Properties of Oils and Natural Gases

r
(
(
(
(
(
Contributions in Petroleum Geology and Engineering
Series Editor: George V. Chilingar, University of Southern California
Volume 1: Geologic Analysis of Naturally Fractured Reservoirs
Volume 2: Applied Open-Hole Log Analysis
Volume 3: Underground Storage of Natural Gas
Volume 4: Gas Production Engineering
Volume 5: Properties of Oils and Natural Gases
Volume 6: Introduction to Petroleum Reservoir Analysis
Volume 7: Hydrocarbon Phase Behavior
Volume 8: Gas Reservoir Engineering
(
(
(
(
(
(
(
(
(
(
(
e
(
(
(
(
(
(
(
(
(
l
(
(
(
~g~Golf Publshng Company
e Book Division
Houston, London, Paris, Tokyo, Zurich
ii
\.
Contributions in Petroleum Geology & Engineering ci)5
Oilsand
r - - -----"'-""""- ' ,,'_. m~".~_
l I\.t;,:':nrm AiiEHHl'W !J Et p[;--' -a.-.
81bUOTEGtl
L~,N"':2:>",,=~J Ct... ~_.. ...J
'1
r
r !APG-Bulgheroni
; l i:lf!ill/!IIIIIII 11/11111/ I/!
622338 C6275 0009329
!:)'-'-~fi,:c. -", . 1: (""n ..-I ~~t('-~JI g3~-?S
~1'1", <:
p-tF'<~),\
,,"r'J -- \.~ ~oI ..
\1S' E~1I0iE{;., .~'i.
'"\--' i--1)
',\';::. ---- 1f
)\ 'rS, .,;,:..
. A.. ~
, ,("{iI./;7
<. 1: //
'."'~'"
,"
Contributions m Petroleum Geology and Engineering
Volume 5
Properties of Oils and Natural Gases
Copyright 1989 by Gulf Publishing Company, Houston,
Texas. Al! rights reserved. Printed in the United States of
...Ainerica. This book, or parts thereof, may not be
reproduced in any form without permission of the
publisher.
Library of Congress Cataloging-in-Publication Data
Pedersen, K. S. (Karen Schou)
Properties of oils and natural gases/K. S. Pedersen, A.
Fredenslund &P. Thomassen.
p. cm.-(Contributions in petroleum geology &
engineering; v. 5)
Ineludes index.
ISBN 0-87201-588-2
1. Petroleum-Analysis. 2. Gas, Natural-Analysis.
1. Fredenslund, Aage, 1941- . n. Thomassen, P. (Per) ,
1950- III. Title. IV. Series.
TP691.P42 1989
665.5-dcl9 88-24659
CIP
ISBN 0-87201-066-X (series)
Nomenclature
l' Ir-
. Iv~ -,"",-
ti
,1 ;:,
r r , \
Contents ~~:~ \i\\'\' i:.
\ "
............................................ ""~ir' ~".-;./
Introduction 1
1. Petroleum Reservoir Fluids 4
Petroleum Reservoir Fluid Constituents, 4; Classification of Petroleum
Reservoir Fluids, 4; Hydrocarbon Fractions, 8; Residue, 9.
2. Compositional Determinations 10
Gas Chromatography and TBP Distillation, 10; Properties of Hydro-
carbon Fractons, 22; Internal Consistency of Analytical Data, 30; Ex-
perimental Extension beyond C20+, 35.
3. Oil and Gas Property Measurements 39
Sampling, 39; Flash Separation and Compositional Analysis, 39; PVT
Measurements, 40; Black Oil Equipment, 40; Gas Condensate Equip-
ment, 54; Volatile Fluid Equipment, 60; Measurement of Viscosity, 63;
Surface Tension, 66.
4. Composition and Property Data 69
5. Equations of State 79
Types of Equations of State, 79; Cubic Equations of State, 81; The
SRK-EquatioiJ of State, 82; Phase Densities, 89.
6. Flash Calculations 99
Two-Phase (P,T)-Flash, 100; Multi-Phase (P,T)-Flash, 104; (P,H) or
(P,S)-Flash Calculations, 105; Three-Phase (P,T)-Flash with a Liquid
Water Phase, 108; Simplified Two-Phase (P,T)-Flash, 109; Phase Enve-
lope Calculations, 112.
7. Characterization Procedures 114
Classification of the Constituents of Oil and Gas Mxtures, 114; TBP
Fractions, 114; The TBP Residue, 116; Characterization Procedures,
120; Grouping of Pseudocomponents, 124; Calculation of the Ideal Gas
Heat Capacity, 126.
8. Simulation of PVT-Experiments 130
Constant Mass Expansion, 130; Differential Liberation, 131; Constant
Volume Depletion, 135; Separator Test, 140; Summary of Results of the
PVT-Simulations, 140.
\
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(,
(
(
(
(
(,
(
(
G
C.
\,
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
9. Comparison Between Experimental and Predicted
Thermodynamic Properties 141
Dew and Bubble Points (Phase Envelopes), 141; Flash Results, 144;
Densities, 144; Enthalpy, 151; Summary of the Results of Thermody-
namic Properties, 154.
10. Tuning of EOS-Parameters 155
The Needsfor Tuning, 155; Parameters Availablefor Tuning, 155; Ex-
amples on the Dangers of Tuning, 156; The Molecular Weight of the
Plus-Fraction as a Tuning Parameter, 166; Recommendations Regard-
ing Tuning, 169.
1l. Viscosity 172
Definition of Viscosity,172; Kinetic GasTheory, 172; ViscosityCorrela-
tions, 173; TheViscosityCorrelation of Lohrenz et al., 173; Calculation
of the ViscosityUsingCorresponding States Theory, 175; Comparison
with Experimental Results, 181.
12. Thermal Conductivity 187
Kinetic GasTheory, 187; Calculation of theThermal Conductivity Us-
ing Corresponding States Theory, 188; Comparison with Experimental
Observations, 194.
13. Surface Tension 196
Estimation of the Surface TensionUsingthe Macleod-Sugden Correla-
tion, 196; Estimation of the Surface Tension from the Viscosity, 199;
Comparison with Experimental Observations, 202.
14. Wax Formation and Inhibition 208
WaxComposition, 208; Computation of WaxFormation, 208; WaxIn-
hibitors, 214; Comparisons Between Calculated and Experimental Re-
sults, 217.
15. Gas Hydrates 220
Types of Hydrates, 220; Hydtate Computation, 223; Hydrate Inhibi-
tors, 228; Comparison with Experimental Observations, 230.
\.
16. Simulation of Miscible Gas Injection 234
ProcessDescription, 234; Calculation of theMinimumMiscibility Pres-
sure (MMP), 238; Cell-to-Cell Simulation, 241; Comparison of Calcu-
lated and Experimentally Determined MMP's, 246.
Index 249
vi
a:
a:
a-a2:
al-aS:
A:
A:
A:
A:
A:
A:
A,,:
Ao-A
2
:
Ao-A
3
:
API:
b:
b:
b-b4:
B:
B:
B:
B:
Ba:
Ba-B:
Ba-B4:
B':
c:
C-C6:
C-Cg:
C:
C:
C:
C:
CF:
CN:
e;
e,
C:
C2:
Nomenclature
Equation of state parameter
Constant defined in Equation 15;18
Constants defined in Equations 5-54 and 5-55
Constants defined in Equation 11-6
Aromatics (Chapter 1)
Constant defined in Equation 5-23
Surface area (Chapter 13)
Parameters defined in Equations 5-51 through 5-53
Coefficients defined in Equation 5-86
API gravity (141.5/SG-131.5)
Equation of state parameter
Constantdefined in Equation 7-13
Shrinkage factor (defined in Equation 3-2)
Parameters defined in Equations 5-57 and 5-58
Peneloux parameter defined in Equation 5-39
Constants defined in Equations 13-16 and 13-17
Critical point (Chapter 1)
Density correlation factor defined in Equation 5-86
Adsorption coefficient (Chapter 15)
Function defined in Equation 7-47
Carbon number
Molar heat capacity at constant pressure
Molar heat capacity at constant volume
Density correlation factor at standard conditions (Chapter 5)
Density correlation factor at actual conditions (Chapter 5)
vii
CO-C3:
.Parameters defined in Equations 5-64 through 5-67 K: Watson Characterization factor (Chapter 7)
C-C4:
Coefficients of ideal gas heat capacity polynomial Equation K: Constant defined in Equation 15-21
5-36 and Equations 7-43 through 7-46 1-17: Constants defined in Equation 12-18
C7+:
Hydrocarbons with 7 and more carbon atoms L: Length of capillary tube Equation 3-10 (
D:
Diameter (Chapter 3)
L: Mean free pass between two molecules
(
DL:
Differential liberation
m: Function of the acentric factor defined in Equation 5-16
e-e3:
Functions defined in Equations 6-54 and 6-56 m: Constant defined in Equation 5-46
e-e6:
Parameters defined in Equations 5-79 through 5-84 m: Number of pseudocomponents
EOS:
Equation of State
M: Molecular weight of inhibitor divided by molecular
Eo-E3:
Parameters defined in Equations 5-75 through 5-78 weight of water
E : Fugacity
MMP: Minimum miscibility pressure
(
E :
Function defined in Equation 12-12 Mr: r'th statistical moment defined in Equation 7-40
E :
Function defined in Equation 12-36 MW: Molecular weight
Ej:
Function defined in Equation 12-24 n: Refractive index (Chapter 7)
(
fx,o:
Function defined in Equation 12-22 n: Number of moles Equation 3-5 and Chapter 6
FA:
Function defined in Equation 12-21 (
n: Constant defined in Equation 5-45
(
F:
Distribution function (Chapter 7)
,
Number of discrete components (Chapter 7) n:
(
FCMP:
First contact minimum miscibility pressure N: Naphthenes (Chapter 1)
F:
N: Stage number (Chapter 3)
(
F2:
N: Number of components (Chapter 6)
g:
Gravitational acceleration
NCOMP: Number of components
g:
NSTEP: Total number of flash stages
g2:
Function defined in Equation 6-28 ' P: Paraffins (Chapter 2)
G:
Gibbs energy
P: Probability function defined in Equation 7-17
GOR: Gas/oil ratio
[PJ : Parachor (Chapter 13)
h.:
PVT: Pressure, volume, temperature
hj:
Functon defined in Equation 12-25
q,,: Function defined in Equation 6-46
hx.o:
Function defined in Equation 12-23 q,s: Function defined in Equation 6-47
H:
qo:
Functian defined in Equation 6-45
H:
Molar enthalpy
q: Heat Ilow per unit area (Chapter 12)
H:
r: Crystal radius (Chapter 14)
HTAN:
r: Distance from lattice wall (Chapter 15)
1:
Characterization Factor defined in Equations 7-21 and 7-22 r*: Critical radius (Chapter 14)
1:
Variable used to define continuous distribution (Chapter 7) R: Gas constant
Ip:
Quadrature point (Chapter 7)
R: Parameter defined in Equation 7-27
i-h:
Rs: Solution gas/oil ratio (defined in Equation 3-3)
J : Phase number (Chapter 6)
s: Number of quadrature points (Chapter 7)
k:
s: Mole fraction in salid phase
k:
Boltzmann's constant (Chapter 14)
. s: Inhibitor concentration (Chapter 15)
kj:
Binary interaction coefficient
S: Standard Conditions (normally 15C and 1atm)
k-k7:
Canstants defined in Equation 11-40
S: Entropy
K:
Equilibrium ratio (vapor mole fraction/liquid mole fraction) SG: Specific Gravity
K:
SRK: Soave-Redlich -Kwong
viii ix
' r
,/';.-.--.
:1/
.... ,
I
I l .
T: Temperature
17: Viscosity
;~'
\ ~ ,
T: Average temperature defined in Equation 15-8
17:
Parameter entering into Equation 7-17 .\;:;- =' r
T*: Transition temperature defined in Chapter 6 17:
Mole fraction of "continuous components" (Chapter 7) '.<'~ I
TBP: True boiling point
. 15 " .
170:
"~. ,//
v: Velocity 171:
""~
v: Parameter defined in Equation 7-23
1117': Function defined in Equation 11-37
v: Volume fraction
8: Shape factor (Chapter 11)
V:
SRK volume defined in Equation 5-41
8:
Parameter defined in Equation 11-39
w: Parameter defined in Equation 7-24
,,:
Thermal conductivity
w: Weight fraction (or weight percent) "o:
w: Weighting function (Chapter 7)
"1:
w: Cell potential (Chapter 15)
,,/:
x: Liquid mole fraction
,,1/ :
(
y: Vapor mole fraction
p. : Chemical potential
y: Variable defined in Equation 6-10
Vi:
Number of cavities of type i (Chapter 15)
( YKi: Probability that a cavity of type i is occupied by gas
~:
Molar vapor fraction (Chapter 7)
(
molecules of type K
~:
Viscosity reducing parameter (Chapter 11)
z: Mole fraction
p: Density
( Z: Compressibility factor
p: Molar density
a: Surface tension
a:
Molecular hard sphere diameter (Chapter 11)
(
CREEK LETTERS
T*: Parameter defined in Equation 6-32
(
T: Shear stress
o' : Parameter defined in Equation 6-43
<1>: Fugacity coefficient
\ a: Parameter entering into Equation 7-17
<1>: Volume fraction (Chapter 14)
(
a: Parameter defined in Equation 11-34
,1
X:
Binary interaction coefficient (Chapter 14)
a: Parameter defined in Equation 16-1
w: Acentric factor
l ac: Riedel parameter defined in Equation 13-4
(
{3: Vapor mole fraction
,1
{3: Parameter entering into Equation 7-17
SUB- AND SUPERSCRIPTS
{3: Constant defined in Equation 13-5
')': Activity coefficient (Chapters 14 and 15)
attr: Attractive
r: Camma function (Chapter 7)
a: Pure water state (Chapter 15)
\
o: Solubility parameter (Chapter 14)
B: Boiling point
~: Step size (Chapter 6)
Br: Reduced boiling point
11"c: Function defined in Equation 12-16
{3:
Empty hydrate lattice (Chapter 15)
11" ': Function defined in Equation 12-16
c: Critical state
11" 1/: Function defined in Equation 12-17
e: Equatorial
1117: Function defined in Equation 13-18
F: Feed (Chapter 5)
1117': Function defined in Equation 11-38
F: Freezing point (Chapter 11)
1117 1/: Function defined in Equation 11-40
C: Caseous phase
1:: Number of moles in the second phase (Chapter 6)
H: Hydrate phase (Chapter 15)
j: Parameter defined in Equation 6-27
i: Component index
x
xi
1:
11:
id:
int:
j:
k:
L:
mix:
M:
N:
o:
o:
p:
pure:
r:
R:
RA:
ref:
rep:
res:
res:
S:
spec:
a:
tr:
V:
vap:
w:
wt%:
x:
y :
..
Phase number
Phase number
Ideal
Internal
Component index
Component index
Liquid
Mixture
Methane
Component identification number (Chapter 16)
Reference state (Chapter 6)
Reference component (Chapters 11 and 12)
Separator conditions (Chapter 2)
Pure component
Reduced state (property at actual conditions/property at critical
conditions)
Reference conditions (normally 15C and 1 atm)
Racket (Equation 5-43)
Reference state
Repulsive
Reservoir conditions
Residual (Chapter 5)
Solid phase (Chapter 14)
Specified value
Surface tension
Translational
Vapor
Vapor
Water
Weight percent
X-direction
Y-direction
Pseudo state (Chapter 5)
Dilute gas (Chapter 11)
xii
ti
I
I
,1
\
1
Introduction
Traditionally, chemical and petroleum engineers engaged in production
of oil and natural gas from reservoirs have not devoted much interest to de-
tails regarding the properties and compositions of the hydrocarbon fluids
produced. Design of production schemes and equipment was mainly based
,on experience. This was, however, insufficient considering the enormous in-
vestments required for oil and gas exploration and production. Today pro-
duction systems and equipment undergo detailed engineering analyses and
careful design 1:0 optimize the amount of hydrocarbons produced and mini-
mize the operating costs and investment.
Analysis and design of oil and gas production facilities and schemes re-
quire thorough knowledge of the thermodynamic and physical properties of
hydrocarbon fluids. To know how much gas and how much oi~arator
produces under given conditions, it is necessary to know the vapor-liquid
equilibrium relationships and the_d~~sities of the mixture. Understanding
.the flow processes in the reservoir requires knowledge of viscosity and SUT-
face tension. Heat exchanger design requires knowledge of the thermal con-
ductivity of the flowing fluids. Other examples abound.
It would, naturally, be best if knowledge of the above mentioned proper-
ties was available from experimental observations. Those are, however, im-
possible to measure for all hydrocarbon fluids in all relevant conditions. The
chief objective of this book is, therefore, to describe in detail how to use
mathematical models to accurately predict transport properties (viscosity
and thermal conductivity), surface tensions, and thermodynamic properties
(density, enthalpy and phase equilibria) of naturally occurring oil and gas
mixtures. The book isintended for the design (chemical or petroleum) engi-
, neer in the oil and gas related industry.
The models described in this book are the results of a collaborative re-
, search and development project executed by three parties: Statoil a.s., the
Norwegian state oil company, Calsep A/S, a Danish engineering consulting
company, and the Department of Chemical Engineering at the Technical
, Universty of Denmark. The research project was carried out on three levels.
The basic level concerns the description of the oil or gas. Detailed compo-
sitional analyses of Statoil's North Sea oils were performed, including chro-
matography, TBP-distillation, and others. The analyses were checked for in-
ternal consistency by, for example, computing and measuring the average
1
2 Properties of Oils and Natural Gases
(
rnolecular weight of the residue. The end result of this basic level research
was a highly detailed and-cornpared with earlier work-rnuch irnproved
analytical description of the hydrocarbon rnixtures.
The second level of research concerns the developrnent of irnproved corre-
lations of Te, P and w (critical ternperature, pressure and acentric factor)
for the hydrocarbon fractions as a function of the rnolecular weight and
other properties. In addition, procedures for estirnating ("characterizing")
the cornposition of the C20+ residue were developed.
The third level concerns choice of equation of state, viscosity rnodel, etc.
by cornparing rneasured and predicted properties. This level of research in-
eludes further developrnent of the rnodels by, for exarnple, irnprovernent o
the mixing rules. It also ineludes application of the rnodels to sirnulate rnac-
roscopic processes such asflash operations, wax forrnation, and reservoir oil!
gas miscibility.
There isagreat deal of interplay between the three levels of research. The
availability of correct and detailed cornpositional analyses is a necessary
condition for developing meaningful and sound improvements in the corre-
lations of Te, Pe, and w in the equations of state. Sound models for Te, Pe,
and w are needed to furnish good predictions of densities and dew points
usng an equation of state. Improvements and adjustments continued on all
three levels throughout the research project.
Because the models presented in this book are based on the most detailed
and correct analytical information possible, the predictive capacity of the
models isvery high. Although the models were developed almost exelusively
on the basis of data for North Sea reservoir fluids, they may also be applied
with confidence to petroleum reservoir fluids from other regions. It is also
important to realize that, although highly detailed analytical data were re-
quired to develop the models, it is notnecessary to have th same degree of
detail available in order to use the models.
The book isdivided into four parts. The first part (up to Chapter 4) intro-
duces and describes oils and gases. Different types of oils are described in
Chapter 1, and laboratory methods for compositional and property mea-
surements are given in Chapters 2and 3. Chapter 4lists compositional data
for several hydrocarbon mixtures, and extensive data ineluding densities,
dew points, viscosities, and others are given.
The second part (Chapters 5-10) describes the modeling of the thermody-
namic properties such as densities, dew points, and flash separations. From
the many different models available, the Soave-Redlch-Kwong equation of
state was chosen asthe basic model for these properties. It isshown in Chap-
ters 8and 9that this model (when modified slightly using the volume trans-
lation concept) gives results that are in excellent agreement with experimen-
tal values. The choice of the Soave-Redlich-Kwong equation of state is,
however, somewhat arbitrary. Equally good results could have been ob-
(
,j :
i:
I
1
1
i
Introduction 3
tained with, for example, the Peng-Robinson equation of state. It should be
emphasized, however, that the characterization procedure of Pedersen et al.
described in Chapter 7 in detail, is tied to the chosen equation of state, al-
though the principIes may be applied with any equation of state. The flash
calculation procedure described in Chapter 6 can be used in conjunction
with any equation of state.
The third part of the book (Chapters 11-13) describes newly developed
correlations for predicting the transport properties (viscosity and thermal
conductivity) and surface tension of hydrocarbon mixtures. The transport
property models are based on the principIe of corresponding states, and may
be applied to both gaseous and liquid phases. One of the surface tension cor-
relations is elosely tied to the viscosity correlation.
The last part of the book (Chapters 14-16) describes some novel, impor-
tant applications of the models developed in this work: wax formation in
pipelines, gas hydrate formation, and studies of miscible gas injection pro-
cesses.
The intent has been as complete a presentation as possible. Thus, many
detailed experimental data are listed, and all details regarding the devel-
oped models are given so that they may be readily applied, based on the
material given in this book.
Chapter 1
Petroleum Reservoir Fluids
PETROLEUM RESERVOIR FLUID CONSTITUENTS
Petroleum reservoir fluids are multi-component mixtures, consisting pri-
marily of hydrocarbons belonging to one of the following three classes:
Paraffns (or alkanes) consist of chains of hydrocarbon segments (e.g., -CH
2
or -CH3). The molecules may be unbranched (normal paraffins), or
branched (iso-paraffins). The atoms are connected by single bonds. Meth-
ane (CH4) is the simplest paraffin and it is also, on molar basis, the most
common constituent of petroleum reservoir fluids.
Naphthenes (or cycloalkanes) are hydrocarbons similar to the paraffins, but
differing from these by containing one or more cyclic structures. The ele-
ments of the cyclic structure (e.g., -CH2) are connected by single bonds,
and usually six carbon atoms are joined in a ring structure.
Aromatcs are hydrocarbons containing one or more ring structures similar
to benzene (C6H6). The atoms are connected by aromatic double bonds.
Polycyclic aromatic compounds (or asphalthenes), like naphthalene
(CIOH8) and anthracene (CI4HIO), may also be presento
In addition to hydrocarbons, the non-hydrocarbons nitrogen (N
2
), carbon
dioxide (C02) , and hydrogen sulfide (H2S) are often found in petroleum
mixtures. Finally, petroleum mixtures may contain helium, mercury, and
metal-organic compounds. These will not be covered in this work. Table 1-1
shows some physical properties of common petroleum reservoir fluid con-
stituents. The molecular structures of some of these components are shown
in Figure 1-1.
CLASSIFICATION OF PETROLEUM RESERVOIR FLUIDS
Petroleum reservoir fluids may beclassified as gases, gas condensates, vol-
atile oils, or black oils.
l
;
f~
.~
t
~
"
Petroleum Reservoir Fluids 5
Table 1-1
Physical Properties of Some Common Petroleum Reservoir Fluid
Constituents
1 Density (g/cm
3
)
Melting Point Normal Boiling at 1 atm and
eomponent Formula (Oe) Point (Oe) 15e
Paraffins
Methane CH. - 184 - 161.5 -
Ethane C2Ha - 172 - 88.3 -
Propane C3Ha - 189.9 - 42.2 -
n-Butane C.HlO - 135 - 0.6 -
Iso-Butane C.HlO - 145 - 10.2 -
n-Pentane
CSH'2
- 131.5 36.2 0.626
n-Hexane
CaH'4
- 94.3 69.0 0.659
Iso-octane CaH,a - 107.4 99.3 0.692
n-Decane C,oH22 - 30 174.0 0.730
Naphthenes
Cyclopentane CSHlO - 93.3 49.5 0.745
Methyl cyclo-
pentane
CaH'2
- 142.4 71.8 0.754
Cyclohexane
CaH'2
6.5 81.4 0.779
Aromatics
Benzene CaHa 5.51 80.1 0.885
Toiuene C7Ha - 95 110.6 0.867
o-Xylene
CaH'0
- 29 144.4 0.880
Naphthalene C,oHa 80.2 217.9 0.971
(
,!,
~.
Table 1-2shows typical molar compositions of mixtures belonging to each
of the four classes. For each composition, the last hydrocarbon is marked
with a " +". It indicates that heavier components are contained in this frac-
tion also. In addition to the composition, the state of a fluid is also deter-
mined by the temperature and pressure.
The differences in the molar compositions between each of the four
classes of reservoir fluids are naturally reflected in differences in the fluid
.properties. Figure 1-2shows schematically typical phase envelopes for each
class. Each phase envelope corresponds to a fixed overall composition. At
the temperatures and pressures inside the phase envelope, two phases (usu-
6
Properties 01Oils and Natural Gases
P ARAF F I NS :
H
I
H-C-H
H I H
, . I
H- C- C- C- H
I I I
H H H
I s o- Bu t a n e
H H H H H H H H H H
"""""
H- C- C- C- C- C- C- C- C- C- C- H
I I I I I I I I I l'
H H H H H H H H H H
n - De c a n e
H
I
H- C- H
I
H
Me t h a n e
NAP HT HE NE S :
H
I
H-C-H
H" t
H, / \~
, C C' H
HJ -!-H
/ \ . \
H H H
Me t h y l
Cy c l ope n t a n e
AROMAT l CS :
H
I
C
H- C' ~C- H
11 I
H- C" ' t
C
-
H
I
H
H / H
'c
H-./,~
H" C C\ j
H--' I H
H}' /(.H
C
H. . . . . . .' H
H H
I I
~C, _ / C~
H-C '\: ' C- H
I 11 I
H- C~ /t , t : C- H
~C \:7'
I I
H H
Cy c l oh e xa n e
Be n z e n e Na ph t h a l e n e
Figure 1-1. Examples of hydrocarbon molecular structures.
e
T E MP E RAT URE -
Figure 1-2. Phase envelopes for reservoir fluids. e is the critical point.
Table 1-2
Typical Molar Compositlons of Petroleum Reservolr Fluids
Gas Volatile Black
Component Gas Condensate 011 Oil
N2 0.3 0.71 1.67 0.67
e 0 2 1.1 8.65 2.18 2.11
e , 90.0 70.86 60.51 34.93
e 2 4.9 8.53 7.52 7.00
e 3 1.9 4.95 4.74 7.82
e . (+n) 1.1 2.00 4.12 5.48
e s (i+n)
0.4 0.81 2.97 3.80
C, (+n) 6+: 0.3
0.46 1.99 3.04
e 7 0.61 2.45 4.39
e a 0.71 2.41 4.71
e 9 0.39 1.69 3.21
e l O 0.28 1.42 1.79
e l l 0.20 1.02 1.72
e ' 2 0.15 12+: 5.31 1.74
c., 0.11 1.74
c., 0.10 1.35
c., 0.07 1.34
c., 0.05 1.06
e l ? 17+: 0.37 1.02
c.,
. 1.00
c., 0.90
e 20 20+: 9.18
f~
~
~
ally gas and liquid) are found. Outside the phase envelope only one phase
(usually gas or liquid) exists. Coming from the single phase region, the phase
envelope represents the conditions where the first trace of a second phase
may be observed. The phase envelope consists of abubble-point branch and
a dew-point branch separated by the critical point, C. The bubble-pont
branch isthe one at the lower temperature and higher pressure. At the bub-
ble-point branch the newly formed phase can be seen as agas bubble. When
reaching the dew-point branch from the single-phase region the first sign of
a second phase will be the formation of a liquid drop. The bubble point
branch lies between the two-phase region and the single-phase liquid re-
gion. Similarly, the dew-pont branch is bounded by the two-phase region
and the single-phase gas region. At the critical point, two identical phases
are in equilibrium. In the near critical two-phase region, the two phases in
equilibrium are almost identical. In the near critical single phase region it is
hard to tell whether the phase is a gas or a liquid, and the expression su-
percritical Huid is often used. The point in Figure 1-2 marked by Tres, Pros
't
...
~.
represents typical reservoir conditions. At those conditions the gas is single
phase gas, and soisthe gas condensate. The volatile oil and the black oil are
both single phase liquids. If at constant temperature the pressure is reduced
as indicated by averticalline in Figure 1-2the gas will remain single phase
gas. A liquid phase will at some stage precipitate from the gas phase, but the
liquid phase will disappear if the pressure is further reduced. Precipitation
of a liquid phase from a gaseous mixture as a resuIt of a reduced pressure is
called retrograde condensation. The pressure interval where two phases ex-
ist is called the retrograde region. When the pressure is reduced, a second
phase is also formed at some stage from avolatile oil, and from a black oil.
The second phase isagas phase, and contains prmarily lighter components.
In a black oil some of the mixture components are so heavy that they will
remain in a liquid form even at very low pressures.
The highest temperature at the phase envelope of agiven mixture iscalled
the cricondentherm, and the highest pressure is called the cricondenbar.
HYDROCARBON FRACTIONS
While the lighter constituents of petroleum reservoir fluids are well-de-
fined individual components, the heavier (C7 +) hydrocarbon fractions are
mixtures of different components. Each carbon number fraction represents
the hydrocarbon mixture having its normal boiling point within a given
temperature interval; the carbon number cIassification isjust used for con-
venience. As is seen from Table 1-1, hydrocarbons of a different chemcal
structure but with the same number of carbon atoms may have very differ-
ent normal boiling points. A carbon number fraction in Table 1-1 may,
therefore, contain hydrocarbons with varying numbers of carbon atoms.
From molar compositions likethose given inTable 1-2it isnot possible to tell
whether the components of a given C7+-fraction are mainly paraffinic (P),
naphthenic (N), or aromatic (A) components. As described in Chapter 2,
components <ClO may be quantitatively identified. A PNA-analysis may
give a rough idea of the dominant molecular structures within the heavier
hydrocarbon fractions. Aromatics have a lower molecular weight than the
paraffins and the naphthenes with the same number of carbon atoms. Paraf-
fins and naphthenes on the other hand have a lower density than the aro-
matics with the same number of carbon atoms. Measurements of the aver-
age molecular weight and the density of each hydrocarbon fraction are,
therefore, of great significance in the characterization procedure (seeChap-
ter 7), which isperformed in order to simulate the propertes of a given pe-
troleum reservoir fluido
RESIDUE
The plus fractions of the petroleum mixtures of Table 1-2 consist of the
hydrocarbons with anormal boilng point above acertain temperature. The
usual technique for separating heavy hydrocarbon fractions is a TBP-distil-
lation as described in Chapter 2. The distillation process cannot always be
continued untillOO % of the petroleum mixture isdistilled off. At some stage
in the process, the temperature may have risen to a leve! where a further
temperature rise will cause cracking of the remaining molecules. The tern-
perature where the distillation process isstopped depends on the distillation
pressure and the sample quantity. This is reflected in the examples on com-
positional data given in Table 1-2where the last hydrocarbon number frac-
tion vares from C
5
to C
19
Components as heavy as Cso may be present in
the TBP-resid~e at a concentration that influences the overall behavior of
the reservoir fluido The average molecular weight and the density of the
TBP-residue may be measured, and this isimportant information used in the
characterization procedures described in Chapter 7.
I
I
}
t'
i
1
Chapter 2
Compositional Determinations
For oil and gas mixtures, the phase behavior and physical properties such
as densities, viscosities, and enthalpies are uniquely determined by the state
of the system, i.e., the temperature, pressure, and the composition. In sirnu-
lating the phase behavior and physical properties of complex hydrocarbon
mixtures accurately, it is necessary to have detai!ed and accurate cornposi-
tional information for each mixture.
GAS CHROMATOGRAPHY AND TBP DISTILLATION
Traditionally, compositional data in the oil industry have only been re-
ported to C7 +, the compositional information being mainly based on low
temperature fractionate distillation data. This level is inadequate for accu-
rate modeling of the phase equilibrium and physical properties of the hydro-
carbon mixtures.
In recent years, new methods have been developed for experimentally de-
termining the composition of hydrocarbon mixtures. These methods yield a
far more accurate and detailed description of the hydrocarbon systems and
are described in this chapter.
Gas Chromatography
The appearance of capillary columns has enabled separation and quanti-
fication of many more individual components than prevously possible.
The compositional description of the reservoir fluid is done by analyzing
separately samples of the gas and liquid phases, which when recombined in
the correct gas to liquid ratio will yield the reservoir fluid composition. The
compositional description of the gas phase is carried out in one step: N2,
CO2, CI-CIO by capillary column chromatography. The liquid phase com-
positional description is carried out in two steps:
1. To C
IO
+ using capillary columnchromatography.
2. From ClO to C20+ using a mini distillation apparatus.
10
Compositional Determinations 11
Gas Phase Analysis
A typical natural gas chromatogram is shown in Figure 2-1, obtained us-
ing a Hewlett Packard 5880 gas chromatograph (Figure 2-2).
o
u o
t-- ; :
-, -;
~ ~
~
. . J Z
" t--
Z t--
" "
;
o
U
t--
A
6
5 7 8
11
12
16
18
~
:t
U
t--
~
V>
88
!I
'1 22
1
1 17 19 23
I ~33 i: 14 38
l' I 32 \
"WI.. "U o ",LL~_
MN ~N
Z
t--
t--
"
Z
t--
t--
"
e
Figure 2-1. Typical chromatogram
and column configuration. Peak
numbers correspond to calibration
numbers given in Table 2-1. The an-
alytical conditions for the organic
and inorganic analyses are given in
Tables 2-2 and 2-3, respectively.
From Osj ord and Malthe-S0rens-
sen, 1983, reprinted from the Jour-
nal of Chromatography.
A --1 PORAPAK R HMOLSIEVE13X~
B
e -j SPLlT 1: 100 HWCOT CP SIL 51--8
The chromatographic separation of oxygen, nitrogen, carbon dioxide,

