Equilibrium Ratio Prediction and Calculation

13
Equilibrium Ratio Prediction and Calculation
CONTENTS
1 INTRODUCTION
1.1 Importance of the Critical Point
1.2 Equilibrium Ratio
2 EMPIRICAL PREDICTION METHODS FOR
EQUILIBRIUM BEHAVIOUR
2.1 Black-Oil System
2.2 Correlation of Experimental Data Through K
Values
2.2.1 Method of Calculating Convergence Pressure
3 METHODS BASED ON THERMODYNAMIC
PRINCIPLES
3.1 Methods Based on Empirical Equations of
State of Fluid Phase Theory
3.2 Prediction of Vapour-Liquid Equilibrium
4. THE COMPOSITION DATA REQUIREMENTS
FOR THE APPLICATION THERMODYNAMIC
EQUILIBRIUM PREDICTION METHODS TO
MULTI-COMPONENT HYDROCARBON
MIXTURES
LEARNING OBJECTIVES
Having worked through this chapter the Student will be able to:
Comment briefly on the approach to vapour equilibrium ( VLE ) calculations

in terms of ; the black oil approach, empirical K values and EOS..
Describe briefly the application of convergence pressure based K values in

VLE calculations.
Be able to determine the correct convergence pressure for VLE application .
Comment on the evolution of EOS in VLE application . ( The student would not
be expected to reproduce the equations.)
Describe briefly the application of EOS in VLE calculations.
13
VAPOUR-LIQUID EQUILIBRIUM PREDICTIONS OF PHYSICAL

PROPERTIES
1 INTRODUCTION
In the previous chapter we introduced the concept of vapour-liquid equilibrium and
its role in attempting to predict the behaviour of multi-component hydrocarbon
mixtures in the two phase region, in particular the ability to predict phase densities and
composition as a function of pressure and temperature. In the chapter we also
reviewed some of the basic equations and procedures which are at the foundation of
these multicomponent calculations
In petroleum engineering applications the situation is complicated in that often very
little detailed information is available on the hydrocarbon mixture. That is, either
individual components are not identified, or, because it is a multi-component mixture
certain component concentrations are difficult to measure accurately even though
they can greatly influence the properties of the mixture.
There are two distinctive features of vapour equilibrium calculations in the context of
petroleum engineering compared to other sectors like process engineering. In process
engineering the calculations are mainly related to either a single process unit or a few
in series, whereas in reservoir engineering the calculations are often part of multi grid
block reservoir simulation where the numbers of process units can be considered in
hundreds or even thousands. In process engineering a full compositional description
is available for the prime objective is to determine compositional split to obtain
specific compositional objectives. In petroleum engineering however, the objective
is not to determine full compositional description of produced phases, but to
determine phase volumes and phase properties. The challenge is to carry this out with
a minimal composition description without compromising the quality of physical
property predictions.
In addition for petroleum engineering applications the calculation method used
should be:
(1) applicable to multi-component mixtures;
(2) be accurate in predicting thermodynamic equilibrium and volumetric properties
over a wide range of conditions of temperature and pressure and specifically be
accurate around and beyond the critical point;
(3) preferably require only pure component data or binary data which is either
readily available from the literature or derivable from available data.
1.1 Importance of the Critical Point

The prediction of the true critical properties of a multi-component system is an
important aspect of the general problem of predicting the overall phase behaviour of
the system. The critical state is the unique condition about which the liquid and vapour
phases are defined, and hence it has theoretical as well as practical significance. In
hydrocarbon processing and producing operations, a knowledge of the critical
condition is of particular significance because many of these operations take place
under conditions which are at or near the bubble point or upper dew point regions and
are frequently accompanied by isobaric or isothermal retrograde phenomena. Fluid
Department of Petroleum Engineering, Heriot-Watt University
property predictions and design calculations in this region are often the most difficult
to make, and a knowledge of a precise location of the critical point for the system under
study is of the utmost assistance.
From a theoretical point of view, the derivatives of many of the thermodynamic and
transport properties take on a special significance as the critical state is approached.
In an empirical way the critical state has formed an integral part of many useful
generalised correlations such as those based on the theorem of corresponding states
or the convergence pressure concept in vapour-liquid equilibrium calculations.
In many ways the characteristics of the critical state that make it theoretically and
practically significant are also the characteristics that make it one of the more difficult
conditions to measure experimentally. The very fact that density differences between
phases vanish, that the rate of volume change with respect to pressure approaches
infinity, or that infinitesimal temperature gradients can be responsible for a transition
from 100% liquid to 100% vapour all make the critical condition one of the more
difficult to measure or observe accurately. For obvious economic reasons, it is a
condition that cannot be obtained by experiment in any practical way for the many
systems for which it is required. Consequently, many attempts have been made to
develop methods for predicting the critical properties based on generalised empirical
or semi-empirical procedures.
1.2 Equilibrium Ratio

In the previous chapter we presented the concept of equilibrium ratio, the measure of
how a component is distributed between different phases however we didnt explain
how these distribution values are obtained.
The literature abounds with methods and equations associated with phase equilibrium
prediction. The methods can be broadly classified as:
(1) methods which involve empirical curve fitting to experimental data;
(2) methods which are based on thermodynamic principles.
2 EMPIRICAL PREDICTION METHODS FOR EQUILIBRIUM

