir: A Case
Study of the Arun Field, paper SPE 28749, presented at
the SPE Asia Pacic Oil & Gas Conference, Melbourne,
Australia, November 710, 1984.
2. For a case study of the Karachaganak eld: Elliott S,
Hsu HH, OHearn T, Sylvester IF and Vercesi R: The
Giant Karachaganak Field, Unlocking Its Potential,
Oileld Review 10, no. 3 (Autumn 1998): 1625.
3. Gas-condensate uids are termed retrograde because
their behavior can be the reverse of uids comprising
pure components. As reservoir pressure declines and
passes through the dewpoint, liquid forms and the
amount of the liquid phase increases with pressure
drop. The system reaches a point in a retrograde
condensate where, as pressure continues to decline,
the liquid revaporizes.
4. Injection of cold or hot uids can change reservoir
temperature, but this rarely occurs near production
wells. The dominant factor for uid behavior in the
reservoir is the pressure change. As will be discussed
later, this is no longer the case once the uid is produced
into the wellbore.
Winter 2005/2006 15
certain conditions of temperature and pressure,
this uid will separate into two phases, a gas and
a liquid that is called a retrograde condensate.
3
As a reservoir produces, formation temper-
ature usually doesnt change, but pressure
decreases.
4
The largest pressure drops occur
near producing wells. When the pressure in a
gas-condensate reservoir decreases to a certain
point, called the saturation pressure or
dewpoint, a liquid phase rich in heavy ends
drops out of solution; the gas phase is slightly
depleted of heavy ends (right). A continued
decrease in pressure increases the volume of the
liquid phase up to a maximum amount; liquid
volume then decreases. This behavior can be
displayed in a pressure-volume-temperature
(PVT) diagram.
The amount of liquid phase present depends
not only on the pressure and temperature, but
also on the composition of the uid. A dry gas, by
denition, has insufcient heavy components to
generate liquids in the reservoir, even with near-
well drawdown. A lean gas condensate generates
>
Phase diagram of a gas-condensate system. This pressure-volume-
ttemperature (PVT) plot indicates single-phase behavior outside the two-
phase region, which is bounded by bubblepoint and dewpoint lines. Lines
of constant phase saturation (dashed) all meet at the critical point. The
numbers indicate the vapor phase saturation. In a gas-condensate
reservoir, the initial reservoir condition is in the single-phase area to the
right of the critical point. As reservoir pressure declines, the uid passes
tthrough the dewpoint and a liquid phase drops out of the gas. The
percentage of vapor decreases, but can increase again with continued
pressure decline. The cricondentherm is the highest temperature at which
ttwo phases can coexist. Surface separators typically operate at
conditions of low pressure and low temperature.
Temperature
P
r
e
s
s
u
r
e
Initial reservoir Initial reservoir
condition condition
Critical point Critical point
Separator Separator
condition condition
C i d th Cricondentherm
Two phase region Two-phase region
60% 60%
70% 70%
80%
90%
100% vapor 00% apo
BBB
uu
bb
bb
ll
ee
pp
oo
inn
tt
l
iinn
ee
DD
e
ww
pp
oo
ii
nn
tt
l
ii
nn
ee
a small volume of the liquid phase, less than
100 bbl per million ft
3
[561 m
3
per million m
3
],
and a rich gas condensate generates a larger
volume of liquid, generally more than 150 bbl
per million ft
3
[842 m
3
per million m
3
] (above).
5
There are no established boundaries in the
definitions of lean and rich, and further
descriptorssuch as very leanare also
applied, so these gures should be taken merely
as indicators of a range.
Determining the fluid properties can be
important in any reservoir, but it plays a
particularly vital role in gas-condensate
reservoirs. For example, condensate/gas ratio
plays a major role in estimates for the sales
potential of both gas and liquid, which are
needed to size surface processing facilities. The
amount of liquid that may be stranded in a eld
is also an essential economic consideration.
These considerations and others, such as the
need for artificial lift and stimulation
technologies, rely on accurate fluid sampling.
Small errors in capturing samples, such as an
incorrect amount of captured liquid, can have
signicant errors in measured behavior, so great
care must be taken in the sampling process (see
Sampling for Fluid Properties, next page).
Once reservoir uids enter a wellbore, both
temperature and pressure conditions may
change. Condensate liquid can be produced into
the wellbore, but liquid also can drop out within
the wellbore because of changes in conditions. If
the gas does not have sufcient energy to carry
the liquid to surface, liquid loading or fallback in
the wellbore occurs because the liquid is denser
than the gas phase traveling along with it. If the
liquid falls back down the wellbore, the liquid
percentage will increase and may eventually
restrict production. Gas lift and pumping
technologies that are used to counter this
behavior will not be discussed in this article.
6
16 Oileld Review
>
Examples of rich and lean gas-condensate behavior. When pressure decreases at reservoir temperature, a rich gas (top left) forms a higher tt
percentage of liquid than a lean gas (top right). The rich gas drops out more condensate than the lean gas ( tt bottom left). The liquid dropout curve tt
assumes the two phases remain in contact with one another. However, in a reservoir, the mobile gas phase is produced; the liquid saturation in the
near-well region builds until it is also mobile. As a result, eventually condensate blockage can affect formations with both lean and rich gases, and
tthe normalized well productivity index (J/ J J
0
) of both can be severely impacted (
0
bottom right). tt
L
i
q
u
i
d

d
r
o
p
o
u
t
,

%
Pressure, psi
0 2,000 3,000 4,000 5,000 6,000 1,000
0
5
10
15
20
25
Lean gas condensate Lean gas condensate Lean gas condensate
Rich gas condensate Rich gas condensate Rich gas condensate
0
0.2
0.4
0.6
0.8
1.0
0 0.2 0.4 0.6 0.8 1.0
Lean gas condensate L d Lean gas condensate g
Rich gas condensate Rich gas condensate Rich gas condensate
P
average
/P
dewpoint
P
r
o
d
u
c
t
i
v
i
t
y

r
a
t
i
o
,

J
/
J
o
Critical Critical
point point
0
1,000
2,000
3,000
4,000
5,000
6,000
150 200 250 300 350 400 450 500 550 600
Temperature, K
Reservoir temperature Reservoir temperature
98 5% 98.5%
99% 99%
99.5%
P
r
e
s
s
u
r
e
,

p
s
i
Lean Gas
Condensate
0
1,000
2,000
3,000
4,000
5,000
6,000
7,000
0 200 300 400 500 600 700 800 900 100
Temperature, K
P
r
e
s
s
u
r
e
,

p
s
i
Critical Critical
point point
Reservoir temperature Reservoir temperature
75% 75%
80% 0
85% 85% 5
90% 90%
95% 95%
Rich Gas
Condensate
5. Gas volumes in this article are given at the conditions that
are considered standard at the measurement location,
which is not the same around the world. Conversions
between metric and oileld units are volumetric.
6. For more on articial lift: Fleshman R, Harryson and
Lekic O: Articial Lift for High-Volume Production,
Oileld Review 11, no. 1 (Spring 1999): 4863.
Winter 2005/2006 17
Fluid composition is determined by capturing
a representative sample of reservoir uid.
Surface samples can be obtained relatively
easily by collecting liquid and gas samples
from test or production separators. The
samples are then recombined in a laboratory.
However, the result can be unrepresentative
of reservoir conditions, particularly when
sampling from a gas-condensate reservoir. A
few examples of potential problems include
recombining the gas and liquid samples at
an incorrect ratio, changing production
conditions prior to or during sampling and
commingling zones with different properties.
If the liquid content is low when capturing
surface samples, a small loss of the liquid in
production tubulars or separators could
render the condensate sample unrepresen-
tative of the formation uid.
Samples can also be collected downhole
from wellbore uids in gas-condensate
reservoirs. This is practical and desirable if
the wellbore owing pressure is above the
dewpoint pressure, but it is generally not
recommended if the pressure anywhere in the
tubing is lower than the dewpoint pressure. In
that condition, there is two-phase ow in the
wellbore. Any liquid forming in the tubing
during or prior to the sampling may segregate
to the bottom of the tubing stringwhere a
bottomhole sampler collects uidspotentially
resulting in an unrepresentative sample with
too much of the heavier components.
Formation testers have improved signi-
cantly over the past decade. The MDT Modular
Formation Dynamics Tester collects uids by
pressing a probe against an uncased borehole
wall and withdrawing uids from a formation.
1
The LFA Live Fluid Analyzer module on the
tool measures the cleanup of contamination
from oil-base drilling or completion uids,
minimizing the wait time and assuring quality
samples.
2
The LFA detector also provides an
indication of the amount of methane, other
light components and liquids. From these
data, the ratio of methane to liquid provides
a measure of the condensate/gas ratio, an
important consideration for early economic
evaluation of a prospect. The analysis can
also show zones with different compositions
or compositional gradients.
Measured data from the MDT tool are trans-
mitted to surface immediately, so sampling
decisions can be made based on knowledge
of approximate composition and reservoir
pressure, another measured parameter. If
desired, uid samples can be collected before
moving to another downhole location.
For gas condensates that are at pressures
above the dewpoint in the reservoir, it is
important to capture and maintain single-
phase uid. If the uid pressure drops below
dewpoint, it may take a long time to
recombine the sample. Even worse, some
changes that occur in a sample on its trip to
surface may be irreversible. By providing
evidence when a uid goes through its
dewpoint, the LFA measurement can indicate
when the pressure drawdown is too large and
should be decreased before sampling to keep
pressure above the dewpoint.
A sample that is single-phase when collected
should be kept in a single phase when brought
to surface. Special MDT sample bottles are
available for this purpose. A single-phase
bottle uses a nitrogen cushion to increase the
pressure in the sampled uid.
3
The sample
cools as it is brought to surface, but the
nitrogen cushion on the sample keeps its
pressure above the dewpoint.
In most cases, the PVT Express onsite well
uid analysis service can provide uid
property measurements at the wellsite in
about 24 hours, saving the weeks or months
that may be needed to get results from a
laboratory.
4
The PVT Express systems can
measure gas/liquid ratio, saturation
pressurebubblepoint or dewpoint
composition to C
30+
, reservoir uid density,
viscosity and oil-base mud contamination.
5
These measurements are critical because an
operating company can use them immediately
to make a decision to complete or to test a
well. Rapid turnaround may be crucial when
drilling exploration or development wells from
an expensive offshore rig. More complete
analyses can be obtained later from samples
sent to a laboratory.
With the basic understanding of where
and how condensate drops out of the gas
phase, engineers can devise ways to optimize
production of gas and condensate.
Sampling for Fluid Properties
1. Andrews RJ, Beck G, Castelijns K, Chen A, Cribbs ME,
Fadnes FH, Irvine-Fortescue J, Williams S, Hashem M,
Jamaluddin A, Kurkjian A, Sass B, Mullins OC,
Rylander E and Van Dusen A: Quantifying
Contamination Using Color of Crude and Condensate,
Oileld Review 13, no. 3 (Autumn 2001): 2443.
2. Betancourt S, Fujisawa G, Mullins OC, Carnegie A,
Dong C, Kurkjian A, Eriksen KO, Haggag M, Jaramillo
AR and Terabayashi H: Analyzing Hydrocarbons in the
Borehole, Oileld Review 15, no. 3 (Autumn 2003):
5461.
3. Jamaluddin AKM, Ross B, Calder D, Brown J and
Hashem M: Single-Phase Bottomhole Sampling
Technology, Journal of Canadian Petroleum
Technology 41, no. 7 (July 2002): 2530.
4. Jamaluddin AKM, Dong C, Hermans P, Khan IA,
Carnegie A, Mullins OC, Kurkjian A, Fujisawa G,
Nighswander J and Babajan S: Real-Time and On-
Site Reservoir Fluid Characterisation Using Spectral
Analysis and PVT Express, Australian Petroleum
Production & Exploration Association Journal (2004):
605616.
5. The nomenclature composition to C
30+
indicates
compounds up to 29 carbon atoms are separately
discriminated, with the remainder combined into a
fraction indicated as C
30+
.
Dewdrops in a Reservoir
When condensate liquid first forms in a gas
reservoir, it is immobile because of capillary
forces acting on the uids. That is, a microscopic
liquid droplet, once formed, will tend to be
trapped in small pores or pore throats. Even
for rich gas condensates with substantial
liquid dropout, condensate mobility, which is the
ratio of relative permeability to viscosity,
remains insignicant away from wellbores. As a
consequence, the condensate that forms in most
of the reservoir is lost to production unless
the depletion plan includes gas cycling. The
effect of this dropout on gas mobility is
typically negligible.
Near a producing well, the situation is
different. Once bottomhole pressure drops
below the dewpoint, a near-well pressure sink
forms around the well. As gas is drawn into the
pressure sink, liquid drops out. After a brief
transient period, enough liquid accumulates
that its mobility becomes significant. The gas
and liquid compete for ow paths, as described
by the formations relative-permeability
relationship. Condensate blockage is a result of
the decreased gas mobility around a producing
well below the dewpoint (right).
Reservoir pressure dropping below the
dewpoint has two main results, both negative:
gas and condensate production decrease
because of near-well blockage, and the produced
gas contains fewer valuable heavy ends because
of dropout throughout the reservoir, where the
condensate has insufficient mobility to flow
toward the well.
Large productivity losses have been reported
for wells in gas-condensate fields. In the Arun
field, which was operated by Mobil, now
ExxonMobil, the loss in some wells was greater
than 50%.
7
In another case, Exxon, now
ExxonMobil, reported two wells that died due to
condensate blockage.
8
Shell and Petroleum
Development Oman reported a 67% productivity
loss for wells in two elds.
9
In another field, the initial productivity
decline has reportedly reversed. The productivity
of wells in the moderately rich gas-condensate
reservoir declined rapidly when bottomhole
pressures dropped below dewpoint. This decline
continued until pressure throughout the
reservoir dropped below dewpoint, then gas
productivity began to increase. Compositional
modeling showed that condensate saturation
increased near the wells to approximately 68%,
decreasing gas permeability and therefore gas
productivity. However, when pressure throughout
the reservoir dropped below dewpoint, some
liquid dropped out everywhere. The gas moving
toward the wellbore was leaner and had less
condensate to drop out in the near-well region,
resulting in decreased condensate saturation to
about 55% and increased gas productivity.
10
The
condensate blockage decreased as the near-well
gas mobility increased.
Condensate Blockage
Not all gas-condensate reservoirs are pressure-
limited because of near-well condensate
blockage, even though all of these fields will
experience condensate blockage. The degree to
which condensate dropout is a production
problem depends on the ratio of the pressure
drop that is experienced within the reservoir to
the total pressure drop from distant areas of the
reservoir to a control point at surface.
If reservoir pressure drop is signicant, then
additional pressure drop due to condensate
blockage can be very important for well
deliverability. This condition typically applies in
a formation with a low kh, the product of
permeability and net formation thickness.
Conversely, if little of the total pressure drop
occurs in the reservoir, typical of high kh
formations, then adding more pressure drop in
the reservoir due to condensate blockage will
probably have little impact on well deliverability.
As a general guideline, condensate blockage can
be assumed to double the pressure drop in the
reservoir for the same ow rate.
Conceptually, flow in gas-condensate fields
can be divided into three reservoir regions,
although in some situations not all three are
present (next page).
11
The two regions closest to
a well can exist when bottomhole pressure is
below the dewpoint of the uid. The third region,
away from producing wells, exists only when the
reservoir pressure is above the dewpoint.
This third region includes most of the
reservoir away from producing wells. Since it is
above the dewpoint pressure, there is only one
hydrocarbon phase, gas, present and flowing.
The interior boundary of this region occurs
where the pressure equals the dewpoint
pressure of the original reservoir gas. This
boundary is not stationary, but moves outward as
hydrocarbons are produced from the well and
the formation pressure drops, eventually
disappearing as the outer-boundary pressure
drops below the dewpoint.
18 Oileld Review
>
Condensate blockage. Once bottomhole pressure in a well falls
below the dewpoint, condensate will drop out from the gas phase.
Capillary forces favor having condensate in contact with the grains
(inset, right). After a brief transient period, the region achieves a tt
steady-state ow condition with both gas and condensate owing
(inset, top). The condensate saturation, S
o
, is highest near the
o
wellbore because the pressure is lower, which means more liquid
dropout. The oil relative permeability, k
ro
, increases with saturation.
o
The decrease in gas relative permeability, k
rg
, near the wellbore
illustrates the blockage effect. The vertical axis, represented by a
wellbore, is schematic only.
Distance from borehole
k
ro
S
o
a
t
i
v
e

p
e
r
m
e
a
b
i
l
i
t
y
R
e
l
0
0.5
1.0
0 0.5 1.0
Condensate saturation
kkkkk
ro ro ro
kkkkk
rg rg rg g
k
rg
Condensate
flow channel
Sand grain
Gas flow
channel
Winter 2005/2006 19
In the second region, the condensate-buildup
region, liquid drops out of the gas phase, but its
saturation remains low enough that it is
immobile; there is still single-phase gas flow.
The amount of liquid that drops out is
determined by the uids phase characteristics,
as indicated by its PVT diagram. The liquid
saturation increases and the gas phase becomes
leaner as gas flows toward the wellbore. This
regions inner-boundary saturation usually is
near the critical liquid saturation for ow, which
is the residual oil saturation.
In the first region, closest to a producing
well, both gas and condensate phases ow. The
condensate saturation here is greater than the
critical condensate saturation. This region
ranges in size from tens of feet for lean
condensates to hundreds of feet for rich
condensates. Its size is proportional to the
volume of gas drained and the percentage of
liquid dropout. It extends farther from the well
for layers with higher permeability than average
since a larger volume of gas has owed through
these layers. Even in a reservoir containing lean
gas with low liquid dropout, condensate
blockage can be significant, because capillary
forces can retain a condensate that builds to a
high saturation over time.
This near-well condensate blockage region
controls well deliverability. The flowing
condensate/gas ratio is essentially constant and
the PVT condition is considered a constant-
composition expansion region.
12
This condition
simplifies the relationship between gas and oil
relative permeabilities, making the ratio
between the two a function of PVT properties.
However, additional relative-permeability
effects occur in the near-well region because the
gas velocity, and therefore the viscous force, is
extreme. The ratio of viscous to capillary forces
is called the capillary number.
13
Conditions of
pressure gradient caused by high velocity or low
interfacial tension have high capillary numbers,
indicating that viscous forces dominate, and the
relative permeability to gas is higher than the
value at lower ow rates.
At even higher flow velocities nearer the
wellbore, the inertial or Forchheimer effect
decreases the gas relative permeability
somewhat.
14
The basis of this effect is the
inertial drag as fluid speeds up to go through
pore throats and slows down after entering a
pore body.
15
The result is a lower apparent
permeability than would be expected from
Darcys law. The effect is usually referred to as
non-Darcy ow.
The overall impact of the two high-velocity
effects is usually positive, reducing the impact of
condensate blockage. Laboratory coreflood
experiments are needed to measure the
inertial and capillary number effects on
relative permeability.
Although the first indication of condensate
blockage is typically a productivity decline, its
presence is often determined by pressure
transient testing. A pressure-buildup test can be
interpreted to show the distribution of liquid
before the well is shut in. The short-time
behavior in the transient test reects near-well
conditions. Condensate blockage is indicated by
a steeper pressure gradient near the wellbore.
With longer test times, the gas permeability far
from the wellbore dominates the response;
permeability can be determined from the
derivative curve on a log-log plot of pseudo-
pressure and shut-in time. If the test continues
long enoughand that shut-in test time
depends on the formation permeabilityflow
properties far from the well will be evident.
Gas-Condensate Reservoir Management
Historically, condensate liquids have been
significantly more valuable than the gas, and
this is still true in a few places far from a gas
market or transport system. The price
differential made gas cycling a common
practice. Injecting dry gas into a formation to
keep reservoir pressure above the dewpoint
slowly displaces valuable heavy ends that are
still in solution in the reservoir gas. Eventually,
the reservoir is blown down; that is, the dry or
lean gas is produced at a low bottomhole pressure.
7. Adick et al, reference 1.
8. Barnum RS, Brinkman FP, Richardson TW and
Spillette AG: Gas Condensate Reservoir Behaviour:
Productivity and Recovery Reduction Due to
Condensation, paper SPE 30767, presented at the SPE
Annual Technical Conference and Exhibition, Dallas,
October 2225, 1995.
9. Smits RMM, van der Post N and al Shaidi SM: Accurate
Prediction of Well Requirements in Gas Condensate
Fields, paper SPE 68173, presented at the SPE Middle
East Oil Show, Bahrain, March 1720, 2001.
10. El-Banbi AH, McCain WD Jr and Semmelbeck ME:
Investigation of Well Productivity in Gas-Condensate
Reservoirs, paper SPE 59773, presented at the SPE/CERI
Gas Technology Symposium, Calgary, April 35, 2000.
11. Fevang and Whitson CH: Modeling Gas-Condensate
Well Deliverability, SPE Reservoir Engineering 11, no. 4
(November 1996): 221230.
12. In a constant-composition expansion condition, the uid
expands with pressure decline and two phases may form,
but no components are removed. This contrasts with the
second region, which is considered a constant-volume
depletion region, because the liquid phase that forms
drops out from the gas phase and becomes trapped.
13. Henderson GD, Danesh A, Tehrani DH and Al-Kharusi B:
The Relative Signicance of Positive Coupling and
Inertial Effects on Gas Condensate Relative
Permeabilities at High Velocity, paper SPE 62933,
presented at the SPE Annual Technical Conference and
Exhibition, Dallas, October 14, 2000.
Whitson CH, Fevang and Svareid A: Gas
Condensate Relative Permeability for Well Calculations,
paper SPE 56476, presented at the SPE Annual Technical
Conference and Exhibition, Houston, October 36, 1999.
14. Forchheimer PH: Wasserbewegung durch Boden,
Zeitschrift ver Deutsch Ingenieur 45 (1901): 17821788.
15. Barree RD and Conway MW: Beyond Beta Factors: A
Complete Model for Darcy, Forchheimer, and Trans-
Forchheimer Flow in Porous Media, paper SPE 89325,
Exhibition, Houston, September 2629, 2004.
>
Three reservoir regions. Gas-condensate eld behavior can be divided
into three regions once bottomhole pressure, P
BH
, drops below the
H
dewpoint pressure, P
D
. Far from a producing well (3), where the reservoir
pressure is greater than P
D
, there is only one hydrocarbon phase present,
D
gas. Closer to the well (2), there is a region between the dewpoint pressure
and the point, r
1
, at which the condensate reaches the critical saturation
for ow. In this condensate-buildup region, both phases are present, but
only gas ows. Once condensate saturation exceeds the critical saturation,
both phases ow toward the well (1).
P
r
e
s
s
u
r
e
P
D
P
BH
r
1
Dewpoint pressure Dewpoint pressure
Reservoir pressure Reservoir pressure
Distance
2 3 2 3 2 3
r
e
o
r
b
o
e
l
l
W
e
W
1
The price of gas has risen to a value that
makes reinjection a less attractive strategy,
unless the uid is very rich in heavy ends. Gas
injection is now more commonly used as a
temporary activity, until a pipeline or other
transport facility is built, or as a seasonal
activity during periods of low gas demand.
Operators also work to overcome condensate
blockage. Some techniques are the same in a
gas-condensate field as they are in a dry-gas
field. Hydraulic fracturing is the most common
mitigating technology in siliciclastic reservoirs,
and acidizing is used in carbonate reservoirs.
Both techniques increase the effective contact
area with a formation. Production can be
improved with less drawdown in the formation.
For some gas-condensate fields, a lower
drawdown means single-phase production above
the dewpoint pressure can be extended for a
longer time.
However, hydraulic fracturing does not
generate a conduit past a condensate saturation
buildup area, at least not for long. Once the
pressure at the sandface drops below the
dewpoint, saturation will increase around the
fracture, just as it did around the wellbore.
Horizontal or inclined wells are also being
used to increase contact area within formations.
The condensate still builds up around these
longer wells, but it takes a longer time. The
productivity of the wells remains high longer,
but the benefit must be weighed against the
increased well cost.
Some operators have tried shutting in wells
to allow time for the gas and condensate to
recombine, but fluid phase behavior generally
does not favor this approach. Separation of a
fluid into a gas and liquid phase in the two-
phase region of the phase diagram happens
quickly, and after this the phases tend to
segregate, either within a pore or on a larger
scale. This phase separation dramatically slows
the reverse process of recombining gas and
liquid. This reversal requires immediate contact
between the gas and liquid phases.
Another method, cyclic injection and
production from one well, sometimes called huff
and puff injection, uses dry gas to vaporize
condensate around a well and then produce it.
This can have short-term benefit for increased
productivity, but the blockage returns when
production begins again and the formation drops
below the dewpoint pressure of the current
gas mixture.
In a eld test, methanol solvent was injected
into Hatters Pond field, Alabama, USA. In this
field, production of a gas condensate comes
mainly from the lower Norphlet sandstone, but
the field also produces from the Smackover
dolomite. Wells in Hatters Pond eld are about
18,000 ft [5,490 m] deep with 200 to 300 ft [60 to
90 m] of net pay. Gas productivity had declined
by a factor of three to ve because of condensate
and water blockage. The operator, Texaco (now
Chevron), pumped 1,000 bbl [160 m
3
] of
methanol down tubing at a rate of 5 to 8 bbl/min
[0.8 to 1.3 m
3
/min] into low-permeability
formations.
16
The methanol treatment removes
both oil and water through a multiple-contact
miscible displacement.
17
As a result of the
treatment, gas production increased by a factor
of three initially, then stabilized at 500,000 ft
3
/d
3
[14,160 m
3
/d], a factor of two over the
pretreatment rate. Condensate production
doubled to 157 bbl/d [25 m
3
/d]. Both gas and
condensate rates persisted for more than
10 months after treatment.
18
Treatment methods have been suggested for
removing condensate blockage through injection
of surfactants mixed with solvents to alter
wetting preference in the reservoir. This topic
will be discussed later in this article.
Remobilizing Stranded Condensate
The Vuktyl gas-condensate field in the Komi
Republic, Russia, has been in production since
1968. Although productivity was not severely
impacted by condensate blockage in the eld, a
signicant amount of condensate dropped out in
the carbonate reservoir. Several condensate
recovery pilots were run in this eld.
The eld is a long anticline with production
from the Middle Carboniferous Moscow and
Bashkir sequences (above left). The 1,440 m
[4,724 ft] thick structure comprises alternating
20 Oileld Review
>
Vuktyl eld, Russia. The Vuktyl eld in the Komi Republic in western Russia (top) is an anticline,
80 km [50 mi] long and up to 6 km [3.7 mi] wide (bottom). The Roman numerals denote gas-processing
facility collection areas. The uid is predominantly methane [C
1
], but with a signicant amount of
intermediate hydrocarbon components and nitrogen (table, right). The eld has three lithotypes tt
(table, left). tt
0 mi 4
Komi
RRRRRR
U S S I I I I

S
S
S S

S
S
S
S
S

U
U
U
U U
AAAAAA
Component
Composition,
% by mole
volume
C
2
C
1
C
3
C
4
C
5+
N
2
74.6
8.9
3.8
1.8
6.4
4.5
Porosity, % Permeability, mD Type
Fine porosity, microvugs, microfractured
Porous, microvugs, microfractured
Fractured, microvugs, porous 0.1 to 4513
0.01 to 0.1
<0.1 0.1 to 3
3 to 6
>6
Winter 2005/2006 21
limestone and dolomite layers, with an average
interbed thickness of 1.5 m [5 ft]. The reservoir
properties vary widely throughout the eld, but
the field has been divided into seven pay
sequences of three basic types. All three types
have microfractures and microvugular porosity.
Fine pores, low permeability and low porosity
distinguish the first type. The third type has
fractures large enough to contribute to
permeability. The other type is intermediate.
At discovery, reservoir conditions were
36 MPa [5,200 psi] and 61C [142F], with 77.5%
initial gas saturation. There is a small rim
containing light oil. Initial gas in place was
about 430 x 10
9
m
3
[15 x 10
12
ft
3
] and initial
condensate was about 142 million metric tons
[1,214 million bbl].
19
The initial, stable,
producing condensate/gas ratio was 360 g/m
3
[87.1 bbl per million ft
3
].
20
The field has an
underlying aquifer, but the water drive was
insignicant and laterally uneven.
The complex geology of the field, including
high-permeability zones that could have acted as
thief zones, led the operator, Gazprom, to develop
the eld with no gas cycling, using depletion gas
drive as the primary production mechanism.
Approximately 170 vertical wells at a typical
spacing of 1,000 to 1,500 m [3,280 to 4,920 ft]
were placed in an irregular triangular grid. Most
of the production wells had 10-in. intermediate
casing and 6
5
5
8 -in. production casing. Several
prolic wells had larger, 7
5
-in. production casing,
allowing 4
5
-in. tubing. Typical completions in the
500- to 800-m [1,640- to 2,625-ft] producing zone
were perforated casing, but some wells used
screen or openhole completions. The deepest
producers were drilled about 100 to 150 m [328
to 492 ft] above the gas/water contact. A
two-stage hydrochloric acid treatment was the
main method of well stimulation.
After nine years, the production plateau was
19 x 10
9
m
3
/yr [671 x 10
9
ft
3
/yr]. A peak stable
condensate production of 4.2 million tons/yr
[36 million bbl/yr] occurred during the
sixth year of development.
Currently, the Vuktyl field is in its final
development phase. Reservoir pressure is 3.5 to
5 MPa [508 to 725 psi]. Approximate field
recoveries are 83% of the gas and 32% of the
condensate, so about 100 million tons
[855 million bbl] of condensate remain in
the eld.
Experts from Severgazprom, a part of the
Gazprom Russian Joint Stock Company, and the
VNIIGAZ and SeverNIPIgaz institutes conducted
a variety of pilot projects in Vuktyl field to
recover additional condensate. In 1988, the
company began the rst pilot experiment, using
a solvent to recover stranded condensate.
21
The
pilot included six producers, one injection well
and three monitor wells (above). The solvent,
25,800 tons [293,000 bbl at formation
conditions] of a mixture of propane [C
3
] and
butane [C
4
], was injected into the formation
followed by 35 million m
3
[1.24 x 10
9
ft
3
] of
separator gas.
22
The intent was to recover
condensate through miscible displacement of
the solvent bank.
Geophysical observations conducted during
the experiment indicated that solvent and
injected gas entered the producing intervals of
the injection well unevenly. Component analyses
of samples from the production and monitor
wells indicated solvent and injected gas broke
through only in the two closest monitor wells
and in none of the production wells. Two events
were seen in these two monitor wells, a change
in condensate/gas ratio from 43 to 65 g/m
3
[10.4
to 15.7 bbl per million ft
3
] with a decline to the
initial ratio, followed by a second increase from
43 to 54 g/m
3
[to 13 bbl per million ft
3
].
Production logging in the monitor wells
revealed two-phase owgas and solventonly
in the bottom part of the productive section.
Overall, 95% of the solvent was produced from the
two monitor wells, but condensate recovery was
only about 0.4%. The pilot study concluded that
the propane and butane solvent bank was not
sufciently effective in recovering condensate.
A different recovery method, injecting dry
gas, began in the Vuktyl eld in 1993. The gas,
from a trunk pipeline that originated in the
Tyumen district, is injected under pipeline
pressure at 5.4 to 7.4 MPa [780 to 1,070 psi]
16. Al-Anazi HA, Walker JG, Pope GA, Sharma MM and
Hackney DF: A Successful Methanol Treatment in a
Gas-Condensate Reservoir: Field Application, paper
SPE 80901, presented at the SPE Production and
Operations Symposium, Oklahoma City, Oklahoma, USA,
March 2225, 2003.
17. In a miscible displacement, a solvent allows uids to mix
freely in a homogeneous mixture. Multiple-contact
miscibility requires sufcient mass transfer between the
solvent and hydrocarbons to achieve miscibility.
18. Al-Anazi et al, reference 16.
19. Zhabrev IP (ed): Gas and Gas-Condensate Fields
Reference Book. Moscow: Nedra, 1983 (in Russian). k
Ter-Sarkisov RM: The Development of Natural Gas Fields.
Moscow: Nedra, 1999 (in Russian).
Conversion from mass to volume is based on condensate
density of 8.55 bbl/ton.
20. Vyakhirev RI, Gritsenko AI and Ter-Sarkisov RM: The
Development and Operation of Gas Fields. Moscow: s
Nedra, 2002 (in Russian).
21. Ter-Sarkisov RM, Gritsenko AI and Shandrygin AN:
Development of Gas Condensate Fields Using
Stimulation of Formation. Moscow: Nedra, 1996
(in Russian).
Vyakhirev et al, reference 20.
22. For more on the role of propane in lowering the dewpoint
of a gas-condensate eld: Jamaluddin AKM, Ye S,
Thomas J, DCruz D and Nighswander J: Experimental
and Theoretical Assessment of Using Propane to
Remediate Liquid Buildup in Condensate Reservoirs,
paper SPE 71526, presented at the SPE Annual Technical
Conference and Exhibition, New Orleans, September 30
October 3, 2001.
>
Plan view with depth to the formation top at a solvent-injection pilot
project near gas-processing facility number 1 (GPF-1). Propane and butane
were injected into Well 103, followed by separator gas. Six production
wellsdesignated 91, 92, 93, 104, 105 and 106and three monitor wells
designated 38, 256 and 257made up the pilot study area. Solvent was
observed and produced only from the two closest monitor wells: Wells 38
and 256.
2 3 2,30000
2 400 2,400
2 500 2,500
2 600 2,600
2 700 2,700
2 800 2,800
2 900 2,900
3 000 m 3,000 m
64
90
95
101 101 1
102
15 15
159
66
93
257
38 38
103
104
256
105
92
91
86
12
19 19
264
106
GPF-1
Pilot area
Production well
Injector well
Monitor well
without local compression.
23
Formation gas,
which is in equilibrium with the retrograde
condensate, is replaced by injected dry gas. The
light C
2
to C
4
components and intermediate C
5+
fractions evaporate into the dry gas.
24
Thus,
recovery is improved both by producing more
formation gas, which still contains components
other than methane, and by vaporizing stranded
liquids and producing them along with the
injected gas. In addition, the injected gas causes
no problems for the production facilities when it
breaks through. However, a signicant volume of
dry gas has to be injected to produce tangible
amounts of condensate.
Engineers monitored the process in both
injection and production wells using gas-liquid
and gas-adsorption chromatography (below).
25
Since the injection gas did not contain nitrogen,
the nitrogen content was used as the indicator
of formation gas.
26
The 1993 pilot test program was expanded to
additional pilot locations in 1997, 2003 and 2004.
By the middle of 2005, the operator had injected
10 x 10
9
m
3
[354 x 10
9
ft
3
] of dry gas into the
pilot wells, and recovered a significant amount
of liquid. Comparing the recovery with estimates
of production through depletion alone showed
that the pilot area produced an additional
785 thousand tons [9.45 million bbl] of C
2
to C
4
and 138 thousand tons [1.22 million bbl] of C
5+
.
27
The operators also ran single-well pilot
projects in Vuktyl field. Although blockage was
not severe enough to cause a dramatic drop in
productivity in this field, the operator sought
ways to counteract the increased saturation that
had formed around wells. The treatment
included injecting solventa mix of ethane and
propaneinto a well, followed by dry gas. After
a sufficient volume of injection, the well was
returned to production.
When the solvent contacts the trapped
condensate, the solvent, formation gas and
condensate mix freely into a single phase. The
dry gas that follows is able to mix freely with the
solvent mixture. Thus, when the well produces
again, the injected gas, solvent and condensate
are produced as a single fluid. As a result,
the condensate saturation is at or near zero in
the treated zone. As formation gas follows the
mixture back through the treated zone, a zone of
increased condensate saturation will reform,
but well productivity can be improved by
periodic treatments.
Treatment volumes ranged from 900 to
2,900 tons [10,240 to 33,000 bbl] of solvent and
1.2 to 4.2 million m
3
[42 to 148 million ft
3
] of dry
gas.
28
Although the effectiveness varied from well
to well, the treatments generally had good
results. The productivity of four of the wells
increased by 20% to 40% over a period ranging
from 6 months to 1.5 years, followed by a decline
to the original production levels (next page).
Modeling Condensate Blockage
Reservoir-simulation models are commonly used
to predict the performance of gas-condensate
fields. The models incorporate rock and fluid
properties to predict the dynamic influence of
condensate blockage on gas and condensate
production. However, a typical gridblock of a
full-eld model (FFM) can be much larger than
the blockage zone, so a coarse grid model may
signicantly overestimate well deliverability.
The most accurate way to determine near-
well behavior of a gas-condensate field is by
using a simulator with a ne grid. There are two
ways to do this: use a FFM with local grid
refinement (LGR), or use a single-well model
with a ne grid near the well.
Modern simulators, such as the ECLIPSE 300
reservoir simulation software, include capability
for LGR. Small gridblocks can be used near
wellbores or other featuressuch as faults
that can signicantly impact local ow. Farther
22 Oileld Review
>
Dry-gas injection pilot. Separator gas injected into three wells designated 269, 270 and 273
vaporized stranded condensate for production from surrounding wells (top). Dry gas (blue) broke
tthrough a few months after the pilot began (middle). Nitrogen in the produced gas (green) gradually
decreased, indicating that less formation gas was being produced. The liquid C
5+
fraction (red)
indicates a slow decline after gas breakthrough. The results show signicant production of formation
gas, light (C
2
to C
4
) and intermediate components (C
5+
) from both produced formation gas and
remobilized stranded condensate (table, bottom).
Oct
93
Jan
94
April
94
July
94
Oct
94
Jan
95
April
95
July
95
Oct
95
Jan
96
April
96
July
96
Oct
96
Jan
97
April
97
July
97
Oct
97
Jan
98
C
o
m
p
o
n
e
n
t

c
o
n
t
e
n
t
,

m
o
l
e

%
0
1
2
3
4 40
30
20
10
0
G
a
s

f
r
a
c
t
i
o
n
,

%
Date
Component from:
Formation gas,
million m
3
Formation gas,
thousand tons
Stranded
condensate,
thousand tons
Dry gas,
million m
3
5,973
1,996
380
238
208
Produced gas
Injected gas
Produced C
2
to C
4
Produced C
5+
10,035
7,366
130
270
129
269
7
195
158
273
133
254
151
128
127
100
2,700 m
2,700 m
2 600 2,600
2,500 2,500
2,400 2,400
2 300 2,300
2,200 2,200
2,100 2,100
2,100 2,100
2,200 2,200
2,300 2 300
131/150
132
Injector well
Production well
Pilot area
Winter 2005/2006 23
away from such features, the gridblocks grow to
a size typical of a FFM. The cost of using LGR
may be a significant increase in computation
time in some cases.
Another way to examine gas-condensate
blockage effects is by using a single-well model.
In many cases, radial symmetry allows a well to
be treated in a two-dimensional model, using the
dimensions of height and radial distance. The
gridblocks nearest the well are small, nominally
half a foot [about 15 cm] in the radial direction.
The radial dimension increases with each
gridblock away from the wellbore, until it
reaches a maximum size used for the rest of the
model. The fine grid provides good resolution
where the flow is highest and the formation
saturation behavior is at its most complex.
Capillary, viscous and inertial forces can be
appropriately modeled. Far from the wellbore,
conditions of pressure and flow can be taken
from a FFM and applied as boundary conditions.
Sometimes, gas-condensate reservoir
simulations can be performed using a black-oil
model. This type of model assumes that there
are only two hydrocarbon components in the
fluid, oil and gas, and it allows for some
pressure-dependent mixing of gas in oil. This
model is inappropriate when the compositions
change signicantly with time, such as through
gas injection, or when there is a significant
compositional gradient. In those cases, a
compositional model with many hydrocarbon
components is necessary. In addition, some
black-oil models do not include capillary
number effects, which are important for
determining well deliverability.
Another way to account for condensate
blockage in a full-eld model is through the use
of pseudopressures. The equation for ow of gas
from a reservoir to a wellbore can be expressed
in terms of a pseudopressure, which is an
integral over pressure. By separately treating
the three regions described beforetwo-phase
flow near the well, gas flow with condensate
buildup next, and single-phase gas ow far from
the wellit is possible to calculate the
pseudopressure from the producing gas/oil ratio,
PVT properties of the fluid, and gas and oil
relative permeabilities.
29
As discussed previously,
the constant-composition expansion condition in
the first region simplifies the relative-
permeability ratios. This pseudopressure
method adds little time to running a FFM.
Pseudopressure methods have also been
implemented in spreadsheet format.
30
These
spreadsheets assume a homogeneous reservoir
and a simple black-oil model. They provide fast
predictions that can be used when many
sensitivity runs are necessary. A similar
semianalytical method was combined with the
effects of non-Darcy flow and permeability
layering. Comparisons using a compositional
simulator with a fine grid showed that
the semianalytical method captured all the
near-well effects accurately and was easy to
embed in a FFM at essentially no increase in
computational time.
31
Modeling Behavior Around a Fracture
Reservoir simulation modeling was used to
determine the effectiveness of fracturing in the
SW Rugeley field in south Texas, USA. This
field produces gas condensate from low-
permeabilityabout 1-mDFrio sand. A well in
this field, which was drilled and completed by
Wagner & Brown, was hydraulically fractured
initially, but a rapid decline in productivity led
the company to refracture the formation about
three months later, in June 2002. Productivity
improved, but then continued to decline over
the next few months. The drawdown in the
vicinity of the well was below the dewpoint
pressure, so the company investigated
the accumulation of condensate saturation
around a fracture.
Engineers at Schlumberger developed a
homogeneous, radially symmetric, single-well
model. This simple model demonstrated that
condensate blockage could result in a rapid
falloff in productivity. It also provided a means
to quickly check the impact of permeability
reduction due to compaction caused by
pressure decline.
23. Ter-Sarkisov RM, Zakharov FF, Gurlenov YM, Levitskii KO
and Shirokov AN: Monitoring the Development of
Gas-Condensate Fields Subjected to Dry Gas Injection.
Geophysical and Flow-Test Methods. Moscow: Nedra, s
2001 (in Russian).
Dolgushin NV (ed): Scientic Problems and Prospects of
the Petroleum Industry in Northwest Russia, Part 2: The
Development and Operation of Fields, Comprehensive
Formation and Well Tests and Logs, A Scientic and
Technical Collection. Ukhta: SeverNIPIgaz, 2005
(in Russian).
Ter-Sarkisov et al, reference 21.
Ter-Sarkisov, reference 19.
24. For a laboratory study of methane injection into cores
with condensate saturation: Al-Anazi HA, Sharma MM
and Pope G: Revaporization of Condensate with
Methane Flood, paper SPE 90860, presented at the
SPE International Petroleum Conference, Puebla,
Mexico, November 89, 2004.
25. Dolgushin, reference 23.
26. Vyakhirev et al, reference 20.
27. Dolgushin, reference 23.
28. Gritsenko AI, Ter-Sarkisov RM, Shandrygin AN and
Poduyk VG: Methods of Increase of Gas Condensate
Well Productivity. Moscow: Nedra, 1997 (in Russian). y
The density of the solvent mixture is 553 kg/m
3
.
29. Fevang and Whitson, reference 11.
30. Mott R: Engineering Calculations of Gas-Condensate-
Well Productivity, SPE Reservoir Evaluation &
Engineering 6, no. 5 (October 2003): 298306.
31. Chowdhury N, Sharma R, Pope GA and Sepehrnoori K:
A Semi-Analytical Method to Predict Well Deliverability
in Gas-Condensate Reservoirs, paper SPE 90320,
Exhibition, Houston, September 2629, 2004.
>
Changes in well productivity as a result of injection of ethane and
propane followed by dry gas. The difference of the squares of the reservoir
pressure, P
R
, and the bottomhole pressure,
R
P
BH
, as the ow rate increases
H
provides a measure of productivity. Before treatment (blue), the well
required a larger pressure difference to produce than it needed after
ttreatment (red). Four months after treatment, productivity had degraded
slightly (green), but it was still signicantly better than productivity before
tthe treatment.
P
2
R
e
s
e
r
v
o
i
r