and approximately 60hydrocarbons from C to CIO+was achieved by using
a combination of packed and capillary columns. The sample was injected
va two time-programmed loops into packed and capillary columns with
thermal conductivity and flame-ionization detection. The packed columns
were molecular sieve 13X, Porapak R, and the capillary columns were
Chrompack Sil 5 (fused silica). The hydrocarbons were analyzed by split in-
jection using the capillary column, and temperature programming starting
at 30C, and the permanent gases were analyzed isothermally at 50C us-
ing the packed column in a special compartment outside the main oyen.
I
I
I
t
12 Propertes 01 Ols and Natural Gases
------- -- ~-~~_'1'::,,;J :iit'*
Figure 2-2. Hewlett Packard 5880A chromatograph. M.ulti-column rigged for one-
step gas analysis.
Through an elaborate calibration program, most of the individual com-
ponents have been identified. Table 2-1shows a resulting list of components
and the quantity of each component (mole fraction) present in the sample
corresponding to Figure 2-1. Details about the analyses are given in Tables
2-2 and 2-3.
This information is usually too detailed for practical purposes, so the
compositional information is regrouped. In the present sample, the re-
grouped gas phase analysis can be either that shown in Table 2-4 or Table 2-
5. Both examples of regrouping will yield satisfactory descriptions of the gas
phase composition.
Liquid Phase Analysis
The liquid hydrocarbon phase, i.e., the oil, is amuch more complex mix-
ture than is the gas phase. Many more components appear in a typical gas
chromatogram of an oil asshown in Figure 2-3 (Osjord et al., 1985). Consid-
erable overlapping of the peaks at the higher carbon numbers occurs, which
makes good quantitative determination of the heavercomponents impossi-
ble from the gas chromatogram alone.
I
UI
CII
UI
ltl
>-
iii
e
..,;:
CID
'O
CII
el
ni
..
CII
~
a
Q ..
";" E
C\I ni
CIIcn
:a UI
! ! ! ~
iii
..
:l
-
ni
Z
ni
-
o
t::
o
i
o
a.
E
o
o
~
s:
CI
~
Wf " - . . ooc n OT " " " ( \ J
NNNNC' )C' )C' )
Compostonal Determnations
VCOC\ J COCDCOO( ( )VOf " - . . C\ J O' l
" , , , , t r i C\ i t r i u : i c xi " " ' : " " ' : 6t r i , , , C\ i
CX: >CO " " , , , , , , , " " " ' T " " "
~N
I , {)l OVNI DO>LOVC\ J O' >O>O>I ' -
<OC' )OOC' )C' )NvNOl NNC' )
10 0 0 0 vOOooooo~
0 0 0 0 0 0 0 0 0 0 0 0 0
6666666666666
Q) Q)
e e
.>. ~-
...c:..c ..c:-
Q5Q) .Q~
E E u u
~ al 'C "C ~ >.
c t s cf7r;= <00)--5
X~V( ' f ) c C~Q)
~ ~ .~ .~ ~ ~ Q 5 .~ ~ ~
.2o.U( ) Q)Q)E O<l l Q)
g.-ciC}l~ ==o+g>.Q )~
u >.(1) CI) Q) >.>."" CI)-X e c o
>' oc : : c : c f ; : - S , c : c oQQl c
...c:>.~ ~ Q) ID Q).~ ~ E +>.0
Q . r : ~- . : ! : 2: 2( )- oxz
:2ii,.:,.:2~dJ ":~-z E:6 C:
C" ? LOCO, . . . . COO>O. , - NI . OI ' - COO)
C' )C' )C' )C' )C' )C' )vvvvvvv
v
oi
N
co
O
Cl
O
'"
Q)
e
<I l
U
Q)
-O
-O
Q)
;;::
~
Q)
-O
'c
::J
-
I ~ ~
NNNOl VONOl vOOOC' )l O COCOT " " " l OLONl . O
ci
0 0 6000~NMM
t r i u : i u : i u : i u : i u : i u : i u : i c xi u : i c r i u : i t r i
C' )
ui
~
ci
-s.
I N O
COWO><OOl Vf ' . . T " " " Ol l ONvC' )N VNT " " " ' COl O- c : t LO
~
VV T " " " ( \ J T " " " T " " " ( \ J " " ' ' ' ' ' ' ' ' OLOO>r - OOO OOC' )OvONV
10 0 COC" )CO( OT " " " OW( O OC' )Nv<OC' ) tDOOtDO<Dr-
CI <OC' ) 10 1OC' )ONOI ' - <O OOOT " " " O" - 0 0 0 0 0 0 0
~
~N 6 t r i C\ i C\ i" , 666 666666 6666666
<O~~
r:i~
c ef-
ci-
CD<Or - LOt OOCO
l l ' i u i Lr i LLt r i c r j
10 10 10 0 0 VO
vNN<O<OOl Ol
OQT " " " QOOC\ J
0 0 0 0 0 0 0
6666666
'E
GI
e
o
Q.
E
o
O
Q) Q)
e e
Q) !9~a5
e Ca . a .
. >. x Q) . - . -
-S -5 Q) .S9 9 el>
Q)Q)J : : OC' )Ne
EE>-' ,f/.cu
oo- E ~c : c : i 5.
I 1 Q )._ ~~Q )
( t )~: 2: 0 _ " " " J :
~~I~~~I
C\ I ~( " , )~~~C: :
e
o
:; .
. .o
,ez
i
O
'E
GI
e
o
Q.
E
o
O
Q)
-o
'x
o
c::a
Q) e
OJ O
e-e
.~ cu
zo
Q)
e
.
J :: Q)
Q ) Q 5cQ )
Q) Q) e Q)E -C:>-cc
-Ecu-Eww
Q)C: Q)0 . 0 . Q)
~. ~~ ~
q. 2qQ Q ~
NUC' ): 2: 2I
N' 0 N' NMC:
e
o
:; .
. . o
,ez
i
O
Q) Q) Q)
Q) e e e
c : : c t I c t I c u
co .................. -
e fi5 fi5 fi5 Q) ~
Q )a.c. a. Ccn
0.>.>. >.Q)~ c:
--5-5 -E@ Cl> ~
>.Q)wwwx~ (j)
~.. e ~ ~ d>
E99~9-Q 5 :
Q5.q.NCC"?~ ro
::2:35C"i6~ ~
e
'" u
o
.
O
E
o
u:
Nv I . [ )COf ' . . c x>O' )O~N
~~~
C' )vl O<OI ' - OO
~~~~~~
O' >O~N( , , ). q. l O
...-NNNNNN
13
s;
1
2
~
B
a
(ij
~
.!:l
Ql
-5
E
,g
U
Ql
e
...
~
M
cc
al
Ul
'
>-
iii
c:
(J
':
111
~
O
~ ..
N O
.!!-;
.c c:
!!! ,2
- '5
c:
o
o
Q)
.o
~
VI
Ol
e
'c
e
Q)
~ .S: E"'" o
O . Q )-ij l ~
:E c S:"E ~:ti ~~
"t:J oE ro ro LO
e -.;;:::::ou >1ija.. ,.- ID
8~EC'loC._ .. ():;
'c o 1:g 8 E ~ O :; (
,Q e: Uj ~ ~ ~ ~ ~ " ~
~I c u ~c . . woT " " " Q)E
ro" Ol~ S ~ :i c.
E
Q)
_ cnQ .CU\U 0)_ .
u.cu1'Q;(/)~c'@ $o
Qj~Q) Clo; 0.'.:':= s ts
g; ~~ E g; ~'5. '5. ~ .!!!.
I-LL~~I-CfJ CfJ CfJ ~E
!I~
e l ) o
E
f $
ea Q)
Il. O
Ol
e
o.
E
ro
CfJ
s,
.f: "&
. E ~
E 2
-D :: E
al N e
~ c. -<:
o~ E; o
s rs ()~ ~-g ~
_ .2 E :Q ~ o Qi E :g
~ ~ <0_ ~~ u)~ EE ~
,- e o =c.ro-ro N:l.....,
>. o o o E o .2 i: g "! ~ J!
iii O g o :J 'O'j[L= 0 _
0'- Q) O - .. o E
c: I c:w"Oc;;-o Q) Q.;() o
.- ~Q) o Q) I E - VI -=
Q i. .. E,?~~Ui Q)Q)"O
:J Q)u=_rorooEe '"
ea o ~'c S >. s: en LO .~ -tj .~
>(0 ~ m$: ~ Q.""i::"'C:C c.
_ . cn .. !Q"OQ) <l> .. Q)
~ ~ (ti g. ~=a. 'S 'E g~E ~~
.- (ti .S L.... >. cu .S! ro Q) (1) = M
Ea:LLl- u...J ...J ELL ~
'
o
~
E
ro
o,
e
Ci i c.
Q)
e l ) ~
E ~
ea ro
~ Q3
Il. c.
E
2
e
Q)
b
e
E
:J
'O
u
( i j
u
.~
( i j
e
-c
'" g
c:
~
~
Q,
.s
-;
::;;
'O
e
lO
'O
o
.;-
O
E
e
LL
Ul
'
>-
iii
c:
(J
':
111
~
o
C? - =
N'"
Q). E
:ti Ul
!!! 15
::;
'5
e
o
o
iii
u
~
iii
c:
e
Q)
E
Q)
Qi
Q)
Ol
.S e c:
I/I
e
VI_'
E
'
E
E
~:::::::: Cii
.2 'S; E E e l ) o
~ ULOo ~:o
e :J NLO roroLO
0'0 - ->-~
O ~~~g.E~
CJ .... oor::l
( i j ~~N- O( j
E
oo .. $.. ~
,"-;::;:::~c.~~
~~o~~'iiiE
1-~]!~CfJ $2
.... ::1 a. .. Q .
~~'8E~E~
~~~~~J 5~
!I~
e l ) o
E
ea Q)
~ ~
1/1
e
~ I e LO " ._,....
e E -
o LO
O _ ~
oo
o ~ o
0 . .0
LOQ)LO
. . : J . .
~(ij s
- >-
ro 'O ro
> e >
( i j o( i j
.- () c:
~ Q) .-
ECfJ LL
VI
e
E
:J
'O
u
'O
Q)
Q) . >c :
u
e l ) ro
E c.
ea -
Q)
~ ~
u
e
:J
~
Q)
c.
E
~
Ol
.S:
o.
E
ro
CfJ
.S:
L:
VI
Q)
E
O .
~ E
s: I ~
VI o '-
~ ~ ~
o X E:
O ~ e
Q) c:..t::
o .> .-
CX) .~ . 'O
ri '; ~ '; E ~
.::s:. -; :- ~ ~ L() 5
~ ~ ~ :ri ~ .f5
ea -Q) _Ola>
o Q )o Q)I~
[LE~~E~i E
,....C")UN.,...U~ ,g
E:5~E~~a; ~
::l O> fA :J Q) (J )'C c:
o~"5oai'5~ ".
O...J oO...J oo ~
M
ex>
al
E
VI
e
E
:J
'O
u
( i j
u
.~
( i j
e
<{
'"
al
al
e;
~
.,
(/)
Q,
.s
-;
::;;
'O
e
lO
'O
o
.;-
O
E
o
u:
".
Table 2-4
Regrouped Composition 01 Separator Gas
Molecular
Component Weight % Weight Mole %
Nitrogen 0.779 28.01 0.659
Carbn dioxide
10.499 44.01 5.652
Methane 46.577 16.04 68.795
Ethane 16.326 30.07 12.863
Propane 14.775 44.10 7.938
;-Butane 2.295 58.12 0.936
n-Butane 4.810 58.12 1.961
;-Pentane 1.050 72.15 0.345
n-Pentane 1.283 72.15 0.421
e, Paraffins 0.728 86.18 0.200
e6 Naphthenes 0.065 70.14 0.022
e, Total 0.793 84.59 0.222
e7 Paraffins 0.159 100.21 0.038
e7 Naphthenes 0.359 85.95 0.099
C, Aromatics 0.055 78.11 0.017
e, Total 0.573 88.59 0.153
ea Paraffins 0.034 114.2 0.007
ea Naphthenes 0.127 99.5 0.030
eaAromatics 0.049 92.1 0.013
ea Total 0.210 99.7 0.050
C, Paraffins 0.009 127.7 0.002
eg Naphthenes
.
0.001 125.0 0.000
C, Aromatics 0.006 106.2 0.001
C, Total 0.016 118.5 0.003
Decanes plus 0.012 140.0 0.002
E
100.000
100.000
Gas molecular weight =23.69
Gas gravity =0.818
In this work the oil phase analysis is divided into two parts:
1. To C
J O
+ using chromatography
2. from CJ O to C20+ using distillation
Liquid Phase Analysis to CLO+
A gas chromatogram of the oil isobtained by use of a chromatograph fit-
ted with a50-m Chrompack Sil 5(fused silica) capillary column. The analy-
sis is temperature programmed from 10 to 300C. The sample volume in-
16
Ptoperties 01 Oils and Natural Gases
jected is 0.1 mI with a split ratio of 1:100. The effluent detector is a flame
ionization type (FID). Further details of the analytical conditions are given
as follows:
Analytical Conditions for Liquid Phase GC Analysis
Column: Type: fused silica capillary
Liquid phase: Chrompack CP Sil 5
Carrier gas: He, =22 cm/s
Length: 50 m
Inside diameter: 0.23 mm
Film thickness: 0.4 ,tm
Detector: Type: flame ionization
Fuel gas: H2, 30 ml!min
Make up gas: N2, 30 ml!min
Temperature: 350C
Sampling: Syringe: Hamilton 7001 N
Sample size: 0:1-0.5 ,tI
Injector: Type: split injector
Ratio: 1:100
Liner: packed J ennings tube
Temperature: 300C
Oyen: Type: temperature programmed
Initial value: +10C, 2 min
Ratel: 3/min -> ll5C
Rate 2: lOo/min -> 300C
Final value: 300C, 60 min
. Total time: - 2h
Source: Osjord et al., (1985).
Toproperly quantify the individual components, an internal standard of
known quantity isadded to the sample. In this way one may compensate for
column losses that are known to occur from carbon numbers of 15 and
higher. The component used as the internal standard is iso-octane, which is
normally not present in naturally occurring oils.
As for the gas sample analysis, a calibration program enables thorough
identification and quantification of the components present between C
2
and
C
g
. Table 2-6 shows a typical example of such an analysis. This detailed an
analysis up to C10+isnormally not needed inpractice. Therefore, aregroup-
ing of components may be carried out as shown in Table 2-7.
Compostonal Determnatons 17
Table 2-5
Regrouped Composition of Separator Gas
(Corresponds to the Analysis in Table 2-4)
1,
,1
Component Mole %
Nitrogen 0.66
Carbon dioxide 5.65
Methane 68.80
Ethane 12.86
Propane 7.94
;-Butane 0.94
n-Butane 1.96
;-Pentane 0.34
n-Pentane
0,42
Hexanes 0.22
Heptanes 0.15
Octanes 0.05
Nonanes 0.00
Decanes plus 0.00
r: 100.00
Gas Molecular Weight 23.69
:1
In addition to the weight fractions of the collective groups such as the
C/s, one may from such a detailed starting point calculate average rnolecu-
lar, weights, average density, critical properties, and PNA distributions for
each of the groups up to C10+' See Tables 2-8 and 2-9.
Identification of the components inthe range CIO to C20on the basis of the
chromatogram in Figure 2-3 is seen to be impractical. Hence the liquid
phase analysis is extended beyond CIO+ by a distillation technique.
1,
i~t
Liquid Phase Analysis from CIO to C20+
A mini-dstillation apparatus isshown in Figure 2-4. It isan adaptation of
the commercially available Fischer HMS 500 automated rnn-dstllaton
system. The required sample volume isabout 100 mI. The distillation proce-
dure used is a modification of the ASTM D-2892 procedure for true boiling
point (TBP) dstllton. The purpose of the mentioned modifications is to
narrow the distillation cut temperature ranges so that they correspond to
those of Katz and Firoozabadi (1978).
The distillation procedure results inan analysis as shown in Table 2-10,
where the analysis up to CIO iscarried out at atmospheric pressure. From CIO
to C20+, the distillation is carried out at a vacuum of 20 mm Hg to reduce
the temperature level and avoid thermal cracking of the sample.
t-
;1
111<
,
l'
18
Properties o] Oils and Natural Gases
u
~
---1
j
~~
: ~~
E'\ ~ ---,=---.3
~=~~
.,
:J
' E
., '" Q. ~
E Cl
., e
>- Q.
u
~
_31
~
~
~
f~
e ;~~
:: :;----." ..
e.,
=.g
- ='"
- -- --- ~ ---=-=
El--=:::::;
".-=-
~---..' g; '>
:-----.;;
~ ~
~
esuodsai al~
~
~
:,..,
::;~
!
CD
N
Q)
:a
~
e
Q)
Ci
E
CO
(J )
'0
Q)
"O
~
Q)
oS
.8
"O
e
o
o.
CIJ
~
o
o
CIJ
ID
.o
E
~
e
~
co
Q)
n,
Q)
Ci
E
co
CIJ
Q)
ro
CIJ
e
Q)
"O
e
O
o
ro
o
E
~
Ol o _
- '"
ro '"
E ~
O -
.c i
O Q
"'O
M ~
N .~
Q J ()
... .,
::l u
Cl :;
.- o
L1. CIJ
Table 2-6
Individual Component Report
(Peak Numbers Correspond with Chromatogram in Figure 2-3) .,,~, t 1
':~\I~' <;:;
" tl /
~
" ,
Molec- Liquid
Weight Mole Volume ular Density
Peak Component % % % Weight (g/cm
3
)
1 C2 0.007 0.058 0.017 30.070 0.3580
2 C3 0.072 0.412 0.122 44.097 0.5076
3 te, 0.051 0.222 0.078 58.124 0.5633
4 o-e, 0.189 0.816 0.276 58.124 0.5847
5 2,2-0M-C. 0.000 0.000 0.000 72.151 0.5967
6 -c, 0.188 0.653 0.257 72.151 0.6246
7 n-Cs 0.285 0.991 0.386 72.151 0.6309
O Light end total 0.792 3.152 1.137 63.092 0.5964
8 2,2-0M-C, 0.012 0.034 0.015 86.178 0.6539
9 cv-c, 0.052 0.185 0.059 70.135 0.7502
10 2,3-DM-C, 0.028 0.081 0.036 86.178 0.6662
11 2-M-Cs 0.165 0.480 0.214 86.178 0.6577
12 3-M-Cs 0.102 0.298 0.131 86.178 0.6688
13 n-C6 0.341 0.993 0.440 86.178 0.6638
O Hexanes total 0.699 2.071 0.895 84.745 0.6687
14 M-CyCs 0.231 0.689 0.262 84.162 0.7534
15 2,4-0M-Cs 0.015 0.038 0.019 100.205 0.6771
16 Benzene 0.355 1.140 0.343 78.114 0.8842
17 Cy-C6
0.483 1.440 0.528 84.162 0.7831
18 2-M-C6 -
0.000 0.000 0.000 100.205 0.6829
66 1, 1-0M-Cy-Cs 0.116 0.298 0.131 98.189 0.7590
19 3-M-C6 0.122 0.307 0.152 100.205 0.6915
20 1, cis-3-DM-Cy-Cs 0.000 0.000 0.000 98.189 0.7493
21 1, trans-3-DM-Cy-Cs 0.052 0.133 0.059 98.189 0.7532
22 1, trans-2-DM-Cy-Cs 0.048 0.122 0.054 98.189 0.7559
25 n-C7
0.405 1.014 0.504 100.205 0.6880
O Unspecified C7 0.171 0.427 0.215 100.205 0.6800
O Heptanes total 1.997 5.609 2.267 89.426 0.7542
26 M-Cy-C6 0.918 2.348 1.016 98.189 0.7737
27 1, 1, 3-TM-Cy-Cs 0.027 0.061 0.031 112.216 0.7526
28 E-Cy-Cs 0.000 0.000 0.000 98.189 0.7708
29 2,2,3-TM-Cy-Cs 0.042 0.093 0.050 114.232 0.7200
30 2,5-DM-C6 0.018 0.039 0.022 114.232 0.6977
31 2,4-0M-C6 0.000 0.000 0.000 114.232 0.7045
32 3,3-0M-C6 0.026 0.057 0.031 114.232 0.7141
33 1, trans-2, cis-3-TM- 0.025 0.056 0.028 112.216 0.7579
cvc,
34 Toluene 0.958 2.610 0.941 92.143 0.8714
35 1, 1, 2-TM-Cy-Cs 0.000 0.000 0.000 112.216 0.7769
36 2,3-DM-C6 0.033 0.073 0.040 114.232 0.7163
37 2-M-C7 0.137 0.300 0.167 114.232 0.7019
38 3-M-C7 0.094 0.206 0.113 114.232 0.7099
39 1, cis-3-DM-Cy-C6 0.190 0.425 0.211 112.216 0.7701
40 1, trans-4-DM-Cy-C6 0.072 0.162 0.081 112.216 0.7668
(table continued on next page)
20 Prop'erties of Oils and Natural Gases
Table 2-6
Continued
Molec- Liquid
Weight Mole Volume ular Density
Peak Component % % % Weight (g/cm
3
)
42 Unspecilied 0.028 0.062 0.031 112.216 0.7700
naphthene
42 Unspecilied 0.013 0.028 0.014 112.216 0.7700
naphthene
42 Unspecilied 0.011 0.025 0.012 112.216 0.7700
naphthene
43 OM-Cy-C6 0.031 0.069 0.034 112.216 0.7700
44 1, trans-2-0M-Cy-C6 0.089 0.199 0.098 112.216 0.7799
45 n-Ca 0.434 0.954 0.526 114.232 0.7065
O Unspecilied C, 0.086 0.190 0.105 114.232 0.7000
O Octanes total 3.231 7.957 3.551 101.978 0.7791
46 Unspecilied 0.047 0.094 0.051 126.243 0.7900
naphthene
47 2,2-0M-C7 0.009 0.018 0.011 128.259 0.7144
48 2,4-0M-C7 0.017 0.033 0.020 128.259 0.7192
49 1, cis-2-0M-Cy-C6 0.024 0.054 0.026 112.216 0.8003
50 E-Cy-C6 +1, 1, 0.281 0.599 0.305 118.000 0.7900
3-TM-Cy-C6
51 Unspecilied 0.047 0.093 0.051 126.243 0.7900
naphthene
52 3,5-0M-C7 0.017 0.034 0.020 128.259 0.7262
53 2,5-0M-C7 0.003 0.006 0.004 128.259 0.7208
54 Ethylbenzene 0.114 0.270 0.112 106.168 0.8714
68 Unspecilied 0.027 0.054 0.029 126.243 0.7900
naphthene
55 m- +p-xylene 0.697 1.649 0.687 106.168 0.86.83
56 4-M-Ca 0.020 0.039 0.024 128.259 0.7242
57 2-M-Ca 0.054 0.106 0.064 128.259 0.7173
58 Unspecilied 0.009 0.018 0.010 126.243 0.7900
naphthene
58 Unspecilied 0.082 0.163 0.089 126.243 0.7900
naphthene
58 Unspecilied 0.007 0.014 0.008 126.243 0.7900
naphthene
59 Ortho-xylene 0.230 0.545 0.223 106.168 0.8844
60 3-M-Ca 0.023 0.045 0.027 128.259 0.7242
61 1-M,3-E-Cy-C6 0.078 0.155 0.083 126.243 0.8000
62 1-M,4-E-Cy-C6 0.034 0.068 0.037 126.243 0.7900
63 Unspecilied 0.006 0.013 0.007 126.243 0.7900
naphthene
63 Unspecilied 0.004 0.007 0.004 126.243 0.7900
naphthene
64 n-Cg 0.471 0.923 0.559 128.259 0.7214
O Unspecilied Cg 0.124 0.243 0.148 128.259 0.7200
O Nonanes total 2.427 5.241 2.598 116.277 0.7995
Source: Osj ord el al. (1985).
Table 2-7
Gas Components and Boiling Point Groups up to C10+ from Capillary
GC-Analysis
l
11
1
1
11
1
::1
.1-1
I
Liquid
Molecular Density
Component Weight % Mole % Volume % Weight (g/cm
3
)
C, 0.000 0.000 0.000 16.0 0.260
C2 0.007 0.058 0.017 30.1 0.358
C3 0.072 0.412 0.122 44.1 0.508
ic, 0.051 0.222 0.078 58.1 0.563
n-C. 0.189 0.816 0.276 58.1 0.585
2,2-0M-C3 0.000 0.000 0.000 72.2 0.597
te, 0.188 0.653 0.257 72.2 0.625
e, 0.285 0.991 0.386 72.2 0.631
Hexanes total 0.699 2.071 0.895 84.7 0.669
Heptanes total 1.j :l97 5.609 2.267 89.4 0.754
Octanes total 3.231 7.957 3.551 102.0 0.779
Nonanes total 2.427 5.241 2.598 116.3 0.799
Oecanes & 90.853 75.971 89.553 300.3 0.868
heavier
' 11 I
11
I
Table 2-8
Distribution of Weight, Mole, and Volume % in a Crude Oil Sample,
from Capillary GC-Analysis
Liquid
Molecular Density
Componen! Weight % Mole % Volume % Weight (g/cm
3
)
Gas components 0.792 3.152 1.137 63.1 0.596
Hexane group 0.699 2.071 0.895 84.7 0.669
Heptane group 1.998 5.612 2.263 89.4 0.756
Octane group 3.231 7.957 3.551 102.0 0.779
Nonane group 2.427 5.241 2.598 116.3 0.799
Oecane group 2.483 4.666 2.654 133.6 0.801
Undecane group 2.239 3.799 2.388 148.0 0.803
Oodecane group 2.569 3.994 2.709 161.5 0.812
Tridecane group 3.127 4.479 3.238 175.3 0.827
Tetradecane group 3.153 4.172 3.214 189.8 0.840
Pentadecane group 3.988 4.890 4.041 204.8 0.845
Hexadecane group 3.417 3.937 3.438 217.9 0.851
Heptadecane group 4.286 4.577 4.355 235.1 0.842
Octadecane group 3.060 3.076 3.099 249.8 0.845
Nonadecane group 3.749 3.604 3.756 261.2 0.854
Eicosane plus 58.781 34.772 56.664 424.5 0.888
Apparent average molecular weight: 254.0
Moiecular weight corrected lor benzene: 251.1
Measured density (g/cm
3
) 15C, 1 atm: 0.856
22 Properties 01 Oils and Natural Gases
Table 2-9
PNA Report
Paraffinic, Naphthenic, and Aromatic Contents of the Boiling Point
Groups up to n-Cg
Molecular Density
Component Weight % Mole % Volume % Weight (g/cm
3
)
Hexane group 0.647 1.886 0.836 86.2 0.663
paraffins
Hexane qroup 0.052 0.185 0.059 70.1 0.750
naphthenes
Heptane group 0.713 1.787 0.889 100.2 0.686
paraffins
Heptane group 0.930 2.682 1.034 87.1 0.769
naphthenes
Heptane group 0.355 1.140 0.343 78.1 0.884
aromatics
Octane group 0.870 1.912 1.054 114.2 0.707
paraffins
Octane group 1.404 3.435 1.556 102.6 0.772
naphthenes
Octane group 0.958 2.610 0.941 92.1 0.871
aromatics
Nonane group 0.739 1.446 0.877 128.3 0.721
paraffins
Nonane group 0.646 1.331 0.699 122.0 0.792
naphthenes
Nonane group 1.042 2.464 1.022 106.2 0.872
aromatics
Table 2-11compares compositnal data obtained by gas chromatography
and by distillation. There is generally satisfactory agreement. One advan-
tage of the analysis by distillation is, however, that it produces a physical
sample of each hydrocarbon fraction present in the oil sample.
PROPERTIES OF HYDROCARBON FRACTIONS
For the hydrocarbon fractions (such as CIl), and the residue (e.g., Cz o+),
it isimportant not only to know their weight fraction, but also their physical
properties such as densities, molecular weights, and PNA distributions.
These properties will be different for each oil and may thus help to distin-
guish an oil of one reservoir from that of another reservoir.
Figures 2-5 and 2-6 show plots of the density vs. carbon number for sev-
eral North Sea oil samples. There are significant differences in the density
,.
I
11,
vs. carbon number profiles, corresponding to different PNA distributions.
This is in agreement with observations by Yarborough et al. (1978).
Similarly, Figure 2-7 shows the molecular weight vs. carbon number for
several different oil samples. As observed by Katz and Firoozabadi (1978),
the average molecular weights of the hydrocarbon fractions are not signifi-
cantly different for the different oils. This can be explained by the fact that
even though the individual molecular structures within each hydrocarbon
fraction may vary considerably, from n-paraffins to isomers to cyclic struc-
tures, their molecular weights wiJ I not be very much different. Different
structures within a given hydrocarbon fraction have very similar ratios of
hydrogen atoms to carbon atoms, and hence the molecular weights are also
similar.
lt is thus feasible to use general, average molecular weights for the frac-
tions from C
g
to C
19
. The average values of the molecular weights of Figure
2-7 and the average densities of Figure 2-5 are shown in Table 2-12.
Unlike Katz and Firoozabadi, general, average density values for each
fraction are not used. Asshown in Figure 2-6, the density vares considera-
bly with the structure (PNA distribution) of the hydrocarbon fraction.
Figure 2-4. Fischer HMS 500 mini-distillation equipment modified for continuous
distillation of crude oils.
24 Properties 01Oils and Natural Gases
Q)
~
c::
Q)
"C
c::
o
U
"C
Q)
~
:c
ea
o
'";' ea
N-
Q) o
- c::
.c O
!!l-.:
~
-.:
.!!l
e
a..
ID
1-
E
~
-
e
~
u U) + +
t! ctj U') W ,... a) 01 ~ ~ :: ~ ~ :! ~ ~ ~ -~ ~ ~
u. Cl
.~~
; Q)
3 E
E-=
8~
o~mw~oor or oWNN~~Wr or o-
W~I D~mvOvmomOvO~ONO
~M~OO~~O~W~~OOONM~WO
~NMmI Do~~r ooooommmmmo
. -
~
Q)I ON~~_ C' )O~or ow_ ~r o_ C' )oC' )
E ",,,,:~'<I;,,,'<I;,,,~~~,,,,,,:'<I;,,,~~,,!~
:J l OCOCOLOCOc n t {)f ' . . LOV( " ' )C' )( \ J . . . . . , . . . . , . . . . . , - C" ' )
_ '1"'""" T""" (")
~
>~
~ME l vCOCOCO( " ' )( " ' )l Oc oc omoo~c ol Ol OV, . . . .
~ WNwC' )wr oNmmONC' )C' )vvv~m
e o ~ww~~~r o~~r or or or or or or or or o
~S cicici6cici6cicici6ci6ci66ci6
~
i ~
3Q)
EO
; ; s : : E
U
NCO( " ' )CJ )COCOOVOOVMf ' . . , . . . . l ! )Vl ! )CJ )
l OCJ )vCOONOl OO, . . . . COc n vv( " )oc o, . . . .
oooom~w~ci~wm~M~w~ooooci
, . . . . C\ J vc o, . . . . . . oc oooCOCJ )CJ )CJ )CJ )CJ )CJ )CJ )O
..
~I N~wwmr ovr owovmvv~m_ v
IV ~~"":~~r:":~'"":~":~~~~~~t.t1
o COOOCOCOOl V<OV( , , )C\ l NT " " " OOOQ' I " ' " " "
:: '1"'""""- T""" T""" C\ J
..
.!!! .
;;sJ :
ld .21
o~
::E
r oNr o~C' )O
~N~~~ml OC' )l Oc o, . . . . ( " , )" " " o<O' I " ' " " " om
l Of ' . . c oc n O . . . . . f ' . . C" )vl Of ' . . COm, . . . . C\ J vl O. , . . . -
" 1" ' : " ' " " " " " " ' ' 1' ' ' " " " ' 1" ' " " " ' 1" ' " " " ' I " ' " " " " - T ' " ' N NNC\ J M
~~
E l :
; ; s CI
E 'j
83:
mf ' . . VN, . . . . . . mOf ' . . c n f ' . . c o" " , V, . . . . MNCOO
~~~~~~~~~~~~~~" : ~~~
vor ov_ oor ovr oNwmoNv~O
' I " ' " " " ' I " ' " " " ( " ' )l OCOOCO" " " " " c oc oCOCJ )CJ )CJ )CJ )O
- -
~
. .l mr o~r o~ _ _ r oNm_ C' )C' )r oNmwN
s: T"""f'..O>T"""OOOOvCOMC\J v"""'f"...mvMv
~ ~w~~~mm~~~~MN~~~~~
' ,.... '1"'""" M
3:
ea
-
ea
e -g 6
Q)
J :~
(ijci.
.a E
o Q)
-
LOC\ J O). . - C" )
c . r i mc c i wr =
I C' )<DO' >~~
f ' . . W( " )LO" " , . , . . . . . mCX)( " ' )C\ l N
oc r i u i " " : w" " : M~6NN
~r ooNC' )~wr - - moo
T " " " T " " " T " " " T " " " ' I " ' " " " T " " " T " " " C\ l C\ I
1\
.F
~~I "'''!~''':~~''''<I;N~~''':~~~~~
c omc oc o. , . . . - . , . . . - ve of ' . . l OM . . . . . f ' . . Mf ' . . . . . . . , . . . .
~~ I ( " ' )I D~N~~~~~M~~OOO~MM
~ ~~~~~NNNNNMMMM
; - 1\ 1\
UU
~
Table 2-11
Comparison Between Results from a TBP Distillation and a Capillary
GC-Analysis. (Average of Six)
Oistillation
GC-analysis
Welght MW Oensity Weight MW Oensity RSO'
Fraction (%) (g/mol) (g/cm
3
) (%) (g/mol) (g/cm
3
) (%)
C5
0.886 65 0.621 0.792 63.1 0.597 0.28
C6
0.737 82 0.695 0.699 84.8 0.669 0.26
C7 2.371 91 0.751 2.000
89.4 0.754 0.14
C8
2.825 103 0.778 3.237 102.0 0.779 0.21
C9
2.539 116 0.793
2.429 116.3 0.799 0.39
ClO+
90.642 306 0.869 90.846 300.3 0.868 0.002
C'0
2.479 132 0.798
2.437 133.6 0.801
0.49
e"
1.916 147 0.803 2.191 148.0 0.803
0.40
C'2
2.352 163 0.817 2.523 161.5 0.812 0.33
C
'3
2.09l 175 0.836 3.106 175.3 0.827 1.06
C14 3.667 190 0.843 3.124 189.8 0.840 1.13
C'5
3.722 205 0.849 3.984 204.8 0.845 1.60
C'6
2.034 215 0.853 3.383 217.9 0.851 6.05
C17 4.135 237 0.844 4.244 235.1 0.842 3.82
C'8
3.772 251 0.846 3.201 249.8 0.845 12.12
C'9
3.407 262 0.857 3.523 261.2 0.854 5.07
C20+ 61.057 426 0.885 59.130 421.6 0.888 1.31
Relative standard deviation on weight % from GCanalysis.
So~rce: Osj ord el al. (1985).
Measurement of Density
Density measurements of the hydrocarbon fractions and residues are
readily carried out using either a pycnometer or a Mettler/Paar densitome-
ter. The latter method [Kratky et al. (1969), and Leopold (1970)1isbased on
density-dependent frequency changes of an oscillating glass U-tube. A Met-
tler/Paar instrument is shown in Figure 2-8.
The instrument iscalibrated against air and water at each temperature of
interest. With the two calibration points, the unknown density isreadily ob-
tained from the measured frequency of the U-tube .
Measurement of Molecular Weight
Several techniques have been proposed for determining average molecu-
lar weights of hydrocarbon fractions and residues. The techniques are based
on boiling point elevation, freezing point depression, vapor phase osmosis,
26 Properties o] Oils and Natural Gases
0.90
0.89
0.88
0.87
0.86
0.85
f'I)
2
0.84
U
0- 0.83
<, 0.82
>-
r-
0.81
(/)
Z
0.80
w
O
0.79
0.78
0.77
0.76
0.75
0.74 .
8
0.98
0.96
iflli
0.94
<.-;:;
t<) 0.92
~
~ 0.90
el
~ 0.88
Vi 0.86
Z
W
O 0.84
0.82
0.80
0.78
9 10 11 12 13 14 15
CARBON NUMBER
17 18 16
Figure 2-5. Density 01distillation Iractions Irom paraffinic North Sea crude oils
and condensates.
Table 2-12
Average Densities and Molecular Weights of 11 Paraftinic
North Sea Crude Oils and Condensates
Molecular Weight Density
Fraction (g/mol) (g/cm
3
)
e9 118.6 0.778
elO 133.6 0.790
e"
148.0 0.792
e, 161.5 0.801
c., 175.3 0.816
c., 189.8 0.829
c., 204.8 0.834
c., 217.9 0.840
c., 235.1 0.837
c., 249.8 0.840
c., 261.2 0.849
19
~.,.
eompositional Determinations 27
---
~...---.> --
f-; .; ; :r.,r-------
/G// '-""-/'---
,;7 ".,' .s-:: '
/,./ //
.~/ -------------
_ ._ "'//''''' ..-........
/
/
/'
--
Aromatic Oils
Medium Paraffinic Oils
Paraffmic Oils
Aromatics
~~e.h-"~I:':' _
Katz & Firoozabadi
---------
0.76I I I I I I I I I I I I I I I I I I I I I I
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
CARBON NUMBER
Figure 2-6. Density profiles of North Sea oils and condensates.
400
380
360
~
340
O
~
el 320
'S...
el 300
':'--.
~ 280
el
W
260
3:
240
o:::
j 220'
::>
o 200
~
O lBO
~
160
140
~//::~.
//'/ /./.. '
/ / / .
//~<,/
,/" /......
// <./
~ /.'
.:
/i~'~""
/jz.... /
/~ .....
Aromatic Oils
0
Medium Paraffinic Oils
Ea!!'t~i.<:.Q iI~. _ _ o __
Aromatic Fraction
~~e.h_ a!i: ~~~~~'!. _
~~~~-
120~ ( I I I I I I I I I I I I I I I I I I I I
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
CARBON NUMBER
Figure 2-7. Molecular weight profiles 01North Sea oils and condensates.
~ 91220 _ ;
...... ,
~k~.J.';
tUl 8Y"T"MTII< I< .
&tl.~, _
Figure 2-8. Mettler/Paar Densitometers.
gel permeation chromatography, and other factors. All of these methods
have limited ranges of accuracy and should be applied to specifc ranges of
molecular weights.
Most oils, oil fractions, and residues have molecular weights in the range
of 100-500. This coversthe hydrocarbons fromC7 to C35. The method most
suitable for determining molecular weights within this range isthat based
on freezing point depression.
The principle, of this method isthat the freezing point of apure solvent
(e.g., benzene or nitrobenzene) islowered when thesolvent iscontaminated
with asmall amount of oil. Figure 2-9shows atypical temperature vs. time
curve whea asolvent iscooled. The horizontalline corresponds to the melt-
ing point of the pure solvent, Tm. The temperature-line curve for the oil-
contaminated solvent will after a certain interval reach a constant value
lower than Tm. The temperature difference, .6. T, isameasure of the molecu-
lar weight per mole of asample of the oil. If the apparatus iscalibrated us-
ing substances of known rnolecular weight, the observed value of .6. T may
r
I
t
yield the average molecular weight of the oil sample. The method is very
sensitive and, therefore, difficult to use, and must be carefully calibrated.
A molecular weight apparatus based onfreezing point depression, and of
type Cryette A isshown in Figure 2-10.
T e mpe r a t u r e
F r e e z i n g T e mp.
Be n z e n e
F r e e z i n g Cu r ve of S a mpl e
1 __
--- liT
--- ~ --= --
T i me
Figure 2-9. Freezing point depression diagram.
<1
Figure 2-10. Apparatus tor measurement 01 molecular weight by freezing point
depression in benzene.
30
Properties o] Oils and Natural Gases
INTERNAL CONSISTENCY OF ANALYTICAL DATA
To this point, most of the analytical data from the gas chromatographic
measurernents and all of the data from the distillation are available in terms
of weight fractions. The molecular weights, therefore, play an important
role in converting the weight fractons to mole fractions. The molecular
weight of the residue is especially important.
The previo LIS section indicates that the molecular weight measurement is
dffeul to perform. Often relatively large errors are associated with experi-
mentally determined molecular weights. Fortunately, when a distillation
analysis has been performed, it ispossible to check the reported weight frac-
tions and associated molecular weights of components, hydrocarbon frac-
tions, and residue, for consistency. Only an additional measurement of the
molecular weight of the whole oil sample is required. .
For example, an analysis by distillation of an oil sample results in the in-
formation shown in Tables 2-13 and 2-14. In order to check the consistency
of the reported C20 + molecular weight, the average molecular weight of the
whole sample is calculated as follows:
MW ~(~ w, I MW,J -'
(2-1)
where w, isthe weight fraction and MW the molecular weight of fraction i.
This value may be compared with ameasured mean molecular weight for
the whole oil sample. The analysis data are considered internally consistent
Table 2-13
Measured and Calculated Molecular Weights and Densities from
North Sea Gas Condensate
Condensate
c-. c-.
Measured molecular
weight (MW)
111 175 319
ealculated MW
using elO+ MW
111
ealculated MW
using e20+ MW
111 175
Measured density (g/cm3)
0.756 0.825 0.891
ealculated density
using elO+ density (g/cm3)
0.752
ealculated density
using e20+ density (g/cm3)
0.752 0.826
111'-, ~
1'1:
l'
),
1,
!I
I
II
if the two values of oil sample average molecular weights agree to within 2-
3 molecular weight units.
In addition, one may back-calculate the values of CI+ and C20+ molecu-
lar weights, which also must agree closely with the experimentally observed
values for Co+ and C20+.
The average density of the oil sample, p, can also be calculated from the
data for the fraction given in Table 2-14 as follows:
p =E w/Ew/p (2-2)
Table 2-14
Data for Computing Table 2-13
1 2 3 4 5 6
Density
Component Weight % MW Mole % g/cm
3
Gas 4.19 54.8 8.52 4.7 0.~64
7.4
es 6.78 72.2 10.45 7.5 0.628 10.8
e6 7.97 84.8 10.46 8.8 0.668 11.9
e7 15.18 91.5 18.46 16.8 0.738 20.6
es 17.05 104.3 18.19 19.0 0.763 22.3
e9 9.81 119.0 9.18 10.9 0.783 12.5
(elO+) (39.01)
.
(175) (24.74) (43.3) (0.825) (47.3)
elO 7.48 133 6.28
8.4
0.796
9.4
el1 5.82 145 4.46 6.5 0.796 7.3
e,
4.39 158 3.10 4.9 0.809
5.4
c., 4.21 171 2.74 4.6 0.820 5.1
c., 3.43 183 2.09 3.8 0.830 4.1
c., 2.73 197 1.54 3.0 0.837 3.3
c., 1.78 210 0.94 2.0 0.848 2.1
c., 1.92 226 0.95 2.1 0.845 2.3
c., 1.49 241 0.69 1.7 0.845 1.8
c., 1.36 250 0.61 1.5 0.854 1.6
e20+
4.42 319 1.54 4.9 0.891 5.0
E 100.00 100.00 111.1 132.9
Colurnn 3 is obtained by: (column l/column 2)/E(column l/column 2)
Colurnn 4 is obtained by: column 2 x column 3/100
Colurnn 6is obtained by: column l/column 5
ealculated MW using elO+ MW =(4.7 +7.5 +8.8 +16.8 +19.0 +10.9 +43.3) =111.0
Calculated MW using e20+ MW =111.1 (see table)
ealculated MW 01 the elO+ Iraction using e20+ MW =100(8.4 +6.5 +4.9 +4.6 +3.8
+3.0 +2.0 +2.1 +1.7 +1.5
+4.9)/24.74 =175
ealculated density using elO+ density =100/(7.4 +10.8 +11.9 +20.6 +22.3 +12.5
+47.3) =0.752 g/cm
3
ealculated density using e20+ density = 100/132.9 = 0.752 g/cm
3
where wisthe weight percent of component i and p the density of compo-
nent or fraction i.
Again, the calculated and measured average oil densities should agree
closely.
Determination of Reservoir Fluid Composition
The reservoir fluid arriving fromthewell totheplatform isseparated into
a "separator lquid" and a "separator gas" at platform separator tempera-
ture Tp and pressure Pp' The separator liquid isin turn flashed at ambient
conditions (T; and Po) inthe laboratory. The principle of the two stage flash
separation isshown in Figure 2-11. The pracedure isdescribed in more de-
tail in Chapter 3.
The reservoir fluid composition isobtained by combining the analyses of
thelaboratory gas and thestock tank oil, fromwhich the composition of the
separator liquid isdetermined. The analysis of theseparator liquid isin turn
combined with the analysis of the separator gas to yield the composition of
the reservoir fluid (or well stream).
Torecombine the previously mentioned analyses correctly, it isnecessary
to know the gas-to-liquid ratio, the so-called gas/oil ratio (GOR), both at
platform and at ambient conditions. In Chapter 3 it isshown how to mea-
sure gas/oil ratios inthe laboratory. The princple of the flash experiment is
shown in Figure 2-11. An example of such data is given in Table 2-15.
Figure 2-11. Principie of two-stage
flash of well stream.
I
Gas
Oil
r, Pp
First Stage
(separator
conditions)
I . Gas
~
r,Po f-------Oil
Feed
Second Stage
(ambient
conditions)
.dli
Compostonal Determinatons 33
Table 2-15
Composition Analyses, Gas/Oil Ratios, and Formation Volume Factors of
Separator Fluids
Stock Tank Stock Tank Separator
Oil Gas Gas
Component Mole % Mole % Mole %
Nitrogen 0.00 0.20 0.66
Carbon dioxide 0.00 3.96 5.65
Methane 0.00 24.86 68.80
Ethane 0.20 20.40 12.86
Propane 2.14 28.42 7.94
;-Butane 1.10 4.78 0.94
n-Butane 4.25 10.97 1.96
;-Pentane 2.68 2.21 0.34
n-Pentane 4.32 2.53 0.42
Hexanes 6.66 1.05 0.22
Heptanes 11.90 0.54 0.15
. Octanes 13.14 0.10 0.05
. Nonanes 7.73 0.00 0.00
Decanes plus 45.89 0.00 0.00
E 100.00 100.00 100.00
Gas/oil ratio at separator, GORo =442 Sm
3
/m
3
separator oil
Gas/oil ratio at ambient flash, GORo =36.9 Sm
3
/m
3
stock tank oil
Flash lormation volume lactor 01 separator liquid =1.165 m
3
/m
3
stock tank oil
Average molecular weight 01 the stock tank oil =160.4 kg/mole
Basis for the recombination can be, e.g., 1std m'' of stock tank oi!. The
total number of moles of each component (or fraction or residue) in this std
rn'' of stock tank oil iscalculated. The molar volume of an ideal gas at atmo-
spheric pressure and OC equals 22.4 m'. The total individual number of
moles of gas sample corresponding to 1std m
3
of liquid sample may beob-
tained fram:
Moles gas =GOR/ (22.4 x T0/273) (2-3)
where GOR is either GORp or GORo [platform (field), or ambient condi-
tions, respectively]. From the moles of gas plus the composition of the gas
samples, the amounts of individual gas-phase components in both gas
streams can be determined.
Finally, the total number of moles of each component fromthe test sepa-
rator phases are added (seeFigure 2-11) and normalized to yield the compo-
sition of the recombined fluid, i.e., the fluid in the reservoir.
Example: Recombination of Stock Tank Oil, Stock Tank Gas, and Separator
Gas. A reservoir fluid isflashed at the field separator, and the separator liq-
uid isflashed to ambient conditions in the laboratory, yielding astock tank
liquid and an evolved gas. The composition analysis and other experimental
information are given in Table 2-15.
From that information, first compute the number of moles in each of the
streams. As a basis, 1m
3
of stock tank oil ischosen.
Moles stock tank oil =860 kg/160.4 kg/mole =5.36 moles
Moles of evolved gas =36.9 m
3
/(22.4 moles/m- x 288K/273K)
=1.56 moles
The volume of gas (Sm
3
) per 1rrr' stock tank oil is442 x 1.165 =515 Sm3
The number of moles of separator gas isthus:
515/ (22.4 x 288/273) =21.79 moles
The total number of moles of the reservoir fluid per m
3
stock tank oil is
thus:
5.36 +1.56 +21.79 =28.71 moles
The number of moles of each component in each stream may now be com-
puted by multiplying the mole % given inthe compositional analysis by the
total number of moles in each stream and dividing by 100. The results are
shown in Table 2-16.
Finally, the number of moles of each component in the reservoir fluid is
obtained by adding the number of moles in each stream:
Moles of n-pentane in the reservoir fluid =0.2316 +0.0395 +0.0915
=0.3626 moles
The mole % of n-pentane in the reservoir fluid =0.3626 x 100/28.71
=1.26 mole %
The final molar composition of the reservoir fluid isshown in Table 2-17.
ir;
',..
f"
t:;
i-ti
:610
''
.~j l
f',y,
~
Table 2-16
Number of Moles of Each Component in Each Separator Stream
Stock Tank Stock Tank Separator
Component Oil Gas Gas
Nitrogen - 0.0031 0.1438
Garbon dioxide - 0.0618 1.2311
Methane - 0.3878 14.9915
Ethane 0.0107 0.3182 2.8022
Propane 0.1147 0.4434 1.7301
i-Butane 0.0590 0.0746 0.2048
n-Butane 0.2278 0.1711 0.4271
i-Pentane 0.1436 0.0345 0.0741
n-Pentane 0.2316 0.0395 0.0915
Hexanes 0.3570 0.0164 0.0479
Heptanes 0.6378 0.0084 0.0327
Octanes 0.7043 0.0016 0.0109
Nonanes 0.4143 - -
Decanes plus 2.4597 - -
;
5.36 1.56 21.79
The recombination computations are normally performed for composi-
tions to CJ O +. However, for equation of state calculations computations to
C20+ arepreferred. The CJ O to C20+ compositions arenormally obtained by
extending the recombination result using the distillation data. Using these
data, the recombined CJ O+-fraction isdivided into its appropriate ratios up
to C20+.
The result at this stage is a well checked, consistent composition analysis
of the reservoir fluido
EXPERIMENTAL EXTENSION BEYOND C20+
In some cases (e.g., dew-point calculations) quantitative extension be-
yond C20+ is advantageous.
Breakdown of the C20+ fraction may beobtained by distillation, either at
high temperatures (upto 550C) or at lowpressures (- 2mmHg). This isa
very time consuming and difficult analysis, because the process conditions
are not easly controlled.
Table 2-17
Reservoir Fluid Composition
Moles in
Reservoir
Component Fluid Mole %
Nitrogen 0.1469 0.51
Carbon dioxide 1.2920 4.50
Methane 15.3793 53.57
Ethane 3.1311 10.91
Propane 2.2882 7.97
i-Butane 0.3384 1.18
n-Butane 0.8260 2.88
i-Pentane 0.2522 0.88
n-Pentane 0.3626 1.26
Hexanes 0.4213 1.47
Heptanes 0.6789 . 2.36
Octanes 0.7168 2.50
Nonanes 0.4143 1.44
Decanes plus 2.4597 8.57
r; 28.71 100.00
An example of such analysis isgiven in Table 2-18. The apparatus used is
a modified Fischer HMS 500 distillation apparatus (seeFigure 2-4), operat-
ing at 2mm Hg. The analysis time is2-3 days. By this distillation it ispossi-
ble to extend the experimental composition analysis to C
30
+.
There is, however, an empirical method of extrapolating the experimental
C20 + analysis, described in Chapter 7. This method performs well for most
purposes and thus saves the time-consuming C20 to C30 + analysis.
-ae
Table 2-18
Extended Compositional Data for a North Sea Gas Condensate
f
Density
PNA Distribution, Mole %
(g/cm
3
)
Component Mole % 1 atm, 15C MW P N A
N2 0.64
CO2 9.16
C, 68.8
C2
8.43
C3 5.11
ic, 0.81
n-C.
1.45
te, 0.52
n-C, 0.53
C6 0.63
C7 0.83 0.741 96 0.50
0.42
0.08
Cs 0.95 0.780 107 0.45 0.38 0.17
C9 0.52 0.807 121 0.48 0.27 0.25
ClO 0.26 0.819 134 0.47
0.30 0.23
Cll 0.20 0.810 147 0.56 0.27 0.17
C'2
0.17 0.828 161 0.55 0.24 0.21
C'J
0.16 0.849 175 0.54 0.22 0.24
C,. 0.15 0.857 190 0.49 0.27 0.24
C'S
0.11 0.868 206 0.52 0.20 0.28
C'6
0.086 0.872 222 0.55 0.19 0.26
C17 0.078 0.859 237 0.57 0.20 0.23
C,s 0.068 0.854 251 0.70 c. 0.11 0.19
C'9
0.050 0.866 263
C20 0.046 0.873 339 C20+
C2, 0.035 0.876
C22 0.025 0.876
C23 0.034 0.875
C2 0.023 0.877
C2S 0.017 0.876
C26 0.018 0.878
C27 0.014 0.882
C28 0.012 0.886
C29 0.013 0.889
CJ O+ 0.047 0.908
,1
38 Properties oj Oils and Natural Gases
REFERENCES
Katz, D. L., and Firoozabadi, A., "Predicting Phase Behavior of Conden-
sate/Crude-Oil Systems Using Methane Interaction Coefficients," lournal
01 Peto Techn., Nov. 1978, pp. 1649-1656.
Kratky, O., Leopold, H., and Stabinger, H., "Density Determination of Liq-
uids and Gases to an Accuracy of 10- 6 g/cm'', with a Sample Volume of
only 0.6 cm"," Z. angew. Physik, 4, 1969, pp. 273-277.
Leopold, H., "Die Digitale Messung der Dichte von Flssigkeiten,"
E lektronik, 19, 1970, p. 297.
Osjord, E. H., and Malthe-Serenssen, D., "Quantitative Analysis of Natural
Gas in a Single Run by the Use of Packed and Capillary Columns,' ]. 01
Chromatography, 297, 1983, pp. 219-224.
Osjord, E. H., Renningsen, H. P., and Tau, L., "Distribution of Weight,
Density, and Molecular Weight in Crude Oil Derived from Computerzed
Capillary GC Analyss,' Journal 01High Res. Chrom. &Chrom. Comm.,
8, 1985, pp. 683-690.
Yarborough, L., "Application of a Generalized Equation of State to Petro-
leum Reservoir Fluds,' Paper presented at the 176th National Meeting of
the American Chemical Society, Miami Beach, FL, 1978.
l"
111'
Chapter 3
Oil and Gas Property Measurements
This chapter describes some of the most commonly used methods for mea-
suring densities, viscosities, saturation points, and other properties of reser-
voir fluids. It is not possible within one chapter to describe all the relevant
methods in detail. An attempt is made, however, to give the reader an un-
derstanding of the methods used to collect the data shown in this book.
SAMPLING
Hydrocarbon samples to be analyzed for the previously mentioned prop-
erties are normally collected at the first stage (field) separator. Representa-
tive samples of both the separator liquid and the separator gas streams are
collected. Sarnples may 'also be obtained from the bottom of the well, so-
called bottom hole samples. These are, however, not very common as they
are only recommended when taken from undersaturated reservoirs.
Before starting an extensive analysis it is important to evaluate the aval-
able samples and to choose the sample thought to be the most representative
for the reservoir hydrocarbon fluids. Samples obtained during the most sta-