BEHAVIOUR
There are a range of approaches to expressing the vapour liquid distribution behaviour
of reservoir fluid hydrocarbon systems, as outlined below.
2.1 Black-Oil System

The black-oil system as described in the chapter on liquid properties treats the fluids
as two components, stock tank oil and solution gas. This concept and associated
calculations were covered in that chapter.
2.2 Correlation of Experimental Data Through K Values

A complicating factor in the application of equilibrium ratios in the context of
reservoir fluids is that the distribution of a component between phases is not only
influenced by the temperature and pressure but is also influenced by the composition.
Calculating therefore how a unique fluid like that from a new exploration well
4
13
performs in these separation calculations is not straightforward, since no one before

would have been able to carry out tests on the fluid!!
An approach is required that the distribution of each component of the mixture
between the liquid and vapour phases be experimentally determined for a range of
temperatures and pressures. The resultant measurements are expressed as an equilibrium
ratio Kj, for component j defined as:
where:
Kj =
yj
xj
(1)
Kj = equilibrium ratio
yj = mole fraction of component j in the vapour phase
xj = mole fraction of component j in the liquid phase
This approach is really only useful for light hydrocarbons. The data is usually
expressed graphically as a plot of Kj versus pressure for a constant temperature.
The oil industry has relied on experimentally determined equilibrium ratios although
increasingly over recent years the move has been towards Equation of State derived
K values. Clearly the empirical K values are obtained from known mixtures, the
challenge is to determine the applicability of these empirical values to new mixtures.
At high pressures the equilibrium ratio is a function of temperature, pressure and also
the composition of the mixture. This compositional influence is of great significance.
Hence, pure component K-values measured for one mixture cannot be accurately
applied to another mixture. Figure 1 and 2 present the K values for a heavy oil and at
the other extreme a condensate at a temperature of 200F. The K values converge to
unity at 5,000 psia and 4,000 psia respectively. This point is called the convergence
pressure. Close examination of these two sets of K values demonstrate that these
distribution coefficients do not have the same values at particular pressures and
temperatures, confirming this compositional influence on K values.
The effect of composition on the K values is achieved by applying the concept of
convergence pressure. The concept of convergence pressure arose from the observation
that for light hydrocarbon mixtures the isothermal component K-values converge to
unity at a specific pressure known as the convergence pressure Pc, see Figures 3 for
a binary mixture and 4 for a light oil . This convergence pressure is a measure of the
composition of the mixture and can be calculated from a knowledge of the effective
boiling point of the lightest and heaviest components in the mixture.
If the temperature at which the equilibrium ratios are presented is the critical
temperature then the convergence pressure would be the critical pressure.
At a given temperature, as the system pressure increases, the K-values of all
components of the system converge to unity when the system pressure reaches the
convergence pressure. In other words, it is the pressure for a system at a given
temperature when vapour-liquid separation is no longer possible. Naturally, it is
equally impossible to have a vapour-liquid separation at a given temperature in which
the system pressure is greater than the convergence pressure.
100
et
h
10
Et
ha
ne
Pr
op
an
Bu
e
ta
ne
Pe
s
nt
an
e
He
s
xa
ne
s
K=1
pt
e
0.1
an
es
an
d
He
av
ie
0.01
10
an
100
1,000
Pressure, psia
10,000
Figure 1
Equilibrium Ratios at
200F for Low-Shrinkage
Oil(Amyx Bass & Whiting)1
40
Equilibrium Ratio, K
20
10
8
6
Et
ha
ne
Pr
op
an
e
Bu
ta n
es
Pe
n ta
ne
s
2
1
0.8
0.6
0.4
0.2
He
xa
ne
et
ha
ne
0.1
0.08
0.06
0.04 He
p
0.02
0.01
100
ta n
es
and
Hea vi e r
1,000
Pressure, psia
10,000
Figure 2
Equilibrium Ratios at
200F for a Condensate
Fluid (Amyx Bass &
Whiting)1
13
CONSTANT
TEMPERATURE
Component 1
In Kj
Kj - 1.0
Figure 3
Variation of K Values with
pressure for a typical
Binary Hydrocarbon
System
Component 2
In Pc
In P
Variations of Equilibrium
Constant With Pressure
at 200 F
10.0
Et
et
ha
ne
Equilibrium Constant, K
an
Pr
e
op
Bu
an
ta
e
ne
Pe
s
nt
H
an
ex
es
an
es
Figure 4
Variation of K Values with
pressure for a crude oil
containing Light
Hydrocarbons
1.0
0.1
ep
ta
ne
an
He
av
ie
r
0.01
10
100
1,000
10,000
Pressure, psia
The apparent convergence pressure is related to the composition of the fluid. This
empirical basis has found favour in the industry for many years as a source of K value
data despite the ability to calculate the information based on equations of state.
However with the increasing availability of computational power and greater confidence
in equations of state the use of convergence pressure based K values is diminishing.
A comprehensive set of equilibrium ratios is published in the GPSA Engineering Data
Book 2. The K values for single components are presented in the NGAA book for a
range of convergence pressures. K values for a convergence pressure of 5,000 psia are
at the back of this chapter for 2 pure components. Figure 9 and 10.
2.2.1 Method of Calculating Convergence Pressure

