Date of Award

12-2012

Document Type

Thesis

Degree Name

Master of Science in Petroleum Engineering (MSPE)

Department

Chemical and Petroleum Engineering

First Advisor

Dr. Samir Abu-Eishah

Second Advisor

Dr. Mohamed A. Nakoua

Third Advisor

Georgios Kontogeorgis

Abstract

Phase behavior modeling in crude oil includes prediction of thermodynamic properties (such as saturation pressure, density, viscosity, thermal conductivity, etc.) for the vapor and liquid phases. The aim of phase behavior modeling is to establish the accuracy and reliability of the developed equation of state model to predict various other fluid phase behavior properties at high pressure and temperature conditions.

In the first part of this work, a new and reliable phase behavior apparatus designed in Brazil was used for the two-stage recombination process and phase behavior measurements. The recombination of the surface sample fluids (first-stage separator gas and stock-tank oil) was used in this work in order to reproduce the original oil composition. Initially, a precise amount of stock-tank oil is introduced into the PVT cell, and then a pre-calculated amount of the gas from the first-stage separator was injected to the PVT cell. The test was started at a pressure well above the bubble point until getting a monophasic fluid, and then it was reduced stepwise until the first bubble was observed. The observed bubble point pressure was almost the same as that of the field reservoir pressure (2277 psia for well A#22 and 2377 psia for well #33, United Arab Emirates). But the composition of the recombined fluid was found a little far from the required reservoir fluid composition; therefore, the vapor molar ratio was varied until a monophasic fluid was obtained with a fluid composition of minimum deviation from the under test reservoir fluid composition. The optimized vapor molar ratio was 0.5183 and 0.5603 for wells A#22 and A#33, respectively, which are almost the same values given by the service provider. Also the value of the vapor molar ratio with minimum deviation in the methane concentration was found to be about 0.42 for each well when compared to that of the reservoir methane.

In the second part of this work, another experimental setup for both recombination and phase behavior of CO2 measurements is described. After recombination process, a precise amount of CO2 was injected into the PVT cell and then the mixture saturation pressure and the swelling factor were measured. The influence of CO2 addition on crude oil properties was investigated using several molar ratios of CO2/crude oil at conditions close to the oil well mixture conditions. The static-synthetic principles method, which consists of preparing a mixture of known overall composition, was used to observe the fluid phase behavior by changing the pressure of the PVT cell at constant temperature. A vapor-liquid-liquid transition was observed at CO2 mass fractions above 0.3. For CO2 mass fractions between 0.0 and 0.6, the swelling factor ranged from 1.0 to 1.74. All the measurements in the above mentioned tests were performed using an infrared device which allows phase transition detection with a precision of less than 1 bar and 3-4 % error.

Lastly, a phase behavior modeling was carried out using the PVTi [1] and PVTpro [2] modules from commercial (Schlumberger) simulators. The recombination process was simulated using the Soave-Redlich-Kwong (SRK) and Peng-Robinson (PR) equations of state including binary-interaction coefficient and volume-shift corrections. The fluid composition and the reservoir saturation pressure and temperature are the main inputs to the phase behavior simulators. The simulation task for compositional analysis was performed and results compared to available field data. After conversion to the bubble point pressure (to be the same as that of field data), and by comparing the results of both EOS, the PR EOS was selected since it gives more accurate results when compared to the experimental results and field data. The other fluid properties such Z-factor, specific volume, density, viscosity, oil formation volume factor, etc., were found to match well with experimental data, except the saturated liquid molar density, therefore, the Rackett equation was used to estimate the liquid saturated volume and gave very close values when compared to the corresponding field values.

The PVTpro simulator was also used to recombine the surface fluids (first-stage separator gas and stock-tank oil) at the reservoir conditions. The predicted recombined fluid composition from the PVTpro was found in very good agreement with the reservoir fluid composition; with an absolute error of 0.17% and 3.11% for wells A#22 and A#33, respectively. The PVTpro simulator was also used to perform the swelling test by injecting CO2 gas to the recombined fluid (to investigate how much oil is going to swell and to establish the relation between CO2 concentration and saturation pressure). The error in the predicted saturation pressure relative to the experimental values was 8.4% for well A#22 and 6.3% for well A#33.

Download
COinS