Asphaltene flocculation
One of the major problems for the oil industry is the precipitation of heavy organics during production, transportation and the refining or processing of crude oil. Asphaltenes are polar compounds that are stabilised in crude oil by the presence of resins. If the oil is diluted by light hydrocarbons, the concentration of resins goes down and a point may be reached where the asphaltene is no longer stabilised and it flocculates to form a solid deposit. Because the stabilising action of the resins works through the mechanism of polar interactions, their effect becomes weaker as the temperature rises, i.e. flocculation may occur as the temperature increases. However, as the temperature increases further the asphaltene re-dissolves in the oil. Thus, depending on the temperature and the composition of the oil, it is possible to find cases where flocculation both increases and decreases with increasing temperature.
The Infochem model for asphaltenes is based on a cubic equation of state but has an additional term to describe the association of asphaltene molecules and their solvation by resins. The parameters for the model were initially developed from a study of nearly thirty sets of experimental measurements of asphaltene deposition which includes both proprietary and public domain data. The model is complex and to ensure reliable results we recommend that you follow the procedure we suggest until you are familiar with the model and the behaviour of your particular fluid.
The asphaltene model in Multiflash is primarily intended for calculating asphaltene flocculation from live oils. We are aware that many users have only titration data for dead (STO) oils. We have investigated using this titration data to set the asphaltene model parameters and this is discussed later in the case study.
Input Data
The ideal input data for the model are:
- A compositional analysis of the live oil
- The amount of asphaltene in the oil and the ratio of resin to asphaltene, often determined from the SARA analysis
- One set of flocculation conditions for the asphaltene flocculation
- Bubble point (optional) to “tune” the petroleum fraction properties.
Recently we modified Multiflash to take into account the fact that the SARA will have been measured for the stock tank oil rather than the total reservoir fluid. If you load files characterised in previous versions of Multiflash, where you have modified the SARA to represent the live oil, the SARA will be taken to be for the STO. To correct this you will need to return to the original measurements or to carry out a P,T flash to re-calculate the STO value.
For some samples you may only know the absolute weight % of asphaltenes and resins in the total live oil but not the full SARA analysis. In these cases you can just use the weight % of resins and asphaltenes in the stock tank oil and ignore the text boxes for saturates and aromatics in the PVT Lab Fluid Analysis dialog box. For the asphaltene model, the saturates and aromatics part of SARA are only used to normalise the weight % of resins and asphaltenes.
The % asphaltene is taken to be that precipitated by heptane. Some laboratories report the wt% asphaltene precipitated by pentane. It is difficult to give exact guidance on how to convert between pentane and heptane precipitation as this can vary from oil to oil and laboratory to laboratory. In general we have found that the wt% asphaltene precipitated by pentane is approximately twice that precipitated by heptane. However, ratios can vary from 1.3 to 2.7.
If you do not have the complete data set we recommend we have developed correlations to assign the required parameters. The minimum set of data in this case are:
- A compositional analysis of the live oil
- Reservoir temperature
- Bubble point (optional), to “tune” the petroleum fraction properties.
Obviously the more data available the better the model predictions.
The prediction of asphaltene flocculation is not as sensitive to the characterisation of the fluid as the wax models. However, we suggest you consider using a common characterisation procedure. We recommend that in the PVT Analysis facility you start the pseudocomponent split at C6 and split the fraction into 15 components, regardless of the original experimental distribution.
Finally, to model asphaltene flocculation successfully you need to adjust one model parameter to match actual flocculation data. If you have such experimental data available then you match to this. If you do not have any data there are two options. We have modified the screening procedure suggested by De Boer et al to adjust the parameter based on a knowledge of the reservoir conditions or you can use STO titration data using heptane if this is available,
Defining the asphaltene model
The cubic equation of state, RKSA, is defined as part of the model, it is not possible to choose a different fluid phase model. This flexibility may be added at a later date.
To set up use the Select/Model Set option.
Select the asphaltene tab, Click on Define Model to select the asphaltene model and Click on Close to load the model.
The default phases for the asphaltene model are gas, liquid1 and asphaltene. However, you can add a water phase if you wish. In Windows this can be done very simply by ticking the water check box. In the asphalt.mfc file we have commented out the lines defining the water phase and the key component for this phase. If you wish amend the .mfc file you must edit it to remove the # sign before the relevant lines. However the asphaltene model parameters should be produced for the fluid composition excluding water, more caution should be taken when plotting the asphaltene phase envelopes as the presence of water in the mixture makes the phase equilibrium calculation more complex.