As was indicated in figures 2 and 3 different mixtures present different K values as
represented by different convergence pressures. The challenge is therefore to select
the convergence pressure value appropriate to the fluid for which we are seeking to
determine its phase separation characteristics.
The procedure is to convert our mixture into a two component mixture based on
methane then seek to identify a compound which has the same physical properties as
our heavier C2+ component
The chart, Figure 5 from the NGPSA Manual, can be used for this2. The method is to
take as the lightest component present in any significant quantity (1.0% or greater in
a raw mixture; it is usually methane) as the light component of a two component
system. The heavy component is estimated from the composition of the remaining
components. A visual estimate is usually good enough. Join the heavy and light
component together as shown on the Figure 5 and read off PK against the operating
temperature.
It has been established that the convergence pressure of systems as encountered
generally in natural gas processes is a function of the temperature and the composition
of the liquid phase. This presumes that the liquid composition had already been known
from a flash calculation using a first approximate guess for convergence pressure.
Therefore, the method of calculating convergence pressure is an iterative procedure.
This calculation is suggested:
Step 1: Assume the liquid phase composition or make an approximation. (If there
is no guide, use the total feed composition).
Step 2: Identify the lightest hydrocarbon component which is present at least 0.1
mole % in the liquid phase.
Step 3: Calculate the weight average critical temperature and critical pressure for the
remaining heavier components to form a pseudo binary system. (A shortcut
approach good for most hydrocarbon systems is to calculate the weight average Tc
only).
Step 4: Trace the critical locus of the binary consisting of the light component and
pseudo-heavy component. When the averaged pseudo heavy component is between
two real hydrocarbons, interpolation of the two critical loci must be made.
8
-200
-100
Triple Phase Locus
Triple Phase Locus
-300
100
200
300
400
500
1,000
900
800
700
600
2,000
3,000
4,000
en
N it r o g
5,000
Eth
ane
0
C -K ens
ol
1
CO2
C1-nC9
100
H2S
C -n C
1
10
C1 -nC
200
SO2
300
Temperature, F
C3 -nC
C 3-nC6
C2 -nC
NH3
N
Bu
ta n
e
I
P
en
ta
ne
N
Pe
nt
an
e
10,000
9,000
8,000
7,000
6,000
ne
Met
ha
le n
Eth
y
e
le n
Pr
op
y
an
Pr
op
ne
I-B
uta
an
NHe
x
20,000
nC 4-nC10
10
C2 -nC
N-
ne
ta
400
He
p
-C 2
H2
e
an
-C 2
H2
NO
ct

Benzene
500
600
nC9
nC7-nC19
Methyl- Toluene
Cyclohexane
-C 3
H2
an
Convergence Pressure, PSIA
Figure 5
Convergence pressure for
Hydrocarbons.2
nC 6
H 2NDe
c
Convergence Pressures for Hydrocarbons

(Critical Locus)
n-
nC12
700
nC11
nC7-nC23
nC5-nC24
nC5-nC22
nC5-nC18
nC5-nC16
nC15
nC14
800
e
an
ec
ad
nC13
He
H2O
nC17
900
nC18
Kensol
13
20,000
C thru' Mid-Continen
t
1
10,000
9,000
8,000
7,000
6,000
5,000
C 1 thru' Distillate Crud

e
4,000
C thr u' G a s o
li
1
3,000
ne
de
ru
tC
en
te
nt
in
ill a
Le
an
O
il M
ol
.
t.
200
16
1
Cr
ud
e
300
Di
st
400
M
id
-C
o
Absorbers, Crude Flashing Towers
500
Lig
ht
Ga
so
lin
1,000
900
800
700
600
2,000
Me
tha
ne
,C
Convergence Pressure, PSIA
C ru d e
Absorber O
verhead
100
-100
100
200
300
400
500
600
700
800
900
Temperature, F
Step 5: Read the convergence pressure (ordinate) at the temperature (abscissa)

corresponding to that of the desired flash conditions, from Figure 5.
Step 6: Using the convergence pressure determined at Step 5, together with the
system, obtain K-values for the components from the appropriate convergencepressure K-charts.
Step 7: Make a flash calculation with the feed composition and the K-values from
Step 6.
Step 8: Repeat Steps 2 through 7 until the assumed and calculated convergence
pressures check within an acceptable tolerance, or until the two successive calculations give the same light and pseudo heavy components check within an acceptable
tolerance.
When the convergence pressure so determined is between the values for which charts
are provided, interpolation between charts may be necessary depending on how close
the operating pressure is to the convergence pressure.
If K-values do not change rapidly with PK(PK>>P) then the set of charts nearest to
calculated Pic may be used. Otherwise, a crossplot of K values versus PK all at constant
temperature and pressure, must be constructed for interpolation.
For those components characterised as a C7+ fraction Katz has suggests using a K value
of 15% of that of C7+. Danesh3 makes reference to other correlations to estimate the
critical properties of C7+ fractions. McCain6 has also presented pseudo critical
properties of C7+ as a function of molecular weight and specific gravity, his correlation
figures are shown below in figure 7.
10
Figure 6
Pseudocritical properties of
Heptane Plus. (NGPSA 5th.
edition 1957)
13
1700
Pseudocritical temperature, R
1600
1500
Specific gravity of heptanes plus = 1.0
1400
.95
.90
1300
.85
.80
1200
.75
.70
1100
1000
900
100
150
200
250
Molecular weight of heptanes plus
300
Pseudocritical temperature, R
500
450
400
1.0 = Specific gravity of heptanes plus
350
.95
.90
300
.85
.80
250
.75
.70
200
150
Figure 7
Pseudocritical properties of
Heptane Plus6.
100
100
150
200
250
Molecular weight of heptanes plus
300
The following example has been presented by McCain in his text . and is helpful as
a worked example in determining the appropriate convergence pressure charts
Example: McCain4. The gas-liquid equilibrium of a high-shrinkage crude oil has
been calculated. The composition of the liquid phase formed at 75F and 100 psia is
given below. A convergence pressure of 2000 psia was used to determine the
equilibrium ratios for the calculations. What value of convergence pressure should
have been used for this mixture?
11
Component
Composition of liquid
mole fraction
Methane
Ethane
Propane
i-butane
n-butane
i-pentane
n-pentane
Hexane
Heavier*
0.0356
0.0299
0.0919
0.0170
0.0416
0.0198
0.0313
0.0511
0.6818
1.0000
* Molecular weight and gravity of C7+

