Computer Modelling Group, Ltd.

Burke Oil 1 plus fraction characterization

Phase Properties (Solvent Mole Fraction = 0.0000)
1.20
212.00 deg F
1.00
Solid Precipitate (wt %)

0.80

0.60

0.40

0.20

0.00
0 1000 2000 3000 4000 5000 6000 7000
Pressure (psia)

TUTORIAL 2007

Asphaltene Precipitation Modelling

Using

WinProp

WinProp Asphaltene 1
This case study describes the procedure for modelling the
precipitation of asphaltene from a live reservoir oil due to pressure
depletion. The thermodynamic model used to describe the
precipitation of asphaltene and the flash algorithm for solving the
equations are given in references 19 and 20. Please refer to the
Flash Calculations chapter for further information on
Asphaltene/wax modelling.

The steps required to develop a precipitation model are:

1. Fluid characterization
2. Regression on fluid PVT
3. Specification of solid model parameters
4. Prediction of precipitation behavior

Fluid analysis and asphaltene precipitation data are taken from

Burke, Hobbs and Kashou, Measurement and Modeling of
Asphaltene Precipitation, Journal of Petroleum Technology,
November 1990, pp. 1440-1446. The following data for Oil 1
taken from Table 1 of Burke et al will be used in this example.

Component Oil 1
Nitrogen 0.57
Carbon Dioxide 2.46
Methane 36.37
Ethane 3.47
Propane 4.05
i-Butane 0.59
n-Butane 1.34
i-Pentane 0.74
n-Pentane 0.83
Hexanes 1.62
Heptanes plus 47.96
Total 100.00
C7+ molecular weight 329
C7+ specific gravity 0.9594
Live oil molecular weight 171.4
Stock tank oil API gravity 19.0
Asphaltene content in stock tank oil, wt% 16.8
Reservoir temperature, F 212
Saturation pressure, psia 2950

WinProp Asphaltene 2
Fluid Characterization

To begin this case study, a data set has been prepared to

characterize the fluid by defining the compositions of components
up to C6 and pseudo-components describing the C7+ fraction.
The data set is named Case_study-3-split.dat and is located in the
WinProp course directory. You can begin by using this file, or by
constructing your own data set as described in Case Study
Number 1. From the table above, the composition data to C6 has
been used, and a Plus Fraction Splitting calculation has been
specified with the C7+ molecular weight and specific gravity. The
plus fraction will be lumped into 4 pseudo-components, and the
Lee-Kesler critical property correlations will be used.

Perform the splitting calculation either by running your own data

set, or by running the Case_study-3-split.dat data set.

Update the System Component Specification

After splitting, the equation of state model can now be tuned to

any available PVT data via regression. First, the component
specification must be updated to reflect the results of the splitting
calculation. This is done by selecting File|Update component
properties from the file menu, which will modify the first three
forms in the data set. Save the data set under a new name to
avoid overwriting the original file by selecting File|Save As ... and
entering a new name. Remove the splitting calculation from the
data set by selecting the row labelled Plus Fraction Splitting and
pressing the Delete key or by selecting Edit|Delete from the main
menu. Open the Titles/EOS/Units form and enter a new comment
and descriptive title. Open the Component Selection/Properties
form and click the check box to Use temperature-dependent
volume shifts. This will allow good initial predictions of the
volumetric properties of the fluid. After closing this form, the
calculation options required to perform the regression can be
added to the data set.

WinProp Asphaltene 3
Set up the Regression Data Set

Select the first blank row underneath the Composition row. Enter
a Regression Parameters form into the data set by selecting
Regression|Start from the main menu or by clicking the Strt Reg
button the toolbar. Then enter a Saturation Pressure calculation
by clicking the Sat Pres button, and an End Regression form by
clicking End Reg. Open the data entry form for the Saturation
Pressure calculation and enter a temperature of 212 F,
saturation pressure estimate of 2500 psia, and an experimental
saturation pressure measurement of 2950 psia. Close this form
and open the Regression Parameters form. Go to the
Interaction coefficients tab and click the option button for Select
from list, then select HCIntCoefExp 1(1.2) from the list. This
selects the hydrocarbon interaction coefficient exponent as a
regression variable. For more information about this parameter,
please refer to the Components section of the Users Guide. Go
to the Regression Controls tab and set the Convergence
tolerance to 0.00001 to achieve a good match to the experimental
data.
Click OK to close the Regression Parameters form. The data set
should now look like the following:

Run the data set, and then view the output from the run. The
regression summary table at the end of the output file shows that
an exact match to the saturation pressure was achieved. The
model is now ready to be modified for asphaltene precipitation
prediction. First update the component properties as before, then
save the file under a new name.