-

P
r
2
B
o
t
t
o
m
h
o
l
e
,

M
P
a
2
0
4
8
12
16
0 50 100 150 200 250
Gas-condensate mixture production, thousand m
3
/d
With these results in hand, Wagner & Brown
had Schlumberger develop a more detailed
reservoir model, using ECLIPSE 300 reservoir
simulation software (above). The model was
refined by history-matching to the gas
production rate, which also provided a good
correlation to the condensate production.
Drawdown in the fracture induced the buildup
of condensate saturation along the fracture
(next page). The average reservoir pressure
dropped below the 6,269-psi [43.22-MPa]
dewpoint pressure during the modeled period.
With a good history-match, Wagner & Brown
could determine whether the fracture provided
signicant gains in productivity. The model was
rerun without the fracture, which resulted in a
production curve that continued the previous
decline rate (left). The difference between the
nonfractured case and the measured production
indicates the success of the fracture job. Over a
seven-month period, the cumulative production
attributed to the fracture job was 256 million ft
3
[7.25 million m
3
] of gas and 15,300 bbl [2,430 m
3
]
of condensate. This modeling study veried the
success of a eld application.
24 Oileld Review
>
History-match of model of SW Rugeley eld with a hydraulic fracture. The ECLIPSE 300 model of one well in the Frio sand has small grids around the
well and along the fracture (top left). Smaller grids were also placed at the fracture tips. The eld gas-rate history was matched by the simulation ( tt top
right), yielding good results for condensate rate ( tt bottom right). The changes in production after the fracture job were due to fracture cleanup and tt
changes in pressure in owlines. The model indicated the average reservoir pressure dropped below the 6,269-psi dewpoint pressure during this
production period (bottom left). tt
Fracture Well
Mar
2002
April
2002
May
2002
June
2002
July
2002
Aug
2002
Sept
2002
Oct
2002
Nov
2002
Dec
2002
Jan
2003
A
v
e
r
a
g
e

r
e
s
e
r
v
o
i
r

p
r
e
s
s
u
r
e
,

p
s
i
13,000
11,000
9,000
7,000
5,000
3,000
G
a
s

p
r
o
d
u
c
t
i
o
n

r
a
t
e
,

m
i
l
l
i
o
n

f
t
3
/
d
0
1
2
3
4
5
Mar
2002
April
2002
May
2002
June
2002
July
2002
Aug
2002
Sept
2002
Oct
2002
Nov
2002
Dec
2002
Jan
2003
C
o
n
d
e
n
s
a
t
e

p
r
o
d
u
c
t
i
o
n

r
a
t
e
,

b
b
l
/
d
0
1
2
3
4
5
Field data
Simulation
>
The hydraulic fracture effect. Rerunning the Frio well model with no fracture generated a simple
decline curve indicating a signicant productivity increase could be attributed to an induced fracture.
Mar
2002
April
2002
Production history
Model with no fracture
G
a
s

p
r
o
d
u
c
t
i
o
n

r
a
t
e
,

m
i
l
l
i
o
n

f
t
3
/
d
May
2002
0
1
2
3
4
5
June
2002
July
2002
Aug
2002
Sept
2002
Oct
2002
Nov
2002
Dec
2002
Jan
2003
Winter 2005/2006 25
2,500
3,750
5,000
6,250
7,500
0
0.1
0.2
0.3
0.2
0.4
0.6
0.8
1.0
R
e
s
e
r
v
o
i
r

p
r
e
s
s
u
r
e
,

p
s
i
C
o
n
d
e
n
s
a
t
e

s
a
t
u
r
a
t
i
o
n
,

f
r
a
c
t
i
o
n
G
a
s

r
e
l
a
t
i
v
e

p
e
r
m
e
a
b
i
l
i
t
y
,

f
r
a
c
t
i
o
n
July 15, 2002 July 25, 2002
2,500
3,750
5,000
6,250
7,500
0
0.1
0.2
0.3
0.2
0.4
0.6
0.8
1.0
R
e
s
e
r
v
o
i
r

p
r
e
s
s
u
r
e
,

p
s
i
C
o
n
d
e
n
s
a
t
e

s
a
t
u
r
a
t
i
o
n
,

f
r
a
c
t
i
o
n
G
a
s

r
e
l
a
t
i
v
e

p
e
r
m
e
a
b
i
l
i
t
y
,

f
r
a
c
t
i
o
n
August 20, 2002 September 20, 2002 December 20, 2002
<
C d bl k d Condensate blockage around a
fracture, Frio model. For each time
step, model results indicate
pressure decline (top), condensate
saturation (middle) and gas relative
permeability (bottom). The rst two
ttime steps in July 2002 (left) focus tt
on the immediate vicinity of the
fracture and the later three time
steps (below) show a wider view w
of the whole model area. Pressure
declines rapidly along the fracture
(left, top). The approximate dewpoint p
prole (oval curves) expands outward
from the fracture. The low gas
permeability around the fracture
at later time steps indicates the
condensate blockage.
Application of Best Practices
Chevron recently completed a study of ve gas-
condensate reservoirs that are at different
stages of development. The objective was to
transfer best practices among various
development teams.
One of the fields in the study, a North Sea
reservoir, is a marine turbidite with gross-pay
interval of more than 120-m [400-ft] thickness.
The average reservoir permeability is 10 to
15 mD, with average porosity of 15%. The
original reservoir pressure of 6,000 psi
[41.4 MPa] is a few hundred psi [a few Mpa]
above the dewpoint pressure, although the
dewpoint varies from east to west.
32
The bottomhole pressure was below the
dewpoint from first production. The
condensate/gas ratio ranged from 70 bbl per
million ft
3
[393 m
3
per million m
3
] in the east to
110 bbl per million ft
3
[618 m
3
per million m
3
] in
the west. Some wells experienced a productivity
reduction of about 80%, most of which occurred
in early production.
Chevron followed a step-by-step procedure to
understand and history-match the fields gas-
condensate behavior. The operator selected core
samples that spanned the range of permeability
and porosity of the field and fluids that mimic
reservoir-fluid behaviorliquid dropout as a
function of pressure, viscosity and interfacial
tensionat lower temperature. The company
measured relative permeability over a range
of flow conditions and fitted those data
to several relative-permeability models for use
in simulators.
A spreadsheet using an analytical
pseudopressure method was used to calculate
deliverability. The calculation showed that
productivity index (PI) decreased from about
80 to about 15 thousand ft
3
/d/psi [33 to
6 thousand m
3
/d/kPa], with little difference
based on bottomhole pressure until late in eld
life (above).
A detailed single-well, compositional flow
simulation using the Chevron CHEARS reservoir
simulator was performed with realistic geology.
Far-eld boundary conditions came from a full-
field model. The simulation honored well
production practices and differential depletion
in the field. The predictions provided a good
match to results from three vertical wells and
one inclined well (next page).
This study led to several initiatives in the
field. Hydraulic fracturing to improve
productivity is an active effort in this field, so
these models are being used to better
understand fracture effectiveness. In addition,
lessons learned from this field regarding the
impact of condensate blocking have been used
extensively in planning for wells in new projects
in other gas-condensate elds.
A Fundamental Alteration
The high price of natural gas on world markets
in recent years has stimulated interest in
developing gas reservoirs. Companies seek new
ways to optimize their gas-condensate resources.
Hydraulic fracturing can mitigate the effect
of condensate blockage, but it does not
eliminate the accumulation of condensate in
26 Oileld Review
>
Spreadsheet model results for a North Sea well. A homogeneous single-well model in a spreadsheet
provided a way to quickly examine different effects. For example, the bottomhole pressure had little
effect on gas productivity index, PI.
0
10
20
30
40
G
a
s

P
I
,

t
h
o
u
s
a
n
d

f
t
3
/
d
/
p
s
i
7,000 6,000 5,000
Reservoir pressure, psi
Bottomhole
pressure, psi
2,000
1,250
1,000
500
4,000 3,000 2,000
50
60
70
80
90
32. Ayyalasomayajula P, Silpngarmlers N and Kamath J:
Well Deliverability Predictions for a Low Permeability
Gas Condensate Reservoir, paper SPE 95529, presented
at the SPE Annual Technical Conference and Exhibition,
Dallas, October 912, 2005.
33. Fahes M and Firoozabadi A: Wettability Alteration to
Intermediate Gas-Wetting in Gas/Condensate Reservoirs
at High Temperatures, paper SPE 96184, presented at
the SPE Annual Technical Conference and Exhibition,
Dallas, October 912, 2005.
34. Kumar V, Pope G and Sharma M: Improving Gas and
Condensate Relative Permeability Using Chemical
Treatments, paper SPE 100529, to be presented at
the SPE Gas Technology Symposium, Calgary,
May 1518, 2006.
Winter 2005/2006 27
areas where the pressure in the formation is
below dewpoint. Dry gas and solvent injections
are able to mobilize some condensate, but the
liquid saturation profile near a producing well
reforms and the blockage effect returns.
New alternatives are being examined in
laboratories. For example, some studies have
focused on ways to prevent fluid buildup by
altering reservoir-rock wettability.
Although mineral surfaces such as quartz,
calcite and dolomite prefer to be wetted by
liquids rather than gas, there are solids that
have a gas-wetting preference. In particular,
fluorinated compounds such as Teflon surfaces
are gas-wetting. So, fluorinated solvents have
been used to alter the wettability of cores.
Recently reported results at high temperature
140C [284F]typical of gas-condensate
reservoirs showed a strong reversal of wetting in
a gas-water-reservoir rock system, but was less
successful in a gas-oil-reservoir rock system.
33
Researchers at the University of Texas at
Austin conducted laboratory tests using 3M
fluorocarbon surfactants.
34
The results on
reservoir core samples blocked by condensate
indicate about a doubling of the gas and
condensate relative-permeability values after
treatment. Based upon these promising
laboratory data, Chevron may test this treatment
in a blocked gas-condensate well sometime in
2006. Treatments such as these must be field
tested under a variety of conditions to fully
develop and prove the technology. If the
technique is ultimately successful, then the cost
of the surfactants used in the treatment will be
very small compared to the benets of increased
gas and condensate production rates.
The alteration these solvents make in the
rock addresses a fundamental cause of
condensate blockage: capillary accumulation of
liquid because of the wetting preference of the
rock. Avoiding liquid buildup alleviates the
problem of choking production, so that a high
production rate can be achieved. MAA
>
Single-well simulation results. The simulator gave a good match to both
gas PI (top) and bottomhole pressure (middle) for behavior in a North Sea
well. Different layer properties resulted in different extents of condensate
saturation buildup (bottom).
G
a
s

P
I
,

t
h
o
u
s
a
n
d

f
t
3
/
d
/
p
s
i
100
90
80
70
60
50
40
30
20
10
0
Time, yr
Simulation Simulation Simulation
Field data Field data Field data
B
o
t
t
o
m
h
o
l
e

p
r
e
s
s
u
r
e
,

p
s
i 2,500
3,000
3,500
2,000
1,500
1,000
500
0
Time, yr
Field data Field data
Si l i Simulation
Oil saturation, fraction
0 0.4 0.2
B
o
r
e
h
o
l
e
3.5 4.0 4.5 5.0 5.5
3.5 4.0 4.5 5.0 5.5

Gas Condensate PVT
Whats Really Important and Why?

Curtis H. Whitson
a,b,c

ivind Fevang
b

Tao Yang
a,b

(a) Norwegian U. of Science and Technology (NTNU), (b) PERA a/s
(c) Corresponding author, curtis@ipt.ntnu.no / whitson@pera.no

Paper ( Curtis H. Whitson) presented at the IBC Conference
Optimisation of Gas Condensate Fields, London, Jan. 28-29, 1999, Cumberland Hotel.
IBC UK Conferences Ltd., Gilmoora House, 57-61 Mortimer Street, London W1N 81X, UK
http://www.ibc-uk.com

ABSTRACT

This paper gives a review of the key PVT data dictating recovery and well
performance of gas condensate reservoirs. The importance of specific PVT data are
put in the context of their importance to specific mechanisms of recovery and flow
behavior. Phase behavior important to gas cycling projects is also covered. Modeling
gas condensate reservoir fluid systems with an equation of state is discussed, as is
EOS modeling of complex fluid systems with strongly varying compositions and PVT
properties.

INTRODUCTION

It could be argued that the engineering of a gas condensate field is 80% traditional
gas engineering, and 20% extra engineering. The numbers could be 90/10 or
70/30 but the majority of engineering of any gas condensate field is always the
same as the engineering of a gas reservoir without condensate.

The main difference between a gas condensate field and a dry gas field is the
additional income derived from surface condensate production. Condensate
production evolves from produced reservoir gas (= produced wet gas = produced
wellstream) as the wellstream is processed at the surface. The production of
reservoir gas can, for the most part, be handled with traditional gas engineering
tools.

Gas Condensate PVT Whats Really Important and Why?
C.H. Whitson, . Fevang, and T. Yang

2
From an engineering point of view, the two extra issues which must be addressed
in a gas condensate reservoir are:

How the condensate yield will vary during the life of a reservoir, and
How two-phase gas/oil flow near the wellbore affects gas productivity.

Both of these issues are strongly related to the PVT properties of the fluid system
(though productivity is more affected by relative permeability effects).

PVT properties important to the engineering of all gas condensate reservoirs
includes:

Z-factor
Gas viscosity

and a few extra properties needed to handle the condensate part of a gas-
condensate reservoir:

Compositional (C
7+
) variation with pressure
Oil viscosity and liquid dropout

Dewpoint pressure is implicitly defined by the pressure dependence of compositional
variation. As discussed below, the dewpoint is less importance than is commonly
thought.

The PVT properties listed above are particularly important to reservoirs produced by
pressure depletion. For gas condensate reservoirs undergoing gas cycling it may
also be important to quantify phase behavior (vaporization, condensation, and near-
critical miscibility) which develops in gas cycling below the dewpoint.

Compositional grading in gas condensate reservoirs may be important to the design
of well placement, estimation of in-place surface volumes, reserves, and prediction
of fluid communication vertically (between geologic layers) and areally (between fault
blocks). Prediction of a potential underlying oil is often required in discovery wells
which are drilled upstructure and encounter only gas which is near-saturated. Here,
accurate sampling and PVT modeling are paramount.

A PVT model
a
should describe accurately the key phase, volumetric, and viscosity
behavior dictating the key processes affecting rate-time performance and ultimate
recoveries of surface gas and oil. Unfortunately, a PVT model may not be capable of
accurately describing all PVT properties with equal accuracy. EOS models often
have difficulty matching retrograde phenomena (compositional variation of gas, and
liquid dropout), particularly when the system is near-critical, or only small amounts of
condensation occur just below the dewpoint (tail-like retrograde behavior). Oil

a
We define a PVT (pressure-volume-temperature) model to include the EOS (equation of state)
model describing phase and volumetric behavior, together with the viscosity model (e.g. the Lorentz-
Bray-Clark
1
or Pedersen et al.
2
); the viscosity model is not formally linked with the EOS model but
makes use of EOS-calculated density.

3
viscosities are also difficult to predict for reservoir condensates, and measured oil
viscosities are not usually available for tuning the viscosity model.

Consequently, it is important to determine which PVT properties are most important
to the accurate engineering of reservoir and well performance for a given field
development. Different fields require different degrees of accuracy for different PVT
properties, dependent on field development strategy (depletion vs. gas cycling), low
or high permeability, saturated or highly undersaturated, geography (offshore vs.
onshore), and the number of wells available for delineation and development.

Consider the following examples.

Example 1. A small offshore satellite reservoir with high permeability (kh=4,000
md-m), initially undersaturated by 400 bar, and with a test yield of 300 STB/MMscf.

Example 2. A large offshore deep-water reservoir with moderate permeability
(kh=1000 md-m), initially saturated or near-saturated (?), large structural relief, and a
test yield of 80 STB/MMscf. A single (and very expensive) discovery well has been
drilled.

Example 3. An onshore old undeveloped gas cap with well-defined initial volume
(by production oil wells and pressure history), uncertain initial composition (estimated
initial yield of 120 STB/MMscf), partially depleted due to long-term production of
underlying oil, and low permeability (kh=300 md-m).

These three examples require significantly different emphasis in the treatment of
PVT data. Why? The reason lies in a coupling of the reservoir and well performance
with PVT properties. Every gas condensate reservoir provides a new example with a
different set of conditions requiring different emphasis on which PVT data are
important.

This paper will attempt to explain when and why various PVT properties are
important to the development of a particular gas condensate field.

PVT EXPERIMENTS

Constant Composition (Mass) Expansion Test
The constant composition expansion (CCE) test, sometimes referred to as a
constant-mass expansion test, is used to measure dewpoint pressure, single-phase
gas Z-factors, and oil relative volume below the dewpoint (liquid dropout curve). A
sample of reservoir fluid is charged in a visual PVT cell and brought to reservoir
temperature and a pressure sufficiently high to ensure single-phase conditions.
Pressure is lowered by increasing cell volume until a liquid phase is visually detected
(through a glass window). Total cell volume and liquid volume are monitored from
the initial reservoir pressure down to a low pressure (dictated by cell and sample
size).


4
Constant Volume Depletion Test
The constant volume depletion (CVD) test is an extremely important laboratory test
which monitors the phase and volumetric changes of a reservoir gas sample (at
reservoir temperature) as the pressure drops below the dewpoint and equilibrium
gas phase is removed. The CVD test simulates closely the actual behavior of a gas
condensate reservoir undergoing pressure depletion, and results from the lab
measurements can be used directly to quantify recoveries of surface gas and
condensate as a function of pressure below the dewpoint. Combined with single-
phase Z-factors from the CCE test, a complete prediction of depletion behavior
(recoveries and liquid-yield variation) can be accurately predicted from initial
pressure to abandonment.

The CVD test involves stepwise lowering the pressure below the dewpoint, with an
associated increase in cell volume. After equilibration at each pressure, enough
equilibrium gas is removed from the top of the cell to bring the cell back to the
original volume occupied at the dewpoint. The amount of gas removed, its
composition and Z-factor, and the remaining oil volume in the cell are measured and
reported. A two-phase Z-factor is also reported for use with the gas material
balance (see discussion below).

Accurate measurement of the removed gas composition is very important to the
prediction of condensate recovery and liquid-yield variation much more important
than accurate measurement of retrograde oil volumes. Special laboratory procedures
should be followed to ensure accurate CVD compositional measurements (e.g.
appropriate heating of tubing used to remove equilibrium gas from the cell).
Measurement of the final low-pressure condensate composition allows an important
backward material balance check.
b

INITIAL FLUIDS IN PLACE AND DEPLETION RECOVERIES

Gas Z-factor
Muz Standing would be happy to hear that the Z-factor is (still) the only PVT property
which always needs accurate determination in a gas condensate reservoir (as in a
dry gas reservoir). The reason is (1) to get an accurate and consistent estimate of
the initial gas (and condensate) in place, and (2) to accurately predict the gas (and
condensate) recovery as a function of pressure during depletion drive.
c

Single-phase gas Z-factors are measured experimentally at reservoir temperature
and pressures from the initial reservoir pressure to the dewpoint. These data are

b
The backward material-balance check is made by starting with the final condensate composition and
amount (using the reported final oil relative volume and properties), adding incrementally the removed
gas from each CVD step, and ending up with a check of the original fluid composition. This check is
insensitive to oil relative volume (except the final value), a big advantage over the traditional forward
material balance which is extremely sensitive to all relative oil volumes and, consequently, can not be
used for leaner gas condensates.

c
It also is important to quantify recovery by depletion for a reservoir which is to be gas cycled,
because the evaluation of economics will always involve comparison of the gas cycling project with
depletion drive.


5
reported as part of the constant composition expansion test. Z-factors used in the
material balance equation below the dewpoint are back-calculated from data in the
constant volume depletion test. These so-called two-phase Z-factors are pseudo
(not-physical) properties which should only be used in the traditional gas material
balance equation.

It is not commonly recognized that condensate recovery is strongly related to the
recovery of gas. Gas Z-factors dictate gas and oil recoveries during depletion
(together with the amount of water expansion and influx), because recoveries are
proportional to (Z
i
/Z) i.e. the term [1 - (Z
i
/Z)(p/p
i
)]. In fact, at average reservoir
pressures above the dewpoint, condensate recovery exactly equals the gas
recovery. Consequently, condensate recoveries are strongly dependent on accurate
description of gas-phase Z-factors (both above and below the dewpoint).

As an example for a high-pressure reservoir, a +5% error in Z
i
and a 5% error in Z
at the dewpoint will result in (a) a +5% error in initial gas and initial condensate in
place, and (b) a +5 to +10 recovery-% error in recovery of gas and condensate at the
dewpoint.

Compositional (C
7+
) Variation During Depletion
As mentioned earlier, the distinguishing characteristic of a gas condensate field is
the added value from condensate production, in addition to gas. Surface condensate
is, for practical purposes, the C
7+
content
d
of the produced wellstream. This
simplification makes the treatment of many engineering calculations easier to
understand without loosing engineering accuracy.

The condensate rate profile is easy to convert into an economic profile, and
engineers can readily relate the two. But how can we readily forecast the condensate
rate profile for a gas condensate reservoir?

For a surface gas-rate production profile q
g
(t), the profile of oil rate versus time is
given (approximately) by
CVD og CVD 7
CVD 7
g o
) C (
1
) y ( 1
) y (
) t ( q ) t ( q

+
+
........................................................................ (1)
o
o
sc
sc
og
M P
T R
C

= ....................................................................................................... (2)
where (
o
/M
o
) = (
7+
/M
7+
)
CVD
. Time dependence of CVD properties must be
correlated with cumulative wet-gas volumes produced,
= dt q G
w pw
............................................................................................................ (3)
where
q
w
= q
g
+ q
o
C
og
........................................................................................................ (4)

d
The simplification of surface condensate being essentially C
7+
is a subjective choice. One could
easily have chosen C
6+
or C
5+
without any change in our comments.


6
Given the q
w
(t) profile and G
pw
(t), this can be translated into cumulative wellstream
produced from the CVD test, (n
p
/n
d
)
CVD
,
(n
p
/n
d
) = G
pw
/G
w
[1 (p/z)
d
/(p/z)
i
] ; p > p
d
............................................................. (5)
where all CVD properties = initial gas properties for p > p
d
. C
og
represents the
surface gas equivalent of one surface oil volume.

Equations for converting CVD results to approximate surface product recoveries,
including the depletion recovery from initial to dewpoint pressure are:
( )
( )
( )
( )
( )
( )
og sk
og si
k
N
1 k d
p
i
d
i
d
gD
C r 1
C r 1
n
n
Z / p
Z / p
Z / p
Z / p
1 RF
+
+
=

=
........................................... (6)
( )
( )
( )
( )
( )
( )
og sk
og si
k
N
1 k d
p
i
d
i
d
oD
C r / 1
C r / 1
n
n
Z / p
Z / p
Z / p
Z / p
1 RF
+
+
=

=
........................................... (7)
og 7
7
s
C
1
z 1
z
r
+
+
...................................................................................................... (8)
A simple spreadsheet calculation using these equations allows quick translation of
laboratory CVD data to important engineering and commercial quantification of in-
place surface volumes, reserves, and production forecasts. Fig. 1 and Table 1
shows an example calculation using Eqs. 6-8 for a rich-gas condensate.

Compositional Variation with Depth
Numerous field case histories have shown that composition varies with depth in
petroleum reservoirs. Component segregation due to gravitational forces is usually
given as the physical explanation for the variation in composition. A theoretical
model for such variation was already defined by Gibbs in the late 1800s for an
isothermal system under the influence of a constant force field such as gravity. The
result of gravitational segregation is that a gas condensate gets richer at greater
depths, with increasing C
7+
mole fraction (and dewpoint pressure)
3
.

Not all fields show compositional gradients with depth as predicted by the isothermal
model. Some fields show practically no gradient over large depths, such as the
Cupiagua field in Columbia
4
where a near-critical gas condensate with more-or-less
constant composition is found over an interval of some 2000 m. Some oil fields have
gradients larger than predicted by the isothermal model. Hier
5,6
made
comprehensive calculations using a number of thermal diffusion models which
indicate that thermal gradients typically reduce compositional gradients in gas
condensate fluids, while for oils a thermal gradient may cause either a reduction or
an increase in compositional gradients.

Our concern with compositional gradients in gas condensates will be limited to three
topics: (1) assessing the effect of a gradient on in-place surface volumes, (2)
assessing the prediction of a gas-oil contact using a theoretical gradient model, and
(3) the impact of compositional gradients on depletion (and cycling) recoveries.


7
Variation in C
7+
composition with depth will obviously affect the calculation of initial
surface condensate in place, compared with a calculation based on a constant
composition. Depending on the location of the sample used in the constant-
composition model, either smaller or larger initial condensate volumes in place can
result when compared with a gradient model
e
.

The gradient model will typically give an optimistic in-place sensitivity when the gas
condensate reference sample is upstructure. If the reference gas sample is
downstructure then the gradient model will provide a pessimistic in-place
sensitivity. For reservoirs with limited control of fluid variation with depth, we
recommend using the in-situ representative samples available, linear interpolation
between these samples, and for sensitivities use both (a) the gradient model for
extrapolation beyond the samples and (b) a constant-composition extrapolation.

Another interesting feature of near-saturated reservoirs with compositional gradients
is that the recoverable condensate volumes by depletion are insensitive to whether
the model is initialized with or without a compositional gradient. This lack of
sensitivity
f
will not be apparent if comparisons are made using recovery factors
(because initial volumes in place can be quite different for the two models of
compositional initialization).

A gradient model may also predict a transition from gas to oil which can dramatically
affect the initial oil in place volumes (see discussion below). However, a predicted
GOC from an upstructure gas sample is, at best, a possibility. Results of predicted
oil zones from gas samples should only be used for sensitivity analysis of a new
discovery, or in a reservoir where additional delineation wells are not planned.

Dewpoint Pressure
Strictly speaking the dewpoint pressure is the pressure where an incipient liquid
phase condenses from a gas phase. Practically, the dewpoint marks the pressure
where (1) reservoir gas phase composition changes and becomes leaner, and (2)
condensate accumulation starts in the reservoir. These two changes can have a
profound effect on reservoir and well performance or, they may have little impact.
The importance of the actual dewpoint pressure will vary from reservoir to reservoir,
but in most situations accurate dewpoint determination is not important.

Why? First, in the context of compositional variation with pressure (and associated
variation of condensate yield with pressure) accurate determination of the
thermodynamic dewpoint pressure is not of particular importance. In fact, we dont
need to know the specific dewpoint at all as long as the variation of composition
(C
7+
content) with pressure is well defined near the thermodynamic dewpoint.

e
The gradient model requires that a composition be specified at a reference depth with a reference
pressure and temperature.

f
This behavior is readily understood in a black-oil model using a solution oil-gas ratio r
s
(p) function.
The initial variation of r
s
(and dewpoint) with depth is short-lived when reservoir pressure drops below
the dewpoint, as r
s
of the equilibrium gas through the reservoir becomes (more-or-less) constant at
the average reservoir pressure. Because only reservoir gas flows into wells, the producing OGR will
reflect the equilibrium r
s
at average reservoir pressure, which becomes independent of depth when
p
R
<p
d
.


8

Second, when the bottomhole flowing pressure (BHFP) drops below the dewpoint
and two phases start flowing near the wellbore, gas relative permeability drops and
well productivity drops. However, as long as BHFP is anywhere near the dewpoint
the well will have excess deliverability i.e. we simply reduce the BHFP to produce
more gas (even though the well productivity is lower).

Only when the BHFP reaches a minimum value (dictated by some delivery-pressure
constraint) will the well no longer be able to deliver the desired rate. At this point,
well productivity is important. However, this occurs at a BH flowing pressure much
lower than the dewpoint. A typical minimum BHFP might be 50-150 bara, while
dewpoint pressures are typically 250-400 bara. Whether the BHFP drops below the
dewpoint at 400 or 350 bara has little impact on what the well will (or wont) produce
when BHFP reaches a minimum constraint of 100 bara.

Another (less common) need for dewpoint pressure is when an underlying saturated
oil zone may exist and a PVT model is used to predict the existence and location of
the gas-oil contact (GOC). In this case, the PVT model dewpoint should be tuned
precisely to an accurately measured dewpoint pressure. It is not uncommon that a
predicted GOC may vary 10s of meters per bar of uncertainty in the (PVT-model)
dewpoint pressure. Thus an accurate description of the dewpoint pressure will have
an impact on the prediction of initial oil and gas in place, placement of delineation
wells, and potential field development strategy. In this situation, accurate dewpoint
measurement and equally-accurate modeling of the measured dewpoint should be
given due attention.

On the other hand, if accurate treatment of the dewpoint pressure is not required (for
estimating a GOC), then we recommend using little if any weighting of the measured
dewpoint pressure when tuning the PVT model. Instead, priority should be given to
matching the variation of C
7+
with pressure in the removed gas from a constant
volume depletion test.

CONDENSATE BLOCKAGE

When BHFP drops below the dewpoint and two-phase gas/oil flow stabilizes in the
near-wellbore region, relative permeability to gas (the primary flowing phase) may
drop dramatically and the well deliverability is lowered accordingly. Saturations in the
near-wellbore region can reach 40-60%, with gas permeability reductions of 0.05-
0.2.

Flow in the near-wellbore region reaches a steady-state condition in a relatively short
time after the BHFP drops below the dewpoint. Flow theory shows that the produced
wellstream mixture is constant throughout the steady-state region, meaning that if
we captured the flowing mixture at any point within this region, its composition
would be the same as the producing wellstream mixture.
g

g
The flowing mixture composition at some point within the steady-state region will not equal to the
overall composition occupying the pore volume (in-situ composition) at that point, as the in-situ
composition will vary throughout the steady-state region.


9

Permeability reduction in the near-wellbore (steady-state) region is particularly
important because most of the pressure drop will be greatest in this region. A relative
permeability reduction of 0.1 in the first 10 meters from the wellbore will have
significantly greater impact than a k
rg
reduction of 0.1 at a 10-m radial interval some
100 m away from the wellbore.

The relative permeability ratio k
rg
/k
ro
in the steady-state (SS) region is given by
k
rg
/k
ro
=(1/V
ro
1)(
g
/
o
)............................................................................................ (9)
where V
ro
= V
o
/V
t
is the CCE oil relative volume of the produced wellstream at any
pressure within the SS region. The pressure in the SS region ranges from the BHFP
to the dewpoint pressure of the produced wellstream, and most deliverability loss
occurs nearest the wellbore where pressures are closer to the BHFP.

Given the k
rg
/k
ro
ratio throughout the SS region from Eq. 9, the relative permeability
to gas k
rg
can be found directly from the relative permeability relationship. That is,
k
rg
(S) = k
rg
(k
rg
/k
ro
(S)). Fevang and Whitson
7
have shown that the k
rg
/k
ro
ratio in the
SS region does not vary more than about one order of magnitude throughout the life
of a reservoir. With regard to uncertainties in PVT properties, and the need to
measure (or predict) their values for accurate description of condensate blockage,
we can conclude that oil viscosity should be given highest priority because it has the
largest uncertainty both experimentally and in predictions. Fig. 2 shows the effect on
k
rg
caused by a 20% error in V
ro
for values of V
ro
ranging from 0.5 for a near-critical
condensate to 0.005 for a very lean condensate (using
g
/
o
=0.02/0.2=0.1 and a
typical Corey relative permeability relation).

Oil Viscosity
As discussed above, oil viscosity is important in the proper modeling of condensate
blockage i.e. the two-phase gas/oil flow effect on gas relative permeability in the
region around the wellbore. Oil viscosity is usually low for reservoir condensates,
ranging from 0.1 to 1 cp in the near-wellbore region.
h
Measurement of condensate
viscosities is not made in routine laboratory tests, and it may be difficult to obtain
measurements for lean condensates (where volumes of condensate are small).
Viscosity correlations are typically unreliable for predicting low oil viscosities, and
some approach is needed to ensure accurate and consistent modeling of this
important property.

We recommend that the oil viscosity model be tuned to measured viscosities of a
separator condensate sample
i
at reservoir temperature and pressures in the range of
100-400 bara. More appropriate condensate viscosity measurements can be
designed (at greater expense), but having oil viscosity data from a separator oil

h
The viscosity of condensate flowing in the near-wellbore region throughout depletion, and
particularly when a well goes on decline, will remain fairly constant at a value close to the condensate
viscosity of the original reservoir fluid in the pressure range 100-200 bara. This viscosity is typically
lower than the viscosity of condensate from a CVD test in the same pressure range, where the
difference may be as much as a factor of 2-3.

i
This idea was suggested by Dr. Jeff Creek, Chevron Oil Company, ca. 1997.


10
sample to tune the viscosity correlation should ensure reasonably accurate oil
viscosity predictions of the condensate actually flowing in the near-wellbore region
when bottomhole flowing pressures drop below the dewpoint.

Gas Viscosity
Gas viscosity for most systems will vary from 0.02 to 0.03 cp for all pressure
conditions. For near-critical gas condensates and high-pressure gases the viscosity
may initially be 0.05 cp, but in most of the near-wellbore region experiencing
significant pressure losses the viscosity will be in the lower range of 0.02-0.03 cp.
Consequently, the absolute value of viscosity does not vary greatly for a given gas,
or from gas to gas system. Viscosity correlations are fairly reliable at predicting
accurate gas viscosities, within 5-10% in most cases.

What is important with respect to gas viscosities is that consistent viscosities be
used in all engineering applications e.g. well test interpretation, well performance
design, reservoir simulation, tubing calculations, pipeline calculations, etc. It is not
uncommon that a 15-25% difference in gas viscosities may result using different
correlations (by different engineering disciplines). This may result in similar
inaccuracies in well performance calculations, even where all flow is single-phase
gas!

Oil Relative Volume (Liquid Dropout Curve)
The oil relative volume or liquid dropout curve is perhaps the most familiar property
to engineers working with gas condensate fields. The maximum liquid dropout
j
is
often used as a subjective measure to characterize the richness or leanness of a gas
condensate fluid system (perhaps even more common than the liquid yield itself!).

Two definitions of oil relative volume V
ro
are used,
V
ro
= V
o
/ V
d
........................................................................................................... (10)
V
ro
= V
o
/ V
t
= V
o
/ (V
o
+V
g
) ..................................................................................... (11)
It is important to differentiate between the two definitions. The first and most
common definition is oil volume relative to the dewpoint volume, where this gives a
direct measure of the actual volume of oil condensed. The second and more
important definition (for engineering purposes) is oil volume relative to total gas+oil
volume, where the change in this V
ro
depends on two effects the change in oil
volume itself and the change in total volume, V
ro
(p) = V
o
(p)/V
t
(p). This latter definition
is more important because it enters directly in the condensate blockage problem,
and at lower pressures (<250 bara) where condensate blockage is particularly
important, the change in total volume V
t
(p) due to gas expansion becomes even
more important than the change in oil volume.

Ironically, the liquid dropout curve has little direct impact on reservoir and well
performance. Only the CCE liquid dropout V
ro
=V
o
/V
t
of a reservoir gas at lower
pressures has a (second-order) effect on the modeling of condensate blockage. The

j
The maximum liquid dropout often occurs at a pressure near 150-250 bara (though higher for near-
critical systems), and is approximately correlated to the minimum in equilibrium K-values (K
i
=y
i
/xi) of
heavier components (C
5+
).


11
average oil saturation in a gas condensate reservoir during depletion, given
approximately by the CVD experiment, is seldom important.

Interestingly, the magnitude of maximum liquid dropout does not determine whether
condensate blockage will or will not be a problem for a given reservoir. It only has a
second-order effect on the relative degree of severity. For example, one reservoir
with 35% maximum liquid dropout may have a condensate blockage effect which has
no importance to well deliverability, while another reservoir with 2% maximum liquid
dropout may have condensate blockage causing a dramatic well deliverability loss.

The importance of condensate blockage on well performance is dictated by the
relative importance of reservoir pressure losses compared to pipe
(tubing+flowline) pressure losses. For a high-kh (kh=10 000 md-m) rich condensate
well the blockage skin may be +30 with a resulting additional pressure loss of only 3
bar, where tubing pressure losses are 300 bar due to high deliverable rates. A lower-
kh (kh=500 md-m) lean condensate well may have a blockage skin of +15 with a
resulting additional pressure loss of 150 bar, where tubing+flowline pressure losses
are 150 bar. Clearly the lean condensate well has a more severe condensate
blockage problem than the rich high-kh well.

GAS CYCLING

Traditional gas cycling with full pressure maintenance is almost completely
unaffected by PVT properties. The gas-gas displacement process is fully miscible,
independent of the injection gas, so only the viscosity ratio of the two gases enters
into the displacement performance. For practical purposes, reservoir heterogeneities
(and mainly layering) dominate recovery performance of a gas cycling project
almost totally for full-pressure maintenance cycling, but also if the reservoir is cycled
below the dewpoint.

As indicated by Coats
8
and others, gas cycling projects below the dewpoint are also
affected by the vaporization characteristics of displacement gas on the retrograde
condensate. For most gas cycling projects the injection gas is fairly lean and
recovery efficiency of the reservoir retrograde condensate depends mostly on
vaporization.

If injection gas is rich in intermediate components C
2
-C
5
and gas cycling occurs
below the (original) dewpoint, an efficient condensing/vaporizing mechanism can
develop and, in some cases, develop full miscibility with the gas+condensate
reservoir system. This mechanism is described by Hier and Whitson
6
.

Gas Cycling Recovery Efficiency in Swept Zone
Let us look at the recovery mechanisms in a gas cycling project in the volume of the
reservoir swept by injection gas the microscopic or pore-level recovery. At a
given time and position in the swept zone, the pressure is either above or below its
original dewpoint when the injection gas front arrives.

If above the dewpoint, a gas-gas miscible displacement will yield 100% recovery of
the current condensate in place. A miscible displacement is guaranteed,

12
independent of the injection gas used, even though the injected gas may be first-
contact immiscible with the original reservoir gas. Miscibility develops by a simple
vaporizing mechanism.

If reservoir pressure is below the dewpoint when the displacement front arrives,
ultimate recovery of condensate is dictated by two processes: (1) gas-gas miscible
displacement of the reservoir gas, and (2) partial vaporization of the retrograde
condensate. The condensate recovery by gas-gas miscible displacement is 100% of
the condensate in solution in the reservoir gas at the time the front arrives. The
recovery efficiency of retrograde condensate by vaporization (E
v
) increases gradually
as increasing volumes of injection gas sweep this point in the reservoir.

We will try to discuss the two mechanisms of condensate recovery for a reservoir
undergoing cycling below the dewpoint. Before the gas front arrives, (1) the amount
of condensate in place in the gas-filled pores continuously decreases below the
dewpoint; and (2) the cumulative retrograde condensate in the oil-filled pores
continuously increases at decreasing pressures.

After the gas front arrives, (1) the gas-gas miscible displacement has a 100%
efficiency of the condensate remaining in solution in the reservoir gas; and (2) the
recovery efficiency of retrograde condensate by vaporization (E
v
) rises quickly but
then flattens quickly after the front passes. This behavior of E
v
is due to the
preferential vaporization of light C
7+
first, leaving behind a heavier condensate
which is less efficiently (more slowly) vaporized. The more volumes of injection gas
passing over the condensate, the less efficient vaporization becomes. Even if
pressure continues to decline, new condensation will not occur because the gas
behind the front is lean and has little dissolved condensate.

