National Technical University of Athens (NTUA)
Thermodynamic modelling of natural gas mixtures
Abstract
dc:descriptionThis work focuses on thermodynamic modeling of natural gas mixtures with Equations of State (EoS), focusing on the midstream and downstream parts of the gas industry. Natural gas is the cleanest amongst fossil fuels. Throughout its production and processing, specifications are set regarding its thermodynamic properties; accurate knowledge of phase behavior and other thermophysical properties of natural gas mixtures enables the compliance with said specifications, lays the foundation for process optimization of the existing infrastructure and aids in the design of new facilities. Apart from phase equilibria, a variety of thermodynamic properties are necessary for the aforementioned calculations, such as volumetric behavior and derivative thermodynamic properties, namely heat capacity, speed of sound and Joule-Thomson coefficient which can be calculated by employing an accurate, robust, and reliable Equation of State (EoS).In this work, several EoS, namely Soave-Redlich-Kwong (SRK), Peng-Robinson (PR) and Perturbed-Chain-Statistical-Associating-Fluid-Theory (PC-SAFT), have been assessed for their capabilities of predicting thermodynamic properties of natural gas mixtures. Extensive comparisons with literature experimental data have been performed, pointing out the superiority of the UMR mixing rules over the one fluid van der Waals (vdW) ones when applied to cubic EoS in the form of UMR-PRU model, mostly in phase-equilibrium. Liquid dropout predictions reveal that UMR-PRU, PR and PC-SAFT are equally accurate and yield better predictions than SRK EoS, with higher deviations appearing close to the upper dew point in the retrograde region, and especially at near critical temperatures. When it comes to derivative thermodynamic properties, cubic EoS are less accurate as compared to aforementioned properties, and they sometimes fail to capture the experimental behavior, especially near the critical point. A new model is proposed for the above calculations. This model is based on PR EoS, since SRK is more susceptible to cancelation of errors in the prediction of derivative properties, and in most properties less accurate than PR. PR is coupled with a modified Mathias-Copeman α function, satisfying consistency constrains to ensure its safe extrapolation to supercritical temperatures, and a Peneloux volume translation. However, it still fails to capture some of the extrema of derivative properties, especially near the critical point. This EoS is combined with original UNIFAC though the universal mixing rules (UMR) forming the so-called UMR-MCPRU model, which yields very good results in all types of multicomponent VLE of natural gas and gas condensate mixtures, with a varying degree of asymmetry. The new model is at least as accurate or superior to the UMR-PRU model and also yields acceptable deviations in the predictions of derivative properties and critical points of multicomponent hydrocarbon mixtures. Use of an EoS to model real petroleum fluids is more challenging than synthetic mixtures, since their composition is rarely fully known. The most common case for reservoir fluids is to perform detailed analysis for the lighter components, while the heavy end of the fluid is usually described by a single plus fraction (C7+), which has a great effect on phase equilibria and other properties and needs to be characterized in order to enable any type of thermodynamic modelling. This procedure is commonly referred to as characterization method of the unidentified plus fraction and comprises three steps: splitting the plus fraction, assigning physical properties to each sub-fraction, and finally, regrouping them into pseudocomponents (lumping) with known properties and composition. Several splitting schemes and methods for the prediction of physical properties were evaluated and tested against suitable experimental data. The most adequate characterization method for UMR-MCPRU was proved to be method of Pedersen et al., using the critical properties predicted by the method of Twu, and the structure of the normal alkane with the closest molecular weight. This way the UMR-MCPRU model was applied successfully to predict gas condensate dew point and liquid dropout. Last but not least, the new model was used to simulate a typical platform operation encountered in midstream natural gas processing that treats a gas condensate well stream for the production of rich natural gas and a stabilized condensate stream. This was performed by incorporation of the model into the Honeywell’s UniSim® process simulator, through an application compliant to the CapeOpen standards.
Degree
thesis:*- Grantor dc:publisher
- National Technical University of Athens (NTUA)
- Year dc:date
- 2018
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Novak, Nefeli-Effrosyni
- Νόβακ, Νεφέλη-Ευφροσύνη
Subjects
dc:subject × 6Rights
dc:rights- Statement dc:rights
-
- BY
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/44920