WinProp Asphaltene 4
Specification of Asphaltene Component

Again, a new title and descriptive comment can be added to the

Titles/EOS/Units form if desired.
The asphaltene component is specified by splitting the heaviest
component of the oil into precipitating and non-precipitating
components. These two components have the same critical
properties and acentric factors; however the precipitating
component has higher binary interaction coefficients with the light
components up to about C5. This is achieved in WinProp by
adding a new component to the component list, then copying the
properties of the heaviest component, onto the newly added
component.
Open the Component Selection/Properties form. Begin by
unchecking the check box for Use temperature-dependent
volume shifts, this allows the volume shifts to be fine tuned to
density data. In the component table, click on the row directly
underneath the component C31+. Enter any hydrocarbon
component from the library by selecting Options|Insert library
component, clicking on FC6, then clicking OK. Select the entire
row of C31+ properties by clicking on the row label 14 at the left
side of the table. Copy the properties to the clipboard by using
Ctrl-C or selecting Edit|Copy from the menu, as shown below.

WinProp Asphaltene 5
Now click on the first column in the row for the component that you
just entered, i.e. click on the component name FC6, and use Ctrl-
V or Edit|Paste to paste the properties onto the new component.
The component names can now be edited, for example C31A+
and C31B+. To be able to specify interaction coefficients with the
other components individually for the C31B+ component, as
opposed to calculating them with the hydrocarbon interaction
coefficient exponent, the HC flag in the column next to the
component name can be set to 0. The component list should now
appear as follows:

Click on the Int. Coef. tab, and click on Show HC Int. Coef. to
view the magnitude of the binaries between C31A+ and the light
components. As described in references 19 and 20, the binaries
for the precipitating component must be considerably higher than
those for the non-precipitating component to give the correct
shape of the precipitation curve below the bubble point. Values on
the order of 0.2 are expected to give good results. For C31B+,
enter the same binaries as C31A+ for CO2 and N2, then enter
values of 0.2 for the interactions with C1 through nC5. All other
values for C31B+ should be left as zero.

WinProp Asphaltene 6
Click OK to complete the component specification.

Specification of Asphaltene Composition

The mole fraction of the asphaltene component can be determined

w Asph MWOil
from the relation: x Asph . From the output of the
MW Asph
regression run, the molecular weight of the oil is calculated as
171.343, as compared to the reported value of 171.4. The
asphaltene content of the stock tank oil is given as 16.8 wt%. As
the weight fraction of gas in the live oil is usually small, the value of
16.8 wt% will be used for the live oil. From the component table, the
molecular weight of the C31B+ component is 665.624. Then the
mole fraction for the precipitating component is:
0.168 171.343
x Asph 0.04324607
665.624
Open the Composition form and enter this value for the mole
fraction of C31B+. Subtract this amount from the original mole
fraction of C31+, and enter 0.0742466 as the mole fraction for
C31A+. Click OK to close the form.

WinProp Asphaltene 7
Specification of Additional Regression Data

When the heaviest component is split into precipitating and non-

precipitating parts, and the binaries for the asphaltene component
are adjusted, the fluid phase behavior predictions will be affected.
For this reason, regression must be performed again to ensure that
the model will predict the correct fluid and solid phase behavior. In
this data set, we will add regression on the stock tank oil API as
well as the saturation pressure. Click on the End Regression row,
then click the SEP button to insert a separator calculation into the
data set. Open the Separator form and enter an estimate for the
saturation pressure of 2500 psia and a temperature of 212 F in the
first column of the separator specification. Go to the Experimental
Data tab and enter 19.0 for the stock tank oil API and click OK.
To match the experimental API, we will modify the volume shift
parameters of the heavy fraction pseudo-components. As the
volume shift of the asphaltene component affects the amount of
precipitate, as well as the liquid density, it will not be adjusted
during regression. Open the Regression Parameters form and
click in the Vol Shift column in the rows corresponding to the
components C07-C15, C16-C25 and C26-C30 to select these
parameters as regression variables. Click OK.