Accurate prediction of the changing vaporization efficiency requires (1) an accurate
description of the C
7+
molar distribution of the condensate, and (2) the K-values of
C
7+
components as a function of pressure and overall composition. Because total
condensate recovery efficiency in the swept region may be dependent on an
accurate description of the component-by-component vaporization process, effort
should be made to obtain compositional data which describes the vaporization
process. Extra effort should also be given towards fitting these compositional data
with the EOS model. However, it is worthwhile to first evaluate the potential for
recovery by vaporization of retrograde condensate below the dewpoint prior to
obtaining extensive (and expensive) laboratory data of the type described above.

Evaluating Gas Cycling Potential
Defining the target of condensate recovery by gas cycling is important to economic
evaluation and field development strategy. The following definitions are useful for
defining the target of gas cycling:

1. RF
oD
is, as previously discussed (Eq. 7), the condensate recovery by pressure
depletion to some reservoir pressure e.g. the end of production p
end
, or at the
pressure of gas cycling p
cycle
.

2. RF
oM
is the recovery of condensate which would be expected at the end of
cycling due to gas-gas miscible displacement with 100% sweep efficiency

13
(E
s
=100%); it is assumed the gas cycling occurs at a constant pressure p
cycling

above or below the original dewpoint.

3. RF
oV
(=100 RF
oM
) is the recovery of condensate which would be expected at
the end of cycling due to vaporization of retrograde condensate with 100% sweep
efficiency (E
s
=100%); it is assumed the gas cycling occurs at a constant pressure
p
cycling
below the initial dewpoint. Note, RF
oV
= 100% for gas cycling above the
initial dewpoint.

4. RF
oDx
is the extra condensate recovery from pressure depletion of the reservoir
volume not swept by injection gas during cycling, (depletion from p
cycling
to p
end
).
Note, RF
oDx
= RF
oD
(p
cycling
) RF
oD
(p
end
).

5. RF
oult
is the ultimate condensate recovery due to (a) depletion prior to cycling, (b)
cycling, and (c) depletion after cycling. For p
cycling
> p
d
,

oDx S oM S oD oult
RF ) E 1 ( RF E RF RF + + = ............................................................ (12)
while for p
cycling
< p
d
,

( )
oDx S oV V oM S oD oult
RF ) E 1 ( RF E RF E RF RF + + + = ......................................... (13)

where E
S
defines the final areal-times-vertical sweep efficiency at the end of cycling,
and E
V
defines the final efficiency of vaporized retrograde condensate (for p
cycling

below the dewpoint).

What components of the ultimate recovery are strongly dependent on PVT
properties?

We already have defined the dependence of RF
oD
on PVT properties (Eq. 7), where
Z-factors and variation of C
7+
in the CVD produced gas phase determine RF
oD
. It is
also easy to show that RF
oM
is given exclusively by data in the CCE and CVD tests,
as is RF
oV
(= 100 RF
oM
).

The only other PVT-dependent parameter is E
V
, which (as mentioned earlier) is
determined by (1) the C
7+
molar distribution of retrograde condensate, and (2) the K-
values of C
7+
components as a function of pressure and overall composition.
However, E
V
is only important for cycling below the dewpoint, and often the
contribution of vaporization to overall condensate recovery is relatively small.

Fig. 3 shows recovery of condensate versus pressure for a high-pressure offshore
gas condensate field. Initial pressure is 900 bara, and dewpoint is 400 bara. The
calculations are based only on CCE and CVD data, as shown in Table 2. The lower
curve gives RF
oD
(p), the recovery due to pressure depletion. The upper curve gives
RF
oM
(p), the recovery due to gas-gas miscible displacement with 100% sweep
efficiency but with no vaporization of retrograde condensate.

In this high-pressure reservoir, the additional recovery due to gas cycling will not be
realized until late into the life of the field, as pressure approaches the dewpoint and

14
recovery is about 30%. In terms of net present value, and depending on when the
investment for compressors etc. are required, cycling does not appear attractive.

For the same reservoir but initially saturated at the dewpoint of 400 bara, additional
condensate recovery by gas cycling is more attractive, as shown in Fig. 4. Here the
rapid decline in producing liquid yields will have a pronounced affect on project
economy, while successful cycling (high sweep efficiency E
S
) can provide prolonged
initial liquid yields and higher ultimate recovery. For this case, net present value is
more positive to a cycling project.

In summary, most of the primary evaluation for potential of gas cycling can be
quantified by CVD and CCE data. Vaporization effects are often less significant than
commonly thought, an observation which follows from the calculation of recovery
factors RF
oD
and RF
oM
based on CVD and CCE data. When vaporization recovery
(RF
oV
) is important, special multi-contact vaporization tests should be conducted and
fit with the PVT model, where variation of C
7+
in the equilibrium gas is the most
important data.

Combined Condensing/Vaporizing Mechanism
Historically it has been assumed that any gas cycling project in a gas condensate
reservoir was miscible only by the vaporizing gas drive mechanism. Consequently,
the MMP has always been assumed equal to the dewpoint pressure. Cycling
projects where reservoir pressure drops below the dewpoint were considered
"inferior" because only partial vaporization of the retrograde condensate could be
expected. For most separator injection gases these traditional assumptions are valid.
However, Hier and Whitson
6
show that miscible displacement of gas condensates
(by the condensing/vaporizing mechanism) can be obtained at pressures far below
the dewpoint for continuous or slug injection of gas enriched with components C
2
-C
5
.

Whether a below-dewpoint miscible displacement can develop in a gas condensate
depends on (1) pressure, (2) composition of the injection gas, (3) composition of the
retrograde condensate ahead of the front, and (4) physical dispersion or fingering
(for slug injection). Although the same conditions also apply for enriched-gas
miscible displacement of an oil reservoir, conditions (3) and (4) are particularly
important for gas condensates.

The most likely candidate for enriched-gas miscible gas cycling below the dewpoint
would be rich or near-critical condensates where injection gas is not available in
sufficient quantities to maintain reservoir pressure.

REPRESENTATIVE SAMPLES

Before a field development starts, the primary goal of sampling is to obtain
"representative" samples of the fluids found in the reservoir at initial conditions. It
may be difficult to obtain a representative sample because of two-phase flow effects
near the wellbore. This occurs when a well is produced with a flowing bottomhole
pressures below the saturation pressure of the reservoir fluids. It is also commonly
thought that bad fluid samples result if gas coning or oil coning occurs during
sampling.

15

The most representative insitu samples are usually obtained when the reservoir fluid
is single phase at the point of sampling, be it bottomhole or at the surface. Even this
condition, however, may not ensure representative sampling. And, as shown by
Fevang and Whitson
9
, samples obtained during gas coning in an oil well can provide
accurate insitu representative samples if a proper laboratory procedure is followed.

Because reservoir fluid composition can vary areally, between fault blocks, and as a
function of depth, we are actually interested in obtaining a sample of reservoir fluid
that is representative of the volume being drained by the well during the test.
Unfortunately, the concept of a "representative" sample is usually a sample that
correctly reflects the composition of reservoir fluid at the depth or depths being
tested.

If we suspect or know that a sample is not "representative" (according to this
definition), then we tend to do nothing with the sample. Or we question the validity of
the PVT analysis done on the "unrepresentative" sample, and consequently dont
include the measured data when developing the PVT model.

We strongly recommend against using this definition of "representivity." First of all, it
is a definition that costs our industry in terms of wasted money and time, and lost
opportunity. An important point to keep in mind is that:

Any fluid sample that produces from a reservoir is automatically
representative of that reservoir. After all, the sample is produced from the
reservoir!

The final EOS fluid characterization of a field should match all (accurate) PVT
measurements of all (uncontaminated) samples produced from the reservoir,
independent of whether the samples are representative of insitu compositions.

Accuracy of PVT Data = Insitu Representivity of Sample

Accurate PVT measurements can be made on both representative and
unrepresentative samples. Inaccurate PVT measurements can also be made on
both types of samples; bad PVT data should be ignored.

Furthermore, an EOS fluid characterization is used to predict compositional changes
during depletion which represent a much greater variation than the compositional
differences shown by "representative" and "unrepresentative" samples.

Another misconception in "representative" fluid sampling of gas condensates is that it
is difficult to obtain insitu-representative samples in saturated gas condensate
reservoirs (with underlying oil). The exact opposite is true! Fevang and Whitson
9

have shown that if a gas condensate is initially saturated and in contact with an
underlying oil zone, then a near-perfect insitu-representative sample can be obtained
(at the gas-oil contact) independent of whether the reservoir gas and reservoir oil
samples collected are insitu-representative.


16
In summary, all uncontaminated samples collected from a reservoir are reservoir
representative and, accordingly, should be described accurately by the PVT model.
Insitu-representative samples may be difficult to obtain. But even when collected,
they may not represent more than a local volume of the reservoir, where significant
variations in fluid composition exist vertically and areally away from the point of
sampling.

EOS MODELING

To make EOS calculations, the minimum required input are (1) molar composition, and
(2) molecular weight and specific gravity of the heaviest fraction (usually C
7+
or C
10+
).
With this minimum information, an EOS can calculate practically any phase and
volumetric property of the mixture - e.g.,

Bubblepoint or dewpoint pressure at a specified temperature
Pressure-temperature phase envelope
Densities and compressibilities of oil and gas phases
Separator gas-oil ratio and surface gravities
Depletion PVT experiments
Multicontact gas injection experiments

Splitting the Plus Fraction
Usually, three to five C
7+
fractions (or 2 to 3 C
10+
fractions) should be used. The
Whitson et al.
10
splitting/characterization procedure is recommended for the Peng-
Robinson EOS. The Pedersen et al.
11
method is recommended for the Soave-Redlich-
Kwong EOS, where each plus fraction has equal mass fraction.

When true boiling point distillation data are available, these data should be used
directly, or to define parameters in the splitting model. TBP data can be used, for
example, to define the molar distribution parameters and in the gamma distribution
model, and constants in the specific gravity correlation.

Tuning the EOS Model
If measured PVT data are available, and they have been checked for accuracy
k
, the
EOS characterization can be modified to improve the predictions of measured data.
Manual adjustments of EOS parameters such as binary interaction parameters (BIPs)
and heavy component critical properties can be used, though this approach is time
consuming (and often frustrating). Nonlinear regression can be used to mathematically
minimize the difference between EOS predictions and measured PVT data. A critical
aspect of the tuning procedure is to properly weigh individual data (and data types)
based on the importance of individual data to reservoir and well performance.

k
Material balance methods are often useful for checking consistency of depletion experiments (CVD
and DLE). The material balance starts at the final stage of the experiment, with known amounts and
compositions. Removed gas is then added back from each depletion stage, arriving at the initial fluid.
Comparison with the initial composition gives a direct measure of consistency. Another approach is to
start the material balance with the initial fluid, back-calculating the oil phase properties and
compositions as depletion progresses; this approach is only useful for medium-rich to rich
condensates because small errors in V
ro
have a dramatic effect on the back-calculated oil properties.


17

Developing a Common EOS Model for Multiple Reservoir Fluids
An important requirement in the development of an EOS model is the need to have one
set of components to describe all reservoir samples in a given field. This is particularly
important if the reservoir fluids from different parts of the reservoir (layers or fault
blocks) mix in the reservoir. It also may be important if mixing only occurs at the
surface.

Our experience has shown that a single set of components and a single set of EOS
component properties can be used to describe a wide range of reservoir fluids, ranging
from leaner gas condensates to low-GOR oils fluids which may or may not be in fluid
communication initially. Whitson et al.
10
propose one method for developing a
common EOS model for multiple reservoir fluids. Another approach is to develop the
EOS model based on a single sample, and then generate the other reservoir fluids by
flash calculations (saturation pressure, two-phase split, or isothermal gradient).

Generating Black-Oil PVT Tables
Once the EOS characterization has been developed, a primary application of the EOS
is to generate black-oil PVT tables for reservoir simulation, material balance and flow
calculations (also pipeflow calculations). The most common application of black-oil
PVT is black-oil simulation.

The procedure proposed by Whitson and Torp
12
is recommended for generating black-
oil PVT tables. They suggest using a reference fluid to conduct a depletion test (e.g.
CVD), sending the equilibrium reservoir phases separately through a surface
separation to obtain (R
s
, B
o
,
o
) for the oil phase and (r
s
, B
gd
, and
g
) for the gas phase.
Surface densities are taken from the surface separation of the reference fluid.

The definition of black-oil PVT properties are:
ratio gas oil- solution =
V
V
=
r
factor volume formation gas dry =
V
V
=
B
ratio oil gas- solution =
V
V
=
R
factor volume formation oil =
V
V
=
B
gg
og
s
gg
g
gd
oo
go
s
o
o
OO

where subscripts are defined as:

o : reservoir oil phase at p and T
g : reservoir gas phase at p and T

oo : surface oil from reservior oil (solution oil)
go : surface gas from reservoir oil ("solution" gas)
og : stock-tank oil (condensate) from reservoir gas
gg : surface gas from reservoir gas

18

The important black-oil PVT properties for IFIP calculations in gas condensate
reservoirs are r
s
/B
gd
= the surface oil in place per reservoir gas volume
l
, and 1/B
gd
= the
surface gas in place per reservoir gas volume. The term r
s
/B
gd
is the quantity required
by geologists to convert reservoir gas pore volumes to surface oil a kind of oil FVF
(B
o
) for the reservoir gas phase. In fact, for compositionally-grading reservoirs with a
transition from gas to oil through an undersaturated (critical) state, the term r
s
/B
gd

should equal B
o
exactly at the undersaturated gas-oil contact, thereby ensuring
continuity and consistency.

Some special problems related to generating black-oil PVT tables with an EOS include:

1. How to extrapolate saturated PVT properties to pressures higher than the original
saturation pressure of the reference fluid.
2. Non-monotonic saturated oil properties B
o
and R
s
for gas condensate systems.
3. Consistency requirements for comparison of black-oil and EOS simulations.
4. Handling saturated reservoirs with a gas overlying an oil, where black-oil PVT
properties from the two fluid systems can be significantly different.
5. Modeling reservoirs with compositional gradients, and how to initialize these
reservoirs in black-oil simulators.

Extrapolating Saturated Tables
Extrapolating saturated black-oil PVT tables can be done in a number of ways,
depending on the reservoir process and why the extrapolation is needed. Extrapolation
is usually required for (a) gas injection studies, (b) reservoirs with compositional
gradients where the reference sample is undersaturated, or (c) ensuring numerical
stability for near-critical fluid systems where pressures may exceed the original
saturation pressure during iteration.

Methods for extrapolating black-oil PVT tables include: (a) mixing the incipient phase
from a saturation pressure of the reference fluid to increase the saturation pressure,
usually in a number of steps (reverse DLE), (b) using a compositional gradient
algorithm, or (c) adding an injection gas in increments and determining the PVT
properties of each incremental mixture, or (d) adding injection gas to a maximum
saturation pressure and then conducting a depletion test either stopping at the
original saturation pressure or continuing all the way to a low (minimum) pressure.

The most appropriate method for extrapolating saturated properties may not be
obvious. It will often depend on the reservoir process. Several extrapolation methods
may be tested in a realistic reservoir simulation model, where results are compared
with a fully-compositional EOS model. The extrapolation method which consistently
gives results most similar to the EOS model can be said to best represent the
reservoir process. Comparisons should include initial fluids in place, recoveries, and
GOR profiles of individual wells.

Non-monotonic Saturated Oil Properties
For medium to lean gas condensates we have often found that the saturated oil black-
oil PVT properties B
o
and R
s
are not monotonic first increasing just below the

l
r
s
/B
gd
is equivalent to the compositional equivalent C
7+
content in the reservoir gas, y
7+
.

19
dewpoint pressure, reaching a maximum, and then decreasing with pressure in a
normal fashion. The physical explanation for this behavior is that the first condensate
which appears is (for lean and medium-lean condensates) quite heavy, with a high
surface-oil density (e.g. low API gravity < 40). The low-gravity or heavy condensate
has, as expected
m
, a relatively low B
o
and R
s
.

As more condensate evolves from the reservoir gas, the total reservoir condensate
becomes lighter with higher (API) gravity. The change in the gravity has a stronger
influence than the decreasing pressure, so the net change in B
o
and R
s
is to increase
with decreasing pressure.

As pressure continues to decrease, the condensate gravity stabilizes and normal
pressure dependence of saturated oil properties results, with B
o
and R
s
both
decreasing at decreasing pressures.

One solution we have found useful in this situation is to generate a separate set of
black-oil PVT tables, starting with the incipient oil phase of the original reservoir gas at
dewpoint, and using a depletion (DLE or CVD) test with this oil. The oil phase PVT
table (B
o
, R
s
, and
o
) from depletion of the incipient oil can then be used together with
the gas phase PVT table (B
gd
, r
s
,
g
) from depletion of the original reservoir gas.

Consistency Between Black-Oil and EOS Models
Coats
8
addresses the need for compositional EOS models for gas condensate
reservoirs. He shows that gas cycling below the dewpoint is the only situation when
black-oil modeling may be inadequate. He suggests that an EOS model with at least 3
C
7+
fractions should be used to properly model vaporization recovery of retrograde
condensate. His results indicate that depletion, single-well modeling, and gas cycling
above the dewpoint are properly modeled with a black-oil model.

Fevang and Whitson
7
show that some modification of black-oil saturated oil viscosities
may be needed to ensure accurate single-well modeling of condensate blockage. They
show that the oil viscosity in the near-wellbore blockage region is the liquid which
condenses from the flowing wellstream, and not a cumulative CVD-type liquid. The
flowing blockage-area liquid can be significantly lighter than the CVD-type condensate,
and with correspondingly lower viscosity (1.5 to 5 times lower).

Handling Saturated Gas/Oil Systems
Black-oil PVT properties for a saturated reservoir gas/oil system (gas cap overlying oil)
may be difficult to generate using a consistent approach. Traditionally we generate a
complete set of PVT tables separately for the reservoir oil and reservoir gas, using a
depletion test for the reservoir oil (e.g. DLE) and a depletion test for the reservoir gas
(e.g. CVD). From the depletion test of each reservoir phase, the complete set of black-
oil PVT tables are consistent only at the initial saturation pressure. That is, the incipient
oil from the dewpoint of the reservoir gas is identical to the reservoir oil; and the
incipient gas from the bubblepoint of the reservoir oil is identical to the reservoir gas.

The saturated oil and gas phases which form from the two depletion tests are different
below the original saturation pressure. This leads to differences in PVT properties

m
Expected, for example, based on a Standing-type saturated correlations for B
o
and R
s
.

20
which are not handled consistently in a black-oil simulator. One solution is to use two
PVT regions, one for the cells originally in the gas cap, and another region for cells
originally in the oil. This solution is incorrect for cells which originally are defined as one
phase but become the other phase due to movement of the gas-oil contact. Still, this
may be the best solution in some reservoirs.

Initializing Reservoirs with Compositional Gradients
Two problems arise when trying to use black-oil simulators for reservoirs with
compositional gradients. The one problem is obtaining correct initial surface fluids in
place (compared with initialization using an EOS model). The second problem is
analogous to that discussed above for saturated gas/oil systems, where PVT
properties of the different reservoir fluids are not the same.

The best way to ensure accurate initialization of surface gas and surface oil in place
is to initialize using R
s
and r
s
versus depth, instead of using saturation pressure
versus depth. The more-common practice of initializing with saturation pressure
versus depth leads to problems of initial fluid in place because of the second
problem mentioned above i.e. though only one PVT table is used, the black-oil
PVT properties of the different reservoir fluids are not the same.

Our recommended method of using R
s
an r
s
versus depth for initialization may lead
to a small error in recoveries near the initial saturation pressures. However, this error
is usually insignificant and always less than errors introduced by wrong initial fluids in
place caused by initialization with saturation pressure versus depth.

Pseudoization (Grouping Components)
Some reservoir processes can not be adequately modeled with a black-oil PVT
formulation. Gas injection, near critical oil and gas condensate systems, and laboratory
simulations may require fully compositional EOS simulation. The mathematical
complexity of integrating an EOS in a reservoir simulator is many times that of using a
simple black-oil PVT formulation. The result is a simulator that runs much slower than
a black-oil simulator. It may be necessary to economize the number of components
used in compositional simulation by "pseudoization" (i.e. reducing the number of
components in an EOS characterization).

The number of components used in an EOS characterization depends both on
computational restraints, and on the desired level of accuracy from the EOS. Some
balance between these two requirements is needed to determine the final number of
components for solving a given problem.

An initial fluid characterization will typically contain from 13 to 20 components, and
sometimes more. For best results, a stepwise pseudoization procedure is
recommended, whereby several pseudoized characterizations are developed
sequentially (e.g. 15, 12, 10, 7, and 5 pseudocomponents). The goal with each
pseudoization is to maintain PVT predictions as close to the original full
characterization as possible. With this stepwise approach, it is readily determined how
few pseudocomponents are necessary to maintain a required similarity to the original

21
full characterization.
a
Reducing the number of components in a stepwise fashion has
three main advantages:

1. It is possible to establish when a further reduction in number of components results
in predicted properties that deviate unacceptably from the original N-component
characterization.

2. The procedure usually results in several alternative characterizations with a
common basis. One simulation might require more components than another (e.g.
radial single-well study versus full-field simulation). Because several
characterizations are available, and they are "related" through the original N-
component characterization, more consistency can be expected.

3. Experience has shown that better results are obtained in going from the N-
component characterization to (for example) a 7-component characterization in
several steps, than going from an N-component to a 7-component characterization
in a single pseudoization.

The recommended stepwise pseudoization procedure is given below:

1. Use regression to develop an EOS characterization so that all relevant and
accurate PVT data are adequately matched. (This is probably the most difficult part
of any fluid characterization).

2. Using this tuned EOS, simulate several PVT experiments. Save the results of these
calculations as data. The experiments should cover as large as possible the
pressure-, temperature-, and composition-space expected in the reservoir during its
development. If gas injection is being considered, multicontact gas injection
experiments should be included, perhaps several with different injection gas
compositions.

3. Reduce the number of components by 2 or 3 by grouping components, e.g., group
iso- and normal-alkanes of butanes and pentanes.

4. Fine tune (by regression) the newly-created pseudo-component EOS parameters.
Recommended parameters include multipliers to EOS constants A and B and
volume shift parameter s for each newly-created pseudocomponent separately; and
BIPs between methane and the C
7+
fractions (collectively).

5. In a subsequent step, regress viscosities for the EOS model with the newly-created
pseudocomponents.

6. Return to step 3, selecting 2 or 3 new components to group.

Fig. 5 summarizes an example pseudoization procedure.

a
The number of pseudocomponents will vary according to the application. Simulation of depletion
processes and water flooding will generally require only 4 or 5 pseudocomponents; immiscible gas injection
may require additional pseudocomponents, and developed miscible gas injection will probably require at
least 6 to 8 pseudocomponents.

22

CONCLUDING REMARKS

We have tried to summarize and detail the importance of various PVT properties on
the reservoir and well performance of gas condensate fields.

1. For calculation of initial gas and condensate in place the key PVT data are (a)
initial Z-factor and (b) initial C
7+
molar content. In terms of black-oil PVT
properties, the two equivalent PVT quantities are (a) B
gd
and (b) r
s
/B
gd
.

2. The constant composition and constant volume depletion tests provide the key
data for quantifying recovery of produced gas and condensate during depletion.
Above the dewpoint depletion recoveries of gas and condensate are equal and
are given by the variation of Z-factor with pressure.

3. For calculation of condensate recovery and varying yield (producing oil-gas ratio)
during depletion it is critical to obtain accurate measurement of C
7+
(r
s
) variation
in the produced gas from a constant volume depletion test.

4. For near-saturated gas condensate reservoirs producing by pressure depletion,
cumulative condensate produced is insensitive to whether the reservoir is
initialized with or without a compositional gradient (even though initial condensate
in place can be significantly different for the two initializations).

5. Oil viscosity should be measured and modeled accurately to properly model
condensate blockage and the resulting reduction in gas deliverability.

6. For richer gas condensates, the oil relative volume (from a constant composition
expansion test) has only a secondary effect on the modeling of condensate
blockage; for lean condensates, V
ro
has a small effect on blockage.

7. For gas cycling projects above the dewpoint, PVT properties have essentially no
effect on condensate recovery because the displacement will always be miscible.
Only the definition of initial condensate in place is important. Gas viscosity has
only a minor effect on gas cycling.

8. For gas cycling below the dewpoint, the key PVT properties are Z-factor variation
during depletion, C
7+
content in the reservoir gas during depletion, and C
7+

vaporized from the reservoir condensate into the injection (displacement) gas.

Nomenclature

B
o
= formation volume factor (FVF) of reservoir oil phase
B
gd
= dry gas FVF of reservoir gas phase
C
og
= conversion factor for gas-equivalent of surface oil
C
5+
= Pentanes-plus
C
6+
= Hexanes-plus

23
C
7+
= Heptanes-plus
C
10+
= Decanes-plus
E
S
= sweep (vertical and areal) efficiency
E
V
= vaporization efficiency of condensate recovery below the dewpoint
k
rg
= gas relative permeability
k
ro
= oil relative permeability
K
i
= equilibrium ratio (K-value) of component i
M
o
= surface oil molecular weight
M
7+
= C
7+
molecular weight
N = total number
n
d
= moles of gas at initial (dewpoint) pressure
n
pk
= incremental CVD moles of gas produced in stage k
p = pressure
p
b
= bubblepoint pressure
p
d
= dewpoint pressure
p
i
= initial pressure
p
s
= saturation pressure
p
sc
= standard pressure
q
o
= surface oil production rate
q
7+
= Heptanes-plus production rate
R = universal gas constant
r
s
= solution oil-gas ratio of reservoir gas phase
r
si
= solution oil-gas ratio at initial pressure
R
s
= solution gas-oil ratio of reservoir oil phase
RF
gD
= depletion recovery factor of gas
RF
oD
= depletion recovery factor of condensate
RF
oDx
= extra depletion recovery factor of condensate (after gas cycling)
RF
oM
= gas-gas miscible recovery factor of condensate
RF
oult
= ultimate recovery factor of condensate
T
sc
= standard temperature
V
d
= oil volume at dewpoint pressure
V
g
= gas volume
V
o
= oil volume
V
ro
= oil volume divided by oil volume at saturation pressure
V
t
= total (gas+oil) volume
x
i
= oil composition
y
i
= gas composition
y
7+
= C
7+
composition in the produced gas
z
7+
= mole fraction of C
7+
of produced wellstream
z
i
= produced wellstream or total mole fraction
Z = Z-factor
Z
i
= Initial Z factor

o
= surface oil density at standard conditions
7+
= surface density of C
7+
at standard conditions
g
= gas viscosity
o
= oil viscosity

REFERENCES

1. Lohrenz, J., Bray, B. G., and Clark, C. R. : Calculating Viscosities of Reservoir Fluids
From Their Compositions, J. Pet. Tech. (Oct. 1964) 1171-1176; Trans., AIME, 231.

24
2. Pedersen, K. S., and Fredenslund, Aa., : "An Improved Corresponding States Model for
the Prediction of Oil and Gas Viscosities and Thermal Conductivities," Chem. Eng. Sci.,
42, (1987), 182.
3. Whitson, C. H. and Belery, P. : "Compositional Gradients in Petroleum Reservoirs,"
paper SPE 28000 presented at the University of Tulsa/SPE Centennial Petroleum
Engineering Symposium held in Tulsa, OK August 29-31, 1994.
4. Lee, S. T., and Chaverra, M. : " Modeling and Interpretation of Condensate Banking for
the Near Critical Cupiagua Field," paper SPE 49265 presented for SPE Annual Technical
Conference and Exhibition held in New Orleans, Louisiana, 27-30 September, 1998.
5. Hier, Lars: " Miscibility Variation in Compositional Grading Petroleum Reservoirs,"
Thesis for dr.ing., Norwegian University of Science and Technology, NTNU, Nov.,1997.
6. Hier, Lars and Whitson, C. H. : Miscibility Variation in Compositional Grading
Reservoirs, paper SPE 49269 presented at the SPE Annual Technical Conference and
Exhibition held in New Orleans, Sep 27-30, 1998.
7. Fevang, ivind and Whitson, C. H. : Modeling Gas Condensate Well Deliverability,
SPE Reservoir Engineering, (Nov. 1996) 221.
8. Coats, K. H. : "Simulation of Gas Condensate Reservoir Performance," JPT (Oct. 1985)
1870.
9. Fevang, ivind and Whitson, C. H. : "Accurate Insitu Compositions in Petroleum
Reservoirs," paper SPE 28829 presented at the EUROPEC Petroleum Conference held
in London, Oct. 25-27, 1994.
10. Whitson, C. H., Anderson, T. F., and Soreide, I. : "C7+ Chracterization of Related
Equilibrium Fluids Using the Gamma Distribution," C7+ Fraction Characterization, L. G.
Chorn and G. A. Mansoori (ed.), Advances in Thermodynamics, Taylor & Francis, New
York (1989) 1.
11. Pedersen, K. S., Thomassen, P. , and Fredenslund, A. : "Characterization of Gas
Condensate Mixtures," C7+ Fraction Characterization, L. G. Chorn and G. A. Mansoori
(ed.), Advances in Thermodynamics, Taylor & Francis, New York (1989) 1.
12. Whitson, C. H. and Torp, S. B. : "Evaluating Constant Volume Depletion Data," JPT
(March, 1983) 610; Trans AIME, 275.

25
Table 1 Approximate Depletion Material Balance Calculations
Based on CVD and CCE Test Results.

Table 2 Approximate Depletion and Gas Cycling Calculations
Based on CVD and CCE Tests.

Pressure r
s
B
gd
Visg npw/nd Z-factor Vro RF
oD
RF
gD
RF
oM
RF
oV
bara Sm3/Sm3 m3/Sm3 cp % % % % % %
900 8.07E-04 3.69E-03 0.062265 0.0 1.928 0.000 0.0 0.00 100.0 0.0
800 8.07E-04 3.83E-03 0.057694 0.0 1.778 0.000 3.6 3.64 100.0 0.0
700 8.07E-04 4.01E-03 0.053105 0.0 1.628 0.000 7.9 7.89 100.0 0.0
600 8.07E-04 4.24E-03 0.04843 0.0 1.477 0.000 13.0 12.97 100.0 0.0
500 8.07E-04 4.56E-03 0.043548 0.0 1.325 0.000 19.2 19.17 100.0 0.0
450 8.07E-04 4.78E-03 0.040966 0.0 1.249 0.000 22.8 22.84 100.0 0.0
398 8.07E-04 5.07E-03 0.038125 0.0 1.171 0.000 27.2 27.21 100.0 0.0
375 7.62E-04 5.21E-03 0.035959 3.3 1.138 2.581 29.5 29.43 96.0 4.0
350 7.09E-04 5.39E-03 0.033592 7.1 1.105 5.222 32.1 32.14 91.5 8.5
325 6.45E-04 5.61E-03 0.031119 11.4 1.075 7.973 34.7 35.22 86.3 13.7
300 5.75E-04 5.89E-03 0.028633 16.1 1.049 10.401 37.3 38.73 80.8 19.2
275 5.06E-04 6.24E-03 0.026314 21.2 1.027 12.042 39.8 42.64 75.8 24.2
250 4.46E-04 6.70E-03 0.024242 26.6 1.008 12.933 42.1 46.89 71.7 28.3
225 3.95E-04 7.28E-03 0.022415 32.5 0.993 13.297 44.3 51.45 68.4 31.6
200 3.52E-04 8.04E-03 0.020816 38.6 0.980 13.317 46.4 56.27 65.9 34.1
175 3.16E-04 9.06E-03 0.019427 45.1 0.970 13.108 48.4 61.35 64.0 36.0
150 2.86E-04 1.04E-02 0.018235 51.9 0.963 12.738 50.2 66.62 62.6 37.4
125 2.64E-04 1.24E-02 0.017226 59.0 0.958 12.246 52.0 72.06 61.7 38.3
100 2.52E-04 1.55E-02 0.016382 66.2 0.957 11.674 53.6 77.62 61.3 38.7
Data from CVD and CCE Calculated

CVD Data Conversion to Surface Oil and Gas Recoveries
Based on Simplified Surface Flash (Surface Gas = C6- and Surface Oil = C7+)
PERA a/s, programmed by Curtis H. Whitson (19981126)
C7+ Mole Weight 161 kg/kmol
C7+ Density 830 kg/m3
Cog 122 Sm3/Sm3 (assumed constant)
(p/z)i/(p/z)d 0.9120 Approx.
Solution
Input (red) OGR
P Z n
p
/n dnp/nd y
7+
r
s RFg RFo
bara % % mol-% Sm3/Sm3 % %
Pi 532 1.2172 0.000 3.996 3.407E-04 0.00 0.00
Pd 430 1.0788 0.000 0.000 3.996 3.407E-04 8.80 8.80
408 2.710 2.710 3.339 2.827E-04 11.29 10.87
372 7.070 4.360 3.366 2.851E-04 15.29 14.22
320 14.720 7.650 2.875 2.423E-04 22.35 19.24
272 24.420 9.700 2.245 1.880E-04 31.36 24.21
221 36.060 11.640 1.742 1.451E-04 42.22 28.83
170 49.130 13.070 1.302 1.080E-04 54.48 32.72
121 62.630 13.500 1.055 8.727E-05 67.17 35.97
62 79.160 16.530 0.675 5.562E-05 82.76 38.52

26

Fig. 1 Approximate Material Balance Calculations
Based on CVD Test Results.

0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.1 1 10 100
V
ro
= V
o
/ V
t
of Flowing Mixture (=Produced Wellstream), %
(V
ro
not equal to Normalized Oil Saturation !)
k
r
g
Normal Range for Near-Wellbore Blockage Region
Lean GC Rich GC
EOS V
ro
is 20% too low
EOS V
ro
is 20% too high

Fig. 2 Effect of error in Vro on gas relative permeability
in near-wellbore condensate blockage zone.

0
100
200
300
400
500
600
0 10 20 30 40 50 60 70 80 90 100
Surface Gas and Surface Oil Recovery Factors
C
V
D

(
R
e
s
e
r
v
o
i
r
)

P
r
e
s
s
u
r
e
,

b
a
r
a
Surface Gas
Surface Oil

27
0
10
20
30
40
50
60
70
80
90
100
0 100 200 300 400 500 600 700 800 900 1000
Pressure, bara
R
e
c
o
v
e
r
y

o
f

C
o
n
d
e
n
s
a
t
e
,

%

I
O
I
P
Initial
Pressure
RF
oM
RF
oD
RF
oDx
RF
oV

Fig. 3 Condensate recoveries for pressure depletion and
gas cycling below the dewpoint in a high-pressure undersaturated reservoir.

0
20
40
60
80
100
0 50 100 150 200 250 300 350 400 450
Pressure, bara
R
e
c
o
v
e
r
y

o
f

C
o
n
d
e
n
s
a
t
e
,

%

I
O
I
P
Initial
Pressure
=
Dewpoint
RF
oM
RF
oV
RF
oD
RF
oDx

Fig. 4 Condensate recoveries for pressure depletion and
gas cycling below the dewpoint in a saturated reservoir.

28

Fig. 5 Example pseudoization procedure reducing an original EOS
characterization with 22 components to multiple pseudoized characterizations.

Component EOS22 EOS19 EOS12 EOS10 EOS9 EOS6 EOS4 EOS3
N2 N2 C1N2 C1N2 C1N2 C1N2 C1N2
CO2 CO2 CO2 CO2 CO2
C1 C1 CO2C2 C02C2
C2 C2 C2 C2 C2 C1N2CO2C2-C6 C1N2CO2C2-C6
C3 C3 C3 C3 C3 C3
IC4 IC4
IC4NC4 IC4NC4 IC4NC4 IC4NC4
NC4 NC4
C3-C6
IC5 IC5
IC5NC5 IC5NC5 IC5NC5 IC5NC5
NC5 NC5
C6 C6 C6 C6 C6 C6
C7 C7 C7
C7C8
C8 C8 C8
C9 C9 C9 C7C8C9F1F2 C7C8C9F1F2 C7C8C9F1F2 C7C8C9F1F2
C10+ F1 F1 C9F1F2
F2 F2
C7C8C9F1-F8
F3 F3
F4 F4 F3-F5
F5 F5
F3-F8 F3-F8 F3-F8 F3-F8
F6 F6
F7 F7 F6-F8
F8 F8
F9 F9 F9 F9 F9 F9 F9 F9
Ecuaciones de Estado
(EOS)
De qu se trata?
Representar el comportamiento volumtrico
de un fluido en funcin de la presin y la
temperatura (PVT) a partir de la
composicin del mismo a travs del
modelado mediante ecuaciones de estado.
Historia de las EOS (1)
1662: Boyles Law - PV = constant at a fixed T
1787: Charles Law - V is proportional to T at constant P
1801: Daltons Law - P = sum of the partial pressure
1802: Cagniard de la Tour - Discovery of critical state
1834: Clapeyron - Combined Boyles Law and Charles Law into
PV=RT
1873: van der Waals - First EOS and idea of corresponding
states
1880: Amagats Law - Volume of mixtures
of gases = sum of pure components Volumes
1901: Lewis Fugacity
1940: Benedict-Webb-Rubin - Eight constant EOS
Historia de las EOS (2)
1949: Redlich-Kwong - Two parameter
cubic EOS
1972: Soave - modification of Redlich-
Kwong Temperature dependent
attraction parameter
1976: Peng-Robinson - Two-parameter EOS
1978: Peng-Robinson - improved the term
in the model for heavier components
1980: Schmidt-Wenzel - Two-parameter
EOS does a better job in predicting liquid
densities
1982: Peneloux et al. - A consistent volume
correction for cubic EOS
Ejemplo: Gases (1)
Gases Ideales:
PV = nRT
- Desprecio de las fuerzas intermoleculares
(atraccin/repulsin)
- Volumen de las moleculas de gas despreciale relativo al
volumen ocupado por el gas.
- Todas las colisiones son perfectamente elsticas.
Gases Reales:
PV = ZnRT
- Todas las no idealidades se corrigen mediante un solo
parmetro: Z
- Z = f(P,V,T,composicin), funcin de las propiedades de los
componentes
Ejemplo: Gases (2)
Ley de los Estados Correspondientes:
Dos sustancias deben tener propiedades intensivas similares a
condiciones correspondientes respecto de su Temperatura
Crtica (Tc) y Presin Crtica (Pc).
Z = f[(T/Tc), (P/Pc), (V/Vc)] = f(Tr, Pr, Vr)
Mezclas:
Definir la regla de mezclado:
Propiedad Pseudo Crtica = Promedio Molar de las Propiedades
Crticas de los Componentes
Determinar las Propiedades Pseudo Reducidas con las
Propiedades Pseudo Crticas.
Utilizacin
Interpolar Datos
Extrapolar Datos por fuera del modelo/laboratorio.
Consistencia/Chequeo de Calidad de informacin obtenida
de Laboratorio o Procesos.
Hacer un PVT virtual de laboratorio
- En fluidos tales como Gas y Condensando / Petrleos
Voltiles, la composicin del fluido vara segn el camino
termodinmico recorrido: Modelado Composicional.
- Esto es difcil de llevar a la prctica en laboratorio (se
necesitan celdas de volumen considerable y variacin de
temperatura durante el ensayo, la cual tarda en estabilizar
para todo el sistema).
Qu datos necesito?
Temperatura de Reservorio.
Composicin del Fluido (o de las Fases)

Propiedades crticas de los componentes (Pc
i
, Tc
i
,
i
):
Para componentes puros y conocidos, las mismas se

encuentran tabuladas. Pero
En todos los petrleos existen HC de alto PM, que no
poseen las propiedades crticas definidas.
Estos se agrupan en un (o varios) pseudo-componente,
generalmente C
7+
, el cual hay que definir Pc, Tc y ,
matcheando el PVT con la salida del simulador. Ms
adelante ampliamos sobre este tema.
Coeficientes de Interaccin Binarios k
ij
.
EOS Cbicas
Van der Waals (1873)
Redlich-Kwong (1948)
Peng-Robinson (1976/1978)
Se denominan cbicas en trmino del
Volumen o el factor Z:
A.V
3
+B.V
2
+C.V+D=0
EOS: Van der Waals (1873)
P=RT/(V-b) - a/V
2
- b is the correction for finite volume of gas molecules
(sometimes referred to as the excluded volume or repulsion
parameter)
- a is the correction for molecular attraction
- a and b are functions of the subject components properties
(Pc, Tc, Vc)
EOS: Peng-Robinson (1976)
Peng-Robinson:
a=0. 457235
( RTc)
2
Pc
P=
RT
V b
a
V (V+b) +b(V b)
a: parmetro que tiene en
cuenta las fuerzas de atraccin
entre las molculas.
b: conocido como co-volumen,
parmetro que tiene en cuenta el
volumen de las molculas.
b=0. 077796
RTc
Pc
0 . 5
=1+m( 1Tr
0 . 5
)
m=0. 37464+1. 542260.26992
2
: factor accntrico, tiene en
cuenta la desviacin de la forma
de la molcula con respecto a
una esfera. Segn Edmister:
=
3
7
log ( Pc/ 14. 7)
Tc/ Tb1
1
EOS en Mezclas
Las EOS fueron desarrolladas para

componentes puros, y luego extendidas a
mezclas. Para ello, se deben definir ciertas
reglas de mezcla. Las ms comunes son:
a
ij
=
a
i
a
j
(
1k
ij
)
b
ij
=
b
i
+b
j
2
k
ij
: parmetro de interaccin binaria
entre el componente i y el
componente j.
Se obtiene de mediciones de
laboratorio en equilibrio gas-lquido
de mezclas binarias.
Usualmente se los iguala todos a 0,
aunque existen varias correlaciones
en la literatura para calcularlos:
Firoozabadi, Nishiumi, etc.
Sistema de Ecuaciones
Falta determinar la variacin de los x
i
e y
i
con
respecto a P y T
Esto se resuelve gracias al concepto de fugacidad

(potencial qumico, energa libre de Gibbs): en un
estado de equilibrio lquido-valor (VLE), las
fugacidades son iguales en ambas fases de un
componente i:
f
i
x
= f
i
y
, i = 1, 2, , n
ln
(
f
i
k
Pk
i
)
=
V
k
{
1
V
k
RT (
P
N
i
k
)
T,P,N
j
}
dV
k
V
k
ln
(
PV
k
RT
)
Caracterizacin C
7+
Se deben definir las propiedades crticas del

pseudo-componente que representa a la fraccin
pesada.
Esto puede ser realizado mediante:
Matchear las curvas PVT obtenidas en laboratorio

variando Pc y Tc de C7+.
Correlaciones:
Riazi-Daubert
Katz-Firoozabadi
Etc
Conviene estimar las propiedades crticas con

una correlacin y luego matchear las curvas. Se
recomienda matchear la curva PV primero (para
ajustar el Pb) y luego comparar con las de Bo y
Rs.
Copyright 2007, Society of Petroleum Engineers

This paper was prepared for presentation at the 2007 SPE Latin American and Caribbean Petroleum Engineering Conference held in Buenos Aires, Argentina, 1518 April 2007.