ble separator conditions are normally chosen. In the laboratory, the satura-
tion pressure (bubble point) of the separator liquid is determined at separa-
tor temperature. The sample is considered to be representative if the
saturation pressure is equal to the test separator pressure at the sampling
time.
The opening pressure of the corresponding gas phase sample may also be
checked. It should be equal to the separator pressure at sampling time.
If both of these pressures agree with the value of the separator pressure,
further processing of the samples may take place.
FLASH SEPARATION AND COMPOSITIONAL ANALYSIS
After quality control, the separator gas sample may be subjected to corn-
positional analysis as described in Chapter 2. The gas sample flask isheated
39
.'
,
\.
"
40 Propertes of Oils and Natural Gases
to separator temperature and left to equilibrate. Figure 3-1 shows a typical
apparatus for this purpose. A sample of the gas is bled through a gas sam-
pling valve and injected into the gas chromatograph. From the cornposi-
tional analysis it ispossible to detect whether the sample has been contami-
nated with, for example, air.
In order to analyze the separator liquid sample, which has been collected
under pressure, it must be stabilized to atmospheric conditions by a single
flash process. The high-pressure sample container is, therefore, connected to
a single-stage flash separator as shown in Figure 3-2. The separator liquid
sample is thus allowed to separate into two phases.
The volume of the gas phase isdetermined. The liquid isallowed to settle
in the bottom of the flash chamber, after which the liquid is drained and
weighed, and its density is determined.
The gas/oil ratio of the separator liquid sample, CORo, may now be cal-
culated:
CORo =volume gas (Sm
3
)/volume oil (Sm
3
) (3-1)
where Sm
3
means the volume of fluids at standard conditions, normally
15C and one atmosphere.
The composition of the gas and liquid phases may be determined as de-
scribed in Chapter 2.
PVT MEASUREMENTS
This section describes the types of PVT measurements (often called PVT
analyses) most commonly performed on reservoir fluids. In order to study
the properties of the reservoir fluid, it isnecessary to recombine the separa-
tor gas and the separator liquid samples. This is done by mixing the two
samples proportionally to their produced rates (CORp).
There are two main types of PVT equipment available: PVT systems for
black oils and PVT systems for gas condensates and volatile fluids. These are
discussed separately.
BLACK OIL EQUIPMENT
A typical schematic diagram of ablack oil PVT.system isshown in Figure
3-3. The main parts of the system are a mercury pump, a calibrated steel
cell, a heating bath with silicone oil, and a precision manometer or dead
weight gauge. The principle of operation is as follows:
1
1
t
;l\
f
/l./
Figure 3-1. Separator gas
sample f1ask rigged for
chromatographic analysis.
:ti
~.;
;~'
i:
t;
;1'1
Oil and Gas Property Measurements 41
Figure 3-2. Single stage

Ruska model 2353-803
flash separator.
"
'----'
Q)
o
CD
1. He a t i n g ba t h wi t h s i l i c on e oi l
2. P VT c e l l ( vol u me n or ma l l y
0.6 - 1.0 1)
3. Va l ve
4. Gr a du a t e d me r c u r y pu mp
5. P r e c i s i on ma n ome t e r
6. Ce l l e n t r y va l ve
Figure 3-3. Schematic diagram of
black oil PVT apparatus.
(0
1. He a t i n g ba t h - wi t h s i l i c on e oi l
2. P VT c e l l
3. l s ol a t i n g va l ve
4. Gr a du a t e d me r c u r y pu mps ( a ) c e l l ( b) c h a r gi n g
5. P r e c i s i on m" ~ome t e r s
6. Ce l l e n t r y va l ve
7. Hy dr oc a r bon s a mpl i n g bot t l e or r e c ombi n a t i on c e l l
Figure 3-4. Black oil PVT apparatus rigged for sample transfer.
A ready, recombined sample is charged to the cell at constant pressure
through valve 6at the top of the cell. The cell may beequipped with oneor
more windows for monitoring the charging and subsequent operations.
The charging process isperformed at the pressure in the sample cylinder.
Constant pressure ismaintained using the mercury pumps shown in Figure
3-4.
:f
f
)
.
.,\
: ..~
~... :~.
t :1,
j "<
* r
.~ "
i ffl'
';,
The mercury pumps can be either hand or machine operated. In both
casesthepumps must begraduated and calibrated to avolumetric scale. By
retracting the cell pump and advancing the charge purnp, aknown volume
of sample can be transferred to the cell.
The cell isimmersed in athermostat and heated to the reservoir tempera-
ture, or any other required ternperature, before the cell ischarged. This is
important because heating after charging could, without proper monitoring
of the oil pressure, lead to excesscell pressures due to thermal expansion of
the mercury and the sample. If the sample has been brought into the two-
phase regin, heating after charging is normally safe but should be moni-
tored.
Bubble-Point Determination
Once the systemhas been equilibrated at the desired (usually reservoir)
temperature, the PVT analyses can start.
Normally, the first analysis performed isthe determination of the bubble
point. The bubble point of asystemisdefined asthe pressure where, at con-
stant ternperature, the oil systemis in equilibrium with an infinitesimally
small gas bubble.
The PVT cell is accessed by opening valve 6 shown in Figure 3-3, and
mercury is injected into the cell by means of the mercury pump. The
amount of mercury isrecorded after having closedvalve3, and thesystemis
equilibrated at the resulting pressure. Based on the amount of mercury in-
jected into the cell, and the amount of sample charged to the cell, onecan
calculate the sample volume at this pressure. This processisrepeated several
times, and the resulting volume pressure data isplotted asshown in Figure
3-5.
Normally, aclear discontinuity isfound. This point isthe bubble point of
the mixture at the given temperature.
Several factors influence the accuracy by which the saturation pressure
can be determined by the graphical technique. The change of slopein Fig-
ure 3-5 depends on the compressibility difference between the two-phase
systemand the single-phase system. At high pressures and temperatures, and
when the systemapproaches the critical region, this difference is not very
large.
Differential Depletion
Other important properties of theoil systemcan bederived through ad-
ferential depletion or differential liberation experimento These properties
are:
44 Properties oi Oils and Natural Gases
(1J
E
:J
~
One Pha se
V
s
P s Pressure
Figure 3-5. Black oil bubble point determination.
1. Shrinkage factor, Bo. This is a measure of the ratio of the volume of
the hydrocarbon system at reservoir conditions to the volume at stock
tank conditions:
Bo =res. m' ol/Sm ' stock tank oil ( 3- 2)
2. Solution gas/oil ratio, Rs. This is a measure of the volume of gas in
solution at given pressures (see also Equation 3-8):
Rs =Sm
3
gas/Sm' stock tank oil ( 3- 3)
3. Density of the reservo ir fluid at different pressures, including the satu-
ration pressure:
p =kg of Fluid/rn'' (3-4)
4. Real gas deviation factor Z, i.e., compressibility factor, for the gas
phase at given pressures and temperatures:
Z =PV/nRT (3-5)
where V is the volume and n the number of moles.
A schematic diagram of the differential depletion experiment isshown in
Figure 3-6. The equipment is shown in Figure 3-7.
The reservoir fluid is, at constant temperature, brought to a pressure
above its saturation pressure. After this, the pressure is brought to a value
Gas off AII gas
displaced
r
. . . . .
. Gas.
t
r
'i
: Gas:
>" ' 1
1 1: Gas.
:o
v
N
>"' >
Oil
11
Oil
I
11
->
Oil >C' l
t
1
11
-;;;
>~
>" '
ro v
>" ' Oil >~I Oil
v
v
>N
>LD
L 1
P
1
- P
sat
P
2
< P
sat
P2 < Psat
P
2
< P
sat
P
3
< P
2
< Psat
Figure 3-6. Schematic representation 01adifferentialliberation experiment.
Figure 3-7. A PVT system used tor differentialliberation analyses. One sees
fromleft to right: the motorized pump, PVT cel!, and bath gas col!ectionsys-
temwithcondensate trapo
below the saturation pressure. The system will then be in the two-phase re-
gion. After the system has been equilibrated, the gas phase isremoved from
the cell at constant pressure. The gas volume is measured with a gasometer
and recorded.
The new volume of the oil iscalculated from the readings on the mercury
pump before and after the discharge of the gas. These readings will also give
the gas phase volume at the given pressure and temperature.
The hydrocarbon system is again brought into the two-phase region, this
time at a lower pressure, and the gas discharge procedure is repeated.
The pressure step intervals are chosen so that 8-10 stages will occur be-
tween the saturation pressure and the atmospheric pressure.
At the last stage, care must be taken to discharge al! of the remaining gas,
whereas the residual oil is left in the cell. After the last part of the gas is
discharged, the cell is closed, its internal pressure is increased in steps, and
the resulting ol volumes are recorded. Normally, one records volumes at
each 50 bar up to 300 bar. The plot of these volumes can be extrapolated
back to atmospheric pressure, sothat ameasure of the residual oil volume at
the given temperature, which is often that of the reservoir, can be deter-
mined.
Then the cell temperature is reduced to 15C, and the oil volume at this
temperature is determined as previously explained. This volume isthe refer-
ence residual volume.
The oil is then discharged from the cell, and its density sdetermined as
explained in Chapter 2.
The data recorded in this analysis can be represented as shown in Figure
3-8. Ps and Vs are the saturation pressure and volume, respectively; VRT the
liquid volume at one atmosphere and temperature, Ti and VR is the refer-
ence liquid volume at 15C. From these data the shrinkage factor of the sys-
tem at the saturation pressurernay be calculated:
Bo,s =VSIVR (3-6)
The whole Bo-curve (see Figure 3-9) may be calculated as follows: .,
Bo =VIVR (3-7)
The solution gas/oil ratio of the reservoir fluid, Rs, at pressure stage N can
be obtained from
NS T E P
Rs = L Vgas,nlV R
n~N +1
(3-8)
where NSTEP isthe total number of flash stages and Vgas,n is the volume of
gas (Sm
3
) liberated at flash stage n. These data can be represented as shown
in Figure 3-10.
i
~';ftt
M
~
Vs.
Ps
~
'"
~
100 200 300
Pressure [Bar J
Figure 3-8. Determination of saturation pressure and volume from a differential
depletion experiment.
~
o
~
100 200 300
Pressuro [Bar)
Bo.s
o
'"
Figure 3-9. The differential liberation Bo-factor as function of pressure.
The density of the reservo ir fluid at each pressure reduction step can be
determined from the known density at 15C, the liquid volume, and the
volume and molecular weight of the gas. The density versus pressure data
may be presented as shown in Figure 3-11.
Deviation from ideal gas behavior may be expressed via the compressibil-
ity factor, Z, as defined in Equation 3-5. Employing Equation 3-5 twice,
48 Properties of Oils and "fatural Gases
1-
Figure 3-10. The solution
gas/oil ratio as function of
pressures.
o, P s
V1
cc
P r e s s u r e
p
Figure 3-11. Oensity 01 reser-
voir fluid as function 01 pres-
sure.
P s P r e s s u r e
i.e., at two different pressures and temperatures and for the same number of
moles, will result in:
Z =PVT2Z2/P2V2T\ (3-9)
Equation 3-9 relates a compressibility factor at condition 1, where it is
unknown, to the compressibility factor at condition 2, where it isassumed to
be known. Z2 is normal!y the compressibility factor at ambient conditions,
and it is usually clase to unity.
Process Simulation by Flash Separation
A three-stage flash can be performed in arder to furnish support data for
process simulation.
Such a flash sequence is shown in Figure 3-12.
Stage three is the stock tank stage, where the pressure is one atm and the
temperature is 15C. For the other two stages, temperature and pressure can
be chosen as desired.
GAS ( 1) GAS ( 2) GAS ( S T )
1st S T AGE
pl' T1
2n d S T AGE
P
2
,T
2
3r d S T AGE
Pst' \t
OI L t o s t oc k t a n k
Figure 3-12. Three-stage flash separation train.
The two first stages are accomplished in the same PVT cell as was used in
the differential liberation experiment, and the last stage is carried out by
transferring the second stage Huid to a single flash apparatus (see "Flash
Separation and Compositional Analysis," discussed earlier). From this ex-
periment the overall COR, COR at individual stages, compositions of the
gas and liquid phases, properties of the stock tank Huid such as density and
molecular weight (seeChapter 2), and also the Bo factor of the oil can be
obtained. Tables 3-1 and 3-2 show a typicallaboratory result of a separator
test experimento
Gas Reinjection Related Analysis
).~
In arder to understand the reservoir Huid behavior under gas injection
processes, two different analyses have been designed:
l. Swelling test
2. Slim tube minimum miscibility pressure determination
Swelling Test. When gas is injected into a reservoir containing an under-
saturated ol, the gas can go into solution. This has the effect of swel!ing the
oil, i.e., the volume of oil becomes larger.
Simulation of this effect can be performed in an ordinary PVT cell, start-
ing with the original reservoir Huid in question. Injection gas has been com-
pressed into a separate steel bottle.
A srnall, known volume of injection gas istransferred into the PVT cell. A
new saturation pressure is determined using the techniques already de-
scribed, and a new saturation volume is recorded. This process is repeated
until the saturation pressure of the fluid is equal to the estimated injection
pressure of the system.
The data from this test can be presented as shown in Figure 3-13.
..
SlimTube Tests. One method of improving oil recovery involves injection of
gas into the oil reservoir. The gas may be nitrogen, carbon dioxde, Hue gas,
(5
..11:
U
ea
al
ea
~
o
- ";" U)
M!: :
CII : : J _ U)
.c CII
~a:
-
U)
~
o
-
ea
~
ea
Q.
CII
CJ )
el)
Cl
ea
00
'" - ea
"'O(!ll:
!5.~i
Q) >J :
a. ea '"
(/) . ea
(!IiL
..
o Q , l .
iE1
.j
ea _ ..,
g. e af
(/)
e
o el)M
'':; E"L
ea j o
E - o
.. > 0 ea
af LL
-"' ..
e ea
~ >0
~ o
- " ' ' ' ' L t l
.., e ...
o el)
000
~
<el
o
~
~
N
Ol
ci
~
: I :
io o
[O ~ I ~
o o
Lt l l " <t
N <Xl
~ ~
o
o
Cl
o
CD
~
N
N
co
co
o
5
E _ O
en;
ea ea
(!l a:
5
C
C : 8
ea ea
(!l a:
o Q)
.. ..
~ ~ :a
a."'1II
Q) f!
(/)a.
"!I'"
Lt l " <t
o N
~
u : r ICD
Oi
Ol
co
CD
<e l
~ ci
" <t
N
N
":
N
'"
Table 3-2
Analyses of Separator Gases from Separator Tests
Component Stage 1 Stage 2 Stage 3
Hydrogen sulfide O O O
Carbon dioxide 1.56 2.05 2.60
Nitrogen 1.65 0.55 0.23
Methane
88.49 85.44 61.49
Ethane 5.79 7.72 17.38
Propane 1.01 1.95 7.70
iso-Butane 0.35 0.57 2.75
n-Butane
0.47 0.66
3.47
iso-Pentane 0.23 0.25 1.44
n-Pentane 0.16 .0.30 0.94
Hexanes 0.12 0.15 0.82
Heptanes plus 0.17 0.36 1.18
E 100.00 100.00 100.00
Q J
E
:::l
o
>
P s
p P r e s s u r e
. .>< .><
~ e ( . 1
:o '" O
t U ~ i
I D . ( . 1 " O
n.O.9 (")
~~ ::
"O - M Q)
e 1 E a.
~ ~ Q; ~
....::.:.a. _
m ~ ~ ~
~ TI ~ ~
uS!" ' " !'
Q)rJl u " ' O
cu - e e
(.1 O ea m
I I I I I I~ME~ 5
tU ~~ ~
. ~ m ~ [
oleD "O e
ME~= * ~ -e-- o u en
1 Ci>~~ :o &.
a. ~.. .!: E
I ~~2 ~ 8
S ~ ~~ ~ g
~Q )~o~~~ ~~o 'O-e
[~~o ;; ~ ~ ~ ~,...~ ~ s
Q,l Q) " ' O~! ' ~ e
(/)F ~rJ~ ~ ~
Clo (.1 s:
~o:g :g .E
I
~ t~.~~*
(f) C/) ID.... ... >-~
T""" tUEO o -;jctS
~ el tl t5 el::
-.~ ea ea cu ctI
O,D- - ...
M ~ ID Q) "'O o
EUE Eci~-;;;
c: .S.2 .2 ~ UtU
I III I l;o~~~~C>
'~~~oca~~
~o~~~o8g
Ci5CEi~x(/)';
cu cu 0- Q) c: ea (1)
ooce es os
! ISfiff ~ ::
P
i
- I n j e c t i on pr e s s u r e
P s - S a t u r a t i on pr e s s u r e
V
s
- S a t u r a t i on vol u me
Vi - Li Qu i d vol u me a t i n j e c t i on
pr e s s u r e
A V - S we l l i n g vol u me
Figure 3-13. Graphical representation of swelling test data.
or natural gas. It is advantageous to have a miscible displacement. This
means that at the gas/oil front only one phase is achieved no matter in what
proportion the oil and the gas are mixed. The injected gas and the reservoir
oil may be miscible by first contact or miscibility may be achieved as aresult
of a multiple contact process. If the gas and oil become miscible, complete
displacement of the reservoir fluid will be accomplished, and the recovery
may be larger than 90%. It is, therefore, of great practical and economical
interest to find out if the gas plus the oil form one phase, i.e., are miscible. If
so, it becomes important to know what the lowest pressure at the reservoir
temperature is where complete miscibility occurs in the displacement pro-
cess. This iscalled the minimum miscibility pressure and it isdetermined in
a slim tube test.
The minimum miscibility pressure may be determined from visual obser-
vations of the displaced fluido Usually it is determined as the breakover
point in the recovery curve from a series of displacement experiments. An
example of a displacement curve is illustrated in Figure 3-14, showing oil
recovery as a function of pressure at a fixed temperature. The concept of
minimum miscibility pressure is further discussed in Chapter 16.
A schematic diagram of a slim tube apparatus is shown in Figure 3-15.
The main part of the equipment is a piece of l(dn. steel tubing, approxi-
mately 10 mlong and packed with sand. Two containers, one for the injec-
tion gas and one for the reservoir oil, are connected to mercury pumps. Us-
ing the mercury pumps, the reservoir fluid and the gas may be forced to
flow through the steel pipe. A sight glass to observe the produced Huids and
a flash separator to determine the amount of fluids produced are provided.
The sand pack isfirst saturated with the oil sample to be studied, and the
initial volume of fluid in the coiled tubing is recorded. The pressure of the
Br e a k ove r
10 0
9 0
>-
Q;
80 >
o
(J
Q)
a::
70
;f.
60
1
10 0
P Oi n \
Figure 3-14. Example of oil
recovery as a function of
pressure in a slimtube experi-
ment.
I
o Mi s c i bl e
O I r n r n i s c i bl e
20 0 30 0
Test Pressure (bar)
system isset, and gas is allowed to displace oil through the sand pack. The
amount of produced fluid is monitored as time proceeds.
The sequence is repeated for several pressures, and the recovery, (pro-
duced oil)/(initial oil), isrecorded for each pressure. The resulting data may
be plotted as shown in Figure 3-16.
GAS
f
o
(])
1. - Me r c u r y pu mp
2. - Oi l r e s e r voi r
3. - Ga s r e s e r voi r
4. - S l l m t u be - s t e e l t u bl n g 1/ 4" a ppr ox. 10 m
5. - S l gh t gl a s s - vi s u a l f l ow c e l l
6. - Ba c k pr e s s u r e va l ve
7. - 0 11 a n d ga s s e pa r a t or
1,
Figure 3-15. Schematic diagram of slim tube apparatus.
....:.::::::::::::::::::::::::::::::/'
//
//
MMP
P
Figure 3-16. Pressure as a
function of recovery from a
slim tube analysis.
50 % 10 0 %
Re c ove r y
The pressure at which 90-100% of the initial fluid in the sand pack is
recovered is defined as the minimum miscibility pressure of the system.
Several modifications of this apparatus can be made. Online density de-
termination of the fluids using a high-pressure densitometer, and viscosity
measurements using a capillary tube viscosimeter, are examples of such
modifications.
GAS CONDENSATE EQUIPMENT
At reservoir conditions, gas condensates usually exist in one single gaseous
phase. Therefore, gas condensates must be studied using types of PVT cells
different from those used for black oils.
There are two main types of gas condensate PVT cells: the long win-
dowed cell (or three windowed cell), and the Sloane type of cell. Figures
3-17 to 3-20 show the different features of the cells.
The windows in the cells are located so that visual observation of the
whole or the bottom part of the hydrocarbon sample can take place.
The dew point isdefined as the pressure at which an infinitesimally small
amount of Iiquid is in equilibrium with a gas phase. At constant tempera-
ture, this can be observed as the pressure at which a small amount of liquid
appears in the system. For gas condensates, normally the upper dew point is
Figure 3-17. A Ruska three-windowed gas condensate/volatile oil cell in its ther-
mostatic enclosure.
'1~"
.:
.~
Oil and Gas Property Measurements
J
Figure 3-19. A cutaway diagram
of a Sloane type gas condensate
cell.
Hola and val ve tor
introduction 01 gas
oline
Spiral lor stirring
gas. The liquid is
stirred by a spiral 5
coil wound around
the axis
Lighting window 8
Liquid voluma 9
measuremant
chamber
55
Figure 3-18. Sloane type PVT
cell in its thermostatic enclosure.
1 Orilice lor introduc-
/ tion 01 mercury
through the upper
cap, lor operation
01 the pisto n
2 Piston with j oist
packing
Hole and valve lor
4 introduction and
sampling 01 gas
6 Sapphire viewing
window
~
7 Viewing window
1O Magnetic agitator
Figure 3-20. Pump and valve arrangement for ~Ioane type cel!.
of interest, i.e., the liquid is formed ("drops out") by pressure reduction.
The separator samples, gas and liquid, are prepared asfor black oil analy-
sisoThe PVT cell is charged, first with the gas, by allowing the gas to ex-
pand into the cell and displace the mercury. The volume of mercury dis-
placed can be measured, and thus the volume of gas in the cell is known.
Constant Mass Expansion
In order to achieve a representative reservoir fluid, separator liquid is
added to the cell in the gas/oil ratio measured during the sampling. Once
this is done, the system in the PVT cell is at reservoir temperature, corn-
pressed to above the recorded reservoir pressure. At this pressure, the cell is
left (usually overnight) for equilibration. The system ischecked to seeif it is
all in single phase, and the total system volume isthen recorded. A constant
mass expansion can start. Figure 3-21 shows a schematic diagram of this
process.
The pressure isreduced in astepwise manner. The system isobserved visu-
ally, and the volume recorded at each pressure level by reading the mercury
pump.
The dew point pressure can be somewhat difficult to observe, and there-
fore the stepwise pressure reduction is continued below the dew point.
"
Below the dew point, the liquid volume is determined at each pressure.
These readings result in the liquid dropout curve. Data from the constant
mass study can be plotted as shown in Figure 3-22. It should be noted that
the dew point is obtained by extrapolating the liquid dropout curve.
The amount of liquid formed is usually relatively small, and therefore
special attention must be paid to the liquid volume determinations. The
available cell designs oHer diHerent methods by which the liquid volumes
can be measured. Best results are obtained in cells where charge volumes are
up to 400 cm'' at the dew point. Significant liquid volumes will then be
formed, allowing relatively accurate liquid volume measurements. The
amount of liquid is best determined in a cell where the window configura-
tion allows light to shine through the system.
For the previous reasons, a Sloane type cell is recommended. It can be
used to study very lean gas condensate systems.
However, a Sloane type cell is not recomrnended for studies of volatile or
near critical fluids.
I NCI P I E NT F ORMAT I ON
, OF usuiu I
y
I
....
I
....
. . . .
1
,
,,,'
....
.... ....
l
.... . . . .
Ga s Ga s
..
Ga s I l. Ga s
vt , '" v
t 2
'" .... . . . . >
1
....
1
....
"
',' ...
f
. . . . . . . . M ....
I
v
t s
..
.... .... > . . . .
.... . . . .
1.....
....
p,' P s a t
P 2 > P s a t P 3 = P s a t P 4 < P s a t P S <P 4<P s a t
Figure 3-21. Schematic representation of constant mass expansion experiment
on a gas condensate.
Constant Volume Depletion
In order to simulate production behavior of the reservoir condensate
fluid, aconstant volume depletion experiment isperformed. A schematic di-
agram of the process is shown in Figure 3-23.
The system is brought to just below its dew point. The dew-point volume
of the system is known from the previous constant mass expansion study.
When the system is equilibrated at pressure P, the excess volume above the
dew-point volume isdisplaced into aflash separation system. Such asystem
is shown in Figure 3-24.
The flash system must be designed in such a way that it ispossible to col-
lect and measure the small amounts of liquid that are condensed out of the
Re l a t i ve vol u me
Q) \ -,-,- Const ant mas s
E
"
_0_0_ o
o
>
.:
Q)
.e.
E
"
Q)
E
o
"
>
'O
e,
~\
>
o
-
4-
'O
o
>
~
Q)
>
+'
+-'
" o
<O
CL
Qj
o
<-
o::
-o
-o
"
cr
~~
--'
""
4 r
'~'--
. 8
I ! I l:a ' -te I
f 40 0 r 50 0
DP 361. 2 P r e s s u r e , [ ba r l
Re s e r voi r P r e s s u r e
10 0 20 0 30 0
Figure 3-22. P-V and liquid dropout curves ter a recombined sample at 121oc.~
11:
)11
;)1
J Ii\
-I,
;;1.,
TR ~ G."II
:::. :::: r:"
~: G" . >".. >'
>"" : : : : l.. :.v G". ',1 G"
_ . Gas. >'" . .. >
>~.... ' V
A
P, - Psat
11
'"
>'"
A", 1: Gas:
> ...
1
1
.:: ..
....
I Gil
P
2
< P
sat
P
2
< P
sat P3<P2<Psat P2 < Psat
Figure 3-23. Schematic representation 01constant volume depletion experiment
on gas condensate.
Figure 3-24. Left side:
Flash separation system
lor gas condensate mix-
tures. Right side: Gaso-
meter tor gas volume
measurements.
produced gas stream. The total gas volume can be measured using agasom-
eter.
From compositional analyses of the gas and the liquid phases, and by
knowing the volume of each phase, aproduced well stream composition can
be determined.
The pressure is further reduced, and more liquid is formed. The amount
of liquid in this phase is measured, and the excess volume above the dew-
point volume isdisplaced out of the cell at constant pressure. The amount of
produced gas is again recorded.
This stepwise pracedure isrepeated until atmospheric pressure isreached.
At the end of the analyses a molar balance is performed to check if some
fluid has been lost during the experimento The balance may be written as:
No moles charged to the cell =No moles produced from the cell.
The agreement should be within 1% in order to avoid errors in the subse-
quent calculations from the anaIysis.
TypicaI data from a constant voIume depletion can be seen in TabIe 3-3.
Figure 3-25 shows the difference in the two liquid. dropout curves from
constant mass expansion and constant volume depIetion, respectiveIy.
VOLATILE FLUm EQUIPMENT
Some of the naturally occuring hydrocarbon mixtures can be classified as
voIatile fluids (see Chapter 1) or near-critical fluids.
The PVT behavior of these systems isdifficuIt to analyze, because the gas
phase and liquid phase are, at reservoir conditions, virtualIy impossible to
distinguish.
These fluids are normally anaIyzed as for gas condensates (see"Gas Con-
densate Equpment").
A long windowed cell, as shown in Figure 3-26, is often used. This cell
allows viewing of the whoIe sample at any stage of the anaIysis.
The cell is charged, and the pressure in the cell is raised by means of a
mercury pump. The sample is Ieft at reservoir conditions to equilibrate.
The pressure isthen reduced in the cell, and observations are made to de-
termine the saturation pressure. In the case of avolatile fluid it may be diffi-
cuIt to seewhether aphase transition isadew or abubble point, becausethe
two phases are nearly identical.
Phase changes are rapid, and a near criticaI system exhibits Iarge changes
in its PVT and phase behavior for relatively small temperature and pressure
Gil and Gas Property Measurements 61
Table 3-3
Constant Volume Depletion Data at 121C Hydrocarbon Analysis of
Produced Wellstream
Component
Reservoir Pressure (bar)
362.2 307.7 274.4 222.4 171.5 120.2 71.8
Carbon dioxide 9.16 9.14 9.29 9.25 9.39 9.31 9.20
Nitrogen 0.64 0.72 0.74 0.69 0.74 0.63 0.64
Methane 68.80 70.05 70.27 70.68 71.30 71.42 69.12
Ethane
8.43 8.54 8.59
8.44 8.34 8.51 8.78
Propane 5.11 5.13 5.18 5.06 4.92 5.02 5.70
iso-Butane 0.81 0.80 0.81 0.79 0.75 0.76 0.97
n-Butane 1.45 1.42 1.45 1.39 1.33 1.35 1.79
iso-Pentane 0.52 0.49 0.50
0.49 0.46 0.46 0.68
n-Pentane 0.53 0.50 0.51 0.49 0.46 0.46 0.72
Hexanes 0.63 0.54 0.55 0.55 0.51 0.53 0.80
Heptanes 0.83 0.62 0.64 0.66 0.62 0.68 0.81
Octanes 0.95 0.55 0.52 0.59 0.59 0.59 0.54
Nonanes 0.52 0.27 0.21 0.26 0.24 0.19 0.16
Decanes + 1.62 1.23 0.74 0.66 0.35 0.09 0.09
L 100.00 100.00 100.00 100.00 100.00 100.00 100.00
Mol. wt. 01C1O+ 200 180 168 157 151 157 147
Density 01C1O+ 0.856 0.828 0.827 0.821 0.819 0.817 0.817
Gas phase compressi-
bility lactor, Z 1.003 0.946 0.922 0.882 0.882 0.903 0.934
Mole % produced O 9.6 7.6 13.4 14.8 16.3 15.5
Cum. mole % produced O 9.6 17.2 30.6 45.4 61.7 77.2
Molecular weight 28.31 26.47 25.60 25.42 24.81 24.49 25.44
Two-phase Z-Iactor 1.003 0.943 0.918 0.887 0.870 0.869 0.872
Viscosity (cP) 0.0332 0.0271 0.0246 0.0215 0.0187 0.0164 0.0147
changes. Figure 3-27 shows the typicaI, Iarge, liquid dropout curve changes
over a temperature range of only 20C.
Once the saturation pressure of the system has been found, and the con-
stant mass study has been performed, the work program of the anaIysis fol-
lows that of a gas condensa te system.
o -- o Constant Mass
Constant Vol ume
~
o
..
~
8. 5%
e
-o
-o
!::v.----~
2~I
1%
I I I I ~.
t 400 t 500
D.P. Res. pressure
373.5
100 200 300
Pressure. [bar)
Figure 3-25. Liquid dropout curves tor a gas condensate mixture trom a constant
mass expansion and a constant volume depletion experiment.
Figure 3-26. Long-windowed cell used tor volatile tluids, coupled to a capillary
tube viscosimeter (measurement ot viscosity section).
100
~
~~
~
~
~M
~
:ro
o
..
: ~
~
cr
: ~
~
~
ew
~
w
~
~
~
10
w M
O_O_O_O--a-O"t
b
.-e_~===-t:;:::::e~A-tr-~tB
e-e
~
Constant mass expansion at 150 'C
Constant mass expansion at 140 "C
Constant mass expans!on at 130 'C
200 280 320 360
P [bar)
240 120 160
Figure 3-27. Liquid dropout curves tor a volatile fluid. Note that a dew point is
obtained at 130C, whereas bubble points are obtained at 140 and 150C.
MEASUREMENT OF VISCOSITY
An important property of ahydrocarbon system isits viscosity at relevant
pressures and temperatures.
A few methods have been developed for these measurements at reservoir
pressure and temperature. The most commonly used methods are:
l. Rolling ball viscosimeter
2. Capillary tube viscosimeter
Rolling Ball Viscosimeter
Figure 3-28 shows aschematic diagram of arolling ball viscosimeter, and
Figure 3-29 shows aphotograph of atypical unit. It consists of the fall tube,
an expansion chamber, and a steel ball of known dimensions.
A hydrocarbon Huid istransferred into the evacuated viscosimeter at con-
stant pressure. The system must be equilibrated at above reservoir pressure
for several hours before the analysis can start.
SOLENOID
Figure 3-28. Schematic dia-
gram of a rolling ball visco-
simeter. ~OUTlET
BARREL SEAL
GAS CHAMBER
INLET
Figure 3-29. Photograph of a rolling ball viscosimeter.
Prior to charging, the viscosimeter must be calibrated with known cali-
bration fluids. The ball sizeshould be chosen so that the calibration curve is
linear in the range of operations, normally 0.2-3 cP. Calibrations are made
at each of the three angular positions in which the viscosimeter can be
placed.
Tostart the viscosity determination, the top valve isclosed in order to seal
the upper end of the ball tube. The apparatus isturned upside down, allow-
ing the steel ball to roll to the topo The valve stem of the top valve is made
magnetic by acoil, sothat the ball will stay at the top when the apparatus is
turned back to measurement position.
f1
v-
Gil and Gas Property Measurements 65
Using aswitch, the magnet is deactivated, and the ball starts to fal!. The
time it takes to reach the bottom of the tube is recorded.
With this time and the known density of the fluid, its viscosity can be
determined Irom the calibration curves.
This method is generally found to be reliable; it is, however, rather cum-
bersome to operate.
Capillary Tube Viscosimeter
With the capillary tube viscosimeter, the viscosity of flowing fluid is cal-
culated using the flow equation:
r =D2 L lP/32 Lv (3-10)
where D =diameter
L =length of the capillary tube
LlP =pressure drop
v =velocity of the fluid flowing through the capillary tube
The equipment used isshown schematically in Figure 3-30. In addition, a
typical capillary tube viscosimeter is shown in Figure 3-26.
Reservoir fluid is charged to the adjoining PVT cell and equilibrated at
reservoir conditions or above.
G)
1. Me r c u r y pu mp
2. P VT c e l l , vi s u a l
3. Coi l e d c a pi l l a r y t u be
4. OP t r a n s du c e r
5. Ba c k pr e s s u r e va l ve
6. Di l a n d ga s s e pa r a t or
Figure 3-30. Schematic diagram of a capillary tube visccsirneter.
Tomeasure viscosity, the fluid ischarged to the coiled capillary tube, and
equilibrated. Flow of fluid through the capillary is established using a mo-
torized pump. The pressure drop across the coil is monitored at constant
flow rateo The temperature and pressure are kept constant.
The process can be repeated at different pressures above the saturation
pressure, and the viscosity calculated from Equation 3-10.
Since the coiled capillary tube does not have a perfectly homogenous in-
ner-diarneter, it must be calibrated with calibration fluids of known vscos-
ity.
SURFACE TENSION
The surface tension between two fluids, as defined in Chapter 13, is an
important fluid property in the evaluation of reservoir Huid flow. Surface
tension exists between a water phase and an oil phase, between a water
phase and a gas phase, and between an oil phase and a gas phase. The sur-
face tension can, with some difficulty, be measured at reservoir conditions
using either apendant drop or aspinning drop apparatus. Both methods are
optical, in the sense that the measurements are made on images of bubbles of
one fluid suspended in the other.
Pendant Drop Tensiometer
The basic apparatus, consisting of ahigh-pressure sight glass or PVT-cell,
is shown in Figure 3-31. This can be charged with brine at reservoir condi-
tions. Through acapillary tube extending into the cell, an oil droplet isseen
suspended frorn the tip of the capillary tube. The shape of the droplet (see
Figure 3-32) will depend on the interfacial tension between the fluds, and
the interfacial tension can be calculated from the droplet's geornetry (Ci-
rault et al., 1984 and Adamson, 1976) using the following expression:
a =/::; .pgDelH (3-11)
where /::; .p =difference in the density of the two fluids
g =gravitational acceleration
De =equatorial diameter (seeFigure 3-32)
D, =distance shown in Figure 3-32
lIH =tabulated function of D,lDe (Adamson, 1976).
,~
{{
Figure 3-31. Pendant drop
apparatus.
Figure 3-32. Pendant
drop.
1\1
!
~
Spinning Drop Tensiometer
This method can only be used for liquid-liquid measurements. A glass
tube is filled with formation brine and an oil droplet is introduced into the
tube. The tube isspun at ahigh speed. The oil bubble can be viewed using a
strobe light. The shape of the bubble will depend on the interfacial tension
between the fluids. The interfacial tension is calculated as described by
Chatenay et al. (1982).
REFERENCES
Adamson, A. W., "Physical Chemistry of Surfaces," J ohn Wiley and Sons,
lnc., 1976.
Chatenay, D. et al., "Measurernent of Low lnterfacial Tension, Comparison
Between a Light Scattering Technique and the Spinning Drop Tech-
nique," I. Dspersion Science and Technology, 3, 1982, pp. 245-260.
Girault, H. H. J ., Schiffrin, D. J ., and Smith, D. D. v., "The Measurement
of lnterfacial Tension of Pendant Drops Using aVideo Image Profile Dig-
itizer," [ournal of Colloid and Interface Science, 101, 1984, pp. 257-266.
~.
"~,/
Chapter 4
Composition and Property Data
Ten petroleum reservoir fluid compositions have been selected to test the
flash calculation procedures of Michelsen described in Chapter 6 coupled
with the characterization procedure of Pedersen et al. described in Chapter
7. The compositions-comprse five gas condensate mixtures, one volatile oil
and four black oils (see Chapter 1 for a description of these kinds of mix-
tures). Six of the mixtures are from the North Sea, two from the U.S., and
two from the Middle East. A survey of the mixtures is given in Table 4-1,
and the mixture compositions are shown in Tables 4-2 through 4-12. The
composition data for mixtures one through four comprise both weight and
molar compositions. For the remaining mixtures only molar compositions
are given.
Si~ulation results for the mixtures are shown in Chapters 8 and 9.
,
i I
11
I
Table 4-1
Survey of the Mixtures tor which Molar Compositions Are Given
t
I
Mixture Composition
No. Type of Mixture Origin in Table
1 Gas Condensate North Sea 4-2
2 Black Oil North Sea 4-3
3 Black Oil North Sea 4-4
4 Volatile Oil North Sea 4-5
5 Gas Condensate North Sea 4-6
6 Black Oil North Sea 4-7 and 4-8
7 Black Oil Texas 4-9
8 Gas Condensate Texas 4-10
9 Gas Condensate Iran 4-11
10 Gas Condensate Bahrain 4-12
' ' 0
Table 4-2
Composition of Mixture 1 (North Sea Gas Condensate)
Molecular Density (g/cm
3
)
Component Weight % Mole % Weight at 15C, 1 atm
N2 0.571 0.60
CO2 5.031 3.34
C, 40.667 74.16
C2 8.126 7.90
C3 6.254 4.15
ic, 1.401 0.71
n-C. 2.855 1.44
ic, 1.306 0.53
n-C, 1.637 0.66
C6 2.355 0.81
C, 3.749 1.20 91.2 0.746
Ca 4.100 1.15 104.0 0.770
.
C. 2.577 0.63 119.0 0.788
Cl O 2.329 0.50 133.0 0.795
Cl 1 1.466 0.29 144.0 0.790
C'2
1.458 0.27 155.0 0.802
C'3 1.624 0.28 168.0 0.814
C,. 1.413 0.22 181.0
.
0.824
c., 1.165 0.17 195.0 0.833
C'6 1.057 0.15 204.0 0.836
C17 1.096 0.14 224.0 0.837
c., 0.729 0.09 234.0 0.839
C,. 1.137 0.13 248.0 0.844
C20+ 5.896 0.47
362.0 0.877
Composition and Property Data 71
Table 4-3
Composition of Mixture 2 (North Sea Black 011)
3
)
N2 0.145 0.56
CO2
1.450 3.55
C, 6.757 45.34
C2 1.531 5.48
C3 1.516 3.70
ic, 0.378 0.70
n-C. 0.891 1.65
ic, 0.489 0.73
n-Cs 0.580 0.87
C6 1.043 1.33
C, 2.276 2.73 89.9 0.757
Ca 3.125 3.26 103.2 0.777
C. 2.342 2.14 117.7 0.796
Cl O 2.379 1.94 133.0 0.796
Cl 1 2.205 1.62 147.0 0.800
C'2 2.179 1.47 160.0 0.815
C
'3
2.693 1.69 172.0 0.833
C, 2.789 1.62 186.0 0.843
C,S 2.937 1.59 200.0 0.849
C
'6
2.553 1.30 213.0 0.858
C17 2.388 1.11 233.0 0.851
c., 2.885 1.26 247.0 0.856
C, 2.571 1.07 258.0 0.868
C20+ 51.898 13.32 421.0 0.914
Table 4-4
Composition of Mixture 3 (North Sea Black 011)
3
)
N2 0.091 0.403
CO2 0.353 1.000
C, 5.845 45.396
C2 1.014 4.202
C3 0.314 0.887
te, 0.262 0.561
c, 0.242 0.518
ic, 0.375 0.647
n-Cs 0.170 0.294
Ca 0.678 1.011
C7 2.118 2.878 91.68 0.745
Ca
3.467 4.114 104.99 0.767
Cg 3.135 3.322 117.60 0.798
ClO+ 81.936 34.767 293.53 0.902
Table 4-5
Composition of Mixture 4 (North Sea Volatile 011)
3
)
N2 0.258 0.58
CO2 2.297 3.27
e, 13.780 53.89
C2 4.108 8.57
C3 4.254 6.05
-c, 0.969 1.05
n-C, 2.263 2.44
ic, 1.013 0.88
n-es 1.348 1.17
Cs 1.970 1.45
-q
Composition and Property Data 73
Table 4-5
Continued
f:
~
~
3
)
e 7
3.489 2.38 91.9 0.742
ea 4.331 2.59 104.7 0.765
Cg 3.329 1.75 119.2 0.788
ClO 3.173 1.50 131.0 0.791
C"
3.666 1.55 147.0 0.796
C' 2
2.408 0.93 161.0 0.811
C' 3
3.125 1.13 171.0 0.826
C"
2.952 1.01 182.0 0.836
c., 2.521 0.80 195.0 0.843
c., 2.878 0.86 208.0 0.848
C17 2.211 0.60 228.0 0.844
c., 2.701 0.68 247.0 0.848
c., 2.184 0.54 252.0 0.859
C20 + 28.773 4.34 411.0 0.903
~,t
~~"
Table 4-6
Composition of Mixture 5 (North Sea Gas Condensate)
3
)
PNA Distribution Mole %
Component Mole % Weight at 15C, 1 atm P N A
N2 0.12
CO2 2.49
C, 76.43
C2
7.46
e3 3.12
ic, 0.59
n-C, 1.21
ic, 0.50
n-Cs 0.59
Cs 0.79
(table continued 011 next page)
Table 4-6
Continued
Molecular Density (g/cm')
PNA Distribution Mol.e %
Component Mole % Weight at 15C, 1 atm P N A
C, 0.95 95 0.726 0.564 0.361 0.076
Cs 10~ 106 0.747 0.113 0.611 0.277
C9 0.78 116 0.769 0.483 0.311 0.206
C'O 0.592 133 0.781 0.530 0.275 0.195
C11 0.467 152 0.778 0.681 0.193 0.'26
C'2 0.345 164 0.785 0.757 0.123 0.120
C
'3
0.375 179 0.802 0.709 0.183 0.108
C, 0.304 193 0.815 0.635 0.209 0.156
C,S 0.237 209 0.817 0.729 0.168 0.103
C'6 0.208 218 0.824 0.624 0.232 0.H4
C'7 0.220 239 0.825 0.668 0.185 0.147
C,S 0.169 250 0.831 0.675 0.192 0.133
C'9 0.140 264 0.841 0.652 0.190 0.158
C20+ 0.833 377 0.873 0.519 0.320 0.161
Source: Pedersen et al. (1985).
Table 4-7
Composition of Mixture 6(North Sea Bla<:kOil)
3
)
Component Mole % Weight at 15C, 1 atm
N2 0.34
CO2 0.84
C, 49.23
C2 6.32
C3
4.46
ic, 0.86
n-C, 2.18
i-C5 0.93
c, 1.33
,
C6 2.06
Composition and Propertq Data
1
,,-
<v,
(:1/
. -1:: I
75"I
') (
'--1
" Table 4-7
Continued
"1
3
)
C, 3.33 99 0.7395
Cs 4.06 106 0.7518
C9 2.76 120 0.7756
ClO 1.33 139 0.7930
C11 1.79 146 0.7902
C
'2
1.70 160 0.8060
C
'3
1.81 174 0.8203
C,
1.46
194 0.8311
C
'5
1.49 205 0.8446
C'6 1.08 218 0.8515
C
'7
1.13 234 0.8542
C,S 0.99 248 0.8561
C'9 0.88 265 0.8663
C20+ 7.64 465 0.9350
.
Table 4-8
TBP Data for Mixture 6
Temp, Cum. Wt % Temp, Cum. Wt % Temp, Cum. Wt %
C Distilled C Distilled C Distilled
100 10.54 218 28.43 288 43.10
127 14.54 237 32.48 304 46.52
152 19.17 255 36.13 318 49.67
176 21.55 272 40.06 332 52.67
197 24.92
Densily 01 TBP sample al 15C, 1 atrn =0.8499 (g/cm
3
)
76 Properiies of Oils and Natural Gases
Table 4-9
Molar Composition of Mixture 7 (Black Oil from Southwest Texas)
C, 52.00
C2
3.81
C, 2.37
ic, 0.76
n-C, 0.96
ic, 0.69
n - Cs 0.51
C6
2.06
C7 2.63 99 0.749
Cs 2.34 110 0.758
Cg 2.35 121 0.779
ClO+
29.52 221 - 0.852
Source: Hoffman el al. (1953).
Table 4-10
Molar Composition of Mixture 8 (Gas Condensate from Southwest Texas)
C, 91.35
C2
4.03
C3
1.53
ic, 0.39
n-C,
0.43
ic, 0.15
e, 0.19
C6
0.39
C7 0.361 100 0.745
Cs
0.285 114 0.753
Cg 0.222 128 0.773
ClO+
0.672 179 0.814
Source: Hoffman el al. (1953).
Compositiol1 and Propertu Data 77
Table 4-11
Molar Composition of Mixture 9(Gas Condensate from Iran)
