Phase behavior of multicomponent alkane mixtures and evaluation of predictive capacity for the PR and RKPR EoS's

New Correlations for Prediction of High-Pressure Phase Equilibria of n-Alkane Mixtures with the RKPR EoS: Back from the Use of lij (Repulsive) Interaction Parameters

“…1 In previous studies 1,4 the superiority of the generalized Redlich−Kwong−Peng−Robinson Equation of State (RKPR EoS) 5 was demonstrated in comparison to the classic PR EoS 3 in the prediction of the vapor−liquid equilibrium (VLE) of the more asymmetric binary n-alkane−n-alkane mixtures and then also for multicomponent mixtures. 6 The present work was conceived as a continuation of those preceding studies and, at the same time, a turn back from the parameter correlations to which they arrived. The reason for such a turn back is the following.…”

“…For simplicity, we will term the obtained predictive models for n -alkane mixtures as RKPR2015 and PR2015. RKPR2015 predictions were excellent for fluid phase equilibria and clearly superior to those of the PR2015 EoS in the more asymmetric cases, not only for binary systems but also for multicomponent fluids . Nevertheless, we then discovered serious limitations in the prediction of mixture liquid densities or volumes and also in the modeling of solid–liquid saturation curves.…”

“…RKPR2015 predictions were excellent for fluid phase equilibria and clearly superior to those of the PR2015 EoS in the more asymmetric cases, not only for binary systems 1 but also for multicomponent fluids. 6 Nevertheless, we then discovered serious limitations in the prediction of mixture liquid densities or volumes and also in the modeling of solid−liquid saturation curves. We concluded that both problems were related, and the reason depended on the predominant role of the lij interaction parameter in the RKPR2015 correlations.…”

“…It is worth noting that even though the correlations of eqs 15 and 16 are here defined by considering the carbon numbers of the n-alkanes, an extrapolation of these correlations to other alkanes (i.e., branched or cyclic alkanes) could be proposed by means of a "scale transformation" as proposed by Tassin et al 6 In the work of by Tassin et al, 6 we have implemented a simple approach to extend the correlations defined for normal alkanes (in terms of carbon numbers) to branched alkanes. In such an approach, an equivalent carbon number (CN EQ ) equal to ω/ 0.05 (i.e., the acentric factor of the branched alkane divided by 0.05) is calculated and then used for the correlations previously defined for normal alkanes.…”

“…Dots: Experimental data from ref . Calculations with the PR EoS (dashed lines) and RKPR2015 EoS (solid line) correspond to quadratic mixing rules with parameters from ref and volume translations according to ref .…”