Phase equilibrium studies of high-pressure natural gas mixtures with toluene for LNG applications
Introduction
Liquefied Natural Gas (LNG) plants are known to be energy and cost-intensive, requiring a large amount of power for the processes of treatment, compression and refrigeration, and with special designed equipment such as the main cryogenic heat exchangers (MCHE), refrigerant compressors and cryogenic distillation towers [1]. Before natural gas is liquefied, impurities such as acid gases, water and heavy hydrocarbons must be removed [2]. Treated natural gas then enters a cryogenic distillation tower known in the industry as a ‘scrub column’ (SC) to remove the heavy hydrocarbons [1]. The lean gas stream then enters the main cryogenic heat exchanger (MCHE) where the liquefaction occurs.
An illustrative schematic for a typical scrub column schematic is shown in Fig. 1. The scrub column operates essentially at a constant pressure, the value of which depends upon the feed pressure and composition. Typically, the scrub column conditions are around 4–6 MPa with its feed gas pre-cooled and partially condensed at a temperature that depends on its composition [3].The treated gas is cooled in the MCHE to about −151 °C (122 K) producing a high pressure LNG, which then enters a turbine or a flash expansion valve to provide further cooling to about −160 °C (113 K) at (or near) atmospheric pressure. Several natural gas liquefaction technologies have been proposed for the production of LNG [[4], [5], [6], [7], [8], [9]] but each stage of that transformation process is designed using predictions of the mixture's thermophysical properties as a function of temperature, pressure and composition. A detailed and accurate simulation is thus essential for the design, operation and optimization of the LNG liquefaction and treatment processes, which is influenced substantially by the accuracy of the property model used in the simulation [10,11].
The scrub column's purpose [3] is to prevent significant concentrations of compounds heavier than propane (C3+) from entering the MCHE, so that (a) the LNG meets its heating value specification and (b) compounds heavier than pentane (C5+) do not freeze-out and block the narrow tubing networks within the MCHE which can leads to severe consequences including unplanned plant shutdowns. The hydrocarbons in natural gas that pose the greatest risk of forming solids are the so called BTEX (benzene, toluene, ethylbenzene and xylenes) compounds. Benzene, for example, has a normal melting temperature of 5.45 °C [12], and a solubility of only around 5 ppm in LNG at (−150 °C, 5 MPa) [13].
In the scrub column, separation of the heavy components from the natural gas destined for the MCHE is achieved by maintaining a temperature gradient along the SC vertical length and by ensuring intimate contact between the liquid and vapor phases either in trays and/or packing distributed along the columns length. The vapor is flashed sequentially at decreasing temperatures allowing the heavier hydrocarbons to be stripped out and condensed. Simulation of the scrub column requires the simultaneous solution of three sets of equations representing the energy balance, material balance and phase equilibrium at each theoretical stage. The solution to these equations for the entire column is achieved iteratively using, for example, the inside/out convergence algorithm discussed by Russell [14]. Central to any solution algorithm is the reliable and efficient calculation of the thermodynamic properties of each mixture on each tray: this includes the distribution at equilibrium of the components between the two phases (the Ki = yi/xi values), the phase enthalpies, and the phase volumes. The most important of these are the Ki values and the phase enthalpies.
However, the uncertainty associated with the thermodynamic model used by the simulator to calculate these quantities has a direct impact on the reliability of the simulation and therefore, the required operating margins employed. Several authors have examined the impact of thermodynamic uncertainties on the reliability of process simulations [16,17]. Kister [18] has reviewed several examples of problems associated with the simulation of distillation columns and identified the inability to predict Ki values as the leading cause of problematic simulations. Thus, the selection of the most appropriate thermodynamic model for use in the process simulation of the scrub column is critical. Determining the most appropriate model requires an assessment of the performance of possible equations of state (EOS) against data obtained in laboratory experiments and/or from operating LNG plants. The latter, however, are difficult to obtain and often have comparatively large uncertainties that make it difficult to distinguish between the EOS used in the process simulations [19].
Most recent efforts to improve LNG process simulations have focussed on the use of complex equations of state capable of more accurately describing the vapor-liquid equilibrium and heat capacities of multi-component fluid mixtures [10,11,15,20,21]. However, the industry still uses cubic equations of state (EOS) such as PR76 [22] developed by Peng and Robinson, RKS [23] developed by Redlich, Kwong and Soave. There are many types of EOS with varying complexity but all are anchored to measured data and the EOS reliability decreases as predictions go beyond the data's range. For example, whilst the reference GERG-2008 EOS by Kunz and Wagner [15] is recommended by NIST for natural gas mixtures, a highly complex multi-parameter equation requiring iterative solution, its VLE predictions are no more accurate than those of the cubic EOS used by process simulators because of their computational efficiency. Dauber and Span [10] examined the influence of different properties models on the simulation of the LNG liquefaction process and LNG transport. They have concluded that GERG-2008 EOS provides the highest potential for accurate calculations of the thermodynamic properties of natural gas such as density, heat capacity and enthalpy. However, in their work, the composition of the LNG considered was representative of the feed entering the MCHE and didn't include heavy hydrocarbons (C6+).
