Road map and optimum procedures for tuning an Equation-of-State based model against reservoir oils PVT lab measurements.Road map and optimum procedures for tuning an Equation-of-State based model against reservoir oils PVT lab measurements.
Μεταπτυχιακή Διατριβή
Master Thesis
2018-08-292018enIn this study, the fundamental properties were tuned first and then the compositional properties of the fluid were to be tuned, such as gas-oil ratio (GOR) and oil-volume factor (Bo).
Before starting the tuning process, one should first look at the composition of the fluid sample he/she got, and this is due to the fact that the composition of the sample tells which components affects more the tuning process and which affects less. In other words, the higher the concentration of a component in the fluid is, the more it affects the tuning process and it has greater effects on the sample itself.
The next step is to check how far the predicted fundamental properties are from the laboratory extracted data. This will give us a clue on how far the EOS model that is predicted by WinProp is from the real EOS model that suits the specific sample.
Then, if the predicted molecular weight of oil at the atmospheric conditions is far from the laboratory extracted, we should start tuning this property first. Sometimes it is hard to tune this property by simply using the molecular weight of the plus fraction, as we may reach to the bounds of this property and without having a great effect on the predicted molecular weight. So, I suggest modifying the weight factor of gas-oil ratio and oil-volume factor measured from the differential vaporization test and the weight factor of the saturation pressure and the oil density calculated at the saturation conditions.
After that, the oil density at atmospheric conditions should be tuned by using the volume shift of the plus fraction as a regression parameter.
In order to tune the gas-oil ratio and the oil-volume factor for the differential vaporization test, the parameters that should be used in the regression is the critical temperature of the plus fraction and the critical temperature of methane (CH4).
Adjusting the critical temperature of the heavy components will affect the GOR and Bo curves at high pressures, in other words, they will shift these curves in an upward or downward movement mainly the mid of these curves. Whereas, the light components will shift GOR and Bo curves in the beginning of these curves, and will bring the predicted values at low pressures closer to the laboratory values. Also, adjusting the critical temperature of methane will shift in an upward or downward movement of the tail of GOR curve, and will affect a lot the Bo curve measured from constant mass study.
Adjusting the molecular weight of the plus fraction affects the API gravity measured from the separator test, the oil density and the residual oil density measured from the differential vaporization test, but it doesn’t affect the oil-volume factor and the gas-oil ratio measured from the separator test.
Sometimes we might need to change the values of some properties in the fluid without making any regression processes, this can help a lot the tuning process, because CMG WinProp software doesn’t give you the flexibility to tune against some fundamental properties directly, for instance, if we want to tune ng from the separator test, or the molecular weight of oil at atmospheric conditions,…etc. This is because these properties are not included as regression parameter that could be tuned, which means that we can’t do it using the regression option, so instead of that, we can chose the most appropriate parameter to adjust its value using the sensitivity table and then we can proceed with the tuning process.
Proper tuning of EOS models to adequately match the measured PVT study values is anything but trivial. For a given compositional characterization, a developed EOS model provides in fact the mapping of each PVT property (Bo, Rs, ρ, etc) versus the ever changing during depletion overall composition and the prevailing operating conditions. These mappings, given the components selected for characterizing the fluid, can be calibrated by performing a multiple regression against several components physical properties and EOS parameters using very few matching points which are concentrated along the PVT depletion study usually at a single temperature whereas the tuned model is subsequently utilized over a fairly wide range of conditions and overall compositions. It is widely known that an inadequately tuned fluid’s model can lead to poor quality of reservoir engineering calculations (material balance, reservoir simulation, etc).
Until now, an EOS model tuning is considered as an art, it relies more than anything else on the operator’s instinct and expertise and no systematic guidelines appear to be available for its accomplishment.
http://creativecommons.org/licenses/by/4.0/Technical University of Crete::School of Mineral Resources EngineeringAl-Ghishan_Georgios_MSc_2018.pdfChania [Greece]Library of TUC2018-08-29application/pdf3.6 MBfree
Al-Ghishan Georgios
Al-Ghishan Georgios
Varotsis Nikolaos
Βαροτσης Νικολαος
Pasadakis Nikos
Πασαδακης Νικος
Marinakis Dimitrios
Μαρινακης Δημητριος
Πολυτεχνείο Κρήτης
Technical University of Crete
PVT simulation
Oil simulation