This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by the author(s). Contents of the paper, as presented, have not
been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material, as presented, does not necessarily reflect any position of the Society of Petroleum
Engineers, its officers, or members. Papers presented at SPE meetings are subject to publication review by Editorial Committees of the Society of Petroleum Engineers. Electronic reproduction,
distribution, or storage of any part of this paper for commercial purposes without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an
abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, Texas 75083-3836 U.S.A., fax 01-972-952-9435.

Summary
To predict the phase and volumetric behavior of hydrocarbon mixtures by using an Equation of State; e.g. the Peng and Robinson
Equation of State PREOS, the critical properties in terms of the critical pressure p
c
and critical temperature T
c
as well as the
acentric factor must be given for each component present in the mixture including the plus-fraction. For pure compounds, the
required properties are well-defined, but nearly all naturally occurring gas and crude oil fluids contain some heavy fractions that are
not well defined and are not mixtures of discretely identified components. These heavy fractions often are lumped and called the
plus-fraction (e.g. C
7+
fraction). Adequately characterizing these undefined plus fractions in terms of their critical properties and
acentric factors has long been a problem. Changing the characterization of the plus fraction can have a significant effect on the
volumetric and phase behavior of a mixture predicted by the PREOS. The inaccuracy of any the cubic equation of state results from
the following two apparent limitations:
1. improper procedure of determining coefficients a, b, and for the plus fraction
2. Equations of state treatment of hydrocarbon components with critical temperatures less than the system temperature
(i.e. methane and nitrogen).

Numerous authors have suggested that the EOS is generally not predictive and extensive splitting of the C
7+
fraction is often required
when matching laboratory data.

This paper presents a practical approach for calculating the coefficients a, b, and of the plus-fraction from its readily available
measured physical properties in terms of molecular weight M and specific gravity with the objective of improving the
predictive capability of equation of state. The predictive capability of the relationship is displayed by matching a set of laboratory data
on several crude oil and gas-condensate systems. In addition; the performance of the proposed method was also compared with
predictive PVT results as generated by using PVTSim
TM
software of Calsep. Additional comparisons are made by comparing the
proposed modified PR EOS results with those of Coats and Smart
4
regression methodology with PR EOS.

This study concludes that when the coefficients of the plus-fraction, i.e. a, b, and , are determined based on the proposed
methodology; splitting of the C
7+
into a number of pseudo-components is essentially unnecessary.

Introduction
An equation of state (EOS) is an analytical expression relating the pressure to the volume and temperature. The expression is used to
describe the volumetric behavior, the vapor/liquid equilibria (VLE), and the thermal properties of pure substances and mixtures.
Numerous EOSs have been proposed since van der Waals
1
introduced his expression in 1873. These equations were generally
developed for pure fluids and then extended to mixtures through the use of mixing rules. The mixing rules are simply a means of
calculating mixture parameters equivalent to those of a pure substance. The PREOS
2
is perhaps the most popular and widely used
EOS. In terms of the molar volume V
m
, Peng and Robinson proposed the following two-constant cubic EOS:

( )
( )
( ) ( ) [ ] b V b b V V
T a
b V
RT
p
m m m m
+ +
= 1

SPE 107331
On Equation of State
Tarek Ahmed, Anadarko Petroleum Corp.
2 SPE 107331
van der Waals observed that for a pure component, the first and second isothermal derivatives of pressure with respect to volume are
equal to zero at the critical point of the substance. This observation can be expressed mathematically as

0 =
C
T
m
V
p
2
and

0
2
2
=
C
T
m
V
p
3

Peng and Robinson imposed the above derivative constraints on Equation 1 and solved the resulting two expressions for the
parameters a(T
c
) and b to give

( )
( )
c
c
a c
p
RT
T a
2
= 4
and

( )
c
c
b
p
RT
b = 5
where the dimensionless parameters
a
and
b
are 0.45724 and 0.07780, respectively. At temperatures other than the T
c
, Peng and
Robinson adopted Soaves
3
approach for evaluating a(T). The generalized expression for the temperature-dependent parameter is
given by
a(T) = a(T
c
)(T) 6
where
( )
2
1 1
+ =
c
T
T
m T 7
with
m = 0.3746 + 1.5423 0.2699
2
8

Introducing the compressibility factor 'Z' into Equation 1 gives

Z
3
+ (B 1)Z
2
+ (A 3B
2
2B)Z (AB B
2
B
3
) = 0 9
where

( )
( )
2
RT
p T a
A= 10
and

( ) RT
bp
B= 11

To use Equation 9 for mixtures with its coefficients as expressed by Equations 10 and 11, Peng and Robinson recommend the
following classic mixing rules:

( ) [ ] ( ) ( ) ( ) ( ) [ ] ( )
=
i
ij
j
j i cj ci j i mix
k T T T a T a x x T a ] 1 [ 12
and
( ) ( )
i i
i
mix
b x b

= 13

In the application of Equations 12 and 13 to a hydrocarbon mixture, a(T) and b are calculated for each component in the mixture with
Equations 4 through 8. Questionable assumptions are made in the application of these equations to the plus fraction and to
SPE 107331 3
hydrocarbon components with critical temperatures less than the system temperature. These assumptions (outlined below) provide the
reasoning for the proposed modification of the popular EOS.

Assumption 1 In the derivation of expressions for a(T) and b, as represented by Equations 4 and 5, the critical isotherm of a
component is assumed to have a slope of zero and an inflection point at the critical point. The assumption, described mathematically
by Equations 2 and 3, is valid only for a pure component. Because the plus fraction lumps millions of compounds that are making up
the fraction, it is unlikely that Equations 4 and 5 would provide an accurate representation of the attraction parameter a(T) and the co-
volume b.

Assumption 2 The coefficients of Equation 7 were developed by regressing vapor-pressure data from the normal boiling point to the
critical point for several pure components. Again, it is unlikely that this equation will suffice for the higher-molecular-weight plus
fractions.

Assumption 3 As pointed out previously, the theoretical
a
and
b
values in the PREOS arise from imposing the van der Waals
critical-point conditions, as expressed by Equations 2 and 3, on Equation 1. These values essentially reflect satisfaction of pure-
component density and vapor-pressure data below critical temperature. At reservoir conditions, methane and nitrogen in particular are
well above their critical points. Coats and Smart
4
pointed out that no theory or clear-cut guide exists to selection or alternation of the
for components well above their critical temperatures.

Wilson et al.
5
showed the distinct effect of the plus fractions characterization procedure on all the PVT relationships predicted by an
EOS. A number of studies
4, 6-10
reported comparisons of EOS and laboratory PVT results for a wide variety of reservoir fluids and
conditions; most of these studies emphasize the plus-fraction characterization as the key element in attaining agreement between EOS
and laboratory results.

Coats and Smart
4
presented numerous examples of matching the measured and calculated data for nine reservoir fluids of various
degrees of complexity. They observed that without regression or significant adjustment of EOS parameters, the PREOS will not
adequately predict observed fluid PVT behavior. Coats and Smart indicated that the adjustment of five parameters in the PREOS is
frequently necessary and sufficient for good data match. These parameters are:

a
and
b
of methane,

a
and
b
of the plus fraction, and
binary interaction coefficient between methane and the C
7+
plus fraction k
C1-C7+

Whitson
9
observed that the method of adjusting the EOS constants
a
and
b
for the plus fraction is essentially the same as altering
the critical properties of the heavy fraction.

Several authors
9, 11-15
showed that the ability of the EOS to predict the phase behavior of complex hydrocarbon mixtures can be
substantially improved by splitting or breaking down the plus fraction into a manageable number of pseudo-components for EOS
calculations.

Description of the Proposed Modification of the PREOS
Because the inadequacy of the predictive capability of the PREOS lies with the three assumptions outlined above, an approach was
devised in this study to remove these assumptions. The approach is based on the fact that the acentric factor and critical properties of
the C
7+
are not well defined and never measure in the laboratory; however; there are measurements that routinely performed and
readily available on the plus fraction that include: the molecular-weight "M", the boiling point "T
b
", and specific gravity "" . Over the
years; many reliable correlations that have developed to cross correlate these parameters from numerous measurements. Therefore; in
the eliminating the first two assumptions, 49 hypothetical heavy petroleum fractions (i.e. plus fractions) with physical properties
(density, molecular weight, and boiling point) governed by the applicability range of Riazi and Daubert
16
equation were generated.
Riazi and Daubert developed a simple two-parameter equation for predicting the physical properties of pure compounds and undefined
petroleum fractions. The proposed correlation; as shown by Equation 14, is applicable in the molecular-weight rage of 70 to 300 and
normal boiling point range of 80 to 650F. The expression correlates the molecular-weight "M" with the boiling point "T
b
" and
specific gravity "" of the substance; it takes the form:

M = a(T
b
)
b

c
exp(dT
b
+ e + T
b
) 14
where

a = 581.960 b = 0.97476
c = 5.43076 x 10
-4
e = 9.53384
= 1.11056 x 10
-3

4 SPE 107331
The specific steps of the proposed modification are outlined below.

Step 1 Each hypothetical heavy fraction with a specified molecular weight, boiling point, and density (specific gravity) is
subjected to 10 temperature and 10
pressure values in the range of 60 to 300F and 14.7 to 7,000 psia. The specified density is then adjusted to account
for the temperature and pressure increases. A total of 100 density values are generated for each hypothetical heavy
fraction.

Step 2 Equation 9 is rearranged and expressed in terms of the density to give

[a(T
c
)(T)bRTb
2
-pb
3
]
3
M[a(T
c
)(T)3pb
2
2RTb]
2

M
2
(pb-RT)pM
3
= 0 15

Step 3 Equation 15 is incorporated into a nonlinear regression model that uses a(T
c
), b, and (T) as regression variables.
For each hypothetical heavy fraction under consideration, Equation 15 is solved for the density by optimizing the
regression variables to match the fraction generated density data.

Step 4 The optimized regression variables [i.e. a(T
c
), b, and (T)] are correlated with M, or T by the following
relationships. For a(T
c
) or b of the plus fraction,

( ) ( ) ( )

7
6
5
4 4
3
0
or
C
C
D
C
D C b T a
i
i
i
i
i
i c
+
+ +
=

=
=
16
with

+
=
c
M
D

Table 1 gives values of C
0
through C
7
(for a(T
c
) and b) of the above expression. For (T), Peng and Robinson (T) function as
expressed by Equation 7 is modified according to

( )
2
5 . 0
520
1 1
+ =
T
m T 17
with

( )

7 2
6 5
4 2
3 2
1 0
C
C C
M
C
M C M C
D C C
D
m + + + + +
+
= 18

Table 1 includes the values C
0
through C
7
for Equation 18.

Table 1 Coefficients of Eqs. 16 and 18

Coefficient a(T
o
) b m
C
0 -2.433525 x 10
7
-6.8453198 -36.91776
C
1 8.3201587 x 10
3
1.730243 x 10
-2
-5.2393763 x 10
-2
C
2 -0.18444102 x 10
2
-6.2055064 x 10
-5
1.7316235 x 10
-2

C
3 3.6003101 x 10
-2
9.0910383 x 10
-9
-1.3743308 x 10
-5
C
4 3.4992796 x 10
7
13.378898 12.718844
C
5 2.838756 x 10
7
7.9492922 10.246122
C
6 -1.1325365 x 10
7
-3.1779077 -7.6697942
C
7 6.418828 x 10
8
1.7190311 -2.6078099

SPE 107331 5
In the elimination of the third assumption, the Peng and Robinson parameters (i.e. a(T
c
), b, and m) for methane and nitrogen are
altered. For this approach, 100 z-factor values for each component were obtained from appropriate gas-compressibility-factor charts.
a(T
c
), b and m (to be used in Equation 17) for methane and nitrogen were optimized by incorporating a regression model in solving
Equation 9 and matching the z-factor data for each fraction. The optimized values are:

For Nitrogen: a(T
c
) = 4,569.3589, b = 0.46825820, and m = -0.97962859
For Methane: a(T
c
) = 7,709.7080, b = 0.46749727, and m = -0.54976500

Several computational schemes
7, 10, 17
for generating binary interaction coefficients were tested for the purpose of providing the
modified PREOS with a systematic and consistent procedure for determining the k
ij
. Petersens
17
computational technique was
adopted and appropriately modified to provide the proper EOS coefficients. The technique is described in the following steps.

Step 1 Set:
k
CO2-N2
= 0.12,
k
CO2-hydrocarbons
= 0.10, and
k
N2-hydrocarbons
= 0.10

Step 2 Estimate the binary interaction coefficient between methane and the heptanes-plus fraction, k
C1-C7+
. To provide this
estimate, the coefficient under consideration was adjusted to minimize the error in calculating the saturation
pressures of 12 hydrocarbon mixtures. Results of the study indicate the strong dependency of the calculated
optimum values of k
C1-C+
on system temperatures. The following linear relationship provides an appropriate
estimate of the parameter:

k
C1-C7+
= 0.00189(T-460) 0.297659 19

where:
T = system temperature,
o
R

Step 3 Calculate the binary interaction coefficients between components heavier than methane (i.e. C
2
, C
3
.etc.) and the
plus fraction according to the following expression
17
:

k
Cn-C+
= 0.8 k
C(n-1) -C7+
20

where:
n = number of carbon atoms

Step 4 Determine the remaining k
ij
from
17

( ) ( ) [ ]
( ) ( ) [ ]
5 5
7
5 5
7
i C
i j
C i ij
M M
M M
k k
=
+
+
21

where:
M
i
= molecular weight of component i
M
C7+
= molecular weight of the heptanes-plus fraction

Application of the Modified EOS and Discussion of Results
To validate the proposed methodology of treating the plus fraction, the modified EOS and PVTSim software were used to simulate a
variety of published laboratory PVT tests and compare their predicted results with actual data. The verification procedure is outlined
below:
1. In applying the proposed modification, the measured physical properties of the C
7+
in terms of molecular weight and specific
gravity were maintained as reported and used to calculate the parameters a(T
c
), b, and (T) of the plus-fraction. No splitting
of the plus fraction is required in performing the PVT analysis when applying the proposed modification.
2. For all various hydrocarbon mixtures used in the study; PVTSim was allowed to split the C
7+
fraction into the following 11
pseudo-components: C
7
, C
8
, C
9
, C
10
-C
12
, C
13
-C
14
, C
15
-C
16
, C
17
-C
19
, C
20
-C
22
, C
23
-C
25
, C
26
-C
30

3. The modified PR EOS model and PVTSim were both tuned to only match the saturation pressure. The PVTSim regressed on
the saturation pressure by adjusting the molecular weight of the plus fraction; while the modified PR model used the binary
interaction coefficient between C
1
and C
7+
as the regression variable to match saturation pressure

6 SPE 107331
For convenience and brevity of presenting and discussing results of the study, the following terms are used in the paper:
CCE = Constant composition expansion
DE = Differential expansion
CVE = Constant composition expansion
C-S = Coats-Smart tuned model
Mod. EOS = Modified Peng-Robinson equation of state
PVTSim = Results as predicted by PVTSim software
Exp. = Experimental data
V/V
s
= relative volume from CCE test; i.e. total volume of hydrocarbon system at any given pressure and
temperature divided by the volume at saturation pressure

Density Predictions To test the modified PREOS for its ability to predict the density of complex hydrocarbon mixtures under a
wide range of pressures and temperatures, the equation was applied to predict densities of the 15 hydrocarbon mixtures used by
Standing and Katz
18
to develop their popular correlation and the densities of 11 crude oil systems reported by Coats and Smart
4

Table 2 and Figure 1 summarize results of the model and compare the predicted densities with those calculated from the Standing-
Katz
18
(S-K) and Alani-Kennedy
19
(A-K) density correlations. In terms of the overall average absolute deviation, the modified EOS
predicted the density of the 26 mixtures with the lowest deviation of 5.58%, which compares favorably with the two density
correlations.
Table 2 Comparison of Predicted Oil Densities with Experimental Data
Density (g/cm
3
)
p
(psia)
T
(F)

Exp.

S-K

A-K
Mod.
EOS
S-K Data
A-1 3,185 120 0.696 0.729 0.720 0.718
A-2 5,270 120 0.745 0.759 0.736 0.745
A-3 8,220 120 0.814 0.817 0.773 0.802
A-4 1,600 120 0.702 0.712 0.701 0.718
B-1 2,915 250 0.697 0.702 0.688 0.663
C-1 2,880 120 0.652 0.655 0.642 0.663
C-2 1,010 120 0.716 0.724 0.712 0.732
C-3 5,330 120 0.712 0.685 0.672 0.712
A-K Data
D-1 4,330 120 0.731 0.729 0.714 0.726
E-1 4,195 120 0.753 0.744 0.728 0.748
F-1 3,185 250 0.654 0.670 0.658 0.638
F-2 4,315 250 0.657 0.664 0.656 0.628
F-3 5,330 250 0.677 0.684 0.672 0.648
G-2 3,485 35 0.679 0.675 0.667 0.701
G-3 4,970 35 0.766 0.764 0.748 0.764

C-S Data
.Oil 1 2,520 180 0.768 0.784 0.762 0.764
Oil 2 4,460 176 0.530 0.509 0.537 0.544
Oil 3 2,2115 140 0.736 0.807 0.809 0.752
Oil 3 2,362 160 0.722 0.804 0.796 0.726
Oil 3 2,597 180 0.708 0.795 0.785 0.701
Oil 3 2,792 200 0.695 0.788 0.772 0.683
Oil 4 2,547 250 0.646 0.647 0.651 0.615
Oil 4 2,283 180 0.679 0.679 0.682 0.672
Oil 4 1,958 110 0.711 0.709 0.712 0.719
Oil 6 2,746 234 0.609 0.620 0.623 0.599
Oil7 1,694 131 0.713 0.717 0.735 0.722

Average absolute
error, %

6.58

6.69

5.58
SPE 107331 7

Figure 1. Comparison Of Predicted Densities
0.5
0.55
0.6
0.65
0.7
0.75
0.8
0.85
0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85
Exp. oil density
O
i
l

d
e
n
s
i
t
y
,

g
m
/
c
u

c
m
Exp. oil density
S-K
A-K
Mod. EOS
Linear (Exp. oil density)

Coats-Smart Hydrocarbon Systems Coats and Smart
4
reported a detailed experimental description of several hydrocarbon
systems. Nine of the hydrocarbon systems, with the compositions given in Table 3, were used in the study. Coats and Smarts C-S
used a regression-based PVT program to match the laboratory by tuning parameters of equation of state. Most of their calculations
were performed by splitting the heptanes-plus into four fractions.

Table 3 Coats and Smart Compositional Data
Gas 2
*
Gas 2
**
Gas 5 Oil 1 Oil 2 Oil 3 Oil 4 Oil 6 Oil 7
CO
2
0.00690 0.00610 0.02170 0.00440 0.00900 0.60310 0.02350 0.01030 0.0008
N
2
0.00420 0.00340 0.00450 0.00300 0.00930 0.00110 0.00550 0.0164
H
2
S 0.00040 0.00040
C
1
0.58320 0.57490 0.70640 0.35050 0.53470 0.07050 0.35210 0.36470 0.2840
C
2
0.13550

0.13450 0.10760 0.04640 0.11460 0.01570 0.06720 0.09330 0.0716
C
3
0.07610 0.07520 0.04940 0.02480 0.08790 0.03060 0.06240 0.08850 0.1048
C
4
0.04030 0.04150 0.03020 0.01660 0.04560 0.03310 0.05070 0.06000 0.0840
C
5
0.02410 0.02330 0.01350 0.01600 0.02090 0.02680 0.05230 0.03780 0.0382
C
6
0.01900 0.01790 0.00900 0.05460 0.01510 0.02580 0.04100 0.03560 0.0405
C
7+
0.11450 0.12200 0.05880 0.48240 0.16920 0.18510 0.34970 0.30430 0.3597

M
+

193

193

153

225

173

189

213

200

252
+
0.8135 0.8115 0.8100 0.9000 0.8364 0.8275 0.8406 0.8366 0.8429
T, F 190 190 267 180 176 179 250 234 131
P, psig 4,450 4,415 4,842 2,520 4,460 2,597 2,547 2,746 1,694

8 SPE 107331
The proposed modification of the Peng-Robinson EOS was tested extensively using the above hydrocarbon system as well as a larger
number of unreported fluid studies. Results of the some of applications of the modified PREOS to a selected number of hydrocarbon
systems are presented below.

Near-Critical Gas Systems Gas 2 is a near-critical gas-condensate fluid at a reservoir temperature of 190F. Coats and Smart
stated that because of the possibility of a small error in gas measurement during well testing, two slightly different separator gas/liquid
ratios are used to obtain the two reservoir fluid compositions given in Table 3. The first sample (Gas 2
*
) exhibited a saturation
pressure and density of 4,465 psia and 28.85 lb
m
/ft
3
and was labeled a dewpoint gas. The second sample (Gas 2
**
) displayed a
bubblepoint of 4,430 psia and a density of 29.54 lb
m
/ft
3
and consequently was labeled a bubblepoint gas. In applying the modified
expression to simulate the volumetric behavior of the two systems, Equations 19 through 21 initially were used to determine the
binary interaction coefficients for each system. Table 4 lists these values for Gas 2
*
. The modified equation predicts a dewpoint
pressure of 3,890 psia (compared with 3,680 psia for the original PREOS) and a saturation density of 28.45 lb
m
/ft
3
. For the
bubblepoint gas, the equation predicts a saturation pressure of 3,824 psia (the original PREOS predicts 3,664 psia) and a saturation
density of 28.86 lbm/ft
3
.

Table 4 Binary Interaction Coefficients for Gas 2
*

Component Component j
i CO
2
N
2
C
1
C
2
C
3
i-C
4
n-C
4
i-C
5
n-C
5
C
6
C
7+
CO
2
0.000 0.012 0.100 0.100 0.100 0.100 0.100 0.100 0.100 0.100 0.100
N
2
0.000 0.100 0.100 0.100 0.100 0.100 0.100 0.100 0.100 0.100
C
1
0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.001 0.061
C
2
0.000 0.000 0.000 0.000 0.000 0.000 0.001 0.049
C
3
0.000 0.000 0.000 0.000 0.000 0.001 0.039
i-C
4
0.000 0.000 0.000 0.000 0.000 0.031
n-C
4
0.000 0.000 0.000 0.000 0.25
i-C
5
0.000 0.000 0.000 0.020
n-C
5
0.000 0.000 0.016
C
6
0.000 0.013
C
7+
0.000

Figure 2 compares experimental CCE data of the dew-point gas in terms of liquid relative volume with those as predicted by PVTSim
and the modified PR model. Both models overestimated the liquid volume just below the dew point pressure. However; PVTSim
predicted a bubblepoint system when the fluid was flashed below the saturation pressure. A constant volume depletion "CVE " test
was also performed on the dew-point gas as shown in Table 5. Results from simulating the test by applying the modified PR EOS are
compared with predicted values from experimental and PVTSim as documented graphically in terms of liquid dropout in Figure 3.
Table 5 and Figure 3 show that PVTSim failed to recognize the system as a dewpoint gas and predicted a 100% liquid at saturation
pressure. The match between the observed data and the modified PR model prediction in terms of liquid dropout is excellent with an
average absolute error of 1.05%. A more detailed documentation of results of the proposed EOS for simulating CCE tests is given in
Reference 20.

SPE 107331 9
Figure 2. CCE for dewpoint Gas 2*
0
10
20
30
40
50
60
70
80
90
100
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Pressure
R
e
l
t
.

V
o
l
,

%
Mod. EOS
Exp
PVTsim

Table 5 CVE at 190F for Dewpoint Gas
Pressure (psig)
Component 4,450.0 3,500.0 2,700.0 1,900.0 1,100.0 500.0 500
*
CO
2
0.00730 0.00767 0.00790 0.00824 0.00874 0.00922 0.00228
N
2
0.00000 0.00000 0.00000 0.00000 0.00000 0.0000 0.00000
C
1
0.58320 0.71198 0.72463 0.72791 0.71235 0.65973 0.07116
C
2
0.13550 0.13745 0.13945 0.14384 0.15417 0.17230 0.06429
C
3
0.07610 0.06804 0.06713 0.06782 0.07388 0.09265 0.07722
i-C
4
0.02015 0.01640 0.01568 0.01528 0.01625 0.02135 0.03199
n-C
4
0.02015 0.01538 0.01447 0.01387 0.01463 0.01961 0.03818
i-C
5
0.01205 0.00822 0.00738 0.00666 0.00658 0.00874 0.03124
n-C
5
0.01205 0.00783 0.00691 0.00611 0.00592 0.00781 0.03397
C
6
0.01900 0.01069 0.00885 0.00719 0.00631 0.00786 0.06607
C
7+
0.11450 0.01634 0.00762 0.00308 0.00117 0.00073 0.58360
Z-Factor
Mod. EOS 0.9605 0.8359 0.8016 0.8031 0.8461 0.9061
Experimental 0.9969 0.8402 0.7966 0.8140 0.8603 0.9108
PVTSim 0.976 0.812 0.784 0.798 0.949 0.980
G
p

Mod. EOS 0.00000 12.89 25.411 41.007 58.989 73.304
Experimental 0.00000 9.589 22.551 39.165 58.225 72.743
PVTSim 0.0000 9.79 22.46 38.950 58.010 73.170
Liquid Dropout (%)
Mod. EOS 0.0 56.53 51.0 46.20 41.35 37.25
Experimental 0.0 52.31 49.40 45.33 40.51 36.82
PVTSim 100 61.37 54.91 48.4 41.87 36.69
*Residual liquid composition
10 SPE 107331
Figure 3. LDO for Dewpoint Gas-2*
0
10
20
30
40
50
60
70
80
90
100
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Pressure, psi
%

L
D
O
Exp
Mod. EOS
PVTsim
Figure 4. CCE for Bubblepoint Gas2**
0
10
20
30
40
50
60
70
80
90
100
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Pressure
R
e
l
.

V
o
l
,

%
Exp
Mod. EOS
PVTsim

Gas-Condensate System 5 Gas 5 is a retrograde gas with a reported upper dewpoint pressure of 4,857 psia at 276F. The model
predicts an excellent value of 4,805 psia for the dewpoint pressure. In simulating the CVE and CCE tests for the system, the modified
equation and PVTSim show excellent predictive capabilities in reproducing the experimental data, as Table 6 illustrates. The
tabulated results compare the observed data with both models in terms of Z factor, liquid dropout, G
p
, and V/V
s
. The modified
expression gives average deviations of 1.6% for the Z-factor, 1.3% for G
p
, and 1.9% for liquid dropout. The near-exact match of the
modified PR EOS with reported liquid dropout data is documented graphically in Figure 5.

SPE 107331 11
Table 6 Expansion Data for Gas 5
Pressure (psig)
Component 4,642.0 3,900.0 3,000.0 2,100.0 1,200.0 700.0 700.0
*
CO
2
0.02170 0.02185 0.02207 0.02237 0.02276 0.02302 0.00611
N
2
0.00340 0.00348 0.00355 0.00359 0.00360 0.00356 0.00032
C
1
0.70640 0.72207 0.73481 0.74357 0.74575 0.73990 0.08684
C
2
0.10760 0.10893 0.11010 0.11124 0.11256 0.11352 0.02989
C
3
0.04940 0.04957 0.04970 0.04997 0.05083 0.05216 0.02532
I-C
4
0.01510 0.01502 0.01493 0.01491 0.01518 0.01583 0.01246
n-C
4
0.01510 0.01490 0.01471 0.01460 0.01490 0.01574 0.01636
I-C
5
0.00675 0.00658 0.00639 0.00625 0.00633 0.00681 0.01184
n-C
5
0.00675 0.00653 0.00629 0.00610 0.00615 0.00668 0.01428
C
6
0.00900 0.00853 0.00802 0.00753 0.00741 0.00814 0.03040
C
7+
0.05880 0.04254 0.02942 0.01986 0.01452 0.01467 0.76619

Z-Factor
Mod. EOS 1.0085 0.9324 0.8870 0.8742 0.8990 0.9290
Experimental 0.985 0.911 0.881 0.882 0.916 0.943
PVTSim 0.990 0.912 0.873 0.870 0.900 0.929

G
p

Mod. EOS 0.00000 13.373 29.34 48.29 69.02 80.561
Experimental 0.00000 12.812 29.341 49.110 69.907 81.220
PVTSim 0.00000 12.890 29.050 48.410 69.370 81.010

Liquid Dropout (%)
Mod. EOS 0.0 6.37 9.24 10.19 9.84 9.20
Experimental 0.0 6.10 9.10 10.40 9.80 9.10
PVTSim 0.0 6.28 10.39 11.16 10.06 9.03

Figure 5- CVE for Gas 5
0
2
4
6
8
10
12
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Pressure, psi
L
D
O
,

%
Mod EOS
Exp
PVTsim

12 SPE 107331
Crude Oil 1 This oil, with the composition given in Table 3, exhibits a saturation pressure and density of 2,535 psia and 47.96
lb
m
/ft
3
at 180F. The modified model predicted excellent values for the bubblepoint pressure and saturation density of 2,501 psia and
47.68 lb
m
/ft
3
. The modified EOS is tested for its predictive ability by a simulating CCE test. Table 7 compares the predicted and
adjusted liquid ratio V/V
s
with the observed data. The observed average absolute deviation is 0.156%. Both models are in excellent
agreement with the reported values

Table 7 CCE for Oil 1 at 180F
p V/V
s

(psig) PVTSim Mod. EOS Experimental
5,000 0.9740 0.9739 0.9782
4,000 0.9832 0.9831 0.9662
3,000 0.9940 0.9940 0.9951
2,900 0.9952 0.9952 0.9961
2,800 0.9964 0.9964 0.9971
2,700 0.9977 0.9977 0.9982
2,600 0.9990 0.9990 0.9992
2,520 1.0000 1.0000 1.0000

Crude Oil 2 This oil is characterized as volatile hydrocarbon system with a reported bubblepoint pressure,
b
, is 4,475 psia at 176F.
Table 8 compares the experimental differential values of R
sbd
, B
obd
, and oil density at the saturation pressure with those of the
PVTSim and modified PR predicted values. PVTSim produced excellent match values at the saturation pressure as compared with
the modified PR EOS. However; as shown in Table 9 and expressed graphically in Figures 6 and 7, the modified PR results compare
extremely well with the DE test data. A further evaluation of the proposed modified EOS is presented in Table 10 for predicting CVE
data on this volatile oil. Both models perform equally well in matching cumulative gas production G
P
, however; both models failed
reproduce acceptable match with the experimental gas deviation factor.

Table 8 Observed and Predicted Data, Crude Oil 2

Approach
p
b
(psia)
R
sbd

(scf/STB)
B
obd
(RB/STB)

(lb
m
/ft
3
)
Experimental 4,475 3,377 2.921 33.10
PVTSim 3,344 (-25.3) 3345 (-0.95) 2.967 (-1.5) 31.10 (-6.0)
Modified EOS 4,502 (0.60) 3,019 (-7.9) 2.675 (-8.4) 33.99 (2.6)
*Numbers in parentheses represent percent error of predictions

Table 9- DE at 176F for Oil 2
R
sd
(scf/STB
od
(RB/STB)
(psig) PVTSim Mod.
EOS
Experimental PVTSim Mod. EOS Experimental
4,460.0 3345.1 3109 3,377 2.967 2.6736 2.921
4,000.0 2723.9 2401 2,351 2.605 2.3039 2.343
3,492.0 2085.4 1909 1,814 2.249 2.0563 2.059
3,003.0 1632.2 1540 1,471 2.004 1.8737 1.886
2,514.0 1265.2 1254 1,205 1.809 1.7351 1.756
2,004.0 985.7 1004 970 1.662 1.6145 1.645
1,534.0 711.4 805 775 1.516 1.5191 1.555
1,001.0 473.1 600 573 1.386 1.4201 1.464
505.0 313.5 409 383 1.291 1.3243 1.372
209.0 285 271 245 1.039 1.2490 1.298
0.0 0 0 000 1.000 1.0601 1.057

SPE 107331 13
Figure 6- DE Rsd for Oil 2 at 176 F
0
500
1000
1500
2000
2500
3000
3500
4000
0.00 500.00 1,000.00 1,500.00 2,000.00 2,500.00 3,000.00 3,500.00 4,000.00 4,500.00 5,000.00
Pressure, psi
R
s
d
,

s
c
f
/
S
T
B
PVTsim
Mod. EOS
Exp Rsd

Figure 7. DE Bod for Oil 2 at 176 F
0
0.5
1
1.5
2
2.5
3
3.5
0.00 500.00 1,000.00 1,500.00 2,000.00 2,500.00 3,000.00 3,500.00 4,000.00 4,500.00 5,000.00
Pressure, psi
B
o
d
,

b
b
l
/
S
T
B
PVTsim
Mod. EOS
Exp

14 SPE 107331
Table 10- CVE at 176F for Oil 2
Pressure (psig)
Component 4,460.0 3,600.0 2,800.0 2,000.0 1,200.0 600.0 600.0
*
CO
2
0.00900 0.00984 0.01020 0.01074 0.01158 0.01256 0.00352
N
2
0.00300 0.00491 0.00479 0.00454 0.00405 0.00333 0.00020
C
1
0.53470 0.73509 0.74529 0.74678 0.73019 0.67668 0.08654
C
2
0.11460 0.11526 0.11767 0.12251 0.13386 0.15557 0.06976
C
3
0.08790 0.07183 0.07101 0.07187 0.07858 0.10006 0.10309
I-C
4
0.02280 0.01624 0.01548 0.01498 0.01568 0.02010 0.03705
n-C
4
0.02280 0.01474 0.01381 0.01311 0.01351 0.01739 0.04213
I-C
5
0.01045 0.00578 0.00514 0.00457 0.00434 0.00533 0.02363
n-C
5
0.01045 0.00541 0.00473 0.00411 0.00381 0.00461 0.02492
C
6
0.01510 0.00643 0.00526 0.00419 0.00348 0.00388 0.04092
C
7+
0.16920 0.01447 0.00662 0.00260 0.00092 0.00050 0.56822

Gas Deviation Factors
PVTSim 0.8900 0.8130 0.7820 0.7910 0.8370
Mod. EOS 0.8388 0.8051 0.8011 0.8361 0.8899
Experimental 0.798 0.783 0.788 0.843 0.913

G

, %
PVTSim 0.00 2.12 13.42 28.84 47.57 63.09
Mod. EOS 0.00000 8.954 19.827 33.422 49.638 63.172
Experimental 0.00000 7.535 17.932 32.371 49.908 63.967

Crude Oil 3 This CO
2
rich system contains 60 mol% CO
2
. The hydrocarbon system exhibits a saturation pressure and density of
2,612 psia and 44.17 lb
m
/ft
3
at 180F. Coats and Smart "C-S" presented the results of CCE tests on the system at four different
temperatures. Coats and Smart used 12 components system to match the saturation pressure over the reported range of temperatures
used in the laboratory. The authors predicted saturation pressure values that are approximately 500 psi lower the experimental values.
The tabulated values shown below compare the observed saturation pressures and densities at these four temperatures with those of
the modified EOS and Coats and Smart 12 components system. The modified EOS shows excellent match with the oil saturation at the
bubble point pressure

Observed and Predicted Data for Crude Oil 3
Temperature (F)
140 160 180 200
Exp.
p
b
2,144 2,392 2,627 2,821
ob
, lb
m
/ft
3
45.9 45.1 44.2 43.3
C-S
p
b
1,776 (-17) -2,000 (-16.5) 2,210 (-16) 2,403 (-14.9)
ob
, lb
m
/ft
3
43.1 (-6) 42.3 (-6.2) 41.5 (-6.1) 40.6 (-6.2)
Mod. EOS
p
b
2,082 (-2) 2,210 (-7.0) 2,304 (-12) 2,375 (-15.3)
ob
, lb
m
/ft
3
46.9 (2.2) 45.3 (0.5) 43.7 (-1.1) 41.8 (-3.4)
Numbers in Parentheses represent percent error of predictions

Crude Oil 4 The reported experimental data available on the system includes differential expansion results at 110 and 250F and
CCE data at 110, 180 and 250F. The system is slightly volatile with B
obd
= 1.671 and R
sbd
= 932 scf/STB at 250F. Table 11 lists the
predicted values for selected PVT properties as compared with the reported values at saturation pressures. Table 12 gives detailed
documentation of results of predicting the DE data at the specified temperatures. With adjustment and in terms of the average
absolute error, the model predicts R
sd
and B
od
data at 110F within 3.5 and 0.77%, respectively. At 250F, the equation gives results
within 6% for R
sd
and 3% for B
od
.

SPE 107331 15
Temperature (F)
110 180 250
Exp.
p
b
1,988 2,313 2,577
R
s
701 932
B
obd
1,341 1,671
ob
, lb
m
/ft
3
44.4 42.4 40.3
C-S
p
b
1,694 (-14.8) 2,033 (-12.1) 2,274 (-11.76)
R
sbd
611 (-12.8) 756 (-18.9)
B
obd
1.294 (-3.5) 1.517 (-9.2)
ob
, lb
m
/ft
3
40.0 (-9.9) 38.4 (9.4) 36.8 (-8.7)
Mod. EOS
p
b
1,695 (-14.7) 2,330 (0.7) 2,655 (3.0)
R
sbd
694 (-0.9) 949 (1.8)
B
obd
1,323 (-1.3) 1.769 (5.9)
ob
, lb
m
/ft
3
44.8 (1.1) 41.96 (-1.0) 38.35 (-4.8)
Numbers in Parentheses represent percent error of the predictions

Table 12 Expansion Data for Oil 4
DE at 110F
p R
sd
(scf/STB) B
od
(RB/STB)
(psig) PVTSim Mod. EOS Experimental PVTSim Mod. EOS Experimental
1,958.0 720.0 694 701 1.368 1.3243 1.341
1,753.0 714.7 629 633 1.366 1.2990 1.313
1,557.0 639.8 570 577 1.333 1.2762 1.291
1,354.0 565.1 509 510 1.301 1.2524 1.264
1,153.0 493.8 450 450 1.270 1.2293 1.240
949.0 423.8 391 389 1.239 1.2063 1.217
748.0 356.9 334 330 1.210 1.1839 1.193
548.0 291.5 278 270 1.181 1.1616 1.168
347.0 225.4 220 209 1.151 1.1382 1.144
157.0 155.4 158 143 1.117 1.1120 1.116
75.0 117.2 123 106 1.097 1.0963 1.097
0.0 0.0 000 000 1.015 1.0234 1.024

DE at 250F
2,547.0 1010.2 948 943 1.724 1.7678 1.671
2,360.0 931.8 863 865 1.680 1.7168 1.636
2,143.0 846.6 784 788 1.633 1.6712 1.595
1,893.0 755.1 690 704 1.583 1.6157 1.553
1,645.0 670.3 604 625 1.536 1.5644 1.512
1,393.0 589.4 522 548 1.491 1.5160 1.473
1,150.0 515.5 450 477 1.450 1.4720 1.436
895.0 441.0 376 407 1.408 1.4275 1.401
647.0 370.2 307 338 1.367 1.3846 1.365
400.0 297.5 237 265 1.323 1.3387 1.326
182.0 219.5 163 190 1.270 1.2858 1.275
87.0 170.0 120 146 1.232 1.2504 1.243
0.0 0.0 000 000 1.054 1.124 1.094

16 SPE 107331
Figure 8- DE Rsd for Oil 4 at 110 F
0
100
200
300
400
500
600
700
800
0.00 500.00 1,000.00 1,500.00 2,000.00 2,500.00
Pressure, psi
R
s
d
,

s
c
f
/
S
T
B
PVTsim
Mod. EOS
Exp.

Crude Oil 6 This oil exhibits a saturation pressure of 2,746 psia at 234F. Table 13 gives the reported R
s
and B
o
and
o
at the
bubblepoint pressure compared with the predicted results. Table 14 shows DE data calculated by the EOS before and after
adjustment. (For detailed results of CCE data, see Table 6 of Reference 20). In a prediction mode, the modified expression
reproduced the entire solution GOR and relative oil volume data with average absolute errors of 6.3 and 1.7%, respectively. (See
Table 7 of Reference 20 for CCE predicted data).