Specification of Reference Fugacity for Asphaltene

Model

The equation describing the fugacity of the solid component

in the solid phase is given in the Flash Calculations chapter
of the Users Guide. Writing this equation for isothermal conditions
gives:
ln f s ln f s* v s ( p1 p o ) / RTo

Where fs* is referred to as the reference fugacity, at the

reference conditions Po and To. vs is the molar volume of the
solid. The reference fugacity is usually set equal to the fugacity of
the precipitating component calculated by the equation of state at
an experimentally determined asphaltene precipitation onset
pressure for a given temperature. This ensures that the model will
predict the correct onset pressure. At other pressure conditions,

WinProp Asphaltene 8
the fugacity of the solid component in the solid phase is compared
to the fugacity of the solid component in the liquid phase as
predicted by the equations of state; if the fugacity of the solid
component in the solid phase is lower then asphaltene will
precipitate.
With modern solid precipitation detection systems, onset
pressures can usually be determined quite accurately. In the
Burke et al data, an exact onset pressure is not given, instead, a
pressure point above the saturation pressure at which a small
amount of asphaltene precipitates is determined. The data given
(in Burke et al Table 5) is 0.402 wt% asphaltene precipitated at
4014.7 psia and 212 F. To use this data, the reference fugacity is
determined at the given pressure and temperature, but with the
precipitated amount of asphaltene removed from the system. The
reported wt% of precipitated asphaltene must again be converted
to a mole fraction to enter into WinProp. Using the same formula
and molecular weights as before:
w Asph MWOil
x Asph
MW Asph

The mole fraction of precipitated asphaltene is:

0.00402 171.343
x Asph 0.001034817
665.624

Click on the Composition form, then use Ctrl-C to copy the form.
Click on the row underneath the End Regression form and use
Ctrl-V to paste a copy of the composition form. Open the new
Composition form, and subtract the amount of precipitated
asphaltene from the mole fraction of component C31B+, the
resulting mole fraction should be 0.042211253. When you click
OK, a warning that the composition no longer sums to 1.0 will be
issued, this can be ignored as all compositions will be normalized.
Click on the row below the new composition form and then click
the ASP/WAX button on the toolbar to enter a new
Asphaltene/Wax Modelling form in the data set. Open the form
and enter the reference conditions of 4014.7 psia and 212 F on
the first tab. Click on the Ref. State tab and select CALCULATE
for the Reference Fugacity Specification. This will set the
reference fugacity equal to the fugacity of the precipitating

WinProp Asphaltene 9
component in the liquid phase calculated by the equation of state.
Click OK. The data set should now appear as follows:

Specification of Solid Molar Volume

As discussed in reference 19, the solid molar volume should be

set to a value slightly higher than the molar volume for the
precipitating component predicted by the equation of state. At this
point we can run the data set to check the regression on the fluid
PVT data and to view the solid molar volume predicted by the
EOS. Run the data set and view the output file. The regression
summary table shows that both the saturation pressure and stock
tank API are matched exactly. At the end of the file, a listing of the
parameters of the solid model is given. The solid molar volume is
given as 0.65946 L/mol. A good initial value to enter for the solid
molar volume is 0.67 L/mol. Return to the data set and open the
Asphaltene/Wax Modelling form and go to the Ref. State tab.
Scroll over in the grid showing the properties of the C31B+
component and enter 0.67 under Molar Vol. This molar volume
will be used by all subsequent asphaltene precipitation
calculations. Click OK close the form.