assumed to be 268 and 0.886
Solution
First, the composition of the liquid must be expressed as weight fraction.
Component
Composition of
liquid mole
C1
C2
C3
i-C4
n-C4
i-C5
n-C5
*C6
*C7+
0.0356
0.0299
0.0919
0.0170
0.0416
0.0198
0.0313
0.0511
0.6818
1.0000
Molecular
weight Mj
Weight xMj
16.0
30.1
44.1
58.1
58.1
72.2
72.2
86.2
263
0.5696
0.9000
4.0528
0.9877
2.4170
1.4296
2.2599
4.4048
179.3134
196.3348
Composition of
liquid, weight
fraction
xjMj/xjMj
0.0029
0.0046
0.0206
0.0050
0.0123
0.0073
0.0115
0.0224
0.9133
0.9999
* Molecular weight and gravity of C7+ assumed to be 263 and 0.886 respectively
Second, adjust weight fraction to exclude methane and calculate weighted-average

critical properties.
Component Composition
excluding
methane, wj
C2
C3
i-C4
n-C4
i-C5
n-C5
C6
*C7+
0.0046
0.0207
0.0050
0.0123
0.0073
0.0115
0.0225
0.9160
0.9999
Critical
temperature
R Tcj
R
wjtcj
psia
pressure, wjpcj
psia pc
psia pcj
549.8
665.7
734.7
765.3
828.8
845.4
899.3
1360
2.53
13.78
3.67
9.41
6.05
9.72
20.23
1245.76
wt avg Tc=
1311R, 851F
707.8
616.3
529.1
550.7
490.4
488.6
445.4
240
* Critical properties of C7+ from figure 7.
12
Critical
3.26
12.76
2.65
6.77
3.58
5.623
10.02
219.84
wt avg pc=
265 psia
13
The calculated values of weight averaged Tc and Pc are close to Kensol and the mid
continent crude point of figures 5 and 6. The location of the temperature 75F with the
methane - component (Kensol or mid - continental crude) is at a convergence pressure
of around 10,000 psia, a value much greater than the assumed value of 2,000. The
calculator needs to be repeated using the higher convergence pressure related K value
data. The calculation procedure will converge when the estimated convergence
pressure is the same as the calculated convergence pressure.
3 EQUATION OF STATE BASED EQUILIBRIUM CALCULATIONS

3.1 Methods Based on Empirical Equations of State of Fluid Phase Theory
The thermodynamic properties of a pure fluid may be represented by an equation of
state of the generalised form:
f (P,V,T) = 0
where the pressure P, temperature T and molar volume V are related by a mathematical
function. Most equations of state have been developed by fitting an analytical
expression to pure component PVT data. To extend the application of the developed
equation to mixtures, the parameters of the equation must take into account the
composition of the mixture and for simplicity require only the insertion of pure
component data. Since most equations are effectively only mathematical models, the
equations tend in general to be more complex, that is contain a large number of
parameters, as the required level of accuracy increases. The basic assumption in the
development of an equation of state is that at a critical point:
2 P
P
= 2 =0
V T V T
Equations of state can be used for the following purposes:
1
representation of PVT data to assist data smoothing and improve interpolation;
prediction of vapour-liquid equilibria of mixtures especially at high pressures;
prediction of gas phase properties of pure fluids and their mixtures using a
minimum amount of experimental data.
In the gas property chapter we reviewed the topic of equations of state. Currently
although a number of different equations could be used, the industry favours two, the
Peng Robinson equation of state and the Soave modification of the Redlich Kwong
equation of state.
3.2 Prediction of Vapour-Liquid Equilibrium