Approach
p
b
(psia)
R
sbd

(scf/STB)
B
obd
(RB/STB)

(lb
m
/ft
3
)
Experimental 2,746 1.230 1,866 38.0
Coats-Smart 2,398 (-12.7) 1,002 (-19) 1,659 (-11.1) 35.6 (-6.3)
Modified EOS 2,916 (6.2) 1,233 (0) 1,909 (2.3) 37.40 (-1.6)

Table 14 Expansion Data for Oil 6
DE at 234F
p R
sbd
(scf/STB) B
obd
(RB/STB)
2,746.0 1,321.0 1,232 1,230 1.917 1.9053 1.866
2,598.0 1,288.0 1,142 1,151 1.898 1.8514 1.821
2,400.0 1,178.5 1,048 1,059 1.836 1.7981 1.771
2,200.0 1,076.1 947 972 1.778 1.7385 1.725
1,897.0 934.1 806 849 1.698 1.6562 1.658
1,600.0 807.3 686 737 1.627 1.5864 1.599
1,300.0 688.9 576 631 1.560 1.5222 1.543
1,000.0 577.7 474 529 1.497 1.4624 1.488
700.0 470.1 377 428 1.434 1.4032 1.433
394.0 355.4 273 321 1.363 1.3374 1.371
195.0 265.7 194 231 1.302 1.2826 1.313
112.0 215.9 152 178 1.265 1.2510 1.274
0.0 0.0 000 000 1.052 1.1079 1.086

SPE 107331 17
Fi gure 9. DE Rsd for Oil-6
0.00
200.00
400.00
600.00
800.00
1,000.00
1,200.00
1,400.00
0.00 500.00 1,000.00 1,500.00 2,000.00 2,500.00 3,000.00
Pressure, psi
R
s
d
,

s
c
f
/
S
T
B
PVTSim
Mod. EOS
Exp

Figure 10. DE Bod for Oil-6
1
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
2
0.00 500.00 1,000.00 1,500.00 2,000.00 2,500.00 3,000.00
Pressure, psi
B
o
d
,

b
b
l
/
S
T
B
PVTSim
Mod. EOS
Exp

Crude Oil 7 This hydrocarbon system is the least volatile of the oil samples in the Coats and Smart study. The oil shows a
bubblepoint of 1,709 psia and a saturation density of 44.5 lb
m
/ft
3
at 131F. Table 15 shows the predicted PVT properties at saturation
pressure compared with the experimental and Coats-Smart model data. Table 16 compares measures DE data with those simulated
with both; the modified PREOS PVTSim model. The modified EOS reproduced the observed PVT with average deviations of 7.5%
and 1.7% for R
sd
and B
od
, respectively.

18 SPE 107331

Approach
p
b
(psia)
R
sbd

(scf/STB)
B
obd
(RB/STB)

(lb
m
/ft
3
)
Experimental 1,709 557 1,324 44.5
Coats-Smart 1,531 (-10.4) 542 (-2.7) 1,296 (-2.10 40.7 (-8.5)
Modified EOS 1,572 (-8.0) 624 (12.0) 1,346 (1.7) 45.0 (1.1)
Numbers in parentheses represent percent error of predictions

Table 16 Expansion data for Oil 7
DE at 131F
p R
sd
(scf/STB) B
od
(RB/STB)
1,694.0 641.2 624 557 1.369 1.3472 1.324
1,550.0 602.8 587 526 1.352 1.3321 1.311
1,400.0 562.9 550 493 1.334 1.3168 1.298
1,252.0 523.7 513 460 1.317 1.3013 1.285
1,100.0 483.3 475 423 1.299 1.2853 1.270
950.0 443.3 437 389 1.280 1.2694 1.256
798.0 402.3 398 349 1.262 1.2529 1.240
643.0 359.9 358 310 1.242 1.2357 1.224
500.0 319.8 320 273 1.223 1.2192 1.209
350.0 275.4 278 229 1.202 1.2005 1.188
200.0 224.2 230 179 1.175 1.1779 1.160
102.0 178.5 188 137 1.149 1.1562 1.136
0.0 000 000 000 1.017 1.0328 1.034

Red River Crude Oil System The crude oil system of the Red River Field (Montana) is slightly volatile with B
obd
= 1.706 and R
sbd

= 921 at 250F. The modified equation reproduced the observed B
od
and R
sd
with average deviations of 5.1 and 4.3%, respectively.
The PVT properties of the crude oil system at saturation pressure were p
b
= 2,392 psia, R
sbd
= 921 scf/STB, B
obd
= 1.706 RB/STB, and
= 38.13 lb
m
/ft
3
. Those predicted by the modified EOS were p
b
= 2,497 psia, R
sbd
= 1,061 scf/STB, B
obd
= 1.8973 RB/STB, and =
36.2 lb
m
/ft
3
for errors of 4.4, 15.2, 11.2 and -5.0% for p
b
, R
sd
, B
od
and , respectively. Table 17 gives detailed documentation of
stimulated DE test for the system. PVTSim and modified give values for B
od
that are considerably higher than the observed data,
however; both models match with the CCE data equally well. In terms of R
sd
, the modified EOS performed much better than PVTSim
in the rang of all the pressures used in conducting the test.

Table 17 Expansion Data for Red River Crude Oil
DE at 250F
p R
sd
(scf/STB) B
od
(RB/STB)
2,377.0 1,146.3 1,060 921 1.849 1.8947 1.706
2,250.0 1,081.5 982 872 1.811 1.8428 1.678
1,950.0 939.0 836 761 1.728 1.7509 1.614
1,650.0 809.5 698 657 1.653 1.6622 1.555
1,350.0 690.3 577 561 1.584 1.5843 1.501
1,050.0 578.5 469 467 1.518 1.5144 1.448
750.0 470.9 366 375 1.453 1.4470 1.395
450.0 360.9 266 274 1.384 1.3780 1.334
225.0 265.7 183 191 1.319 1.3170 1.279
125.0 212.0 139 140 1.279 1.2822 1.241
0.0 0.0 000 000 1.059 1.1341 1.102

Wyomings Heavy Crude Oil, Teapot Dome Field This heavy oil contains 9.61 mol% C
1
and 75.01 mol% C
7+
. The plus fraction
is characterized by a molecular weight of 223.5 and a specific gravity of 0.8429. The oil exhibits a bubblepoint of 505 psia at 162F.
The PVT properties were p
b
= 505 psia, R
sbd
= 110 scf/STB, B
obd
= 1.1098 RB/STB, and = 48.77 lb
m
/ft
3
. The model predicted p
b
=
520 psia, R
sbd
= 113 scf/STB, B
obd
= 1.1060 RB/STB, and = 48.77 lb
m
/ft
3
with errors of 2.97, 2.7, 0.34, and 0% for p
b
, R
sbd
, B
obd
and
, respectively. Table 18 shows the close match with the reported data.

SPE 107331 19
Table 18 Expansion Data for Teapot Dome Crude Oil
P,
(psig)
Mod. EOS
R
sd
, scf/STB
Experimental
R
sd
, scf/STB
Mod. EOS
B
od
, RB/STB
Experimental
B
od
, RB/STB
490.3 113 110 1.1060 1.1098
385.3 97 94 1.0991 1.1025
285.3 80 77 1.0920 1.0962
185.3 63 58 1.0840 1.0886
85.3 42 36 1.0740 1.0782
35.3 28 23 1.0665 1.0701
0.0 00 00 1.0481 1.0491

North Sea Gas-Condensate System The system is characterized by a dewpoint of 6.750 psia at 280F. The gas contains 73.19
mol% C
1
and 8.21 mol% C
7+
. The plus fraction is characterized by a molecular weight of 148 and a specific gravity of 0.16. The
maximum liquid dropout with a value of 21.6% occurs at 3,100 psia. Whitson and Torp
10
gave a detailed compositional and
experimental analysis of the system. The modified equation predicts a dewpoint pressure of 6,189 psia. Table 19 gives complete
results of the CVE simulation. The tabulated values show that the EOS predicts the entire liquid dropout and Z-factor data with
average absolute deviations of 4.7 and 1.8%, respectively. Results of the model are generally in excellent agreement with reported
data.

Table 19 CVE at 280F for the North Sea Reservoir
Pressure (psig)
Component 6,750.0 5,500.0 4,300.0 3,100.0 2,100.0 1,200.0 700.0 700.0
*
CO
2
0.02370 0.02379 0.02399 0.02438 0.02493 0.02568 0.02630 0.00748
N
2
0.00310 0.00320 0.00327 0.00334 0.00338 0.00337 0.00332 0.0037
C
1
0.73190 0.75710 0.77263 0.78619 0.79345 0.79296 0.78412 0.10353
C
2
0.07800 0.07950 0.08034 0.08126 0.08227 0.08376 0.08523 0.02357
C
3
0.03550 0.03574 0.03577 0.03585 0.03619 0.03730 0.03903 0.01904
I-C
4
0.00710 0.00706 0.00700 0.00694 0.00696 0.00721 0.00774 0.00600
n-C
4
0.01450 0.01426 0.01403 0.01381 0.01379 0.01438 0.01570 0.01572
I-C
5
0.00640 0.00620 0.00601 0.00580 0.00569 0.00590 0.00659 0.01079
n-C
5
0.00680 0.00652 0.00627 0.00600 0.00583 0.00604 0.00682 0.01355
C
6
0.01090 0.01023 0.00964 0.00893 0.00839 0.00847 0.00966 0.03245
C
7+
0.08210 0.05641 0.04105 0.02749 0.01912 0.01492 0.01549 0.76750

Z- Factor
PVTSim 1.225 1.068 0.966 0.913 0.904 0.922 0.943
Mod. EOS 1.2380 1.0890 0.9720 0.9130 0.9140 0.9370 0.9600
Experimental 1.2618 1.1105 0.9958 0.9239 0.9029 0.9182 0.9407

G
p (%)
PVTSim 0.00000 9.310 21.89 38.620 55.42 71.920 81.270
Mod. EOS 0.00000 9.024 21.744 38.674 55.686 72.146 81.301
Experimental 0.00000 9.512 21.782 38.093 54.800 71.473 80.940

Liquid Dropout (%)
PVTSim 0.0 9.31 22.52 23.52 22.19 20.20 18.85
Mod. EOS 0.0 14.10 19.70 21.60 21.30 20.20 19.30
Experimental 0.0 12.89 19.07 20.92 20.56 19.23 18.13

Conclusions

1. Improved correlations for calculating a(T
c
), b, and (T) of the plus fraction, methane, and nitrogen are presented.
2. The modified PREOS gives hydrocarbon liquid density predictions that are compatible with or better than, the S-K and the A-K
density correlations.
3. The proposed modifications eliminate the need for splitting the heptanes-plus fraction into pseudo-components.
4. The proposed modifications significantly improve the ability of the PREOS to predict the PVT properties of complex
hydrocarbon mixtures.
20 SPE 107331
5. With such significant improvement in the predictive capability of the modified EOS, the equation is recommended for calculating
the volumetric behavior of crude oil and condensate systems

Nomenclature

a, b, A
,
B = EOS constants
B
o
= oil FVF obtained from differential expansion, RB/STB
B
ob
= oil FVF at bubblepoint pressure, RB/STB
J
L
= volume fraction of liquid in expansion cell
G
p
= volume fraction of gas removed from a laboratory CVE cell
k
ij
= binary interaction coefficient between Components i and j
m = characteristic constant
M = molecular weight, lbm/lbm mol
p = pressure, psi
p
b
= bubblepoint pressure, psi
p
c
= critical point, psi
p
s
= saturation pressure, psi
R = universal gas constant, 10.73 psia ft
3
/mol R
R
s
= solution GOR obtained from differential expansion, scf/STB
T = temperature, R
T
c
= critical temperature, R
V = laboratory expansion cell total volume, ft
3
V
L
= volume of liquid in expansion cell, ft
3

V
m
= molar volume
V
s
= volume of expansion cell at saturation pressure, ft
3
x = mole fraction
z = compressibility factor
= correction factor for a
= specific gravity
= density, lbm/ft
3

ob
= density at bubblepoint, lbm/ft
3
= acentric factor
a
,
b
= EOS constants

Subscripts

c = critical
i, j = component number
mix = mixture
o = oil
ob = bubblepoint

References

1. van der Waals, J.D.: On the Continuity of the Liquid and Gaseous State, PhD dissertation, Sigthoff U. Leiden (1873)
2. Peng, D.Y. and Robinson, D.B.: A New Two-Constant Equation of State, Ind. & Eng. Chem. (1976) 15, No. 1, 59-64
3. Soave, G.: Equilibrium Constants from a Modified Redlich-Kwong Equation of State, Chem. Eng. Sci. (1972) 27, 1197-1203
4. Coats, K.H. and Smart, G.T.: Application of a Regression Based EOS PVT Program to Laboratory Data, SPERE (May 1986)
277-99.
5. Wilson, A., Maddox, R.N. and Erbar, J.H.: C Fractions Affect Phase Behavior, Oil & Gas J. (Aug. 21, 1978) 76-81
6. Katz, D.L. and Firoozabadi, A.: Predicting Phase Behavior of Condensates/Crude Oil Systems Using Methane Interaction
Coefficients, JPT (Nov. 1978) 1649-55; Trans., AIME 265
7. Firoozabadi A., Hekim, Y., and Katz, D.L.: Reservoir Depletion Calculations for Gas Condensates Using Extended Analyses in
the Peng-Robinson Equation of State, Cdn. J. Chem. Eng. (1978) 56, 610-15
8. Yarborough, L.: Application of a Generalized Equation of State to Petroleum Reservoir Fluids, Equations of State in
Engineering, Advances in Chemistry Series, K.C. Choa and R.L. Robinson (eds.), American Chemical Soc., Washington DC
(1979) 182, 385-435
9. Whitson, C.H.: Characterizing Hydrocarbon Plus Fractions, SPEJ (Aug. 1983) 683-94
10. Whitson, C.H. and Torp, S.B.: Evaluating Constant Volume Depletion Data: JPT (March 1983) 610-20
SPE 107331 21
11. Lohrenz, J., Bray, B.G., and Clark, C.R.: Calculating Viscosities of Reservoir Fluids From Their Compositions, JPT (Oct.
1964) 1171-76; Trans., AIME 231
12. Katz, D.L.: Overview of Phase Behavior in Oil and Gas Production, JPT (June 1983) 1205-14
13. Ahmed, T., Cady, G., and Story, A.: An Accurate Method of Extending the Analysis of C
7+
, paper SPE 12916 presented at the
1984 SPE Rocky Mountain Regional Meeting, Casper, WY, May 21-23
14. Ahmed, T., Cady, G. and Story, A.: A Generalized Correlation for Characterizing the Hydrocarbon Heavy Fractions, paper SPE
14266 presented at the 1985 SPE Annual Technical Conference and Exhibition, Las Vegas, Sept. 22-25
15. Kenyon, D. and Behie, G.: Third SPE Comparative Solution Project: Gas Cycling of Retrograde Condensate Reservoirs, JPT
(Aug. 1987) 981-97
16. Riazi, M.R. and Daubert, T.E.: Characterization Parameter for Petroleum Fractions, Ind. & Eng. Chem. Res. (197) 26, No. 4
17. Petersen, C.S.: A Systematic and Consistent Approach to Determine Binary Interaction Coefficients for the Peng-Robinson
Equation of State, SPERE (Nov. 1989) 488-96
18. Standing, M.B. and Katz, D.L.: Density of Crude Oils Saturated with Natural Gas, Trans., AIME (1942) 146, 150-65
19. Alani, G.H. and Kennedy, H.T.: Volumes of Liquid Hydrocarbons at High Temperatures and Pressure, Trans., AIME (1960)
219, 288-92
20. Ahmed, T.: Supplement to a Practical Equation of State, SPE 22219 available from SPE Book Order Dept., Richardson, TX.


bbl x 1.589 873 E
-01
= m
3
ft
3
x 2.831 685 E
-02
= m
3

F (F-32)/1.8 =C
lbm x 4.535 924 E
-01
= kg
psi x 6.894 757 E
+00
= kPa

HEPTANES-PLUS CHARACTERIZATION 1
Chapter 5
HeptanesPlus Characterization
5.1 Introduction
Some phase-behavior applications require the use of an equation of
state (EOS) to predict properties of petroleum reservoir fluids. The
critical properties, acentric factor, molecular weight, and binary-in-
teraction parameters (BIPs) of components in a mixture are required
for EOS calculations. With existing chemical-separation techniques,
we usually cannot identify the many hundreds and thousands of com-
ponents found in reservoir fluids. Even if accurate separation were
possible, the critical properties and other EOS parameters of com-
pounds heavier than approximately C
20
would not be known accu-
rately. Practically speaking, we resolve this problem by making an
approximate characterization of the heavier compounds with exper-
imental and mathematical methods. The characterization of heptanes-
plus (C
7)
) fractions can be grouped into three main tasks.
13
1. Dividing the C
7)
fraction into a number of fractions with
known molar compositions.
2. Defining the molecular weight, specific gravity, and boiling
point of each C
7)
fraction.
3. Estimating the critical properties and acentric factor of each C
7)
fraction and the key BIPs for the specific EOS being used.
This chapter presents methods for performing these tasks and
gives guidelines on when each method can be used. A unique char-
acterization does not exist for a given reservoir fluid. For example,
different component properties are required for different EOSs;
therefore, the engineer must determine the quality of a given charac-
terization by testing the predictions of reservoir-fluid behavior
against measured pressure/volume/temperature (PVT) data.
The amount of C
7)
typically found in reservoir fluids varies from
u50 mol% for heavy oils to t1 mol% for light reservoir fluids.
4
Average C
7)
properties also vary widely. For example, C
7)
molec-
ular weight can vary from 110 to u300 and specific gravity from
0.7 to 1.0. Because the C
7)
fraction is a mixture of many hundreds
of paraffinic, naphthenic, aromatic, and other organic compounds,
5
the C
7)
fraction cannot be resolved into its individual components
with any precision. We must therefore resort to approximate de-
scriptions of the C
7)
fraction.
Sec. 5.2 discusses experimental methods available for quantify-
ing C
7)
into discrete fractions. True-boiling-point (TBP) distilla-
tion provides the necessary data for complete C
7)
characterization,
including mass and molar quantities, and the key inspection data for
each fraction (specific gravity, molecular weight, and boiling point).
Gas chromatography (GC) is a less-expensive, time-saving alterna-
tive to TBP distillation. However, GC analysis quantifies only the
mass of C
7)
fractions; such properties as specific gravity and boil-
ing point are not provided by GC analysis.
Typically, the practicing engineer is faced with how to character-
ize a C
7)
fraction when onlyzC
7)
the mole fraction, ; molecular
weight, M
C
7)
; and specific gravity, g
C
7)
, are known. Sec. 5.3 re-
views methods for splitting C
7)
into an arbitrary number of sub-
fractions. Most methods assume that mole fraction decreases expo-
nentially as a function of molecular weight or carbon number. A
more general model based on the gamma distribution has been suc-
cessfully applied to many oil and gas-condensate systems. Other
splitting schemes can also be found in the literature; we summarize
the available methods.
Sec. 5.4 discusses how to estimate inspection properties g and T
b
for C
7)
fractions determined by GC analysis or calculated from a
mathematical split. Katz and Firoozabadis
6
generalized single car-
bon number (SCN) properties are widely used. Other methods for
estimating specific gravities of C
7)
subfractions are based on forc-
ing the calculated g
C
7)
to match the measured value.
Many empirical correlations are available for estimating critical
properties of pure compounds and C
7)
fractions. Critical properties
can also be estimated by forcing the EOS to match the boiling point and
specific gravity of each C
7)
fraction separately. In Sec. 5.5, we review
the most commonly used methods for estimating critical properties.
Finally, Sec. 5.6 discusses methods for reducing the number of
components describing a reservoir mixture and, in particular, the
C
7)
fraction. Splitting the C
7)
into pseudocomponents is particu-
larly important for EOS-based compositional reservoir simulation.
A large part of the computing time during a compositional reservoir
simulation is used to solve the flash calculations; accordingly, mini-
mizing the number of components without jeopardizing the quality
of the fluid characterization is necessary.
5.2 Experimental Analyses
The most reliable basis for C
7)
characterization is experimental
data obtained from high-temperature distillation or GC. Many ex-
perimental procedures are available for performing these analyses;
in the following discussion, we review the most commonly used
methods. TBP distillation provides the key data for C
7)
character-
ization, including mass and molar quantities, specific gravity, mo-
lecular weight, and boiling point of each distillation cut. Other such
inspection data as kinematic viscosity and refractive index also may
be measured on distillation cuts.
Simulated distillation by GC requires smaller samples and less
time than TBP distillation.
7-9
However, GC analysis measures only
the mass of carbon-number fractions. Simulated distillation results
can be calibrated against TBP data, thus providing physical proper-
ties for the individual fractions. For many oils, simulated distillation
2 PHASE BEHAVIOR
Fig. 5.1Standard apparatus for conducting TBP analysis of
crude-oil and condensate samples at atmospheric and subat-
mospheric pressures (after Ref. 11).
Dp
Dp
N
2
N
2
provides the necessary information for C
7)
characterization in far
less the time and at far less cost than that required for a complete
TBP analysis. We recommend, however, that at least one complete
TBP analysis be measured for (1) oil reservoirs that may be candi-
dates for gas injection and (2) most gas-condensate reservoirs.
5.2.1 TBP Distillation. In TBP distillation, a stock-tank liquid (oil
or condensate) is separated into fractions or cuts by boiling-point
range. TBP distillation differs from the Hempel and American Soc.
for Testing Materials (ASTM) D-158 distillations
10
because TBP
analysis requires a high degree of separation, which is usually con-
trolled by the number of theoretical trays in the apparatus and the
reflux ratio. TBP fractions are often treated as components having
unique boiling points, critical temperatures, critical pressures, and
other properties identified for pure compounds. This treatment is
obviously more valid for a cut with a narrow boiling-point range.
The ASTM D-2892
11
procedure is a useful standard for TBP
analysis of stock-tank liquids. ASTM D-2892 specifies the general
procedure for TBP distillation, including equipment specifications
(see Fig. 5.1), reflux ratio, sample size, and calculations necessary
to arrive at a plot of cumulative volume percent vs. normal boiling
point. Normal boiling point implies that boiling point is measured
at normal or atmospheric pressure. In practice, to avoid thermal de-
composition (cracking), distillation starts at atmospheric pressure
and is changed to subatmospheric distillation after reaching a limit-
ing temperature. Subatmospheric boiling-point temperatures are
converted to normal boiling-point temperatures by use of a vapor-
pressure correlation that corrects for the amount of vacuum and the
fractions chemical composition. The boiling-point range for frac-
tions is not specified in the ASTM standard. Katz and Firoozabadi
6
recommend use of paraffin normal boiling points (plus 0.5C) as
boundaries, a practice that has been widely accepted.
Fig. 5.2TBP curve for a North Sea gas-condensate sample il-
lustrating the midvolume-point method for calculating average
boiling point (after Austad et al.
7
).
Cutoff (n-paraffin) boiling point
Midvolume (normal) boiling point
Fig. 5.2
7
shows a plot of typical TBP data for a North Sea sample.
Normal boiling point is plotted vs. cumulative volume percent.
Table 5.1 gives the data, including measured specific gravities and
molecular weights. Average boiling point is usually taken as the val-
ue found at the midvolume percent of a cut. For example, the third
cut in Table 5.1 boils from 258.8 to 303.8F, with an initial 27.49
vol% and a final 37.56 vol%. The midvolume percent is
(27.49)37.56)/2+32.5 vol%; from Fig. 5.2, the boiling point at
this volume is [282F. For normal-paraffin boiling-point intervals,
Katz and Firoozabadis
6
average boiling points of SCN fractions
can be used (see Table 5.2).
The mass, m
i
, of each distillation cut is measured directly during
a TBP analysis. The cut is quantified in moles n
i
with molecular
weight, M
i
, and the measured mass m
i
, where n
i
+m
i
M
i
. Volume
of the fraction is calculated from the mass and the density,
i
(or spe-
cific gravity, g
i
), where V
i
+m
i
i
. M
i
is measured by a cryoscop-
ic method based on freezing-point depression, and
i
is measured
by a pycnometer or electronic densitometer. Table 5.1 gives cumula-
tive weight, mole, and volume percents for the North Sea sample.
Average C
7)
properties are given by
M
C
7)
+
N
i+1
m
i
N
i+1
n
i
and
C
7)
+
N
i+1
m
i
N
i+1
V
i
, (5.1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
where
C
7)
+g
C
7)
w
with
w
+pure water density at standard
conditions. These calculated averages are compared with mea-
sured values of the C
7)
sample, and discrepancies are reported as
lost material.
Refs. 7 and 15 through 20 give procedures for calculating proper-
ties from TBP analyses. Also, the ASTM D-2892
11
procedure gives
details on experimental equipment and the procedure for conducting
TBP analysis at atmospheric and subatmospheric conditions. Table
5.3 gives an example TBP analysis from a commercial laboratory.
TABLE 5.1EXPERIMENTAL TBP RESULTS FOR A NORTH SEA CONDENSATE
Fraction
Upper
T
bi
(F)
Average
T
bi
*
(F)
m
i
(g) g
i
**
M
i
(g/mol)
V
i
(cm
3)
n
i
(mol)
w
i
(%)
x
Vi
%
x
i
%
Sw
i
%
Sx
Vi
% K
w
C
7
C
8
C
9
C
10
C
11
C
12
C
13
C
14
C
15
C
16
C
17
C
18
C
19
C
20
C
21)
208.4
258.8
303.8
347.0
381.2
420.8
455.0
492.8
523.4
550.4
579.2
604.4
629.6
653.0
194.0
235.4
282.2
325.4
363.2
401.1
438.8
474.8
509.0
537.8
564.8
591.8
617.0
642.2
90.2
214.6
225.3
199.3
128.8
136.8
123.8
120.5
101.6
74.1
76.8
58.2
50.2
45.3
427.6
0.7283
0.7459
0.7658
0.7711
0.7830
0.7909
0.8047
0.8221
0.8236
0.8278
0.8290
0.8378
0.8466
0.8536
0.8708
96
110
122
137
151
161
181
193
212
230
245
259
266
280
370
123.9
287.7
294.2
258.5
164.5
173.0
153.8
146.6
123.4
89.5
92.6
69.5
59.3
53.1
491.1
0.940
1.951
1.847
1.455
0.853
0.850
0.684
0.624
0.479
0.322
0.313
0.225
0.189
0.162
1.156
4.35
10.35
10.87
9.61
6.21
6.60
5.97
5.81
4.90
3.57
3.70
2.81
2.42
2.19
20.63
4.80
11.15
11.40
10.02
6.37
6.70
5.96
5.68
4.78
3.47
3.59
2.69
2.30
2.06
19.03
7.80
16.19
15.33
12.07
7.08
7.05
5.68
5.18
3.98
2.67
2.60
1.87
1.57
1.34
9.59
4.35
14.70
25.57
35.18
41.40
48.00
53.97
59.78
64.68
68.26
71.96
74.77
77.19
79.37
100.00
4.80
15.95
27.35
37.37
43.74
50.44
56.41
62.09
66.87
70.33
73.92
76.62
78.91
80.97
100.00
11.92
11.88
11.82
11.96
11.97
12.03
11.99
11.89
12.01
12.07
12.16
12.14
12.11
12.10
Sum 2,073.1 2,580.5 12.049 100.00 100.00 100.00
Average 0.8034 172 11.98
Reflux ratio+1 : 5; reflux cycle+18 seconds; distillation at atmospheric pressure+201.2 to 347F; distillation at 100 mm Hg+347 to 471.2F; and distillation at 10 mm Hg+471.2
to 653F.
V
i
+m
i
/g
i
/0.9991; n
i
+m
i
/M
i
; w
i
+100m
i
/2073.1; x
Vi
+100Vi/2580.5; x
i
+100n
i
/12.049; Sw
i
+Sw
i
; Sx
Vi
+Sx
Vi
; and K
w
+(T
bi
+460)
1/3
/g
i
.
*Average taken at midvolume point.
**Water+1.
Boiling points are not reported because normal-paraffin boiling-point
intervals are used as a standard; accordingly, the average boiling
points suggested by Katz and Firoozabadi
6
(Table 5.2) can be used.
5.2.2 Chromatography. GC and, to a lesser extent, liquid chroma-
tography are used to quantify the relative amount of compounds
found in oil and gas systems. The most important application of
chromatography to C
7)
characterization is simulated distillation by
GC techniques.
Fig. 5.3 shows an example gas chromatogram for the North Sea
sample considered earlier. The dominant peaks are for normal paraf-
fins, which are identified up to n-C
22
. As a good approximation for
a paraffinic sample, the GC response for carbon number C
i
starts at
the bottom response of n-C
i *1
and extends to the bottom response
of n-C
i
. The mass of carbon number C
i
is calculated as the area under
the curve from the baseline to the GC response in the n-C
i *1
to n-C
i
interval (see the shaded area for fraction C
9
in Fig. 5.3). As Fig. 5.4
7
shows schematically, the baseline should be determined before run-
ning the actual chromatogram.
Because stock-tank samples cannot be separated completely by
standard GC analysis, an internal standard must be used to relate GC
area to mass fraction. Normal hexane was used as an internal stan-
dard for the sample in Fig. 5.3. The internal standards response fac-
tor may need to be adjusted to achieve consistency between simu-
lated and TBP distillation results. This factor will probably be
constant for a given oil, and the factor should be determined on the
basis of TBP analysis of at least one sample from a given field. Fig.
5.5 shows the simulated vs. TBP distillation curves for the Austad
et al.
7
sample. A 15% correction to the internal-standard response
factor was used to match the two distillation curves.
As an alternative to correcting the internal standard, Maddox and
Erbar
15
suggest that the reported chromatographic boiling points be
adjusted by a correction factor that depends on the reported boiling
point and the paraffinicity of the composite sample. This correc-
tion factor varies from 1 to 1.15 and is slightly larger for aromatic
than paraffinic samples.
Several laboratories have calibrated GC analysis to provide simu-
lated-distillation results up to C
40
. However, checking the accuracy
of simulated distillation for SCN fractions greater than approxi-
mately C
25
is difficult because C
25
is usually the upper limit for reli-
able TBP distillation. The main disadvantage of simulated distilla-
tion is that inspection data are not determined directly for each
fraction and must therefore either be correlated from TBP data or es-
timated from correlations (see Sec. 5.4).
Sophisticated analytical methods, such as mass spectroscopy,
may provide detailed information on the compounds separated by
GC. For example, mass spectroscopy GC can establish the relative
amounts of paraffins, naphthenes, and aromatics (PNAs) for car-
bon-number fractions distilled by TBP analysis. Detailed PNA in-
formation should provide more accurate estimation of the critical
properties of petroleum fractions, but the analysis is relatively cost-
ly and time-consuming from a practical point of view. Recent work
has shown that PNA analysis
3,19-23
may improve C
7)
characteriza-
tion for modeling phase behavior with EOSs. Our experience, how-
ever, is that PNA data have limited usefulness for improving EOS
fluid characterizations.
5.3 Molar Distribution
Molar distribution is usually thought of as the relation between mole
fraction and molecular weight. In fact, this concept is misleading be-
cause a unique relation does not exist between molecular weight and
mole fraction unless the fractions are separated in a consistent man-
ner. Consider for example a C
7)
sample distilled into 10 cuts sepa-
rated by normal-paraffin boiling points. If the same C
7)
sample is
distilled with constant 10-vol% cuts, the two sets of data will not
4 PHASE BEHAVIOR
TABLE 5.2SINGLE CARBON NUMBER PROPERTIES FOR HEPTANES-PLUS (after Katz and Firoozabadi
6
)
Katz-Firoozabadi Generalized Properties
T
b
Interval*
Lee-Kesler
12
/Kesler-Lee
13
Correlations Riazi
14
Defined
Fraction
Number
Lower
(F)
Upper
(F)
Average T
b
(F) (R) g* M
Defined
K
w
T
c
(R)
p
c
(psia)
w
V
c
(ft
3
/lbm mol) Z
c
6 97.7 156.7 147.0 606.7 0.690 84 12.27 914 476 0.271 5.6 0.273
7 156.7 210.0 197.4 657.1 0.727 96 11.96 976 457 0.310 6.2 0.272
8 210.0 259.0 242.1 701.7 0.749 107 11.86 1,027 428 0.349 6.9 0.269
9 259.0 304.3 288.0 747.6 0.768 121 11.82 1,077 397 0.392 7.7 0.266
10 304.3 346.3 330.4 790.1 0.782 134 11.82 1,120 367 0.437 8.6 0.262
11 346.3 385.5 369.0 828.6 0.793 147 11.84 1,158 341 0.479 9.4 0.257
12 385.5 422.2 406.9 866.6 0.804 161 11.86 1,195 318 0.523 10.2 0.253
13 422.2 456.6 441.0 900.6 0.815 175 11.85 1,228 301 0.561 10.9 0.249
14 456.6 489.0 475.5 935.2 0.826 190 11.84 1,261 284 0.601 11.7 0.245
15 489.0 520.0 510.8 970.5 0.836 206 11.84 1,294 268 0.644 12.5 0.241
16 520.0 548.6 541.4 1,001.1 0.843 222 11.87 1,321 253 0.684 13.3 0.236
17 548.6 577.4 572.0 1,031.7 0.851 237 11.87 1,349 240 0.723 14.0 0.232
18 577.4 602.6 595.4 1,055.1 0.856 251 11.89 1,369 230 0.754 14.6 0.229
19 602.6 627.8 617.0 1,076.7 0.861 263 11.90 1,388 221 0.784 15.2 0.226
20 627.8 651.2 640.4 1,100.1 0.866 275 11.92 1,408 212 0.816 15.9 0.222
21 651.2 674.6 663.8 1,123.5 0.871 291 11.94 1,428 203 0.849 16.5 0.219
22 674.6 692.6 685.4 1,145.1 0.876 305 11.94 1,447 195 0.879 17.1 0.215
23 692.6 717.8 707.0 1,166.7 0.881 318 11.95 1,466 188 0.909 17.7 0.212
24 717.8 737.6 726.8 1,186.5 0.885 331 11.96 1,482 182 0.936 18.3 0.209
25 737.6 755.6 746.6 1,206.3 0.888 345 11.99 1,498 175 0.965 18.9 0.206
26 755.6 775.4 766.4 1,226.1 0.892 359 12.00 1,515 168 0.992 19.5 0.203
27 775.4 793.4 786.2 1,245.9 0.896 374 12.01 1,531 163 1.019 20.1 0.199
28 793.4 809.6 804.2 1,263.9 0.899 388 12.03 1,545 157 1.044 20.7 0.196
29 809.6 825.8 820.4 1,280.1 0.902 402 12.04 1,559 152 1.065 21.3 0.194
30 825.8 842.0 834.8 1,294.5 0.905 416 12.04 1,571 149 1.084 21.7 0.191
31 842.0 858.2 851.0 1,310.7 0.909 430 12.04 1,584 145 1.104 22.2 0.189
32 858.2 874.4 865.4 1,325.1 0.912 444 12.04 1,596 141 1.122 22.7 0.187
33 874.4 888.8 879.8 1,339.5 0.915 458 12.05 1,608 138 1.141 23.1 0.185
34 888.8 901.4 892.4 1,352.1 0.917 472 12.06 1,618 135 1.157 23.5 0.183
35 901.4 915.8 906.8 1,366.5 0.920 486 12.06 1,630 131 1.175 24.0 0.180
36 919.4 1,379.1 0.922 500 12.07 1,640 128 1.192 24.5 0.178
37 932.0 1,391.7 0.925 514 12.07 1,650 126 1.207 24.9 0.176
38 946.4 1,406.1 0.927 528 12.09 1,661 122 1.226 25.4 0.174
39 959.0 1,418.7 0.929 542 12.10 1,671 119 1.242 25.8 0.172
40 971.6 1,431.3 0.931 556 12.10 1,681 116 1.258 26.3 0.170
41 982.4 1,442.1 0.933 570 12.11 1,690 114 1.272 26.7 0.168
42 993.2 1,452.9 0.934 584 12.13 1,697 112 1.287 27.1 0.166
43 1,004.0 1,463.7 0.936 598 12.13 1,706 109 1.300 27.5 0.164
44 1,016.6 1,476.3 0.938 612 12.14 1,716 107 1.316 27.9 0.162
45 1,027.4 1,487.1 0.940 626 12.14 1,724 105 1.328 28.3 0.160
*At 1 atmosphere.
**Water+1.
produce the same plot of mole fraction vs. molecular weight. How-
ever, a plot of cumulative mole fraction,
Q
zi
+
i
j+1
z
j
N
j+1
z
j
, (5.2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
vs. cumulative average molecular weight,
Q
Mi
+
i
j+1
z
j
M
j
i
j+1
z
j
, (5.3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
TABLE 5.3STANDARD TBP RESULTS FROM COMMERCIAL PVT LABORATORY
Component mol% wt%
Density
(g/cm
3
)
Gravity
g
API
Molecular
Weight
Heptanes
Octanes
Nonanes
Decanes
Undecanes
Dodecanes
Tridecanes
Tetradecanes
Pentadecanes plus
1.12
1.30
1.18
0.98
0.62
0.57
0.74
0.53
4.10
2.52
3.08
3.15
2.96
2.10
2.18
3.05
2.39
31.61
0.7258
0.7470
0.7654
0.7751
0.7808
0.7971
0.8105
0.8235
0.8736
63.2
57.7
53.1
50.9
49.5
45.8
42.9
40.1
30.3
96
101
114
129
144
163
177
192
330
*At 60F.
Note: Katz and Firoozabadi
6
average boiling points (Table 5.2) can be used when normal paraffin boiling-point intervals are used.
should produce a single curve. Strictly speaking, therefore, molar
distribution is the relation between cumulative molar quantity and
some expression for cumulative molecular weight.
In this section, we review methods commonly used to describe
molar distribution. Some methods use a consistent separation of
fractions (e.g., by SCN) so the molar distribution can be expressed
directly as a relationship between mole fraction and molecular
weight of individual cuts. Most methods in this category assume that
C
7)
mole fractions decrease exponentially. A more general ap-
proach uses the continuous three-parameter gamma probability
function to describe molar distribution.
5.3.1 Exponential Distributions. The Lohrenz-Bray-Clark
24
(LBC) viscosity correlation is one of the earliest attempts to use an
exponential-type distribution for splitting C
7)
. The LBC method
splits C
7)
into normal paraffins C
7
though C
40
with the relation
z
i
+z
C
6
exp[A
1
(i *6) )A
2
(i *6)
2
], (5.4) . . . . . . . . . . . . .
where i+carbon number and zC
6
+measured C
6
mole fraction.
Constants A
1
and A
2
are determined by trial and error so that
z
C
7)
+
40
i+7
z
i
(5.5) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Fig. 5.3Simulated distillation by GC of the North Sea gas-con-
densate sample in Fig. 5.2 (after Austad et al.
7
).
and z
C
7)
M
C
7)
+
40
i+7
z
i
M
i
(5.6) . . . . . . . . . . . . . . . . . . . . . . . .
are satisfied. Paraffin molecular weights (M
i
+14i)2) are used in
Eq. 5.6. A Newton-Raphson algorithm can be used to solve Eqs. 5.5
and 5.6. Note that the LBC model cannot be used when z
C
7)
tz
C
6
and M
C
7)
uM
C
40
. The LBC form of the exponential distribution
has not found widespread application.
More commonly, a linear form of the exponential distribution is
used to split the C
7)
fraction. Writing the exponential distribution
in a general form for any C
n)
fraction (n+7 being a special case),
z
i
+z
C
n
exp A[(i *n)], (5.7) . . . . . . . . . . . . . . . . . . . . . . . .
where i+carbon number, z
C
n
+mole fraction of C
n
, and
A+constant indicating the slope on a plot of ln z
i
vs. i. The constants
z
C
n
and A can be determined explicitly. With the general expression
M
i
+14i ) h (5.8) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
for molecular weight of C
i
and the assumption that the distribution
is infinite, constants z
C
n
and A are given by
z
C
n
+
14
M
C
n)
*14(n *1) *h
(5.9) . . . . . . . . . . . . . . . . . .
and A +ln
1 *z
C
n
(5.10) . . . . . . . . . . . . . . . . . . . . . . . . . . .
so that
R
i+n
z
i
+1 (5.11) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Fig. 5.4GC simulated distillation chromatograms (a) without
any sample (used to determine the baseline), (b) for a crude oil,
and (c) for a crude oil with internal standard (after MacAllister
and DeRuiter
9
).
(a)
(b)
(c)
6 PHASE BEHAVIOR
Fig. 5.5Comparison of TBP and GC-simulated distillation for
a North Sea gas-condensate sample (after Austad et al.
7
).
and
R
i+n
z
i
M
i
+M
C
n)
(5.12) . . . . . . . . . . . . . . . . . . . . . . . . . . .
are satisfied.
Eqs. 5.9 and 5.10 imply that once a molecular weight relation is cho-
sen (i.e., h is fixed), the distribution is uniquely defined by C
7)
molec-
ular weight. Realistically, all reservoir fluids having a given C
7)
mo-
lecular weight will not have the same molar distribution, which is one
reason why more complicated models have been proposed.
5.3.2 Gamma-Distribution Model. The three-parameter gamma
distribution is a more general model for describing molar distribu-
tion. Whitson
2,25,26
and Whitson et al.
27
discuss the gamma dis-
tribution and its application to molar distribution. They give results
for 44 oil and condensate C
7)
samples that were fit by the gamma
distribution with data from complete TBP analyses. The absolute
average deviation in estimated cut molecular weight was 2.5 amu
(molecular weight units) for the 44 samples.
The gamma probability density function is
p(M) +
(M*h)
a*1
exp
M*h
b
a
G(a)
, (5.13) . . . . . . . .
where G+gamma function and b is given by
b +
M
C
7)
*h
a
. (5.14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The three parameters in the gamma distribution are a, h, and
M
C
7)
The key parameter a defines the form of the distribution, and
its value usually ranges from 0.5 to 2.5 for reservoir fluids; a+1
gives an exponential distribution. Application of the gamma dis-
tribution to heavy oils, bitumen, and petroleum residues indicates
that the upper limit for a is 25 to 30, which statistically is approach-
ing a log-normal distribution (see Fig. 5.6
28
).
The parameter h can be physically interpreted as the minimum
molecular weight found in the C
7)
fraction. An approximate rela-
tion between a and h is
h [
110
1 *
1 )4a
0.7
(5.15) . . . . . . . . . . . . . . . . . . . . . . . . .
Fig. 5.6Gamma distributions for petroleum residue (after
Brul et al.
28
).
700 to 1,000F Distillate
1,000 to 1,250F Distillate
1,200F Residue
for reservoir-fluid C
7)
fractions. Practically, h should be considered
as a mathematical constant more than as a physical property, either
calculated from Eq. 5.15 or determined by fitting measured TBP data.
Fig. 5.7 shows the function p(M) for the Hoffman et al.
29
oil and
a North Sea oil. Parameters for these two oils were determined by fit-
ting experimental TBP data. The Hoffman et al. oil has a relatively
large a of 2.27, a relatively small h of 75.7, with MC
7)
+198; the
North Sea oil is described by a+0.82, h+93.2, and MC
7)
+227.
The continuous distribution p(M) is applied to petroleum frac-
tions by dividing the area under the p(M) curve into sections (shown
schematically in Fig. 5.8). By definition, total area under the p(M)
curve from h to R is unity. The area of a section is defined as
normalized mole fraction z
i
z
C
7)
for the range of molecular
weights M
bi *1
to M
bi
. If the area from h to molecular-weight
boundary M
b
is defined as P
0
(M
b
), then the area of Section i is
P
0
(M
bi
)*P
0
(M
bi *1
), also shown schematically in Fig. 5.8. Mole
fraction z
i
can be written
z
i
+z
C
7)
P
0
M
b i
*P
0
M
b i*1
. (5.16) . . . . . . . . . . . . . . .
Average molecular weight in the same interval is given by
M
i
+h ) ab
P
1
M
b i
*P
1
M
b i*1
P
0
M
b i
*P
0
M
b i*1
, (5.17) . . . . . . . . . . .
where P
0
+QS, (5.18) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
and P
1
+Q
S *
1
a
, (5.19) . . . . . . . . . . . . . . . . . . . . . . . . . . .
Fig. 5.7Gamma density function for the Hoffman et al.
29
oil
(dashed line) and a North Sea volatile oil (solid line). After Whit-
son et al.
27
a +2.273
h +75.7
M
C
7)
+198.4
a +0.817
h +93.2
M
C
7)
+227
where Q +e
*y
y
a
G(a), (5.20) . . . . . . . . . . . . . . . . . . . . . . . . .
S +
R
j+0
y
j
j
k+0
(a )k)
*1
, (5.21) . . . . . . . . . . . . . . . . . . .
and y +
M
b
*h
b
. (5.22) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Note that P
0
(M
b0
+h)+P
1
(M
b0
+h)+0.
The summation in Eq. 5.21 should be performed until the last
term is t110
*8
. The gamma function can be estimated by
30
G
x )1
+1 )
8
i+1
A
i
x
i
, (5.23) . . . . . . . . . . . . . . . . . . . . .
where A
1
+*0.577191652, A
2
+0.988205891, A
3
+*0.897056937,
A
4
+0.918206857, A
5
+*0.756704078, A
6
+0.482199394, A
7
+
*0.193527818, and A
8
+0.035868343 for 0xxx1. The recurrence
formula, G(x)1)+xG(x), is used for xu1 and xt1; furthermore,
G(1)+1.
The equations for calculating z
i
and M
i
are summarized in a short
FORTRAN program GAMSPL found in Appendix A. In this simple
program, the boundary molecular weights are chosen arbitrarily at
increments of 14 for the first 19 fractions, starting with h as the first
lower boundary. The last fraction is calculated by setting the upper
molecular-weight boundary equal to 10,000. Table 5.4 gives three
sample outputs from GAMSPL for a+0.5, 1, and 2 with h+90 and
M
C
7)
+200 held constant. Fig. 5.9 plots the results as log z
i
vs. M
i
.
The amount and molecular weight of the C
26)
fraction varies for
each value of a.
The gamma distribution can be fit to experimental molar-distribu-
tion data by use of a nonlinear least-squares algorithm to determine
a, h, and b. Experimental TBP data are required, including weight
fraction and molecular weight for at least five C
7)
fractions (use of
more than 10 fractions is recommended to ensure a unique fit of mod-
el parameters). The sum-of-squares function can be defined as
F
a, h, b
+
N*1
i+1
(D
Mi
)
2
, (5.24) . . . . . . . . . . . . . . . . . . . . . . .
where D
Mi
+
(M
i
)
mod
*(M
i
)
exp
(M
i
)
exp
. (5.25) . . . . . . . . . . . . . . . . . .
Subscripts mod and exp+model and experimental, respectively. This
sum-of-squares function weights the lower molecular weights more
than higher molecular weights, in accordance with the expected accu-
racy for measurement of molecular weight. Also, the sum-of-squares
function does not include the last molecular weight because this mo-
lecular weight may be inaccurate or backcalculated to match the mea-
sured average C
7)
molecular weight. If the last fraction is not in-
cluded, the model average molecular weight, (M
C
7)
)
mod
+h )ab,
can be compared with the experimental value as an independent
check of the fit.
A simple graphical procedure can be used to fit parameters a and
h if experimental M
C
7)
is fixed and used to define b. Fig. 5.10
shows a plot of cumulative weight fraction,
Q
wi
+
i
j+1
w
i
, (5.26) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
vs. the cumulative dimensionless molecular-weight variable,
Q
*
Mi
+
Q
Mi
*h
M
C
7)
*h
. (5.27) . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 5.5 and the following outline describe the procedure for deter-
mining model parameters with Fig. 5.10 and TBP data.
1. Tabulate measured mole fractions z
i
and molecular weights M
i
for each fraction.
2. Calculate experimental weight fractions, w
i
+(z
i
M
i
)
B(z
C
7)
M
C
7)
), if they are not reported.
3. Normalize weight fractions and calculate cumulative normal-
ized weight fraction Q
wi
.
4. Calculate cumulative molecular weight Q
Mi
from Eq. 5.3.
5. Assume several values of h (e.g., from 75 to 100) and calculate
Q
*
Mi
for each value of the estimated h.
6. For each estimate of h, plot Q
*
Mi
vs. Q
wi
on a copy of Fig. 5.10
and choose the curve that fits one of the model curves best. Read the
value of a from Fig. 5.10.
7. Calculate molecular weights and mole fractions of Fractions i
using the best-fit curve in Fig. 5.10. Enter the curve at measured val-
ues of Q
wi
, read Q
*
Mi
, and calculate M
i
from
M
i
+h )
M
C
7)
*h
Q
wi
*Q
wi*1
Q
wi
Q
*
Mi
Q
wi*1
Q
*
Mi*1
.
(5.28) . . . . . . . . . . . . . . . . . . .
Fig. 5.8Schematic showing the graphical interpretation of areas under the gamma density
function p(M) that are proportional to normalized mole fraction; A+area.
h
+P
0
M
bi
*P
0
M
bi*1
hM
bi
A +P
0
M
bi
A +P
0
M
bi*1
hM
bi*1
p(M)
A +z
i
z
C
7)
8 PHASE BEHAVIOR
TABLE 5.4RESULTS OF GAMSPL PROGRAM FOR THREE DATA SETS WITH DIFFERENT
GAMMA-DISTRIBUTION PARAMETER a
a+0.5 a+1.0 a+2.0
Fraction
Number
Mole
Fraction
Molecular
Weight
Mole
Fraction
Molecular
Weight
Mole
Fraction
Molecular
Weight
1 0.2787233 94.588 0.1195065 96.852 0.0273900 99.132
2 0.1073842 110.525 0.1052247 110.852 0.0655834 111.490
3 0.0772607 124.690 0.0926497 124.852 0.0852269 125.172
4 0.0610991 138.758 0.0815774 138.852 0.0927292 139.038
5 0.0505020 152.796 0.0718284 152.852 0.0925552 152.963
6 0.0428377 166.819 0.0632444 166.852 0.0877762 166.916
7 0.0369618 180.836 0.0556863 180.852 0.0804707 180.883
8 0.0322804 194.848 0.0490314 194.852 0.0720157 194.859
9 0.0284480 208.857 0.0431719 208.852 0.0632969 208.841
10 0.0252470 222.864 0.0380125 222.852 0.0548597 222.826
11 0.0225321 236.870 0.0334698 236.852 0.0470180 236.814
12 0.0202013 250.875 0.0294699 250.852 0.0399302 250.805
13 0.0181808 264.879 0.0259481 264.852 0.0336535 264.797
14 0.0164152 278.883 0.0228471 278.852 0.0281813 278.790
15 0.0148619 292.886 0.0201167 292.852 0.0234690 292.784
16 0.0134879 306.888 0.0177127 306.852 0.0194514 306.778
17 0.0122665 320.890 0.0155959 320.852 0.0160543 320.774
18 0.0111762 334.892 0.0137321 334.852 0.0132017 334.770
19 0.0101996 348.894 0.0120910 348.852 0.0108204 348.766
20 0.1199341 539.651 0.0890834 466.000 0.0463166 420.424
Total 1.0000000 1.0000000 1.0000000
Average 200 200 200
For all three cases h +90 and M
C
7)
+200.
Mole fractions z
i
are given by
z
i
+z
C7)
Q
wi
Q
*
Mi
*
Q
wi*1
Q
*
Mi*1
. (5.29) . . . . . . . . . . . . . . . . . . .
For computer applications, Q
wi
and Q
*
Mi
can be calculated exactly
from Eqs. 5.16 through 5.23 with little extra effort.
M
C
7)
Fig. 5.9Three example molar distributions for an oil sample
with =200 and h=90, calculated with the GAMSPL program
(Table A-4) in Table 5.4.
M
C
7)
+200
h + 90
DM
b
+ 14 a + 2.0
a + 0.5
a + 1.0
}
f
V
Fig. 5.11 shows a Q
*
Mi
*Q
wi
match for the Hoffman et al.
29
oil
with h+70, 72.5, 75, and 80 and indicates that a best fit is achieved
for h+72.5 and a+2.5 (see Fig. 5.12).
Although the gamma-distribution model has the flexibility of
treating reservoir fluids from light condensates to bitumen, most
reservoir fluids can be characterized with an exponential molar dis-
tribution (a+1) without adversely affecting the quality of EOS pre-
M
C
7)
Fig. 5.10Cumulative-distribution type curve for fitting exper-
imental TBP data to the gamma-distribution model. Parameters
a and h are determined with held constant.
1.0
0.0
0.8
0.6
0.4
0.2
0.0 0.2 0.4 0.6 0.8 1.0
Cumulative Normalized Mole Fraction, Q
zi
(