Prediction of Asphaltene Precipitation

To perform predictions of asphaltene precipitation, we need to use

the whole live fluid composition, not the composition adjusted for
the reference fugacity calculation. Click on the first Composition
row in the data set (row 3) and use Ctrl-C to copy it to the

WinProp Asphaltene 10
clipboard. Now click on the row underneath the Asphaltene/Wax
Modelling form and use Ctrl-V to paste the Composition form.
Then add a new Asphaltene/Wax Modelling form at the end of
the data set. Open the form and enter a temperature of 212 F.
We want to predict the amount of asphaltene precipitated at
various pressures. Enter a Pressure of 14.7 psia, a Pressure
Step of 200 psia and a No. of pressure steps of 31. This
specification results in flashes being performed every 200 psi from
14.7 to 6014.7 psia.
Click on the Ref. State tab and set the Reference Fugacity
Specification to PREVIOUS. This specifies that the reference
fugacity for the asphaltene model will be set to that value
determined in the previous asphaltene flash, where the reference
fugacity was set to CALCULATE. The molar volume, reference
pressure and reference temperature set previously will also be
used. Now go to the Plot Control tab, click on the X-Y Plots
option button, then check the Plot weight % solid phase
checkbox. Click OK to close the form. The final data set is now
ready to run, and should appear like the one below.

Run the data set, and click on the button with the Excel icon on the
toolbar to create the plots. You should see that the shape of the
asphaltene precipitation curve from the upper onset pressure to
the saturation pressure shows the expected trend of increasing
precipitation with decreasing pressure. Note also that the
predicted amount of asphaltene at the reference pressure of
4014.7 psia is exactly equal to the experimental value of 0.402
wt%. The shape of the curve at lower pressures is incorrect. The
final step in this case study discusses how to adjust the solid

WinProp Asphaltene 11
model parameters to achieve the correct shape of the precipitation
curve.

Burke Oil 1 plus fraction characterization

Phase Properties (Solvent Mole Fraction = 0.0000)
1.20

212.00 deg F
1.00
Solid Precipitate (wt %)

0.80

0.60

0.40

0.20

0.00
0 1000 2000 3000 4000 5000 6000 7000
Pressure (psia)

Adjustment of Solid Model Parameters

As the solid precipitation model used in WinProp is

thermodynamic, as opposed to kinetic, reversibility of precipitation
is possible, i.e. precipitated solids can redissolve in the liquid
phase. This phenomena has been observed in the laboratory for
pressure depletion. Usually the maximum amount of precipitation
occurs around the saturation pressure of the fluid. Below this
pressure, liberation of gas from the oil changes the solubility
parameter of the liquid phase and allows redissolution of the
precipitated asphaltene. It is possible that all of the precipitated
asphaltene will go back into solution at sufficiently low pressures.
The parameters that control this behavior in the solid model are
the solid molar volume and the interaction parameter between
the precipitating component and the light ends of the oil.
Increasing the solid molar volume increases the maximum amount
of precipitation at the saturation pressure. Increasing the
interaction parameter with the light ends will force the asphaltene
to redissolve at lower pressures.
The experimental data given in Burke et al (Table 5) indicates the
maximum amount of precipitation from this fluid should be about

WinProp Asphaltene 12
1%, and that the amount of precipitation should decrease to 0.403
wt% at 1014.7 psia.
Judging from the results of the initial run, it is seen that the molar
volume of the solid must be increased slightly to increase the
maximum amount of precipitation from approximately 0.8 wt% to 1
wt%. The interaction parameter between the precipitating
component and the light ends must also be increased to give the
correct shape of the precipitation curve below the saturation
pressure.
As noted earlier, performing the regression within the asphaltene
modelling data set allows the model to predict the correct fluid
PVT behavior when the interaction parameters for the asphaltene
component are changed.
Adjust the solid molar volume first to achieve the desired
maximum amount of precipitation. In this case, a value of 0.675
L/mol was found to give good results.

WinProp Asphaltene 13
After this step, adjust the interaction parameters between the
C31B+ component and the components C1 through nC5.
Using a constant value of 0.224 for all of these interactions gives a
good shape to the precipitation curve, although it doesnt match
exactly the experimental data.

For comparison with your work, the final model is given in the
template data set Case_study-3-asph.dat. The asphaltene
precipitation plot from this data set is shown below.

Burke Oil 1 plus fraction characterization

Phase Properties (Solvent Mole Fraction = 0.0000)
1.20
212.00 deg F
1.00
Solid Precipitate (wt %)

0.80

0.60

0.40

0.20

0.00
0 1000 2000 3000 4000 5000 6000 7000
Pressure (psia)

WinProp Asphaltene 14