,N2 0.08
CO2
2.44
C, 82.10
. C2 5.78
e , 2.87
- C, 0.56
-c, 1.23
ic, 0.52
e, 0.60
C6 0.72
C7+ 3.10 132 0.774
Source: Firoozabadi el al. (1978).
Table 4-12
Molar Composition of Mixture 10 (Gas Condensate from Bahrain)
N2 11.71
CO2 6.50
1; 125 0.05
C, 79.06
C2 1.62
C3 0.35
- C, 0.08
c, 0.10
ic, 0.04
n - e s 0.04
e 6 0.06
C7 0.06 97 0.7243
Cs 0.05 111 0.7413
Cg 0.04 124 0.7611
ClO+ 0.24 195 0.8238
Source: AI-M::.hrnnc: ::Inri Ti,,!:! (1 Q Cl7\
78 Propertieso} Oils and Natural Gases-
REFERENCES
Al-Mahroos, F. M. and Tjoa, G. H., "Analysis of the Phase Behavior of
Khuff Gas System in Bahrain Field," SPE Paper 15766, 1987.
Firoozabadi, A., Hekim, Y, and Katz, D. L., "Reservor Depletion Calcula-
tions for Gas Condensates Using Extended Analyses in the Peng-Robinson
Equation of State," The Canadian lournal o] Chemcal E ngneerng, 56,
1978, pp. 610-615.
Hoffmann, A. E., Crump, J . S., and Hocott, C. R., "Equilibrium Con-
stants for aGas-Condensate System," Petroleum Transactons AIME ,198,
1953, pp. 1-10.
Pedersen, K. S., Thomassen, P, and Fredenslund, Aa., "Thermodynamics
of Petroleum Mixtures Containing Heavy Hydrocarbons. 3. Efficient
Flash Calculation Procedures Using the SRK Equation of State," Ind.
E ng. Chem. Process Des. Dev., 24, 1985, pp. 948-954.
,':Z)'
Chapter 5
Equations of State
Equations of state (EOS) are, for pure substances, mathematical relations
between volume, pressure, and temperature. For mixtures, EOS in addition
include composition. EOS are very versatile tools for engineering applica-
tions. They can be used for all states of matter (mostly gas, vapor, and liq-
uid), and they can describe transitions between states.
The development of EOS for the representation of real systems started in
1873with the publication of the van der Waals (vdW) EOS. Since then hun-
dreds of different EOS have been proposed. It is beyond the scope of this
book to review in detail all these developments. In fact, all of the oil/gas
pressure-volume-temperature and phase equilibria computations are based
on one equation of state. For broad reviews of EOS, the reader isreferred to
Abbott (1979), Leland (1980) and Tsonopoulos and Heidman (1986). In ad-
dition, Tzouvaras (1986) gives a particularly clear and detailed account of
the different types of EOS.
TYPES OF EQUATIONS OF STATE
Leland (1980) distinguished between four families of EOS:
l. van der Waals family
2. Benedict-Webb-Rubin family
3. Reference-fluid equations
4. Augmented-rigid-body equations
The vdW family encompasses simple, mostly cubic EOS. Their main
characteristic is the separation between the repulsive and attractive effects.
79
Despite their simplicity, these EOS display quantitatively correct perfor-
mance, even being able to describe multiphase equilibria, tricritical points,
and other complicated phenomena (Michelsen and Hedemann, 1988). Cu-
bic EOS are discussed further in the next section.
The Benedict-Webb-Rubin (BWR) family includes complicated EOS and
are empirical extensions of the virial EOS. Besides the Benedict-Webb-Ru-
bin (1940) EOS itself, the most significant members of this family for oil/gas
application are the Starling (1973) EOS, and the Lee and Kesler (1975)
EOS.
The reference-fluid equations aim at accurate representation of a large
amount of PVT data for pure substances. The equations contain many pa-
rameters. An important application of reference-fluid equations istheir use
as reference fluids in corresponding state theories, see e.g., MolIerup and
Rowlinson (1974).
The augmented-rigid-body family combines description of repulsive
forces between hard molecules of different shapes with expressions for the
molecular attractions. The sound theoretical basis of the repulsive term is
the most important and promising characteristic of these EOS. The Cama-
han and Starling (1972) expression for the repulsion between hard spheres
has been repeatedly used. Approximations to th~ augmented-rigid-body
EOS include simplfcatons of the expression for the repulsive term, such as
in the Cubic-Chain-of-Rotators EOS (Kim et al., 1986).
Today alI four families of EOS receive considerable attention. Recent lit-
erature, however, mainly contains developments in EOS of the vdW family,
because of their simplicity and practicality, and of the augrnented-rigid-
body EOS, because of their theoretical foundation and potential for accu-
rate extensions.
The oil/gas PVT- and phase equilibrium computations in this book are
based on a cubic equation of the vdW family. The reasons for this choice
are:
1. Cubic EOS of the vdW family yield relatively simple expressions for
the thermodynamic properties and phase equilibrium relationships of
interest.
2. The EOS of the other three families, though more complicated in na-
ture, do not give quantitatively better descriptions of mixture phase
transitions than do cubic EOS (Tsonopoulos and Heidman, 1986).
It is welI known that the more complicated EOS yield better predictions
of liquid densities than the cubic EOS. This disadvantage of the cubic EOS
is taken into account.
E quations o] State 81
CUBIC EQUATIONS OF STATE
Cubic EOS are explicit regarding pressure and may be written in terms of
a repulsive and an attractive contribution to the pressure P:
P =prep _ pattr
(5-1)
Classical examples of cubic EOS are those of van der Waals (1873):
P =RT/(V - b) - a/V2
(5-2)
and Redlich and Kwong (1940) (R-K):
P =RT/(V - b) -;- a/[T05 V(V +b)]
(5-3)
where R = gas constant
T =temperature
V =molar volume
There are two parameters; b represents the "real," or hard-sphere volume of
the molecules, and a represents the intermolecular attraction.
The parameters a and b in Equations5-2 and 5-3 are normalIy deter-
mined by imposing the critical conditions:
[(aPlaVh =(a2Pla
v2
h =O]crit. point (5-4)
The vdW and R-K EOS may be extended by altering their ternperature-
dependence and by including parameters in addition to a and b. J ensen
(1987) defines two generalized cubic EOS, from which many of the com-
monly used equations may be derived:
Type 1: The repulsive-terrn modified R-K EOS
P =[RT(2V +b(T) )]/[V(2V - b2(T))] - a(T)/[V(V +b
3
(T))] (5-5)
Type 2: The attractive-term modified vdW EOS
P =RT/[V - b(T)] - a(T)/[(V +b2(T))(V +b
3
(T))]
(5-6)
Equations 5-5 and 5-6 each contain four temperature-dependent pararne-
ters: a, b, b2, and b
3
Some of the EOS that may be derived from Equations 5-5 and 5-6 are:
~
,~!
Soave-Redlich-Kwong (SRK), (Soave, 1972)
P =RT/(V - b) - a(T)/[V(V +b)]
Type 2: a = a(T), b, = b, b2 = O, b, = b
The Peng-Robinson EOS (PR), (Peng and Robinson, 1976)
P =RT/(V - b) - a(T)/[V(V +b) +b(V - b)]
Type 2: a = a(T), b = b, b2 = (1 +205)b, b, = (1 - 205)b
The Adachi-Lu-Sugie EOS (ALS), (Adachi et al., 1983)
P =RT/(V - b) - a(T)/[(V - b2)(V +b3)]
Peneloux's modified SRK EOS, (Peneloux et al., 1982)
P =RT/(V - b) - a(T)/[(V +c)(V +b +2c)] (5-10)
Type 2: a = a(T), b = b, b2 = e, b3 = b +2c
The Ishikawa-Chung-Lu EOS, (Ishikawa et al., 1980)
,
i
P =[RT(2V +b(T) )]/[V(2V - b(T))]
- a(T)/[T05V(V +b(T))]
Type 1: a = a(T), b, = b(T),b2 = b(T), b, = b(T)
(5-11)
Many more examples exist (J ensen, 1987). The presence of parameters in
addition to a and b as in the R-K and vdW EOS permits inclusion of con-
straints other than those given by Equation 5-4. Careful evaluations in-
formed the choice to base the oil/gas computations on the SRK EOS. No
other cubic EOS tested yielded better results for naturally occurring oil and
gas mixtures. The choice of the SRK EOS (and modifications thereof) as the
basis for the oil/gas property and phase equilibrium calculations isjustified
by the many results shown in subsequent chapters.
THE SRK-EQUATION OF STATE
The analytical form of the SRK-equation of state isgiven in Equation 5-7.
The a and b parameters are found from the pressure-volume relationship at
(5-7)
(5-8)
(5-9)
the critical point expressed in Equation 5-4. By inserting the expression for
the pressure given in Equation 5-7 into Equation 5-4 the following relations
may be derived for the pure component a-parameter at the critical point
and for the pure component b-parameter:
llc=0.42747 . R2. T2e/Pe
(5-12)
b, =0.08664 . R . Te/Pci (5-13)
Values for Te and P, of some common petroleum mixture constituents are
given in Table 5-1. The temperature dependence of the a-parameter enter-
Table 5-1
Critical Temperature (Tc), Critical Pressure (pc) and Acentric Factor (w) of
Some Common Petroleum Mixture Constituents
Te (Kelvin) Pe (atm) w
N2 126.2 33.5 0.040
CO2 304.2 72.8 0.225
H2S 373.2 88.2 0.100
C, 190.6 45.4 0.008
C2 ' 305.4 48.2 0.098
C3 369.8 41.9 0.152
te, 408.1 36.0 0.176
n-C. 425.2 :f7.5 0.193
ic, 460.4 33.4 0.227
n-Cs 469.5 33.3 0.251
n-C6
507.4 29.3 0.296
Source: Reid el al. (1977).
ing Equation 5-7 is expressed in the form of a term ct(T) which multiplied
with aci gives the final expression for the a-parameter of the SRK-equation:
a(T) =acct(T) (5-14)
The a-parameter is obtained from the following expression:
ct(T) =(1 +m(l - T~i5))2 (5-15)
(5-16) where m =0.480 +1.574w - 0.176w 2
It is seen from Equation 5-15 that O'(T) =1 at the critical temperature,
where a therefore becomes equal to 8.c.In Equation 5-15 Wis the acentric
factor which is defined as follows (Pitzer, 1955):
W= - IOglOPri
ap
(at Tr = 0.7) - 1 (5-17)
where Pri
ap
is the reduced vapor pressure (P/Pci) of component i.
The acentric factor may be interpreted as the non-sphericity of molecules
of type i. Values for Wof a number of different components are given in
Table 5-1. The dependence of mon Wgiven in Equation 5-16 isfound from
experimental vapor pressure data for aliphatic hydrocarbons ranging from
CitO C10 (Soave, 1972).
Table 5-2
Non-Zero Binary Interaction Coefficients for Use with the SRK-Equation
of State Between Some Common Petroleum Mixture Constituents
Component
-
Pair N, CO, H, S
N2 0.00 0.00 0.00
CO2 0.00 0.00 0.12
H2S 0.00 0.12 0.00
C, 0.02 0.12 0.08
C2 0.06 0.15 0.07
C3 0.08 0.15 0.07
ic, 0.08 0.15 0.06
n-C. 0.08 0.15 0.06
i-C5 0.08 0.15 0.06
n-C5 0.08 0.15 0.06
n-C6 0.08 0.15 0.05
Source: Reid el al. (1977).
For a mixture, a and b are found as follows:
a =E E zza
j
(5-18)
b =E z.b, (5-19)
E quatons of State 85
-
where z, and z are mole fractions of components i and j , respectively and:
a=(aa)1I2( 1- k)
(5-20)
kij is a binary interaction coefficient, which is usually consdered equal to
zero for hydrocarbon-hydrocarbon interactions, and different from zero for
nteractions between a hydrocarbon and a non-hydrocarbon, and between
unlike pairs of non-hydrocarbons. Table 5-2 shows the non-zero binar y in-
teraction coefficients to be recommended for use with the SRK-equation
(Reid et al., 1977) for some lighter petroleum mixture constituents.
Equation 5-7 can be rewritten in terms of the compressibility factor, Z:
Z3 - Z2 +(A - B +B2)Z - AB =O
(5-21)
where Z is defined ;ts:
Z =PV/RT
(5-22)
and A and B are given by the following expressions:
A =a(T)P/R2T2
(5-23)
B =bP/RT
(5-24)
.~
~:
The polynomial of Equation 5-21 may have one or three real roots. When
for apure component or asingle-phase mixture only one real root exists, this
root equals the compressibility factor of the phase presento The compress-
ibility factor of a dilute gas is clase to unity. The compressibility factor of a
near critical pure component is clase to 0.33. Using the SRK-equation of
state the compressibility factor of apure component at the critical point will
always be found equal to 0.33. In the case where the polynomial of Equa-
tion 5-21 has three real roots, one or more phases may be found. The num-
ber of phases, and the amounts and compositions of each phase may be de-
termined as described in Chapter 6.
In the case of two phases, each component will have equal fugacities, f,
in both phases:
~., fr = ff"
(5-25)
where V and L refer to the vapor and liquid phases, respectively. The crite-
rion of equal fugacities indicates equality in the driving forces in the transfer
of components from one phase to the other. The fugacity is related to the
chemical potential, jl, as follows:
86 Properties oj Oils and. NaturalCases
.t=.t?+RT In f (5-26)
where .t0 is a standard-state chemical potential for component i. Equation
5-25 can be rewritten:
cfJry = cfJ! -x or K = cfJrN r (5-27)
where Xand y =mole fractions of component i in the vapor and liquid
phases, respectively
cfJ =fugacity coefficient
K =(y/x)iscalled the equilibrium ratio. The following general thermody-
namic relationship exists for determination of the fugacity coefficient of
component i in the mixture:
v
In cfJ =- l/RT J 00 ((ap/anh,v,nj - RT/V) dV - In Z (5-28)
where n, is the number of moles of type i. When the SRK-equation is used
Equation 5-28 yields: .
In cfJ =(b/b)(Z - 1) - In Z +In [V/(V - b)]
+a/bRT [b/b - 2 ~ Zj(l- kj)(aaj)05/alln [~V +b)/V] (5-29)
The phase compositions are related to the total composition as follows:
x, =z/(l +(K - 1)/J ) (5-30)
y =Kz/(l +(K - 1)13) (5-31)
where Z=mole fraction of component i in the total mixture
13 =the molar vapor phase fraction
The SRK-equation may also be used for calculation of other thermody-
namic properties, e.g., the enthalpy and the entropy. The enthalpy, H of a
E quations oj State 87
pure component, or a mixture of given composition, may be calculated as
the sum of two contributions, namely the ideal gas enthalpy and the resid-
ual enthalpy:
H =Hd+H""
(5-32)
and similarly for the entropy:
s =Sd+Sres
(5-33)
The ideal gas terms at temperature T are usually calculated from the follow-
ing equations:
T
Hd= EZHi
d
= EZ J qd dT
i i Tref
(5-34)
Sd=E z(Si
d
- R In z.)
= E Z(J T (qd/T) dT - R In P/Pref - R In z)
i Ttef
(5-35)
where Tref is a reference temperature, here OC, and Pref is the reference
pressure, here 1 atm, and C~the ideal gas heat capacity, which is usually
approximated by a third-degree polynomial in the temperature
C~ =C +C2T +C3T2 +C4T3 (5-36)
Values for the coefficients C-C4 of the lighter petroleum mixture constit-
uents are given by Reid et al., (1977). A procedure for estimating C-C4 of
heavy hydrocarbons ispresented in Chapter 7. The residual terrns of Equa-
tions 5-32 and 5-33 may be derived from the SRK-equation using the follow-
ing general thermodynamic relations:
H'es = - RT2 a In cfJ / aT
(5-37)
where cfJisthe fugacity coefficient of the mixture and the derivative istaken
at constant pressure and total composition. Also
sres=Hres/T - R In cfJ
(5-38)
88 Propertes 01 Oils and Natural Gases
A summary of the equations used for computing the thermodynamic prop-
erties from the SRK equation of state is shown in the box:
Summary of Equations Used for Calculating Fugacity Coefficients, Enthalpies and
Entropies from the SRK-EOS (Edmister and Lee, 1984)
SRK-EOS: P = RT/(V - b) - a(T)/[V(V +b))
a;(T) =ao;",;(T)
",;(T) = [1 +m;(1 - T~5)F
m; = 0.480 +1.574w; - 0.176w; 2
ad = 0.42747 R
2
To; 2/Po;
b, = 0.08664 R T,; lP,;
a =L: L: (a;a)o'\z;z(1 - k;)
i j
b =L: z.b,
Fugacity coefficient at constant composition, </>:

In </> =Z - 1 - In (Z - B) - (A/B)(1 +B/Z)
Z =rvIRT; B/Z =blV; A/B =a/bRT
Fugacity coefficient of component i in a mixture, </>;:
In </>; = - In (Z - B) +(Z - l)b;lb
- (A/B) [~ (2a; ~ zaf5(l - k;)) - bJb] In (1+B/Z)
Enthalpy at constant composition, H:
H;d = " zH;d
'-' ' ,
(id means ideal gas)
T
Hi
d
= 1 c~dT
Tref
H"" =RT[Z -1 - (AlE) [1 -:; :;]ln (1 +B/Z)]
T da = - L: L: z;zm(a;a,T,)05(l - k;)
dT j
H =H;d +H""
Entropy at constant composition, S:
[
T Cid P ]
S;d=L: z;(S;;d- R In z;) =L: z, 1. ~ dT - R In - - R In z,
i i lref T Pref
S,~=In (Z - B) +(A/B) [:!: da] In (1 +B/Z)
a dT
S =S;d +S'"
E quations 01 State 89
PRASE DEN SITIES
The phase densities may be determined from the SRK-equation by using
Equation 5-7 or Equation 5-21. In general, good results are obtained for
vapor phase densities, whereas the liquid phase densities are somewhat un-
derestimated (Pedersen et al., 1984). Using the SRK-equation for density cal-
culations in cases where the liquid phase densities are of importance isthere-
fore not recommended.
It is generally advantageous to use the same density correlation for both
gas and liquid phases (e.g., to avoid discontinuities in the near critical re-
gion). Peneloux et al. (1982) have presented a consistent method for correc-
tion of SRK-volumes using an equation of state of the following form:
] \
1,
!
P =RT/(Y - b) -:-a/[(Y +c)(Y +b +2c)] (5-.39)
where e is a constant; which for a mixture is calculated as follows:
e =LCZ
iI
(5-40)
e, isthe value of efor component i. The SRK-equation (Equation 5-7), and
Equation 5-.39give identical results for saturation points and K-factors, but
different results for gas and liquid phase volumes. The SRK-volume, V, and
the Peneloux-volume, Y, are related as follows:
11
:
1 '
I'
.) ,
:i
"
1
I
'i_,
l'
,
Y=V-c (5-41)
The parameter e can therefore be regarded as a volume translation pa-
rameter. Martin (1979) has a discussion on the volume translation concepto
For non-hydrocarbons and for hydrocarbons < C7 the following expres-
sion is used for c:
e, =0.40768 (RTcJPci) (0.29441 - (ZRA)) (5-42)
where (ZRA) isthe Racket compressibility factor (Spencer and Danner, 197.3)
of component i for which the following approximation is used:
(ZRA) =0.29056 - 0.08775 W (5-4.3)
Peneloux et al. (1982) suggest finding the c-value of paraffinic,
naphthenic, and aromatic C7 +-components, respectively, from fifth-degree
polynomials in the carbon number. Pedersen et al. (1984) have found that
this procedure works reasonably well for gas and gas condensate mixtures,
9 0 Properties oj Oils and Natural Gases
while inaccurate results are obtained for heavy oil mixtures. The c-parame-
ter of C7+-components (or fractions) may instead be found as the difference
in the results for the molar volume at atmospheric pressure and 15C, calcu-
lated with the SRK-equation and determined experimentally (Pedersen et
al., 1988).
Far from the critical point it is of lessimportance to get consistent results
for gas and liquid phase densities. Liquid density correlations developed
specifically for petroleum mixtures are then often superior to the Peneloux
procedure. Examples of such procedures follow.
The Alani-Kennedy Equation
The form of the Alani-Kennedy equation (1960) is:
y3 - (RT/P +b)y2 +aY/P - ab/P =O (5-44)
For pure substances a =K exp (n/T) (5-45)
(5-46) and b =mT +C
where K, n, m, and C are constants. Values for the lighter hydrocarbons are
given in Table 5-3. It issuggested using the methane values also for N2, CO2,
and H2S.
Table 5-3
Constants ot the Alani-Kennedy Equation tor Pure Hydrocarbons
Component K n m x 10' C
C, (70-300F) 9160.6413 61.893223 3.3162472 0.50874303
C, (301-460F) 147.47333 3247.4533 - 14.072637 1.8326695
C2 (100-249F) 46709.573 - 404.48844 5.1520981 0.52239654
C2 (250-460F) 17495.343 34.163551 2.8201736 0.62309877
C3 20247.757 190.24420 2.1586448 0.90832519
ic, 32204.420 131.63171 3.3862284 1.1013834
n-C. 33016.212 146.15445 2.9021257 1.1168144
n-Cs 37046.234 299.62630 2.1954785 1.4364289
n-C6 52093.006 254.56097 3.6961858 1.5929406
Units: P in psa: T in R; V in ft3/lb-mol; R =10.7335 Ib-ft3/in.2R Ib-mol.
Source: Alani and Kennedy (1960).
'l\'
~"I
E quations oi State 9 1
The values of a and b of a C7+-fraction are found from the following ex-
pressions:
In aC7+=3.8405985 X 1O-
3
MW - 9.5638281 X 1O-4MWlp
+2.6180818 x 10
2
/T +7.3104464 x 1O-6MW2
+10.753517
(5-47)
bC7+=3.4992740 X 1O-
2
MW - 7.2725403 p +2.2323950
X 1O-
4
T - 1.6322572 x 1O-
2
MWlp +6.2256545
(5-48)
where MW =molecular weight of the total C
7
+-fraction
p =density (in g/cm'') of the total C7+-fraction at 1 atm and
15C
aand b of amixture are found as molar averages. The Alani-Kennedy equa-
tion gives accurate results for liquid densities, but it is not applicable to va-
por phase densities.
The Standing-Katz Procedure
This was originally developed as a graphical procedure (Standing and
Katz, 1941). Analytical expressions (Pedersen et al., 1984) approximating
the graphical form petmit the correlations to be applied in computer calcu-
lations also.
The procedure involves the following steps (the densities are expressed in
g/cm'):
l. (H2S +C3 +) Density. The density af standard conditions of the
(H2S +C3 +) fraction is calculated as follows
p(H2S +C3+) =1:MWx/ 1:MW x/p
i i
(5-49)
The index i refers to H2S, C3, and heavier components. The pure com-
ponent densities can betaken from Table 5-4, whereas ameasured C
7
+
density must be available.
2. Correction for C2. The weight fraction W2 of C
2
of the (H
2
S +C
2
+)
fraction is determined. The density at standard conditions of the
(H2S +C2+) fraction is then calculated from the equation:
p(H2S +C2+) =p(H2S +C3+) - Aa- Alal - A
2
a2 (5-50)
where Aa =0.3158 W2
(5-51)
9 2 Properties 01 Oils and Natural Gases
Table 5-4
Standing-Katz Pure Component Densities at 1 atm and 15C
Density
Component (g/cm
3
)
H2S 0.7970
C3 0.5072
ic, 0.5625
n-C. 0.5836
i-C5 0.6241
n-C5 0.6305
n-C6 0.6850
Sauree: Standing and Katz (1941).
A = - 0.2583 W2 (5-52)
(5-53)
(5-54)
A2=0.01457 W2
a =3.3 - 5.0p (H2S +C3+)
a2 =1+15(p(H2S +C3+)
- 0.46) (2.5p(H2S +C3+) - 2.15) (5-55)
3. Correction for CO2. Thedensity of the (C02 +H2S +C2+) fraction at
standard conditions is calculated on additive volume basis using
p(H2S +C2+) and a CO2 density of 0.8215 g/cm
3
.
4. Correction for C and N2 The weight fraction Wof (C +N2)of the
total mixture iscalculated. The density Po of the total mixture at stan-
dard conditions is then calculated from the equation:
Po =p(C02 +H2S +C2+) - Bo- Bb (5-56)
where Bo= 0.088255 - 0.95509 b2 +0.007403 b3
- 0.00603 b4 (5-57)
B =0.142079 - 0.150175 b2 +0.006679 b
3
+0.001163 b, (5-58)
(5-59)
(5-60)
b =p(C02 +H2S +C2+) - 0.65
b2 =1 - 10 W
E quations 01 State 9 3
b
3
=1 +30 w(5 W - 1) (5-61)
(5-62) b, =1 - 60 W +750 w} - 2,500 w3
5. Pressure Correction. The density Pp at the actual pressure P (in psia)
and 60C is found from the equation:
Pv =Po - Co - Cjc - C2C2- C3C3 (5-63)
where Co = - 0.034674 +0.026806 C4+0.003705 C5
+0.000465 Cs (5-64)
C = - 0.022712 +0.015148 C4+0.004263 Cs
+0.000218 Cs (5-65)
C
2
= - 0.007692 +0.003521 C4+0.002482 e,
. I
+0.000397 Cs (5-66)
I
C
3
= - 0.001261 - 0.000294 C4+0.000941 Cs
(
+0.000313 C6 (5-67)
\
Cj = 1 - 2(P - 500)/10,000 (5-68)
(
C2= 1 +6(P - 500)[(P - 500)/10,000 - 1]/10,000 (5-69)
C3= 1 - 12(P - 500)/10,000 +30[(P - 500)/10,000]2
I
- 20[(P - 500)/10,000]3 (5-70)
C4= 3.4 - 5 Po (5-71)
Cs= 1 +15(po - 0.48)(2.5 Po - 2.2) (5-72)
Cs= 1 - 30(po - 0.48) +187.5(po - 0.48)2
- 312.5(po - 0.48)3 (5-73)
6. Temperature Correction. Finally, the density p of the total mixture at
the actual temperature T (in F) and pressure can be calculated
p =Pp - Eo - Eje - E2e2- E3e3 (5-74)
where Eo =0.055846 - 0.060601 e, +0.005275 es
- 0.000750 es (5-75)
94 Properties 01otts and~Natural Gases
E = 0.037809 - 0.049262 e4+0.012043 es
+0.000455 e6 (5-76)
E2 = 0.021769 - 0.032396 e4+0.011015 es
+0.000247 e6 (5-77)
E3 = 0.009675 - 0.015500 e4+0.006520 es
- 0.000653 e6 (5-78)
e = 1 - 2[(T - 520)/200] (5-79)
e2= 1 +6[(T - 520)/200][(T - 520)/200 - 1] (5-80)
e3 = 1 - 12[(T - 520)/200] +30[(T - 520)/200]2
- 20[(T - 520)/200P (5-81)
,
e, = 3.6 - 5 pp (5-82)
es = 1 +15(pp - 0.52)(2.5 pp - 2.3) (5-83)
e6 = 1 - 30(pp - 0.52) +187.5(pp - 0.52)2
- 312.5(pp - 0.52)3 (5-84)
The Standing-Katz procedure gives accurate density results for liquid
phases with a large concentration of heavy components. The procedure is
not applicable to vapor phases.
The API Method
Calculation of the liquid density using the API method (1982) involves the
following steps:
1. Calculation of the pseudocritical temperature (T~) and pressure (P~)
for the mixture in question. These values are obtained as molar aver-
ages of the pure component critical temperatures and pressures.
2. Calculation of the pseudoreduced temperature (T; = T/T~) and the
pseudoreduced pressure (P; =P/P~).
3. Calculation of the average density for the mixture at 60F and 1atm
from the following formula:
n n \
p =E xMW/ E(xMW/p? )
i ~1 i~ 1
(5-85)
Table 5-5
API Pure Component Densitles at 1 atm and 15C
Density
Component (g/cm
3
)
N2 0.804
CO2 0.809
H2S 0.834
C, 0.300
C2 0.356
C3 0.508
i-C4 0.563
n-C4 0.584
te, 0.625
n-C, 0.631
C6 0.664
Source: API (1982).
p-values for non-hydrocarbons and for C-Ce are given in Table 5-5.
The measured densities at standard conditions are used for the C7 +-
components.
4. Calculation of the density correlation factor, C, at standard conditions
and at the actual conditions. C is given by the following expression:
C =A +A2x T; +A3x T;2 +A4x T;3 (5-86)
Each of the coefficients A-~ in Equation 5-86 is determined by an
expression of the following form:
A=B +B2 X P; +B3X P; 2 +B4X P; 3+Bs x P; 4 (5-87)
where the Bj-values are given in Table 5-6.
5. Calculation of the unknown density from the equatin:
p =p(C/C2) (5-88)
where the C
I
and C2 are the density correlation factors (Equation
5-86) at standard conditions, and at the actual conditions, respec-
tively.
9 6 Properties 01 Oils and Natural Gases
Table 5-6
Coefficients of Equation 5-87 for Calculation of Liquid Denslties Using the
API-Method
B, B2 B3 B. B5
A, 1.6368 - 0.04615 2.1138(10-
3
) - 0.7845(10-
5
) - 0.6923(10-
6
)
A2 - 1.9693 0.21874 - 8.0028(10-
3
) - 8.2328(10-
5
) 5.2604(10-
6
)
A3
2.4638 - 0.36461 12.8763(10-
3
) 14.8059(10-
5
) - 8.6895(10-
6
)
A. - 1.5841 0.25136 - 11.3805(10-
3
) 9.5672(10-
5
) 2.1812(10-
6
)
Source:API (1982).
The API-density method is only applicable to liquid phases. Results for
vapor and liquid phase densities using each of the correlations mentioned
previously are given in Chapter 9.
REFERENCES
Abbott, M. M., "Cubic Equations of State: An Interpretive Beview," Chap-
ter 3 in: Chao, K. C. and Robinson, R. L., E quations oj State in E ngi-
neering and Research, Advances in Chemistry Series 182, Am. Chem.
Soc., Washington, D.C., 1979.
Adachi, Y., Lu, B. C.-Y., and Sugie, H., "A Four-Parameter Equation of
State," Fluid Phase E quilibria, 11, 1983, pp. 29-48.
Alani, H. G. and Kennedy, H. T., "Volumes of Liquid Hydrocarbons at
High Temperatures and Pressures," ]. Peto Tech., Nov. 1960, pp. 272-273.
American Petroleum Institute, Technical Data Book-Petroleum Reiining,
API, New York, 1982.
Benedict, W, Webb, G. B., and Rubin, L. C., "An Empirical Equation for
Thermodynamic Properties of Light Hydrocarbons and their Mixtures. I.
Methane, Ethane, Propane, and Butane," ]. Chem. Phys., 8, 1940, pp.
334-345.
Carnahan, N. F. and Starling, K. E., "Interrnolecular Repulsions and the
Equation of State for Fluids," AIChE i., 18, 1972, pp. 1184-1189.
Edmister, W. C. and Lee, B. I., Applied Hydrocarbon Thermodynamics,
Volume I, Houston: Gulf Publishing Company, 1984.
Ishikawa, T., Chung, W K., and Lu, B. C.-Y., "A Cubic Perturbed, Hard
Sphere Equation of State for Thermodynamic Properties and Vapor-Liq-
uid Equilibrium Calculations," AlChE t., 26, 1980, pp. 372-378.
E quations 01 State 9 7
J ensen, B. H., "Densites, Viscosities and Phase Equilibria in Enhanced Oil
Recovery," Doctoral Dissertation, Technical Univ. of Denmark, Lyngby,
1987.
Kirn, H., Lin, H.-M. and Chao, K.-C., "Cubic Chain-of-Rotators Equation
of State," Ind. E ng. Chem. Fundam., 25, 1986, pp. 75-84.
Lee, B. I. and Kesler, M. G., "A Generalized Thermodynamic Correlation
Based on Three-Pararneter Corresponding States," AIChE [., 21, 1975,
pp. 510-527.
Leland, T. W., Phase E quilibria and Fluid Properties in the Chemical In-
dustry, Frankfurt/Main: DECHEMA, 1980, pp. 283-333.
Martin, J . J ., "Cubic Equations of State-Which?" Ind. E ng. Chem. Fun-
dam., 18, 1979, pp. 81-97.
Michelsen, M. L. and Heidemann, R. A. "Calculaton of Tricritical Points,"
Fluid Phase E quilibria, 39, 1988, pp. 53-74.
Mollerup, J . and Rowlinson, J . S., "The Prediction of the Densities of Lique-
fied Natural G~s and Lower Molecular Weight Hydrocarbons," Chem.
E ng. Sci., 29, 1974, pp. 1373-1381.
Pedersen, K. S., Thomassen, P., and Fredenslund, Aa., "Therrnodynamcs of
Petroleum Mixtures Containing Hydrocarbons. 2. Flash and PVT Calcu-
lations with the SRK Equation of State," Ind. E ng. Chem. Process Des.
Dev., 23, 1984, pp. 566-573.
Pedersen, K. S., Thomassen, P, and Fredenslund, Aa., "Characterzaton of
Gas Condensate Mixtures," paper presented at the 1988 AIChE Spring
National Meeting, New Orleans, LA, March 6-10, 1988.
Peneloux, A., Rauzy, E. and Frze, R., "A Consistent Correction for Red-
lich-Kwong-Soave Volumes," Fluid Phase E quilihria, 8, 1982, pp. 7-23.
Peng, D.-Y. and Robinson, D. B., "A New Two-Constant Equation of
State," Ind. E ng. Chem. Fundam., 15, pp. 59-64.
Pitzer, K. S., "Volumetric and Thermodynamic Properties of Fluids. 1. The-
oretical Basis and Virial Coefficents," ]. Am. Chem. Soc., 77, 1955, pp.
3427-3433.
Redlich, O. and Kwong, J . N. S., "The Thermodynamics of Solutions. V. An
Equation of State. Fugacities of Gaseous Solutons," Chem. Rev., 44,
1949, pp. 233-244.
Reid, R. C., Prausnitz, J . M., and Sherwood, T. K., The Properiies o] Gases
and Liquids, 3rd ed., New York, McGraw-Hill, 1977.
Soave, G., "Equilbrium Constants from a Modified Redlich-Kwong Equa-
tion of State," Chem. E ng. Sci., 27, 1972, pp. 1197-1203.
Spencer, C. F. and Danner, R. P., "Prediction of Bubble-Point Density of
Mixtures," ]. Chem. E ng. Data, 18, 1973, pp. 230-234.
Standing, M. B. and Katz, D. L., "Density of Crude Oils Saturated with
Natural Gas," Am. Inst. Min. Metal/. E ng., Meeting in Los Angeles, Oct.
1941, Tech. Pub. No. 1397, pp. 159-165.
(
(
(
(
(
(
(
(
(
(
Starling, K., Fluid Thermadynamic Properties far Lght Petroleum Systems,
Houston, Gulf Publishing Company, 1973.
Thomassen, P., Pedersen, K. S., and Fredenslund, A., "Adjustment of the
C7 +-Molecular Weight in the Characterization of Petroleum Mixtures
Containing Heavy Hydrocarbons," SEP Report 8617, Instituttet for
Kemiteknik, The Technical University of Denmark, 1986.
Tsonopoulos, C. and Heidman, J . L., "Hgh-Pressure Vapor-Liquid Equlb-
ria with Cubic Equations of State," Fluid Phase Equilibria, 29, 1986, pp.
391-414.
Tzouvaras, N., "Equatons of State for Vapor-Liquid Equilibrium Data Re-
duction," Ph.D. Thesis, Rensselaer Polytechnic Institute, Troy, NY, 1986.
van der Waals, J . D., Doctoral Dissertation, Leiden, Holland, 1873.
Chapter 6
Flash Calculations
The flash process isillustrated in Figure 6-1. The feed stream has N com-
ponents, i, with the composition z., i =1,2, ... , N (mole fraction). The feed
stream is initially in the single-phase region. After expansion through the
valve into the flash chamber, which here operates at fixed temperature, T,
and pressure, P, two or more phases are formed. In our illustration, avapor
phase of composition y, and a liquid phase of composition Xi, i =1, 2, ... ,
N, are formed. The fraction, on molar basis, which enters into the vapor
phase, iscalled 3. The vapor phase is assumed to be in equilibrium with the
liquid phase, i.e., the flash chamber contains exactly one equilibrium stage.
- .
ed T, P
ol e)
~-
( f i xed)
r
i
i = 1, 2N
-
Fe
(1 m
Vapor a mol es
Y
i
z
Li qui d (1-a) mol es
xi
Figure 6-1. Schematic diagram of a ~LE flash chamber.
A calculation of the equilibrium conditions of a given mixture at specified
pressure and temperature iscalled a (P,T)-flash calculation. It is used to de-
termine:
1. The number of phases
2. Molar amounts of each phase
3. Molar phase compositions
4. Phase densities
5. Molar enthalpy, entropy, and heat capacity of each phase
A flash problem may alternatively bestated as a(P,H)-flash, or as a (P, S)-
flash. In the former case, in addition to the molar composition, the pressure
and the molar enthalpy are specified; in the latter case the pressure and the
99
molar entropy are specified. When solving the flash problem in those cases,
the temperature iscalculated in addition to the properties mentioned above.
(P,T), (P,H), and (P,S)-flashes each have a unique solution. This is in con-
trast to (T,H), (T,S), (P, (3), and (T, (3)-flashes, which may have two, one, or
no solutions. (T,H) and (T,S)-flash calculations will not be covered here, as
they are seldom used in practice. Calculation of the pressures and tempera-
tures corresponding to a fixed molar vapor fraction are treated as a phase
envelope problem (see Phase Envelope Calculations), and not as a flash
problem.
The equilibrium conditions are those where the Gibbs energy (G) is at a
minimum. The Gibbs energy is defined as follows:
N
G =E n, .t (6-1)
i=1
where ni and .tare the number of moles and the chemical potential of com-
ponent i, respectively. Procedures for computing G and .tifrom equations of
state are given in the Two-Phase (P,T)-F1ash section.
The chemical potential can be regarded as the "escapng tendency" of
component i, and one way to escape is to form an acditional phase. Only
one phase exists if the total Gibbs energy increases for all possible trial com-
positions of an additional phase. Two or more phases exist if it ispossible to
split the mixture into two phases having a total Gibbs energy that is lower
than that of the single phase. The final number of phases and phase compo-
sitions are determined to be those with the lowest total Gibbs energy.
TWO-PHASE (P, T)-FLASH
The calctilation of whether a given mixture at specified (P,T) splits into
two or more phases is called a stability analysis (Michelsen, 1982a). The
starting point is the Gibbs energy (Go) the mixture would have if there was
only one phase present:
C, =G(n, n2, n3, ... , nN) (6-2)
where ni stands for the number of moles of type i present in the mixture, and
N is the number of different components. The total number of moles
N
(=~ ni) is called n. Secondly, the situation is considered where the mix-
ture splits into two phases (1and 11)with the compositions (n - e, n2 - e2,
n3 - e3, ... , nN - eN)and (e, e2, e3, ... , eN),where ei issmall. The Gibbs en-
Flash Calculations 101
ergy of phase 1may be approximated by aTaylor series expansion truncated
after the first order term:
G
I
=G _ ~ . (aGi)
o 1.J el -
i= 1 ani n
(6-3)
The Gibbs energy of the second phase is found to be:
Gn =G(e, e2, e3, ... , eN)
(6-4)
The change in the Gibbs energy due to the phase split is hence:
N N
~G =G
I
+G
n
- Vo =Ee ((.ti)n- (.t)o)=e E y ((.t)n- (.t)o) (6-5)
i= 1 i= 1
N
where e=E e, y is the mole fraction of component iin phase 11 and the
i=l
subindices o and 11refer to the single phase and to phase 11, respectively.
Only one phase exists if ~G is greater than or equal to zero for all possible
trial compositions of phase 11. The chemical potential, .t,may be expressed
in terms of the fugacity, f, as follows:
: (
(
i I
i (
I i
.t=.tp+RT In f =.tP+RT (In Z+In +In P)
(6-6)
\
~
where .t0is astandard state chemical potential, a fugacity coefficient, z a
mole fraction, P the pressure, and the subindex istands for component i.
The standard state is in this case the pure component iat the temperature
and pressure of the system. Equation 6-5 may then be rewritten:
~G N
RT
=E y (In y +In ()n- In Z- In ()o)
e i= 1
(6-7)
where z isthe mole fraction of component i in the total mixture. The stabil-
ity criterio n can now be expressed in terms of mole fractions and fugacity
coefficients. Only one phase exists if:
N
E y (ln y +ln()n- In Z- In()o) ~ O
i= 1
(6-8)
for all trial compositions of phase 11.
A minimum in the expression on the left hand side of Equation 6-8will at
the same time be astationary point. It can be shown that astationary point
must satisfy the equation:
t'
ti
1,
r
In y +ln(<l>)n- In Z- ln(<l>)o=k
where k isindependent of the component index. Introducing new variables,
Y, given by:
In Y =In y - k (6-10)
,
'1
,1
l
i
1
1
l'
Equation 6-9 may be rewritten:
In Y =In Z+In(<l>)o- ln(<l>)n (6-11)
Michelsen suggests using the following initial estimate for the ratio K, be-
tween the mole fraction of component i in the vapor phase, and inthe liquid
phase:
y K Pci (5 2 (1 Te))
~= =pexp.4 -T (6-12)
,
where Tci is the critical temperature, and Pci is the critical pressure of com-
ponent L Asinitial estimates for Y, use Kz, f phase o is aliquid, and z;lK,
if phase o is a vapor. The fugacity coefficients, (<l>)11' corresponding to the
initial estimates for Y, are determined. Based on these fugacity coefficients,
new Yj-values are determined, and so on. For a single phase mixture this
direct substitution calculation converges to either the trivial solution (i.e., to
two identical phases) or to Yj-values-Iulfillng the criterion:
N
E Y :5 1 (6-13)
=I
i
11
l
l
1:
!
which corresponds to a non-negative value of the constant k of Equation
6-9. A negative k-value would be an indication of the presence of two or
more phases. The molar composition of phase II obtained by solving Equa-
tion 6-11 isthen usually a good starting point in the calculation of the phase
compositions. For two phases in equilibrium, the following set of equations
must be satisfied:
Material Balance Equations:
1.
I
!i
{3y +(1 - (3)X=Z (i = 1, 2, 3, oo., N) (6-14)

(6-9)
Equilibrium Equatons:
y <l>r =X<l>r (i =1, 2, 3, oo., N) (6-15)
Summation of Mole Fractions:
N
E (y - Xi)=O
=I
(6-16)
In Equations 6-14 through 6-16 Xi,y, and Zare mole fractions in the liq-
uid phase, the vapor phase, and the total mixture, respectively. {3isthe mo-
lar vapor fraction. <l>v and <l>L are the fugacity coefficients of component i in
the vapor and liquid phases, computed as shown in Chapter 5. Equations
6-14 through 6-16 contain (2N +1) equations with (2N +3) variables,
namely (Xl, X2,X3, oo., XN),(y, Y2,Y3, oo., YN),{3,T, and P. With T and P
specified, the number of variables equals the number of equations, and
Equations 6-14 through 6-16 can be solved for Xi, y, and {3.These three
equations can be simplified by introducing the equilibrium ratio K =y;lx.
The following expressions may then be derived for Xand y:
x, =z;l(1 +(3(K- 1))
(6-17)
y =Kx,
(6-18)
and for K:
K =<l>f"/<l>r
(6-19)
The (2N +1) equations of Equations 6-14 through 6-16 may then be re-
duced to the following (N +1) equations:
In K =In <l>r - In <l>r (i =1, 2, 3, .", N)
(6-20)
N
E (y- Xi)=E Z(K - 1)/(1 +(3(K- 1)) =O
=1
(6-21)
For a given total composition, a given, (T, P), and K estinated from the
stability analysis, an estimate of {3may be derived from Equation 6-21. New
estimates of Xand y may be calculated from Equations 6-17 and 6-18, and
the K-factors are recalculated using Equation 6-20. A new {3-valueis calcu-
lated from Equation 6-21. This direct substitution calculation may be re-
peated until convergence. However, as described by Michelsen (1982b),
convergence may be accelerated by using, e.g., the general dominant eigen-
value method followed by a Newton-Raphson iteration.
MULTI-PHASE (P,T)-FLASH
A stability analysis as described in the previous section may be performed
on each of the two phases determined in a two-phase (P,T)-flash calcula-
tion. If the stability analysis shows that one or both of these phases are un-
stable, it is an indication that more than two phases existo A three-phase
(P,T)-flash calculation as sketched below may then be carried out. A stabl-
ity analysis on each of the three resulting phases may reveal that the system
isstill unstable and afour-phase (P,T)-flash calculation may be performed.
This calculation scheme can be continued until the stage where the stability
analysis shows that all the calculated phases are stable.
For a system consisting of J phases, the analogue of Equation 6-21 is:
N
Ez, (Kim- 1)/Hi =O (m=1, 2, ... , J - 1) (6-22)
i= 1
J -l
where n, =1+E {3m(Kim- 1)
me I
(6-23)
{3mis the molar fraction of phase m. In Equation 6-23 Kimis defined as the
ratio of the mole fraction of component i in phase mto its mole fraction in
phase J . Initial estimates for the K-factors are available from the stability
analysis for the preceding J -1 phase solution. Using a Newton-Raphsorr
iteration, the corresponding molar phase fractions, {3,may be found from
Equations 6-22 and 6-23. The phase compositions are subsequently found
from:
Yim=Zi Kim/Hi (m =1, 2, ... , J - 1) (6-24)
(6-25) YiJ =z/H
where Yim,and Yu, are the mole fractions of component i in phase m and
phase J , respectively. The corresponding Fugacity coefficients may be calcu-
lated (from Equation 5-29 if the SRK-equation of state is used) and a new
estimate of the K-factors isestablished using an analogue of Equation 6-20.
As is the case for the two-phase (P,T)-flash calculations previously de-
scribed, the general dominant eigenvalue method may be used for accelera-
tion of the direct substitution procedure which has just been outlined. How-
Flash Calculatums 105
ever, problems which cannot be converged in this manner occur relatively
frequently in multi-phase systems. According to Michelsen (1982b), New-
ton-Raphson iteration is not well suited for final convergence of difficult
problems. He instead recornrnends Gibbs energy minimization using the
method discussed by Murray (1972).
(P,H) OR (P,S)-FLASH CALCULATIONS
When pressure and enthalpy (P,H), or pressure and entropy (P,S), are used
as specification variables, the procedures outlined in prior sections are not
readily applicable since they require that the temperature, T, be known.
When performing (P,H) or (P,S)-flash calculations it isessential to provide a
good initial estmate-of the temperature. Figure 6-2 shows the enthalpy de-
pendence of temperature for a mixture of 50% methane and 50% ethane at
60 atm (from Michelsen, 1986a). The SRK equation as described in Chapter
5, isused for calculation of all thermodynamic properties, and the enthalpy
is chosen such that the enthalpy for an ideal gas equals zero at OC. Ideal