)
(

)
TABLE 5.5CALCULATION OF CUMULATIVE WEIGHT FRACTION AND
CUMULATIVE MOLECULAR WEIGHT VARIABLE FOR HOFFMAN et al.
29
OIL
Q
*
Mi
Component
i z
i
Sz
i
M
i
z
i
M
i
Sz
i
M
i
Q
wi
Q
Mi
h+70 h+72.5 h+75 h+80 h+85
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
0.0263
0.0234
0.0235
0.0224
0.0241
0.0246
0.0266
0.0326
0.0363
0.0229
0.0171
0.0143
0.0130
0.0108
0.0087
0.0072
0.0058
0.0048
0.0039
0.0034
0.0028
0.0025
0.0023
0.0091
0.0263
0.0497
0.0732
0.0956
0.1197
0.1443
0.1709
0.2035
0.2398
0.2627
0.2799
0.2941
0.3072
0.3180
0.3267
0.3338
0.3396
0.3444
0.3483
0.3517
0.3545
0.3570
0.3593
0.3684
99
110
121
132
145
158
172
186
203
222
238
252
266
279
290
301
315
329
343
357
371
385
399
444
2.604
2.574
2.844
2.957
3.497
3.882
4.570
6.067
7.371
5.093
4.079
3.596
3.466
3.008
2.526
2.152
1.811
1.582
1.351
1.196
1.039
0.963
0.926
4.049
2.604
5.178
8.021
10.978
14.475
18.357
22.928
28.995
36.366
41.458
45.538
49.134
52.600
55.607
58.133
60.285
62.097
63.679
65.031
66.227
67.265
68.228
69.154
73.203
0.036
0.071
0.110
0.150
0.198
0.251
0.313
0.396
0.497
0.566
0.622
0.671
0.719
0.760
0.794
0.824
0.848
0.870
0.888
0.905
0.919
0.932
0.945
1.000
99.0
104.2
109.6
114.8
120.9
127.2
134.2
142.5
151.7
157.8
162.7
167.0
171.2
174.9
178.0
180.6
182.9
184.9
186.7
188.3
189.8
191.1
192.5
198.7
0.225
0.266
0.308
0.348
0.396
0.445
0.499
0.563
0.634
0.682
0.720
0.754
0.787
0.815
0.839
0.859
0.877
0.893
0.907
0.919
0.931
0.941
0.952
1.000
0.210
0.251
0.294
0.335
0.384
0.434
0.489
0.555
0.627
0.676
0.715
0.749
0.782
0.811
0.836
0.857
0.875
0.891
0.905
0.918
0.929
0.940
0.951
1.000
0.194
0.236
0.280
0.322
0.371
0.422
0.478
0.546
0.620
0.669
0.709
0.744
0.778
0.808
0.832
0.854
0.872
0.889
0.903
0.916
0.928
0.939
0.950
1.000
0.160
0.204
0.249
0.293
0.345
0.398
0.457
0.526
0.604
0.655
0.697
0.733
0.769
0.799
0.825
0.847
0.867
0.884
0.899
0.913
0.925
0.936
0.948
1.000
0.123
0.169
0.216
0.262
0.316
0.371
0.433
0.506
0.586
0.640
0.683
0.722
0.758
0.791
0.818
0.841
0.861
0.879
0.894
0.909
0.921
0.933
0.945
1.000
Total 0.3684 198.7 73.203
dictions. Whitson et al.
27
proposed perhaps the most useful applica-
tion of the gamma-distribution model. With Gaussian quadrature,
their method allows multiple reservoir-fluid samples from a com-
mon reservoir to be treated simultaneously with a single fluid char-
acterization. Each fluid sample can have different C
7)
properties
when the split is made so that each split fraction has the same molec-
ular weight (and other properties, such as g, T
b
, T
c
, p
c
, and w), while
Fig. 5.11Graphical fit of the Hoffman et al.
29
oil molar distribu-
tion by use of the cumulative-distribution type curve. Best-fit
model parameters are a=2.5 and h=72.5.
1.0
0.0
0.8
0.6
0.4
0.2
0.0 0.2 0.4 0.6 0.8 1.0
J
Cumulative Normalized Mole Fraction, Q
zi
h + 70
h + 75
h + 80
h + 65 Y
F
X
the mole fractions are different for each fluid sample. Example ap-
plications include the characterization of a gas cap and underlying
reservoir oil and a reservoir with compositional gradient.
The following outlines the procedure for applying Gaussian
quadrature to the gamma-distribution function.
1. Determine the number of C
7)
fractions, N, and obtain the
quadrature values X
i
and W
i
from Table 5.6 (values are given for
N+3 and N+5).
2. Specify h and a. When TBP data are not available to determine
these parameters, recommended values are h+90 and a+1.
3. Specify the heaviest molecular weight of fraction N (recom-
mended value is M
N
+2.5M
C
7)
). Calculate a modified b
*
term,
b
*
+
M
N
*h
X
N
.
Fig. 5.12Calculated normalized mole fraction vs. molecular
weight of fractions for the Hoffman et al.
29
oil based on the best
fit in Fig. 5.11 with a=2.5 and h=72.5.
10 PHASE BEHAVIOR
TABLE 5.6GAUSSIAN QUADRATURE FUNCTION
VARIABLES, X, AND WEIGHT FACTORS, W
X W
Three Quadrature Points (plus fractions)
1
2
3
0.415 774 556 783
2.294 280 360 279
6.289 945 082 937
7.110 930 099 2910
*1
2.785 177 335 6910
*1
1.038 925 650 1610
*2
Five Quadrature Points (plus fractions)
1
2
3
4
5
0.263 560 319 718
1.413 403 059 107
3.596 425 771 041
7.085 810 005 859
12.640 800 844 276
5.217 556 105 8310
*1
3.986 668 110 8310
*1
7.594 244 968 1710
*2
3.611 758 679 9210
*3
2.336 997 238 5810
*5
Quadrature function values and weight factors can be found for other quadrature numbers
in mathematical handbooks.
30
4. Calculate the parameter d.
d +exp
ab
*
M
C
7)
*h
*1. (5.30) . . . . . . . . . . . . . . . . . . .
5. Calculate the C
7)
mole fraction z
i
and M
i
for each fraction.
z
i
+z
C
7)
W
i
f (X
i
)
,
M
i
+h ) b
*
X
i
,
and f(X) +
(X)
a*1
G(a)
1 ) ln d
a
d
X
. (5.31) . . . . . . . . . . . . . . . . . .
6. Check whether the calculated M
C
7)
from Eq. 5.12 equals the
measured value used in Step 4 to define d. Because Gaussian quad-
rature is only approximate, the calculated MC
7)
may be slightly in
error. This can be corrected by (slightly) modifying the value of d,
and repeating Steps 5 and 6 until a satisfactory match is achieved.
When characterizing multiple samples simultaneously, the values
of M
N
, h, and b
*
must be the same for all samples. Individual sample
values of M
C
7)
and a can, however, be different. The result of this
characterization is one set of molecular weights for the C
7)
frac-
tions, while each sample has different mole fractions z
i
(so that their
average molecular weights MC
7)
are honored).
Specific gravities for the C
7)
fractions can be calculated with
one of the correlations given in Sec. 5.4 (e.g., Eq. 5.44), where the
characterization factor (e.g., F
c
) must be the same for all mixtures.
The specific gravities, g
C
7)
, of each sample will not be exactly re-
produced with this procedure (calculated with Eq. 5.37), but the av-
erage characterization factor can be chosen so that the differences
are very small ( g"0.0005). Having defined M
i
and g
i
for the C
7)
fractions, a complete fluid characterization can be determined with
correlations in Sec. 5.5.
5.4 InspectionProperties Estimation
5.4.1 Generalized Properties. The molecular weight, specific grav-
ity, and boiling point of C
7)
fractions must be estimated in the ab-
sence of experimental TBP data. This situation arises when simulated
distillation is used or when no experimental analysis of C
7)
is avail-
able and a synthetic split must be made by use of a molar-distribution
model. For either situation, inspection data from TBP analysis of a
sample from the same field would be the most reliable source of M,
g, and T
b
for each C
7)
fraction. The next-best source would be mea-
sured TBP data from a field producing similar oil or condensate from
the same geological formation. Generalized properties from a pro-
ducing region, such as the North Sea, have been proposed.
31
Katz and Firoozabadi
6
suggest a generalized set of SCN proper-
ties for petroleum fractions C
6
through C
45
. Table 5.2 gives an ex-
tended version of the Katz-Firoozabadi property table. Molecular
weights can be used to convert weight fractions, w
i
, from simulated
distillation to mole fractions,
z
i
+
w
i
M
i
N
j+7
w
j
M
j
. (5.32) . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
However, the molecular weight of the heaviest fraction, C
N
, is not
known. From a mass balance, M
N
is given by
M
N
+
w
N
w
C
7)
M
C
7)
*
N*1
i+7
w
i
M
i
, (5.33) . . . . . . . . . . . .
where M
i
for i+7,, N*1 are taken from Table 5.2. Unfortunately,
the calculated molecular weight M
N
is often unrealistic because of
measurement errors in MC
7)
or in the chromatographic analysis and
because generalized molecular weights are only approximate. Both
w
N
and MC
7)
can be adjusted to give a reasonable M
N
, but caution
is required to avoid nonphysical adjustments. The same problem is
inherent with backcalculating M
N
with any set of generalized molec-
ular weights used for SCN Fractions 7 to N*1 (e.g., paraffin values).
During the remainder of this section, molecular weights and mole
fractions are assumed to be known for C
7)
fractions, either from
chromatographic analysis or from a synthetic split. The generalized
properties for specific gravity and boiling point can be assigned to
SCN fractions, but the heaviest specific gravity must be backcalcu-
lated to match the measured C
7)
specific gravity. The calculated g
N
also may be unrealistic, requiring some adjustment to generalized
specific gravities. Finally, the boiling point of the heaviest fraction
must be estimated. T
bN
can be estimated from a correlation relating
boiling point to specific gravity and molecular weight.
5.4.2 Characterization Factors. Inspection properties M, g, and T
b
reflect the chemical makeup of petroleum fractions. Some methods
for estimating specific gravity and boiling point assume that a par-
ticular characterization factor is constant for all C
7)
fractions.
These methods are only approximate but are widely used.
Watson or Universal Oil Products (UOP) Characterization Fac-
tor. The Watson or UOP factor, K
w
, is based on normal boiling point,
T
b
, in R and specific gravity, g.
32,33
K
w
5
T
13
b
g
. (5.34) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
K
w
varies roughly from 8.5 to 13.5. For paraffinic compounds,
K
w
+12.5 to 13.5; for naphthenic compounds, K
w
+11.0 to 12.5;
and for aromatic compounds, K
w
+8.5 to 11.0. Some overlap in K
w
exists among these three families of hydrocarbons, and a combina-
tion of paraffins and aromatics will obviously appear naphthenic.
However, the utility of this and other characterization factors is that
they give a qualitative measure of the composition of a petroleum
fraction. The Watson characterization factor has been found to be
useful for approximate characterization and is widely used as a pa-
rameter for correlating petroleum-fraction properties, such as mo-
lecular weight, viscosity, vapor pressure, and critical properties.
An approximate relation
2
for the Watson factor, based on molecu-
lar weight and specific gravity, is
K
w
[4.5579 M
0.15178
g
*0.84573
. (5.35) . . . . . . . . . . . . . . . . . .
This relation is derived from the Riazi-Daubert
14
correlation for
molecular weight and is generally valid for petroleum fractions with
normal boiling points ranging from 560 to 1,310R (C
7
through
C
30
). Experience has shown, however, that Eq. 5.35 is not very ac-
curate for fractions heavier than C
20
.
K
w
calculated with MC
7)
and g
C
7)
in Eq. 5.35 is often constant
for a given field. Figs. 5.13A and 5.13B
7
plot molecular weight vs.
specific gravity for C
7)
fractions from two North Sea fields. Data
for the gas condensate in Fig. 5.13A indicate an average
KwC
7)
+11.99"0.01 for a range of molecular weights from 135 to
150. The volatile oil shown in Fig. 5.13B has an average
KwC
7)
+11.90"0.01 for a range of molecular weights from 220 to
Fig. 5.13ASpecific gravity vs. molecular weight for C
7)
frac-
tions for a North Sea Gas-Condensate Field 2 (after Austad et al.
7
).
Molecular Weight, M
C
7+
255. The high degree of correlation for these two fields suggests ac-
curate molecular-weight measurements by the laboratory. In gener-
al, the spread in KwC
7)
values will exceed "0.01 when measure-
ments are performed by a commercial laboratory.
When the characterization factor for a field can be determined,
Eq. 5.35 is useful for checking the consistency of C
7)
molecular-
weight and specific-gravity measurements. Significant deviation in
K
wC
7)
, such as "0.03 for the North Sea fields above, indicates pos-
sible error in the measured data. Because molecular weight is more
prone to error than determination of specific gravity, an anomalous
K
wC
7)
usually indicates an erroneous molecular-weight measure-
ment. For the gas condensate in Fig. 5.13A, a C
7)
sample with spe-
cific gravity of 0.775 would be expected to have a molecular weight
of [141 (for K
wC
7)
+11.99). If the measured value was 135, the
Watson characterization factor would be 11.90, which is significant-
ly lower than the field average of 11.99. In this case, the C
7)
molec-
ular weight should be redetermined.
Eq. 5.35 can also be used to calculate specific gravity of C
7)
frac-
tions determined by simulated distillation or a synthetic split (i.e.,
when only mole fractions and molecular weights are known). As-
suming a constant K
w
for each fraction, specific gravity, g
i
, can be
calculated from
g
i
+6.0108 M
0.17947
i
K
*1.18241
w
. (5.36) . . . . . . . . . . . . . . . . .
K
w
must be chosen so that experimentally measured C
7)
specific
gravity, (g
C
7)
)
exp
, is calculated correctly.
g
C
7)
exp
+
z
C
7)
M
C
7)
N
i+1
z
i
M
i
g
i
. (5.37) . . . . . . . . . . . . . . . . . . . . .
The Watson factor satisfying Eq. 5.37 is given by
K
w
+
0.16637g
C
7)
A
0
z
C
7)
M
C
7)
*0.84573
, (5.38) . . . . . . . . . . . . . . .
where A
0
+
N
i+1
z
i
M
0.82053
i
. (5.39) . . . . . . . . . . . . . . . . . . . . . .
Fig. 5.13BSpecific gravity vs. molecular weight for C
7)
frac-
tions for a North Sea Volatile-Oil Field 3B(after Austad et al.
7
).
Molecular Weight, M
C
7+
Fig. 5.14Specific gravity vs. molecular weight for constant val-
ues of the Jacoby aromaticity factor (solid lines) and the Watson
characterization factor (dashed lines). After Whitson.
25
J
a
K
w
Jacoby Correlation
(Aromaticity Factor, J
a
)
Present Correlation
(Watson Factor, K
w
)
Boiling points, T
bi
, can be estimated from Eq. 5.36.
T
bi
+(K
w
g
i
)
3
. (5.40) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Unfortunately, Eqs. 5.36 through 5.40 overpredict g and T
b
at mo-
lecular weights greater than [250 (an original limitation of the
Riazi-Daubert
14
molecular-weight correlation).
Jacoby Aromaticity Factor. The Jacoby aromaticity factor, J
a
, is
an alternative characterization factor for describing the relative
composition of petroleum fractions.
34
Fig. 5.14
2
shows the original
Jacoby relation between specific gravity and molecular weight for
several values of J
a
. The behavior of specific gravity as a function
of molecular weight is similar for the Jacoby factor and the relation
for a constant K
w
. However, specific gravity calculated with the
Jacoby method increases more rapidly at low molecular weights,
flattening at high molecular weights (a more physically consistent
behavior). A relation for the Jacoby factor is
J
a
+
g *0.8468 )
15.8M
0.2456 *
1.77M
(5.41) . . . . . . . . . . . . . . . . . .
12 PHASE BEHAVIOR
Fig. 5.15Specific gravity vs. carbon number for constant val-
ues of the Yarborough aromaticity factor (after Yarborough
1
).
or, in terms of specific gravity,
g +0.8468 *
15.8
M
) J
a
0.2456 *
1.77
M
. (5.42) . . . . . .
The first two terms in Eq. 5.42 (i.e., when J
a
+0) express the relation
between specific gravity and molecular weight for normal paraffins.
The Jacoby factor can also be used to estimate fraction specific
gravities when mole fractions and molecular weights are available
from simulated distillation or a synthetic split. The Jacoby factor
satisfying measured C
7)
specific gravity (Eq. 5.37) must be calcu-
lated by trial and error. We have found that this relation is particular-
ly accurate for gas-condensate systems.
27
Yarborough Aromaticity Factor. Yarborough
1
modified the
Jacoby aromaticity factor specifically for estimating specific gravi-
ties when mole fractions and molecular weights are known. Yarbo-
rough tries to improve the original Jacoby relation by reflecting the
changing character of fractions up to C
13
better and by representing
the larger naphthenic content of heavier fractions better. Fig. 5.15
shows how the Yarborough aromaticity factor, Y
a
, is related to spe-
cific gravity and carbon number. A simple relation representing Y
a
is not available; however, Whitson
26
has fit the seven aromaticity
curves originally presented by Yarborough using the equation
g
i
+exp
A
0
)A
1
i
*1
)A
2
i )A
3
ln(i)
, (5.43) . . . . . . . . . .
where i+carbon number. Table 5.7 gives the constants for Eq. 5.43.
The aromaticity factor required to satisfy measured C
7)
specific
gravity (Eq. 5.37) is determined by trial and error. Linear interpola-
tion of specific gravity should be used to calculate specific gravity
for a Y
a
value falling between two values of Y
a
in Table 5.7.
Sreide
35
Correlations. Sreide developed an accurate specific-
gravity correlation based on the analysis of 843 TBP fractions from
68 reservoir C
7)
samples.
g
i
+0.2855 ) C
f
(M
i
*66)
0.13
. (5.44) . . . . . . . . . . . . . . .
C
f
typically has a value between 0.27 and 0.31 and is determined for
a specific C
7)
sample by satisfying Eq. 5.37.
5.4.3 Boiling-Point Estimation. Boiling point can be estimated
from molecular weight and specific gravity with one of several cor-
relations. Sreide also developed a boiling-point correlation based
on 843 TBP fractions from 68 reservoir C
7)
samples,
T
b
+1928.3 *
1.695 10
5
M
*0.03522
g
3.266
exp
4.922 10
*3
M*4.7685g
)
3.462 10
*3
Mg
, (5.45) . . . . . . . . . . . . . . . . . . . . .
with T
b
in R.
Table 5.8 gives estimated specific gravities determined with the
methods just described for a C
7)
sample with the exponential split
given in Table 5.4 (a+1, h+90, M
C
7)
+200) and g
C
7)
+0.832.
The following equations also relate molecular weight to boiling
point and specific gravity; any of these correlations can be solved
for boiling point in terms of M and g. We recommend, however, the
Sreide correlation for estimating T
b
from M and g.
Kesler and Lee.
12
M +
*12, 272.6 )9, 486.4g )(4.6523 *3.3287g)T

b
)