The next step is to characterise your fluid. Go to the PVT Analysis form, described in detail in “PVT Analysis” on page 35. Enter the fluid analysis and set the Start pseudocomponents to C6 and the number of pseudocomponents to 15. Enter any data you have on the molecular weight and/or specific gravity.
The final step in the characterisation is to enter any data you have on the weight of asphaltene in the oil and the ratio of resin to asphaltene. The example we are going to look at is based on a supplied problem set up file, asphex.mfl. We have assumed in this example that a full SARA analysis is available, but later in the case study we will go back and look at the options if you don’t have this much data.
If you load asphex.mfl the PVT Analysis form will look like this
Once you are happy the data is correct Click on Do Characterisation. In this case a message box will tell you the characterisation has been successfully completed and show a plot of the data and fitted distribution. You can then Click on OK and on Close to go back to the main window.
If you have any warnings or errors associated with your own examples then the relevant message box will appear. Since the release of the asphaltene model we have had feedback from several users. In our original database of asphaltene measurements the resin/asphaltene ratio was always greater than 2.5. However, some of our users have reported values below this, possibly because the SARA analysis was based on a different experimental technique. Depending on the actual R/A entered you may find that model parameters cannot be generated. We have added a warning message if R/A is below 2. Initially, on characterisation you will be warned of a possible problem
You can either ignore this warning message and see if model parameters can be produced, you can increase the resin/asphaltene ratio manually or delete the Resin amount from the SARA and tick the Estimate RA box and the model will generate a default R/A ratio.
After characterisation the asphaltene component is named as ASPHALTENE, and the resin component(s) by an “R” prefix, e.g. R65+.
The next stage is to use the matching facility to “tune” the pseudocomponent properties and fix the one adjustable model parameter. We recommend that if you have bubble point data available you tune the petroleum fraction properties to match this. If you have multiple bubble points then we suggest you use the Tools/Matching Bubble point form. However, if you have only one measured bubble point we have incorporated this into the asphaltene matching form.
In practice we have found that you do not always have to match to bubble point data to model the asphaltene data, although you almost always need such matching for light oils.
However, we do recommend use of the matching procedure for the asphaltene phase. Although we have supplied a default for the case where you have no information to fix the adjustable model parameter we cannot recommend any of the subsequent results. But we do recognise that you may want to use the asphaltene model for screening purposes and we have developed a procedure to adjust this parameter based on a knowledge of the reservoir conditions or STO titration data.
For this example we have several bubble points. Enter these into the Tool/Matching/Bubble point table
Click on match to generate the comparison of predicted and matched data.
Use the Tools\Matching\Asphaltene Phase to display the dialogue box, and enter the values to obtain the asphaltene model parameters. Initially, we have used the reservoir conditions (241°F, 8500 psi), although the dialogue box allows for two flocculation points or titration data.
Click on Match and Close: the model parameter values will be displayed in the main window. The asphaltene model has now been defined.
Calculating asphaltene flocculation conditions
Once the asphaltene model has been defined and the parameters generated you can carry out flocculation calculations.
If you are starting from the position where you only know the reservoir conditions and have no information on specific flocculation conditions we suggest you should use the phase boundary tracer to get an overall picture. This is extremely useful, but for these complex calculations it can be difficult to find convergence or even starting values. We recommend you follow this procedure.
First plot the bubble point line. Use the Phase Envelope facility and plot the boundary for zero gas phase, in this case with the initial values for pressure set to 1000 psi.
Click on Plot to display the phase boundary. This will usually show a point, or points, of discontinuity at high pressure, labelled D. This is the point where the asphaltene deposition envelope crosses the bubble point line.
These points can be very useful for setting an appropriate starting pressure for the asphaltene deposition envelope or providing starting values if these are required. For this example go back to the Phase Envelope and this time set the phase to asphaltene, the fraction to zero, the solution type to unspecified and the Initial value for pressure to 3500 psi with pressure increasing. Ask for more points to be plotted until the asphaltene boundary is complete.
For other examples you may have to set the pressure to decrease or to plot the upper and lower boundaries separately. The upper boundary uses the Unspecified solution or Upper retrograde type solution, the lower boundary the Lower retrograde type of solution. Alternatively you can try specifying temperature rather than pressure and/or providing a starting value. We have found asphaltene boundaries most difficult to plot for very light oils.