H/R (K)
o
-500
-1000
-1500 ' I I , , , ! !
210 220 230 250 260 270 240
T( K)
Figure 6-2. Molar enthalpy in dependence oftemperature lor an equimolar mix-
ture 01 methane and ethane at 60 atm. Source: Michelsen (1986a).
gas heat capacities are represented by third-degree temperature polynomials
as described in Chapter 5.
It is observed that the enthalpy increases monotonically with tempera-
ture, for this mixture. Hence, the temperature corresponding to any speci-
fied enthalpy iseasily located. The enthalpy-temperature variation isfor the
same mixture shown at 30 atm in Figure 6-3. In the temperature interval
223.7K <T <240.1K, two different results may be obtained for the single-
phase enthalpy of the mixture. (These results correspond to two different
real roots in the third-degree polynomial of Equation 5-21, for the compress-
ibility factor). The lower branch corresponds to aliquid state and the upper
branch to a vapor state. At temperatures below T =233.0 K, the liquid
state has the lower Gibbs energy (.e., it isthe more stable state), and above
T*, the vapor phase has the lower Gibbs energy (and is the more stable
state). At the transition temperature, T* ( =233.0 K), the Gibbs energies of
the two phases are equal. It is found that the molar enthalpies of the two
phases at T* are H/R = - 65 K, and HIR = - 455 K, respectively. If
the specified enthalpy is outside this interval, determining the correspond-
ing temperature is straightforward. For a specification in the interval
( - 1165 K to - 455 K) the transition temperature, T*, may be used as an
initial temperature estimate. In the latter situation the system is known to
H/R (K)
o i I I
I
I
11
l'
!!
!I
il
q
1:
1,
-500
form at least two phases at equilibrium, and a stability analysis (seeTwo-
Phase (P,T) Flash Section) is not required. As initial estimates for the phase
compositions at T* , two phases of identical compositions are used, but with
different molar volumes (corresponding to the smallest and the largest root
in the compressibility factor polynomial of Equation 5-21).
Michelsen (1986a) has described a general multi-phase flash procedure
for solving a (P,H) specification for a given mixture. The procedure is out-
lined as follows for the case with only two phases.
Two object functions are defined, namely:
N
g =LZ(K - 1)/l
=1
(6-26)
where L = 1 +(3(K - 1)
(6-27)
and ~ =H - Hspec
(6-28)
where H isthe molar enthalpy corresponding to the estimated phase cornpo-
sitions, and Hspecisthe specified molar enthalpy. At convergence both g and
~ are zero (g equals the left side of Equation 6-21). The molar enthalpy H
is calculated as follows:
H =(3HV +(1 - (3)HL (6-29)
where HV and HL are the molar vapor and liquid phase enthalpies, respec-
tively. These properties are calculated as shown in the Summary of Equa-
tions section in Chapter 5. The iterations proceed as follows:
Calculate (3from Equation 6-2l.
Calculate In <pand a In <p/aT for each of the two phases (for the
SRK-equation of state, <pmay be derived from Equation 5-29).
Calculate K from Equation 6-20.
Calculate g from Equation 6-26.
l.
--~-
2.
""
~ -1000 L
I
I
-
3.
I
'L
I
I
I~
4.
-1500 I I I I I I I I I
210 220 230 I 240 250 260 270 280
I
T* T(K)
Figure 6-3. Molar enthalpy in dependence ot temperature for an equimolar mix-
ture of methane and ethane at 30 atm. Source: Michelsen (1986a).
N
5. Calculate ag/a(3 =- LZ(K - 1)2/lr
i= 1
anda~/a(3 =HV - HL
6. Calculate ag =V. Z (K a In K - (K - l)T*/I-.)/I-.
aT 1.J 1 1 aT 1 1 ~1 ~1
1=
(6-30)
(6-31)
where Tt =(3K a In K/aT (6-32)
7. Calculate g2from Equation 6-28.
8. Calculate ag2/OT ={3C~+(1- (3)q +
N
RT2 1:zH! (Ti" - (T*)2Iti)
ie l
(6-33)
(6-34) where T* * ={3K, (a In K/aT)2
and C,v and CpL are the molar vapor and liquid phase heat capacities
( =the temperature derivatives of the molar phase enthalpies).
9. Use a Newton-Raphson iteration procedure to determine the size of
the next {3and temperature steps (fl{3 and flT).
10. Calculate new estimates for {3and for the K-factors:
{3new={3+fl{3 (6-35)
1
_ a In K,
n Kinew- In K, +-- fl T
aT
(6-36)
11. CaIculate the new phase compositions from Equations 6-17 and 6-18.
12. Return to step 1.
A (P,S)-flash calculation may be performed using aprocedure that isquite
analogous to what has just been outlined for a (P,H)-flash calculation.
In most cases the final equilibrium temperature T deviates less than 30K
from the initial temperature estimate, T*. In the case of difficult flash prob-
lems, (e.g., near critical fluids) it can be advantageous to use adifferent cal-
culation procedure. (P,H) and (P,S)-flash problems can both be expressed as
minimization problems. A (P,H)-flash may be solved by locating the mini-
mum of ( - S) subject to H =Hspec'Similarly a (P,S)-flash calculation may
be performed by locating the minimum of H subject to S =Sspec'
The (P,H) and (P,S)-flash algorithms outlined require approximately the
same computer time as the equivalent (P,T) algorithm.
THREE-PHASE (P,T)-FLASH WITH A LIQUID WATER PHASE
The need for performing three-phase flash calculations on petroleum
mixtures often arises when water is present, and where an almost pure liq-
uid water phase potentially may be formed. If the solubilities of other com-
ponents in the aqueous liquid phase are neglected, (P,T)-flash calculations
can be very much simplified (Michelsen, 1981) as compared with what was
described in the Multi-Phase (P,T)-Flash section.
~\r:'
Initial testing establishes whether apure liquid water phase is separated
from the feed. This isdone by comparing the chemical potential (seeEqua-
tion 6-6) of water in a pure water phase:
(J Lw)pure =JLe.,+RT(ln P +ln(<pw)pure) (6-37)
with the fugacity of water in the total mixture:
(JLw)mix =J L~v +RT(ln P +ln(<pw)mix+In zw) (6-38)
Zw is the mole fraction of water in the total mixture.
If (JLw) pure < (JLw) mix (6-39)
or In (<Pw)mix +eln z., >In (<pw)pure (6-40)
apure liquid water phase will precipitate. The amount of water remaining
in the mixed phase is found as the amount fulfilling the criterion:
ln(<pw)mix+In x., - ln(<pw)pure=O (6-41)
where x., isthe mole fraction of water in the mixed phase, after removal of
the pure water phase. A stability analysis, as described under the Two-Phase
(P,T)-Flash heading, may then be carried out for the remaining mixed
phase. If it splits into two phases, the amount of the pure water phase must
be adjusted, considering that Equation 6-41 must be fulfilled for each of the
mixed phases.
Initially comparing the chemical potentials of water in the pure liquid
form with those of water in the mixed phase may alternatively reveal that
water does not condense from the feed. In that case, aconventional stability
analysis is performed for the total feed. If it is found that more than one
phase exists, atwo-phase (P,T)-flash calculation isperformed as described in
the section so named. Each iteration step is checked to see whether water
will precipitate from the mixed phases, i.e., whether Equation 6-39 is ful-
filled. The final result of the calculation may well be that three phases are
present, even though the initial test for water precipitation from the feed
was negative.
SIMPLIFIED TWO-PHASE (P,T)-FLASH
When using the SRK-equation of state, the binary interaction coefficients
of Equation 5-20 are usually assumed equal to zero for hydrocarbon-hydro-
carbon interactions, while non-zero values are used for interactions between
two non-hydrocarbons, and between a hydrocarbon and a non-hydrocar-
bono For mixtures containing up to 2% non-hydrocarbons, binary interac-
tion coefficients equal to zero are usually a good approximation. For such
mixtures, the two-phase (P,T)-flash calculation can be greatly. simplified
(Michelsen, 1986b).
When, with the SRK equation, all binary interaction coefficients equal
zero, Equation 5-20 can be reduced to:
a =(a')2 (6-42)
N
where a' =1: z, a'
i= 1
with a'= a?5 (6-43)
and the expression for the fugacity coefficient given in Equation 5-29 can be
rewritten:
In <p=qo +q"a' +Clab (6-44)
where qo= -In (:T (V - b))
(6-45)
(
(
(
q" =_ 2 ~ In ((V +b))
bRT V
(6-46)
l(PV ) a"
q~=b RT - 1 +2b q"
(6-47)
The following relation exists between the mole fractions of a given corn-
ponent in two phases in equilibrium, and in the total mixture:
z, =(3y + (1 - (3)x (6-48)
Multiplication by a' and summation over all components yields the identity:
N N N
1: z, a'= (31: y a'+ (1- (3)1: x, a'
i= 1 i= 1 i= 1
(6-49)
or a{: =(3a{r + (1 - (3)a{ (6-50)
Flash Calculatons 111
where F refers to the total mixture, or feed. Similarly, for the b-pararneter:
bF =(3by + (l - (3)b
L (6-51)
al, and bL may be expressed in terms of these three parameters:
aL =a
r
- (3 ay
1 - (3
(6-52)
b
L
=bF - (3by
1-(3
(6-53)
Irrespective of the number of components, the flash problem has now
been reduced to involve only three parameters, namely ay', by, and (3. The
calculation procedure is as follows:
l. Calculate aF' and bF
2. Estimate ay', by, and (3
3. Calculate aL' and bL from Equations 6-52 and 6-53
4. Calculate the gas and liquid phase fugacities for each component
5. Calculate the K-factors ( =Y/Xi)
6. Calculate the phase compositions using Equation 6-17 and 6-18
7. Evaluate the check-Iunctions:
N
el =1:(y- Xi)
=1
(6-54)
N
'\"' ' r
e2=1..J y a - ay
i=l
(6-55)
N
e3=1:y b, - by
i=1
(6-56)
If el, e2, and e3are all equal to zero, the phase distributions at equilb-
rium have been determined. If not equal to zero, a Newton-Raphson itera-
tive correction of ay, by, and (3isperformed, and the calculation returns to
step 3. '
The calculation time isalmost independent of the number of components.
Thus, for multi-component mixtures, much computation time can be saved
as compared with the usual two-phase (P,T)-flash previously described. Pe-
dersen et al. (1985) have reported relative computation times for (P,T)-flash
calculations on petroleum mixtures using the usual (P,T)-flash, and the
k=O flash described in this section. The results are presented in Table
6-1.
Table 6-1
Relative Computer Time tor Flash Calculations on Petroleum Mixtures
3 Hydro- 6 Hydro- 10 Hydro- 20 Hydro- 40 Hydro-
Program carbons carbons carbons carbons carbons
k;J *0 0.08 0.16 0.25 0.46 1.00
k;J =O 0.05 0.06 0.08 0.09 0.15
k;*O: Flash program with some 01the binary interaction coefficients different lrom zero.
k,=O: Flash program with all binary interaction coefficients equal to zero.
Source: Pedersen el al. (1985).
PHASE ENVELOPE CALCULATIONS
The (P,T)-values which for a given mixture correspond to a fixed vapor
mole fraction, (3, are collectively called a phase envelope. Michelsen (1980)
has described a procedure for phase envelope calculations, and it is briefly
outlined in the following paragraph.
400.0 i i
~ 300.0
"'
B = o
~
~
!
200.0
B = 0.2
-
-
- -
/
/'
-
/
6 = 0.4
-
- /'
/
-
-
-:
/
- .fo.6
,/
16 = I
/ /
./
/'
/
./'
./'6 = 0.8
/
/
-
- -
100.0
0.0 ' , , , I , I ,
-100.0 0.0 300.0 500.0 600.0
Temperaturo ('e),
100.0 200.0 400.0
Figure 6-4. Phase envelope of a North Sea oil mixture calculated using the SRK
equation. {:J is the mole fraction of the vapor phase.
To initialize a phase envelope calculation, it is convenient to specify a
fairly low pressure, e.g., P < 20 atm. Equation 6-12 may be used to find
initial estimates of the K-factors, and using this simple K-factor model,
Equation 6-21 may be solved for the temperature T. Equations 6-20 and
6-21 are then solved simultaneously for T and In K, where now the actual
K-factor model, (e.g., the SRK equation) is used. In addition, the deriva-
tives of T and In K, with respect to P, can becalculated and used to generate
an estimate of the second point on the phase envelope at a slightly higher
pressure than the first one. From the third point, and on, the extrapolation is
based on the two last calculated points, and the corresponding derivatives.
Using this procedure the complete phase envelope can be constructed in a
stepwise manner, and even passing the critical point creates no problems.
Figure 6-4 shows the phase envelope calculated for a typical North Sea oil.
REFERENCES
Michelsen, M. L., "Calculaton of Phase Envelopes and Critical Points for
Multicomponent Mixtures," Fluid Phase Equilibria, 4, 1980, pp. 1-10.
Michelsen, M. L., "Three-Phase Envelope and Three-Phase Flash Algo-
rithms with a Liquid Water Phase," SEP Report 8123, Instituttet for
Kemiteknik, The Technical University of Denmark, 1981.
Michelsen, M. L., "The Isothermal Flash Problem. Part 1. Stability," Fluid
Phase Equlibria, 9, 1982a, pp. 1-19.
Michelsen, M. L., "The Isothermal Flash Problem. Part II. Phase-Split Cal-
culation," Fluid Phase Equilibria, 9, 1982b, pp. 21-40.
Michelsen, M. L., "Multphase Isenthalpic and Isentropic Flash Algo-
rithms," SEP Report 8616, Instituttet for Kemiteknik, The Technical Uni-
versity of Denmark, 1986a.
Michelsen, M. L., "Simplfed Flash Calculations for Cubic Equations of
State," Ind. Eng. Chm, Process. Des. Dev., 25, 1986b, pp. 184-188.
Murray, W., "Methods for Unconstrained Optmzation," Academic Press,
London, 1972.
Pedersen, K. S., Thornassen, E, and Fredenslund, Aa., "Therrnodynamics of
Petroleum Mixtures Containing Heavy Hydrocarbons. 3. Efficient Flash
Calculation Procedures Using the SRK-Equation of State," Ind. Eng.
Chem. Process. Des. Dev., 24, 1985, pp. 948-954 .
Chapter 7
Characterization Procedures
CLASSIFICATION OF THE CONSTITUENTS OF OIL AND GAS
MIXTURES
As described in Chapter 2, naturally occurring oil and gas mixtures will
often be too complex for acomplete identification of components. The com-
ponents of a reservoir fluid may instead be elassified as follows:
Defined components are those for which there exists a quantitative gas chro-
matographic analysis and for which Te, P and w are known. This in-
eludes non-hydrocarbons and hydrocarbons up to at least C6.
TBP fractions cover the true boiling point fractions from C7and up to max.
C30,each containing many different components. The density and the av-
erage molecular weight may be measured on each fraction.
The TBP resulue consists of components that are too heavy to be separated
using a TBP distillation procedure. The density and the average molecu-
lar weight may be measured for the total sample., but neither boiling
point nor molecular weight distributions are determined.
TBP FRACTIONS
Because pure component properties are not available, equation of state
(EOS) calculations on mixtures ineluding heavy hydrocarbon fractions re-
quire a characterization procedure for determination of the EOS-parame-
terso Each TBP fraction may be represented as one pseudo-component. The
critical temperature (Te), and pressure (Pe), and the acentric factor (w) to be
used in the EOS-calculation can be found from the boiling point, the molec-
ular weight, and the density using empirical correlations. Examples of such
correlations are given below:
1. Lee-Kesler Correlations (Kesler and Lee, 1976): .
Te =341.7 +811SG +(0.4244 +0.1l74SG)TB
+(0.4669 - 3.2623SG) X 105/TB
114
(7-1)
Characterization Procedures 115
In P, =8.3634 - 0.0566/Se
- (0.24244 +2.2898/SG +0.1l857/SG2) X 10-
3
T
B
+(1.4685 +3.648/SG +0.47227/SG2) x 10-
7
T~
- (0.42019 +1.6977/SG2) x 10-10 ~ (7-2)
w =(In PBr- 5.92714 +6.09648/TBr
+1.28862 In TBr - 0.169347T~r)/(15.2518 - 15.6875/TBr
- 13.4721 In TBr +0.43577Tt) (for TBr<0.8) (7-3)
w = - 7.904 +0.1352K - 0.007465K2 +8.359TBr
+(1.408 - 0.01063K)/T Br (for TBr>0.8) (7-4)
The boiling point, TB, and the critical temperature, Te, are in R, and
SG s the 60F/60F specific gravity (the density of the fraction at at-
mospheric pressure and at 60F, divided by the density of water at the
same conditions; i.e., se is approximately equal to the density of the
fraction in g/crn" at atmospheric pressure and 60F). The critical pres-
sure, P is in psia. TBr =TB/Te, and PBr=PB/Pe, where PBsthe pres-
sure where TBwas measured. K isthe Watson characterization factor
which equals T~3/se. Thomassen et al. (1986) have suggested that
Equation 7-3 be modified to:
w =(In PBr- 5.77881 +6.05615/TBr +1.37882 In T
Br
- 0.173914 T~r)/(15.5523 - 15.7915/TBr
- 12.7855 In TBr +0.43487 Tt) (for TBr<0.8) (7-5)
2. Winn-Relations (Winn, 1957; Sim and Daubert, 1980):
Te =exp [4.2009 ~.08615SG004614]/1.8 (7-6)
P, =6.1483 X 1012TB'2.3177SG24853
(7-7)
TB and Te are in K, SG in 15C/15C, and Pein Pa.
3. Cavett (1964) Relations: *
Te =768.071 +1.7134 TB - 0.10834 x 1O-2T~
+0.3889 X 1O-6T~- 0.89213 X 1O-2T
B
API
+0.53095 X 1O-6T~API +0.32712 X 1O-7T~APJ 2 (7-8)
The term 0.53095 X 10-
6
T~API in Equation 7-8 was in Cavett's work originally
0.53095 x 10-
5
T~API.
loglOPe=2.829 +0.9412 X 1O-3TB - 0.30475 X 1O-5T~
+0.15141 X 1O-8T~- 0.20876 X 1O-4T
B
API
+0.11048 X 1O-7T~API +0.1395 X 1O-9T~APJ 2
- 0.4827 X 1O-7T
B
APJ 2 (7-9)
In Equations 7-8 and 7-9, API equals 141.5/Se - 131.5, where se
isthe 60F/60F specific gravity. The critical temperature, Te, and the
boiling point, TB, of the fraction are given in F and the critical pres-
sure, Pe, in psia.
4. Edmister's (1958) Acentric Factor:
w =~[ log Pe ] _ 1
7 Te/TB-1
(7-10)
where P, is in atmospheres.
5. The correlations of Riazi and Daubert (1980a):
Te =24.2787 ~.58848seO.3596
(7-11)
r,=3.12281 X 109T
B
2.3125se2.3201
(7-12)
where Te and TB are in R, P, in psia, and se in 60F/60F.
THE TBP RESIDUE
Neither boiling point nor rnolecular weight distributions are available for
the TBP residue, i.e., Equations 7-1through 7-12 are not readily applicable.
Figure 7-1shows, for atypical North Seaoil whose analytical data are given
in Table 7-1, the mole fraction from C7to C29versus molecular weight. Asis
seen from the Figure, the composition of the C7 +-fraction is an almost con-
tinuous distribution in molecular weight. The boiling point fractions may be
represented using an appropriate analytical expression for the mole fraction
versus molecular weight (or carbon number). This analytical expression may
then be applied to estimate the molar composition of the TBP residue. Vari-
ous distribution functions are suggested in the literature. Yarborough (1978)
and Pedersen et al. (1984) have suggested for gas condensate mixtures to as-
sume a logarithmic distribution of the molar fraction Z versus the carbon
number CN:
In z =A +B X CN (7-13)
Characterization Procedures
117
Moa
-,
3.0+ I I ~\
\
\
:\
2,Or ""
l l HH
(
1.0
O.
10 15 20 25
Carbon Number
Figure 7-1. Molar disiribution of C7+-fractions in a North Sea oil (molar composi-
tion in Table 7-1).
Table 7-1
Molar Composition 01the North Sea Oi!
whose C7+-Composition is Plotted in Figure 7-1
Density
Molecular
(g/cm
3
) at
Component
Mole %
Weight
15C, 1 atm
N2 0.57
CO2 1.20
C,
45.85
C2 3.68
C3 1.23
-C4 0.63
n-C4 0.87
te,
0.65
n-e; 0.43
C6 0.87
86.0
0.644
Cl 3.28
93.0
0.733
Ca 3.55
107.0
0763
Cg 3.44.
120.0
0.785
C10 2.53
131.6
0.808
Cll 2.48
147.6
0.815
C'2 2.13
158.6
0.836
C'3 2.45
172.4
0.850
C'4 2.20
184.8
0.861
c.,
2.06 197.4
0.873
C'6 1.66
209.0
0.882
c.,
1.69
226.5
0.873
Table 7-1
Continued
Density
Molecular (g/cm
3
) at
Component Mole % Weight 15C, 1 atm
c., 1.45 242.6 0.875
c., 1.27 253.7 0.885
e20 1.21 262.0 0.903
e; 0.90 281.1 0.898
e22 0.95 293.2 0.898
e23
0.77 307.3 0.899
e24 0.75 319.6 0.900
e25 0.66 333.3 0.905
e26 0.57 345.9 0.907
e27 0.66 361.1 0.910
e28 0.62 373.5 0.915
e29 0.60 380.8 0.920
C30+ 6.14 624.0 0.953
A and B are constants determined from the measured mole fraction and the
measured molecular weight of the plus-fracton, Figure 7-2 shows, for a gas
condensate (molar composition in Table 4-6), aplot of ln Z of the TBP frac-
tions versus the carbon number. Also sketched in Figure 7-2 is the distribu-
tion estimated for the TBP residue. To get agreement with the experimental
mole fraction and the average molecular weight of the TBP residue, the con-
stants A and B of Equation 7-13 may for the TBP residue deviate slightly
from those determined by a least squares fit to the data. for the TBP frac-
tions. For oil mixtures, the composition of the TBP residue may alternatively
be estimated by extrapolating the TBP distillation curve to 100% distilled
off (Pedersen et al., 1984) by fitting a fifth-degree polynomial to the experi-
mental TBP data, plus the following artificially generated data:
TSOwt%=1O-02127T;ii~~%se-o 6495 (7-14)
TIOO,m =1OI.OS3 TO.7097seO.6717
\\t-,Cl 50wt%
(7-15)
T50wt%,TSOwt%,and TIOOwt%are the temperatures in K at which 50,80, and
100 wt % of the crude oil has been distilled off. se is the specific gravity of
the TBP sample that roughly corresponds to the C7+-fraction of the total
mixture. The cumulative weight % distilled off versus temperature isshown
in Figure 7-3 for the North Sea oil of Table 4-7, Also sketched is the ex-
trapolated TBP curve. Mole fractions below 10- 6are neglected, and the
heaviest component considered is Cso.
O. Or ' - - - - - - - - - - - ,
-2.5
-sJ

Measur ed mol e f r act i ons
-- -
Est i mat ed mol e f r act i ons
e

o
+-'
<,
u -7.5 <,
'"
e,
<,
4-
~
<,
o
<,
.s
10.0 <,
.s
<,
<,
<,
.12.51-
<,
<,
<,
-IS.J
I I I I
, ,
I
lO 20 30 40 50 60 70 80
Car bon Number
Figure 7-2. Measured and estimated (Pedersen et al., 1984) C7+-composition for
the gas condensate of Table 4-6.
800.
700J
I
A
/
.'
Measur ed TBP- dat a
/
600.11-
Ar t i f i ci al l y gener at ed TBP- dat a
/
. :
/
~'""l
- - -.
Ext r apol at ed TBP- cur ve
,,-
,/
/'
+-' 400.
/'
'"
s,
/'
QJ
O-
E
(!!. 300.
200.0~
100l

0.0
0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0
Wei ght %di st i l l ed of f
Figure 7-3. Measured TBP-distillation data for the oil mixture of Table 4-7, and
estimated extrapolation to 100% distilled off. (Pedersen et al., 1984).
Whitson (1983) suggested describing the C7vmolar composition as a
gamma-distribution. The molar fraction, Z,of component i, with amolecu-
lar weight of MW, is given by the expression:
Z=P(MW) - P(MW - 1) (7-16)
where
P(MW) =(MW - 1)),,-lexp( - (MW - 1))1(3)
(3"r(ex)
(7-17)
I' is the gamma function, MW isthe molecular weight, and ex, (3, and 1) are
parameters defining the distribution. These parameters are determined
from the avalable analytical information. In cases where the analytical in-
formation islimited, exis used as a regression variable, which is determined
from measured PVT-data.
CHARACTERIZATION PROCEDtJ RES
Pedersen et al. (1984) use Equations 7-8 and 7-9 to determine Teand Pe,
and Equations 7-4 and 7-5 to determine w. The equations are used sepa-
rately for the paraffinic (P), naphthenic (N), and aromatic (A) compounds
within aTBP fraction. Te, Pe, and w of the total fraction are found as molar
averages. The specific gravities of the P, N, and A components are found
from the specific gravity of the total fraction, and the following relations:
SGN =1.12 SGp (7-18)
(7-19) SGA =(1.21 +0.008 (CN - 6)) SGp
where CN is the carbon number. In those cases where no PNA analysis is
available, the PNA distribution of a given TBP fraction is estimated. The
refractive index, n, of each TBP fraction iscalculated from the specific grav-
ity, the normal boiling point, and the molecular weight, using the correla-
tions of Riazi and Daubert (1980b):
n= (~)O.5
1- 1
(7-20)
1isacharacterization factor, which isfound from one of the following corre-
lations:
1=3.583 X 1O-3T~OI47(MWlp)-04787
1=1.4 X 1O-3T~o9(MWlp)-O.3984
for MW<200 (7-21)
for MW>200 (7-22)
where MW is the molecular weight, TB is the boiling point in R, and p is
the liquid density ing/cm
3
at 20C and 1atm. Based on the refractive index,
the specific gravity, and the molecular weight, the PNA distribution (in mole
%) can be estimated as described by Nes and Westerns (1951):
v =2.51 (n - 1.4750) - p +0.8510 (7-23)
w =p - 0.8510 - 1.11 (n - 1.4750) (7-24)
%A =430 v +3,660/MW for v>O (7-25)
%A =670 v +3,660/MW for v-cO (7-26)

R =820 w +10,000/MW for wo-O (7-27)
R =1,440 w +10,600/MW for w-c
(7-28)
%N =R - %A (7-29)
%P = 100 - R (7-30)
Touse Equations 7-4,7-5,7-8, and 7-9 to determine Te, Pe, and w, T
B
and
SG of the pseudocomponents of the TBP residue must beestimated. Pedersen
et al. use a logarithmic distribution for the specific gravity versus carbon
number. Experimental and extrapolated specific gravities versus carbon
numbers are shown in Figure 7-4 for theoil mixture of Table 4-7. For the
boiling points, the values of Katz and Firoozabadi (1978) are used, up to
C45.When alogarithmic distribution according to Equation 7-13 isassumed
for the TBP residue, the following relation is used for the boiling points (in
K) of the pseudocomponents of the TBP residue:
T
B
=97.58 MWO3323SG004609
(7-31)
When, for oil mixtures, the characterization is based on a TBP distillation
curve, 6 K boiling-point intervals per carbon number are assumed for the
C45+-components. The following expression is used for the molecular
weights of the components of the TBP residue:
MW = - 4.0 +14.0 CN (7-32)
122 Propertes oj Ols and Natural Gases
1.1 I i
1.05
E
1.0 +-'
ro
-
1..'-
.95
~
+-'
'" .9
"" E
u
'<,
~ .85
c-,
+-'
v>
.8
e
Q!
el
. 75
.7
-
-
-
-
-
-
-
-
/
/'
,,/
./
./
.fIII'
. : Measur ed densi t i es
Est i mat ed densi t i es
t 10 20 30 40 50 60 70 80
Car bon Number
Figure 7-4. Measured and estimated (Pedersen et al., 1984) C7+-densities for
the oil mixture of Table 4-7.
Pedersen et al. use binary interaction coefficients equal to zero for interac-
tions between two hydrocarbons. For interactions between an unlike pair of
non-hydrocarbons, and between a non-hydrocarbon and hydrocarbons
ranging from el to e6, the interaction coefficients of Tahle 5-2 are used. For
interactions between a e7 +-component and a non-hydrocarbon, the kij"val-
ues given in Table 5-2 for e6 are used.
Whitson (1983) assumes that the Watson characterization factor:
K =T
B
1/3/se (7-33)
is aconstant for each e7+ fraction, and that it can be related to the molecu-
lar weight using the following formula:
K =4.5579 MW015178 se - 084573
(7-34)
TB is in R, and se in 60F/60F.
It is necessary to choose K such that the specific gravity of the total e7+-
fraction equals the measured value. The boiling points of the e7+-samples
may then be calculated using the definition of the Watson characterization
factor. Teand P, are found from Equations 7-11 and 7-12, and w from Equa-
~: : ~
~I\''''
123/!r'-;' :" f t ,,(.
! \"
tion 7-10. Whitson useshis characterization procedure with the Peng- Robn- , ' ' ..
son equation of state (seeEquation 5-8). Other examples of characterization J t\lXllJ f-j. ; o
procedures are given by Edmister (1988). .. I ;
Pedersen et al. and Whitson only use the distribution functions as appro-. .
priate procedures for selecting pseudo-components. Using continuous ther- \,;
modynamics (see e.g., Cotterman and Prausnitz, 1985), the distribution '-.
functions are maintained in the phase equilibrium calculations. The molar
composition is described by a molar distribution function in some variable,
1, often the molecular weight or the boiling point. The molar fractions of a
continuous mixture must add up to 1:
Characterzaton Procedures
J 1 F(I)dl =1
(7-35)
For a naturally occurring petroleum mixture, non-hydrocarbons, and the
hydrocarbons e1-e4, are handled as discrete components. If the total mole
fraction of the continuous components is r and there are n discrete compo-
nents, Equation 7-35 is changed to:
n
1:z, +r J F(l)dl =1
i=1 1
(7-36)
where z, isthe molar fraction of the -th discrete component. For two phases
in equilibrium, with amolar vapor fraction of tthe following material bal-
ances can be set up for the discrete components:
z, =~Yi+(1 - ~) x, (7-37)
where Yiand x, are the mole fractions of component i in the vapor and liquid
phases, respectively. Similarly, for the continuous components:
rFFF (1) = ~rv FV (1) +(1 - ~)rLFL (1) (7-38)
F refers to the feed, V to the vapor phase, and L to the liquid phase. Cotter-
man et al. (1985) usecontinuous thermodynamics with the SRK-equation of
state (Equation 5-7). From C, on, the components are treated as three con-
tinuous distributions in molecular weight: one for the paraffins, one for the
naphthenes, and one for the aromatics. The a and b parameters entering
into the SRK-equation are, for the continuous distributions, determined
from the molecular weight. Cotterrnan et al. have shown that when the
SRK-equation isused, the distribution functions for the two phases in equi-
librium can be related to one another. For example, if a gamma distribution
describes the compositon of one phase, the composition of the other phase is
also described by a gamma distribution. One major drawback of this
method isthat the total mixture is not described by a gamma function, sim-
ply because the sumof two gamma functions isgenerally not agamma func-
tion. Therefore, the material balance cannot be satisfied, and the flash
problem can only be solved approximately. Cotterman et al. suggest an ap-
proximate solution, where the total mixture is also described by a gamma
distribution. The parameters are chosen so that the following equation is
satisfied for r =0, r = 1and r =2:
l1FMt =~l1vM~+(1 - ~)l1LM~ (7-39)
M, is the r-th statistical moment about the origin of the distribution F(I):
M, =11 I'F(I)dI (7-40)
Tosatisfy the material balance exactly, Equation 7-39 should have been sat-
isfied for r =0, 1,2, ... , oo , The errors introduced in the material balance
for two flash calculations on a mixture of 40% CO2 and 60% paraffins are
shown schematically in Figure 7-5. Cotterman et al. nave also suggested an
alternate flash calculation procedure, the quadrature method, which is ac-
tually a pseudocomponent procedure. For each pseudocomponent, the mo-
lecular distribution, f, is described by the following expression:
s
1 f(I)dI =E w(Ip) f (Ip)
1 p =1
(7-41)
s is the number of quadrature points, w is a weighting function, and Ipis a
quadrature point. Equation 7-36 is then changed to:
n m s
E z, +E l1kE wk(Ip)fk(Ip) =1
i=l k=l p=l
(7-42)
The continuous distribution has thus been replaced by m pseudocornpo-
nents.
GROUPING OF PSEUDOCOMPONENTS
The characterized mixture may consist of more than 80 components or
pseudocomponents. Obviously, for extensive process calculations, it isneces-
sary to drastically reduce this number. Pedersen et al. (1985) have developed
0. 012 I I I I i i i i i
Feed: 40% CO
2
, 60% par af f i ns ( cont i nuous)
0. 008
e
O
+-'
:J
.D
\....
+-'
Vl
0. 004
el
s,
'"
O
::E
0. 008
0. 006
o
O
O
P
T
, (
50 bar
512 K
- - speci f i ed
sum of l i qui d
and vapor
( cal cd~
I
Feed
p = 10 bar
T = 422 K
'I
1 (
100
Mol ecul ar Wei ght
Er r or i n mat er i al bal ance i ndi cat ed by shaded ar ea
150
200
Figure 7-5. Examples of the errors in the material balance when using continu-
ous thermodynamics (Cotterman and Prausnitz, 1985).
a procedure whereby all hydrocarbon segments of the C
7
+- fractions are
given approximately equal importance. This is accomplished via a grouping
of carbon number fractions on weight basis. The C
7
+-fraction is divided
into three (or more) groups that, by weight, are of approximately equal size.
The group critical properties are computed as weight-mean values. Two ex-
amples of grouping are given in Tables 7-3 and 7-4, with analytical data in
Table 7-2. It is seen that N2, CO2, and CH4 (C
l
) are given separate status,
ethane and propane are grouped together (C2-C
3
), and so are the C/s, Cs's,
and the Ca's.
Table 7-2
Analytical Data for North Sea Gas Condensate
Density
PNA-Distribution (Mole %)
(g/cm
3
) at Molecular
Components Mole % 15C, 1 atm Weight P N A
N2 0.64
CO2 9.16
C, 68.80
C2
8.43
C3 5.11
ic, 0.81
n-C.
1.45
i-C5 0.52
n-C5 0.53
C6 0.63
C7 0.83 0.741 96 0.50 0.42 0.08
C8 0.95 0.780 107
0.45
0.38 0.17
C9 0.52 0.807 121 0.48 0.27
0.25
C10 0.26 0.819 134
0.47 0.30 0.23
C11 0.20 0.810 147 0.56 0.27 0.17
C
'2
0.17 0.828 161 0.55 0.24 0.21
C
'3
0.16 0.849 175 0.54 0.22 0.24
C'4 0.15 0.857 190 0.49 0.27 0.24
C
'5
0.11 0.868 206 0.52 0.20 0.28
C'6 0.086 0.872 222 0.55 0.19 0.26
C17 0.078 0.859 237 0.57 0.20 0.23
C'8
0.068 0.854 251 C,8.O.70 0.11 0.19
C
'9
0.050 0.866 263
C20 0.046 0.873 C2o.339
C21 0.035 0.876
..
C22 0.025 0.876
C23 0.034 0.875
C24 0.023 0.877
C25 0.017 0.876
C26 0.018 0.878
C27 0.014 0.882
C28 0.012 0.886
C29 0.013 0.889
C30 0.047 0.908
CALCULATION OF THE IDEAL GAS HEAT CAPACITY
As described in the SRK equation of state section in Chapter 5, the ideal
gas heat capacity is calculated from a third-degree polynomial in the tem-
perature (seeEquation 5-36). For the lighter petroleum mixture constituents
C
7
), the coefficients C-C4 of the temperature polynomial are attributed
eharacterization Procedures 127
Table 7-3
Characterization and Grouping for North Sea Gas Condensate
(Analytical Data in Table 7-2)
Input to Flash Program
Components Mole % Weight % Te(K) Pe (atm) w
N2 0.64 0.63 126.2 33.5 0.040
CO2 9.16 4.19 304.2 72.8 0.225
C, 68.80 38.83 190.6 45.4 0.008
C2-C3 13.54 16.85 335.7 45.2 0.123
C,-C6 3.94 9.20 450.7 34.3 0.225
C,-C1O 2.30 8.57 565.0 31.6 0.408
C11-C
'5
0.94 5.13 653.3 25.1 0.656
C,6-C'8 0.68 6.60 766.2 14.8 1.085
Source: Pedersen el al. (1985).
Table 7-4
Results of the Characterization and Grouping of the Oil Mixture from
Data Shown in Table 4-7
Input to Flash Program
Components Mole % Weight % Te(K) Pe (atm) w
N2 0.34 0.10 126.2 33.5 0.040
CO2 0.84 0.41 304.2 72.8 0.225
C, 49.13 8.79 190.6 45.4 0.008
C2-C3 10.76 4.30 338.2 45.0 0.126
C,-C6 7.35 5.76 464.5 33.2 0.242
C7-C
'5
19.69 30.20 619.4 25.3 0.581
C,6-C26 7.75 23.85 760.2 15.0 1.046
C27-C76 4.14 26.58 925.6 11.3 1.585
fixed values, For C7 +-components the procedure suggested by Kesler and
Lee (1976) may be used (the calculated heat capacity is in Btu/lb):
C = - 0.33886 +0.02827K - 0.26105CF +0.59332wCF
C
2
= - (0.9291 - L1543K +0.0368K2) 10-
4
+CF(4.56 - 9.48w)10-
4
(7-43)
(7-44)
C3 = - 1.6658 X 10-
7
+CF(0.536 - 0.6828w) 10-
7
(7-45)
C4 =O (7-46)
(7-47) where CF =((12.8 - K)(lO - K)/(lOw))2
and K is the Watson characterization factor defined in Equation 7-33.
REFERENCES
Cavett, R. H., "Physcal Data for Distillation Calculation, Vapor-Liquid
Equilibria," 27th Midyear Meeting, API Division of Refining, San Fran-
cisco, CA, 'May 15, 1964.
Cotterman, R. L. and Prausnitz, J . M., "Flash Calculations for Continous
or Semicontinous Mixtures Using an Equation of State," Ind. Eng. Chem.
Process Des. Dev., 24, 1985, pp. 434-443.
Edmister, W. C., "Appled Hydrocarbon Thermodynamics. Part 4. Com-
pressibility Factors and Equations of State," Peto Refiner, 37, 1958, pp.
173-179.
Edmister, W. C., Applied Hydrocarbon Thermodynamics, Volume 2, Hous-
ton, Gulf Publishing Company, 1988.
Katz, D. L. and Firoozabadi, A., "Predicting Phase Behavior of Conden-
sate/Crude-Oil Systems Using Methane Interaction Coeffcients," f. Peto
Technol., 20, 1978, pp. 1649-1655.
Kesler, M. G. and Lee, B. l., "Improve Prediction of Enthalpy of Frac-
tions," Hydrocarbon Processing, 55, 1976, pp. 153-158.
Nes, K. and Westerns, H. A. van, Aspects of the Consttution of Mineral
Oils, Elsevier, New York, 1951.
Pedersen, K. S., Thomassen, P, and Fredenslund, Aa., "Thermodynamcs
of Petroleum Mixtures Containing Heavy Hydrocarbons. 1. Phase Enve-
lope Calculations by Useof the Soave-Redlich-Kwong Equation of State,'
Ind. Eng. Chem. Process Des. Dev., 23, 1984, pp. 163-170.
Pedersen, K. S., Thomassen, P., and Fredenslund, Aa., "Therrnodynamcs
of Petroleum Mixtures Containing Heavy Hydrocarbons. 3. Efficient
Flash Calculation Procedures Using the SRK Equation of State,' Ind.
Eng. Chem. Process Des. Dev., 24, 1985, pp. 948-954.
Riazi, M. R. and Daubert, T. E., "Simplfy Property Predctions,' Hydro-
carbon Processing, 59, 1980a, pp. 115-116.
Riazi, M. R. and Daubert, T. E., "Predicton of the Composition of Petro-
leum Fractions,' Ind. Eng. Chem. Process Des. Dev., 19, 1980b, pp.
289-294.
Sim, W. J . and Daubert, T. E., "Prediction of Vapor-Liquid Equilibria of
Undefined Mixtures,' Ind. Eng. Chem. Process Des. Dev., 19, 1980, pp.
386-393.
Thomassen, P., Pedersen, K. S., and Fredenslund, Aa. "Adjustment of C
7
+_
Molecular Weights in the Characterization of Petroleum Mixtures Con-
taining Heavy Hydrocarbons,' SEP Report 8617, Instituttet for Kerni-
teknik, The Technical University of Denmark, 1986.
Whitson, C. H., "Characterizing Hydrocarbon Plus Fractions,' SPE [our-
nal, 23, 1983, pp. 683-694.
Winn, F. W., "Physical Properties by Nornogram,' Peto Refiner, 36, 1957,
pp. 157-159.
Yarborough, L. "Application of a Generalized Equation of State to Petro-
leum Reservoir Fluds,' paper presented at the 176th National Meeting of
the American Chemical Society, Miami Beach, FL, 1978.
Chapter 8
Simulation of PVT- Experiments
The simulation results of this chapter are all based on the Soave-Redlich-
Kwong equation of state (Chapter 5) and the characterization procedure of
Pedersen et al. described in Chapter 7. The procedure for adjusting the mo-
lecular weight used for some mixtures is described in Chapter 10. AlI the
density calculations of this chapter have been performed using the proce-
dure of Peneloux et al. described in Chapter 5. The compositions of the mix-
tures referred to are shown in Chapter 4.
CONSTANT MASS EXPANSION
Table 8-1 shows the results of a constant mass expansion study (described
in Chapter 3) at T = 155C on mixture l. The results comprise saturation
point, liquid dropout curve, and single-phase compressibility factors. The
liquid dropout curve is plotted in Figure 8-1, together with calculated re-
sults. The results of using 6 and 20 hydrocarbon fractions to represent the
mixture are shown. With 20 hydrocarbon fractions twq, different kinds of
results are shown: in one case the measured molar composition is used, and
in the other case the molar cornposition has been changed corresponding to
a decrease in the molecular weight of the plus-fraction (C20 +) of 5.4 % .
With this adjustment, agreement with the measured dew-point pressure is
obtained. When adjusting the C2o+-molecular weight, the weight composi-
tion of Table 4-2 was maintained. A molecular weight adjustment was also
carried out when only 6 hydrocarbon fractions were used to represent the
mixture. It isseen from Figure 8-1 that too much liquid precipitation is pre-
dicted in the pressure range of 250-400 bar, whereas the measured and the
calculated results agree well at lower pressures. Torepresent this mixture, it
matters little whether 6or 20 hydrocarbon fraetons are used. Also it makes
little difference whether the molecular weight of the plus-fraction is ad-
justed or noto
Table 8-2 shows constant mass expansion results at T =71.6C for mix-
ture 3. The data include the bubble-point pressure, and the volumes at vari-
ous pressure stages relative to the volume at the bubble point. The results are
plotted in Figure 8-2. Also shown in Figure 8-2 are simulated results in
130
Simulation of PVT-Experiments 131
Table 8-1
Experimental Constant Mass Expansion Results tor 'Mixture 1
(Molar Composition In Table 4-2) at T =155C
(The Results are Plotted Together with Calculated Results in Figure 8-1)
L1quid Vol.
Pressure (% of Compressibility
(bar) Dew Pt. Vol.)
Factor
597.1
1.3729
577.8
1.3450
560.9
1.3208
540.5
1.2902
519.5
1.2596
495.1
1.2244
479.8
1.2027
462.7
1.1779
449.9
1.1589
434.8
1.1367
412.0
1.1043
393.0
1.0793
*388.0 0.00 1.0740
385.1 0.05
368.6 0.75
345.1 2.43
320.7 4.52
300.5 - 6.11
278.2 7.75
255.6 9.06
238.6 9.89
229.3 10.29
206.7 11.03
183.7 11.58
161.3 11.80
146.2 11.89
Dew point
which the characterized mixture contains 10hydrocarbon fractions, and the
flash is performed using the kij =Oflash versin described in Chapter 6.
The calculated and the simulated results agree very well.
DIFFERENTIAL LIBERATION
Table 8-3 shows the results of a differentialliberation study (described in
Chapter 3) at T =92.8C on mixture 2. The results comprise saturation
pressure, oil formation volume factors, gas/oil ratios, densities of the oil
phase, and gas-phase compressibility factors. The results for the oil forma-
15.0 i
Q)
E
:>
o
e>
+-'
c::
2!. 10.0
~
Q)
C>
4-
o
..
Q)
E
:>
5.0
o
:>
"O
:>
CT
--'
0.0
0.0
es' __
Exper i ment
20 CH- f r act i ons
20 CH- f r act i ons. MW adj ust ed
300.0 '00.0 100.0 200.0

Pr essur e ( bar )
Figure 8-1. Liquid dropout curvos lor mixture 1 at a constant mass expansion
study at T =155.0C. The experimental results are shown in Table 8-1. The den-
sity is calculated using the procedure 01 Peneloux et al. (see Chapter 5).
2.0 I i
+-'
c::
'C; 1.5
e,
c::
o
+-'
-o
'-
:>
+-'
'"
V1 1.0
..,
'"
Q)
E
:>
o
2:.
~ ,.5
o
:>
\ .
~--- --
Exper i ment
10 CH- f r act i ons. k
i j
= o
0.0 100.0 200.0 300.0 '00.0 500.0
Pr essur e ( bar )
Figure 8-2. Results lor the relative volumes 01 mixture 3 (see Table 4-4) in a con-
stant mass expansion study at T =71.6C. The experimental results are shown
in Table 8-2. The density is calculated using the procedure 01 Peneloux et al. (see
Chapter 5).
Simulation 01 PVT-Experments 133
Table 8-2
Experimental Constant Mass Expansion Results tor Mixture 3
(Molar Composition in Table 4-4) at T =71.6C
(The Results are Plotted Together with Calculated Results in Figure 8-2)
Pressure Volume/Volume
(bar) at Sato Pt.
401.7 0.9811
380.1 0.9836
362.5 0.9853
343.8 0.9873
325.3 0.9895
307.4 0.9915
291.4 0.9934
274.9 0.9955
254.3 0.9981
'239.0 1.0000
235.0 1.0033
215~1 1.0230
182.4 1.0678
142.8 1.1651
113.3 1.2946
91.7 1.4568
77.1 1.6193
65.7 1.8158
Sat ur al i onpol nt
Table 8-3
Experimental Dlfferen.tial Liberation Results tor Mixture 2
(Molar Composition in Table 4-3) at T =92.8C
(The Results lor the Oil Formation Factor and the GOR are
Plotted Together with Calculated Results in Figures 8-3 and 8-4)