1 *0.77084g *0.02058g
2
1.3437 *720.79T
*1
b
10
7
T
*1
b
1 *0.80882g )0.02226g
2
1.8828 *181.98T
1
b
10
12
T
*3
b
. (5.46) . . . . . . . .
Riazi and Daubert.
14
M +(4.5673 10
*5
)T
2.1962
b
g
*1.0164
. (5.47) . . . . . . . . . . . .
American Petroleum Inst. (API).
36
M +
2.0438 10
2
T
0.118
b
g
1.88
exp
0.00218T
b
*3.07g
.
(5.48) . . . . . . . . . . . . . . . . . . . .
Rao and Bardon.
37
ln M +(1.27 )0.071K
w
) ln
1.8T
b
22.31 )1.68K
w
.
(5.49) . . . . . . . . . . . . . . . . . . . .
Riazi and Daubert.
18
M +581.96T
0.97476
b
g
6.51274
exp
5.43076 10
*3
T
b
*9.53384g )
1.11056 10
*3
T
b
g
. (5.50) . . . . . . . . .
TABLE 5.7COEFFICIENTS FOR YARBOROUGH AROMATICITY FACTOR CORRELATION
1,26
Y
a
A
0
A
1
A
2
A
2
0.0 *7.4385510
*2
*1.72341 1.3805810
*3
*3.3416910
*2
0.1 *4.2580010
*1
*7.0001710
*1
*3.3094710
*5
8.6546510
*2
0.2 *4.4755310
*1
*7.6511110
*1
1.7798210
*4
1.0774610
*1
0.3 *4.3910510
*1
*9.4406810
*1
4.9370810
*4
1.1926710
*1
0.4 *2.7371910
*1
*1.39960 3.8056410
*3
5.9200510
*2
0.6 *7.3941210
*3
*1.97063 5.8727310
*3
*1.6714110
*2
0.8 *3.1761810
*1
*7.7843210
*1
2.5861610
*3
1.0838210
*3
TABLE 5.8COMPARISON OF SPECIFIC GRAVITIES WITH CORRELATIONS BY USE OF
DIFFERENT CHARACTERIZATION FACTORS
g
C
7)
+0.832
g
i
for Different Correlations With Constant Characterization
Factor Chosen To Match
Fraction z
i
M
i
K
w
+12.080 J
a
+0.2395 Y
a
+0.2794 C
f
+0.2864
1 0.1195 96.8 0.7177 0.7472 0.7051 0.7327
2 0.1052 110.8 0.7353 0.7684 0.7286 0.7550
3 0.0926 124.8 0.7511 0.7849 0.7486 0.7719
4 0.0816 138.8 0.7656 0.7981 0.7660 0.7856
5 0.0718 152.8 0.7789 0.8088 0.7813 0.7972
6 0.0632 166.8 0.7913 0.8178 0.7951 0.8072
7 0.0557 180.8 0.8028 0.8253 0.8075 0.8161
8 0.0490 194.8 0.8136 0.8318 0.8189 0.8241
9 0.0432 208.8 0.8238 0.8374 0.8294 0.8314
10 0.0380 222.8 0.8335 0.8423 0.8391 0.8380
11 0.0335 236.8 0.8426 0.8466 0.8482 0.8442
12 0.0295 250.8 0.8514 0.8505 0.8567 0.8500
13 0.0259 264.8 0.8597 0.8539 0.8646 0.8554
14 0.0228 278.8 0.8677 0.8570 0.8722 0.8604
15 0.0201 292.8 0.8753 0.8598 0.8793 0.8652
16 0.0177 306.8 0.8827 0.8623 0.8861 0.8697
17 0.0156 320.8 0.8898 0.8646 0.8926 0.8740
18 0.0137 334.8 0.8966 0.8668 0.8988 0.8782
19 0.0121 348.8 0.9033 0.8687 0.9048 0.8821
20 0.0891 466.0 0.9514 0.8805 0.9468 0.9096
Total 1.0000 200.0 0.8320 0.8320 0.8320 0.8320
5.5 CriticalProperties Estimation
Thus far, we have discussed how to split the C
7)
fraction into
pseudocomponents described by mole fraction, molecular weight,
specific gravity, and boiling point. Now we must consider the prob-
lem of assigning critical properties to each pseudocomponent. Criti-
cal temperature, T
c
; critical pressure, p
c
; and acentric factor, w, of
each component in a mixture are required by most cubic EOSs.
Critical volume, v
c
, is used instead of critical pressure in the Bene-
dict-Webb-Rubin
38
(BWR) EOS, and critical molar volume is
used with the LBC viscosity correlation.
24
Critical compressibility
factor has been introduced as a parameter in three- and four-constant
cubic EOSs.
Critical-property estimation of petroleum fractions has a long his-
tory beginning as early as the 1930s; several reviews
22,25,26,39,40
are available. We present the most commonly used correlations and
a graphical comparison (Figs. 5.16 through 5.18) that is intended
to highlight differences between the correlations. Finally, correla-
tions based on perturbation expansion (a concept borrowed from
statistical mechanics) are discussed separately.
The units for the remaining equations in this section are T
b
in R,
T
bF
in F+T
b
*459.67, T
c
in R, p
c
in psia, and v
c
in ft
3
/lbm mol.
Oil gravity is denoted g
API
and is related to specific gravity by
g
API
+141.5/g*131.5.
5.5.1 Critical Temperature. T
c
is perhaps the most reliably corre-
lated critical property for petroleum fractions. The following criti-
cal-temperature correlations can be used for petroleum fractions.
Roess.
41
(modified by API
36
).
T
c
+645.83 )1.6667
T
bF
)100
0.7127 10
*3
T
bF
)100
2
. (5.51) . . . . . . . . . . .
Kesler-Lee.
12
T
c
+341.7 )811g )(0.4244 )0.1174g)T
b
)(0.4669 *3.2623g) 10
5
T
*1
b
. (5.52) . . . . . . . . . . . .
Cavett.
42
T
c
+768.07121 )1.7133693T
bF
*
0.10834003 10
*2
T
2
bF
*
0.89212579 10
*2
g
API
T
bF
)
0.38890584 10
*6
T
3
bF
)
0.5309492 10
*5
g
API
T
2
bF
)
0.327116 10
*7
g
2
API
T
2
bF
. (5.53) . . . . . . . . . . . . . . .
Riazi-Daubert.
14
T
c
+24.27871T
0.58848
b
g
0.3596
. (5.54) . . . . . . . . . . . . . . . . . .
Nokay.
43
T
c
+19.078T
0.62164
b
g
0.2985
. (5.55) . . . . . . . . . . . . . . . . . . . .
5.5.2 Critical Pressure. p
c
correlations are less reliable than T
c
cor-
relations. The following are p
c
correlations that can be used for pe-
troleum fractions.
Kesler-Lee.
12
ln p
c
+8.3634 *
0.0566
g
*0.24244 )
2.2898
g
)
0.11857
g
2
10
*3
T
b
14 PHASE BEHAVIOR
Fig. 5.16Comparison of critical-temperature correlations for
boiling points from 600 to 1,500R assuming a constant Watson
characterization factor of 12.
)1.4685 )
3.648
g
)
0.47227
g
2
10
*7
T
2
b
*0.42019 )
1.6977
g
2
10
*10
T
3
b
. (5.56) . . . . .
Cavett.
42
log p
c
+2.8290406 )
0.94120109 10
*3
T
bF
*
0.30474749 10
*5
T
2
bF
*
0.2087611 10
*4
g
API
T
bF
)
0.15184103 10
*8
T
3
bF
)
0.11047899 10
*7
g
API
T
2
bF
*
0.48271599 10
*7
g
2
API
T
bF
)
0.13949619 10
*9
g
2
API
T
2
bF
. (5.57) . . . . . . . . . . .
Riazi-Daubert.
14
p
c
+
3.12281 10
9
T
*2.3125
b
g
2.3201
. (5.58) . . . . . . . . . . . . .
5.5.3 Acentric Factor. Pitzer et al.
44
defined acentric factor as
w 5*log
p
*
v
p
c
*1, (5.59) . . . . . . . . . . . . . . . . . . . . . . . .
where p
*
v
+vapor pressure at temperature T+0.7T
c
(T
r
+0.7).
Practically, acentric factor gives a measure of the steepness of the
vapor-pressure curve from T
r
+0.7 to T
r
+1, where p
*
v
/p
c
+0.1 for
w+0 and p
*
v
/p
c
+0.01 for w+1. Numerically, w[0.01 for meth-
ane, [0.25 for C
5
, and [0.5 for C
8
(see Table A.1 for literature val-
ues of acentric factor for pure compounds). w increases to u1.0 for
petroleum fractions heavier than approximately C
25
(see Table 5.2).
The Kesler-Lee
12
acentric factor correlation (for T
b
/T
c
u0.8) is
developed specifically for petroleum fractions, whereas the correla-
tion for T
b
/T
c
t0.8 is based on an accurate vapor-pressure correla-
tion for pure compounds. The Edmister
45
correlation is limited to
pure hydrocarbons and should not be used for C
7)
fractions. The
three correlations follow.
Fig. 5.17Comparison of critical-pressure correlations for boil-
ing points from 600 to 1,500R assuming a constant Watson
characterization factor of 12.
Fig. 5.18Comparison of acentric factor correlations for boiling
points from 600 to 1500R assuming a constant Watson charac-
terization factor of 12.
Lee-Kesler.
13
(T
br
+T
b
/T
c
t0.8).
w +
ln
p
c
14.7
)A
1
)A
2
T
*1
br
)A
3
ln T
br
)A
4
T
6
br
A
5
)A
6
T
*1
br
)A
7
ln T
br
)A
8
T
6
br
,
(5.60) . . . . . . . . . . . . . . . . . . . .
where A
1
+*5.92714, A
2
+ 6.09648, A
3
+ 1.28862, A
4
+
*0.169347, A
5
+ 15.2518, A
6
+*15.6875, A
7
+*13.4721,
and A
8
+ 0.43577.
Kesler-Lee.
12
(T
br
+T
b
/T
c
u0.8).
w +*7.904 )0.1352K
w
*0.007465K
2
w
)8.359T
br
)(1.408 *0.01063K
w
)T
*1
br
. (5.61) . . . . . . .
Edmister.
45
w +
3
7
log
p
c
14.7
T
c
T
b
*1
* 1. (5.62) . . . . . . . . . . . . . . . . . . . . .
5.5.4 Critical Volume. The Hall-Yarborough
46
critical-volume
correlation is given in terms of molecular weight and specific grav-
ity, whereas the Riazi-Daubert
14
correlation uses normal boiling
point and specific gravity.
Hall-Yarborough.
46
v
c
+0.025M
1.15
g
*0.7935
. (5.63) . . . . . . . . . . . . . . . . . . . . . .
Riazi-Daubert.
14
v
c
+
7.0434 10
*7
T
2.3829
b
g
*1.683
. (5.64) . . . . . . . . . . . . .
Critical compressibility factor, Z
c
, is defined as
Z
c
+
p
c
v
c
RT
c
, (5.65) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
where R+universal gas constant. Thus, Z
c
can be calculated directly
from critical pressure, critical volume, and critical temperature. Reid
et al.
40
and Pitzer et al.
44
give an approximate relation for Z
c
.
Z
c
[0.291 *0.08w. (5.66) . . . . . . . . . . . . . . . . . . . . . . . . .
Eq. 5.66 is not particularly accurate (grossly overestimating Z
c
for
heavier compounds) and is used only for approximate calculations.
5.5.5 Correlations Based on Perturbation Expansions. Correla-
tions for critical temperature, critical pressure, critical volume, and
molecular weight have been developed for petroleum fractions with
a perturbation-expansion model with normal paraffins as the refer-
ence system. To calculate critical pressure, for example, critical
temperature, critical volume, and specific gravity of a paraffin with
the same boiling point as the petroleum fraction must be calculated
first. Kesler et al.
47
first used the perturbation expansion (with n-al-
kanes as the reference fluid) to develop a suite of critical-property
and acentric-factor correlations.
Twu
48
uses the same approach to develop a suite of critical-prop-
erty correlations. We give his normal-paraffin correlations first,
then the correlations for petroleum fractions.
Normal Paraffins (Alkanes).
T
cP
+T
b
0.533272 )
0.191017 10
*3
T
b
)
0.779681 10
*7
T
2
b
*
0.284376 10
*10
T
3
b
)
(0.959468 10
2
)
0.01T
b
13
*1
, (5.67) . . . . . . . . . . . . . . . . . .
p
cP
+(3.83354 )1.19629a
0.5
)34.8888a
)36.1952a
2
)104.193a
4
)
2
, (5.68) . . . . . . . . . . . . . . .
v
cP
+[ 1 *(0.419869 *0.505839a *1.56436a
3
*9481.7a
14
)]
*8
, (5.69) . . . . . . . . . . . . . . . . . . . . . . . . .
g
P
+0.843593 *0.128624a *3.36159a
3
*13749.5a
12
, (5.70) . . . . . . . . . . . . . . . . . . . . . . . . . . . .
and T
b
+exp(5.71419 )2.71579q *0.28659q
2
*39.8544q
*1
*0.122488q
*2
)
*24.7522q )35.3155q
2
, (5.71) . . . . . . . . . . . . . . . . .
where a +1 *
T
b
T
cP
(5.72) . . . . . . . . . . . . . . . . . . . . . . . . . . . .
and q +ln M
P
. (5.73) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Paraffin molecular weight, M
P
, is not explicitly a function of T
b
, and
Eqs. 5.67 through 5.73 must be solved iteratively; an initial guess
is given by
M
P
[
T
b
10.44 *0.0052T
b
. (5.74) . . . . . . . . . . . . . . . . . . . . .
Twu claims that the normal-paraffin correlations are valid for C
1
through C
100
, although the properties at higher carbon numbers are
only approximate because experimental data for paraffins heavier
than approximately C
20
do not exist. The following relations are
used to calculate petroleum-fraction properties.
Critical Temperature.
T
c
+T
cP
1 )2f
T
1 *2f
T
2
,
f
T
+Dg
T
*0.362456
T
0.5
b
)0.0398285 *
0.948125
T
0.5
b
Dg
T
,
and Dg
T
+exp[5(g
P
*g)] *1. (5.75) . . . . . . . . . . . . . . . . . .
Critical Volume.
v
c
+v
cP
1 )2f
v
1 *2f
v
2
,
f
v
+Dg
v
0.466590
T
0 .5
b
)*0.182421 )
3.01721
T
0.5
b
Dg
v,
and Dg
v
+exp
g
2
P
*g
2
*1. (5.76) . . . . . . . . . . . . . . . . . .
Critical Pressure.
p
c
+p
cP
T
c
T
cP
V
cP
V
c
1 )2f
p
1 *2f
p
2
,
f
p
+Dg
p 2.53262 *
46.1955
T
0.5
b
*0.00127885T
b
) *11.4277 )
252.14
T
0.5
b
)0.00230535T
b
Dg
p,
and Dg
p
+exp[0.5(g
P
*g)] *1. (5.77) . . . . . . . . . . . . . . . . .
Molecular Weight.
ln M +ln M
P
1 )2f
M
1 *2f
M
2
,
f
M
+Dg
M
|x| )*0.0175691 )
0.193168
T
0.5
b
Dg
M
,
x +0.012342 *
0.328086
T
0.5
b
,
and Dg
M
+exp[5(g
P
*g)]. (5.78) . . . . . . . . . . . . . . . . . . . . .
Figs. 5.16 through 5.18 compare the various critical-property cor-
relations for a range of boiling points from 600 to 1,500R.
5.5.6 Methods Based on an EOS. Fig. 5.19
28
illustrates the impor-
tant influence that critical properties have on EOS-calculated proper-
ties of pure components. Vapor pressure is particularly sensitive to
critical temperature. For example, the Riazi-Daubert
19
critical-tem-
perature correlation for toluene overpredicts the experimental value
16 PHASE BEHAVIOR
Fig. 5.19Effect of critical temperature on vapor-pressure pre-
diction of toluene with the PR EOS; AAD+absolute average devi-
ation (after Brul et al.
28
).
T
c
underpredicted T
c
overpredicted
Deviation From Experimental Value, %
by only 1.7%. Even with this slight error in T
c
, the average error in
vapor pressures predicted by the Peng-Robinson
49
(PR) EOS is 16%.
The effect of critical properties and acentric factor on EOS calcula-
tions for reservoir-fluid mixtures is summarized by Whitson.
26
In principle, the EOS used for mixtures should also predict the be-
havior of individual components found in the mixture. For pure
compounds, the vapor pressure is accurately predicted because all
EOSs force fit vapor-pressure data. Some EOSs are also fit to satu-
rated-liquid densities at subcritical temperatures. The measured
properties of petroleum fractions, boiling point, and specific gravity
can also be fit by the EOS, as discussed later.
For each petroleum fraction separately, two of the EOS parame-
ters (T
c
; p
c
; w; volume-shift factor, s; or multipliers of EOS constants
A and B) can be chosen so that the EOS exactly reproduces exper-
imental boiling point and specific gravity. Because only two inspec-
tion properties are available (T
b
and g), only two of the EOS parame-
ters can be determined. Whitson
50
suggests fixing the value of w
with an empirical correlation and adjusting T
c
and p
c
to match nor-
mal boiling point and molar volume (M/g) at standard conditions.
Critical properties satisfying these criteria are given for a wide range
of petroleum fractions by the PR EOS and the Soave-Redlich-
Kwong (SRK) EOS.
22,23
A better (and recommended) approach for
cubic EOSs is to use the volume-shift factor s (see Chap. 4) to match
specific gravity or a saturated liquid density and acentric factor to
match normal boiling point.
Other methods for forcing the EOS to match boiling point and
specific gravity have also been devised. Brul and Starling
51
pro-
posed a method that uses viscosity as an additional inspection prop-
erty of the fraction for determining critical properties. This ap-
proach proved particularly successful when applied to the BWR
EOS for residual-oil supercritical extraction (ROSE).
28
5.6 Recommended C
7)
Characterizations
We recommend the following C
7)
characterization procedure for
cubic EOSs.
1. Use the Twu
48
(or Lee-Kesler
12
) critical property correlation
for T
c
and p
c
.
2. Choose the acentric factor to match T
b
; alternatively, use the
Lee-Kesler
12
/Kesler-Lee
13
correlations.
3. Determine volume-translation coefficients, s
i
, to match specific
gravities; alternatively, use Peneloux et al.s
52
correlation for the SRK
EOS
22,23
or Jhaveri and Youngrens
53
correlation for the PR EOS.
49
When measured TBP data are not available, a mathematical split
should be made with either (1) the gamma distribution (default
a+1, h+90) with Gaussian-quadrature or equal-mass fractions or
(2) the exponential distribution (Eq. 5.7). Specific gravities should
be estimated with the Sreide
35
correlation (Eq. 5.44), choosing C
f
to match measured C
7)
specific gravity (Eq. 5.37). Boiling points
should be estimated from the Sreide correlation (Eq. 5.45).
For the PR EOS, we recommend the nonhydrocarbon BIPs given
in Chap. 4 and the modified Chueh-Prausnitz
54
equation for C
1
through C
7)
pairs,
k
ij
+A
1 *
2v
16
ci
v
16
cj
v
13
ci
)v
13
cj
, (5.79) . . . . . . . . . . . . . . . .
with A+0.18 and B+6.
5.6.1 SRK-Recommended Characterization. Alternatively, the
Pedersen et al.
55
characterization procedure can be used with the
SRK EOS.
1. Split the plus fraction C
n)
(preferably nu10) into SCN frac-
tions up to C
80
using Eqs. 5.7 through 5.11 and h+*4.
2. Calculate SCN densities
i
(g
i
+
i
/0.999) using the equation
i
+A
0
)A
1
ln(i), where A
0
and A
1
are determined by satisfying the
experimental-plus density,
n)
, and measured (or assumed) densi-
ty,
n*1
(
6
+0.690 can be used for C
7)
).
3. Calculate critical properties of all C
7)
fractions (distillation
cuts from C
7
to C
n*1
and split SCN fractions from C
n
through C
80
)
using the correlations
T
c
+163.12 )86.052 ln M)0.43475M*
1877.4
M
,
ln p
c
+*0.13408 )2.5019 )
208.46
M
*
3987.2
M
2
,
and m
SRK
+0.48 )1.574w *0.176w
2
+0.7431 )0.0048122M)0.0096707
*
3.7184 10
*6
M
2
. (5.80) . . . . . . . . . . . . . . . . . .
Note that the use of acentric factor is circumvented by directly calcu-
lating the term m used in the a correction term to EOS Constant A.
4. Group C
7)
into 3 to 12 fractions using equal-weight fractions
in each group; use weight-average mixing rules.
5. Calculate volume-translation parameters for C
7)
fractions to
match specific gravities; pure component c values are taken from
Peneloux et al.
52
6. All hydrocarbon/hydrocarbon BIPs are set to zero. SRK BIPs
given in Chap. 4 are used for nonhydrocarbon/hydrocarbon pairs.
The two recommended C
7)
characterization procedures out-
lined previously for the PR EOS and SRK EOS are probably the best
currently available (other EOS characterizations, such as the Re-
dlich-Kwong EOS modified by Zudkevitch and Joffe,
56
and some
three-constant characterizations should provide similar accuracy
but are not significantly better). Practically, the two characterization
procedures give the same results for almost all PVT properties (usu-
ally within 1 to 2%). With these EOS-characterization procedures,
we can expect reasonable predictions of densities and Z factors ("1
to 5%), saturation pressures ("5 to 15%), gas/oil ratios and forma-
tion volume factors ("2 to 5%), and condensate-liquid dropout
("5 to 10% for maximum dropout, with poorer prediction of tail-
like behavior just below the dewpoint).
The recommended EOS methods are less reliable for prediction
of minimum miscibility conditions, near-critical saturation pressure
and saturation type (bubblepoint or dewpoint), and both retrograde
and near-critical liquid volumes. Improved predictions can be ob-
tained only by tuning EOS parameters to accurate PVT data cover-
ing a relatively wide range of pressures, temperatures, and composi-
tions (see Sec. 4.7 and Appendix C).
5.7 Grouping and Averaging Properties
The cost and computer resources required for compositional reser-
voir simulation increase substantially with the number of compo-
nents used to describe the reservoir fluid. A compromise between
accuracy and the number of components must be made according
to the process being simulated (i.e., according to the expected effect
that phase behavior will have on simulated results). For example, a
detailed fluid description with 12 to 15 components may be needed
to simulate developed miscibility in a slim-tube experiment. With
current computer technology, however, a full-field simulation with
fluids exhibiting near-critical phase behavior is not feasible for a
15-component mixture. The following are the main questions re-
garding component grouping.
1. How many components should be used?
2. How should the components be chosen from the original fluid
description?
3. How should the properties of pseudocomponents be calculated?
5.7.1 How Many and Which Components To Group. The number
of components used to describe a reservoir fluid depends mainly on
the process being simulated. However, the following rule of thumb
reduces the number of components for most systems: group N
2
with
methane, CO
2
with ethane, iso-butane with n-butane, and iso-pen-
tane with n-pentane. Nonhydrocarbon content should be less than
a few percent in both the reservoir fluid and the injection gas if a
nonhydrocarbon is to be grouped with a hydrocarbon.
Five- to eight-component fluid characterizations should be suffi-
cient to simulate practically any reservoir process, including (1) reser-
voir depletion of volatile-oil and gas-condensate reservoirs, (2) gas
cycling above and below the dewpoint of a gas-condensate reservoir,
(3) retrograde condensation near the wellbore of a producing well,
and (4) immiscible and miscible gas-injection. Coats
57
discusses a
method for combining a modified black-oil formula with a simplified
EOS representation of separator oil and gas streams. The oil and
gas pseudocomponents in this model contain all the original fluid
components in contrast to the typical method of grouping where each
pseudocomponent is made up of only selected original components.
Lee et al.
58
suggest that C
7)
fractions can be grouped into two
pseudocomponents according to a characterization factor deter-
mined by averaging the tangents of fraction properties M, g, and J
a
plotted vs. boiling point.
Whitson
2
suggests that the C
7)
fraction can be grouped into N
H
pseudocomponents given by
N
H
+1 )3.3 log(N *7), (5.81) . . . . . . . . . . . . . . . . . . . . .
where N+carbon number of the heaviest fraction in the original
fluid description. The groups are separated by molecular weights M
I
given by
M
I
+M
C
7
M
N
M
C
7
1N
H
, (5.82) . . . . . . . . . . . . . . . . . . .
where I+1,..., N
H
. Molecular weights, M
i
, from the original fluid
description (i+7,..., N) falling within boundaries M
I*1
to M
I
are in-
cluded in Group I. This method should only be used when C
7)
frac-
tions are originally separated on a carbon-number basis and for N
greater than [20.
Li et al.
59
suggest a method for grouping components of an origi-
nal fluid description that uses K values from a flash at reservoir tem-
perature and the average operating pressure. The original mixture
is divided arbitrarily into light components (H
2
S, N
2
, CO
2
, and C
1
through C
6
) and heavy components (C
7)
). Different criteria are
used to determine the number of light and heavy pseudocompon-
ents. Li et al. also suggest use of phase diagrams and compositional
simulation to verify the grouped fluid description (a practice that we
highly recommend).
Still other pseudoization methods have been proposed
60,61
; Schlij-
pers
61
method also treats the problem of retrieving detailed composi-
tional information from pseudoized (grouped) components. Behrens
and Sandler
62
suggest a grouping method for C
7)
fractions based
on application of the Gaussian-quadrature method to continuous
thermodynamics. Although a simple exponential distribution is
used with only two quadrature points (i.e., the C
7)
fractions are
grouped into two pseudocomponents), Whitson et al.
27
show that
the method is general and can be applied to any molar-distribution
model and for any number of C
7)
groups.
In general, most authors have found that broader grouping of C
7)
as C
7
through C
10
, C
11
through C
15
, C
16
through C
20
, and C
21)
is
substantially better than splitting only the first few carbon-number
fractions (e.g., C
7
, C
8
, C
9
, and C
10)
). Gaussian quadrature is recom-
mended for choosing the pseudocomponents in a C
7)
fraction;
equal-mass fractions or the Li et al.
59
approach are valid alternatives.
5.7.2 Mixing Rules. Several methods have been proposed for calcu-
lating critical properties of pseudocomponents. The simplest and
most common mixing rule is
q
I
+
iI
z
i
q
i
iI
z
i
, (5.83) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
where q
i
+any property (T
c
, p
c
, w, or M) and z
i
+original mole frac-
tion for components (i+1,..., I) making up Pseudocomponent I. Av-
erage specific gravity should always be calculated with the assump-
tion of ideal solution mixing.
g
I
+
iI
z
i
M
i
iI
z
i
M
i
g
i
. (5.84) . . . . . . . . . . . . . . . . . . . . . . . . . . .
Pedersen et al.
55
and others suggest use of weight fraction instead
of mole fraction. Wu and Batyckys
63
empirical mixing-rule ap-
proach uses both the molar- and weight-average mixing rules and
a proportioning factor, F, to calculate p
cI
, T
cI
, and w
I
.
q
I
+
iI
f
i
q
i
, (5.85) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
where q
I
represents p
cI
, T
cI
, and w
I
and f
i
+average of the molar and
weight fractions,
f
i
+Fq
i
z
i
)(1 *F) q
i
w
i
and w
i
+
z
i
M
i
N
j+1
z
j
M
j
, (5.86) . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
with 0xFx1.
A generalized mixing rule for BIPs can be written
k
IJ
+
iI
jJ
f
i
f
j
k
ij
, (5.87) . . . . . . . . . . . . . . . . . . . . . . . . .
where f
i
is also given by Eq. 5.86.
On the basis of Chueh and Prausnitzs
54
arguments, Lee-Kesler
13
proposed the mixing rules in Eqs. 5.88 through 5.92.
v
cI
+
1
8
iI
jJ
z
i
z
j
v
13
ci
)v
13
cj
iI
z
i
2
, (5.88) . . .
T
cI
+
1
8v
cI
iI
jJ
z
i
z
j
T
ci
T
cj
12
v
13
ci
)v
13
cj
B
iI
z
i
2
, (5.89) . . . . . . . . . . . . . . . . . . . . . . . . . . . .
w
I
+
iI
z
i
w
i
iI
z
i
, (5.90) . . . . . . . . . . . . . . . . . . .
Z
cI
+0.2905 *0.085w
I
, (5.91) . . . . . . . . . . . . . . . . . . . . . .
and p
cI
+
Z
cI
RT
cI
v
cI
. (5.92) . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
18 PHASE BEHAVIOR
TABLE 5.9EXAMPLE STEPWISE-REGRESSION PROCEDURE FOR PSEUDOIZATION
TO FEWER COMPONENTS FOR A GAS CONDENSATE FLUID UNDERGOING DEPLETION
Original
Component
Number
Original
Component
Step 1 Step 2 Step 3 Step 4 Step 5
1 N
2
N
2
)C
1
* N
2
)C
1
N
2
)C
1
N
2
)C
1
)CO
2
)C
2
* N
2
)C
1
)CO
2
)C
2
2 CO
2
CO
2
)C
2
* CO
2
)C
2
CO
2
)C
2
C
3
)i-C
4
)n-C
4
)i-C
5
)n-C
5
)C
6
*
C
3
)i-C
4
)n-C
4
)i-C
5
)n-C
5
)C
6
3 C
1
C
3
C
3
C
3
)i-C
4
)n-C
4
*
F
1
F
1
4 C
2
i-C
4
i-C
4
)n-C
4
* i-C
5
)n-C
5
)C
6
*
F
2
F
2
)F
3
*
5 C
3
n-C
4
i-C
5
)n-C
5
*
F
1
F
3
6 i-C
4
i-C
5
C
6
F
2
7 n-C
4
n-C
5
F
1
F
3
8 i-C
5
C
6
F
2
9 n-C
5
F
1
F
3
10 C
6
F
2
11 F
1
F
3
12 F
2
13 F
3
Regression Parameters
k
ij
1, 9, 10, and 11 1, 7, 8, and 9 1, 5, 6, and 7 1, 3, 4, and 5 1, 3, and 4
W
a
1 4 3 1 3
W
b
1 4 3 1 3
W
a
2 5 4 2 4
W
b
2 5 4 2 4
*Indicates the grouped pseudocomponents being regressed in a particular step.
Lee et al.
58
and Whitson
2
consider an alternative method for cal-
culating C
7)
critical properties based on the specific gravities and
boiling points of grouped pseudocomponents.
Coats
57
presents a method of pseudoization that basically elimi-
nates the effect of mixing rules on pseudocomponent properties.
The approach is simple and accurate. Coats requires the pseudoized
characterization to reproduce exactly the volumetric behavior of the
original reservoir fluid at undersaturated conditions. This is
achieved by ensuring that the mixture EOS constants A and B are
identical for the original and the pseudoized characterizations. First,
pseudocritical properties ( p
cI
, T
cI
, and w
I
) are estimated with any
mixing rule (e.g., Kays
64
mixing rule). Then W
aI
and W
bI
are deter-
mined to satisfy the following equations.
W
aI
+
iI
jJ
z
i
z
j
a
i
a
j
1 *k
ij
iI
z
i
R
2
T
2
cI
p
cI
a
I
(T
rI
, w
I
)
and W
bI
+
iI
z
i
b
i
iI
z
i
RT
cI
p
cI
b
I
(T
rI
, w
I
)
, (5.93) . . . . . . . . . . . . . . . .
where a
i
+W
ai
R
2
T
2
ci
p
ci
a
i
(T
ri
, w
i
)
and b
i
+W
bi
RT
ci
p
ci
b
i
(T
ri
, w
i
) . (5.94) . . . . . . . . . . . . . . . . . . . . .
W
ai
and W
bi
may include previously determined corrections to the
numerical constants W
o
a
and W
o
b
. This approach to determining
pseudocomponent properties, together with Eq. 5.87 for k
I J
, is sur-
prisingly accurate even for VLE calculations. Coats also gives an
analogous procedure for determining pseudocomponent v
cI
for the
LBC
24
viscosity correlation.
Coats approach is preferred to all the other proposed methods. It
ensures accurate volumetric calculations that are consistent with the
original EOS characterization, and the method is easy to implement.
5.7.3 Stepwise Regression. A reduced-component characterization
should strive to reproduce the original complete characterization
that has been used to match measured PVT data. One approach to
achieve this goal is stepwise regression, summarized in the follow-
ing procedure.
1. Complete a comprehensive match of all existing PVT data with
a characterization containing light and intermediate pure compo-
nents and at least three to five C
7)
fractions.
2. Simulate a suite of depletion and multicontact gas-injection
PVT experiments that cover the expected range of compositions in
the particular application.
3. Use the simulated PVT data as real data for pseudoization
based on regression.
4. Create two new pseudocomponents from the existing set of
components. Use the pseudoization procedure of Coats to obtain
W
aI
and W
bI
values, and use Eq. 5.87 for k
I J
.
5. Use regression to fine tune the W
aI
and W
bI
values estimated
in Step 4; also regress on key BIPs, such as (N
2
)C
1
)*C
7)
,
(CO
2
)C
2
)*C
7)
, and other nonzero BIPs involving pseudocom-
ponents from Step 4.
6. Repeat Steps 4 and 5 until the quality of the characterization
deteriorates beyond an acceptable fluid description. Table 5.9
shows an example five-step pseudoization procedure.
In summary, any grouping of a complete EOS characterization
into a limited number of pseudocomponents should be checked to
ensure that predicted phase behavior (e.g., multicontact gas injec-
tion data, saturation pressures, and densities) are reasonably close
to the predictions for the original (complete) characterization. Step-
wise regression is the best approach to determine the number and
properties of pseudocomponents that can accurately describe a res-
ervoir fluids phase behavior. If stepwise regression is not possible,
standard grouping of the light and intermediates (N
2
)C
1
,
CO
2
)C
2
, i-C
4
)n-C
4
, and i-C
5
)n-C
5
) and Gaussian quadrature
for C
7)
(or equal-mass fractions) is recommended; a valid alterna-
tive is the Li et al.
59
method. The Coats
57
method (Eqs. 5.93
and 5.94) is always recommended for calculating pseudocompon-
ent properties.
References
1. Yarborough, L.: Application of a Generalized Equation of State to Pe-
troleum Reservoir Fluids, Equations of State in Engineering and Re-
search, K.C. Chao and R.L. Robinson Jr. (eds.), Advances in Chemistry
Series, American Chemical Soc., Washington, DC (1978) 182, 386.
2. Whitson, C.H.: Characterizing Hydrocarbon Plus Fractions, SPEJ
(August 1983) 683; Trans., AIME, 275.
3. Pedersen, K.S., Thomassen, P., and Fredenslund, A.: SRK-EOS Cal-
culation for Crude Oils, Fluid Phase Equilibria (1983) 14, 209.
4. Craft, B.C., Hawkins, M., and Terry, R.E.: Applied Petroleum Reservoir
Engineering, second edition, Prentice-Hall Inc., Englewood Cliffs,
New Jersey (1991).
5. McCain, W.D. Jr.: The Properties of Petroleum Fluids, second edition,
PennWell Publishing Co., Tulsa, Oklahoma (1990).
6. Katz, D.L. and Firoozabadi, A.: Predicting Phase Behavior of Conden-
sate/Crude-Oil Systems Using Methane Interaction Coefficients, JPT
(November 1978) 1649; Trans., AIME, 265.
7. Austad, T. et al.: Practical Aspects of Characterizing Petroleum
Fluids, paper presented at the 1983 North Sea Condensate Reservoirs
and Their Development Conference, London, 2425 May.
8. Chorn, L.G.: Simulated Distillation of Petroleum Crude Oil by Gas
ChromatographyCharacterizing the Heptanes-Plus Fraction, J.
Chrom. Sci. (January 1984) 17.
9. MacAllister, D.J. and DeRuiter, R.A.: Further Development and Ap-
plication of Simulated Distillation for Enhanced Oil Recovery, paper
SPE 14335 presented at the 1985 SPE Annual Technical Conference
and Exhibition, Las Vegas, Nevada, 2225 September.
10. Designation D158, Saybolt Distillation of Crude Petroleum, Annual
Book of ASTM Standards, ASTM, Philadelphia, Pennsylvania (1984).
11. Designation D2892-84, Distillation of Crude Petroleum (15.Theoreti-
cal Plate Column), Annual Book of ASTM Standards, ASTM, Philadel-
phia, Pennsylvania (1984) 8210.
12. Kesler, M.G. and Lee, B.I.: Improve Predictions of Enthalpy of Frac-
tions, Hydro. Proc. (March 1976) 55, 153.
13. Lee, B.I. and Kesler, M.G.: A Generalized Thermodynamic Correla-
tion Based on Three-Parameter Corresponding States, AIChE J.
(1975) 21, 510.
14. Riazi, M.R. and Daubert, T.E.: Simplify Property Predictions, Hydro.
Proc. (March 1980) 115.
15. Maddox, R.N. and Erbar, J.H.: Gas Conditioning and ProcessingAd-
vanced Techniques and Applications, Campbell Petroleum Series, Nor-
man, Oklahoma (1982) 3.
16. Organick, E.I. and Golding, B.H.: Prediction of Saturation Pressures
for Condensate-Gas and Volatile-Oil Mixtures, Trans., AIME
(1952) 195, 135.
17. Katz, D.L. et al.: Handbook of Natural Gas Engineering, McGraw-Hill
Book Co. Inc., New York City (1959).
18. Riazi, M.R. and Daubert, T.E.: Analytical Correlations Interconvert
Distillation-Curve Types, Oil & Gas J. (August 1986) 50.
19. Riazi, M.R. and Daubert, T.E.: Characterization Parameters for Petro-
leum Fractions, Ind. Eng. Chem. Res. (1987) 26, 755.
20. Robinson, D.B. and Peng, D.Y.: The Characterization of the Heptanes
and Heavier Fractions, Research Report 28, Gas Producers Assn., Tul-
sa, Oklahoma (1978).
21. Riazi, M.R. and Daubert, T.E.: Prediction of the Composition of Petro-
leum Fractions, Ind. Eng. Chem. Proc. Des. Dev. (1980) 19, 289.
22. Pedersen, K.S., Thomassen, P., and Fredenslund, A.: Thermodynam-
ics of Petroleum Mixtures Containing Heavy Hydrocarbons. 1. Phase
Envelope Calculations by Use of the Soave-Redlich-Kwong Equation
of State, Ind. Eng. Chem. Proc. Des. Dev. (1984) 23, 163.
23. Pedersen, K.S., Thomassen, P., and Fredenslund, A.: Thermodynam-
ics of Petroleum Mixtures Containing Heavy Hydrocarbons. 2. Flash
and PVT Calculations with the SRK Equation of State, Ind. Eng.
Chem. Proc. Des. Dev. (1984) 23, 566.
24. Lohrenz, J., Bray, B.G., and Clark, C.R.: Calculating Viscosities of
Reservoir Fluids From Their Compositions, JPT (October 1964)
1171; Trans., AIME, 231.
25. Whitson, C.H.: Effect of C
7)
Properties on Equation-of-State Predic-
tions, paper SPE 11200 presented at the 1982 SPE Annual Technical
Conference and Exhibition, New Orleans, 2629 September.
26. Whitson, C.H.: Effect of C
7)
Properties on Equation-of-State Predic-
tions, SPEJ (December 1984) 685; Trans., AIME, 277.
27. Whitson, C.H., Andersen, T.F., and Sreide, I.: C
7)
Characteriza-
tion of Related Equilibrium Fluids Using the Gamma Distribution,
C
7)
Fraction Characterization, L.G. Chorn and G.A. Mansoori
(eds.), Advances in Thermodynamics, Taylor & Francis, New York
City (1989) 1, 3556.
28. Brul, M.R., Kumar, K.H., and Watansiri, S.: Characterization
Methods Improve Phase-Behavior Predictions, Oil & Gas J. (11
February 1985) 87.
29. Hoffmann, A.E., Crump, J.S., and Hocott, C.R.: Equilibrium
Constants for a Gas-Condensate System, Trans., AIME (1953) 198, 1.
30. Abramowitz, M. and Stegun, I.A.: Handbook of Mathematical Func-
tions, Dover Publications Inc., New York City (1970) 923.
31. Haaland, S.: Characterization of North Sea Crude Oils and Petroleum
Fractions, MS thesis, Norwegian Inst. of Technology, Trondheim,
Norway (1981).
32. Watson, K.M., Nelson, E.F., and Murphy, G.B.: Characterization of
Petroleum Fractions, Ind. Eng. Chem. (1935) 27, 1460.
33. Watson, K.M. and Nelson, E.F.: Improved Methods for Approximat-
ing Critical and Thermal Properties of Petroleum, Ind. Eng. Chem.
(1933) 25, No. 8, 880.
34. Jacoby, R.H. and Rzasa, M.J.: Equilibrium Vaporization Ratios for Ni-
trogen, Methane, Carbon Dioxide, Ethane, and Hydrogen Sulfide in
Absorber Oil/Natural Gas and Crude Oil/Natural Gas Systems, Trans.,
AIME (1952) 195, 99.
35. Sreide, I.: Improved Phase Behavior Predictions of Petroleum Reser-
voir Fluids From a Cubic Equation of State, Dr.Ing. dissertation, Nor-
wegian Inst. of Technology, Trondheim, Norway (1989).
36. Technical Data BookPetroleum Refining, third edition, API, New
York City (1977).
37. Rao, V.K. and Bardon, M.F.: Estimating the Molecular Weight of Pe-
troleum Fractions, Ind. Eng. Chem. Proc. Des. Dev. (1985) 24, 498.
38. Benedict, M., Webb, G.B., and Rubin, L.C.: An Empirical Equation
for Thermodynamic Properties of Light Hydrocarbons and Their Mix-
tures, I. Methane, Ethane, Propane, and n-Butane, J. Chem. Phys.
(1940) 8, 334.
39. Reid, R.C.: Present, Past, and Future Property Estimation Tech-
niques, Chem. Eng. Prog. (1968) 64, No. 5, 1.
40. Reid, R.C., Prausnitz, J.M., and Polling, B.E.: The Properties of Gases
and Liquids, fourth edition, McGraw-Hill Book Co. Inc., New York
City (1987) 1224.
41. Roess, L.C.: Determination of Critical Temperature and Pressure of
Petroleum Fractions, J. Inst. Pet. Tech. (October 1936) 22, 1270.
42. Cavett, R.H.: Physical Data for Distillation Calculations-Vapor-Liq-
uid Equilibria, Proc., 27th API Meeting, San Francisco (1962) 351.
43. Nokay, R.: Estimate Petrochemical Properties, Chem. Eng. (23 Feb-
ruary 1959) 147.
44. Pitzer, K.S. et al.: The Volumetric and Thermodynamic Properties of
Fluids, II. Compressibility Factor, Vapor Pressure, and Entropy of Va-
porization, J. Amer. Chem. Soc. (1955) 77, No. 13, 3433.
45. Edmister, W.C.: Applied Hydrocarbon Thermodynamics, Part 4:
Compressibility Factors and Equations of State, Pet. Ref. (April
1958) 37, 173.
46. Hall, K.R. and Yarborough, L.: New, Simple Correlation for Predict-
ing Critical Volume, Chem. Eng. (November 1971) 76.
47. Kesler, M.G., Lee, B.I., and Sandler, S.I.: A Third Parameter for Use
in Generalized Thermodynamic Correlations, Ind. Eng. Chem. Fund.
(1979) 18, No. 1, 49.
48. Twu, C.H.: An Internally Consistent Correlation for Predicting the
Critical Properties and Molecular Weights of Petroleum and Coal-Tar
Liquids, Fluid Phase Equilibria (1984) No. 16, 137.
49. Peng, D.Y. and Robinson, D.B.: A New-Constant Equation of State,
Ind. Eng. Chem. Fund. (1976) 15, No. 1, 59.
50. Whitson, C.H.: Critical Properties Estimation From an Equation of
State, paper SPE 12634 presented at the 1984 SPE/DOE Symposium
on Enhanced Oil Recovery, Tulsa, Oklahoma, 1518 April.
51. Brul, M.R. and Starling, K.E.: Thermophysical Properties of Com-
plex Systems: Applications of Multiproperty Analysis, Ind. Eng.
Chem. Proc. Des. Dev. (1984) 23, 833.
20 PHASE BEHAVIOR
52. Peneloux, A., Rauzy, E., and Freze, R.: A Consistent Correction for
Redlich-Kwong-Soave Volumes, Fluid Phase Equilibria (1982) 8, 7.
53. Jhaveri, B.S. and Youngren, G.K.: Three-Parameter Modification of
the Peng-Robinson Equation of State To Improve Volumetric Predic-
tions, SPERE (August 1988) 1033; Trans., AIME, 285.
54. Chueh, P.L. and Prausnitz, J.M.: Calculation of High-Pressure Vapor
Liquid Equilibria, Ind. Eng. Chem. (1968) 60, No. 13.
55. Pedersen, K.S., Thomassen, P., and Fredenslund, A.: Characterization
of Gas Condensate Mixtures, C
7)
Fraction Characterization, L.G.
Chorn and G.A. Mansoori (eds.), Advances in Thermodynamics, Tay-
lor & Francis, New York City (1989) 1.
56. Zudkevitch, D. and Joffe, J: Correlation and Prediction of Vapor-Liq-
uid Equilibrium with the Redlich-Kwong Equation of State, AIChE J.
(1970) 16, 112.
57. Coats, K.H.: Simulation of Gas-Condensate-Reservoir Performance,
JPT (October 1985) 1870.
58. Lee, S.T. et al.: Experiments and Theoretical Simulation on the Fluid
Properties Required for Simulation of Thermal Processes, SPEJ (Oc-
tober 1982) 535.
59. Li, Y.-K., Nghiem, L.X., and Siu, A.: Phase Behavior Computation for
Reservoir Fluid: Effects of Pseudo Component on Phase Diagrams and
Simulations Results, paper CIM 84-35-19 presented at the 1984 Petro-
leum Soc. of CIM Annual Meeting, Calgary, 1013 June.
60. Newley, T.M.J. and Merrill, R.C. Jr.: Pseudocomponent Selection
for Compositional Simulation, SPERE (November 1991) 490;
Trans., AIME, 291.
61. Schlijper, A.G.: Simulation of Compositional Processes: The Use of
Pseudocomponents in Equation-of-State Calculations, SPERE (Sep-
tember 1986) 441; Trans., AIME, 282.
62. Behrens, R.A. and Sandler, S.I.: The Use of Semicontinuous Descrip-
tion To Model the C
7)
Fraction in Equation of State Calculations, pa-
per SPE 14925 presented at the 1986 SPE/DOE Symposium on En-
hanced Oil Recovery, Tulsa, Oklahoma, 2323 April.
63. Wu, R.S. and Batycky, J.P.: Pseudocomponent Characterization for
Hydrocarbon Miscible Displacement, paper SPE 15404 presented at
the 1986 SPE Annual Technical Conference and Exhibition, New Or-
leans, 56 October.
64. Kay, W.B.: The Ethane-Heptane System, Ind. & Eng. Chem. (1938)
30, 459.
ft
3
/lbm mol 6.242 796 E*02+m
3
/kmol
F (F*32)/1.8 +C
F (F)459.67)/1.8 +K
psi 6.894 757 E)00+kPa
R 5/9 +K
This paper was prepared for presentation at the 2007 SPE International Student Paper Con-
test at the SPE Annual Technical Conference and Exhibition being held in Anaheim, California,
11-14 November, 2007.

This paper was selected for presentation by merit of placement in a regional student paper
contest held in the program year preceding the International Student Paper Contest. Contents
of the paper, as presented, have not been reviewed by the Society of Petroleum Engineers
and are subject to correction by the author(s). The material, as presented, does not necessar-
ily reflect any position of the Society of Petroleum Engineers, its officers, or members.

Abstract
A compositional simulator uses an Equation of State (EOS) to
predict the pressure-volume-temperature (PVT) behavior of
gas and crude oil fluids, which are very complex hydrocarbon
mixtures.
In a hydrocarbon mixture, the critical properties (critical
pressure P
c
, critical temperature T
c
and accentric factor
) must be given for each component. These properties are
well known for pure compounds (like methane, ethane, etc.),
but nearly all naturally occurring gas and crude oil fluids con-
tain some heavy fractions that are not well defined. These
heavy fractions are lumped and called the plus-fraction
(C
7+
). There arises the need of adequately characterizing these
undefined plus fractions in terms of their critical properties.
This work presents a correlation for the critical properties
of the plus-fraction C
7+
needed in the characterization of crude
oil fluid samples for a compositional simulator using Peng-
Robinsons
1
EOS. The correlation is a function of the frac-
tions molecular weight (MW).
Twenty (20) PVT laboratory tests made in black oil fluid
samples taken from fields of the Neuqun Basin (Argentina)
were utilized for the adjustment of the bubble point (Pb), gas
oil ratio (GOR), constant composition expansion pressure-
volume relation (PV) and differential expansion curves (Rs,
Bo and o) of the compositional simulator.
A comparative study was performed against the Riazi-
Dauberts
2
correlation. The correlation here presented gives
better results than those of Riazi-Dauberts correlation and
also gives excellent liquid density predictions when there are
high MW hydrocarbons present in the fluid.

Introduction
One of the biggest existing problems when using a composi-
tional simulator is the lack of good critical properties estima-
tion for the heavy fraction of the fluid.
3
Furthermore, for the
characterization of the fraction few data are available, usually
MW, SG and normal boiling point (T
b
). For this correlation
MW and SG were utilized, as both are habitually given or can
be obtained from the molar composition of the fluid.
There are various methods for the characterization of
heavy fractions: Riazi-Daubert, Kessler-Lee
4
, Twu
5
and Sta-
mataki-Magoulas
6
. This study only compares to Riazi and
Daubert, since it is the one mostly used in the industry; Stama-
taki-Magoulas is presented as a better alternative to Twu and
Kessler-Lee, but its performance was deficient, consequently
these last three studies are discarded from the comparison.
The trigger of this work was that estimations generated
from the Riazi-Dauberts correlation were not close enough to
the PVT laboratory tests, and fraction parameters adjustment
was very time consuming in the simulator. A correlation was
created that could solve this problem, giving closer results to
laboratory curves and as a consequence, less time spent in the
critical parameters adjustment.

Equation of State
Today Peng-Robinson EOS (PR-EOS, 1976) is one of the
most widely used EOS in the chemical and petroleum indus-
try. It is superior when considering liquid density predictions
to the Soave-Redlich-Kwong EOS (SRK-EOS, 1972), al-
though all cubic EOS experience difficulties in the liquid den-
sity calculations, which brings to the application of certain
modifications to these equations. We have used two modifica-
tions presented below. The EOS is presented here:

2 2
2
m m m
RT a
P
V b V bV b
=
+
(1)

2 2
0.457235528921
c
c
R T
a
P
= (2)

0.077796073904
c
c
RT
b
P
= (3)

( )
2
1 1
r
T

= +

(4)

2
0.37464 1.54226 0.26992 = + (5)

For the application of EOS in mixtures (this is the case of
petroleum fluids), certain mixture rules must be utilized. In
this study the quadratic mixture rule is used, since it is the
simplest:

SPE-113026-STU (Student 6)
Heavy Fraction C
7+
Characterization for PR-EOS
Gastn Fondevila Sancet, Buenos Aires Institute of Technology
2 SPE Student Paper
( )
1
ij i j ij
a a a k = (6)

2
i j
ij
b b
b
+
= (7)

Binary interaction parameters k
ij
of the Eq. 6 were ob-
tained from Nishiumis
7
correlation.

EOS Modifications
In our case the compositional simulator utilizes two modifica-
tions: Twu-Coon-Cunningham
8
and Twu-Tilton-Bluck
9
.

Twu-Coon-Cunningham. This modification considers a new
relation of the attraction term: the parameter, temperature
and accentric factor dependent. They propose the next equa-
tions:

( ) ( ) ( )
( )
0 1 0
= + (8)

( )
( )
0 0.171813 1.77634
exp 0.125283 1
r r
T T

=

(9)

( )
( )
1 0.607352 2.20517
exp 0.511614 1
r r
T T

=

(10)

Twu-Tilton-Bluck. Presents a volume-change factor for im-
proving the liquid density predictions. They propose the fol-
lowing equations:

, , s CEOS s RA
c v v = (11)

( )
2 7
1 1
,
Tr
c
s RA RA
c
RT
v Z
P

+

=

(12)

CEOS i i
v v x c =
(13)

Where:
v
s,CEOS
= volume of saturated liquid calculated by a cubic EOS.
v
s,RA
= volume of saturated liquid calculated by Racketts
equation (Eq. 12).
Z
RA
= Racketts parameter (see Table 3).

In the case of pseudo-components, the value of Z
RA
is equal to
the critical compressibility factor Z
c
.

Riazi-Daubert Correlation
Riazi-Dauberts correlation utilizes MW and SG as heavy
fraction parameters. They propose the same type of equation
for all the physical properties, varying only the coefficients in
each particular case. For calculating the accentric factor, Ri-
azi-Daubert uses the Edmisters
10
correlation. The proposed
relationship is next:

( ) exp
b c
aMW SG dMW eSG f MW SG = + +

(14)

Where:
= some physic property.
a-f = coefficients for each physic property (Table 1).