An equation of state capable of predicting behaviour in the liquid phase and vapour
phase is sufficient in itself for vapour-liquid equilibrium predictions. Unfortunately
as we have indicated although modifications to the Van der Waals type equation have
predicted vapour properties the equations in general have not been so accurate in
predicting the liquid behaviour.
13
Many vapour-liquid prediction methods have therefore gone to an equation of state,

for example the Redlich-Kwong equation, for the prediction of the vapour phase
fugacity and have employed a liquid theory to evaluate the liquid phase properties.
We will now go through the steps in the application of equation of states in vapour
liquid equilibrium calculations. It is worth noting that the steps described are those
which would be taking place within each grid block of a compositional reservoir
simulator. The implication therefore is a very large computational load for large
numbers of grid blocks.
The application of equation of states is based on the fact that at
equilibrium, the fugacity of the gas and liquid phases are identical i.e:
fiv = fiL
(2)
where fiv and fiL are the fugacities of component in the vapour and liquid phases
respectively.
The ratio of the fugaciy to pressure is called the fugacity coefficient
where i = fi/(pzi)
(3)
where zi is the composition of the component in the system.
The fugacities can be expressed for a vapour and liquid therefore by:
fiL = i xip
and:
fiv = i yip
p
yi & xi
- is the system pressure;

- are the mole fractions of i in the vapour and liquid phase.
Therefore:
fiL
xp
y
K i = iL = i = i
f
n i( )
ig
ig
yi p
The fugacity coefficients for the liquid and gas can be calculated using the following
equation
ln i =
1 RT p
dV ln z
RT V n i T , V, n
i
(4)
McCain5 and also Ahmed6 give good descriptions of the application of EOS in
equilibrium calculations.
At the present time the preferred EOS are the 3 parameter Peng-Robinson (PR) 8,9 and
the Soave-Redlich Kwong (SRK)10. Other equations exist and are more accurate in
predicting some properties. The considerable investment in binary interaction
14
13
parameters for the preferred equations is such that there is a reluctance to use some
recently developed EOSs. Danesh's3 text gives a good review of the EOS.
Since the applications of the equations are applied to mixtures, mixing rules are
required to determine the values for the parameters in the particular EOS being used.
We will use the Peng Robinson equation as our basis but others could be used.
The Peng Robinson equation is
p=
RT
aT
Vm b Vm (Vm + b) = b(Vm b)
(5)
The equation is set up for both liquid and gas using the following mixing rules to
calculate b, and aT . The rules are presented in the context of gas ie. y, clearly for
liquids, x values are used.
and
b = yi bi
(6)
a T = y i y j (a Ti a Tj )0.5 (1 k iy )
i
(7)
where kij are the binary interaction coefficients, and kii = kjj = o and kij = kji.
The value of bi and aTi for the individual components are calculated as follows
b i = 0.07780
and
RTci
Pci
aTi = acii and
a c i = 0.45724
R 2 T 2 ci
Pci
(8)
(9)
(10)
i is a temperature dependant factor where
i 0.5 = 1 + m(1.0 Tri0.5 )
(11)
where
m = 0.37476 + 1.54226 i 0.26992 2i
(12)
The Peng Robinson equation can be written as a cubic equation in terms of z, the
compressibility factor,
15
z3 - (1-B)z2 + (A-2B -3B2)z - (AB - B2 -B3) =0
(13)
where
A=
aP
bP
B=
2 2
RT
RT
(14)
The solution to the above equation gives three roots for z. The highest value is the
liquid phase z factor and when the equation is solved using vapour compositions then
the lowest root is the z factor for the vapour.
If the Peng Robinson equation is combined with the fugacity coefficient equation the
following equation results for the fugacity coefficient of each component.
z + (21.5 + 1)B
A
ln i = ln(z B) + (z 1)Bi 1.5 ( A i Bi ) ln
1.5
2
z (2 + 1)B
Where B'i = bi/b
(15)
(16)
and
A i =
1
aT
0.5
0.5
2a Ti y i a Ti 1 k ij
1
(17)
Following all these steps independently for the liquid and gas phases the fugacities of
the gas and liquid phases can be calculated.
fLi = xipLi and fvi = yipvi
When fLi=fvi then equilibrium is achieved and calculations are complete.
Having presented these equations we will now describe the process to calculate K
values , vapour liquid equilibrium ratios and compositions given a system composition,
temperature and pressure.
Step 1: Estimate the K values of the system. The Wilson11 equation is good for this
purpose.
p
T
K i = ci exp 5.37(1 + i )1 ci
T
p
(18)
wi = acentric factor for component i