Di! Compro
Pressure 011Form Solution Density Factor
(bar) Vol. Fact. GOR*' (g/cm
3
) Z
*267.9 1.396 132.9 0.718
239.6 1.365 119.4 0.726 0.889
209.4 1.329 103.7 0.735 0.888
179.8 1.297 89.6 0.744 0.876
150.0 1.265 75.7 0.753 0.880
112.9 1.227 58.6 0.765 0.894
72.4
1.186
40.4 0.778 0.904
33.1 1.144 21.9 0.791 0.949
1.0 1.067 0.816
Sal ur at i onpoi nl
St d, m3 gas par m3 01 r esi dual oi l al 15C, 1 bar .
-
o
2.0 I .
0
1 , , ! I , I
o.
0.0 100.0 200.0 300.0
Pr essur e ( bar )
Figure 8-3. Results for the oil lormation volume factor 01mixture 2 (see Table
4-3) in a differentialliberation study at T =92.8C. The experimental results are
shown in Table 8-3. The density is calculated using the procedure 01Peneloux et
al. (see Chapter 5).
<'")
o E
.- .
+' ' "
'" '" IX O>
-<'")
.- E
o "O
' " +'
'" '"
'" -
'" ::J
"O
Exper i ment
100.0
Pr essur e ( bar )
Figure 8-4. Results tor the gas/oil ratio 01mixture 2 (see Table 4-3) in a differen-
tial liberation study at T =92.8C. The experimental results are shown in Table
8-3. The density is calculated using the procedure 01Peneloux et al. (see Chapter 5).
'" Ql
'-
<'")
~
~ --- - -, o
""'E
'-
o
+'
u
'"
"-
1.0
Ql
E
::J
o
,.
e;
o
+'
ro
E
'-
.l'
o
200.0 I
o
MW adj ust ed
MW adj ust ed
-o
::J
"O
fiF
-;:/.;:- -'
-;?''-'
-;?'/ /
~ .~/.-.
; ' l " . : : : : : - A
:; .~"'"
~.~....
.. ~ .... /
"....
'" <l>
L 100.0
Exper i ment
20 CH- f r act i ons.
6 CH- f r act i ons.
0.0
0.0
200.0 300
f)
~<,
. . ~J!...I\I(i
Simulation 01 PVT-Experiments 135/;/ --,.~()~
<, - ("
tion factor are plotted in Figure 8-3 and the. results for the gas/oil ratio in~(BllUOlt,;j/ ;.'
Figure 8-4. Also shown in Figures 8-3 and 8-4 are the calculated results of-=>\ ;~
using 6and 20hydrocarbons to represent the mixture. Two different kinds or,. -- ti' P
results are shown, produced with the 20 hydrocarbon fractions. One corre- ",-'" ",flr)
sponds to the measured composition. The other one is obtained by treating ~~,
the molecular weight of the plus-fraction as an adjustable parameter to
which is attributed the value that makes the measured and the calculated
saturation points agree. The latter procedure was also used when the mix-
ture was represented using only 6hydrocarbon fractions. The experimental
and the calculated results agree very well. The best results are, in fact, ob-
tained when the mixture is characterized using only 6 hydrocarbon frac-
tions.
Table 8-4 shows, for mixture 4, the compositions of the gas phase libera-
ted at each stage of adifferentialliberation study at T = 141C. The results
are shown schematically in Figure 8-5. Also shown in Figure 8-5 are the cal-
culated results where the characterized mixture comprised 6 hydrocarbon
fractions. The measured and the calculated results agree very well.
Table 8-4
Experimental Results of Differential Liberation Study on Mixture 4
(Molar Composition in Table 4-5) at T =141oC.
Composition of Differentialiy Liberated Gas (Mole %)
(The Results are Plotteq Together with Calculated Results in Figure 8-5)
P (bar) 317.3 292.4 249.3 214.9 174.5 138.4 91.4 49.3 16.9 1.0
N2 1.06 1.06 1.06 1.05 0.95 0.78 0.54 0.27 0.07 0.01
CO2 3.85 3.87 3.97 4.12 4.23 4.28 4.57 4.99 4.46 1.21
C, 71.31 71.93 73.67 75.72 75.64 75.01 71.39 64.44 41.40 7.15
C2 9.01 9.00 9.13 9.37 9.58 10.27 11.58 14.75 20.13 9.39
C3 5.55 5.50 5.42 5.32 5.32 5.76 6.79 9.33 18.31 14.93
- C. 0.88 0.88 0.84 0.79 0.77 0.82 0.97 1.33 3.01 3.45
n-C. 1.92 1.91 1.79 1.61 1.56 1.67 2.05 2.70 6.72 9.34
ic, 0.59 0.59 0.54 0.43 0.42 0.43 0.54 0.66 1.79 2.29
c, 0.74 0.75 0.66 0.50 0.49 0.49 0.66 0.75 2.11 6.39
Cs 0.68 0.68 0.57 0.35 0.36 0.30 0.47 0.45 1.24 9.27
C7 0.84 0.79 0.60 0.25 0.28 0.16 0.33 0.26 0.65 14.80
Ca 0.80 0.70 0.45 0.14 0.14 0.03 0.10 0.07 0.11 12.79
C9
0.49 0.42 0.24 0.06 0.05 0.00 0.01 0.00 0.00 5.44
ClO+ 2.27 1.92 1.06 0.29 0.21 0.00 0.00 0.00 0.00 3.55
eONSTANT VOLUME DEPLETION
Table 8-5 displays the liquid dropout curve, and the gas phase compress-
ibility factors, measured at a constant volume depletion experiment (de-
100.0 I I
75.0
. . . .--.
--
..
ClJ
o
E
e Measur ed mol e %of C,
Measur ed mol e % of C
2
- C
6
Measur ed mol e% of C
7
+
- - - - - Cal cul at ed r esul t s. 6 CH- f r act i ons
'00.0
e
o
..
'"
\...
..
e
ClJ
u
e
o
u
o....... 0- - 0
"-.....~O----o----o---o
_6--.
~
200.0
Pr essur e ( bar )
100.0 300.0
Figure 8-5. Results of a differential liberation study on mixture4(see Table4-5)
at T =141C. Hydrocarbon concentrations inthe liberated gas. The experimen-
tal results are shown inTable8-4.
Table 8-5
Experimental Results of a Constant Volume Depletion Study
at T =82.5C on Mlxture 9 (Molar Compositlon in Table 4-11)
(The LiquidVolumeis inPercent of the 8aturation Point Volume.
The LiquidDropout Curve is Plotted inFigure 8-6Together with
Calculated Results)
Gas Phase Cumulative
Pressure Volume % Compressibility Mole Fraction
(bar) Liquid Factor Removed
281 0.0 0.893 0.0000
235 1.3 0.852 0.1191
180 3.4
0.836 0.3077
125 4.6 0.858 0.5205
84 4.7 0.893 0.6841
49 4.4
0.932 0.8121
scribed in Chapter 3) on mixture 9 at T =82.5C. The liquid dropout curve
isplotted in Figure 8-6. Alsoshown in Figure 8-6 are calculated liquid drop-
out curves, where the measured C7 vmolecular weight is used, and where
the C7+-molecular weight has been decreased by 8.8%, to get agreement
Simula tia n 01 PVT-Experiments 137
10.0 I I
ClJ
E
::J
~ 7.5
..
e
o
o..
"
ClJ
Cl
4-
o 5.0
...

..... ,
-,
<,
,
\
. \
\
\
\
\
.\
"
"
....-
,..,
,.., .
ll!
::J
o
:> 2.?
"O
::J
O"
0.0 0.0
100.0 200.0
Pr essur e ( bar )
300.0 400.0
Figure 8-6. Liquiddropout curves for mixture 9 ina constant volume depletion
study at T =82.5C. The experimental results are shown inTable8-5. The den-
sity is calculated using the procedure of Peneloux et al. (see Chapter 5).
Table 8-6
Experimental Results of a Constant Volume Depletion Study
at T =82.5C on Mixtute 9 (Molar Composition InTable 4-11)
(Molar Compositions of the Gas Mixtures Removed at Each Depletion 8tage)
P (bar) 235 180 125 84 48
N2 0.08 0.08 0.09 0.09 0.08
CO2
2.45 2.46 2.47 2.49 2.52
C, 82.97 83.67 84.28 84.35 83.88
C2 5.77 5.77 5.79 5.84 5.94
C3 2.82 2.78 2.77 2.82 2.96
-C, 0.54 0.53 0.52 0.54 0.57
n-C, 1.19 1.15 1.13 1.16 1.22
te, 0.49 0.46 0.45 0.46 0.49
e, 0.56 0.52 0.50 0.51 0.56
C6 0.65 0.57 0.53 0.53 0.58
C7+
2.49 2.01 1.47 1.21 1.20
with the measured dew point. The calculations were performed using 20
hydrocarbon fractions. The measured and the calculated results agree rea-
sonably well after the adjustment of the C7 +-molecular weight. Table 8-6
shows the measured molar composition of the gas removed at each depletion
100.0 .,--------------------------------,
25..0
~ o -D o o
...-------.-- - - . .. .. .
75.0
..
'"
o
E
~
c:
:! 50.0
+-'
'"
...
+-'
e
'"
v
e
o
u
Measur ed mol e %of C
1
o Measur ed mol e %of C
2
- C
6
Measur ed mol e %of C
7
+
Cal cul at ed r esul t s, measur ed MW
Cal cul at ed r esul t s, adj ust ed MW
0.00.0
50.0 100.0 150.0 200.0 250.0
Pr essur e ( bar )
300.0
Figure 8-7. Results 01 aconstant volume depletion study onmixture9(see Table
4-11) at T =82.5C. Hydrocarbon concentrations inthe removed gas. The ex-
perimental results are shown inTable8-6.
1.0
Exper i ment
Cal cul at ed usi ng~measur ed MW
Cal cul at ed usi ng adj ust ed MW
"
'"
E; .75
E
'"
a:
c:
o
+-'
v
re
....
u- .5
'"
o
:>:
'" >
+-'
'"
~ .25
=>
u
I ! I ! '-. ! .......! I
0.0
0.0 50.0 100.0 300.0 350.0 150.0 200.0 250.0
Pr essur e ( bar )
Figure 8-8. Cumulative mole Iractions 01 mixture 9 (molar composition inTable
4-11 removed as a lunction 01 pressure ina constant volume depletion study at
T =82.5C. The experimental results are shown inTable8-5.
'00.0
Simulation Di PVT-Experiments 139
stage. The results are illustrated in Figure 8-7, where the calculated results
are also shown. It isseen from Figure 8-7 that it makes very little difference
for the gas-phase compositions whether the molecular weight of the C7 +-
fraction is adjusted or noto Measured and calculated results for the cumula-
tive mole fraction removed at each depletion stage are shown in Table 8-5.
These results are plotted in Figure 8-8 together with the calc~lated results.
Table 8-7
Molar Compositions of Separator Fluids of Mixture 5
(Molar Composition in Table 4-6)
Stage 1:T =44C. P =67.2 bar
Gas Liquid
Measured Calculated Measured Calculated
Component Mole % Mole % % Dev. Mole % Mole % % Dev.
N2 0.13 0.14 - 0.01 0.01 -
CO2 2.65 2.69 1.5 1.35 1.24 - 8.1
C, 84.50 85.15 0.8 18.84 21.62 14.8
C2 7.64 7.52 - 1.6 6.14 7.09 15.5
C3 2.83 2.63 - 7.1 5.12 6.19 20.9
C. 1.32 1.14 - 13.6 5.18 5.98 15.4
-
C5 0.53 0.42 - 20.8 5.04 5.27 4.6
C6 0.20 0.16 - 4.91 4.78 - 2.6
C7+ 0.20 0.15 - 53.41 47.82 - 10.5
Stage 2:T =1SoC, P =1 bar
Gas Liquid
Measured Calculated Measured Calculated
Component Mole % Mole % % Dev. Mole % Mole % % Dev.
N2 0.01 0.03 - 0.00 0.00 -
CO2 3.72 3.03 - 18.5 0.00 0.03 -
C, 51.92 53.67 3.3 0.00 0.28 -
C2 16.60 16.82 1.3 0.18 0.60 -
C3
12.44 12.75 2.5 0.96 1.82 89.6
C. 8.64 8.65 0.1 :3.21 4.20 30.8
C5 3.92 3.55 - 9.4 5.67 6.42 13.2
C6
1.44 0.99 - 31.3 6.89 7.31 6.1
C7+ 1.30 0.51 - 60.8 83.09 79.34 - 4.5
% Dev. =100 x (Calculated - Measured)/Measured.
(Only stated,for measured mole %'5 above 0.5.)
An almost-perfect agreement with the experimental results is observed,
when the molecular weight of the plus-fraction is adjusted to match the
measured dew point.
SEPARATOR TEST
Table 8-7 shows the results of a two-stage separator test (described in
Chapter 3) on mixture 5. The results comprise measured and calculated gas
and liquid-phase compositions. The characterized mixture consisted of 10
hydrocarbon fractions. The measured and calculated gas-phase composi-
tions agree very well, whereas larger deviations are seen for the liquid
phases.
SUMMARY OF THE RESULTS OF THE PVT-SIMULATIONS
The PVT-simulations provide evidence that good agreement with mea-
sured PVT-results can be obtained using the SRK equation of state in con-
junction with the C7+-characterizaton procedure-of Pedersen et al. (pre-
sented inChapter 7). Sixhydrocarbon fractions are, inmost cases, sufficient
to represent the hydrocarbons of areservoir fluido The procedure described
in Chapter 10, for adjusting the molecular weight of aplus fraction against
a measured saturation point, isvery useful for obtaining reasonable results
for the gas/liquid ratios of a gas condensate mixture. An adjustment of less
than 10% of the molecular weight of theplus fraction doesnot influence the
phase properties significantly.
Chapter 9
Comparison Between Experimental
and Predicted Thermodynamic
Properties
The calculated results of this chapter areall based on the flash and phase
envelope algorithms"of Michelsen described in Chapter 6 and on the SRK
equation of state (Chapter 5) coupled with the C7 +-characterization proce-
dure of Pedersen et al. (seeChapter 7). The procedure, used in some cases
for adjusting the molecular weight of aplus-fraction, isdescribed in Chap-
ter 10. Thecompositions of the mixtures mentioned areshown inChapter 4.
OEW ANO BUBBLE POINTS (PHASE ENVELOPES)
Table 9-1 presents measured and calculated dew and bubble points for
mixtures 1-10. Calculated results, obtained using the standard flash (non-
zero binary interaction coefficients for interactions where non-hydrocar-
bons take part), and 20 hydrocarbon and 6 hydrocarbon fractions, respec-
tively, to represent the mixture, are shown. The simplified two-phase
(P,T)-flash (ki=O), described inChapter 6, isalsoused to yield results. The
last column of Table 9-1 shows the needed adjustment to the molecular
weight of the plus-Iracton (maintaining themeasured weight composition),
to get agreement with the measured saturation (dew or bubble) point. Re-
sults of using a characterized mixture consisting of 6 and 20 hydrocarbon
fractions areof comparable quality. Generally, it makes very little difference
whether the standard flash or the ki=O flash is used. For all mixtures ex-
cept mixture 7, an adjustment of lessthan 10% of the molecular weight of
theplus-fraction issufficient to get agreement with the measured saturation
point. In Figure 9-1 the calculated phase envelope of mixture 10isshown
together with four measured saturation points. The characterized mixture
consists of 20 hydrocarbon fractions. Figure 9-2 shows, for mixture 3, the
results of three different phase envelope calculations corresponding to the
results in Table 9-1.
141
Table 9-1
Measured and Calculated Saturation Pressures for 10 Different Petroleum Mixtures
(The Molar Compositions are Given in Chapter 4)
Calculated Calculated Calculated
Measured p, with Psa' with p, with MW-
Mixture Temperature
P
oo
'
20 CH 6 CH kij = O adjustment
No. (0C) (bar) (bar) % Dev. (bar) % Dev. (bar) % Dev. (%)
1 155.0 388.0 398.3 2.7 386.4 - 0.4 392.4 1.1 - 5.4
2 92.8 267.9 259.5 - 3.1 263.5 - 1.6 251.4 - 6.2 9.1
3 71.6 239.0 235.9 - 1.3 243.6 1.9 233.1 - 2.5 2.6
4 141.0 340.3 346.5 1.8 347.2 2.0 339.5 - 0.2 - 8.3
5 129.0 464.0 463.7 - 0.1 458.5 - 1.2 457.6 -1.4
0.1
6 93.3 274.5 276.5 0.7 282.1 2.8 274.2 - 0.1 - 3.0
7
93.4
264.7 248.0 - 6.3 252.1 - 4.8 248.0 - 6.3 18.5
8
93.4
264.7 279.6 5.6 272.9 3.1 279.6 5.6 9.6
9 82.5 281.0 334.5 19.0 312.2 11.1 330.5 17.5 - 8.8
10 32.2 462.8 435.8 - 5.8 415.3 - 10.3 405.1 - 12.5 7.6
10 67.8 401.6 412.6 2.7 387.2 - 3.6 390.1 - 2.9 - 2.4
10 125.4 303.5 337.3 11.1 303.5 0.0 324.3 6.9 - 6.0
10 169.4 229.6 246.2 6.7 198.6 - 13.5 238.1 3.7 - 2.2
%AAD - - - 5.0 - 4.3 - 5.1 6.4
% BIAS - - - . 2.6 - - 1.1 - 0.2 0.9
Notetion:
CH:
kij = o:
% Dev.:
MW-adjustment:
% AAD:
% BIAS:
Number of hydrocarbon fractions in the characterized mixture.
The kij = Oflash described in Chapter 6 has been used on a characterized mixture consisting of 20 CH-fractions.
100 x (Calculated - Measured)/Measured.
The adjustment needed of the molecular weight of the plus-fraction (see Chapter 10) to get agreement with the measured saturation point using 20
CH-fractions and standard k-values (see Chapter 7).
% Average absol ute deviation.
% Average deviation.
Pressure (bar)
Pressure (bar)
N
o o
o o
,~ o g o g N g
gr-~~~ - ; o ~ ~o~ o~ ~o
o
o
o
o
o
o
o
W
::J
a.
o
w
C)
e
Ol
ID
a.
'O
:;y
W
en
ID
ID
::J
<
ID
O'
'O
ID
en
~
o
o
o
o
-<
ru
3
' O N
~~
r-e-
e
-s
ro
r l g
n n n t ' Tl
Q Q.I Q ) (
'O
n n n ro
e e e -s
Q Q Q 3
r-+ . r + ro
ro ro ro :::J
o. o. . o. r +
N N
<> <:>
- < o
ro
3 o
'O
ro
;;)
rt
e o
..,
ro o
g
o
o
o
n ",
'" x 'O
n '" e ..,
'" 3
rt '"
~ ~
N
o
o
o
:3
X
e
ID
~
o
o~ ~
N
o
o
o
o
o
144 Propertes 01 Ols and Natural Gases
FLASH RESULTS
Measured and calculated liquid-phase mole fractions, at T =93.4DC and
various pressures, are shown in Table 9-2 for mixture 7. The calculations
were performed, using 20 and 10 hydrocarbon fractions, respectively, to
represent the mixture. Table 9-3 shows measured and calculated phase com-
positions at T =93.4
D
C and P =201.1 bar. The calculated results for the
totalliquid fr~ction agree well with the experimental observations, whereas
larger deviations are found for the compositions of each of the two phases. It
makes little difference whether the characterized mixture consists of 20or of
10 hydrocarbon fractions.
Table 9-2
Measured and Calculated Liquid Mole Fractions
at T =93.4C of Mixture 7
(Molar Composition in Table 4-9)
Calculated Calculated
Measured Liquid Mole Liquid Mole
Pressure Liquid Mole Fraction Fractlon
(bar) Fraction Using 20 CH % Dev.
~
Using 10 CH % Dev.
201.1 0.871 0.861 - 1.1 0.855 - 1.8
173.4 0.775 0.788 1.7 0.784 1.2
138.9 0.681 0.704
3.4 0.701 2.9
104.4 0.606 0.625 3.1 0.624 3.0
70.0 0.537 0.550
2.4
0.550
2.4
35.5 0.479 0.475 - 0.8 0.477 - 0.4
%AAD - - 2.1 - 2.0
% BIAS - - 1.5 - 1.2
% Dev. is defined in Table 9-1.
CH: Hydrocarbon fractions used in the calculation.
DEN SITIES
The experimental gas-phase compressibility factors (of Table 8-1), are
plotted in Figure 9-3 for mixture l. Also shown are the calculated results of
using the SRK-equation, and the procedure of Peneloux et al. (a11density
correlations in this section are described in Chapter 5). The results cover the
pressure range 400-600 bar. The SRK-equation gives the better results at the
highest pressures, whereas at lower pressures, the better results are obtained
with the procedure of Peneloux et al.
Figure 9-4 shows the results for the gas-phase compressibility factor of the
gas liberated, in adifferentialliberation study on mixture 2. The experimen-
Comparing Experimental and Predicted Thermodynamic Properties 145
..
lO
.c
:> o <J : al
(") al C\J Cl
CI) M C\J ~ M N
l1i
al
O
I
~ ~ ~ ~ C\J
~
I I I I I
e
'g
CI)
o; X
r-, C\J C\J al ll) ll)
Cl co ll)
~ ~
(")
o
c::
" 5 U
ll) M N ~ ~ N M
s
~<D v v
'"
. U
o
Q.
E :> ~ v C\J C\J ID co (")
o CI) N N N
<i (") ll) <Xi
U O
I
~ ~ ~ ~ C\J
CI) ;f. I I I I I
1/)
'" s:
Q.
'g
-
~x r-- C\J r-, -e- (") ll)
'5 . . ! ! ! U
<J : co
.n
co (") (") t-,
CT ::lo M N N N
:::
ll) ~ ~
~N
v v
U
'g
CI)
ll) ID ID co ll) v ID
;;
v
(") co
~
V
1/)
cri ~ ~ ~
ll) r-, (") r-,
CI) ~ o
(")
,...:
ll) l1i <Xi
O
-r-
I I I I
ll)
;f. I I
'g
CI)
o; X
r-- '<f o v
(") ID ll)
" 5 U eo r-, ll) f'-:
(") (")
v
N
(") ~ o 6 6 o
e ~<D al
o
'" .. U
.
o
Q.
:> al ID co
(") r-,
ID co
E CI)
6 o
(") cri l1i
N <i
o O ~
I I I I
C\J
U
~
I I
CI) ..
1/)
'" s: 'g
Q.
CI)
ch
" X al
(")
o ll) (") r-, C\J
'"
. . ! ! ! U
v
r-,
~
r-, (") (")
<Xl
"
i:lo
N M ~ 6 6 o 6
N
al
U
'g
CI)
ll) r-- ID o ll) co al
;;
ID ~ ll) eo
(") (")
o
1/)
a;
<
:B
-o
(")--
cf, QI
Ql o
:c:::
~:.
en
e
o
E
s
Q.
E
o
o
QI
:G
.r:.
o-
"C
QI
-
lO
'5
u
~
"C
e
lO
"C
QI
..
::J
en
lO
QI
:::
e
O
.~
:;
u
-;
u
Q)
:S
S
1:)
Q)
~ en
'" :J
Q) en _ e
.c O
~ "<3
~~
1: ) e
Q) O
e.c
~~
1: ) o
en ~
- 1: )
~: t
I
;f.o
450.0 500.0
Pr essur e ( bar )
Figure 9-3. Gas phase compressibility factor of mixture 1(molar composition in
Table4-2) inaconstant mass expansion study at T = 155.0
o
C. The experimental
results are shown inTable8-1.
146
Properties 01 O/s and Natural Gases
1.5
o-,
+'
Exper i ment
Penel oux
SRK
<.
o 1.4
+'
u
ro
u,
.,-
. . . . . . . . . . . .
...
... ... .
... .
.a 1.3
...
. . . . . . . . . . . . .
...
... .
'"
'" <lJ
<.
o.
E
o
u 1.2
<lJ
'" ro
s:
e,
...
...
... """ .
... .
'" ro
C,!) 1.1
I I I I 1 I
1.0
350.0 400.0
550.0
1.25
1.1
s, 1. 15
o
+'
u
ro
u; 1.1
'>,
+'
-
1.05
.o
'"
1.0
'" <lJ
<.
o.
E
.95 o
u
~
ro .9
s:
e,
'"
'"
.85
<.O
.8
.75
Exper i ment
__ __ __20 CH- f r act i ons, MW adj ust ed
__ . - 6 CH- f r act i ons, MW adj ust ed
-----------'-'--=--::.---- --. .

600.0
100.0 300.0
100.0
Pr essur e ( bar )
Figure 9-4. Results lor the gas phase compressibility tactor 01 mixture2(molar
composition inTable4-3) ina differential liberation study at T =92.8C. The ex-
perimental results are shown inTable8-3. Thecompressibility lactor iscalculated
using the procedure 01 Peneloux et al. (see Chapter 5).
0.0
tal results are listed in Table 8-3. Also shown in Figure 9-4 are calculated
results where the procedure of Peneloux et al. is used. The characterized
mixture consists of 20 and of 6hydrocarbon fractions, respectively. With 20
hydrocarbon fractions, results produced using the measured molar composi-
tion are shown, and also when the molecular weight of the plus-fraction is
adjusted to match the measured bubble point. The results with 6hydrocar-
bon fractions are obtained with an adjusted molecular weight. Figure 9-5
shows the corresponding results for the liquid-phase density in the same dif-
ferential liberation experiment. For both the gas and the liquid phases it
makes very little difference whether 20 or 6 hydrocarbons are used in the
calculations or whether the molecular weight of the plus-fraction isadjusted
or noto The results of liquid density calculations using 5 different density
correlations are shown by Figure 9-6. The best results are obtained with the
procedure of Peneloux et al., with the procedure of Standing and Katz, and
with the Alani-Kennedy equation. Somewhat larger deviations are found
using the API-procedure, but the deviations from the measured results are
stillless than 3 to 4%. Much larger deviations are found when the liquid-
phase density is calculated using the SRK-equation.
For mixture 9, avery large deviation was observed between the measured
and the calculated dew point (see Table 9-1). This deviation can be elimi-
1.0
Exper i ment
20 CH- f r act i ons ( and 6 CH- f r act i ons) , MW
adj ust ed.
.9
'"E
..!::.
01
.8
~--:.;:::.
>,
+'
.7
'" e
<lJ
'"
.6
.5 I I I ,1 I
0.0 100.0 100.0
Pr essur e ( bar )
Figure 9-5. Results tor the density 01the liquidphase 01 mixture2(molar compo-
sitioninTable4-3) inadifferentialliberation study at T =92.8C. The experimen-
tal results are shown inTable8-3. The density is calculated using the procedure
01 Peneloux et al. (see Chapter 5).
300.0
.85 , I
ro
E
~
00
---'-
'-'--.
I I I I I ---,"- I Pressure
.85
0.0 50.0 100.0 150.0 100.0 150.0 300.0 ( bar )
Exper i ment
St andi ng- Kat z
Penel oux
APl
- - SRK
<-, :::::-.::.::::-.-....: ..
81-', .
. '.
" . ......
" ...... ' .
", " ~..".'.",'<:::::... ..
.......... . ~.
. . . . . . . . . .
" . . . . . . . . . .
. 75
c-,
+-'
-.
'"
'" e
Q)
Cl
---
- . " " - . . . . . !
.7
-'-

Figure 9-6. Measured and calculated liquidphase densities of mixture 2(molar
composition inTable4-3) ina differential liberation stndy at T =92.8C. The ex-
perimental results are shown inTable8-3. The employed density correlations are
described inChapter 5.
...
o
+-'
:;; .95
u..
c-,
+-'
-
.o
'"
'" Q)
... .9
o.
E
o
u
~
-o
s:
a..
'"
.85
'" <.!)
Exper i ment al
Cal cul at ed usi ng measur ed MW
__ - Cal cul at ed usi ng adj ust ed MW
~
~
~
~
"
"
-:
<:
<,
...
....
.....
/
/ .
/
/
/
,/
,/
/
-----
.8 I ! I I I ,
0.0 50.0 100.0 150.0 100.0 150.0 300.0
Figure 9-7. Gas phase compressibility factor of mixture 7(molar composition in
Table4-9) inaconstant volume depletion experiment at T =82.2C. The experi-
mental results are shown inTable8-5.
Pr essur e
( bar )
Comparing Experimental and Predicted Thermodynamc Properties 149
nated by an adjustment of the molecular weight of the ClO+-fraction of
- 8.8 %. Figure 9-7 shows measured (see Table 8-5) and calculated gas-
phase compressibility factors of the constant volume depletion experiment
on this mixture. Calculated results are shown with and without adjustment
of the molecular weight. The procedure of Peneloux et al. was used in the
density calculations. Better results are obtained with the adjusted molecular
weight than with the measured molecular weight, and thus the adjustment
has in this case not only served to improve the dew point result, but it has
also improved the results for the gas-phase density and, as shown in Figure
8-6, for the gas/liquid ratio .
The liquid densities of mixture 6are shown in Table 9-4, as measured in a
differentialliberation experiment at T =93.3C. The results are plotted in
Figure 9-8, which also illustrates calculated results produced by using five
different density correlations. The best results are obtained with the API-
procedure, but none of the calculated results deviate by more than about
5% from the measured ones.
Table 9-5 shows measured liquid phase densities for mixture 8 at
T =93.4C and various pressures. These measured results are plotted in
Table 9-4
Measured Liquid-Phase Denslties 01 Mixture 6 in a Differential
Liberation Study at T =93.3C.
(The Results are Shown Graphically in Figure 9-8, Together with
Calculated Results)
Pressure (bar) Density (g/cm
3
)
389.8 0.677
352.6 0.670
319.7 0.665
288.3 0.661
'274.5 0.660
227.0 0.680
193.7 0.693
148.1 0.711
109.9 0.727
70.6 0.742
31.4 0.759
1.0 0.814
Sat ur at i onpoi nt
.90 i I
.85
.8
M
E
~
CJ >
. 75
>.
..,
V>
e
'"
D
. 7
Exper i ment
St andi ng- Kat z
Penel oux
Al ani - Kennedy
- - - API
_ _ SRK
-..;;:::
---- # - ~ -..:=-: -===-
_/ .

....
..........
. . . . . . .
<,
<,
..... .
<,
..... ---
<,
'<,
...... - __e __
---
---
-<;
.65
-...... _ ..-..-_.-_.- ---
.6
0.0 50.0 100.0 150.0 200.0 250.0 300.0 350.0 400.0
Pr essur e ( bar )
Figure 9-8. Measured and calculated liquidphase densities of mixture 6(molar
composition inTable4-7) inadifferentialliberation study at T =93.3C. The ex-
perimental results are shown inTable9-4. The employed density correlations are
described inChapter 5.
Table 9-5
Measured Liquid-Phase Densities 01 Mixture 8 at T =93.4C
(The Results are Shown Graphically inFigure 9-9: Together with
Calculated Results)
Pressure (bar)
Density (g/cm
3
)
201.1
0.6565
173.4 0.6536
138.9
0.6538
104.4 0.6753
70.0
0.7160
35.5
0.7209
Figure 9-9, together with the calculated ones. Five different density correla-
tions were tried. Except with the SRK-equation, all the calculated results
deviate less than 5% frorn the measured results. The best results are, in this
case, obtained with the procedure of Peneloux et al., and with that of Stand-
ing and Katz.
.75

,
<,
''-'-,-
<, ..............
<, <,
.............
...........................
.--- .
..--.~-:---:-::::--- --
..--~
.7
M
E
u
.....
~
.65
.>,
..,
V>
e
'"
D
.6
-,
<,
<,
<,
'"

Exper i ment
........
........
St andi ng- Kat z
Penel oux
API
Al an i - Kennedy
SRK
........
........
--- =--,
[ I ! !
.55 I !
0.0 50.0 100.0 150.0 200.0 250.0
Pr essur e ( bar )
Figure 9-9. Measured and calculated liquidphase densities of mixture8(molar
composition inTable4-10) at T = 93.4C. The experimental results are shown in
Table9-5. The ernployer density correlations are described inChapter 5.
ENTHALPY
No published data have been Iound for the enthalpy of petroleum reser-
voir fluids. The data presented here are for hydrocarbon boiling point frac-
tions, shown in Table 9-6 as boiling point curves (Lenoir and Hipkin, 1973).
To use the characterization procedure of Pedersen et al. it is necessary to
"convert' the distillation data of Table 9-6 to concentrations of carbon nurn-
ber fractions. This isdone using the data presented by Katz and Firoozabadi
(seeChapter 7) for boiling point ranges of carbon nurnber fractions. A usual
C7 +-characterization is then carried out. The ideal gas enthalpy was calcu-
lated using the procedure of Kesler and Lee described at the end of Chapter
7. The residual enthalpy was calculated from the SRK-equation of state us-
ing Equation 5-37. To facilita te cornparisons between measured and calcu-
lated results, the enthalpies given by Lenoir and Hipkin have been adjusted
to the same base level as the one used in the calculations (i.e., the ideal gas
enthalpy equals Oat T =273.15 K),
IIII,I!
I
111
!'I
"~!
11
ni
I ~ i
i
,,/:
ii','l' ,1
,!
!
,.1
I!I
,II" ",r, , l'l
j I
1 '
I,t'
'In
u
'1
,
I I
i
l
,1"
'i' ,[
i;
!I,
!
"
:1
152 Propertes 01Oils and Natural Gases
N
....
en
.c
Cl
::J
o
..
.c
- o
....
en
1/1
CII
..
::J
Cl
:
.:
c::
;:
o
s:
en
1/1
- :;
1/1
CII
a::
>-
l Oa.
a,(ij
CII:5
- c:: .c
w
t!!!
~
1/1
CII
..
::J
X
:
E
::J
CII
'O
..
-
CII
a.
..
o
-1/1
CII
>
..
::J
U
c::
.2
~
:;:;
1/1
e
.~;:;'
VI E
c: (.)
( 1) -
COI
-
>-
=i1:
c: (.)
( I ) [ j ,
c_
>-
'1:
c: (.)
( I ) [ j ,
c_
00 11{) 1(') 1"

C\J (') '<1" '<1"11{)
~ ~ ~ ~ al
I{)lr.o
1{) 1{)
00 00
r
m
::::-
-s.
_"O
3: ~
E . ~
8
c
011{)
e
.~s
a.I
"-u
-i e
'"
60.0
Comparing Experimental and Predicted Thermodynamic Properties
50.0
40.0
30.0
<lJ
o
E
<,
.."
..::, 10.0
20.0
'>,
o.. 0.0
ro
.c
;g -10.0
-20.0
-30.0
-40.0
-50.0
250.0
153
//
/
/
//
/ Measured
./ - Calculated
300.0 350.0 400.0 450.0
Temperature (K)
500.0 550.0 600.0
~I I I I I I I I I I I
"" <.!)
Oli:"
"O'=~IO I~ 10 11{) 1'<1"1'<1"1 1'<1"I~
~'5c ~ '!f ~ ~ U; -;, g g ~
ID '0
a..
"" .s
s:
c.
"" Z
I I I I I I I I I
;gl - O) u.
~"OC:~Ol{)ool{)r.oo>r.oC\J C\J
.3 c:::_II{)IC\J lr.oloIC\J I"IC\J loolo
w~.g ~~C\J C\J C\J (')(')'<I"
a..
'<1"10>10>10> 11{) IllJ 11{) 11{)
LO CX) T""" (") I D f'.. f'.. co
r.o ) ": ": " ": " "
Figure 9-10. Measured and calculated enthalpy results at 1400 psia tor the jet
naphtha Ior which the boiling point curve is shown in Table 9-6.
90.0 fl ----------'--------- -,
70.0
50.0
<lJ
o 30.0
E
<,
.."
"'"
:r: 10.0
-10.0
-30.0
Measur ed
-- Calculated
I I I ! I I I I I
-50.0
250.0 300.0 350.0
"" s:
E
g-I I I I I I I I
z
-(1)1 -
" ' ) O) u.
"O'=~IC\J 10>1(') 1
00
1" 10>I~ 1" 1" c:::_ v lO r-, 00 o C\J v io CD
W 0.= C\J C\J C\J C\J (') (') (') (') (')
ID o
a..
C\J IO>I"I~I(')IOI~IO I D O> o O' > o T""" T""" C\ J
f'.. t--: ~ r---: ~ ~ ~ aJ
ol ~l gl ~I Rl gl ~I ~~
'" .,
u....J
o ..
o .,
I D t'
- ::J
o
(f)
400.0 450.0 500.0
Temperature (K)
550.0 600.0 650.0
Figure 9-11. Measured and calculated enthalpy results at 30 psia tor the low-
boiling naphtha tor which the boiling point curve is shown in Table 9-6.
154 Properties o] Oi/s and Natural Gases
100.0 r - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ,
20.0
80.0
60.0
'" a 40.0
..!:.
->
"'"
- 20.0
,
el.
~ 0.0
+'
B
.40.0
Measur ed
- - - Cal cul at ed
60.0
-80.0 I .,.. I I 1 I I I
250.0 300 350.0 400.0 450.0 500.0 550.0 600.0
Temper at ur e ( K)
Figure 9-12. Measured and calculated enthalpy resulls al 700 psia for lhe gas/oil
for which lhe boiling poinl curve is shown in Table 96.
SUMMARY OF THE RESULTS FOR THERMODYNAMIC
PROPERTIES
Dew and bubble point pressures can usually be predicted to within 5%.
Sixhydrocarbon fractions are suHicient to represent the eornponents of ape-
troleum reservoir fluido The fast flash version (kij =O) for the mixtures in
this chapter (which all contain lessthan 15% non-hydrocarbons) giveresults
comparable in quality to those found with the standard flash.
The SRK-equation and the procedure of Peneloux et al. may both, with
good results, be used for calculation of gas-phase densities. The SRK-equa-
tion predicts insufficient liquid-phase densities, however, good results are
obtained with the procedure of Peneloux et al., with the API-procedure,
with the procedure of Standing and Katz, and with the Alani-Kennedy
equation.
The enthalpy is accurately predicted using the procedure of Kesler and
Lee for the ideal gas enthalpy, and the SRK-equation for the residual en-
thalpy.
REFERENCE
Lenoir, J . M. and Hipkin, H. G., "Measured Enthalpies of Eight Hydrocar-
bon Fractions," t. Chem. Eng. Data, 18, 1973, pp. 195-202.
Chapter 10
Tuning of EOS-Parameters
THE NEEDS FOR TUNING
Accurate data for the phase behavior of oil and gas mixtures are needed
in, for example, the design of process plants, and in reservoir simulation
studies. Often, experimental PVT-data such as those described in Chapter 3
will be available, but in practice, not very many PVT-measurements are
carried out for a given mixture. It is, therefore, necessary to rely on thermo-
dynamic models when planning the production strategies for a given petro-
leum reservoir. This raises the question of what accuracy can be obtained
using acubic equation of state for phase equilibrium calculations, for exam-
ple, at the conditions where oil and gas are being produced.
Asseen from Chapters 8 and 9, a reasonable accuracy can usually be ob-
tained for saturation points, gas/oil ratios, and phase densities, using the
SRK equation of state, with aproper procedure for characterizing the unde-
fined, heavy hydrocarbons. The only major deviation between measured
and calculated results isfound in the gas/liquid ratio for gas condensa temix-
tures, at pressures slightly below the dew-point pressure. The calculated
amounts of liquids are, in most cases, somewhat larger than those observed
experimentally (see, for example, Figure 8-1). Unless a different thermody-
namic model is used, the only way to improve the agreement between the
measured and the calculated results isto adjust the parameters entering into
the equation of state. This kind of adjustment is called tuning.
PARAMETERS AVAILABLE FOR TUNING
Asdescribed in Chapter 5, the parameters entering into the SRK equation
of state are:
1. The critical temperatures (Te), the critical pressures (Pe), and the acen-
tric factors (w) of each component of the mixture
2. The binary interaction coefficients between each pair of components
(kij)
3. The molar composition
155
For the welI-defined components of the mixture (N2, CO2, Cl, C2, C3, i-
C4, n-C4, i-C5, n-C5, and n-C6), accurate measured data are available for
Te, Pe, and w, and it is usually not considered justified to usethese as tuning
parameters. Te, Pe, and w of the C7 +-fractions are usualIy determined from
empirical correlations (see Chapter 7). Each C7+-fraction consists of many
different components, each having different critical temperatures, critical
pressures, and acentric factors. It is not obvious that the empirical correla-
tions represent the best possible choice of mixture properties. Therefore, at
least in principle it seems justified to useTe, Pe, and w of the C7 +-fraction as
tuning parameters.
The binary interaction coefficients, kij, are introduced to account for the
effect of polar forces in the interactions between two components. Usually,
kij for hydrocarbon-hydrocarbon interactions are considered equal to zero,
or at least clase to zero. Cubic equations of state were originally not in-
tended for use on components deviating from sphericity as much as the
heaviest components of a petroleum reservoir Huid do. One possible way to
compensate for the non-sphericity of the heavy hydrocarbons could be to
introduce binary interaction coefficients differing from zero for the interac-
tions including a heavy hydrocarbon.
Coats and Smart (1986) present aprocedure where, normally, Te, Pe, and
w for the heaviest fraction, and/or kij between methane and the heaviest
fraction, are adjusted using non-linear regression to laboratory data. The
data might include saturation pressures, densities, K-values, constant vol-
ume expansions, differentialliberations, etc. While the desired stated objec-
tives may be obtained using this approach, it is shown in the Dangers of
Tuning section that highly undesirable effects may result for properties not
included in the objective function for the regression.
The molar composition of the petroleum mixtures may be determined as
described in Chapter 2. Inaccuracies in the analytical techniques will, of
course, influence the results of the phase equilibrium calculations. It is,
therefore, considered of interest to analyze the effect, on the results of the
phase equilibrium calculations, of a small variation in the least accurately
determined molar component concentrations.
EXAMPLES ON THE DANGERS OF TUNING
The experimentalliquid dropout curve of aconstant volume depletion ex-
periment is presented in Figure 10-1 for the gas condensate for which the
molar composition is given in Table 10-1. It is a heavy gas condensate,
which, at pressures >200 bar and temperature around 136.1C (the tem-
perature of the depletion experiment), has almost dentical gas and liquid
phases. Thus, it exhibits near critical behavior. The problems of handling
Tuning 01EOS-Parameters 157
35.0 j I
::l
O"
(
\
30.0
'<,
........ ~
::l
~ 25.0
+'
e;
o
e,
~ 20.0
Q)
o
4-
o
15.0
Q)
E
::l
o
t>
"O 10.0
~---
<,
-,
" \
<, \
<,
',. \
" \
\ \
\ \
...
\ \
\i
\.
,
/'
---.---!-

./
./
~
/ .............
........
"' ....
....
400.0
(
Exper i ment
St d. Char act er i zat i on
" Tuned" C
7
+- Pr oper t i es
k
ij
~ f or C, - C
7
+
5.0
~
0.0 100.0 200.0
Pr essur e ( bar )
300.0
Figure 10-1. Constant volume depletion experiment tor gas condensate (molar
composition in Table 10-1).
Table 10-1
Molar Composition of North Sea Gas Condensate
(
(
(
Density
Molecular (g/cm
3
) at
N2
0.42
e02 2.98
e, 66.36
e2
8.44
e3 5.12
ic, 1.04
e, 2.35
te, 0.84
n-es 1.12
e6 1.36
e7 2.14 93 .743
ea 2.20 107 .753
eg 1.43
120 .776
e1Q 0.60 134 .790
el 1 0.48
147 .795
158 Propertes o] Ols and Natural Gases
Table 10-1
Continued
Density
Molecular (g/cm
3
) at
C'2
0.40 161 .807
C'3
0.39 174 .823
C'4
0.34 188 .831
C'5
0.30 202 .836
C'6
0.21 217 .843
C
'7
0.23 234 .840
C'B
0.16 251 .845
C
'9
0.14 259 .856
C20+ 0.95 397 .890
Sauree: Pedersen el al. (1988a).
two almost identical phases are reflected in the deviation, as presented in
Figure 10-1, between the measured liquid dropout curve and the one calcu-
lated using the characterization procedure of Pedersen et al., described in
Chapter 7. The data used to represent the C7+-fractions are given in Table
10-2. In order to improve the agreement between the measured and the cal-
culated liquid dropout curve, tuning of Te, Pe, and W of the C7+-components
was attempted. Binary interaction coefficients of zero were maintained for
the hydrocarbon-hydrocarbon interactions. By employing for the C7 +-com-
ponents the tuned critical temperatures, critical pressures and acentric fac-
tors of Table 10-3, the agreement between the measured'and the calculated
liquid dropout curves was very-much improved, as compared with the re-
sults of using the parameters of Table 10-2. The liquid dropout curve calcu-
lated using the tuned parameters is also shown in Figure 10-1. Figure 10-2
shows measured and calculated results for the methane concentration at
various depletion levels. Figure 10-3portrays measured and calculated corn-
pressibility factors at the conditions of the depletion experimento Figures
10-2 and 10-3show that the tuned predictions are slightly inferior to the re-
sults found using the characterization procedure of Pedersen et al. This
means that the tuning has only served to improve the results for the liquid
volume which was used in the tuning process, and less accurate results are
obtained for other thermodynamic properties.
Figure 10-1 also shows the liquid dropout curve resulting from atuning of
the binary interaction coefficients for interactions between methane and the
C
7
+-components. It was found that the liquid dropout curve could be much
improved when, maintaining the parameters of Table 10-2, the following
non-zero binary interaction coefficients were introduced:
(~~
Tunng o] EOS-Parameters 159;;:JY -~~<'l()"
Q: ( \'0
-c -- ('
Th~1~ o: . ;
Results of the Characterization of the C7+-Fraction of the Gas condensat~ t;\ ~\~\.I'.ln:C~ .
of Table 10-1 \ _ ,
\. /J ....'
, ",
,
Te Pe Acentric
Component Mole % (K) (Atm) Factor
C7-9 5.77 560.7 30.07 .4327
ClO_'6 2.72 658.3 21.00 .7479
C'7-60
1.48 802.3 12.16 1.2993
Sauree: Pedersen el al. (1988a) .
Table 10-3
Results of Tuned Characterization of the C7+-Fraction of the Gas
Condensate of Table 10-1
Te Pe Acentric
Component Mole % (K) (Atm) Factor
C
7
_
9 5.77 571.0 27.00 .3000
ClO-16 2.72 629.0 21.00 .6000
C,7-60
1.48 782.0 14.50 1.3500
80
G>
ti>
'" .J ::
a.
lb
'" (!l
.,
75
.J ::
-
s
~
o
::E 70
I

Exper i ment
St d. Char act er i zat i on
" Tuned" C
7
+- Pr oper t i es

---- ...
...
...
"- <,
<,
-,
-,
<,
-,
"- ...
<,
<,
<,
<,
<,
-,
65 LI ----~----~----~----~----~----~-- ~ __ ~
O 100 200 400 300
Pressure, bar
Figure 10-2. Experimental and calculated methane concentration correspond-
ing to the constant volume depletion experiment of Figure 10-1.
~1 10
o
-
U
!\l
U.
c-,
- .-
-
.c
'"
1.00
'"
.,
c.
E
o
u
.,
'" !\l
s:
el.
0. 90
'" !\l
o
l
O
~~~.
..... _- - -
l.
I
I
/
I
/
/
1/
/
/
./
/,
t
100 200 300 400
Pressure, bar
Figure 10-3. Measured and calculated gas phase compressibility factors corre-
sponding to the constant volume depletion experiment of Figure 10-1.
CJ - C7~9 :kj = - 0.25
CI - CIO~16 :k;j =0.075
The phase envelopes shown in Figure 10-4 are calculated using the C
7
+_
properties of Table 10-2 and
l. kij"values of zero for all hydrocarbon-hydrocarbon interactions
2. The non-zero k(values indicated above and kj-values of zero for all
other hydrocarbon-hydrocarbon interactions.
The phase envelopes eorresponding to the tuned parameters have a mini-
mum in pressure at approximately 27C. There is no physical evidence for
this behavior, and the example ShOWSthat equation of state parameters de-
termined for one temperature may lead to highly erroneous predietions at
temperatures that are only slightly different.
Only the molecular weight, and the density at 15C and 1atm, have been
measured for the C7 +-fraction of the oil for whieh analytical data are pre-
sented in Table 10-4. Gas (of the eomposition listed in Table 10-5) was grad-
ually added to this oil, and after each addition of gas, the saturation point
pressure at 111.1C was measured. The addition of gas was continued until
the saturation point shifted from a bubble point to a dew point. The addi-
tion of gas was simulated using the characterization procedure of Pedersen
et al. (Chapter 7). Experimental and simulated results are shown graph-
Tuning of EOS-Parameters 161
600.0
500.0
400.0
;:-
'"
:=:
'"
300.0
'-
"
Vl
Vl
Q)
'- e,
200.0
100.0
0.0
-100.0
\
\
\
\
\
\
"
<,
Exp. De~ Poi nt
St d. ehar act er i zat i on
- - - k
ij
~f or el - e
7
+
~ __ - f : : _" ,
-c
""
<::,
100.0 200.0
Temper at ur e " C
Figure 10-4. Calculated phase envelope for the gas condensate mixture for
which the molar composition is given inTable 10-1.
0.0 300.0 400.0
Table 10-4
Molar Composition of the Oll Mixture Mixed with Gas Mixture 01
Table 10-5
Density
Molecular (g/cm
3
) at
Nz 0.31
COz 0.69
C, 47.69
Cz 6.64
Cl 4.59
-C. 0.85
n-C. 1.74
ic; 0.70
c, 0.46
Ce 1.69
C7+ 34.64 234 .8688
Source: Pedersen el al. (lg88a).
cally in Figure 10-5. When the characterization procedure of Pedersen et al.
is used on a C7 +-fraction for which no detailed composition is available, a
logarithmie dependenee according to Equation 7-13 isassumed The molee-
ular weight of the C7 +-fraction of the oil described in Table 10-4 is 234,
Table 10-5
Molar Composition of the Gas Mixture Added to Oil Mixture of
Component
Mole %
N2
0.55
CO2
1.43
C,
81.63
C2
10.12
C3
5.20
-c,
1.06
n-C.
0.01
Source: Pedersen et al. (1988a).
650.0 I
600.0
'- 550.0
-o
.D
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/ .
/ /.
/
/
/;
~
~
?
Q
Q)
'-
~ 500.0
~
'" '-
o-
... '50.0
e
o
o-
e 400.0
o
';;;
'-
B 350.0
m
V1
Exper i ment al Bubbl e Poi nt s