If you have a known set of conditions and want to see if, and how much, asphaltene is present you can use a simple P,T flash. Enter the temperature and pressure, for example 200F and 4000 psi, Click on the P,T icon, or Select the P,T flash from the Calculate\Standard flash menu. The phases present, and the composition and amount of each phase, will be reported. Before doing this you may find it useful to set the units for amounts to mass as this usually reflects the units of measurement.
Flash at fixed P and T:
T (degF) = 241.000 P (psi ) = 4000.00
NO. PHASES = 2 CONVERGED STABLE
COMPONENT OVERALL PHASE1 PHASE2
LIQUID1 ASPHALTENE
g g g
NITROGEN 3.000000E-02 2.999870E-02 1.301098E-06
H2S 2.000000E-02 1.999677E-02 3.233066E-06
CO2 0.550000 0.549932 6.832952E-05
METHANE 8.32000 8.31950 5.045115E-04
ETHANE 3.97000 3.96974 2.631473E-04
PROPANE 3.73000 3.72978 2.202586E-04
ISOBUTANE 0.680000 0.679967 3.257557E-05
N-BUTANE 2.50000 2.49987 1.334123E-04
ISOPENTANE 1.07000 1.06995 5.100751E-05
N-PENTANE 1.84000 1.83991 8.916160E-05
C6-10 18.6155 18.6142 1.292815E-03
C10-13 9.13541 9.13500 4.095346E-04
C13-15 7.33287 7.33265 2.159847E-04
C15-18 6.49355 6.49344 1.111808E-04
C18-20 5.67482 5.67476 5.992178E-05
C20-23 5.15517 5.15514 2.989052E-05
C23-26 4.64417 4.64416 1.253549E-05
C26-29 4.18401 4.18400 4.609279E-06
C29-33 3.77167 3.77167 1.435925E-06
C33-37 2.96233 2.96233 3.057482E-07
R33-37 0.358907 0.358903 4.260846E-06
R37-42 2.88963 2.88962 6.401622E-06
R42-49 2.43232 2.43232 5.633672E-07
R49-59 1.93459 1.93459 2.033747E-08
R59-65 0.608010 0.608010 2.496835E-10
R65+ 0.702108 0.702108 7.285689E-11
ASPHALTENE 0.394936 0.250979 0.143957
Total(g ) 100.000 99.8525 0.147474
If you want to know the pressure at which asphaltene will start to deposit at any given temperature then you should use a flash at fixed phase fraction and temperature. Set the temperature, in this case 200°F. Again Click on the icon or select the calculation option and the dialogue box will appear.
Set the molar phase fraction to zero. To calculate the pressure at which asphaltene will first appear for pressures above the bubble point, select Unspecified or Upper retrograde as the solution type and Click on Do Flash. Multiflash will calculate the pressure on the upper asphaltene phase boundary, in this case 7930 psi. To obtain the pressure for the lower asphaltene phase boundary, below the bubble point, follow the same procedure but set the Type of solution to Lower retrograde. In this case the reported pressure is 1879 psi.
You can determine the amount of asphaltene flocculation at any set of P,T conditions using an isothermal flash as described earlier, You may like to get an idea of the amount of flocculation along any isotherm by carrying out a series of 4 or 5 calculations. The maximum flocculation will occur at the boiling point.
Sensitivity of calculations to variation in input data
Choice of Analysis method
The calculations were carried out by matching to reservoir conditions. The plots below show the effect of including the n-paraffin distribution. After any re-characterisation the data must be matched again, in this case the bubble points and the asphaltene parameters.
The inclusion of the n-paraffins for this case makes a small difference to the predicted ADE.
Data Availability
This example of asphaltene flocculation was based on a data set which comprised the compositional fluid analysis, a SARA analysis, bubble points and the reservoir conditions. In this case, the bubble points were close to the unmatched predictions and so matching to bubble points might not be expected to make a major difference to the asphaltene predictions. However, even in this case you can see a noticeable effect from the matching.
There are other data that may be missing and have an effect on predictions.
No information on the amount of asphaltene in the oil
We have provided a procedure to estimate the weight % of asphaltene in the oil if this data are not available. In this case in the PVT Analysis box you should tick the box for estimate RA, and then Click on Do Characterisation and Close.
