This paper investigates the effectiveness of Equations of State (EOS) when used to simulate Pressure-Volume-Temperature (PVT) experiments. The phase behaviour and volumetrics of six EOS and three EOS modifications (9 models in total) has been examined using a newly developed package inside the MATLAB software. The simulator is an expandable system designed to compare different EOS and fluid characterisation schemes. Comparison will be made between the Soave-Redlich-Kwong (SRK), Peng-Robinson (PR), Adachi-Lu (AL), Patel-Teja (PT), Schmidt-Wenzel (SW) and Esmaeilzadeh-Roshanfekr (ER) EOS, for which the SRK, PR and ER equations optionally include published volumetric shift parameters. Local Australasian fluid samples are studied to illustrate strengths and weaknesses of different EOS, using several fluid types. It was found that the Esmaeilzadeh-Roshanfekr EOS is most accurate for predicting gas-condensate properties, while the original SRK and PR equations remain reliable for oil samples.
The purpose of the study is twofold:
Increasingly companies are optimising experimental PVT studies by simulating subsequent experiments using EOS1,2, or they are using preliminary simulation to evaluate the need for further experiments.
This is particularly true for multi-contact miscibility experiments such as the Rising Bubble Apparatus or slim-tube test, which are time-consuming and costly. Furthermore, (2) there is reason to believe that thermodynamic models which provide better representation of PVT experiments without matching will perform best after complete matching has occurred. Matching EOS simulation with experiment, known as tuning or regression, is used in industry before application in compositional simulation. Obviously, this is only possible if PVT studies are available.
Some work has been done previously on EOS comparison. Recently, Bonyadi and Esmaeilzadeh3 compared the Esmaeilzadeh-Roshanfekr EOS against the popular Peng-Robinson EOS for gas-condensates. This was done in a purely predictive manner, i.e. no EOS tuning took place before simulation begun. It was reported that the ER equation was superior for saturation pressure predictions and had lower average deviation on PVT experiment simulation. Ahmed4 compared 8 EOS in a "semi-predictive" fashion; each EOS was matched to a known saturation pressure before the simulation of PVT experiments started. Again, gas-condensates were chosen for the study. It was reported that certain EOS were advantageous in different areas: compressibility factors were best represented by the PT and SW equations of state, phase compositions were best predicted by the PR equation and the volumetric behaviour of condensates was best predicted by the SW equation. Martin5 considered pure petroleum components and compared EOS. From the Australian School of Petroleum, Beinke6 compared different EOS in the compositional simulation of CO2 sequestration. There are many more examples in the literature7–9 where EOS are compared.
The findings all highlight the fact that there is currently no "all in one" EOS that will favourably predict all thermodynamic properties. It seems that some EOS can outperform others only for certain reservoir fluid types, and some EOS are advantageous for particular properties, e.g. the PR EOS for superior phase compositions. This may also lead to entire PVT experiments being better represented by particular EOS, as is suspected. These are all subtle points for the engineer to consider when modelling fluids under either limited data or laboratory PVT experiment appraisal. Often the peculiarity of phase behaviour (e.g. in the near-critical area, for oil-CO2 systems under asphaltene precipitation or for oil with significant non-hydrocarbon components3,15,36) can encourage the development of EOS which are hoped to increase this accuracy of predictive simulation.
The objective of this study is thus to compare the new and old EOS, using a larger and more comprehensive set of samples than those already present in the literature. There will be 7 case studies, comprising of six new samples from South Australian oil companies and one previously published oil sample. A spectrum of fluids is studied: Oil, volatile oil, gas condensates and gas. These will provide a small survey of Australian fluid types from currently producing fields. In some instances, practical applications of EOS are also considered in addition to the comparison aspect, according to which company provided the sample and their interests.