Exper i ment al Dew Poi nt
St d_ Char act er i zat i on
.t.
300.0 Tuned C7+- pr oper ~i es
250. o ! ! I ! I I J I I I I
0.0 .\ .2 .3 .4 .5 .6 .7 .8 .9 1.0
Mol e Fr act i on of Gas Added per Mol e Di l
Figure 10-5. Experimental and calculated saturation points at T =111.1e for
mixtures of the oil of Table 10-4 and the gas of Table 10-5.
which roughly corresponds to that of a Cl7-fraction. The influence of using
a distribution function to represent the C7 +-fraction may be illustrated by
calculating the bubble-point pressure of the mixture of Table 10-4, where
the properties of n-C
l7
are used to represent those of the total C7+-fraction.
In that case, a bubble-point pressure at 111.1C is calculated, which is
34.1 % lower than the measured value. This allows the introduction of a
binary interaction coefficient of 0.145 for the interaction between C and
the C
7
+-fraction having the properties of n-Cl7, in which case a perfect
Tuning o] EOS-Parameters 163
Table 10-6
Molar Composition of the Oil Mixture Used in the Separator Calculation
Reported in Tables 10-8 and 10-9
Density
Molecular (g/cm
3
) at
N2 0.79
CO2 0.17
C, 45.90
C2 5.97
C3 4.37
ic, 0.92
n-C. 2.35
ic, 1.06
n-Cs
1.46
C6 2.31
C7 4.07 91.8 .734
Ca 4.65 104.9 .756
C9
3.40 121.0 .775
CID. 22.59 271.6 .881
Source: Pedersen et al. (1988a).
match of the measured bubble point of the original oil isachieved. However,
as is shown in Figure 10-5, the estimated saturation points of the
oil +gas mixtures are, in this case, much different from the experimental
values. This means that the kij-values which are appropriate for the original
oil givesaturation points for the ol +gas mixtures that deviate considerably
from the measured results.
As explained in Chapter 2, the composition of a reservoir oil mixture is
usually determined from the compositions of the gas and liquid phases re-
sulting from a single stage flash to standard conditions (i.e., 15C and 1
atm). This is also the case for the oil composition given in Table 10-6. The
final composition is, of course, dependent on the assumed ratio for the
amounts of gas and liquido The influence, on the calculated phase behavior,
of assuming an erroneous gas/liquid ratio, has been investigated for the oil
mixture of Table 10-6. Experimental and calculated bubble-point pressures
at T =74.5C are shown in Table 10-7. The calculated bubble-point pres-
sure using the correct molar composition agrees very well with the measured
bubble point, whereas the bubble-point pressure calculated using the erro-
neous composition deviates by 8.70/0 from the experimentally determined
bubble point. Hydrocarbon-hydrocarbon k(values, which depend slightly
on the molecular weights (MW) of the constituent components in the fol-
lowing manner, were introduced:
Table 10-7
Measured and Calculated Bubble-Point Pressures at T =74.5C for the
Oil Mixture for Data Given in Table 10-6
Pressure (bar) % Dev.
Experimental bubble
point
235.0 -
Calculated bubble
point based on
233.6 - 0.6
measured composition
Calculated bubble
point based on
214.5 - 8.7
composition with
erroneous GOR
Calculated bubble
points based on
composition with
235.0 0.0
erroneous GOR using
k;-values different
from zero
% Dev. =100 x (Cale. - Exp.)/Exp.
k=0.00145 MW
MW
k=0.00145 MW
MW
for MW>MW
(10-1)
for MW>MW
(10-2)
With these binary interaction coefficients a perfect agreement with the
measured bubble point is obtained (seeTable 10-7). Results of a three-stage
separation process are available for the oil mixture of Table 10-6. The results
are summarized inTables 10-8 and 10-9. Simulated results are shown for the
correct composition, and for the composition with a gas/liquid ratio 5%
lower than what was measured. Results are shown using hydrocarbon-hy-
drocarbon interaction coefficients equal to zero, and determined from
Equations 10-1 and 10-2. The results in Tables 10-8 and 10-9 indicate that
the erroneous analytical data are not accounted for by tuning to ameasured
bubble point, using k-values different from zero, for hydrocarbon-hydro-
carbon interactions.
The examples just given show that equation of state parameters tuned to
one specific property cannot be expected to give reliable predictions for
other therrnodynamc properties. If the analytical data are inaccurate, tun-
ing to the properties reported for these erroneous compositions may lead to
highly uncertain predictions.
(O
o
...
:> C'!
C\J (')
e
Q)
v al ci
o
e ~ ~ ~
-
'R I I I
.
o
.
&. i!-
E ~
o
u
a: C\J al C\J
I/J
O r-: en C\J
~
CJ
r-, C\J ~
o
Q)
e
e
...
e
w
:8
o v :> al
E lO Q) r-,
,.j r-:
~
.!:! e ~ ~
~ 'S- I I I
a:
ti
O lO
CJ :;;
s:
u
j
a: al C'!
10
U 'O
O (') C\J N
'jj
C!J J :'-
(') ~
e
o
:>
E~ '"
CXl
""
Q)
cri
,.j
e
C\J
o o
. .o.
'S-
I I I
E
a: o
u
CJ 'C
'Q)
( ) ...
j ~ a:
~
LO
U I/J
O r-, ..t ,.j
lO
CJ CXl (') ~
Q)
::E
'O
~a:
o o 10
~O
ci r-:
,.j
QC!J
al (') ~
Q)
::E
6 C'!
10 C!
z,
::;:
cri LO
1-
CXl ~
~ al LO C!
e
a:i
N ~
10 C\J
a..
ti
- Q)
lO el
(') . lO
~ C\J
lO_
g.!J l
!J l
Q)
:c
t!!!
'O
Q)
..
:1
-><
s
s
..
o
-1/1
o
(1);
, ni
a:: ...
Q) =
- o
.0--
t!!! :
o
..
o
-
1i
~
..
o
-
ni
..
ni
o.
Q)
!J l
-o
1/1
::
:1
1/1
Q)
a::
M
Q;
C
'"
s:
o
fE
<ti
Ol
Ol
~
6 .
: o~~
~~ffi
~~~
...e Cii
( l ) . - _
a. U) Ol
U) '" c:
"'-oOl
OJ OlU)
c: ~
E~~
"E~~
(tJ > ..
-o Ol Ol
c: O u
'" ~
5'if.g
(J)
Tuning 01 EOS-Parameters
(O
O
...
:> CXl al O
Q) ..t
e
LO (')
c:
I I I
a
'S-
o
.
jj.
o
o.~
E
o CXl C\J
<O u
al
o o
I/J
(') ~ o
~ ~ ~ ~
o
Q)
e
o
...
c:
W
o
:> (') al o
E
; Q)
10
,.j ,.j
N e
o .. I I I
~
!
'S-
o
o
ID
e
() lO
j
s:
U
U
10 al C\J
al
o o 10
'O (') ~ o
'jj ~ ~ ~
e
o
:>
-
r-,
'"
10
E .~
Q) ,.j
e
C\J C\J
o o.
'S-
I I I
~ E
o o
IDU
''0
~ Q)
lO .
C\J r-, LO
U ~
al
LO ~ 10
I/J (') ~ C!
lO
Q) ~ ~ ~
::E
'O
~ al
""
LO
~ o CXl V o
I/J ID (') ~
'"':
lO
~ ~ ~
Q)
::E
6 LO C\J 10
~ v
::;:
cri
1-
t-, CXl
...
o al LO
lO
N
e
LO CXl
(') 10 C\J
a.. C\J
ti
- Q)
f el
~ C\J (')
lO lO
0. -
Q)!J l
!J l 1,
Q)
:c
t!!!
'O
Q)
..
:1
-
><
:
5
..
s
1/1
..
o
-
u
ni
m u. .
o e
...o
Q,):;:
- ni
.o E
t!!! .
~
s
..
o
-
-
1/1
~
..
o
-
C\I
..
ni
C.
Q)
en
-o
1/1
- :;
1/1
Q)
a::
165
M
Q;
C
'"
s:
o
o
U)
<ti
Ol
Ol
~
-o
:;
g
a
E
-o
~~~
~~~
U) Ol co
(fi:O~
a.~ __
"2ccij
::J .- _
cr U) Ol
~~s
; 2 I D Vl
'" c: ~ :;:: (l)
'" Ol -o
a. -o Ol
Ol . "-
ln a; "
o ~
~" o
(J)
Gani and Fredenslund (1987) have developed a two-part tuning proce-
dure that overcomes some of these problems:
l. A sensitivity analysis, to select the parameters to tune, and afeasibility
study, to check whether tuning can/should be done
2. If tuning isfeasible, it is accomplished by regression; and evaluation of
the results takes place
While this tuning policy often yields reasonable results, and often will
warn the user not to tune, it is also not completely safe. The dashed curve in
Figure 10-4 is drawn with kij-values for C1-C7+ computed from this proce-
dure. The reason for the incorrect rise of the bubble-point branch isthat the
sensitivity of the low-temperature bubble points to the adjusted kifvalues
was not taken into account.
THE MOLECULAR WEIGHT OF THE PLUS-FRACTION AS A
TUNING PARAMETER
Asexplained in Chapter 2, the compositional analysis for oil and gas rnix-
tures is most often given in weight units. Conversion to molar compositions
requires knowledge of the molecular weights of each component and carbon
number fraction. The experimental inaccuracy in determining the molecu-
lar weights of the plus-fractions is of the order 5-10%. Deviations in that
order of magnitude can considerably influence acalculated dew-point pres-
sure of agas condensate mixture. An obvious application of this fact isto use
the molecular weight of the plus fraction as an adjustabl parameter. Ad-
justments within the experimental uncertainty can, e.g., be used to match a
measured saturation point, as suggested by Pedersen et al. (1988b).
Table 10-10 presents experimental results for the dew-point pressures of 10
gas condensate mixtures. Also shown are the dew-point pressures calculated
using the characterization procedure of Pedersen et al. as described in Chap-
ter 7. The average absolute deviation between the experimental and the cal-
culated dew-point pressures is 4.9%. To improve the dew-point calcula-
tions, Pedersen et al. attempted to treat the molecular weight of the plus
fraction as an adjustable parameter, maintaining the measured weight com-
position. Table 10-10 also shows the adjustments needed in the molecular
weights of the plus-fraction to get agreement with the measured dew point.
The maximum adjustment required is 12.1% (gas condensate 10). The mo-
lar compositions of this mixture before and after 'ths adjustment are shown
in Tables 1O-1l and 10-12. Figure 10-6 shows the measured liquid dropout
curve of aconstant mass expansion experiment (described in Chapter 3), on
gas condensate 1, in comparison with those calculated using:
Tuning 01EOS-Parameters 167
Table 10-10
Experimental and Calculated Dew Point Results for 10 North Sea Gas
Condensate Mixtures
Temp. Exp. Dew Pt. Calc. Dew Pt. Adjust- Plus-
No. (OC) (bar) (bar) % Dev. ment of MW fraction
1 96.6 282.0 311.6 10.5 - 8.8 e,,.
2 93.8 235.0 235.2 0.1 0.0 e1O+
3 118.9 398.0 377.1 - 5/3 8.3 e1O+
4 119.7 394.0 369.7 - 6.2 9.6 e1O+
5 150.3 381.5 396.3 3.9 -
5.4
e20+
6 154.1 385.5 410.2
6.4
- 9.6 e1O+
7 131.5 367.0 378.7 3.0 - 9.0 e20+
8 129.0 464.0 454.2 - 2.1 5.6 e20+
9 136.1 386.4 375.6 - 2.8
8.4
e20+
10 148.9 542.5 493.5 - 9.0 12.1 e20+
%AAD 4.9 7.7
% BIAS - 0.2 1.1
% adjustrnent needed in Ihe molecular weighl 01Ihe plus Iraction lo get agreement wilh the measured
dew point maintaining Ihe measured weight composition.
% dev. delined as in Table 10-7.
% AAD =% average absolute deviation.
% BIAS =% average deviation.
Source: Thomassen et al. (1986).
Table 10-11
Measured Molar Composition of Gas Condensate 10 of Table 10-10
Densily
Molecular (g/cm') al
Componenl Mole % Weighl 15C, 1 atrn
N2 .39
e02
3.47
e, 80.17
e2 6.28
e3 2.75
-c,
.43
n-C, .88
te, .31
n-es .35
C6 .54
C7 .72 96 .743
ea .88 106 .761
eg .55 118 .779
e10 .33 132 .787
C
"
.24 149 .788
C'2 .20 163 .799
Table 10-11
Continued
Density
Molecular (g/cm
3
) at
C'3 .21 175 .820
C
'4
.18 194 .831
C,S .15 203 .833
C'6 .11 217 .838
C17 .12 235 .836
C,S .09 247 .842
C'9 .08 255 .852
C2D+ .57 396 .867
Source: Thomassen el al. (1986).
Table 10-12
Molar Composltion of Gas Condensate 10 of Table 10-10 After an
Adjustment of the Molecular 'Weight of the C20 +-Fraction by 12.1%
- Density
Molecular (g/cm
3
) at
N2 .39
CO2
3.47
C, 80.22
C2 6.28
C3 2.75
i-C4
.43
n-C4 .88
ic, .31
c, .35
C6 .54
C7 .72 96 .743
Cs .88 106 .761
C9 .55 118 .779
ClO .33 132 .787
Cll .24 149 .788
C'2 .20 163 .799
C,3 .21 175 .820
C
'4
.18 194 .831
C,S .15 203 .833
C
'6
.11 217 .838
C17 .12 235 .836
C,S .09 247 .842
C
'9
.08 255 .852
C20+ .51 444 .867
Source: Thomassen el al. (1986).
Tuning 01 EOS-Parameters 169
8.0 I i
::>
o-
Exper i ment al Resul t s
Cal cul at ed Based on Measur ed MW
Cal cul at ed Based on Adj ust ed MW
(j)
5 6.0
o
:>
..,
e
o
e,
"
(j)
'"
4-
o
U
---- .
/~~"
. .." -,
-,
"-,
.\ \
\
\
\
\
\
\ I
..\ ,\ I , ,
0.0
0.0
4.0
~ ,.0
o
r>
100.0 200.0
Pr essur e ( bar )
300.0 400.0
Figure 10-6. Liquid dropout curve (constant mass expansion) lor the gas con-
densate mixture of Tables 10-1 and 10-12.
l. The measured molecular weight of the plus-fraction and
2. The molecular weight of the plus-fraction adjusted to get agreement
with the measured dew point.
Figure 10-7 shows measured and calculated results for the compressibi!ity
factor of the gas phase liberated during the constant volume depletion ex-
periment. The results indicate that the adjustment of 12.1% in the molecu-
lar weight of the plus-fraction has very little influence on the calculated gas
phase density.
Pedersen et al. find that an accurate molecular weight of the plus-fraction
is less important for the bubble point of an oi! mixture than for the dew
point of a gas condensate mixture. Adjustment of the molecular weights is,
therefore, a less effective tool for oil mixtures than it is for gas condensate
mixtures. On the other hand, there is also less need for improvements in the
calculations on ol mixtures, because accurate results can in most cases be
obtained based on the measured compositions .
RECOMMENDATIONS REGARDING TUNING
Agreement between the results of PVT-experiments, and those obtained
with a cubic equation of state, may be improved by adjustment of the criti-
'-
o
+-1 1.2
u
-o
u;
- - --==----- -
e-,
+-'
-
/
'1
'1
'1
'1
'1
'"
'1
'1
:
'1
'1
jI
'1
>-
Experimental Results
--- Calculated Based on Adjusted MI<
- - - CaI cu I ated Based on Measured MI<
D
~ 1.1
QJ
'- el.
5
u
~
-o
.c;
o.. 1.0
'/
'/
'/
'/
.'/
/
. /-
/
V>
-c
co
'~
0.9
0.0 400.0 500.0 600.0 200.0 300.0
Pressure (bar)
Figure 10-7. Gas phase compressibility factor of the gas condensate mixture of
Tables 10-11 and 10-12.
100.0
cal temperature, the critical pressure, the acentric factor, or the binary in-
teraction coeHicients. It is, however, unlikely that parameters tuned to one
specific property will provide accurate results for other properties. Pararne-
ters determined in a limited pressure and temperature range may lead to
highly inaccurate results when considered at pressures and temperatures not
included when performing the parameter estimation.
Rather than adjusting the equation of state parameters, it is recom-
mended treating the molecular weight of the plus-fraction as an adjustable
parameter. The experimental uncertainty on the molecular weight of the
plus-fraction is of the order 5-10 %, and it seems, therefore, justified to ad-
just the molecular weight of the plus fraction (e.g., against ameasured satu-
ration point) within these limits.
REFERENCES
Coats, K. H. and Smart, G. T., "Application of a Regression-Based EOS
PVT Program to Laboratory Data," SPE Reserooir Engineering, May
1986, pp. 297-298.
Gani, R. and Fredenslund, Aa., "Therrnodynarnics of Petroleum Mixtures
Containing Heavy Hydrocarbons: An Expert Tuning Systern,' Ind. Eng.
Chem . Research, 26, 1987, pp. 1304-1312.
/e:"';' ''\()~
./a:! ~~
f
'<' - \
Tuning 01 EOS-Parameters 17 o: \1
. 1-- ~\al\Oll:~~~
Pedersen, K. S., Thomassen, P, and Fredenslund, Aa., "On the Dangers of\") \ PII _ ~
Tuning Equations of State Pararneters," Chem. Eng. Sci., 43, 1988a, pp. \:,.;\... , / h7A
269-278. .~
, ,...
Pedersen, K. S., Thomassen, P, and Fredenslund, Aa., "Characterization of ' "
Gas Condensate Mixtures,' paper presented at the 1988 AIChE Spring
National Meeting, New Orleans, LA, March 6-10, 1988b.
Thomassen, P, Pedersen, K, S., and Fredenslund, Aa., "Adjustment of C
7
+-
Molecular Weights in the Characterization of Petroleum Mixtures Con-
taining Heavy Hydrocarbons,' SEP Report 8617, Instituttet for Kemi-
teknik, The Technical University of Denrnark, 1986.
Chapter 11
Viscosity
DEFINITION OF VISCOSITY
An external shear stress applied to a portion of a fluid will introduce a
movement of the molecules of the affected part of the fluid in the direction
of the applied shear stress. The moving molecules will interact with the
neighboring molecules. These will start moving too, but with alower veloc-
ity than that of the molecules exposed to the stress. The dynamic vscosity, 1/,
of aNewtonian fluid is defined as the following rato for the flowing fluid:
1/ = - Txy/(aVx/ay)
(11-1)
where Txy=shear stress
V
x
=velocity of the fluid in the x-direction (the direction of the
applied stress)
avx/ay =the gradient of V
x
in the y-direction (perpendicular to the
x-direction) .
KINETIC GAS THEORY
For adilute gas, one may derive an approximate analytical expression for
the viscosity, 1/. It is:
1/ =1/3 n v MW L
(11-2)
where n =number of molecules per unit volume
v =' average molecular speed
MW = molecular weight
L =mean free path between two molecules.
It can beshown (seee.g., Hirschfelder et al., 1954) that the average speed, v,
is proportional to (RT/MW)1/2 and the mean free path.to l/(na
2
), where a is
172
Viscosty 173
the molecular hard sphere diameter. The following dependence of 1/ on the
temperature and the molecular weight may then be derived:
1/ =constant (T1I2MW1I2/a
2
) (11-3)
a3 is often associated with the critical molar volume, Ve' If it is further as-
sumed that Veis proportonal to RTe/Pe, the following expression may be
obtained for 1/ at the critical point:
n: =constant (MW1I2p~/3/nI6) (11-4)
Though kinetic gas theory is not applicable to the near critical region, this
expression plays an important role in vscosity calculations. It is convenient
to introduce the viscosity reducing parameter, ~:
~=T~/6/MW1I2 p~/3
~
(11-5)
VISCOSITY CORRELATIONS
Several methods have been suggested for estimating the viscosity of hydro-
carbon mixtures. In the simulation of processes related to oil and gas produc-
tion, vscosity correlations are needed that are applicable to a wide range of
hydrocarbon mixtures and process conditons. Methods limited to narrow
ranges of compositon and/or temperature and pressure will not be covered
here. Only calculaton procedures that can be used for both gas and liquid
phases, giving consistent results for each, are discussed.
THE VISCOSITY CORRELATION OF LOHRENZ ET AL.
The viscosity correlaton that has the most widespread use in flow models
for petroleum mixtures isprobably the correlation of J ossi et al. (1962) in the
form suggested by Lohrenz et al. (1964). Gas and liquid viscosites are re-
lated to a fourth-degree polynomial in the reduced density, o, =p/ Pe:
[(1/ - 1/')~ +10-
4
]1/4 =al +a2Pr +a3Pr
2
+a4Pr
3
+a5Pr
4
(11-6)
where al =0.10230
a2=0.023364
a3 =0.058533
a4= - 0.040758
a5=0.0093324
174
Properties oj Oils and Natural Gases
TI' is the low-pressure gas mixture viscosity. t is the viscosity-reducing pa-
rameter, which for a mixture is given by the following expression:
[
N ll/6[ N l- 1/2[ N l- 2/3
t = .1: zTci .1: zMW1: zPe
1=1 1=1 1= 1
(11-7)
where N is the number of components in the mixture, and Zthe mole frac-
tion of component i. Pure components have a well-defined Pe, but it is not
obvious how to determine the critical density of a complex hydrocarbon
mixture. Lohrenz et al. (1964) suggest calculating the critical density of a
petroleum fluid as follows:
p, =(Ve)- 1 =(t (z.Vci)+ZC7+ VeC7+)- 1
1=1
i*C7-t
(11-8)
where the critical molar volume (in ft
3
/lb mole) of the C7+-fraction isfound
from the expression:
V,C7+=21.573 +0.015122 MWC7+- 27.656 X SGC7+
+0.070615 MWC7+x SGC7+
In this expression, MW isthe molecular weight, SG the specific gravity, and
the sub-index C
7
+indicates that the mentioned property isan average value
for the hydrocarbons with 7 and more carbon atoms.
The dilute gas mixture viscosity TI*isdetermined as f.ollows (Herning and
Zippener, 1936): .
N N
r* =1: zTI*MWI/2/1: ZMWI/2
=1 =1
(11-10)
For the dilute gas viscosity of the individual components, TI~the following \
expressions are used (Stiel and Thodos, 1961):
TI*=34 X 10-
5
~ T0
94
t n
for T; <1.5
(11-11)
TI*=17.78 X 10-
5
~ (4.58Tr - 1.67)5/8
~
for T
r
>1.5 (11-12)
where ~is given by the expression in Equation 11-5.
(11-9)
Viscosity 175
By expressing the viscosity as a function of the density, the calculated vis-
cosity becomes very sensitive to the results obtained for the density. Espe-
cially for high viscous fluids this may lead to severe errors for the calculated
viscosity (see the examples in the Comparison with Experimental Results
section) .
CALCULATING VISCOSITY USING THE CORRESPONDING
STATE S THEORY
The viscosity may al terna tivel y becalculated using amodified form of the
corresponding states method. The starting point-is the properties at the criti-
cal point (e.g., the critical viscosity, Tle,the critical density, Pe, the critical
pressure, Pe, and the critical temperature, Te)' Away from the critical point,
the properties may be expressed in terms of reduced properties (TI"p" P" T"
etc.), which are defined as the properties at the actual conditions, divided
by those at the critical point. A group of substances obey the corresponding
states principIe with respect to viscosity, if the functional dependence of Tlr
on, e.g., o, and T, is the same for all substances within the group.
Tlr(P, T) =f(p" Tr) (11-13)
In that case, compreheflsive viscosity data are only needed for one of the
components of the group. That component is then used as reference sub-
stance (o), and the viscosity of another component (x) within the group can
easily be calculated. In general, the critical viscosity (TI e) will not be known.
The inverse of tfor which the expression isgiven in Equation 11-5, may be
used as an approximate value. When the corresponding states principle is
expressed in terms of p" and not P" it isconvenient to express ~in terms of
Ve, instead of Pe' The final expression for the reduced viscosity (Tlr)then be-
comes:
Tlr(P, T) =TI(p, T) (T
e
)-1/2 (V
e
)2/3 (MW)-1/2 (11-14)
The viscosity of component x, at the temperature T, and apressure where
it has the density P, is then found from the expression:
Tlr(P, T) =(Te,/Teo)1I2 (Vex/Veo)-2/3 (MW,/MWo)I/2 ro(Po, To) (11-15)
where Po =P Peol Pe,
T, =T Teo/Te,
Tlo=viscosity of the reference substance at temperature T, and den-
sity Po
The corresponding states principie, in this simple form, may be used suc-
cessfully for, e.g., mixtures of hydracarbons of approximately the same
chemical structure and molecular weight. Oil and gas mixtures consist of a
large number of components, including paraffinic, naphthenic, and aro-
matic components. For oil mixtures, the molecular weights may range fram
16(methane), to about 1100 (Cso). It isnot to be expected that methane and
Cso should both belong to a group where the simple corresponding states
principie applies. Very comprehensive experimental viscosity data exist for
methane, and methane is therefore a natural choice as the reference sub-
stance. Toapply the corresponding states principie to the remaining compo-
nents found in oil and gas mixtures, some modifications must be intraduced,
as compared with Equation 11-15. Ely and Hanley (1981) have suggested
expressing the deviations from the simple corresponding states principie in
terms of two shape factors, e(T" V" w), and (T"V" w), which enter into
Equation 11-16 as follows:
1/x(p, T) =(eT
ex
/T
co
)1I2 (VexlVco)-213(MWx/MWo)1I2
1/o(po<l>,Tole) (11-16)
Vr isthe reduced molar volume and w isthe acentric factor. Vr (and thereby
p) is itself found by a corresponding states model, and the determination of
e and <1> therefore involves an iterative procedure. Touse the corresponding
states principie on a mixture, it is necessary to represent the mixture as a
fluid consisting of one hypothetical pure component with agiven Te, Pe, and
MW. Ely and Hanley adapted the mixing rules of Mo and Gubbins (1976)
for Teand P, (the expression for Teis given in Equation 11-27, and the ex-
pression for P, in Equation 11-30). The average molecular weight isformed
from considerations regarding the radial distribution function for a fluid in
non-equilibrium due to the presence of ashear. Baltatu (1982 and 1984) ex-
tended the pracedure of Ely and Hanley to cover boiling point fractions of
petraleum mixtures also. The critical properties and the molecular weights
of these fractions are estimated using the method suggested by Riazi and
Daubert (1980).
Pedersen et al. (1984 and 1987) presented acorresponding states model for
viscosity calcuIations where the reduced viscosity isexpressed in terms of the
reduced pressure, P" and the reduced temperature, Tr:
n, =f(P" Tr) (11-17)
where the following expression is used for the reduced viscosity:
1/r=1/~=1/1(Te-1I6 p~13MWII2) (11-18)
Viscosty 177
The deviation from the simple corresponding states principIe isexpressed
in terms of a ratationaI coupling coefficient, Q'(Tham and Gubbins, 1970).
The final expression for the viscosity of a given mixture is the following:
1/m.(P,T) = (Te,mi.lTeo)-116(Pe,mixIPeo)213(MWmix/MWo)1I2
Q'mix/Q'o1/o(Po, To) (11-19)
P PeoQ'o
Po=P Q"
c,mix rrux
(11-20)
T TeoQ'o
T, =T mixQ'mix
e,
(11-21)
The critical temperature and the critical molar volume for unlike pairs of
molecuIes (i and j) are found using the following formulas:
"
Tcij =(TciTej)1I2 (11-22)
V .. =!(V )13+V )13)3
el) 8 el eJ
(11-23)
The critical molar volume of component i may be related to the critical tem-
perature and the critical pressure as follows:
Vci =RZciTe;lP ci (11-24)
where Zci is the compressibility factor of component i at the criticaI point.
Assuming that Z is the same for all components, Equation 11-23 may be
rewritten:
1 ((T )113 (T) 113)3
Vcij =- constant ~ + ~
8 r, P
el
(11-25)
The critical temperature of a mixture isfound fram the following expres-
sion:
1: 1: ZiZj Tcij Vcf
i j
Te,mix= 1: 1: Z;ZjV
cij
i j
(11-26)
where Zand Zj are mole fractions of components i and j, respectively. Using
Equations 11-22 and 11-25, Equation 11-26 may be rewritten:
"" [(T)113 (T )113
p
T . ='r' 'T' ZiZj P:: + P:; j [Tci
T
Cj]112
C,OlIX
1: 1: z.z [(Tel)113+(Te
j
)1131
3
I J Pel Pej
(11-27)
For the crtical pressure of amxture, Pe,mithe followng relation is used:
Pe,mix=constant Te,mi)Ve,mix
(11-28)
where the constant equals the one enterng nto Equaton 11-25 and Ve,mixis
found as follows:
Ve,mix=1: 1: ZiZjVeij
i j
(11-29)
From Equations 11-27, 11-28, and 11-29, the followng expresson may be
derved for Pe,mix:
"" [(T)113 (T )
113
1
3
P . =8 'r' 'T' ZiZj ~ + ~ [Tci
T
ej]112
c.nux (~ ~ [(!::)", +(!;)"']'l'
(11-30)
The mxing rules of Equations 11-27 and 11-30 are those recommended by
Mo and Gubbns (1976).
The mixture molecular weght is found as follows:
MWmix=1.304 x 10- 4(MW~;303- MW~303) +MWn (11-31)
where MW
w
and MWn are the weght average and number average molecu-
lar weghts, respectively:
N N
MWw =1:ZiMWr/1: ZiMWi
i~ 1 i~ 1
(11-32)
N
MWn =1:ZiMWi
i~ 1
(11-33)
ti
"i
i
~f
:1.
1,
~
.i
')
l~!
Viscosity 179
Ths mixng rule is derved emprcally on the basis of available vscosty
data, and assigns (as one may expect) to the heaver components arelatvely
large nfluence on the mixture vscosty. The o-parameter of the mixture is
found from the expresson:
amix=1.000 +7.378 x 1O-3p;.847 MWmi~S173 (11-34)
and the o-parameter of the reference substance (methane) as follows:
a
o
=1.000 +0.031 p;847 (11-35)
The constants and exponents in Equatons 11-34 and 11-35 are estimated on
the bass of experimental viscosity data. In Equations 11-34 and 11-35:
p; =Po(T TeofTe,mix,P Peo/P e,mix)1 Peo (11-36)
By expressing the viscosity in terms of the reduced pressure, nstead of the
reduced densty asin Equaton 11-13, it ispossible to perform adirect calcu-
laton of the viscosty. The calculations become much smpler, compared
wth those of Ely and Hanley (1981), where the densty is calculated usng
an terative procedure. The use of methane as the reference substance pres-
ents problems when methane is in a solid form in its reference state. Peder-
sen and Fredenslund (1987) have suggested a procedure for mproving the
viscosity predictions in those cases. The reference vscosty correlation is
based on the methane viscosity model of Hanley et al. (1975):
1)'(p,T) =1)o(T)+1)l(T)p +~1)'(p,T) (11-37)
where 1)0' 1)1,and ~1)' are functions defined in the above reference. The
methane densty isfound using the BWR-equation in the form suggested by
McCarty (1974). In the dense liqud regon ths expression is mainly gov-
erned by the term ~1)' (p, T):
~1)' (p,T) =expfj +i4/T) [exp [p01(h +i~/2)
+8pOS(is+~+;2)] - 1.0]
(11-38)
where in the work of Hanley (1975), the coeffcients j-h have the followng
values:
180
t = - 10.3506
iz =17.5716
b = - 3019.39
j4 =188.730
s =0.0429036
j6 =145.290
h =6127.68
f) is given by:
f) =(p - Pc)/Pc
(11-39)
Pedersen and Fredenslund (1987) have estimated an additional set of coeffi-
cients of Equation 11-38 using viscosity data measured at reduced tempera-
tures below 0.4, i.e., corresponding to a methane reference state where
methane is in a solid formo The data include viscosities of oil mixtures and
oil fractions (Baltatu, 1984). The estimation gives the following results:
~1)"(p,T) =expk +k4/T)[exp[po+2 +;~2)
+OP"(k, +~+~:)]- 10].
(11-40)
with k = - 9.74602
k2 = 18.0834
k3 = - 4126.66
k, =44.6055
ks =0.976544
~ =81.8134
k7 =15649.9
Continuity between viscosities above and below the freezing point of
methane issecured by introducing ~1)" as a fourth term in Equation 11-37:
1)(p,T) =1)o(T)+1)(T)p +Fl~1)'(p,T) +F2~1)"(p,T)
(11-41)
F =HTAN +1
2
F
2
=1- HTAN
2
-.*:-
/. '\
18;;>-----;
.,
Vscosty
HTAN =exp(~T) - exp( - ~T)
exp(~T) +exp( - ~T)
with ~T =T - TF
TF =freezing point of methane
".-
.-
iIl':.
j
~.
COMPARISON WITH EXPERIMENTAL RESULTS
Table 11-3 gives experimental viscosity results (measured as described in
Chapter 3) for three North Sea oil mixtures whose compositions are given in
Tables 11-1, 4-7, and 11-2. The results are presented graphically in Figures
11-1 through 11-3, in comparison with the results calculated using the
method of Pedersen et al. (1984, 1987) and the method of Lohrenz et al.
(1964). The latter results were obtained using densities calculated from the
,~.
~
[i
1~'
.,
Table 11-1
Molar Composition of the North Sea Oll (Mixture 1) for which Viscosity
Data are Shown in Table 11-3
iil.~;
~
~\,
~
~.
Density
(g/cm
3
) at Molecular
Component Mole % 15C, 1 atm Weight
N2
0.41
CO2
0.44
C, 40.48
C2 7.74
C3 8.20
ic, 1.23
n-C. 4.22
ic, 1.43
n-C, 2.21
C6 2.83
C7 4.13 100 0.7294
Ca 4.31 106 0.7492
C9 3.13 121 0.7697
ClO
2.439 135 0.7861
C"
1.880 148 0.7919
C'2 1.674 161 0.8037
C'3 1.573 175 0.8191
C, 1.207 196 0.8331
C,S 1.232 206 0.8359
C' 6
0.985 224 0.8429
Cl ? 0.977 236 0.8400
c., 0.911 245 0.8458
C' 9
0.585 265 0.8575
C20+ 6.382 453 0.9183
Table 11-2
Molar Composition of the North Sea Oil (Mlxture 1) for which Viscosity
Data are Shown in Table 11-3
Density
(g/cm') at Molecular
Component Mole % 15C, 1 atm Weight
Nz 0.33
COz 0.19
C, 35.42
Cz 3.36
C3 0.90
-C, 0.69
n-C, 0.26
ic, 0.26
n-Cs 0.14
C6 0.72
C,. 57.73 255 0.9165
Table 11-3
Measured Viscosities of Oil Mixtures for which Molar Compositions are
Given in Tables 11-1, 4-7, and 11-2
(Below the Saturation Point Pressure, the Results for the Viscosity Apply to
the Liquid Phase)
Mixture 1 Mixture 2 Mixture 3
at T =97.8C at T =93.3C atT=71.1C
Pressure Viscosity Pressure Viscosity Pressure Viscosity
(bar) (cP) (bar) (cP) ~bar) (cP)
401.9 0.359 389.8 0.469 345.7 2.64
376.8 0.352 347.7 0.447 276.7 2.44
325.7 0.340 302.1 0.425 242.3 2.34
275.3 0.322 285.9 0.413 207.8 2.24
251.3 0.316 275.6 0.406 *158.9 2.10
223.8 0.306 *274.5 0.404 145.5 2.21
*203.5 0.299 260.9 0.424 125.1 2.41
180.8 0.316 225.1 0.484 104.4 2.65
151.6 0.360 143.2 0.651 83.7 2.93
126.8 0.404 107.9 0.761 63.0 3.29
101.1 0.448 64.7 0.911 42.4 3.70
77.1 0.512 35.8 1.051 21.7 4.23
58.2 0.570 7.9 4.82
26.6 0.715
Experimental Saturation Points
1.75
1.5
1.25
~
.::.
1.0
>,
+'
'"
.75
o
~
>
.5
.25
0.0
0
.
0
Viscosity 183
Exper i ment al r esul t s
Figure 11-1. Measured and calculated viscosities at T =92.8C of the liquid
phase of oil mixture 1for which the composition is given in Table 11-1. The mea-
sured data are tabulated in Table 11-3.
3.0
2.5
1.0
e,
.::.
c-,
+'
'"
1.5
o
u
'"
>
1.0
.5
0.0
0.0
\
\
\
\
\
'"
" --
" --
<, "-._----:-
Peder sen et al .
- - - - - Lohr enz et al . usi ng SRK densi t y
- - - - - - Lohr enz et al . usi ng Penel oux densi t y
...-e--e-----e----e
--------------
100.0 100.0 300.0
Pr essur e ( bar )
400.0
400.0
Figure 11-2. Measured and calculated viscosities at T =93.3C of the liquid
phase of oil mixture 2 for which the composition is given in Table 4-7. The mea-
\
\
\
\
' " - - - - - Lohr enz et al . usi ng Penel oux densi t y
'"
.""'-
. . . . . . . . .
~
<;
... "-
............ <, --.--"
-----~--_._.
--
--
- - - - - - ~! . . - - ~
- - - - - - - - Peder sen et al .
- - - - - Lohr enz et al . usi ng SRK densi t y
100.0 100.0
Pr essur e ( bar )
300.0
184
Properties of Oils and Natural Gases
11.0 i I

Lohr enz et al . usi ng SRK densi t y
Lohr enz et al . usi ng Penel oux densi t y
10.0
8.0
a.
~
>,
+'
'"
6.0
o
u
'"
".
4.0
\
'"
".."
'" ----
'- -
'<, --
-"----'-
Pr ocedur e of Peder sen et al .
~- ---=- ------------
-.-....--e--- e -----
1.0
0.0
0
.
0
100.0 100.0 300.0 400.0 Pr essur e
( bar )
Figure 11-3. Measured and calculated viscosities at f =71.1e 01the liquid
phase 01oil mixture 3 lor which the composition is given in Table 11-2. The mea-
Table 11-4
Measured Vlscosities in Atmospheric Pressure of Oklahoma Crude Oil
Boiling Point Fractions (Baltatu, 1984)
Average Boiling
Exp. Kinematic
Point 01 Oi! API Gravlty 01Oil
Temperature Viscosity
Sample (Oe) Sample
(Oe) (mm
2
/s)
0.0
0.770
10.0
0.692
87.6
63.3
20.0 0.627
30.0 0.567
40.0 0.523
50.0
0.484
0.0 3.020
10.0
2.420
20.0 1.990
30.0
1.741
212.6 42.3
40.0 1.440
50.0 1.256
60.0
1.100
80.0 0.880
100.0
0.738
API gravity =141.5 - 131.5/specific gravity.
Kinematic viscosity (mm2/s) =dynamic viscosity (cP)/density (g/cm
3
)
Viscosity 185
4.0 I I
~ 3.0
N
E
E
Exper i ment al r esul t s f or cr ude o I wi t h

T
B
= 87. 6 "C
Exper i ment al r esul t s f or cr ude oi l wi t h
T
B
= 212. 6 "C
Peder sen et al . ( 1987)
Ba It at u ( 1984)
>,
+'
.,
'-.
'-.
".-......:~
" ' - ' - . . . ; : : : : ,
o
'"
8 2.0
'"
".
u _
L ---------~
<V
e;
" i.o [ -7-"=- =-=- D , "J
1 75.0 I 50.0 0.00.0 15.0
Temper at ur e
("C )
Figure 11-4. Measured and calculated viscosities 01two Oklahoma crude oils at
atmospheric pressure. The measured data are tabulated in Table 11-4.
SRK-equation and the procedure of Peneloux (both of which are described
in Chapter 5). The results obtained with the procedure of Pedersen et al.
agree very well with the experimental observations. Larger deviations are
found using the procedure of Lohrenz et al. The results obtained using den-
sities calculated from the Peneloux procedure are found to be worse than
those obtained using densities calculated from the SRK-equation. This is
surprising, becausethe Peneloux liquid densities are known to be more accu-
rate than the SRK densities (see Chapter 9).
Presented in Table 11-4 are measured viscosities for two Oklahoma oil
boiling-point fractions (i.e., mixtures of hydrocarbons with the boiling
points within a narrow interval) at atmospheric pressure (Baltatu, 1984).
The results are plotted versus temperature in Figures 11-4, which alsoshows
the results obtained using the procedures of Pedersen et al. and Baltatu. The
calculated resultsof both procedures are in good agreement with the mea- .
sured results.
REFERENCES
Baltatu, M. E., "Prediction of the Liquid Viscosity for Petroleum Frac-
tons,' Ind. Eng. Chem. Process. Des. Dev., 21, 1982, pp. 192-195.
(
Baltatu, M. E., "Predicton of the Transport Properties of Petroleum Frac-
tions," paper presented at the 1984 Winter National Meeting AIChE, At-
lanta, Ceorgia, 1984.
Ely, J . F. and Hanley, H. J . M., "Prediction of Transport Properties. 1. Vis-
cosity of Fluids and Mixtures," Ind. Eng. Chem. Fundam., 20, 1981, pp.
323-332.
Hanley, H. J . M., McCarty, R. D., and Haynes, W. M., "Equation for the
Viscosity and Thermal Conductivity Coefficients of Methane,' Cry-
ogencs, 15, 1975, pp. 413-417.
Herning, F. and Zippener, L., "Calculation of the Viscosity of Technical Cas
Mixtures fram the Viscosity of the Individual Cases," Gas u. Wasserfach,
79, 1936, pp. 69-73.
Hirchfelder, J . O., Curtiss, C. F., and Bird, R. B., Molecular Theory of
Cases and Liquids, J ohn Wiley &Sons, New York, 1954.
J ossi, J . A., Stiel, L. 1., and Thodos, C., "The Viscosity of Pure Substances
in the Dense Caseous and Liquid Phases,' AIChE l., 8, 1962, pp. 59-63.
Lohrenz, J ., Bray, B. C., and Clark, C. R., "Calculating Viscosities of Res-
ervoir Fluids from Their Compositions,' l. Peto TechnoZ., Oct. 1964, pp.
1171-1176.
McCarty, R. D., "A Modified Benedict-Webb-Rubin Equation of State for
Methane Using Recent Experimental Data," Cryogencs, 14, 1974, pp.
276-280.
Mo, K. C. and Cubbins, K. E., "Conformal Solution Theory for Viscosity
.and Thermal Conductivity of Mixtures," MoZ. Phys., 31, 1976, pp. 825-
847.
Pedersen, K. S., et al., "Viscosity of Crude Ol,' Chem. Eng. Sc., 39, 1984,
pp. 1011-1016.
Pedersen, K. S. and Fredenslund, Aa., "An Improved Corresponding States
Model for the Prediction of Oil and Cas Viscosities and Thermal Conduc-
tivities,' Chem. Eng. Sc., 42, 1987, pp. 182-186.
Riazi, M. R. and Daubert, T. E., "Smplify Property Predictions," Hydro-
carbon Process., 59, 1980, pp. 115-116. .'
Stiel, L. 1. and Thodos, C., "The Viscosity of Nonpolar Cases at Normal
Pressures,' AIChE i.. 1961, pp. 611-615.
Tham, M. J . and Cubbins, K. E., "Correspondence Prnciple for Transport
Properties of Dense Fluids, Nonpolar Polyatomic Fluids,' Ind. Eng.
Chem. Fundam., 9, 1970, pp. 63-70.
Chapter 12
Thermal Conductivity
When atemperature gradient existsin afluid, the fluid molecules will try
to eliminate the temperature gradient by transferring energy from regions of
higher temperature to regions of lower temperature. The thermal conduc-
tivity isdefined as the proportionality constant, A, in the following relation
(Fouriers law):
q =-A (~)
(12-1)
where q isthe heat flow per unit area, and (dT/dX) is the temperature gra-
dient in the direction of the heat flow.
Most of the procedures that have been suggested for calculation of the
thermal conductivity of hydrocarbon mixtures (e.g., Reid et al., 1987) are
only applicable to mixtures of well-defined components, or may only be
used for either gas or liquid phases. This chapter deals with calculation pra-
cedures that are applicable also to petroleum mixtures containing undefined
heavy hydrocarbon fractions, and that give consistent results for both gas
and liquid phases.
KINETIC GAS THEORY
Asfor viscosity (seethe Kinetic Cas Theory section in the previous chap-
ter), one may for the thermal conductivity, A, of a dilute gas, derive an ap-
proximate analytical expression. It is given by:
1
A=-nvLC
3 v
(12-2)
where n =number of molecules per unit volume
v =average molecular speed
C, =molar heat capacity at constant volume
L =mean free path between two molecules
187
If it is assumed that the average speed, v, isproportional to (RT/MW)1I2,
the mean free path proportional to 1I(na
2
) where a is the molecular hard-
sphere diameter, and C, isproportional to the molecular weight, the follow-
ing equation may be derived:
" =constant x T1I2/(MW
1I2
a
2
) (12-3)
If it is furthermore assumed that a3is proportional to the critcal molar
volume, Ve, and that Veisproportional to RTe/Pe, the following expression
may be derived for " at the critical point:
"e =constant x P~/3/(MW1I2Te1l6) (12-4)
This expression isused in the calculation of the thermal conductivity using
the corresponding states theory.
CALCULATION OF THE THERMAL CONDUCTIVITY USING THE
CORRESPONDING STATES THEORY
The most accurate procedures for calculation of thermal conductivities of
hydrocarbon mixtures are those based on acorresponding states method sim-
ilar to that used for viscosity calculations (seeChapter 11). In terms of pres-
sure and temperature, the corresponding states theory applied to thermal
conductivity can be expressed as follows:
x, =f(P" Tr) (12-5)
where f is the same function for a group of substances obeying the corre-
sponding states principie. For the reduced thermal conductivity, "" the fol-
lowing equation is used:
"r(P, T) =,,(P, T)/(Te-1I6P~/3MW-1I2) (12-6)
The nominator of the expression at the right side of Equation 12-6 may be
regarded as an approximation of the thermal conductivity at the critical
point (seeEquation 12-4). Using the simple corresponding states theory, the
thermal conductivity of component x at the temperature T, and the pressure
P, may be found from the following equation:
"x(P, T) =(Tex/Teo)-1I6(pex/Peo)2/3(MWx/MWo)-1I2,,0(Po, To) (12-7)
~\
!/rflY"" ~f'~
Thermal Conductivty 89/ ..... \\
( ~ 1 nIIUQ1Et;A ) ~
where Po =P Peo/Pex '. ~\ __ ~
r,=T Teo/Tex ~ V$\:-,.. ~?;"'"
"o =thermal conductivity of the reference substance at the tempera~~.1' '-, o'J
T d P
'i:;" 'r 41
ture oan pressure o '--,.."
As was the case in the discussion of viscosity, methane appears to be acon-
venient choice as a reference substance, however, some corrections must be
introduced as compared with the simple corresponding states principie ex-
pressed in Equation 12-7. The thermal conductivity of poi yatomic sub-
stances (Hanley, 1976) can be separated into two contributions, one due to
transport of translational energy, and one due to transport of internal en-
ergy:
x ="tr +"int
(12-8)
.
Christensen and Fredenslund (1980) have suggested that the correspond-
ing states theory should be applied to the translational term only. They use a
term a to correct for the deviations from the simple corresponding states
model. The final expression for calculation of the thermal conductivity of a
component mixture at the temperature, T, and the pressure, P, isthe follow-
ing:
(
(
(
(
(
(
(
(
, (
"mix(P,T) = (Te,mix/Teo) -1/6 (Pe,mix/Peo)2/3 (MWmix/MWo) -1/2
X (amix/(l!o)("o(To,Po) - "int,o(To)) +"int,mix(T)
(
T ,a,) (p ,a,)
where T; =TI e,mlXmlX and Po=PI e,mlXmlX
Teoao Peoao
(12-9)
Te,mixand Pe,mixare found from Equation 11-27 and Equation 11-30, re-
spectively. The critical temperatures Te, and critical pressures Pe, of the indi-
vidual components are found using the characterization pracedure of Peder-
sen et al. described in Chapter 6. The mixture molecular weight MWmixis
found from the Chapman-Enskog theory as described by Mo and Gubbins
(1976):
MWmix=~[EE (ziZj(lIMWi +lIMWj)1I2(TeiTej)1/4)1
8 i j
]
- 2
((Te;lPci)1I3+(Tc/PCj)1I3)2 Te~~~~P::!iX (12-10)
where Z are mole fractions, and i and j component indices. The interna! en-
ergy contributions to the thermal conductivity ("int,o,of the reference sub-
stance, and "int,mixof the mixture) are both given by:
(
190 Properties o} Oils and Natural Gases
(
(
(
(
Aint=1.18653 7} (C~- 2.5R)f(Pr)/MW
f(Pr) =1 +0.053432 Pr - 0.030182 P; - 0.029725 P~
(12-11)
(12-12)
7}isthe gas viscosity at the actual temperature and apressure of 1atm, C~d
the ideal gas heat capacity at the temperature T (qd isfound as described in
the sections on the SRK Equation of State and the Ideal Gas Heat Capac-
ity). R isthe gas constant. The correlations 12-11 and 12-12 are founded on
the assumption that the internal energy contribution to the thermal conduc-
tivity isclosely related to diffusion of molecules possessing an internal energy
deviating from the average level. Tham and Gubbins (1970) have reported
o-values for N
2
, CO
2
, and the hydrocarbons < C7 Pedersen and Fre-
denslund (1987) extended the model to cover heavy hydrocarbons also. They
use the following expression for a of component i, derived from thermal
conductivity data:
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
ai =1+0.0006004 p;;043 MW~OB6 (12-13)
where Pri =Po(T Tco/Tci, P Pco/Pci)/Pco
(12-14)
amix is found using the following mixing rule:
amix=E E ZZj(aiaj)O.5
i j
(12-15)
Using this mixing rule, the components having small o-values, i.e., small
molecules, are attributed more importance than those having larger a-val-
ues. The smaller molecules are more mobile than the larger ones, and
thereby contribute relatively more to the transfer of energy than do the
larger ones. The calculation of the thermal conductivity of the reference
substance, methane, is based on a model of Hanley et al. (1975) which has
the following form:
A(p,T) =Ao(T) +A(T)p +LlA'(p,T) +LlAc(p,T)
(12-16)
\
In the dense liquid region the major contribution to this expression comes
from LlA'(p, T), which has the same functional form as the expression for
Ll7}' (p,T) given in Equation 11-38. The coefficients h-h have the following
values:
(
h = - 7.0403639907
h =12.319512908
3 = - 8.8525979933 X 10
2
Thermal Conductivity 191
j4=72.835897919
~=0.74421462902
j6= - 2.9706914540
h = 2.2209758501 X 10
3
Pedersen and Fredenslund (1987) have introduced a "low temperature
term" in Equation 12-16, paralleling that in the discussion of viscosity (see
Equation 11-41). Their final expression for the thermal conductivity of
methane is the following:
A(p,T) =Ao(T) +A(T)p +FLlA' (p,T) +F2LlA"(p,T)
+LlAc(p,T) (12-17)
F and F2 are defined in Equation 11-41. Based on thermal conductivity
data measured at reduced temperatures below 0.4, i.e., corresponding to a
methane reference state, where methane is in' a solid form, the following
results were obtained for LlA"(p, T):
LlA"(p,T) =exp(I +14/T) [ exp (pO (12+;~2)
+ 8p" 11, +~ +~)) - 101
where 1 = - 8.55109
12=12.5539
13= - 1020.85
14=238.394
15=1.31563
16= - 72.5759
17= 1411.60
(12-18)
Baltatu et al. (Baltatu, 1984; Baltatu et al. 1985) have suggested that the
thermal conductivity iscalculated from asomewhat different corresponding
states model than the one previously described. They also calculate the ther-
mal conductivity of a mixture as the sum of a translational and an internal
contribution:
Amix(p,T) =A:(p,T) +A~ix(T) (12-19)
A: (p, T) is the translational contribution to the thermal conductivity of a
hypothetical pure fluid, x, with the density, p, and at the temperature, T.
A~ix(T)is the internal contribution to the thermal conductivity of the mix-
ture at the temperature, T. The translational contribution isfound by apply-
ing the corresponding states principie:
A;(p,T) =A~(Po,To)FA
(12-20)
where A~(Po, To) isthe translational contribution to the thermal conductivity
of the reference substance, o (methane), at the temperature To, and the den-
sity Po' FAis found from the expression:
FA=(MW /MW )112 fl/2 h-2/3
o x X,O X,O
(12-21)
In this equation:
fx,o=h~; E E zzjfjhj
i j
(12-22)
and hx,o=E E zzjhj
i j
(12-23)
where fj =(ffj)1/2
(12-24)
1
and h.. =- (h~/3 +h/3)3
1) 8 1 )
(12-25)
The parameters f and h are, for component i, calculated as follows:
f =8 Te/Teo
(12-26)
h = Vd/Veo
(12-27)
where 8 ando are shape factors identical to those entering into Equation
11-16 for calculation of the viscosity. Teis the critical temperature, and Ve
the critical molar volume. The properties are found from the correlations
mentioned in Chapter 11under the Corresponding States Theory heading as
those used by Baltatu to extend the viscosity model of Ely and Hanley to
apply to mixtures containing heavy hydrocarbons. The density Po, and the
temperature T of Equation 12-20 are found as follows:
Po = phx,o; To =T/fx,o
(12-28)
The internal contribution to the thermal conductivity of Equation 12-19
is calculated as follows:
Thermal Conductivity
~~- i - +- t - t - - +- ~- +- - ~~
Gil
e Sl
5l
:u ..<..
'C :::
cf ..
..<
11I
Q)
:;:
~
::::J
"O
e
o
O
cu
E
Q)
s:
....
"O
Q)
;
... -
::::J
N U
... cu
Q)O
:c"O
~e
111
~
Q)
E
.:
~
><
W
-
o
e
s
.:
111
a.
E
o
O
C\!I~
I X) ( " )
I
~I f ' - : Ol U' )
a C\ J
~ ~
a
M
~
a
ci
a
~
C\!
Ol
I
Ol
M
I X)
.,
el)
~I I I I I I I I I I
:;
en
~
(ij
E
ID
E
.~
x
~
=-
:;
en
~
(ij
E-
U1Q) ~
Ea>
~.~ :::.
~ a. ::l
I ~ 1
~I~
'> !:: al
ro ::J ..
c,~~
U ~ "O
;:: "O _
c~~~
s g--5 ~
~~~ .~
c::: . o Q)
.- <t- - a.
~ ::: x ~
~ ti 8 Ql-
ctS >- T""" U
E'~ 11 :;
g ~ > ~
J I~~~
I I I I I I I I I I~~ ~8
;::
el)
'"
~
~I ( " ) I " ' I C\ ! e r- __ C\l
:f I I
.,
~I C\ ! I ~ ( " ) I X) Ol
( " ) I X) ~
~ ~
~I ~I ~I ~
e a I X) Ol
-s. I
~ISl
!:: E
81 ..< ~
..
f ' - : 1c>1~
' <t i'- i'-
( " ) Ol C\ J
~ ~
,.<?
c..E
>e-
w:::
..
~I " ' I ~
U') Ol <O
( " ) I X) ~
~ ~
I-g
~I ~I ~ ( " ) ( " ) <O
i'- I X) Ol
C\ J ( " ) C\ J
-~
Q, >
I!
CJ
f ' - : I C\ ! I ~
U') i'- i'-
U') U') ' <t
!'Sl
c>1~1~ U') v Ol
U') I X) <O
( " ) ( " ) ' <t
ti
:::J
'C
~
Q,
E
:::J
QI
~
~
~
al a ,
al r-, U)
~ c:h 1-_
'" - QI
Q) al c:
...J c: al
e g e
I D tU al
Q, (!J :.::
Q) ~
c: al
al " C
(/l : : J
o ~
~
~'~IO E ~ (/l
el::J '"
::J ~ (!J
' <l : 1f ' - :
~ C\ J
I ~
;;;I~
~ ( " )
~ ~
f ' - : I c>
Ol ~
~ C\ J
~I~ ( " ) ( " )
( " ) Ol
( " ) C\ J
" ' I ' <l :
U') U')
' <t ( " )
' <l : 1~
' <t I X)
a C\ J
U') U')
u
~
E
e
- '" al ><
c: '"
Q) : : E
(/l
e~
ala
: ' : : C\ J
C\ J
M
I
r-,
,...:
'"
N
~
(")
M
(")
(")
U')
~
C\!
~
(")
U')
.~
Qi o
~ E
_ ::J
al : : J
(/l u
.!!!~
o~
I X)
M
~
U')
Ol
' <t
N
Ol
(")
M
~
I X)
~
(")
' <l :
<O
I X)
U')
193
~
~
"!
~
~
<O
N
~
Ol
(")
C\ J
~
(
(
a
ci
~
(")
c
Ol
(")
I X)
,...:
~
r-,
,...:
<O
<O
ti
al _
I O
u ~
~~
E ~
e '"
<~
A,~x(T)=1: 1: zz/\j(T)
j
(12-29)
where Aj(T) =2 [l/A "(T) +l/A "(T)]- 1 (12-30)
For apure component, i, the internal contribution to the thermal conduc-
tvity is found from the expression:
(
(
A"(T)=1.32 (C:~i (T) - ~R)r,(T)/MW (12-31)
where C~~i =ideal gas heat capacity of component i at the temperature T
(C~ may be found as described in the section on the SRK-
Equation of State in Chapter 5, and the Calculation of the
Ideal Gas Heat Capacity, in Chapter 7.)
R =gas constant
r'(T) =dilute gas viscosity (see section on Viscosity Correlation of
Lohrenz et al. in Chapter 11) of component i at the tempera-
ture T
MW =molecular weight of component i
COMPARISON WITH EXPERIMENTAL OBSERVATIONS
It has not been possible to find thermal conductivity data for petroleum
reservoir fluids or any other rnult-component hydrocarbon mixtures. The
most interesting data are those published by BaItatu (1984), covering ther-
mal conductivity data for petroleum fractions, and each containing hydro-
carbons with the normal boiling point within a narrow temperature inter-
val. Some of these data are shown in Table 12-1, in a comparison with
resuIts obtained using the calculation procedures of Baltatu et al., and Pe-
dersen et al. Calculated results of comparable quality are obtained using ei-
ther of the two calculation procedures.
REEERENCES
Baltatu, M. E., "Prediction of the Transport Properties of Petroleum Frac-
tons," paper presented at the 1984 Winter National Meeting AIChE, At-
lanta, Georgia, 1984.
Baltatu, M. E. et al., "Therrnal Conductivity of Coal-Derived Liquids and
Petroleum Fractions,' Ind. Eng. Chem. Process Des. Dev., 24, 1985, pp.
325-332.
Thermal Conductivity 195
Christensen, P. L. and Fredenslund, Aa., "A Corresponding States Model for
the Thermal Conductivity of Gases and Liquds," Chem. Eng. Sci., 35,
1980, pp. 871-875.
Hanley, H. J . M., McCarty, R. D., and Haynes, N. M., "Equation for the
Viscosity and Thermal Conductivity Coefficients of Methane,' Cry-
ogenics, 15, 1975, pp. 413-417.
Hanley, H. J . M., "Predction of the Viscosity and Thermal Conductivity
Coefficients of Mxtures,' Cryogenics, 16, 1976, pp. 643-651.
Mo, K. C. and Gubbins, K. E., "Conformal Solution Theory for Viscosity
and Thermal Conductivity of Mxtures,' MoZ. Phys., 31, 1976, pp. 825-
847.
Pedersen, K. S. and Fredenslund, Aa., "AnImproved Corresponding States
Model for the Prediction of Oil and Gas Viscosities and Thermal Conduc-
tivities,' Chem. Eng. Sci., 42, 1987, pp. 182-186.
Reid, R. C., Prausnitz, J . M., and Poling, B. E., The Properties oj Gases and
Liquids, McGraw-Hill, New York, 1987.
Tham, M. J . and Gubbins, K. E., "Correspondence Principie for Transport
Properties of Dense Fluids. Nonpolar Polyatomic Fluids," Ind. Eng.
Chem. Fundam., 9, 1970, pp. 63-70.
Chapter 13
Surface Tension
Knowledge of the oil/gas surface tension is needed to understand the
pore-level flow processes in areservoir. The molecules located in the surface
of a liquid in equilibrium with a gas are exposed to forces differing from
those acting between molecules in the bulk liquid or gas phases. The attrac-
tion from a neighboring gas molecule is less than the attraction from a
neighboring liquid molecule. The surface layer is in tension and tends to
contract to the smallest area compatible with the mass of material, con-
tainer restraints, and external forces. The surface tension, a, is defined as
follows: ~
a =(aGlaAh,v,N (13-1)
where G =Gibbs energy
A =surface area
T =temperature
y =molar volume
N =number of molecules
The surface tension can be interpreted as the resistance from the liquid
against an imposed increase in the surface area.
ESTIMATION OF THE SURFACE TENSION USING THE
MACLEOD-SUGDEN CORRELATION
Brock and Bird (1955) have shown that the surface tension of pure non-
polar compounds can be approximated by the following relation:
a =Ac (l - Tr)O (13-2)
where (J =11/9, T, is the reduced temperature ( =T/Tc) and
Ac =p~/3 n
/3
(0.133 (Xc - 0.281) (13-3)
r
I
I
Surface Tension 197
(Xc is the Riedel parameter (Riedel, 1954):
(Xc =0.9076 (1 +TBr In Pc)/(l - TBr) (13-4)
P, is the critical pressure in atm, T, the critical temperature in Kelvin,
and TBr the reduced boiling point (TB/Tc)' For mixtures, T, and P, will not
always be known, and it is also questionable whether the surface tension of
amixture isrelated to the critical properties in the same way as isthe surface
tension of apure component. For amixture, it ismore convenient to express
the surface tension in terms of phase properties, which can be measured, or
accurately calculated; ascan, for example, phase densities. Totransform the
temperature dependence to density dependence, the following approxima-
tion is used (Fisher, 1967):
PL - Pv =B'(~c - T)1i
(13-5)
where PL and Pv are the liquid and vapor phase densities, respectively. {3 is a
constant, which, for most pure components, has a value between 0.3 and
0.5. The coefficient B' has the dimension (density/temperature). It istrans-
forrned into a dimensionless constant B by rewriting Equation 13-5 to (Lee
and Chien, 1984):
(PL - Pv)1 Pc =B(l - Tr)1i
(13-6)
where Pc isthe critical density, and T, the reduced temperature (T/T
c
)' Sub-
stitution of Equation 13-6 into Equation 13-2 gives:
!J.IiIOy _ )
alilO ='''c_ c(PL - Pv
B
(13-7)
where Yc isthe critical molar volume, and PL and Pv are the molar densities
of the liquid and vapor phases, respectively. The coefficient:
NgIOyc
B
is called the Parachor [P]. Equation 13-7 can then be rewritten to:
a
lilO
=[P](PL - Pv)
(13-8)
which is often referred to as the Macleod-Sugden relation (Macleod, 1923,
and Sugden, 1924). The equation can be extended for application to rnix-
tures, as proposed by Weinaug and Katz (1943) who use {31(J =1/4:
al/
4
=.E (,odP]x - ,ov[PJ y)
i
(13-9)
where i is a component index, and x and y are liquid and vapor mole frac-
tions.
A second modification of Equation 13-8 with application to mixtures has
been presented by Lee and Chien (1984). The surface tension is found from
the following expression:
al/
4
=,odP]L - ,ov[P]v (13-10)
The liquid and vapor phase Parachors, [P]L and [P]v are found from the
expressions:
A
1
/
4
[PJ L = eL X VeL
B
L
;
A
1
/
4
[P]V = eV X V
eV
Bv
(13-11)
where VeL =.E XVe
i
B
L
=.E xB
2/3 1/3
AeL=P
eL
T
eL
(0.133 aeL - 0.281)
P
eL
=.E xPcj
i
,.
T
eL
=.E xTcj
i
aeL =.E xacj
i
x stands for liquid mole fractions, and i is a component index. The same
mixing rules are applied to the vapor phase, i.e., Xis replaced with the va-
por mole fraction, y. B-values for a number of oil and gas constituents are
given in Table 13-1. The critical properties of the same components are
given in Table 5-1. The properties entering into Equation 13-11 for the C7+-
fractions may be determined using the C7+-characterizaton procedure of
Pedersen et al. described in Chapter 7. Te and Peof the total C7+- fraction
are found as molar averages.
Pedersen et al. (1988) have suggested that the critical molar volume of the
C7+-fraction is estimated from the following equation of Riedel (1954):
Surjace Tension 199
Table 13-1
Values ot the B-parameter (Lee and Chien, 1984) Entering into Equation
13-11 tor Determining the Parachor
Compound B-parameter
Water 4.310
Carbon dioxide 3.505
Nitrogen 3.414
Methane 3.403
Ethane 3.591
Propane 3.602
n-Butane 3.652
n-Pentane 3.690
n-Hexane 3.726
Ve=RTe/[Pe(3.72 +0.26(ae - 7.0))] (13-12)
and that the Parachor of the C7+-fraction is calculated from the following
equation (Nokay, 1959):
IOglO[P]= - 8.93275 +3.68894 logo(Te/Seo6676) (13-13)
Te is the critical temperature in R, and se the C7+-specific gravity
(60F/60F). Knowing fP] it is possible to determine B of the C7+-fraction
by rearranging Equation 13-11:
B =A~/4Ve
[P]
(13-14)
The liquid and vapor phase Parachors may now becalculated from Equa-
tion 13-11. If the phase densities are not known experimentally, they can be
calculated as shown in Chapter 5, and the surface tension may finally be
calculated from Equation 13-10.
ESTIMATION OF THE SURFACE TENSION FROM THE VISCOSITY
It has been experimentally observed, e.g., by Pelofsky (1966) and by Pe-
dersen et al. (1988), that for pure hydrocarbons, and for oil mixtures at at-
mospheric pressure, there isan almost linear dependence between the recip-
rocal of the viscosity, r, and the logarithm of the surface tension:
In a =In A +B/r (13-15)
where A and B are constants depending on composition only, Pedersen et al.
have suggested that the A and B parameters entering into Equation 13-15
for a hydrocarbon mixture should be determined from the following rela-
tions:
A =EW[(Clp) +(C2N) +(C3A)]
(13-16)
B =EW[(C4pMW~7)+(CsNMW?) +(C6AMw~g)]
(13-17)
where i = a given carbon number fraction
w = weight fraction
MW = molecular weight
Table 13-2
Coefficients Entering into Equations 13-16 and 13-17 for Calculating
Surface Tension of a Hydrocarbon Mixture (Pedersen et al., 1988)
(Surface Tension is in mN/m ano, 1) in cP)
Coefficient
Numerical Value
e,
3.118 x 10
e2
2.805 x 10
e,
3.673 x 10
e,
- 3.282 x 10-
2
e,
- 6.096 x 10-
6
e,
- 6.098 x 10-
4
e-
- 3.641 x 10-'
e,
2.191
eg
1.229
I
111
P, N, and A are the mole fractions of paraffinic, naphthenic and aromatic
components, respectively. The recommended values for the coefficients Cl-
Cgare given inTable 13-2. At higher pressures, where the vapor phase contri-
bution must be taken into 'consderation, Pedersen et al. recommend that the
following formula be used:
ln a =ln A +BI Ll1'/
(13-18)
(
o.s 0.S)2
where Ll1'/ = 1'/L - 1'/v
The subindices L and V refer to the liquid and vapor phases, respectively.
;j
M ti) ti)
I e GI
M .2 ..
ti) :::1
GI e )(
: a~: E
!!! GI
U
.e
..
:::1
en
..
o
-ti)
- :;
ti)
GI
a:
'O
GI
;j
:;
U
~
'O
e
1'1:I
ea
-e
GI
E
.;:
GI
Q.
><
W
Surface Tension 201
~~ I I
~
(5
1'1:I
GI
en
.c
-
..
o
z
o
~
El e .,z ,...:
E C\I
Cl
e
r-,
oi
C\I
~
Cl
io
al
C\I
r-,
cD
o
o
lO
M
o
,...:
C\I
v
q:
C\I
~
~, I
lO
,...:
C\I
~I <' : !
<O lO
C\ I C\ I
;r:J I I
~
-
O
GI
..
:::1
ti)
ti)
GI
..
o..
U
.;:
GI .,
.c
Q.
ti)
O
E
<
El f ' - . ., Z ,...:
E C\ I
en
O
~
f'-.
O
C')
lO
N
v
01 I
I
C\ I f'-.
cD ,...:
~
~ al
O , . . . :
C') C\I
~
~
'<1:
lO
C\ I
:>
~ al
~~I I
11
~ EI C\ I
o Z cri
E C\ I
en
N
~
~
c;:;
v
<Xi
C\I
<OI~
f'-. lO
C\ I C\ I
<O lO
<Xi oi
C\ I
C')
C\I
,...:
~
en
C\I
C\I
<Xi
C\ I
al
M :::1 1
~t!JI I I
11 I . I <' : !
~ "z al
E C\ I
tq
~
~
r-,
C\I
C')
C')
oi
C\I
~I <' : !
en <O
C\ I C\ I
o
<Xi
~
N
C')
C\ I
oi
r-,
q:
r-,
al
C\I
lO
<Xi
C\ I
r-,
N ~I 1
'O I ci.
'l:I fII ><
e al W
!;2II:
al
el>
c. i c: :
"j .Q
010
::l
cr
W
o
C')
' 
.2~
(3 g I a.
'" ) (
g. ~
W a.
x
Q)
in
;:;;
.2 ~I ci,
'" O ><
0';::: W
'" ::l
rr
W
o
~
el>
uc.2c:
- o '" O
(3 ~ 0~
::l ::l
rr rr
W W
al
el>
~
al
..
::l
)(
:li
<5
c::
O
:;: Q)
'00 :o v
8.~el>
Ec::~
O. -
o
e
el>
~
c. i c: :
"j O
o.;:::
'" ::l
rr
W
c::
O
"';:::; ID
"Ci515 ll)
8.~'"
E c:: ~
O . -
o
u
Cii
.!.
x
i I I I ~
:>
Q)
O
202 Properties of Oi/s and Natural Gases
Table 13-3 gives experimental and calculated results for two North Seaoil
mixtures. Calculated results of using Equations 13-9, 13-10, and 13-15, re-
spectively, are shown. The best results are obtained with Equations 13-15
and 13-10. The logarithm of the experimental surface tension is plotted
against the recprocal of the viscosity in Figure 13-1.
Table 13-4
Molar Composition of on Mixture for whlch Surface Tension Results are
Shown in Table 13-3
Molecular Density, (g/cm
3
)
Component Mole % Weight at 15C, 1 atm.
C,-Cs 3.70 51 -
C6 3.25 80 0.644
C7 8.38 93 0.733
C8 7.28 107 0.763
C9 7.28 120 0.785
C'O
5.26 131.6 0.808
Cll 5.17 147.6 0.815
C'2
4.44 158.6 0.836
C'3
5.11 172.4 0.850
C'4
4.58 184.8 0.861
C'S
4.29 197.4 0.873
C'6
3.45 209.0
.~
0.882
C'7
3.53 226.5 0.873
C'8
3.01 242.6 0.875
C'9
2.64 253.7 0.885
C20 2.51 262.0 0.903
C2, 1.87 281.1 0.898
C22 1.97 293.2 0.898
C23 1.60 307.3 0.899
C24 1.56 319.6 0.900
C2S 1.38 333.3 0.905
C26 1.19 345.9 0.907
C27 1.38 361.1 0.910
C28 1.30 373.5 0.915
C29 1.26 380.8 0.920
C30+ 12.80 624.0 0.953
Surface Tension 203
In the literature, experimental surface tension data are scarce for petro-
leum mixtures at elevated pressures. Simon et al. (1978) presented high-pres-
sure data for surface tensions of a reservoir fluid to which considerable
amounts of carbon dioxide were added. The composition of the reservoir
fluid isgiven in Table 13-6, and the measured results for the surface tensions
are presented in Table 13-7. Lee and Chien (1984) have calculated the sur-
face tensions using Equations 13-9 and 13-10, respectively. Their results are
Table 13-5
Molar Composition of 011 Mixture for whlch Surface Tenslon Results are
Shown in Table 13-3
Molecular Density, (g/cm')
Component Mole % Weight at 15C, 1 atm.
C,-Cs 12.35 61.7 -
C6 4.66 82.0 0.685
C7 11.76 98.7 0.737
C8 7.96 109.6 0.754
C9 7.84 121.9 0.774
C1Q 5.73 133 0.789
Cll 5.06 148 0.794
C'2
~
4.72 163 0.806
C'3
4.07 177 0.819
C'4
4.41
190 0.832
C'S
3.50 204 0.834
C'6
2.59 217 0.844
C17 3.06 235 0.841
C'8
2.15 248 0.847
C'9
2.45 260 0.860
C20 0.81 269 0.874
C2, 1.98 283 0.870
C22 1.31 298 0.872
C23 1.10 310 0.875
C24 0.84 322 0.877
C2S 1.19 332 0.881
C26 0.96 351 0.886
C27 0.76 371 0.888
C28 0.81 382 0.895
C29 0.59 394 0.898
C30+ 7.33 612 0.935
'pamseaw(pamseaw - paiajnojao x 001 =UOIeIAaa %
Z'v ~ 588'0 ZZZ - 809'0 5LL'0 8' ~OZ 17'175
OZ/08
6'8 - 888'0 0'917 - 9617'0 6~6'0 ~'99~ 17'175
OZ/08
["6 - ~66'0 5'917 - L85'0 L60' ~ Z'68~ 17'175
OZ/08
v' ~ - 8ZV'0 6'58 - 8Lc:'0 17817'0 6'L8~ 17'175 517/55
UO!le!lI<la % O ~-E ~ UO!lenb3 UO!le!/laa % 6-E ~ UO!lenb3 paJnseaw (req) d (::>0) 1 I!O/" O::>
palelnole::> palelnoleo O!leH 6u!x!W
(WlNw) UO!SUal aoe,JnS
ap!xo!o uoqJeO pue (9-E ~ alqe.l U! uoursoduiorj I!O J !oI\J asa~e ~osaJnlX!W
~ouojsuej aoepns aln J O~("86~'Ua!4:> pue aal) sllnsa~palelnole:> pue (8L6~ 'Ie la UOW!s) pamseaw
z-cr alqe.l
"-
u
e
-- e
c:
o
~ -e
VI .),
o
o-
l5 '"
U
.D
... '"
'" ....
/e
L- L- - - - - - - ~- - - - ~~
( W/ NW Ul D) ( D) U
~
e
.2
<') N en O
c.; O
t-, N
~
io
r::.
N <O
O O ~ 00 <') lO
f'-. <t <')
N
<O
-e-
'"
r-, <t N lO <t N N
O
<')
at
O ~ O
-r-
~
<') ~ ~ O
~
O O O O O O ~
I
G> O 6 6 6 O O O O O O O O 6 6
"O
::E
e
G>
c:
o
Q.
E
+
o
o
'"
~
o
eS
.s