Step 2: Carry out vapour equilibrium calculations using estimated K values using
the iterative procedure outlined previously. That is estimate the V/L ratio and iterate
until convergence is obtained, that is when compositions sum to unity. We now have
liquid and vapour compositions to use in equation of state calculations.
16
13
Step 3: Using liquid compositions, calculate the A & B values for the EOS and then
solve the z-value form of the EOS, to determine z. The lowest root (value) is the z
value for the liquid.
Step 4: Calculate the compositional coefficients A'i and B'i for the liquid components
and calculate the fugacity coefficients of the components of the liquid
Step 5: Repeat steps 3 & 4 using the vapour phase compositions.
Step 6: Calculate fgi and fLi fgi=yipi and fLi= xipI . Check if fgi= fLi.
If this value is greater than 10-12 then the whole process has to be repeated from step
1, except that the K values used are the calculated K values arising from step 5 i.e.
Rather than set up the tolerance check on fugacity
equivalence the tolerance can be
i = fLi - fgi
based on K values.
A value of of 0.001 can be used for the sum of the errors.
Ki =
Li
gi
(K
=
E
i
(19)
KiC )
KiEKiC
(20)
The iteration is complete when these tolerance limits are met and the compositions of
the repective phases are those which have been been determined at the last iteration.
Calculations can then proceed to provide volumetric and density data for the
respective phases.
Danesh3 has given a flow diagram for the above flash calculation.
An example follows to illustrate the calculation process.
17
Start
Input zi, P, T,
Component Properties
Estimate Ki, Using Eq.(18)
Calculate xi, yi, Using Figs. 20-22 (Ch 12)
Set-Up EOS
For Liquid
Set-Up EOS
For Vapour
Calculate ZL
Calculate ZV
Calculate fiL
Calculate fiV
Adjust Ki =
Ki old (fiL/fiV)
NO
NO
Is
(1- fiL/fiV)2<10-12
?
YES
YES
Print xi,yi,vL,vV,nL,nV
End
EXERCISE 1.
Calculate the liquid and vapour phase composition when the mixture with the
composition given in Table 1 is flashed to 2000 psia and 100C. Use the PengRobinson equation of state with binary interaction parameters given in Table 2.
Table 1: Multicomponent system.
Component
Methane
n-Pentane
n-Decane
n-Hexadecane
Composition
mole fraction
0.55100
0.11400
0.14600
0.18900
Table 2: Binary interaction parameters of Peng- Robinson Eq.
18
Figure 8
Flow chart of flash
calculations using equation
of state.2
13
No.
1
2
3
4
Component
Methane
n-Pentane
n-Decane
n-Hexadecane
1
0.0000
0.0236
0.0489
0.0600
0.0000
0.0000
0.0070
0.0000
0.0000
0.0000
4. THE COMPOSITION DATA REQUIREMENTS FOR THE APPLICATION

OF THERMODYNAMIC EQUILIBRIUM PREDICTION METHODS TO
MULTI-COMPONENT HYDROCARBON MIXTURES
In designing a versatile phase equilibrium prediction computer program, the available
experimental data for the mixture whose performance is to be simulated, must be
considered. Most laboratory analyses of crude oil are predominantly concerned with
the composition. This is normally reported as a percentage weight. This presents few
problems for lighter hydrocarbons such as methane which can have no isomers nor for
other light hydrocarbons which do have isomers, for example butane, since the
number of hydrocarbons having the same number of carbon atoms is relatively few.
However, for heavier hydrocarbons the number of hydrocarbons which possess the
same number of carbon atoms can be very large due to the presence of not only alphatic
or straight chain compounds but also of cyclic compounds such as aromatics or
naphthenes, or combinations of these structural types. Obviously the structure of
molecules greatly influences the forces between them and hence their deviations from
ideality. This is most significant when binary interaction parameters are employed in
the mixing. The analysis of the heavier components in crudes is further complicated
by the fact that most of the individual components are present in very small quantities
which makes them difficult to identify and quantify.
Ideally what is required is a full analysis of a crude oil in which all components are
quantified and identified so that the effects on the mixture behaviour due to their
structural and chemical properties can be taken into consideration. The phase
equilibrium prediction method which is employed, must match the quality and extent
of the experimental data available, for example, the use of a highly accurate method
might be invalidated if the composition analysis has been carried out on a basis of
generalised crude fractions such as boiling range. However, this is where flexibility
must be built into the program.
At present there is a tendency to increase the extent to which a crude oil composition
is analysed and this can only serve to make phase equilibrium prediction more
accurate. Condensate well fluids, however, are still generally reported on a basis of
a heptanes plus weight fraction, that is the collective weight fraction of heptane and
heavier components, because these heavier components are generally only present in
minute quantities.
It is thus desirable that any proposed phase equilibrium method should be able to
utilise detailed experimental data based on identifying the individual components of
a crude oil as well as data derived from a more generalised approach as discussed
above.
19
Most of the work carried out on fluid phase equilibria has concerned itself solely with
binary or tertiary systems and occasionally mixtures containing pure components
whose behaviour is quite well understood. Much of the published work on the
adaptation of experimental data for prediction techniques has been related to petroleum
refinery operations.
Figure 9
K value chart for Ethane.
20
13
Pressure, PSIA
10
1,000
30
50
6 7 8
9 100
300 4 500 6
7 8 9 1,000
9
8
7
6
3,000 4
6 7 8 9
Plotted from 1947 tabulation

of G. G. Brown, University of
Michigan. Extrapolated and
drawn by The Fluor Corp. Ltd.
in 1957.
10,000
1,000
9
8
7
6
5
100
100
9
8
7
6
9
8
7
6
Temperature F
2
0
0
30 20
25
10
K= y/x
50
40 0
0
10
9
8
7
6
9
8
7
6
10
0
80
5
4
60
40
20
K= y/x
0
-2
0
1.0
-4
1.0
-6
0
9
8
7
6
9
8
7
6
5
0.1
0.1
9
8
7
6
9
8
7
6
10
30
50
6 7 8 9 100
300 4 500 6
7 8 9 1,000
Pressure, PSIA
3,000 4
6 7 8 9
10,000
ETHANE
CONV. PRESS. 5000 PSIA
21
Pressure, PSIA
10
1,000
30
50
6 7 8
9 100
300 4 500 6
7 8 9 1,000
9
8
7
6
3,000 4
6 7 8 9
Plotted from 1947 tabulation

