What Is the Optimal Activity Coefficient Model To Be Combined with the translated–consistent Peng–Robinson Equation of State through Advanced Mixing Rules? Cross-Comparison and Grading of the Wilson, UNIQUAC, and NRTL aᴱ Models against a Benchmark Database Involving 200 Binary Systems

Piña-Martinez, Andrés; Privat, Romain; Nikolaidis, Ilias K.; Economou, Ioannis G.; Jaubert, Jean-Noël

What Is the Optimal Activity Coefficient Model To Be Combined with the translated–consistent Peng–Robinson Equation of State through Advanced Mixing Rules? Cross-Comparison and Grading of the Wilson, UNIQUAC, and NRTL aᴱ Models against a Benchmark Database Involving 200 Binary Systems

2021

Piña-Martinez, Andrés | Privat, Romain | Nikolaidis, Ilias K. | Economou, Ioannis G. | Jaubert, Jean-Noël

The extension of the translated–consistent Peng–Robinson (tc-PR) equation of state (EoS) to mixtures has been investigated. For this purpose, advanced EoS/aᵣₑₛᴱ,ᵞ mixing rules are used to combine the tc-PR EoS with the residual part of an activity coefficient model chosen among the Wilson, nonrandom two-liquid (NRTL), and universal quasi-chemical (UNIQUAC) models. The performances of the three resulting EoS versions are compared against a high-quality reference database containing binary system data, specifically selected for the cross-comparison of thermodynamic models and the assessment of their accuracy. It is shown that the best choice to extend the tc-PR EoS to mixtures along with EoS/aᵣₑₛᴱ,ᵞ mixing rules is to use the residual part of the Wilson activity coefficient model (aᵣₑₛᴱ,ᵂⁱˡˢᵒⁿ). After grading, the resulting model, named “tc-PR-Wilson”, received a mark of 12.4/20. It has a clear advantage over the other investigated models in this study, which all have marks below 12/20. Observing that some experimental data could not be compared to a calculated value (typically, when the model does not predict the right topology of a phase diagram such as, e.g., predicting a heterogeneous azeotrope instead of a homogeneous one), a new performance indicator (“success ratio”) was introduced as the ratio of the number of data points for which the model can reproduce qualitatively the experimental data behavior over the total number of experimental data. The success ratio of the tc-PR-Wilson model reaches 0.96; this model outperforms especially in the correlation of liquid- and gas-phase compositions, azeotropic points, and three-phase pressures.

显示更多 [+]