.s c3

c3 c3'

u
o
E
u
C,
l!)
al
<D
o
I
e
o
c.
.,
:
.D
:J
.D
1i
Z-
;
"
.,
o
Table 13-8
Molar Composition of Reservoir Fluld for which Surface Tenslon Results
are Given In Table 13-9
Component Mole %
N2 0.03
CO2 2.02
C, 51.53
C2 8.07
C3 5.04
-C. 0.83
n-C. 2.04
ic, 0.84
r-e, 1.05
C6 1.38
C7 27.17
e7 molecular weight: 217.
Source: Firoozabadi el al. (1988).
Table 13-9
Measured and Calculated Results at T :::;82.2C for Surface Tension
(dyn/cm) of Reservoir Fluid of Composition Given in Table 13-8
Pressure Measured
Equation 13-10 Equation 13-15
(bar) (mN/m) (mN/m) % Dev. (mN/m) % Dev.
263.0 1.3 1.3 0.0 1.3 0.0
228.5 2.3 2.0 - 13.0 2.2 - 4.3
194.0 3.3 3.1 - 6.1 3.3 0.0
159.6 4.6 4.7 2.2 5.2 13.0
"
% Dev. =100 x (calculated - measured)/measured.
Sources: Firoozabadi el al. (1988) and Pedersen el al. (1988).
also given in Table 13-7. Also, in this case, Equation 13-10 is superior to
Equation 13-9. Equation 13-15 is not applicable due to the high concentra-
tion of non-hydrocarbons (C02).
Firoozabadi et al. (1988) have presented experimental results for the sur-
face tension of the reservoir fluid of Table 13-8, for pressures ranging from
160 bar to 263 bar. The results are shown in Table 13-9 in comparison with
the results obtained using Equations 13-10 and 13-15. Positive agreement
with the measured results is obtained with both correlations.
Surface Tension
207
REFERENCES
Brock, J . R. and Bird, R. B., "Surface Tension and the PrincipIe of Corre-
sponding States," AIChE t.. 1, 1955, pp. 174-177.
Firoozabadi, A., Katz, D. L., Sonoosh, H., and Sajjadian, V. A., "Surface
Tension of Reservoir Crude Oil-Cas Systems Recognizing the Asphalt in
the Heavy Fraction," SPE 13826, SPE Reservor Engneerng, 3, 1988,
pp. 265-272.
Fisher, M. E., "The Theory of Equilibrium Critical Phenomena," Report on
Progress in Physics, A. C. Stickland, editor, 1967, p. 616.
Lee, S.-T. and Chien, M. C. H., "A New Multicomponent Surface Tension
Correlation Basedon Scaling Theory," SPE/DOE Fourth Symposium on
Enhanced Oil Recovery, Tulsa, Oklahoma, April16-18, 1984, pp. 147-
158.
Macleod, D. B., "On a Relation Between Surface Tension and Density,"
Trans. Faraday Soc., 19, 1923, pp. 38-42.
Nokay, R., "Estimate Petrochemical Propertes," Chem. Eng., 66, 1959, pp.
147-148.
Pedersen, K. S., Lund, T., and Fredenslund, Aa., "Surface Tensions of Petro-
leum Mixtures," f. Can. Peto Tech., in press 1988.
Pelofsky, A. H., "Surface Tenson. Viscosity Relation for Liquids,' f. Chem.
Eng. Data, 11, 1966, pp. 394-397.
Riedel, L., "A New Universal Vapor-Pressure Equation. 1. The Extension of
the Theories of Corresponding States,' Chem. Eng. Tech., 26, 1954, pp.
83-89.
Simon, R., Rosman, A., and Zana, E., "Phase-Behavior Properties of CO
2
-
Reservoir Oil Systems,' Soco Peto Eng. s.. Feb. 1978, pp. 20-26.
Sugden, S., "The Variation of Surface Tension with Temperature and Some
Related Functions,' f. Chem. Soc., 1924, p. 32-41.
Weinaug, C. F. and Katz, D. L.,. "Surface Tensions of Methane-Propane
Mixtures," lnd. Eng. Chem., 35, 1943, pp. 239-246.
Chapter 14
Wax Formation and Inhibition
WAX COMPOSITION
The majority of crude oils contain a certain proportion of heavy hydro-
carbon compounds, which may precipitate as awaxy salid phase if the oil is
cooled below a certain temperature. Wax precipitation may take place at
temperatures far above the freezing point of water, and istherefore apoten-
tial problem when petroleum mixtures are transportd in pipelines, e.g., un-
derwater, or treated in aprocess planto Wax precipitation may result in plug-
ging of pipes and process equipment. When designing pipelines and
separation plants it is, therefore, of importance to be able to determine the
conditions where wax precipitation takes place, and the amount of wax
likely to formo
Wax precipitated from oil mixtures consists primarily of closed packed lat-
tices of aligned paraffinic and naphthenic molecules, shown schematically
in Figure 14-1. The most comprehensive wax-composition analyses have
been made on waxes extracted from petroleum mixtures. They can be di-
vided into two main categories: paraffinic waxes and microcrystalline
waxes. The basic characteristics of each category are shown in Table 14-1
(Gilby, 1983).
COMPUTAT.ION OF WAX FORMATION
The chemical potential of component i in aliquid phase at temperature T
may be expressed as follows:
L
POi p,9
RT =R~+lo (xn+')
(14-1)
and the chemical potential of component i in a wax (salid) phase as follows
(Prausnitz et al., 1986):
Wax Formaton and Inhibition 209
CH3 CH2 CH2 CH2 CH2 CH2 CH2 CH2 CH2 CH3
\ / \ / \ / \ / \ / \ / \ / \ / \ /
CH2 CH2 CH2 CH2 CH2 CH2 CH2 CH2 CH2
NORMAL PARAFFIN
CH2 CH2 CH2 CH2 CH2 CH2 CH2 CH2 CH2
/ \ / \ / \ / \ / \ / \ / \ / \ / \
CH3 CH2 CH2 CH2 CH2 CH2 CH2 CHz CH CH3
\
CHz
/
CH3
ISOPARAFFIN
\)\I\/\;\I\lVV\
CH2 CHz CH2 CH2 CH2 CH2 CH2 CH CH2
I I
CH2 CH2
~~
CYCLOPARAFFIN
Figure 14-1. Typical formulae of petroleum waxes.
Table 14-1
Typical Compositions and Properties of Paraffin and
Microcrystalline Waxes
Paraffin Microcrystalline
Waxes Waxes
Normal paraffins (%) 80- 95 0- 15
Branched paraffins (%) 2- 15 15- 30
eycloparaffins (%) 2- 8 65- 75
Melling poinl range (Oe)
50- 65 60- 90
Average molecular
weighl range
350- 430 500- 800
Typical carbon number
range
18- 36 30- 60
eryslallinity range (%) 80- 90 50- 65
Source: Gilby (1983).
J -Lr_ J -Lf
RT - RT +In (s-yr)- .6Hf (1 - T/TD/RT
(14-2)
In Equations 14-1 and 14-2:
(
(
(
(
(
i =a component index
J-L0 =a standard state chemical potential
.6Hf =the molar enthalpy of melting
Tf =the pure component melting point temperature
x =the mole fraction in the liquid phase
s =the mole fraction in the wax phase
-yL=activity coefficient in liquid phase
-ys=activity coefficient in wax phase
R =the gas constant.
If at a given temperature the change in Gibbs energy associated with the
formation of a wax phase is negative for all possible wax phase composi-
tions:
.6G =I.>M +1:nrJ-Lr - 1:n1J-L1 ~ O
i i i
(14-3)
\
(
(
(
no wax is formed, i.e., the temperature T is higher than the cloud point
(Twax)' In Equation 14-3 n means mole number, Zis the mole fraction of
component i in the total mixture and F refers to the total mixture (feed).
The cloud point is defined as the temperature at which the first small
amount of solid (wax) isformed. At thermodynamic eqnilibrium between a
liquid and awax phase, the following equation must be satisfied for all com-
ponents:
.6J-L/RT =(J-Lr - J -Lr)/RT=O
(14-4)
Won (1985, 1986) uses the following expression to find the activity coeffi-
cient of component i in both the liquid phase, and in the wax phase:
V(b- D)2
ln-y= RT
(14-5)
where D is Hildebrand's solubility parameter of component i (Prausnitz et
al., 1986) and
b =1:D
(14-6)
li~
~~.
r
Wax Formaton and Inhibition 211
 is the volume fraction of component i, and is found as follows:
<l>r = xV
1:xV/
J
sV s 1
 - 1:xV
j
(14-7)
V is the molar volume. Won gives values for the liquid phase solubility pa-
rameters, OL,up to C40. These values are shown in Table 14-2.
Table 14-2
Liquid Phase Solubility Parameters Recommended by Won (1986)
b
L
b
L
b
L
Component (cal/cm3)o .s Component (cal/cm3)o .s Component (cal/cm3)o .s
CO2 7.12 C
'5
7.96 C29 8.24
C, 5.68 C'6 7.99 C30 8.25
C2 6.60 C17 8.02 C31 8.26
C3 6.65 C'8 8.05 C32 8.27
C4 6.65 C'9 8.07 C33 8.28
C5 7.02 C20 8.09 C34 8.29
C6 7.25 C21 8.11 Ca5 8.30
C7 7.41
C22 8.13 C36 8.31
C8 7.53 C23 8.15 Ca7 8.32
C9 7.63 . C24 8.17 C38 8.33
C10 7.71 C25 8.18 Ca9 8.34
C11 7.78 C26 8.20 C40 8.35
C'2 7.83 C27 8.21
C13 7.88 C28 8.22
C
'4
7.92
According to Won, the solubility parameters of the components in the
solid phase, os, are found from the following expression:
(or)2 =(oh2 +.6Hf
V
(14-8)
where .6Hf =O.1426MW . Tf (14-9)
MW is the molecular weight of component i.
By comparing Equations 14-1, 14-2, and 14-4, the conditions for thermo-
dynamic equilibrium of Equation 14-4 can be rewritten:
s/x =(-yr/-yf)exp(.6Hf(l - T/TD/RT) (14-10)
The cloud points are much toa high when predicted using the model of
Won (see section on Comparisons Between Calculated and Experimental
Results). To account for that it is necessary to drastically reduce either the
liquid phase activity coefficients or the melting enthalpies. Hansen et al.
(1988) have chosen the former alternative. They have suggested using poly-
mer solution theory as described by Flory (1953) for the liquid phase. The
following general thermodynamic relation is used:
L\Gmix =L\Hmix - TL\Smx (14-11)
where L\Gmix, L\Hmix, and L\Smix are the changes of mixing of the Gibbs en-
ergy, the enthalpy, and the entropy. For an ideal mixture:
L\Hmix =O (14-12)
(14-13) L\Smx =- k E nln(nl E nj)
i j
For a real mixture, Flory uses the following expressions:
L\Hmix =kT E E nvXj;
j <j
Xii =O (14-14)
where k =the Boltzmann's constant
Vj =the volume fraction of component j
Xj = a binary interaction parameter between components i and j.
The volume fraction of component i can be approximated as follows:
v =MWnl E MWjnj
j
(14-15)
For the binary interaction coefficients, the following relation applies
(Flory, 1953):
Xj =Xj(MWlMW) (14-16)
The entrapy of mixing is for a polymer soluton expressed as:
L\Smix = - k E n.lnv, (14-17)
The term (nl~nj), employed in the ideal case (Equation 14-13), is
::'"';~~
~~II\tO~
~~ " tI;- 1
.. q - .
Wax Formation and Inhibition 213( ~( n\tUO~
i y\DI~-.
replaced by Vin Equation 14-17. The following expression may now be de- ... ~~I~
rived for the Gibbs energy of mixing: \ 1-, e
*
L\Gmx =kT E E nvXj+kT E n.Inv,
j i<j i
(14-18)
By partal differentiation of this expression with respect to ni, it ispossible
to derive an expression for the chemical potential of component iin the lq-
uid phase:
(IIJ.. - 119)/RT =lnv, -t 1- v. - " (MW/MW)v. ,.,...1 r-l 1 1 1 J... I J J
j*i
+E VjXjE Vk- E E (MWlMWj)VkVjXkj (14-19)
j*i k*i j*i k<j
..
The molecular interactions are divided into three classes:
1. PN-PN
2. PN-A
3. A-A
where P, N, and A represent paraffinic, naphthenic, and aromatic mole-
cules, respectively. It is assumed that the interactions between paraffinic
components with other paraffins are identcal to those with naphthenic
components, and vice versa. The interacton coefficients are found fram the
following relations, based on experimental wax formation data for real pe-
traleum mixtures (Hansen et al., 1988):
PN-PN:
Xlj =2.17884 X 1O-
3
IMW - MWjl
Xj = - 1.76960 x 10- 4 (MW . MW
j
)
Xj = - 8.53350 X 1O-
3
IMW - MWjl
(14-20)
(14-21)
PN-A:
A-A:
(14-22)
where the molecular weight of component i, MW, islarger than the molec-
ular weight of component j, MWj (see aIso Equation 14-16).
Both the liquid and the wax phases have so far been treated as bulk
phases, and interfacial contributions are neglected.
According to Mandelkern (1964), however, crystallization consists of two
basic processes, nucleation and crystal growth .. Nucleation is a necessary
condition for a crystal to grow. Crystalline structures below a certain size
will be unstable and decompose, because the surface tension of the crystal
contributes to the chemical potental of the crystalline phase. The definition
of the surface tension, asexpressed in Equation 13-1, may be used to find the
214
contribution from the surface tension to the change in Gibbs energy when a
crystal forms. Assuming a spherical crystal of radius r, this contribution
equals:
.::lG
o
=47fr
2
0
(14-23)
Asthe volume of the crystal equals 4/3 7fr
3
, the corresponding change in
chemical potentialbecomes:
47fr
2
0
.::lJ l..o =-- V =30 V/r (14-24)
4/37fr
3
where V is the molar volume of the crystal.
The contribution from the surface tension to the Gibbs energy will, in the
absence of other solid particles, cause supercooling. If the temperature is
lowered, the size of the spontaneously formed crystalline structures in-
creases, and at some stage crystallization takes place. The smallest radius
where the formed crystals are stable is called the critical radius, r".
Hansen et al. (1988) have modified Equation 14-2 to include the contri-
bution from the surface tension:
s
-'2-- =-'2-- +In s- .::lH{(1- T/TD/RT +30 VI(r*RT)
RT RT
(14-25)
In this equation, all activity coefficients in the solid phase are assumed
equal to unity, i.e., the solid phase is assumed to be ideal. The cloud point
may now be located by inserting into Equation 14-4 the expression for J I..L
given in Equation 14-19 and the expression for JI..s give~in Equation 14-25.
The critical radius, r", is assumed equal to 10-
8
m, and the surface tension
of the wax particle is found from Equation 13-15. In the limit, as 1/-+00, o
(in Equation 13-15) approaches A, and in the work of Hansen et al. o is as-
sumed equal to the paraffinic value of A (= ci, i~Equation 13-16).
WAX INHIBITORS
A variety of different additives, usually polymeric, have been shown to
modify the wax crystal size and crystal structure in a way that lowers the
pour point of the oil (the temperature where the oil .may no longer be
poured). Materials that have this effect include polyalkyl acrylates and
methacrylates, low molecular weight polyethylene waxes, and ethyl-vinyl
acetate (EVA) copolymers. The EVA copolymers are prabably the most com-
monly used of these wax inhibitors. According to Gilby (1983), the main
structural features which influence their performance as wax inhibitors are:
Wax Formation and lnhibition 215
l. Ratio of methylene to vinyl acetate (usually expressed as the vinyl ace-
tate content).
2. Average molecular weight.
3. Hydrocarbon chain branching.
Vinyl Acetate Contento As the vinyl acetate content increases, three impor-
tant trends can be identified:
The degree of crystallinity is reduced
The polarity of the polymer increases
At very high levels of vinyl acetate, chain stiffness is increased
The EVA copolymers used as wax inhibitors are low crystalline compounds
usually having an EVA content between 20 and 40 weight percent.
Molecular Weight. The solubility of EVA copolymers in oil mixtures de-
creases with the molecular weight of the copolymer. On the other hand, the
high molecular weight copolymers are those exhibiting the greatest effect as
wax inhibitors. The EVA copolymers used as wax inhibitors usually have an
average molecular weight between 10,000 and 30,000 (the heaviest oil mix-
ture constituents have molecular weights of the order 1,200).
Hydrocarbon Chain Branching. EVA copolymers having low levels of hy-
drocarbon chain branching are usually more effective aswax inhibitors than
EVA copolymers of similar vinyl acetate content and average molecular
weight, but containing higher levels of hydrocarbon branches.
The EVA copolymers include in their structure polymethylene segments
identical with those of the molecules forming the wax phase. When, prior to
the wax formation, an ordering of the paraffinic molecules takes place in the
liquid phase, the paraffinic branches of the EVA copolymers may enter these
pre-crystalline structures. This is illustrated in Figure 14-2. The vinyl ace-
tate groups of the EVA copolymer are very unlike the paraffinc branches,
and disturb further structuring of the paraffinic molecules. This is the rea-
son why the EVA molecules act as pour point depressants.
Table 14-3 shows the effect on the pour point of adding various amounts
of EVA copolymer to various crude oils. The pour point can be lowered by
approximately 25C, by adding between 25 and 100ppm EVA copolymer to
the crude oils. The addition of 200 ppm of the inhibitor, on the other hand,
increases the pour point.
It is not absolutely clear frorn the literature whether the wax inhibitors
mentioned depress only the pour point, or whether they also depress the
cloud point.
216
Properties 01Oils and Natural Gases
I I
CH2CH2CH2CH2CH2CH2CH2
~/ \ / \ / V\ / \ / \ / \
I CH2CH2CHz CHz CHz CHz CH2 I
I I
J \?\~\/\A/\?\/H2 /\/\/\/\A
CH CH2CH2CH2CH2CH2CHz ~2 CH2CH2CH2CH
/ I i I
\ ~\/\/\J \/\A/\/2 \
co I CH2CH2CH2CH2CH2CH2CH2 I c=o
/ I I /
CH3 I I C
H
3
I I
I I
AMORPHOUS I CRYSTALLINE REGIONS I AMORPHOUS
I
.1
I
REGlON
REGlON
Figure 14-2. The wax inhibition mechanism.
Table 14-3
Effect of Addltlon of EVA-Polymer on the Pour Point of Varlous Crude Oils
Amount of
Pour Point
on No.
EVA (ppm)
'C
1
O
30
1
100
9
2
O
8
2
25
- 18
3
O
- 1
3
25
- 26
4
O
21
4
200
25
EVA characteristics: VAcontent: 28%
Molecular weight: 13,000
Hydrocarbon branches per 1,000 carbon atoms: 9
No details have been given about the oil mixtures.
Source: Gilby (1983).
Wax Formation and Inhibition 217
COMPARISONS BETWEEN CALCULATED AND EXPERIMENTAL
RESULTS
In Tables 14-4 and 14-5, molar compositions are shown for two oil mix-
tures for which the cloud points have been measured at atmospheric pres-
sure. The results are given in Table 14-6. In addition, the cloud points have
been calculated, using the model of Won (1985, 1986), and using the model
of Hansen et al. (1988). (Both models are described in the Computation of
Wax Formation section). The mixtures were characterized using the proce-
dure of Pedersen et al. as described in Chapter 7. When using the model of
Won, the liquid-phase solubility parameters given in Table 14-2 were extra-
polated to components heavier than C40, using the following formula:
o~=7.971 +0:36132 In MW (14-26)
The calculated results are given in Table 14-6. The cloud points are much
too high when calculated using Won's procedure. On the other hand, very
Table 14-4
Molar Compositlon of Oil Mlxture 1 for Measured Cloud Point Given in
Table 14-6
Density
Molecular (g/cm
3
) at
Component Mole % Welght 15'C, 1 atm
C.
1.47 58
C5 17.73 72
e. 6.01 86
e7 13.97 93 .743
e. 10.74 107 .753
e9 8.65 120 .776
elO 5.90 134 .790
e"
4.72 147 .795
c., 3.96 161 .807
c., 3.86 174 .823
c., 3.39 188 .831
c., 2.92 202 .836
e, . 2.05 217 .843
en 2.25 234 .840
c., 1.62 251 .845
C'9
1.36 259 .856
C2()+
9.41 379 .890
Table 14-5
Molar Composition ot Oil Mixture 2 tor Measured Cloud Point Given in
Table 14-6
(
(
Density
Molecular 15C, 1 atm
Component Mole % Weight (g/cc)
e4 .11 58
es 3.01 72
e6 5.06 86
e7 4.55 93 .741
ea 7.21 104 .762
e9 6.02 121 .780
e1Q
3.45 134 .795
en 4.90 149 .801
c., 3.61 164 .815
c., 4.56 176 .828
e"
4.06 188 .838
c., 4.71 203 .845
c., 2.96 214 .853
c., 3.78 232 .848
c., 3.23 248 .851
c., 3.85 259 .861
e20+ 34.92 446 .922
Table 14-6
Measured and Calculated Cloud Points at Atmospheric Pressure
(The Molar Compositions 01the Oil Mixtures are Given in Tables 14-4 and
14-5)
Calculated Cloud Point (OC)
Measured
Cloud Point Model of Won Model of Hansen
Mixture
.
et al.
No. (OC)
Twax
. 1T
Twax
. 1T
1 16 56 49 17 1
2 25 83 58 24 - 1
IIT =Calculated cloud point-measured cloud point.
accurate results are obtained with the procedure of Hansen et al. Observe
that the experimental data of Table 14-6were included in the data base used
when the values of the interaction parameters (Equations 14-20 through
14-22) were estimated. Due to lack of experimental data it is not possible to
check the predictive capabilities of the model by Hansen et al.
Wax Formation and Inhibition
219
REFERENCES
Flory, P. J ., Principies 01Polymer Chemistry, Cornel! University Press, New
York, 1953.
eilby, e. W., "The Useof Ethylene-Vinyl Acetate Copolymers as Flow Im-
provers in Waxy Crude Oil," Chemicals in the Oil Industry, special publ-
cation, Proceedings, Manchester, 1983.
Hansen, J . H., Hanningssn, H. P., Pedersen, K. S., and Fredenslund, Aa.,
"A Thermodynamic Model for Predicting Wax Formation in Crude Oils,'
AIChE i.. 1988.
Mandelkern, L., Crystallization 01 Polymers, Mceraw-Hill, New York,
1964.
Prausnitz, J . M., Lichtenthaler, R. N., and de Azevedo, E. e., Molecular
Thermodynamics 01 Fluid-Phase Equilibria, 2nd ed., Prentice-Hal!,
1986.
Won, K. W., "Continuous Thermodynamics for Solid-Liquid Equilibria:
Wax Formation from Heavy Hydrocarbon Mixtures," Paper 27A pre-
sented at AIChE Spring National Meeting, Houston, TX, March 26,
1985.
Won, K. W., "Thermodynamics for Solid-Liquid- Vapor Equilibria: Wax
Phase Formation from Heavy Hydrocarbon Mixtures," Fluid Phase Equi-
libria 30, 1986, pp. 265-279.
Chapter 15
Gas Hydrates
Gas hydrates are solid, semi-stable compounds which sometimes have
been found plugging natural gas transmission pipelines and other gas han-
dling equipment.
Gas hydrates consist of geometric lattices of water molecules containing
cavities occupied by light hydrocarbons, or other light gaseous compounds
(e.g., nitrogen or carbon dioxide). Gas hydrates may be formed where the
mentioned gaseous compounds are in contact with water at temperatures
below approximately 35C. Gas hydrates may cause plugging of not only
pipes, but also nozzles, and separation equipment. H istherefore important
that the gas hydrate formation conditions are taken into consideration when
designing separation plants and pipelines for use in the oil and gas industry.
Figure 15-1 illustrates the phase behavior of apure hydrate forming gas.
The fully drawn curves are three phase loci. Line AB represents equilibria
between hydrate (H), gas in gaseous form (G), and ice (1). Line BC repre-
sents equilibria between hydrate (H), gas in gaseous form (G), and liquid
water (L1). Line CD represents equilibria between hydrate (H), gas in liq-
uid form (L2), and water in liquid form (L1).
For a mixed gas, the situation may become slightly more complicated,
due to the possible intersection of the hydrate curve with the phase envelope
of the gas mixture. This isillustrated in Figure 15-2. The lines AB, BC, and
DE correspond to the lines AB, BC, and CD in Figure 15-1. At the line CD
in Figure 15-2, hydrate (H), gas in gaseous form (G), gas in liquid form
(L2), and water in liquid form (Ll) are in equilibrium.
TYPES OF HYDRATES
Water can form two different types of hydrate lattices, which are referred
to as structure 1and structure I1, respectively. Both types of lattice contain
numerous small and large cavities. In a stable hydrate, some of the cavities
are occupied by gaseous compounds called guest molecules. Some physical
constants for the two different types of hydrate lattices are given in Table
15-1.
, "",""
> \~
~
,'1i
221
Gas Hydrates
1;
\ ~
;
LN(P)
.
I
I
I
.
I. L---
_-------7--
1f
F
--- .
/
/
/
H + L2
L1 + L2
H + G
.
T
273.15
Figure 15-1. Coexistence curve for hydrate (H), gas (G), liquid hydrate forming
compound (L2), ice (1), and liquid aqueous solution (L1) in the case of apure hy-
drate forming compound (- ). Coexistence curve when the hydrate forma-
tion is inhibited by methanol (_. -' -l. Vapor pressure of the pure hydrate forming
compound (----------).
LN(P)
L1 + L2
H + L2
-----.~...
'o
\
G+L1+L2 1
I
/
/
/
/
,/
.'-c
,./'/
...../" " . I G+L1
.-
.-
.-
.-
H + G + L2
_..-
273.15
Figure 15-2. Coexistence curve for hydrate (H), gas (G), liquid hydrate forming
compounds (L2), ice (1), and liquid aqueous solution (L1) in the case with a rnlx-
ture of hydrate forming compounds ( ). Phase envelope for the hydrate
forming compounds (----------). Critical point (o).
222 Propertes o] Oils and Natural Gases
Only gases of rather rnodest size and appropriate geornetry can enter the
cavities. Table 15-2indicates which of the gaseous cornpounds found in typi-
cal petroleum mixtures may enter hydrate cavities. The cavities may contain
one kind of gaseous compounds only, or they rnay contain gas rnolecules of
different chemical species. Not all cavities need to be filled in a stable hy-
drate.
Table 15-1
Some Physical Constants of the Two Possible Water Structures in
Hydrates
Property Structure I Structure 11
Number 01H20
molecules per unit cel! 46 136
Number 01smal!
cavities per unit cel! 2 16
Number 01large
cavities per unit cell 6 8
o
Cavity diameter/A
Small 7.95 7.82
Large 8.60 9.46
Table 15-2
Gases of Interest to Natural Gas Hydrate Formation, and Their
Occurrence in the Different Cavities of the Water Structure
Structure I Structure 11
Small Large Small Large
Compound cavities cavities cavities cavities
Methane + + + +
Ethane - + - +
Propane - - - +
Butane - - - +
Isobutane - - - +
Carbon
dioxide + + + +
Nitrogen + + + +
Hydrogen
sullide + + + +
/Ir' /
I
~:t\
~"
\;'i
Gas Hydrates 22:l,
HYDRATE COMPUTATION
Hydrates are formed when the hydrate state is energetically favorable,
compared to a pure water state (water in ice, liquid, or vapor forms). The
transformation from apure water state to ahydrate state can be regarded as
consisting of two steps:
pure water (a) - empty hydrate lattice ({3)- filled hydrate lattice (H)
where a, {3,and H are used to identify each of the three states considered.
The {3-stateis purely hypothetical, and only considered to facilitate the hy-
drate calculations. Whieh state is energetically favorable depends on which
state has the lowest chemical potential. The difference between the chemi-
cal potential of water in the hydrate state (H), and in apure water state (a)
can be expressed as:
.tH - .t0 = (.tH - .ti3) +(.ti3 - .t0)
(15-1)
The first term on the right side (.tH - .t(3) can be regarded as the stabiliz-
ing effect on the hydrate lattice, caused by the adsorption of the gas mole-
cules. This latter effect depends on the tendency of the gas molecules to en-
ter the cavities in the hydrate lattice. van der Waals and Platteeuw (1959)
have proposed expressing this tendency using asimple gas adsorption model.
The difference between the chernical potential of water in the empty hy-
drate lattice and the filled one is expressed as follows:
(.tH - .ti3) =RT ~ ViIn (1 - ~ YKi)
i = 1, 2, ... , NCAV
K =1, 2, "', NCOMP
(15-2)
where Viis the number of cavities of type i, and YKidenotes the probability
that acavity of type i isoccupied by a gas molecules of type K. NCAV isthe
number of cavities per unit cell in the hydrate lattice (seeTable 15-1), and
NCOMP isthe number of gaseous cornpounds present that may enter acav-
ity in the hydrate lattice. The probability YKi is calculated using the
Langmuir adsorption theory:
yKi=CKJ K+ +~ C) (15-3)
j =1,2, ... , NCOMP
where fK is the fugacity of component K. CKi s a temperature-dependent
adsorption constant, specific for the cavity of type (i), and for the gas corn-
ponent (K). The adsorption constant accounts for the water-gas interactions
in the hydrate lattice. In the work of van der Waals and Platteeuw and sev-
erallater works (e.g., Parrish and Prausnitz, 1972, Anderson and Prausnitz,
1986), the adsorption constant C iscalculated from a cell potential w (r):
C(T) =4 ll/kT J ; exp ( - w(r)/kT)r
2
dr
(15-4)
In this equation k isBoltzmann's constant, T isthe absolute temperatura,
and r is the distance from the lattice wall (the water molecule) to the guest
(gas) molecule. The lower the cell potential in a given distance, r, fromthe
lattice wall, the more likely it is to find a guest molecule at that position.
The cell potential depends on the sizeof the molecule, and the nature of the
gas-water interaction. The latter is expressedin terms of acharacteristic en-
ergy, and a characteristic distance (the distance Frornthe lattice wall in
which thecell potential becomes zero). The distance and energy parameters
can befound from experimental hydrate-formation data by minimizing the
deviation between experimental and calculatated hydrate-formation condi-
tions.
A somewhat simpler approach for calculating C is to use the following
temperature dependence (Parrish and Prausnitz, 1972, and Munck et al.,
1988):
CKi =(AKJT) exp (BKJT)
(15-5)
For each compound (K) capable of entering acavity of type (i), A
Ki
and
BKi must bedetermined from experimental data. The A and Bvalues recom-
mended by Munck et al. are' given in Table 15-3.
The fugacity of component K, fK, may be determined by aflash calcula-
tion, asexplained inChapter 6, using theSRK equation of state asdescribed
in Chapter 5.
The second term on the right sideof Equation 15-1, (.ti3 - .ta), isequal to
the difference between the chemical potentials of water in the empty hy-
drate lattice (the fl-state), and water in the form of liquid or ice (the a-
state). Anexpressionfor this difference inchemical potentials can bederived
using the following thermodynamic relation:
d( : ~) =
_ dH dT +dV dP
RT2 RT
(15-6)
Gas Hydrates 225
Table 15-3
A and B Parameters for Calculating Langmuir Constants of Equation 15-5
Small cavity Large cavity
A x 10
3
B A X 10
3
B
Gas Structure (Klatm) (K) (K/atm) (K)
C, 0.7228 3187 23.35 2653
0.2207 3453 100.0 1916
C2 0.0 0.0 3.039 3861
0.0 0.0 240.0 2967
C3
.
0.0 0.0 5.455 4638
-C. 0.0 0.0 189.3 3800
C. 0.0 0.0 30.51 3699
N2
..
1.617 2905 6.078 2431
0.1742 3082 18.00 1728
CO2 I 0.2474 3410 42.46 2813
11 0.0845 3615 851.0 2025
H2S I 0.0250 4568 16.34 3737
11 0.0298 4878 87.2 2633
Source: Munck el al. (1988).
where R isthe gas constant, and dH and dV are the changes in molar en-
thalpy and molar volume associated with the transition. The following ex-
pression may beobtained for the difference between thechemical potentials
of water inthe fl and inthe o-states at the temperature, T, and the pressure,
P:
.ti3- .ta =d.t(T,P)
RT RT
rdH dT +r dV dP
r, RT2 Po RT
d.t(To,Po)
RT
(15-7)
where (To, Po) indicates a reference state where d.t is known. In Equation
15-7 it isassumed that dH isindependent of pressure, and the temperature
dependence of the second term on the right sideof Equation 15-6 has been
approximated by the average temperature:
T =(T +To)/2 (15-8)
If thereference pressure, Po, ischosenequal to zero, Equation 15-7can be
rewritten to:
226 Properties oj Ols and Natural Gases
{t!3- p." =~{t(T,P) =~{t(To,P =O) _ J T ~H dT +P ~~
RT RT RT 'r, RT2 . RT
(15-9)
~H may becalculated from the difference, ~Cp, in the molar heat capacities
of the (3 and the o-states:
T
~H(T) =J Cp(T)' dT
'r,
(15-10)
The constants needed in the calculation of ~{tof the (3 - a transition at a
given temperature and pressure are given in Table 15-4.
Table 15-4
Physical Constants Used for Evaluating Equation 15-9
Property Unit Structure I Structure 11
t>/4o(liq) J /mol 1264 883
t>Ho(liq) J /mol - 4858 - 5201
t>Ho(ice) J /mol 1151 808
t>Vo(liq) cm
3
/mol 4.6 5.0
t>Vo(ice) cm
3
/mol 3.0
3.4
t>Cp(liq) J /mol/K 39.16 39.16
Sources: Erickson (1983). Munck el al. (1988).
Using this procedure, the difference in chemical potentials ({tH- p.") be-
tween water in a hydrate state, and in a pure water state, snay now be cal-
culated. In Figures 15-1 and 15-2 the unbroken lines indicate corresponding
values of temperature and pressure where:
{tH- {t" =O (15-11)
At those conditions, the hydrate state and the pure water state are equally
favorable. To the left of the unbroken lines:
p.H - p." <O (15-12)
and the water will, at equilibrium, be in a hydrate formo Whether this is a
structure I or a structure 11hydrate depends on which of the two structures
has the lowest chemical potential, in the presence of the actual gas cornpo-
nents as potential guest molecules. To the right of the unbroken lines:
{tH- {t" > O (15-13)
Gas Hydrates 227
i.e., at equilibrium at those conditions, no hydrate can exist, and the water
will be in the form of either liquid or ice.
Hydrates are not always formed from acondensed phase. The water mol-
ecules of a gas phase may also form hydrate lattices. The hydrate formation
process may, in that case, be regarded as consisting of the following succes-
sive steps:
water in gaseous form (G) - water in condensed form (a) -
empty hydrate lattice ({3) - filled hydrate lattice (H)
The difference between the chemical potential of water in a hydrate
phase and in a gas phase may be expressed as follows:
{tH- p'e =({tH- {t") - ({t" - {te)
=({tH- {t(3) - ({t/3 - p.") - ({t" - {te)
=({tH- {t(3) - ({t/3 - {te) (15-14)
In this equation ({tH- {t(3) is known from Equation 15-2 and ({t!3 - p.,,)
from Equation 15-9. The problem that remains is, in arder to calcula te
({t!3 - {te) and thereby {tH- {te, finding an expression for ({t'" - {te). The dif-
ference between the chemical potential of water in the o-state and in gas-
eous form may be found from the following expression:
({t" - {te) =RT(ln f~- In f~) =RT In (f~If~) (15-15)
where f~is the fugacity of water in the gas phase, and f~the fugacity of
water in the o-phase. The fugacity of water in the gas phase may be ob-
tained using an equation of state (e.g., the SRK-equation which s described
in Chapter 5):
f~=4>wywP (15-16)
P s the total pressure, and 4>w and Yw are the fugacity coefficient and the
mole fraction, respectively, of water in the gas phase. The fugacity of water
in apure liquid form is calculated from the equation:
v"'P
f~=p~t exp RT
(15-17)
p~t isthe saturation pressure of water, and may be found from the equation:
In p~t =a +bIT (15-18)
a and b may be estimated from experimental data. Munck et al. have pro-
posed to use the following values:
Liquid water: a =17.372;
Ice: a =14.484;
b = - 6141 K
b = - 5351 K
where p~t is in atm. In Equation 15-17, Va is the molar volume of water
(Va =19.6 cm" for ice, and 18.0 cm" for liquid water). Equation 15-15 may
now be modified to:
," - ,G - RT In [~wYwP/(l"':'"'p~;)l (15-19)
By combining Equation 15-9 and Equation 15-19, the following expression
may be derived for the difference between the chemical potential of water
in the {3-state, and in a gaseous state:
('";/) - +WYW+':' exp ~;:)l
+~,t(T0' P =O) _ J T ~H dT +P ~~
RT To RT2 RT
(15-20)
The fugacity coefficient <Pwof water in the gas phase may be calculated us-
ing the SRK-equation (fugacity expression given in Equation 5-29).
By combining Equations 15-2 and 15-20, it ispossible to evaluate whether
hydrate formation in the gas phase is thermodynamically favorable or not,
i.e., whether (,tH- ,tG)has a positive or a negative value.
HYDRATE INHIBITORS
The hydrate-formation curve may beshifted towards lower temperatures
(the dashed-dotted Iines in Figure 15-1) by addition of a hydrate inhibitor.
Commonly used hydrate inhibitors are methanol, ethanol, glycols, sodium
chloride, and calcium chloride. The formula of Hammerschmidt (1969)
presents a simple way to calculate the decrease, ~T, in temperature during
the formation of a gas hydrate. It has the following form:
Ks
~T =M(lOO _ s)
(15-21)
Gas Hydrates 229
where ~T = C
K =constant that depends on the nature of the inhibitor and on
the heat of formation of the hydrate
s =concentration of the inhibitor in weight percent in the aque-
ous phase
M =molecular weight of the inhibitor divided by the molecular
weight of water
Values of the constant, K, for some commonly used inhibitors are given in
Table 15-5.
Table 15-5
Values tor Various Inhibitors ot the Constant, K, Entering the Formula
Proposed by Hammerschmidt (Equation 15-21) tor Calculating Inhibitor
Ettect
(
Inhibitor K
Methanol 2335
Ethanol 2335
Ethylene glycol 2700
Diethylene glycol 4000
Triethylene glycol 5400
Source: Hammerschmidt (1969).
A more advanced procedure in the calculation of the inhibitor effect is to
use an activity coefficient model to calculate the influence of the inhibitor
on the chemical potential of water in the o-state. When inhibitors are con-
sidered, an additional term must be included in Equation 15-9, which then
becomes:
~,t(T, P) _ In (x ) =~,t(To, P =O) _ J T ~H dT +P S,! (15-22)
RT w'Yw RT r, RT2 RT
where x and I'ware, respectively, the molar concentration, and the activity
coefficient of water in the water rich phase. ~,tstill stands for the difference
in chemical potential between the hydrate state and the pure water state.
The activity coefficient and the fugacity of water are interrelated as follows:
I'w=fw(T, P, xI. X2,... , xw)/(f;:'(T,Pix; (15-23)
where fw=fugacity of water at the actual conditions
Xl, X2,etc. =molar inhibitor concentrations in the water rich phase
f;:'=fugacity of pure liquid water at the same temperature
and pressure
230
Properties oi Oils and Natural Gases
The aetivity coefficent of water may be ealeulated using theJ J NIQUAC ae-
tivity eoeffieient model (Abrams and Prausnitz, 1975). This approaeh has
been used by Anderson and Prausnitz (1986) and by Munek et al. (1988). In
the latter work the modified UNIQUAC-model of Sander et al. (1986) is
used to model the inhibition eHeet of eleetrolytes (e.g., NaCI and CaCI
2
).
Figure 15-3 shows experimental results for the deeomposition eonditions
of hydrates formed from aseven-eotnponent gas mixture (Ng and Robinson,
1984) whose molar eomposition isgiven in Table 15-6. Alsoshown are calcu-
lated results using the model of Munek et al. (1988). Figures 15-4 and 15-5
show experimental hydrate results for a gas eondensate mixture, and an oil
mixture, respeetively. The molar eompositions are given in Table 15-7. Also
shown in Figure 15-4 are experimental observations of methanol's influenee
as an inhibitor on the hydrate formation from the gas eondensate mixture.
Figure 15-5 shows the results of the hydrate formation from the oil mixture
E
...,
<O
50~
Exper i ment al dat a l.
poi nt
- - ea 1cu 1at ed hydr at e
cur ves
- - . - - . Phase envel ope
01 wt
MEOH
35~
200.0
150.0
~ 100.0
=>
~ 1---'--
~ -- ---
e,
'-
<,
_"
\
o)
50.0
o.o~-=-=~
230.0 I e
I 1 I
240.0 250. o 260. o 270. o 280. o
Temper at ur e, [ Kel vi nl
290.0 300.0
Figure 15-3. ealculated hydrate curve for a seven component gas mixture.
Methanol (MEOH) is used as inhibitor (Munck et al., 1988). Experimental data
points. (Ng and Robinson, 1984).
Gas Hydrates 231
having been inhibited by ethylene glyeol (EG). Also shown in Figures 15-4
and 15-5are results ealculated using the model of Munek et al. Figures 15-3
through 15-5show that the hydrate formation temperatures where no inhib-
itor is present are predieted to within O.5C. The aeeuraey is apparently
Table 15-6
Molar Composition of the Seven Component Mixture for Hydrate
Formation Results Shown in Figure 15-3
Component Mole %
N2 5.26
CO2 13.37
C, 73.90
C2 3.85
C3 2.02
n-C. 0.80
nc, 0.80
Source: Ng and Robinson (1984).
300.0
Exper i ment al dat a poi nt
__ Cal cul at ed hydr at e cur ves
250.0
200.0
E
...,
<O
.;; '50.0
1-
~
V>
'"
1-
e,
100.0
50.0
t===;;; ---=== ~I 1 I 0.0
260.0 270.0 280.0
Temper at ur e, [ Kel vi nl
290.0 300.0
Figure 15-4. Experimental (Ng et al., 1987) and calculated (Munck et al., 1988)
results for the hydrate formation conditions for the gas condensate mixture of Ta-
ble 15-7. Methanol (MEOH) is used as inhibitor.
300.0
O.U EG
250.0
. : Exper i ment al hydr at e poi nt ,
. : Exper i ment al bubbl e poi nt
Cal cul at ed hydr at e cur ve
Cal cul at ed bubbl e poi nt cur ve
200.0
E
..,
'"
~ 150.0
'"
s,
::l
'"
'"
'" k100.0
50.0
275.0 290.0 295.0
Temper at ur e, ,
[ Kel vi nl
300.0
0.0
285.0 270.0 280.0
Figure 15-5. Experimental (Ng et al., 1987) and calculated (Munck et al., 1988)
results for the hydrate formation conditions for the oil mixture of Table 15-7. Ethyl-
ene glycol (EG) is used as inhibitor.
Table 15-7
Compositions of the Gas Condensate Mixture for Hydrate Results Shown
In Figure 15-4 and of the 011Mixture for Hydrate Results Shown in Figure
15-5
Gas
Condensate
Component Mole % 011Mole %
N2 0.64 0.16
CO2 3.11 2.10
C, 73.03 26.19
C2 8.04 8.27
C3 4.28 7.50
te, 0.73 1.83
n-C. 1.50 4.05
ic, 0.54 1.85
n-C5 0.60 2.45
C6+ 7.53 45.60
MW 32.4 90.2
Sour ce: Ng el al . ( 1987) .
I
I
Gas Hydrates 233
maintained after addition of considerable amounts of methanol to the water
phase. The hydrate formation temperatures from a gas condensate mixture
in the presence of EG ispredicted to within 1C, when the water phase con-
tains 20 weight percent EG.
REFERENCES
Abrams, D. S. and Prausnitz, J . M., "Statistical Thermodynamics of Liquid
Mixtures: A New Expression for the Excess Gibbs Energy of Partly or
Completely Miscible System," AIChE l., 21, 1975, pp. 116-128.
Anderson, F..E. and Prausnitz, J . M., "Inhibition of Gas Hydrates by Meth-
anol," AIChE t., 32, 1986, pp. 1329-1333.
Erickson, D. D., "Development of aNatural Gas Hydrate Prediction Com-
puter Prograrn," M.Sc. Thesis, Colorado School of Mines, 1983.
Hammerschmidt, E. G., "Possble Technical Control of Hydrate Formation
in Natural Gas Pipelnes," Brennstoff-Chemie, 50, 1969, pp. 117-123.
Ng, H.-J . and Robinson, D. B., AIChE National Meeting, Atlanta, Geor-
gia, 1984.
Ng, H.-J ., Chen, C.-J ., and Saeterstad, T., "Hydrate Formation and Inhibi-
tion in Gas Condensate and Hydrocarbon Liquid Systems," paper pre-
sented at the AIChE Meeting, Houston, Texas, April 1987.
Parrish, W. R. and Prausnitz, J . M., "Dssociation Pressures of Gas Hydrates
Formed by Gas Mxtures," I&EC Process Des. Dev., 11, 1972, pp. 26-35.
Sander, B., Fredenslund, Aa., and Rasmussen, F., "Calculation of Vapor-
Liquid Equilibria in Mixed Solvent/Salt Systems Using an Extended UN-
IQUAC Equation," Chem. Eng. Sci., 1986, pp. 1171-1183.
Munck, J ., Skjold-jergensen, S., and Rasmussen, F., "Computations of the
Formation of Gas Hydrates," Chem. Eng. Sci., in press 1988.
van der Waals, J . H. and Platteuw, J . C., "Clathrate Solutions," Adv. Chem.
Phys., 2, 1959, pp. 1-57.
Simulation oj Miscible Gas lnjection 235
PROCESS DESCRIPTION
weight components will be in contact with original reservoir oil. With the
increasing content of intermediate molecular weight components in the gas
phase, miscibility may be achieved even though this was not possible with
the original injection gas.
In case of a condensing gas drive, the original reservoir fluid and the in-
jection gas are not miscible, but miscibility may be achieved at alater time,
at the injection well. The reservoir fluid takes up intermediate molecular
weight components from the injection gas. Since the liquid phase has a
lower mobility than the gas phase, liquid which has taken up intermediate
molecular weight components will besought and contacted by the injection
gas. Further transference of intermediate molecular weight components
from the gas to the liquid phase may then occur. After some time, miscibility
may be obtained between the liquid phase and the injection gas.
Ternary diagrams like the one shown in Figure 16-1 are often used to il-
lustrate the mixing processes at a given temperature and pressure. The corn-
position of apetroleum mixture is represented using three groups of compo-
Chapter 16
Simulation of Miscible Gas Injection
Reinjecting a gas produced from an oil field into the reservoir is a com-
mon practice. While previously this was mainly regarded as a convenient
way of getting rid of the gas, much interest is now being devoted to gas in-
jection as an enhanced oil recovery (EOR) process. Naturally, reinjection of
gas reduces the pressure drop associated with producing oil from afield over
aperiod of time. However, the injected gas also influences the oillgas phase
equilibrium compositions in the reservoir. The gas may selectively take up
components from the oil phase (vaporizing gas drive), the oil may take up
components from the gas phase (condensing gas drive), or the gas and the oil
may exhibit first contact miscibility. This means that only one phase is
achieved, no matter in what proportion the oil and the gas are mixed. A
miscible gas displacement of the oil isadvantageous, because valuable heavy
components will at the end be contained in aphase with afairly high mobil-
ity. The mobility isinversely proportional to the viscosity, and the viscosity is
reduced when the oil takes up gaseous components. The status of- miscible
displacement has been extensively reviewed by Stalkup (1984).
Miscibility may be achieved as a result of:
l. First contact miscibility
2. A vaporizing gas drive
3. A condensing gas drive
First contact miscibility isfound when the original reservoir fluid and the
injection gas are miscible. This means that all possible mixtures of the two
compositions are single-phase fluids at the specified conditions.
With avaporizing gas drive, miscibility isachieved at some distance from
the injection well. At the injection well, both a gas and a liquid phase are
present at equilibrium. Compared to the injection gas, the equilibrium gas
phase contains more intermediate molecular weight components, which
have been extracted from the oil phase. The gas phase has ahigher mobility
than the oil phase. Hence, gas which has taken up intermediate molecular
Critical Tie-Line
Critical Mixture
""cOmpOSitiOn Ffath
Between A ancl B
C7
Molel e7+ C
2
-C
6
Figure 16-1. Ternary diagram. Mixture A and Mixture B are miscible by first con-
tact.
234
Simulation 01Miscible Gas Injecton 237
nents, C +CO2 +N2, C2-C6, and C7+. Each corner of the diagram
represents 100% of a given group of components. Any concentration be-
tween O and 100% is represented at a proportionate distance between the
bottom of the triangle and the opposite corner. Point A in Figure 16-1corre-
sponds to an oil mixture, and point B to a gas mixture. Also indicated in
Figure 16-1 is the location of the two-phase area. The dashed lines are the
tie-lines, connecting gas and liquid phase compositions that are in thermo-
dynamic equilibrium. The tangent to the two-phase area in the critical
point iscalled the critical tie-line. The locations of the points corresponding
to the reservoir Huid composition and the injection gas, relative to the crit-
cal tie-Iine, determine whether miscibility can be achieved.
One may distinguish between the following cases:
l. The injection gas and the reservoir oil compositions are both located
on the two-phase side of the critical tie-line: Miscibilty cannot be
achieved.
2. The composition of the original reservoir Huid is located on the two-
phase side of the critical tie-Iine, and the composition of the injection
gas ison the other side: Condensing gas drive ispossible, and miscibl-
ity (multi-contact) can be achieved.
3. The cornpositon of the injection gas islocated on the two-phase side of
the critical te-line, and the composition of the original oil is on the
other side: Vaporizing gas drive ispossible, and miscibility (multi-con-
tact) can be achieved.
4. The composition of the original oil and the injection gas are both 10-
cated on the sngle-phase side of the critical tie-lne: The oil and the
gas exhibit first contact miscibility.
C, + COz + NZ
New Gas Compositions
Critica! He-Une
..
011
C
z
- C
6
e7+
Figure 16-2. The principies of a vaporizing gas drive iIIustrated in a ternary dia-
gramo
Compositions A and B in Figure 16-1 are both located on the single-phase
side of the critical tie-line and are, therefore, miscible by first contacto Misci-
bility between the oil and the injection gas of Figure 16-2 may be achieved
as a result of a vaporizing process. When the original oil is mixed with an
appropriate amount of injection gas, two phases are formed. The equilb-
rium phases can be represented as two points on the borderline between the
single and two-phase areas in the ternary diagram. The two points are con-
nected by a tie-Ine in Figure 16-2. The new gas composition is brought in
contact with original reservoir oil, and a new gas phase is formed, which
has a composition more like the composition at the critcal point. At some
stage, miscibility may be obtained between the original oil and the gas. The
mechanisms of a condensing drive are illustrated in Figure 16-3. The oil
takes up intermediate molecular weight components from the gas phase,
and the composition of the oil phase approaches that of the mixture at the
critical point. After some time, miscibility may be achieved between the oil
and the injection gas.
Ternary diagrams represent the phase equilibrium picture at afixed tem-
perature and pressure. Figure 16-4 illustrates qualitatively the development
in the two-phase area as a function of pressure. The size of the two-phase
area decreases with increasing pressure, meaning that the chances of achiev-
ing miscibility are better at ahigher pressure. The lowest pressure where, at
a fixed temperature, miscibility may be achieved between a given reservoir
Huid and a given injection gas is called the minimum miscibility pressure
(MMP). The lowest pressure where the reservoir oil and the injection gas are
miscible by first contact iscalled the first contact minimum miscibility pres-
sure (FCMP).
In reality, ternary diagrams can only give an approximate picture of the
processes that take place when gas isbeing injected into an oil reservoir. The
reason is that the reservoir Huid, and in many cases, also the injected gas,
are multi-component mixtures. The two-phase area shown in aternary di a-
gram is based on one specific distribution of pure components within the
groups C
I
+CO
2
+N
2
, C
2
-C
6
, and C7 +. The distribution of components in
C
l
+ cO
2
+ N
2
Cr-l t c al
He-Une
Composltions
Oi 1
C7+
C
2
- C
6
Figure 16-3. The principies of a condensing gas drive illustrated in a ternary dia-
gramo
the oil mixture may differ markedly from that in the injection gas. When an
oil ismixed with a gas of a different distribution, the distribution of compo-
nents within one group, e.g., C2-CS, will change. This behavior is further
explained in the section dealing with Cell-to-Cell Simulation.
CALCULATION OF THE MINIMUM MISCIBILITY PRESSURE
(MMP)
With a total of only three real components present in the reservoir Huid
and the injection gas, it ispossible to give aprecise definition of MMP, using
aternary diagram. The first contact miscibility pressure (FCMP) isthe low-
est pressure where a straight line between the reservoir fluid and the injec-
tion gas nowhere crosses the two-phase area. With avaporizing gas drive in
the three component case, the critical tie-line, at MMP, will pass through
the original oil composition. Similarly, with acondensing gas drive at MMP,
'1
,]{
Smulaton oj Mscible Gas lnieciion 239
J ncreasing Pres sure
C7+ C
2
-C
6
Figure 16-4. Development in the two-phase area of a petroleum mixture as a
function of pressure.
the critical te-line will pass through the injection gas composition. This def-
inition may, in principle, be extended to the multi-component case, but it is
then no longer possible to represent the results accurately in a ternary dia-
gram, due to the exchange of real components among the groups of compo-
nents representing the corners of the triangle.
Turek et al. (1986) suggested a procedure for calculating the MMP of a
multi-component mixture. The principle issketched in Figure 16-5for ava-
porizing mechanism. Liquid of composition L is initially mixed with gas of
composition G. Two phases are formed. The tie-line passes through the
point MI. As afirst approximation it isassumed that all the tie-lines are par-
allel. An estimate of the tie-line passing through the originalliquid composi-
tion L can, therefore, easily be established. From what isknown, the critical
tie-line at MMP passes through the point L. The line connecting the points L
and M2 must, therefore, at MMP be a good estimate of the location of the
critical tie-line. In reality the tie-lines are not exactly parallel, and an itera-
tive procedure is therefore needed to determine the relative locations of the
tie-line through L and the critical te-line. If they are coincident, the se-
lected pressure equals MMP. If the critical te-line is located to the right of
the tie-line through L, the selected pressure is lower than MMP, and vice
versa if the critical tie-line lies on the left side of the tie-Iine through L.
J ensen and Michelsen (1987) have shown that the generalization of the
ternary-dagram approach to the multi-eomponent case may in some cases
give unexpected results. If a te-lne passes through one of the original corn-
positions, the following relation must be fulfilled:
Z=ay +(1 - a)x i =1, 2, 3, ... , NCOMP (16-1)
where Z=mole fraction of component i in the original reservoir oil
for avaporizing gas drive, and the mole fraction of corn-
ponent i in the injection gas for a condensing gas drive
y, X=mole fractions of component i in the gas and liquid
phases in equilibrium
a = a parameter that is usually less than O or greater than 1
NCOMP =number of components
Equation 16-1 can be rewritten to:
x, =z/(l +a(K - 1)) i = 1, 2, 3, ..., NCOMP (16-2)
where the equilibrium factor K =y/x. J ensen and Michelsen considered a
vaporizing drive where the original reservoir fluid contains C but no CO2
The injection gas, on the other hand, consists of pure CO2. CO2 is given
G
Vapor
Figure 16-5. A schematic lllustra-
tion of a single step inthe MMP-
calculation procedure of Turek et
al. (1986).
Vapor
+
Liquid
Li quid
L
~
Simulation of Miscible Gas lniection 241
component number N, .e., ZNequals zero. The critical point can be re-
garded as a te-line of zero length. If at MMP the critical tie-line passes
through the composition of the original oil, Equations 16-1 and 16-2must be
fulfilled for Xand yequal to the mole fraction of component i in the critical
mixture. Since ZNequals zero, one of the following equations must also be
fulfilled:
XN=YN=O (16-3)
or
a = - lI(KN - 1) (16-4)
By combining Equations 16-2 and 16-4, the following relation is ob-
tained:

x =z/(l - (K - l)/(KN - 1)) (16-5)
The oil contains methane (M), which is more volatile than CO2 (N) and,
therefore, has a higher equilibrium ratio:
KM>KN (16-6)
XMcalculated from Equation 16-5 will be negative. This isobviously wrong,
and Equation 16-4 must therefore be rejectedl This means that Equation
16-3must befulfilled at MMP, i.e., at MMP no CO2 ispresent at the point in
the reservoir where miscibility is achieved.
Based on considerations quite analogous to those given, it can be shown
that a condensing gas drive at MMP does not contain any components not
present in the injection gas. With an injection gas consisting of a mixture of
Cl-C5, the miscible fluid at MMP will contain no components heavier than
C5, and this is obviously not the desired resulto
J ensen and Michelsen have shown results for other multi-cornponent mix-
tures in which the ternary diagram analogue gives results which are compa-
rable to what can be obtained with other procedures for determining the
minimum miscibility pressure.
CELL- TO-CELL SIMULATION
Gas injection into a petroleum reservoir can be simulated using a cell-to-
cell model. The calculation procedures were described by Metcalfe et al.
(1972). The program simulates cells of equal volumes in aseries, asshown in
CELL 1 CELL 2
GAS.lNQ 04ll GASIANO cu.t
I
BATCH 1
INJECTlOH GAS
'-
~ll
C;.t.SIANQOlll
I
BATCH 2
INJE" C,TlOH GAS
'-
uA!> ...u Vil. I

BATCH NN
INJECTlOH GAS
L-
CELL N
w[LL STROrI4
I
Wl:lL ST"[A"
I
I
I
I
I
I
I
I
I
I wEll STRQ14
Ir-
---
Figure 16-6. Principies of cell-to-cell calculation.
A STAGNANT Oll B. MOVING EXCESS Oll
GAS
GASTO
NEXT CEll
GAS IANO Oll) ~AS
h Oll
8;~Oll
FINAL CELl ORIGINAL CEll
CONomON CONOITION
ORIOINAl CEll
CONOITION
FINAL CEll
CONomON
C. Oll ANO OAS MOVED BY PHASE MOBIUTIES
GAS ANO Oll 10NEXT CELl
I
OAS (ANO Oll) ~AS
~
Oll
GAS ~
Oll
ORIGINAL CEll
CONOITION
Figure 16-7. Different criteria for moving excess fluid from one cell to the next
one.
J
l~
Simulation o] Miscible Gas lniection 243
Figure 16-6. The temperature and the pressure are the same in each cell,
and the volume is kept constant. Initially, all the cells contain the same
fluido A specified amount of gas isadded to cell 1. It isassumed that perfect
mixing occurs, and that thermodynamic equilibrium isreached. This means
that the conditions in the cell can be described by a (P,T) flash calculation
(as discussed in Chapter 6). After the injected gas and the cell fluid mix, the
gas-plus-liquid volumes will be larger than the assumed cell volume. The
excess volume is transferred to cell 2. Metcalfe et al. have, as illustrated in
Figure 16-7, used three different criteria for defining the excess volume:
1. Stagnant Oil-AlI the gas formed in cell 1 is transferred to cell 2,
whereas all the oil remains in cell 1.
2. Moving Excess=-Al] the gas formed in cell I istransferred to cell 2. If
the volume of the remaining oil phase exceeds that of the original cell,
the excess volume of oil is also transferred to cell 2.
3. Phase Mobility Criterion- The cell volume remains constant through-
out the calculation, and the excess volume is transferred to cell 2. If
two phases are present, gas and liquid are moved according to their
relative phase mobilities. These are calculated from permeability data
and from the phase viscosities.
Similarly, the excessvolume from ce1l2 istransferred to cell3, etc. When
one batch calculation has been completed, a new injection into cell 1 can
take place, and the cell-to-cell calculation continued. The composition of
the injected gas may be changed Irom batch to batch.
Computer simulation of miscible gas injection may thus be carried out by
a series of flash and material balance calculations, accounting for the mass
transfer between cells.
In case of only one phase in a cell, it can be difficult to decide whether
miscibility has been achieved or noto Ternary diagrams are commonly used
for this purpose, but as mentioned in the Process Description section, ter-
nary diagrams can only give an approximate picture of the thermodynamic
state of the reservoir fluido This is exemplified in Figure 16-8 (Pedersen et
al., 1986). The ternary diagram shows the two-phase area of atypical North
Sea oil. When initially determining the location of the two-phase area, the
distributions of components within the groups Cl +N2 +CO2, C2-C6, and
C7 + were assumed to beconstant. Alsoshown in Figure 16-8 are the te-lines
between the gas and liquid compositions in cells 1through 5 after injection
of 6.6 moles of a hydrocarbon gas mixture (74 mole % Cl, 13 mole % C3,
and 13mole % n-Cg). The results indicate atwo-phase area outside the two-
phase area, constructed on the basis of the initial component distributions
within the groups C +N2 +CO2, C2-C6, and C7 +. This example shows
that ternary diagrams should not be used for mult-cornponent mixtures to
determine the minimum miscibility pressure. Pedersen et al. (1986) have
244 Propertes oj Oils and Natural Gases
suggested that the definition of the miscibility should be based on the loca-
tion of the critical point (Te, Pe) of the reservoir Huid, relative to the reser-
voir conditions (Tres,Pres). Miscibility isachieved when the reservoir Huid is
in the single-phase region, and Tres, Presis located on the dew-point side of
the phase envelope. This isillustrated in Figure 16-9, which shows the phase
envelopes of the Huids in cell1 through cell5, after injection of an apprecia-
ble amount of a hydrocarbon gas (of the same composition as that given
above) into a North Sea oil. In cell 1, Te<Tres, i.e., miscibility has been
achieved according to the definition of Pedersen et al. This definition does
not completely agree with the one given in the beginning of the chapter.
This isbecause two phases may be formed in cell1 at a later stage. In most
cases the newly formed phase, however, constitutes only a very small frac-
tion of the total volume, and the definition of miscibility holds from aprac-
tical point of view. This agrees with the observations of Zick et al. (1986).
J ensen and Michelsen suggested asomewhat different kind of cell calcula-
tion. The starting point, for a vaporizing drive, is a container filled with
injection gas. Original reservoir oil is continuously added to the container
and equilibrated with the gas. The oil that isnot dissolved in the gas phase is
C
1
+ CO
2
+ N
2
Figure 16-8. Two-phase area of original reservoir fluid and the tie-line in cells 1-
5 after injection of 6.6 mole gas per initial cell mole.
l
Simulation oi Miscible Gas Iniection 245
o Crltlcal Polnt
Reservoir Condltions
--- Crltical Locus
~
.c
e
~
~
.t
Cell 4
Cell 5
400
300
200
100
Temperature (el
Figure 16-9. Phase envelopes of mixtures of a reservoir oil and an injection gas,
present in cells 1-5of a cell-to-cell simulation immediately after miscibility has
been achieved in cell 1. Source: Pedersen et al. (1986).
Table 16-1
Results for Mlnimum Miscibillty Pressures Obtained from Slim-tube
Experiments vs. Those from Simulation Model of J ensen and Michelsen
(1987)
Reference
to Exp. InJ ection MIscibility Measured Calculated
Results Gas Mechanlsm MMP (atm) MMP (atm) % Dev.
Glasa (1986)
Natural
Vaporizing 474 445 -6
Gas
Hagoort and Natural
Condensing 125 121 -3
Dumore (1985) Gas
Firoozabadi
Natural
and Aziz
Gas
Vaporizing 396 449 14
(1986)
Firoozabadi
Natural
and Aziz
Gas
Vaporizing 409 496 21
(1986)
Firoozabadi
and Aziz Methane Vaporizing 430 474 10
(1986)
% Dev. =100 x (calculated - experimental)/experimental.
246 Properties oj Ols and Natural Gases
continuously let out. If, at some later stage, no further oil is formed in the
container, miscibility has been achieved. Tosimulate acondensing drive the
container isinitially filled with original reservoir oil. Injection gas isadded,
and the gas which isnot dissolved in the oil phase iscontinuously let out. If,
at some later stage, all the added gas goes into solution in the container
fluid, miscibility has been achieved. As compared with the situation in the
reservoir, the container fluid reflects the situation right at the gas front.
COMPARISON OF CALCVLATED AND EXPERIMENTALLY
DETERMINED MMP's
Slim-tube experiments, as described in Chapter 3, are the most commonly
used experimental technique for carrying out experimental gas displace-
ment studies.
Some results of slim-tube experiments are found in the literature. It isnot
obvious what the relation is between an MMP determined by a slirn-tube
experiment, and the results of an MMP calculation. In Table 16-1, experi-
mental MMP results (from aslim-tube experiment) are compared with those
simulated using the container model of J ensen and Michelsen as described in
the section on Cell-to-Cell Simulation. The results typically agree within
10% .
REFERENCES
Firoozabadi, A. and Aziz, K., "Analyss and Correlation of Nitrogen and

Lean-Gas Miscibility Pressure," SPE Reservoir Engineering, Nov. 1986,
pp. 575-582.
Clase, 0., "Blandbar Fortrengning rsrapport 1986," report (in Norwe-
gian) from IKV, Trondheim, Norway, 1986.
Hagoort, J . and Dumore, J . M., "Deterrnination of Minimum Miscibility
Pressure with an Equation of State Program," 3rd European Meeting on
Improved Oil Recovery, Rome, Italy, 1985.
J ensen, F. and Michelsen, M. L., "Calculation of First Contact and Multiple
Contact Minimum Miscibility Pressure," paper presented at the 4th Euro-
pean Symposium on Enhanced Oil Recovery, Hamburg, Germany, Oct.
27-29, 1987.
Metcalfe, R. S., Fussel, D. D., and Shelton, J . L., "A Multicell Equilibrium
Separation Model for the Study of Multiple Contact Miscibility in Rich-
Gas Drives," paper presented at the SPE-AIME 47th Annual Meeting,
San Antonio, Texas, Oct. 8-ll, 1972.
Simulation o] Mscible Gas Iniection
247
Pedersen, K. S., Fjellerup, J ., Thomassen, P., and Fredenslund, Aa., "Stud-
ies of Gas Injection into Oil Reservoirs by a Cell-to-Cell Simulation
Model," paper presented at the 61st Annual Conference of the Society of
Petroleum Engineers, New Orleans, LA, October 5-8, 1986.
Stalkup, F. L, Miscible Displacement, Monograph Volume 8, H. L. Doherty
Series, Society of Petroleum Engineers, 1984.
Turek, E. A., Luks, K. D., and Baker, L. E., "Calculation of Minimum Mis-
cibility Pressure," paper presented at the 5th Symposium on Enhanced Oil
Recovery of the Society of Petroleum Engineers, Tulsa, OK, April 20-23,
1986.
Zick, A. A., "A Combined CondensinglVaporizing Mechanism in the Dis-
placement of Oil by Enriched Gases," paper presented at the 61st Annual
Conference of the Society of Petroleum Engineers, New Orleans, LA, Oc-
tober 5-8, 1986.
A
Acentric factor (Pitzer's)
correlations for, 115-116
tuning of, 15~158
values of, 83
Alani-Kennedy equation, 90-91
calculation results, 147-151
API density calculation, 94-96
Aromatics, 4-5
B
Binary interaction coefficients
Soave-Redlich- Kwong, 84
tuning of, 155,159-160
Black oils
bubble-point determination, 43
composition, 7,71-72,74-76
C7+-characterization, 114-128
phase envelopes, 6
PVT measurements, 40-54
PVT simulations, 130-140
thermodynamic properties,
141-150
Ba,44
calculation results, 165
definition, 44
measurement of, 46
Bubble points
measurement of, 43
Index
C
Composition data, 69-77
Composition determination, 10-37
extension beyond C20+, 35-37
gas chromatography, 10-17
liquid phase analysis, 12-22
recombination, 32-35
sampling, 39
TBP distillation, 17
Compressibility factor
experimental results, 131,133,
measurement of, 44,48
from Soave-Redlich-Kwong, 85,88
Constant mass expansion
experiment, 56-57
simulation, 130-131
Constant volume depletion
experiment, 58-60
simulation, 135-140,156-160
Corresponding states theory
principie, 175
thermal conductivity, 188-194
viscosity, 175-181
Critical properties
correlations for Teand Pe, 114-116
tuning of, 155-158
values of Teand Pe, 83
C7+-Characterization, 114-128
binary interaction coefficients, 122
continuous thermodynamics, 123
density distribution, 121-122
249
grouping of pseudocomponents,
124-125
molar distribution, 116-120
molecular weight distribution, 121
TBP fractions, 114-116
TBP residue, 116-120
D
Density
Alani-Kennedyequation,
90-91,147-151
API method, 94-96,147-151
calculation of, 89-96
experimental determination, 25
internal consistency, 30
Peneloux,89-90,144-150
Standing-Katz,91-94,147-151
volume translation concept, 89-90
Dew points
experimental determination, 54-56
Differential depletion (liberation)
experiment,43-48
simulation, 131-135
E
Enthalpy
experimental data, 151-153
ideal gas, 87
residual, 87-88
from Soave-Redlich-Kwong, 87-88
Entropy
ideal gas, 87
residual, 87-88
Irom Soave-Redlich-Kwong, 87-88
Equations of state, 79-96
cubic, 81-91
Peneloux, 89-90
Soave-Redlich- Kwong (SRK),
82-88
types of, 79-82
F
Flash calculation, 99-112
multiphase, 104-105
(P,H), 105-108
(P,T), 100-105
(P,T) with water, 108-109
results, 144
simplified (P,T) (kij=O), 109-112
stability anal ysis, 100-102
two phase, 100-104
Flash separation, 48-49
calculation results, 144
Fugacity coefficients
Soave-Redlich- Kwong, 86,88
G
Gas chrornatographic analyses, 10-17
Gas condensates
composition, 7,70,73,76-77
dew point determination, 54-56
phase envelopes, 6
PVT measurements, 54-60
141-150
Gas hydrates, 220-2~
computation, 223-228
experimental results, 230-233
inhibitors, 228-238
types, 220-222.
Gas/oil ratio (GOR)
separator (GOR), 33,40
solution (GOR), 44,46
Gas reinjection, 49-54,234-246
H-N
Heat capacity
ideal gas, 87,126-128
Hydrates (See Gas hydrates)
Minimum miscibility pressure
(MMP)
experimental results, 246
measurement of, 49-54
Miscible gas injection
condensing drive, 234-238
simulation of, 238-246
vaporizing drive, 234-237
Molecular weight
experimental determination, 25-29
internal consistency, 30
astuning parameter, 166-169
Naphthenes,4-5
P-R
Paraffins, 4-5
Peneloux density, 89-90
Phase envelopes, 6
calculation pracedure, 112,113
of different reservoir fluids, 6
Phase compositions r
from Soave-Redlich- Kwong, 86
measurements of (see
Composition Determination)
PNA distribution
estimation of, 120-121
experimental determination of,
10-17
PVT measurements
black oils, 40-54
Bo,44,46
bubble points, 43
compressibility factors, 44
constant mass expansion, 56-57
constant volume depletion, 58-60
dew points, 54-56
differential depletion, 43-48
differentialliberation, 43-48
flash separation, 48-49
formation volume factors, 50
gas condensates, 54-60
gas reinjection, 49-54
lndex 251
minimum miscibility pressure,
49-54
real gas deviation factor, 44
separator gas/oil ratio, 40
separator test, 48-49
shrinkage factor, 44,46
slimtube test, 49-54
solution gas/oil ratio, 44,46
swelling test, 49
volatile fluids, 60-63
volume factors, 50
constant mass expansion, 130-131
constant volume depletion,
135-140
differential depletion, 131-135
differentialliberation, 131-135
phase envelopes, 141-144
separator test, 140
Reservoir fluids
composition, 7
constituents, 4-6
phase envelopes, 6
PNA-distribution, 8
s
Separator test
experiment, 48-49
simulation, 140
Shrinkage factor, 44
definition, 44
measurement of, 46
Slimtube test, 49-54 .
simulation of, 241-246
Soave-Redlich- Kwong (SRK)
equation, 82-88
binary interaction coefficients, 84
compressibility factors frorn, 85,88
enthalpies from, 87-88
entrapies from, 87-88
fugacity coefficients from, 86,88
phase compositions from, 86
252 Properties o} Oils and Natural Gases
Standing-Katz density, 91-94
Surface tension
definition, 196
relation to viscosity, 199-200
Swelling test, 49
T- Z
TBP-distillation, 17
Thermal conductivity
corresponding states theory,
188-194
, definition, 187
experimental results, 194
Transport properties (See Surface
Tension, Thermal conductivity,
and Viscosity)
Tuning, 155-170
binary interaction coefficients,
155,159-166
~~"""' __ """''''''.Ir.''>~...,,.. __ ~,.kLW--''''''''-'l.,~,---,,,,,,,,",,:,~:,,"'I"~~~-~~
'It..L ..r~T!ITn '-,';r.:1l";!l .'p r. 'l!
A',U' "'U, U t-hlJ '-H l,,'"f r: tU" "

r~I nI I n T '. (' ;, ~
~
'~ '-,I~t~.~~~<~ '~'_1<' I
, w 'NVENT~_~~~.':J.339c ._._, .....,...~
MATe:?I,'. r
r~:-I~'i~~-~-~'(;~ 24-8'f- --'1
~ , __ _~._ ,_. _ "_k" o,, _" ~_,. '.
\.. ~-~...."..'",:..~,,-.,.", ....,..,\"' .... -'.:--,.
molecular weight of plus fraction,
166-169
Tc,Pc and w, 155-158
Viscosity
corresponding states theory,
175-181
definition, 172
Volatile fluids
composition, 7,72
phase envelopes, 6
PVT measurement, 60-63
141-150
~Volume factor, 50
Wax, 208-218
calculation of wax formation,
208c214,217-218
composition, 208
inhibitors, 214-216
J J
..

Properties of Oils and Natural Gases

Enviado por

Dados do documento

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

Properties of Oils and Natural Gases

Enviado por

Direitos autorais:

Formatos disponíveis

r

The chromatographic separation of oxygen, nitrogen, carbon dioxide,

Figure 3-2. Single stage

Fugacity coefficient at constant composition, </>:

{3y +(1 - (3)X=Z (i = 1, 2, 3, oo., N) (6-14)

%A =670 v +3,660/MW for v-cO (7-26)

300.0 '00.0 100.0 200.0

00 11{) 1(') 1"

Exper i ment al Bubbl e Poi nt s

Exper i ment al r esul t s

Exper i ment al r esul t s f or cr ude o I wi t h

uA!> ...u Vil. I

Firoozabadi, A. and Aziz, K., "Analyss and Correlation of Nitrogen and

A',U' "'U, U t-hlJ '-H l,,'"f r: tU" "

Você também pode gostar