of G. G. Brown, University of
Michigan. Extrapolated and
drawn by The Fluor Corp. Ltd.
in 1957.
10,000
1,000
9
8
7
6
5
Temperature F
100
9
8
7
6
45
40
0
38
0
36
0
34
0
32
0
30
0
28
0
26
0
24
0
22
0
200
4
3
10
K= y/x
100
50
9
8
7
6
9
8
7
6
5
4
3
4
3
10
9
8
7
6
K= y/x
5
4
3
18
16
14
1.0
1.0
9
8
7
6
12
9
8
7
6
100
80
60
0.1
9
8
7
6
40
0.1
20
9
8
7
6
-20
10
30
50
6 7 8 9 100
300 4 500 6
Pressure, PSIA
22
7 8 9 1,000
3,000 4
6 7 8 9
10,000
OCTANE
CONV. PRESS. 5000 PSIA
13
Solution to Exercise
EXERCISE
Calculate the liquid and vapour phase composition when the mixture with the
composition given in Table 1 is flashed to 2000 psia and 100C. Use the PengRobinson equation of state with binary interaction parameters given in Table 2.
Table 1: Multicomponent system.
Component
Methane
n-Pentane
n-Decane
n-Hexadecane
Composition
mole fraction
0.55100
0.11400
0.14600
0.18900
Table 2: Binary interaction parameters of Peng- Robinson Eq.

No.
1
2
3
4
Component
Methane
n-Pentane
n-Decane
n-Hexadecane
1
0.0000
0.0236
0.0489
0.0600
0.0000
0.0000
0.0070
0.0000
0.0000
0.0000
SOLUTION
Step 0: Calculate the coefficients of components in the mixture, using the properties
(critical temperature Tc and critical pressure Pc) of pure compounds:
ac i = 0.45724
b i = 0.07780
R 2 T 2 ci
Pci
(10)
RTci
Pci
(8)
If <0.5
m= 0.37464 + 1.54226 - 0.26992 2
If >0.5
m= 0.3796 + 1.485 - 0.1644 2 - 0.016673
(12)
w - accentric factor from thermodynamic property tables for pure components
i = 1.0 + m 1.0 T r )
ai = ci i
(11)
(9)
Figure 10
K value chart for Octane.
23
Component Tci
K
Pci
atm
C1
n-C5
n-C10
n-C16
45.39
33.26
20.82
13.82
190.6
469.7
617.7
723
i
acentric
factor
0.0115
0.2515
0.4923
0.7174
aci
ai
bi
2.4632
20.4230
56.4253
116.4577
0.7112
1.1686
1.5327
1.9176
1.7518
23.8663
86.4825
223.3236
.0268
.0902
0189
.3340
Step 1: Select initial values for equilibrium ratios (K-values) and calculate the trial
composition of liquid and vapour phases at equilibrium. The procedure is iterative as
follows.
Iteration One
Initial K-values can be calculated from Wilson Equation:
Ki =
Pci
T
Exp (5.37)(1.0 + i )1.0 ci ))
P
T
Component
Methane
n-Pentane
n-Decane
n-Hexadecane
Feed (Zi)
0.55100
0.11400
0.14600
0.18900
(15)
K-Values (Ki)-Wilson Eq.

4.7644
4.290 E-02
7.9917 E-04
1.7768 E-05
Step 2: Determine the trial value of phase fraction (vapour fraction here) by solving
the vapour equilibrium equation:
N
L
i=l
Zi
=V
(5)
+ KL
Vapour Fraction (Vf) = 0.4378

Use the determined vapour fraction (Vf) to calculate the liquid and the vapour phase
composition from material balance:
Component
Methane
n-Pentane
n-Decane
n-Hexadecane
Feed (Zi)
0.55100
0.11400
0.14600
0.18900
Liquid (xi)
0.2081
0.1962
0.2595
0.3362
Vapour (yi)
0.9914
0.0084
0.0002
0.000006
Step 3, 4 and 5: Calculate composition dependent coefficients for compressibility

factor (Z - factor) calculations for both liquid and vapour:
N
Liquid:
i=l J=l
0.5
( ) (1. k )
a = x i x j a ia j
ij
(7)
b = xi bi
i=l
24
(8)
13
Vapour:
0.5
( ) (1. k )
a = y i y j a ia j
i=l J=l
(7)
ij
b = xi bi
(8)
i=l
A=
aP
R2T2
Phase
Liquid
Vapour
B=
bP
RT
a
74.88118
1.83396
(14)
b
0.18470
0.02737
A
10.86872
0.26619
B
0.82089
0.12165
Calculate Z - factors of the liquid and vapour phases (Peng - Robinson EOS):
Z3 - (1 - B)Z2 + (A - 2B - 3B2)Z - (AB - B2 - B3) = 0.
(13)
Vapour Phase, take the highest root: Zv = 0.92034