Rematch the bubble point and asphaltene phase at the reservoir conditions as before and plot the asphaltene deposition envelope (ADE). It is important to include this step; the matching procedure is cancelled when the fluid is re-characterised.
The default procedure estimates both the weight % of asphaltene and the resin/asphaltene ratio. For this particular example the predicted weight % of asphaltene is very close to the reported value with 0.7 wt% asphaltene predicted compared to the experimental data of 0.5 wt%.
No resin – asphaltene ratio
Even if you do not have a SARA analysis you may have the weight % asphaltene and only need to estimate the Resin/Asphaltene ratio. Proceed as before, enter the wt% asphaltene (0.5 %) in the correct text box in the PVT Analysis but still tick the Estimate RA box. Repeat the matching of bubble point and reservoir conditions again and plot the ADE.
Although the estimated R/A ratio is lower, at 13, than the original, at 22, the resultant ADE are very close. This is the result of two factors: once you reach a certain level of R/A the effect of increasing the R/A has a reduces plus matching to a specific flocculation point or reservoir condition compensates to some extent for changes in R/A.
No reservoir pressure
If you only have the reservoir temperature we have included a facility to estimate this. Simply enter the bubble point data and reservoir temperature as before and initiate the matching procedure.
For this particular example the resultant ADE is reasonably close to the ADE calculated from the real reservoir conditions.
As a corollary to this we have noticed that you usually generate very conservative ADE when you have a very over-pressured reservoir. If you have a bubble point measurement at the same temperature as the reservoir temperature and the reservoir pressure is more than 2.5 times the bubble point a warning message is generated.
You can continue to match to your reservoir conditions although it may also be beneficial to generate the model parameters with an estimated reservoir pressure to see the likely sensitivity.
If the bubble point is matched at a different temperature to the reservoir temperature no warning is issued.
No reservoir or flocculation conditions
If you do not have either reservoir or flocculation conditions then there are two default options for generating the asphaltene model parameters.
If you have entered a bubble point in the Asphaltene matching but nothing for Reservoir conditions or Asphaltene flocculation then Multiflash will assume that the reservoir temperature is the same as the bubble point temperature and proceed to estimate the reservoir pressure as above.
If nothing is entered for bubble point, reservoir conditions or asphaltene flocculation the model parameters are generated from correlations based on data held in our database. The results from using this route are vary variable, depending on the fluid analysis and we cannot recommend its use. In this case the result would be a much more conservative ADE.
Matching to asphaltene deposition data
The assumption in this case is that you have more data than our basic example, real deposition data either from field conditions or a asphaltene flocculation measurement. In this case we have two flocculation points at 241°F and 6921 psi and 120°F and 9150 psi. Simply enter these in the Asphaltene matching box instead of the Reservoir conditions.
For comparison purposes we have matched to each point individually and then to both points.
With the new graphics package it is easy to add the actual data to the plot. Using the Add Data option in the phase envelope plotter you can add the measured flocculation points and the reservoir conditions.
Gas injection
It is known that as gas is injected into a reservoir the likelihood of asphaltene flocculation is increased. The asphaltene model predicts this trend correctly. Return to the original ADE, calculated from the asphex.mfl input file with matched bubble points and reservoir conditions. You can mimic gas injection by increasing the amount of methane by adding more moles of methane in the drop down composition box. If you increase the amount of methane from 8.32g to 12g and replot the ADE you will see that the fluid bubble point line is at higher pressures and the ADE has expanded.
When looking at the effect of gas injection you should, of course, not rematch the fluid bubble point or asphaltene phase deposition as doing this will alter the petroleum fraction properties and model parameters.
You should not use the PVT Analysis GOR option to add the injection gas to the reservoir fluid. Any re-characterisation cancels the properties and parameters derived from earlier matching and, as you now have a different fluid, the values of bubble point and reservoir conditions used for matching are no longer valid. If you have a complex injection gas and want to study the effect of different gas injection rates then we suggest the use of an Excel spreadsheet or use the blending procedure in Multiflash to blend the injection gas and the reservoir fluid. Please note that the reservoir fluid with asphaltene model tuned and bubblepoints matched should be selected for the model definition in the blending form so that the asphaltene model parameters are based on the original reservoir fluid. If the asphaltene flocculation data are available for the blended mixtures, the data should be matched after blending.