The Edmisters correlation for the accentric factor is P
c

and T
c
dependent. Is given by:

( ) log 14.7
3
1
7 1
c
c b
P
T T
=
(15)

TABLE 1 Riazi and Dauberts coefficients
T
c
P
c
T
b

a
544.4 45203 6.77857
b 0.2998 -0.8063 0.401673
c 1.0555 1.6015 -1.58262
d -0.00013478 -0.0018078 0.00377409
e -0.61641 -0.3084 2.984036
f 0.0 0.0 -0.00425288

PVT Data
A selection of twenty (20) PVT studies was realized in the
Neuqun-Basin, mostly Black-Oil samples from these reser-
voirs: Centenario, Mulichinco, Petrolfera, Quintuco, Sierras
Blancas and Tordillo. The samples posses an API range be-
tween 30 and 60, a GOR range between 50 and 250 m
3
/m
3
and
a C7+ mole % range between 25% and 50%. Their character-
istics are given in Table 4 and their heavy fraction properties
are given in Table 5.

Methodology
The compositional simulator utilized is the software UniTest
(property of DeltaP). It uses the PR-EOS with the modifica-
tions previously presented, the quadratic mixing rule and the
Nishiumis binary interaction parameters. The outputs of the
simulator are the following curves:

- Solution gas oil ratio (Rs), DE.
- Oil formation volume factor (Bo), DE.
- Oil density (o), DE.
- Pressure volume relation (PV), CCE.

DE stands for Differential Expansion and CCE stands for
Constant Composition Expansion.
The software has an optimization function for comparing
the laboratory data with the output of the simulator, varying
parameters selected by the user and adjusting the data set by
the least squares method. This function was used to adjust
the density curve, varying the plus-fraction MW and obtaining
an array of adjusted MW (called MW*).

SPE Student Paper 3
Each crude oil sample was represented by a composi-
tional model of the reservoir fluid. This model is defined by
the mole composition of: Nitrogen (N
2
), Carbon Dioxide
(CO
2
), Methane (CH
4
) through Hexane (C
6
H
14
); the critical
properties of these elements were obtained from Reid
11
and
are presented in Table 3. Lastly the pseudo-component
(which represents the heptanes plus-fraction of the fluid, in
this case C
7+
) is defined with the correlation presented next.

Proposed Correlation
First a simple curve matching, using the previously presented
software function, between the compositional simulator and
PVT tests data, varying the critical properties of the C
7+
, was
tried. The results were not promising: an excellent matching
for each sample was accomplished, but there was no relation-
ship encountered between the MW and SG with the critical
properties, the matching took several minutes of computer
calculations and no physical meaning or explanation was
taken into account.
Then it was decided to try something simpler: the Reids
known critical properties of pure compounds between n-C
7

and n-C
20
. The reason is that for this set of black oil samples,
the MW of the heptanes plus fraction is between the MW n-C
7

and n-C
20
, so it is not radical to think the C
7+
as an average
compound between these two.
A relationship was established between the MW and the
critical properties using the data set provided by Reid. The
functions for obtaining P
c
(Eq. 16) and T
c
(Eq. 17) from MW
were working well but the obtained brought some problems:
for samples with high MW of the C
7+
, the was greater than
unity and the vapor pressures obtained by the EOS at low
temperatures diverged to infinity.
So it was decided to represent the accentric factor using the
correlation proposed by Edmister (Eq. 15). This decision is
based on the fact that the values of calculated accentric factors
are less than unity and the previous problem disappears.
Using Edmisters correlation for the calculation brings
us an additional problem: the normal boiling point temperature
T
b
needs to be measured or obtained from the heavy fraction.
For solving this inconvenient it was also appealed to Reids
data set and a relation between T
b
and T
c
was found (Eq. 18).
The correlation for the C
7+
is presented next:

( ) [ ] 82.82 653exp 0.007427
c
P psia MW = + (16)

( ) [ ] 778.5 383.5ln 4.075
c
T R MW = + (17)

( )
1.869
[ ] 194 0.001241 [ ]
b c
T R T R = + (18)

The results of using this correlation in the compositional
simulator are promising: in all the samples the bubble point is
very close to the PVT value so the matching is sometimes not
needed, and also the PV, Rs and Bo curves are very close to
their PVT values. Graphics of the previous equations are pre-
sented in Figure 1 to Figure 3.
But here arises another problem: the liquid density
curves calculated by the simulator are far beyond from the
laboratory measured curves, although the PV, Rs and Bo
curves are adjusted. The solution: match the oil density curve
given by the simulator varying the MW of the plus-fraction.
This new MW is called the adjusted MW (MW*) and is
only used for calculating the oil density in the simulator. A
relation between MW* and MW (Eq. 19) was found:

( ) * 39.44 47.47exp 0.006701 MW MW = + (19)

This equation is presented in Figure 4.

Correlation Comparison
The comparison of both correlations was done using the rela-
tive average error RAE, calculated between the PVT data
points and the output points of the simulator, for each correla-
tion. Table 2 presents the RAE comparison between the two
correlations, Table 6 and Table 7 shows the RAE of Pb,
GOR, PV, Rs, Bo and o for each sample. The relative aver-
age error (RAE) is given by:

1
calc data
data
x x
RAE
N x
=

(20)

Where:
N = number of data points
x
calc
= calculated variable
x
data
= known data variable

TABLE 2 RAE Comparison

RAE
This Work
RAE
Riazi-Daubert
Pb 2.29% 3.43%
GOR 7.37% 11.10%
PV 1.42% 1.47%
Rs 9.24% 12.59%
Bo 2.24% 2.20%
o 1.50% 12.63%

In Figure 5 and Figure 6 the comparisons between cal-
culated and real values of Pb and GOR are presented, respec-
tively. From Figure 7 to Figure 26 the PVT and simulated
curves of the following samples can be appreciated: Sample 1
(GOR = 250 m
3
/m
3
, m% C
7+
= 26%), Sample 6 (GOR = 180
m
3
/m
3
, m% C
7+
= 33%), Sample 10 (GOR = 130 m
3
/m
3
, m%
C
7+
= 38%), Sample 15 (GOR = 76 m
3
/m
3
, m% C
7+
= 44%),
Sample 20 (GOR = 50 m
3
/m
3
, m% C
7+
= 51%). The PVT
properties presented are PV, Rs, Bo and o. The contrast in oil
density predictions between this correlation and RD can be
clearly seen, showing that this work gives better oil density
prediction performance.
4 SPE Student Paper
Conclusions
This work presents a new set of correlations for obtain-
ing the critical properties of a pseudo-component repre-
senting the heptanes plus fraction C
7+
of a crude oil fluid.
These correlations are function of the fractions MW.
A plus-fractions MW adjustment function is presented
for improving the liquid density predictions of the simu-
lator. This function shows to be a solution to the problem
of the cubic EOS when predicting liquid densities in res-
ervoirs fluids with high content of heavy hydrocarbons.
This work is compared to the Riazi-Dauberts correla-
tion: the two correlations performed equally when pre-
dicting the PV and Bo curves; in the calculation of the Rs
curve this work obtained better results than the other. In
oil density calculations, this work demonstrates very
good predictions, better than the previous correlation.

Future Work Recommendations
Add to the adjustment, phase envelope curves obtained
in the laboratory, since the simulator can generate this
type or curves. This point should be taken into considera-
tion because the adjustments are realized at constant tem-
perature, equal to reservoir temperature, having as a re-
sult an infinite number of phase diagrams that coincide in
the same isotherm, and the user must be careful when ex-
trapolating to other temperatures.
The data base of this study should be enlarged with a
large number of samples, and from more fields: Golfo
San Jorge, Noroeste, Cuyo and Australbasin. Also cor-
relations for different types of fluid should be created:
volatile oil and gascondensate.
Enlarge this work with the study of more heavy frac-
tions: C
10+
and C
20+
. Also the effect of dividing the heavy
fraction in various pseudo-components with a determined
distribution of molecular weights and compositions
should be studied. In this point it should be taken into ac-
count the computer power needed, because with more
components used, the time of calculations increase expo-
nentially.

Acknowledgments
I want to express my most sincerely gratefulness to all the
people that have helped me to do this work:

Rubn Caligari, Petrobras Energa SA. For granting
me the access to the PVT laboratory information needed.
Marcelo Crotti, I nLab. For his excellent predisposition
in helping me with the selection of the information and
the PVT tests.
Juan A. Rosbaco, I TBA. For being my mentor in this
work and for the excellent person that he is.
Miguel Schindler, DeltaP. For letting me use the soft-
ware UniTest. Also for helping me evaluate certain
theory aspects of this study.
Nomenclature
Bo = oil formation volume factor
CCE = constant concentration expansion
DE = Differential expansion
EOS = equation of state
GOR = gas oil ratio, m
3
/m
3

k
ij
= binary interaction coefficient, k
ij
= k
ji

MW = molecular weight
P = pressure, Pa
P
b
= bubble point, kg/cm
2

P
c
= critical pressure, psia
PR = Peng-Robinson
PV = pressure volume ratio
PVT = pressure volume temperature
R = universal gas constant, 8.3143 J mol
-1
K
-1

RAE = relative average error, %
RD = Riazi and Daubert
Rs = solution gas oil ratio, m
3
/m
3

SG = specific gravity relative to water at 1 atm and 60 F
T = temperature, K
T
b
= normal boiling point temperature, R
T
c
= critical temperature, R
T
r
= reduced temperature, where T
r
= T / T
c
T
res
= reservoir temperature, C
V
m
= molar volume, m
3

Z
c
= critical compressibility factor
Z
RA
= Racketts parameter

Greek letters

o
= oil density, g/cm
3

= accentric factor

References
1. Peng, D.Y. and Robinson, D.B.: A New Two-Constant
Equation of State, Ind. and Eng. Chem. Fund. (1976)
15, No. 1, 59-64.
2. Riazi, M. R., Daubert, T. E., Characterization Parame-
ters for Petroleum Fractions, Ind. Eng. Chem. Res.,
1987, Vol. 26, No. 24, pp. 755-759.
3. Whitson, C.H.: Effect of C7+ Properties on Equation-
of-State Predictions, 1982 SPE Annual Technical Con-
ference and Exhibition, SPE 11200.
4. Kesler, M.G. and Lee, B.I.: Improve Prediction of En-
thalpy of Fractions, Hydrocarbon Processing (1976) 55,
59.
5. Twu, C.H.: An internally Consistent Correlation for
Predicting the Critical Properties and Molecular Weights
of Petroleums and Coal-Tar Liquids, Fluid Phase Equi-
libria (1984) 16, 137.
6. Stamataki, S.K. and Magoulas, K.G., Characterization
of Heavy Undefined Fractions, 2001 SPE International
Symposium on Oilfield Chemistry, SPE 64996.
7. H. Nishiumi and T. Arai, Generalization of the Binary
Interaction Parameter of the Peng-Robinson Equation of
State by Component Family, Fluid Phase Equilibria
(1988), 42, 43-62.
SPE Student Paper 5
8. Twu, C.H., Coon, J.E. and Cunningham, J.R., A New
Generalized Alpha Function for a Cubic Equation of
State. Part 1: Peng-Robinson EOS, Fluid Phase Equi-
libria (1995), Volume 105, Number 1.
9. Twu, C.H., Tilton B. and Bluck D., The Strengths and
Limitations of Equations of State Models and Mixing
Rules.
10. Edmister, W.C., Applied Hydrocarbon Thermodynamic,
Part 4: Compressibility Factors and Equations of State,
Petroleum Refiner, April 1958, Vol. 37, pp. 173-179.
11. Reid, R.C., Prausnitz, J.M. and Poling, B.E.: The Prop-
erties of Gases and Liquids, McGraw-Hill Companies;
4th edition (April 1987).

GOR: m
3
/m
3
x 35.31467 = scf/stb
Pressure: kg/cm
2
x 14.22334 = psia
Temperature: C x 1.8 + 32 = F
Density: g/cm
3
x 62.42796 = lb/ft
3

Appendix
In the Tables 6 and 7, GF stands for this work (authors name)
and RD stands for Riazi-Daubert.

TABLE 3 Reids Critical Properties
Element MW
P
c

[psia]
T
c

[R]
Z
RA

C
1
16.0 666 343 0.0115 0.2892
C
2
30.1 707 550 0.0908 0.2808
C
3
44.1 616 666 0.1454 0.2766
n-C
4
58.1 551 766 0.1928 0.2730
i-C
4
58.1 528 735 0.1756 0.2754
n-C
5
72.2 489 846 0.2510 0.2684
i-C
5
72.2 490 829 0.2273 0.2717
C
6
86.2 437 914 0.2957 0.2635
C
7
100.2 397 973 0.3506 0.2604
C
8
114.2 361 1024 0.3978 0.2571
C
9
128.3 334 1071 0.4437 0.2543
C
10
142.3 306 1109 0.4902 0.2507
C
11
156.3 285 1150 0.5349 0.2499
C
12
170.3 264 1185 0.5622 0.2466
C
13
184.4 250 1216 0.6231 0.2473
C
14
198.4 235 1245 0.6797 0.2430
C
15
221.4 220 1272 0.7060 0.2418
C
16
226.4 206 1297 0.7418 0.2388
C
17
240.5 191 1320 0.7699 0.2343
C
18
254.5 176 1341 0.7895 0.2275
C
19
268.5 197 1397 0.8270 0.2278
C
20
282.6 161 1381 0.9070 0.2281

TABLE 4 PVT Samples Properties
Sample API
GOR
[m
3
/m
3
]
T
res
[C]
P
b

[kg/cm
2
]
Sample 1 50.8 249.9 90.0 150.11
Sample 2 47.7 213.1 82.6 146.59
Sample 3 38.0 208.8 105.0 241.53
Sample 4 41.2 204.8 73.9 269.70
Sample 5 60.2 185.4 81.7 69.10
Sample 6 57.2 181.2 81.2 127.00
Sample 7 31.8 163.6 95.6 253.39
Sample 8 42.8 138.1 156.0 198.83
Sample 9 34.9 130.7 86.7 201.48
Sample 10 47.9 129.5 62.5 161.71
Sample 11 35.7 128.2 63.9 159.20
Sample 12 31.7 88.1 61.7 163.73
Sample 13 33.4 85.8 63.0 141.28
Sample 14 32.5 77.9 54.0 140.43
Sample 15 32.4 75.8 59.8 135.30
Sample 16 35.1 75.2 91.5 113.60
Sample 17 35.2 72.7 70.0 111.92
Sample 18 31.8 64.6 52.0 121.39
Sample 19 41.4 57.8 76.0 66.51
Sample 20 32.8 50.2 59.0 82.08

TABLE 5 C
7+
Properties of Samples
Sample MW SG
x

[mole %]
Sample 1 163.4 0.797 26.13%
Sample 2 170.3 0.799 27.82%
Sample 3 224.5 0.844 28.97%
Sample 4 218.7 0.829 27.55%
Sample 5 144.4 0.779 29.52%
Sample 6 136.7 0.773 32.62%
Sample 7 256.8 0.870 30.67%
Sample 8 217.5 0.821 38.54%
Sample 9 241.2 0.861 36.50%
Sample 10 182.8 0.801 38.52%
Sample 11 234.9 0.859 36.05%
Sample 12 272.1 0.879 40.98%
Sample 13 238.2 0.871 41.70%
Sample 14 251.5 0.873 45.35%
Sample 15 266.8 0.875 44.19%
Sample 16 236.7 0.865 47.30%
Sample 17 223.9 0.863 48.45%
Sample 18 268.5 0.879 46.57%
Sample 19 221.0 0.836 52.60%
Sample 20 255.5 0.878 50.93%
6 SPE Student Paper
TABLE 6 RAE of Pb, GOR and PV
Pb [%] GOR [%] PV [%]
Sample
GF RD GF RD GF RD
Sample 1 2.1 4.6 5.5 4.6 1.5 1.4
Sample 2 4.3 5.9 0.5 0.6 1.4 1.2
Sample 3 0.4 0.6 14.2 18.3 0.9 0.9
Sample 4 3.0 2.0 3.8 7.1 0.9 0.9
Sample 5 1.1 4.3 20.2 17.9 1.9 1.7
Sample 6 0.0 5.1 3.2 4.5 1.7 1.3
Sample 7 0.6 3.1 11.8 18.3 1.1 1.6
Sample 8 1.6 0.0 15.6 17.7 2.9 2.6
Sample 9 5.5 6.7 11.7 17.3 2.0 2.2
Sample 10 1.5 1.8 3.9 3.1 1.1 1.1
Sample 11 5.8 3.2 5.9 11.8 1.4 1.5
Sample 12 3.7 8.8 7.5 15.2 2.5 2.9
Sample 13 0.1 2.0 5.1 19.9 1.4 1.6
Sample 14 2.3 1.5 3.7 10.9 1.3 1.6
Sample 15 0.1 4.0 5.6 0.4 1.2 1.5
Sample 16 1.6 1.8 8.8 15.3 1.2 1.2
Sample 17 0.3 0.2 3.2 10.3 1.0 1.1
Sample 18 0.4 4.6 1.7 10.1 1.5 1.9
Sample 19 6.0 6.0 10.3 14.5 0.6 0.7
Sample 20 5.2 2.4 5.2 4.1 0.8 0.9

TABLE 7 RAE of Rs, Bo and o
Rs [%] Bo [%] o [%]
Sample
GF RD GF RD GF RD
Sample 1 9.5 9.7 2.0 2.5 1.5 5.6
Sample 2 5.1 5.7 2.3 2.3 1.8 6.5
Sample 3 17.9 21.7 3.4 5.3 1.3 11.0
Sample 4 8.0 10.8 1.7 1.1 2.0 9.4
Sample 5 21.6 19.8 8.8 8.6 0.8 3.9
Sample 6 2.0 3.8 1.2 1.6 1.6 3.3
Sample 7 16.3 21.2 1.4 3.4 1.6 16.4
Sample 8 15.3 18.1 1.9 2.9 2.5 8.6
Sample 9 14.2 19.2 1.0 1.4 1.2 15.3
Sample 10 4.5 3.6 3.0 2.7 3.3 5.4
Sample 11 11.5 16.0 1.2 1.5 1.5 13.6
Sample 12 5.3 10.6 2.7 1.2 0.9 21.4
Sample 13 3.0 4.0 3.6 1.9 1.5 17.0
Sample 14 4.9 10.8 1.9 0.6 0.7 17.9
Sample 15 7.4 13.5 2.0 0.6 1.0 19.7
Sample 16 11.4 17.7 0.8 1.9 0.8 15.0
Sample 17 5.3 11.2 0.9 0.9 0.7 13.8
Sample 18 2.9 9.7 1.9 0.6 0.7 20.7
Sample 19 15.2 19.2 1.3 2.2 3.5 9.8
Sample 20 3.7 5.3 1.7 0.6 1.0 18.3

MW
50 100 150 200 250 300 350
P
c

[
p
s
i
a
]
100
150
200
250
300
350
400
450

Figure 1. Pc vs MW from Reid, Eq. 16.

MW
50 100 150 200 250 300 350
T
c

[
R
]
900
1000
1100
1200
1300
1400
1500

Figure 2. Tc vs MW from Reid, Eq. 17.

Tc [R]
900 1000 1100 1200 1300 1400 1500
T
b

[
R
]
600
700
800
900
1000
1100
1200

Figure 3. Tb vs Tc from Reid, Eq. 18.

Pc = 82.82 + 653exp(0.007427MW)
R
2
= 0.9966
Tc = 778.5 + 383.5ln(MW 4.075)
R
2
= 0.9989
Tb = 194 + 0.001241Tc
1.869
R
2
= 0.9999
SPE Student Paper 7
MW
120 140 160 180 200 220 240 260 280 300
M
W
*
140
160
180
200
220
240
260
280
300
320
340
360

Figure 4. MW* vs MW, Eq. 19.

Pb [kg/cm
2
]
50 100 150 200 250 300
P
b

[
k
g
/
c
m
2
]
50
100
150
200
250
300
PVT Data
Pb This Work
Pb Riazi-Daubert

Figure 5. Comparison of predicted bubble points.

GOR [m
3
/m
3
]
50 100 150 200 250 300
G
O
R

[
m
3
/
m
3
]
50
100
150
200
250
300
PVT Data
GOR This Work
GOR Riazi-Daubert

Figure 6. Comparison of predicted gas oil ratios.
Pressure [kg/cm
2
abs]
50 100 150 200 250 300
R
e
l
a
t
i
v
e

V
o
l
u
m
e
0.90
0.95
1.00
1.05
1.10
1.15
1.20
1.25
1.30
PVT
This Work
Riazi-Daubert

Figure 7. CCE PV relation, Sample 1.

Pressure [kg/cm
2
abs]
0 50 100 150 200 250 300
R
s

[
m
3
/
m
3
]
0
50
100
150
200
250
300
PVT
This Work
Riazi-Daubert

Figure 8. DE Rs curve, Sample 1.

Pressure [kg/cm
2
abs]
0 50 100 150 200 250 300
B
o

[
m
3
/
m
3
]
1.0
1.2
1.4
1.6
1.8
2.0
2.2
PVT
This Work
Riazi-Daubert

Figure 9. DE Bo curve, Sample 1.

MW* = 39.44 + 47.47exp(0.006701MW)
R
2
= 0.9957
GOR = 250 m3/m3
mole % C7+ = 26%
GOR = 250 m3/m3
mole % C7+ = 26%
GOR = 250 m3/m3
mole % C7+ = 26%
8 SPE Student Paper
Pressure [kg/cm
2
abs]
0 50 100 150 200 250 300
o

[
g
/
c
m
3
]
0.50
0.55
0.60
0.65
0.70
0.75
0.80
PVT
This Work
Riazi-Daubert

Figure 10. DE Oil Density curve, Sample 1.

Pressure [kg/cm
2
abs]
50 100 150 200 250 300
R
e
l
a
t
i
v
e

V
o
l
u
m
e
0.90
0.95
1.00
1.05
1.10
1.15
1.20
1.25
1.30
PVT
This Work
Riazi-Daubert


Pressure [kg/cm
2
abs]
0 50 100 150 200 250 300
R
s

[
m
3
/
m
3
]
0
50
100
150
200
250
PVT
This Work
Riazi-Daubert

Pressure [kg/cm
2
abs]
0 50 100 150 200 250 300
B
o

[
m
3
/
m
3
]
1.0
1.2
1.4
1.6
1.8
2.0
PVT
This Work
Riazi-Daubert


Pressure [kg/cm
2
abs]
0 50 100 150 200 250 300
o

[
g
/
c
m
3
]
0.50
0.55
0.60
0.65
0.70
0.75
PVT
This Work
Riazi-Daubert


Pressure [kg/cm
2
abs]
50 100 150 200 250
R
e
l
a
t
i
v
e

V
o
l
u
m
e
0.95
1.00
1.05
1.10
1.15
1.20
1.25
1.30
PVT
This Work
Riazi-Daubert


GOR = 250 m3/m3
mole % C7+ = 26%
GOR = 180 m3/m3
mole % C7+ = 33%
GOR = 180 m3/m3
mole % C7+ = 33%
GOR = 180 m3/m3
mole % C7+ = 33%
GOR = 180 m3/m3
mole % C7+ = 33%
GOR = 130 m3/m3
mole % C7+ = 38%
SPE Student Paper 9
Pressure [kg/cm
2
abs]
0 50 100 150 200 250
R
s

[
m
3
/
m
3
]
0
20
40
60
80
100
120
140
160
PVT
This Work
Riazi-Daubert


Pressure [kg/cm
2
abs]
0 50 100 150 200 250
B
o

[
m
3
/
m
3
]
1.0
1.1
1.2
1.3
1.4
1.5
1.6
PVT
This Work
Riazi-Daubert


Pressure [kg/cm
2
abs]
0 50 100 150 200 250
o

[
g
/
c
m
3
]
0.55
0.60
0.65
0.70
0.75
0.80
0.85
PVT
This Work
Riazi-Daubert

Pressure [kg/cm
2
abs]
50 100 150 200 250
R
e
l
a
t
i
v
e

V
o
l
u
m
e
0.95
1.00
1.05
1.10
1.15
1.20
PVT
This Work
Riazi-Daubert


Pressure [kg/cm
2
abs]
0 50 100 150 200 250
R
s

[
m
3
/
m
3
]
0
20
40
60
80
100
PVT
This Work
Riazi-Daubert


Pressure [kg/cm
2
abs]
0 50 100 150 200 250
B
o

[
m
3
/
m
3
]
1.00
1.05
1.10
1.15
1.20
1.25
1.30
PVT
This Work
Riazi-Daubert

GOR = 130 m3/m3
mole % C7+ = 38%
GOR = 130 m3/m3
mole % C7+ = 38%
GOR = 130 m3/m3
mole % C7+ = 38%
GOR = 76 m3/m3
mole % C7+ = 44%
GOR = 76 m3/m3
mole % C7+ = 44%
GOR = 76 m3/m3
mole % C7+ = 44%
10 SPE Student Paper
Pressure [kg/cm
2
abs]
0 50 100 150 200 250
o

[
g
/
c
m
3
]
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
PVT
This Work
Riazi-Daubert


Pressure [kg/cm
2
abs]
0 50 100 150 200 250 300
R
e
l
a
t
i
v
e

V
o
l
u
m
e
0.95
1.00
1.05
1.10
1.15
1.20
1.25
1.30
PVT
This Work
Riazi-Daubert


Pressure [kg/cm
2
abs]
0 50 100 150 200 250 300
R
s

[
m
3
/
m
3
]
0
10
20
30
40
50
60
PVT
This Work
Riazi-Daubert

Pressure [kg/cm
2
abs]
0 50 100 150 200 250 300
B
o

[
m
3
/
m
3
]
1.00
1.05
1.10
1.15
1.20
1.25
PVT
This Work
Riazi-Daubert


Pressure [kg/cm
2
abs]
0 50 100 150 200 250 300
o

[
g
/
c
m
3
]
0.60
0.65
0.70
0.75
0.80
0.85
0.90
PVT
This Work
Riazi-Daubert


GOR = 76 m3/m3
mole % C7+ = 44%
GOR = 50 m3/m3
mole % C7+ = 51%
GOR = 50 m3/m3
mole % C7+ = 51%
GOR = 50 m3/m3
mole % C7+ = 51%
GOR = 50 m3/m3
mole % C7+ = 51%
Chequeo PVT
(Resumen Apunte de Rafael Cobeas)
Petrofsica y Fluidos de
Reservorio
G. Fondevila
Chequeo de Consistencia y Validacin
de Datos de un Informe PVT
Por qu es necesario realizarlo?:
En la etapa de caracterizacin de un
reservorio nuevo (o no conocido), es de vital
importancia obtener datos confiables, ya que
luego el reservorista utiliza a diario
informacin del fluido, que debe ser
coherente y representativa.
Esquema de Trabajo
Balance de Masa:
Balance de Masa Global
Balance de Masa por Etapas
Balance de Masa Composicional
Grficos de Control
Log K VS Tb
Hoffman
Bashbush
Chequeos Generales de Parmetros
Simulacin Termodinmica
Esquema de Trabajo
Balance de Masa:
Grficos de Control
Log K VS Tb
Hoffman
Bashbush
Balance de Masa
Balance de Masa Global:
Chequea que la densidad de la muestra medida sea

consistente con la calculada mediante los valores
reportados de:
Densidad de TNK
Gravedad especfica de los gases liberados en cada etapa
de separacin.
Relacin de gas producido por etapa (Rs).
Factor volumtrico de Formacin del Petrleo (Bo).
Balance de Masa por Etapas:
dem anterior, excepto
Se va chequeando etapa por etapa de la liberacin.

Balance de Masa Global (1)
En este chequeo se suman las masas

correspondientes a las salidas de los efluentes
en cada etapa de la liberacin diferencial junto
con la masa de petrleo residual.
La masas total dividida por el volumen en fondo dede

resultar igual a la densidad de la muestra en fondo.
Asumiendo que el volumen residual a condiciones de

tanque es Vtk, entonces la masa de lquido residual
(m_res) resulta igual a:
m_res = dens_res . Vtk
Balance de Masa Global (2)
La masa de gas en la isima de la liberacin diferencial

es igual a:
mgas
i
= (Rs
i-1
Rs
i
) . densgas
i
. Vtk
Por lo que, la masa total del sistema resulta:

m = [(Rs
i-1
Rs
i
) . densgas
i
] . Vtk + dens_res . Vtk
Y el volumen total a presin de burbuja y temperatura de

reservorio:
V = Vtk . Bodb
Entonces la densidad a esas condiciones es:

dens_Pb = ( [(Rs
i-1
Rs
i
) . densgas
i
] + dens_res ) / Bodb
Balance de Masa
Balance de Masa Composicional:
En cada etapa de la liberacin uno tiene el

dato de la fraccin molar de cada
componente tanto en la fase lquida como
gaseosa.
Se puede calcular las constantes de equilibrio

para cada componente en cada etapa
(presin) y compararles con valores de
referencia.
Esquema de Trabajo
Balance de Masa:
Grficos de Control
Log K VS Tb
Hoffman
Bashbush
Grficos de Control
Constantes de Equilibrio VS Temperatura

de Ebullicin:
En una determinada etapa de la liberacin (ya

sea flash o diferencial) uno puede calcular
para cada elemento la constante de equilibro
K (fraccin molar en fase gaseosa xi /
fraccin molar en fase lquida yi).
Luego se grafica en escala semi-log K vs Tb.

La misma debe tender a una lnea recta.
K vs Tb
Grficos de Control
Grfico de Hoffman (1952):
Nuevamente para una determinada etapa de

liberacin (presin), se grafica en escala
semi-log:
Absisas: Funcin especial de Hoffman que

depende de la inversa de la Temperatura del
Ensayo y de la Temperatura de Ebullicin del
componente i. f_Hoffman = b * (1/Tbi - 1/T)
Ordenadas: El producto Constante de Equilibrio

del componente i por Presin (Ki * P).
Debe aproximarse a una lnea recta.

Grfico de Hoffman
Grficos de Control
Bashbush (1981):
El mtodo consiste en la aplicacin de un

balance de masa a los moles del fluido original
presente en la celda a presin de saturacin.
El objetivo es obtener la composicin molar del

lquido, que junto con la composicin molar del
gas (cromatografa), nos permite calcular las
constantes de equilibrio para cada componente
en cada etapa de la liberacin.
Bashbush
Esquema de Trabajo
Balance de Masa:
Grficos de Control
Log K VS Tb
Hoffman
Bashbush

Chequeo de Parmetros
Entre todos los parmetros de un ensayo PVT existen

interdependencias.
Aunque estas interdependencias parezcan coherentes,

no es suficiente para indicar que los resultados
obtenidos sean representativos.
En esta etapa tiene mayor peso la experiencia del

analista.
Algunos ejemplos:
Valores anmalos de expansin trmica.

Curva Bo con concavidad errnea.
Curva de viscosidad (P > Pb) no muestra la pendiente
esperada.
Las constantes de equilibrio se entrecruzan.

Etc
Esquema de Trabajo
Balance de Masa:
Grficos de Control
Log K VS Tb
Hoffman
Bashbush

El objetivo es hallar un modelo que represente

el comportamiento PVT del fluido de reservorio.
Como vimos anteriormente, uno tiene dos

opciones para evaluar el modelo:
Utilizar correlaciones para los parmetros crticos del

pseudo-componente que representa a la fraccin
pesada.
Ajustar con los datos PVT el modelo variando los

parmetros crticos del pseudo-componente que
representa a la fraccin pesada
Luego del ajuste uno puede corroborar la

coherencia de los datos.

Ejemplo: Modelo Petrleo Volatil
Chequeo PVT
Hiptesis de Alta y Baja Volatilizacin
Objetivo
Es simplemente otro mtodo para poder chequear la
consistencia de un PVT.
Consiste en utilizar la composicin del Fluido de
Reservorio Recombinado y plantear dos hiptesis:
Alta Volatilidad
Baja Volatilidad
En cada caso se calcula el Bob y el Rsb.
El Bob y Rsb del informe PVT debe estar dentro del
intervalo obtenido:
Bob (min) < Bob (informe PVT) < Bob (max)
Rsb (min) < Rsb (informe PVT) < Rsb (max)
Hiptesis
Alta Volatilizacin: consideramos que los moles de
los componentes ms livianos (<C4) al bajar la
presin pasan todos a la fase gaseosa. Luego
progresivamente vamos repartiendo las fracciones
molares en cada fase hasta considerar que los
componentes pesados (>C10) estn solamente en la
fase lquida.
Baja Volatilizacin: consideramos que los moles de
los componentes ms livianos (<C2) al bajar la
presin pasan todos a la fase gaseosa. Luego
progresivamente vamos repartiendo las fracciones
molares en cada fase hasta considerar que los
componentes pesados (>C6) estn solamente en la
fase lquida.
Ejemplo de Hiptesis
Componente Fase Gas Fase Lquida
C1 100% 0%
C2 100% 0%
C3 80% 20%
C4 80% 20%
C5 60% 40%
C6 60% 40%
C7 20% 80%
C8 20% 80%
C9 0% 100%
C10 0% 100%
C11 0% 100%
Componente Fase Gas Fase Lquida
C1 100% 0%
C2 100% 0%
C3 60% 40%
C4 60% 40%
C5 20% 80%
C6 20% 80%
C7 0% 100%
C8 0% 100%
C9 0% 100%
C10 0% 100%
C11 0% 100%
Alta Volatilizacin Baja Volatilizacin
Ejemplo de TP
1. Planteo las Hiptesis
A B C D E
Comp
Fluido de
Reservorio
[%molar]
Gas Liq Gas Liq
C1 28 100% 0% 100% 0%
C2 10 100% 0% 100% 0%
C3 8 80% 20% 60% 40%
C4 8 80% 20% 60% 40%
C5 6 60% 40% 20% 80%
C6 6 60% 40% 20% 80%
C7 8 20% 80% 0% 100%
C8 8 20% 80% 0% 100%
C9 10 0% 100% 0% 100%
C10 10 0% 100% 0% 100%
Alta Volatilizacin Baja Volatilizacin
Ejemplo de TP
2. Calculo Moles de Lquido y Gas para cada Hiptesis
A B C D E A x C A x E A x B A x D
Comp
Fluido de
Reservorio
[%molar]
Gas Liq Gas Liq
Alta
[moles]
Baja
[moles]
Alta
[moles]
Baja
[moles]
C1 28 100% 0% 100% 0% 0,0 0,0 28,0 28,0
C2 10 100% 0% 100% 0% 0,0 0,0 10,0 10,0
C3 8 80% 20% 60% 40% 1,6 3,2 6,4 4,8
C4 8 80% 20% 60% 40% 1,6 3,2 6,4 4,8
C5 6 60% 40% 20% 80% 2,4 4,8 3,6 1,2
C6 6 60% 40% 20% 80% 2,4 4,8 3,6 1,2
C7 8 20% 80% 0% 100% 6,4 8,0 1,6 0,0
C8 8 20% 80% 0% 100% 6,4 8,0 1,6 0,0
C9 10 0% 100% 0% 100% 10,0 10,0 0,0 0,0
C10 10 0% 100% 0% 100% 10,0 10,0 0,0 0,0
Alta Volatilizacin Baja Volatilizacin Moles del Lquido
[mol]
Moles del Gas [mol]
Ejemplo de TP
A B C D E A x C A x E A x B A x D A x C A x E A x B A x D
Comp
Fluido de
Reservorio
[%molar]
Gas Liq Gas Liq
Alta
[moles]
Baja
[moles]
Alta
[moles]
Baja
[moles]
Alta
[moles]
Baja
[moles]
Alta
[moles]
Baja
[moles]
C1 28 100% 0% 100% 0% 0,0 0,0 28,0 28,0 0,00 0,00 0,46 0,56
C2 10 100% 0% 100% 0% 0,0 0,0 10,0 10,0 0,00 0,00 0,16 0,20
C3 8 80% 20% 60% 40% 1,6 3,2 6,4 4,8 0,04 0,06 0,10 0,10
C4 8 80% 20% 60% 40% 1,6 3,2 6,4 4,8 0,04 0,06 0,10 0,10
C5 6 60% 40% 20% 80% 2,4 4,8 3,6 1,2 0,06 0,09 0,06 0,02
C6 6 60% 40% 20% 80% 2,4 4,8 3,6 1,2 0,06 0,09 0,06 0,02
C7 8 20% 80% 0% 100% 6,4 8,0 1,6 0,0 0,16 0,15 0,03 0,00
C8 8 20% 80% 0% 100% 6,4 8,0 1,6 0,0 0,16 0,15 0,03 0,00
C9 10 0% 100% 0% 100% 10,0 10,0 0,0 0,0 0,25 0,19 0,00 0,00
C10 10 0% 100% 0% 100% 10,0 10,0 0,0 0,0 0,25 0,19 0,00 0,00
40,8 52,0 61,2 50,0
Composicin Lquido
[%molar]
Composicin Gas
[%molar]
Alta Volatilizacin Baja Volatilizacin Moles del Lquido
[mol]
Moles del Gas [mol]
3. Totalizo Moles de Lquido y Gas, calculo Composicin Molar de ambos
Ejemplo: %C2 lquido = Moles C2 lquido / Cn Moles Lquido
Clculo de Parmetros PVT
Podemos calcular los parmetros PVT Bo y Rs a travs de la composicin de los fluidos:
Tener cuidado con las unidades, siempre chequear consistencia.
Donde:
X: Fraccin molar total del lquido.
Y: Fraccin molar total del gas.

CS
liq
: Densidad del lquido de TNK a CS (utilizar composicin del lquido)

Pb
: Densidad del fluido en el punto de burbuja (utilizar composicin total)
PM
liq
: Peso Molecular del lquido de TNK (utilizar composicin del lquido)
PM
Fres
: Peso Molecular del fluido en el punto de burbuja (utilizar composicin total)
R: Constante universal de los gases (0.082 lt atm/ K mol)
Pcs: Presin a condicin estndar (1 atm)
Tcs: Temperatura a condicin estndar (60 F)
Utilizando las definiciones de los parmetros PVT:
Bo = Vol liq fondo (fluido de reservorio) / Vol liq CS (comp liquido)
Rs = Vol gas disuelto CS (comp gas) / Vol liq CS (comp liquido)
Podemos desarrollar las ecuaciones anteriores. Practicar el desarrollo de las mismas utilizando las
siguientes relaciones:
Vol liq = nro moles liq x PM liq / dens liq = nro moles x xi . PMi / xi . Densi
Vol gas = nro moles gas x R x T / P
Para cada hiptesis calculamos Bo y Rs, as obtenemos un intervalo. Los valores informados en el PVT
deberan estar dentro del mismo.
Pb
liq
CS
liq
Fres
PM
PM
X
Bo
1
=
CS
liq
CS
liq
CS
P
PM
RT
X
Y
Rs

Fluid Os 2012

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fluid Os 2012

Uploaded by

Copyright:

Available Formats

Fluidos de Reservorio

based on the rele-

and the liquid oil gradient is 47.37/144 = 0.329 psi/ft.

These values are listed in column 6 of table 2.2.

Gas FVF+1.424 Mscf/bbl.

+(2, 638 lbm mol)

379 scflbm mol

757.87 )96.68 )24.23 Mscflbm mol

379 scflbm mol

+0.8788 lbm mollbm mol, (6.43) . . . . . . . . . . . . . . .

757.87 )96.68 )24.23 Mscflbm mol

135.7 STBlbm mol

[(301.57 )20.75 )5.61) *(124.78 )12.09 )3.16)]

0.762(186 )460)2, 115

+2, 299 galMMscf (6.54a) . . . . . . . . . . . . . . . . . . . .

3,500 2,900 2,100 1,300 605 0

Stock-tank liquid, bbl 135.7 0 6.4 15.4 24.0 29.7 35.1

Propane, gal 1,198 0 67 204 513 910 1,276

Propane, gal 669 0 33 86 149 205 286

Propane, gal 2,296 0 121 342 750 1,229 1,706

Recovery assumes 100% plant efficiency.

Composicin del Fluido (o de las Fases)

Para componentes puros y conocidos, las mismas se

Las EOS fueron desarrolladas para

Esto se resuelve gracias al concepto de fugacidad

Se deben definir las propiedades crticas del

Esto puede ser realizado mediante:

Matchear las curvas PVT obtenidas en laboratorio

Conviene estimar las propiedades crticas con

*12, 272.6 )9, 486.4g )(4.6523 *3.3287g)T

Por qu es necesario realizarlo?:

Balance de Masa Global

Balance de Masa por Etapas

Balance de Masa Composicional

Chequeos Generales de Parmetros

Balance de Masa Global

Balance de Masa por Etapas

Balance de Masa Composicional

Chequeos Generales de Parmetros

Balance de Masa Global:

Chequea que la densidad de la muestra medida sea

Balance de Masa por Etapas:

dem anterior, excepto

Se va chequeando etapa por etapa de la liberacin.

En este chequeo se suman las masas

La masas total dividida por el volumen en fondo dede

Asumiendo que el volumen residual a condiciones de

La masa de gas en la isima de la liberacin diferencial

Por lo que, la masa total del sistema resulta:

Y el volumen total a presin de burbuja y temperatura de

Entonces la densidad a esas condiciones es:

Balance de Masa Composicional:

En cada etapa de la liberacin uno tiene el

Se puede calcular las constantes de equilibrio

Balance de Masa Global

Balance de Masa por Etapas

Balance de Masa Composicional

Chequeos Generales de Parmetros

Constantes de Equilibrio VS Temperatura

En una determinada etapa de la liberacin (ya

Luego se grafica en escala semi-log K vs Tb.

Grfico de Hoffman (1952):

Nuevamente para una determinada etapa de

Absisas: Funcin especial de Hoffman que

12, 272.6 )9, 486.4g )(4.6523 3.3287g)T