To determine which EOS is best for the description of heavy hydrocarbon carry-over in LNG scrub columns requires accurate experimental data. However, such data are scarce, particularly for the multi-component mixtures and high-pressure conditions of most relevance to industrial LNG columns. In 2015, May et al. [20] reported reference-quality p,T,x,y measurements describing binary mixture VLE of methane + ethane, + propane, + 2-methylpropane (isobutane), and + butane that clearly identified deficiencies in EOS commonly used by industry, and in several of the archival literature data to which those models have been tuned. We have also studied the VLE of methane + toluene binary mixtures at temperatures from (179–313) K [24] as well as methane + pentane, and methane + hexane binary mixtures at temperatures from (173–330) K [25]. However, while these results can be of used to tune equations of state, they do not provide stringent tests of the phase compositions found in an LNG plant where a range of intermediate compounds is also present. Accordingly, in this work we investigated the VLE of the methane + propane + toluene system with the objective of investigating conditions where toluene is a minor component in both the liquid and vapor phases. Measurements of the VLE of this ternary mixture were conducted at temperatures between 213 and 298 K and pressures up to 8.3 MPa.
The measured VLE data for this ternary system were compared to results calculated with the AspenTech HYSYS Peng Robinson (PR) equation of state (EOS) [26]. While both phases were measured and the all data obtained are reported here, the focus of the comparisons was on the ability of the EOS to correctly predict the toluene content of the vapor phase, as this relates directly to the risk of cryogenic solids formation in the MCHE and the operational performance of the scrub column.
Section snippets
Materials
The suppliers and supplier-analyzed purities of all components used in this work are listed in Table 1. No further purification was applied.
Experimental setup
Two apparatus were used for the VLE measurements, and the consistency of their results further increased the confidence in the results obtained. The first apparatus was described in detail previously [20,25,[27], [28], [29]] and referred to hereafter as the ‘Cryostat VLE cell’. Only a summary is given here for completeness. The apparatus (Fig. 2) comprised
Results and model comparison
Measurements were completed along isochoric pathways with each apparatus: isochores 1, 2 and 3 were measured with the Cryostat VLE cell at temperatures between 213 K and 273 K. Isochores A, B, C and D were measured with the Bath VLE cell at temperatures range between 257 K and 298 K. In Table 2, Table 3 a total of 38 new dew point and 38 new bubble point VLE data for (methane + propane + toluene) ternary system are presented. The focus of the Cryostat VLE cell in particular was to investigate
Conclusion
New VLE data for the ternary system [methane + propane + toluene] have been measured at conditions relevant to the operation of LNG scrub columns. The ternary system was studied over a wide range of conditions with toluene as the minor component in both the liquid and vapor phases. Measurements were conducted along different isochoric paths at temperatures between (213 and 298 K) and pressures up to 8.3 MPa. These data provide an opportunity to test the performance of the equations of state
CRediT authorship contribution statement
Saif ZS. Al Ghafri: Investigation, Writing - review & editing, Visualization. Thomas J. Hughes: Investigation, Writing - review & editing, Visualization. Fernando Perez: Investigation, Methodology. Corey J. Baker: Investigation, Methodology. Arman Siahvashi: Investigation, Methodology. Armand Karimi: Investigation. Arash Arami-Niya: Investigation. Eric F. May: Supervision, Writing - review & editing.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This work was funded by the ARC Training Centre for LNG Futures (Australian Research Council grant number IC150100019). We thank Stanley Huang and Jeff Buckles of Chevron for helpful discussions and Martin Khamphasith (UWA) for his assistance with the measurements.
References (39)
- et al.
Low temperature techniques for natural gas purification and LNG production: an energy and exergy analysis
Appl. Energy
(2016) - et al.
Utilization of the cryogenic exergy of LNG by a mirror gas-turbine
Appl. Energy
(2004) - et al.
Low temperature techniques for natural gas purification and LNG production: an energy and exergy analysis
Appl. Energy
(2016) A thermodynamic review of cryogenic refrigeration cycles for liquefaction of natural gas
Cryogenics
(2015)- et al.
Knowledge based decision making method for the selection of mixed refrigerant systems for energy efficient LNG processes
Appl. Energy
(2013) - et al.
Performance enhancement of propane pre-cooled mixed refrigerant LNG plant
Appl. Energy
(2012) - et al.