Titration
The Infochem asphaltene model was intended for use in predicting the asphaltene flocculation of live oils and the model parameter generation based on asphaltene studies of live fluids. However, live oil asphaltene studies can be expensive, particularly with the requirement to obtain and transport bottom hole samples. Some of our users have asked whether titration measurements on dead oils can be used to generate the model parameters. To date we have only been able to obtain limited samples of titration data and have traced only one oil, in the public domain, where there is any information on both asphaltenes in the live oil and reported titration on the associated stock tank oil (STO), enabling us to compare results. However, we understand that some of our users have applied this approach successfully, and the procedure for using titration data has been automated.
The studies have been limited to titration with heptane.
Our example is based on the titration.mfl file provided. The file includes the live oil composition and the wt% of asphaltene and resins. The reported value of asphaltene was 1.9 wt%, that of resins is 16.1 wt% for the STO. Characterise the fluid composition as usual and the return to the main menu, Tools/Matching/Asphaltene phase. The reported onset amount of heptane to just cause asphaltenes to flocculate from the STO at ambient conditions is 1.4 cm3 per g tank oil. This has been converted to .962 g heptane using the known density. Enter this value and click on match.
The asphaltene model parameters will be reported in the main window as usual and the ADE plotted. The resultant ADE is compared to those generated from matching to a known flocculation point of 54.4 °C and 200 bar and to a combination of reservoir temperature (54.4°C and bubble point (54.4°C and 156.2 bar)
The ADE predicted from matching to titration of the STO is very close to the ADE from flocculation measurements and both are less conservative than using reservoir conditions to provide the model parameters. It is believed this has been the experience for other fluids.
If your titration data does not include the amount of heptane just to initiate flocculation and it has to be deduced from the other titration results then the procedure for parameter generation is slightly more complicated and requires the use of an Excel spreadsheet. In Multiflash for Windows either characterise the STO, if this composition is provided or flash the characterised live oil to STO conditions, using the RKSA model set to ensure that no separate asphaltene phase is formed. If you have bubble point data it is important that you match to this before flashing to STO conditions. Using the STO composition, change the model set to asphaltene, match the asphaltene flocculation to ambient conditions and save the problem using the File/Save Problem Setup option .
You then need to create an Excel worksheet to read in this .mfl file. Details of how to do this are described in the Excel manual, but we have provided an example spreadsheet, titration_sto.xls.
For our example we have generated the file STO.mfl from the fluid used in our titration example. This is the file that should be called from the Excel spreadsheet. In the spreadsheet you then need to do two things: add a new component to the list, heptane, and add a command line describing the asphaltene parameters. This can be copied from Multiflash for Windows by using Tools/Show/Problem to display the commands. The command line can be copied and pasted to the spreadsheet but for fitting purposes it must be set up so that the RAP parameter appears in a single cell so that it can be optimised using the Excel Solver, e.g.
include c:\work directory\exampleoil.mfl";
model MREFASPHALTENE RAEQUIL DATA AAPREEXP 1.00000000 AAEXP 1.00000000 RAPREEXP
.634756
RAEXP
0.96636
;
component heptane;
The spreadsheet, titration_sto.xls, is set up to optimise the value of RAPREEXP (RAP) using the Excel Tools/Solver by comparing the calculated wt% of asphaltene deposited for given amounts of heptane to the experimental wt% deposited.
For our particular example the data reported in the paper included a live oil and a STO composition, a wt% asphaltene for the STO, an asphaltene flocculation point and five points for the heptane titration.
The reported titration data are plotted below
Fitting to the onset flocculation point using the matching facility produced the following parameters
| RAP: | 0.94568 |
| RAE: | 0.96636 |
Whereas fitting to the other four points in Excel gave parameters
| RAP: | 0.95534 |
| RAE: | 0.96636 |
These parameters represent the STO titration data well, but the amount of heptane to just initiate flocculation is slightly higher. You can check the predicted value for the amount of heptane required for the onset of asphaltene flocculation using a tolerance calculation with heptane as the second fluid. The predicted amount is .986 g/g oil rather than .962.
It is clearly preferable to generate the live oil ADE from live oil data. Some predictions are possible from titration data but is important that all data are compatible, particularly the compositions of the STO and the flashed liquid and the physical properties of pseudo components by matching to bubble point data in both cases if you have it.
For this example we also had a measured flocculation point and you can see the ADEs resulting from the different approaches.