Liquid Phase, take the smallest root: Zl = 0.96602
Step 6: Calculate the fugacity of each component in the liquid and vapour phase:
ln i = ln( Z B) + ( Z l)
bi
A 1 0.5 N
0.5
a
y
a
l
k
2
i
j j
ij
b 2 2 B a
j= l
(15)
Z + ( 2 + 1)B
ln
Z ( 2 1)B
Fugacity of components in the vapour phase:
fi v = y i P i v
Fugacity of components in the Liquid phase:
fi l = x i P i l
Component
Methane
n-Pentane
n-Decane
n-Hexadecane
Liquid
il
fil
.23811E+01 .67429E+02
.66312E-01 .17708E+01
.19023E-02 .67189E-01
.43075E-04 .19707E-02
Vapour
iv
fiv
.90058E+00 .12150E+03
.36516E+00 .41838E+00
.14686E+00 .41455E-02
.51625E-01 .41965E-04
End of Iteration One
l
Next Iteration: K i = i
iv
(16)
Iterate till the fugacity of every component in the liquid phase is equal to that of vapour
phase (iterate from Step 2.):
25
Final Iteration
Component
Methane
n-Pentane
n-Decane
n-Hexadecane
Feed (Zi)
0.55100
0.11400
0.14600
0.18900
K-Values (Ki)
2.44097
0.19991
1.6774 E-02
1.25099 E-03
Then: Vapour Fraction (Vf) = 0.2710
Component
Methane
n-Pentane
n-Decane
n-Hexadecane
Feed (Zi)
0.5510
0.1140
0.1460
0.1890
Liquid (xi)
0.3962
0.145665
0.199039
0.259147
Vapour (yi)
0.967236
0.029100
0.003339
0.000324
Calculate composition dependent coefficient for z - factor calculations for both liquid
and vapour:
Liquid:
(7)
0.5
( ) (1. k )
a = x i x j a ia j
i=l J=l
ij
b = yi bi
(8)
i=l
Vapour:
i=l J=l
0.5
( ) (1. k )
a = y i y j a ia j
(4)
ij
b = yi bi
(8)
i=l
aP
R2T2
B=
Phase
Liquid
Vapour
a
48.03656
2.11326
A=
bP
RT
(14)
b
0.14799
0.02929
A
6.97233
0.30673
B
0.6577
0.1301
Calculate z - factors of Liquid and Vapour and Vapour Phase (Peng - Robinson EOS):
Z3 - (1 - B)Z2 + (A - 2B - 3B2)Z - (AB - B2 - B3) = 0.
Vapour Phase, take the highest root: Zv = 0.90039
Liquid Phase, take the smallest root: Zl = 0.81539
Calculate the fugaciy of each component in the liquid and vapour phase:
26
(3)
13
ln i = ln( Z B) + ( Z l)
bi
bi
A 1 0.5 N
0.5
a
y
a
l
k
2
i j j
ij
x
b 2 2 B a
j= l
b
Z + ( 2 + 1)B
ln
Z ( 2 1)B
Fugacity of components in the vapour phase:
Fugacity of components in the liquid phase:
Component
Methane
n-Pentane
n-Decane
n-Hexadecane
Molecular volume
P2
Phase
Liquid
Vapour
fi v = y i P i v
fi l = x i P i l
Liquid
il
fil
.22032E+01 .11881E+03
.66155E-01 .13105E+01
.19873E-02 .53832E-01
.45319E-04 .15983E-02
Vm 2
Vapour
iv
fiv
.90269E+00 .11882E+03
.33093E+00 .13106E+01
.11847E+00 .53831E-01
.36226E-01 .1598E-02
2 RT
P
PM
2 RT
2
099039
.81539
Vm gr mole/cc
0.05451
0.04936
Ma g/mole/cc
103.86
18.165
Ps/cc
0.56613
0.08967
Compositions are for final iteration
Component
Methane
n-Pentane
n-Decane
n-Hexadecane
Feed (Zi)
0.5510
0.1140
0.1460
0.1890
Liquid (xi)
0.3962
0.145665
0.199039
0.259147
Vapour (yi)
0.967236
0.029100
0.003339
0.000324
27
REFERENCES
1. Amyx, J.W, Bass, D.M, Whiting, RL, Petroleum Reservoir Engineering, McGraw
Hill. New York 1960
2. Gas Processors Suppliers Association. Engineering Data Book 9th Edition. Tulsa
3. Danesh, A, PVT and Phase Behaviour of Petroleum Reservoir Fluids. 1998
Elsevier. pp 66-77
4. Wilson,G: A Modified Redlich-Kwong EOS,Application to General Physical
Data Calculations,Paper Nu 15C, presented at the AIChE 65th National Meeting
(May 1968)
5. McCain, W.D. The Properties of Petroleum Fluids Penwell. 1st. Edition 1973
Starling, K.E., A.I.Ch.E.Jnl, 1972, 18, 6, 1184-1189.
6. McCain, W.D. The Properties of Petroleum Fluids. Pennwell, 2nd Ed., 1990, p
414-436.
7. Ahmed, T. Hydrocarbon Phase Behaviour, Gulf Pub., 1989.
8. Peng, D.Y., Robinson, D.B., I.E.C. Fundamentals 1976, 15, 1, 59-64.
9. Jhaveri B.S. and Youngren G.K. Three-Parameter Modification of the PengRobinson Equalising State to Improve Volumetric Predictions SPE 13118 (1984)
10. Soave, G., Chem. Eng. Sci., 1972, 27, 1197-1203.
11. Wilson, G.M., "A Modified Redlich - Kwong EOS Application to General
Physical Data Calculations." Paper No 15c presented at the AIChE 65th National
Meeting (May 1968) AI ChE.
28

Equilibrium Ratio Prediction and Calculation

Uploaded by

Copyright:

Available Formats

Equilibrium Ratio Prediction and Calculation

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Equilibrium Ratio Prediction and Calculation

Uploaded by

Copyright:

Available Formats

13