Modelling liquefied-natural-gas processes using highly accurate property models
Appl. Energy
(2012) - et al.
Achieving higher accuracies for process simulations by implementing the new reference equation for natural gases
Comput. Chem. Eng.
(2012) - et al.
Extending the GERG-2008 equation of state: improved departure function and interaction parameters for (methane + butane)
J. Chem. Thermodyn.
(2016) Equilibrium constants from a modified redlich-kwong equation of state
Chem. Eng. Sci.
(1972)
Phase equilibrium measurements of (methane + benzene) and (methane + methylbenzene) at temperatures from (188 to 348) K and pressures to 13 MPa
J. Chem. Thermodyn.
Simulating the capture of CO2 from natural gas: new data and improved models for methane + carbon dioxide + methanol
Int. J. Greenhouse Gas Control
Phase equilibria of (Methylbenzene + Carbon dioxide + Methane) at elevated pressure: experiment and modelling
J. Supercrit. Fluids
Advanced predictions of solidification in cryogenic natural gas and LNG processing
J. Chem. Thermodyn.
Fundamentals of Natural Gas Processing
Improvement of High Heating Value of Commercialised Liquefied Natural Gas of GL1Z Plant by Optimising the LPG Extraction
Current status and perspectives of Liquefied Natural Gas (LNG) plant design
Ind. Eng. Chem. Res.
Equation of state for benzene for temperatures from the melting line up to 725 K with pressures up to 500 MPa
High. Temp. - High. Press.
Rapid simulation of solid deposition in cryogenic heat exchangers to improve risk management in liquefied natural gas production
Energy Fuels
Cited by (6)
Natural gas density measurements and the impact of accuracy on process design
2021, FuelCitation Excerpt :Thus, in terms of ternary or multi-component gas mixtures, their accuracy needs to be further tested with new high-quality experimental data [12]. Our recent studies of thermodynamic properties such as phase equilibrium [14,20-32], heat capacity [33,34] and density [24,35–37] clearly identified deficiencies in EOS commonly used by industry, and in several of the archival literature data to which those models have been tuned [14]. Therefore, in this work, the density of gas and liquid phases at high pressures and over a wide-range of temperatures, including near the mixture’s critical point (where gas and liquid become indistinguishable and the existing predictive fluid property models used by engineers breakdown) were investigated.
Experimental study of impurity freeze-out in ternary methane + ethane + benzene mixtures with applications to LNG production
2021, Journal of Natural Gas Science and EngineeringCitation Excerpt :However, once these binary interactions are characterized, the predictive accuracy of models describing solid-fluid-equilibrium (SFE) conditions for LNG-like mixtures must be examined via experiments involving higher-order mixtures such as ternary, quaternary, and multi-component systems. It is well known that the presence of other compounds in the solvent such as ethane or propane (in addition to methane) can appreciably enhance the solubility of the heavier hydrocarbons (Kohn et al., 1977; Brew, 1977; Kohn and Luks, 1976; Al Ghafri et al., 2020; Sampson et al., 2020). In the 1970s attempts were made to investigate and measure the solubility of heavier hydrocarbons such as hexane, cyclohexane, heptane, octane, and benzene in the presence of methane-dominant mixtures containing ethane and propane.
Avoiding costly LNG plant freeze-out-induced shutdowns: Measurement and modelling for neopentane solubility at LNG conditions
2021, EnergyCitation Excerpt :The Kyowa pressure transducer uncertainty, u(p), was estimated by combining in quadrature the manufacturer’s stated uncertainty (0.2% of the reading), the r.m.s. deviation of the transducer’s calibration (2.8 kPa), and the standard deviation of three repeat measurements (typically around 20 kPa). As detailed by Al Ghafri et al. (2020) [22], the uncertainty associated with the in-situ synthetic mixture preparation using pure methane gas and pure neopentane liquid was calculated by combining in quadrature the uncertainties arising from the syringe pump (ISCO) volume injections, visual cell volume (69.9 ± 0.1 mL), the fluid density at the pump’s T and p conditions, dead volumes of the injection lines, the purity of the chemicals used in this study, and the mole fraction of neopentane in the gas mixture. The neopentane density was obtained from the equation of state of Lemmon and Span [23], with an estimated relative uncertainty of 0.2% for the conditions studied in this work.
The Removal of Benzene and Toluene in Natural Gas with Cryogenic Liquid Propane: Effects and a Cyclic Purification Process
2024, Korean Journal of Chemical EngineeringBarrier Discharge Ionization Detector in Gas Chromatography: A Review on Applications
2023, Critical Reviews in Analytical ChemistryHigh Pressure Thermal Conductivity Measurements of Ternary (Methane + Propane + Heptane) Mixtures with a Transient Hot-Wire Apparatus
2021, International Journal of Thermophysics