Procedure for the correlation of normal appearance VLE data, where the classical models dramatically fail with no apparent reason

Abstract

The necessity of using a pressure (P) or temperature (T) dependence in the Gibbs energy of mixing of the liquid phase (g M,L ) for some isothermal (or isobaric) VLE data sets, respectively, is discussed in this paper. A graphical representation, directly obtained from the experimental data, is proposed as a method to clearly classify the behaviour of the systems and select the adequate model (i.e. including when appropriate the T or P dependence) to correlate them. The result is that many of the poorly fitted systems in the literature could be precisely modelled using the correct procedure with adequate functions of P or T. Some suitable examples are used to illustrate the validity of these ideas, providing satisfactory correlation results for those systems in the literature. The ideas presented in this paper reveal important aspects related to the inappropriate application of some thermodynamic consistency (TC) tests to evaluate the quality of VLE data sets.

What does the NRTL model look like? Determination of boundaries for different fluid phase equilibrium regions

“…On the one hand, that a model presents great flexibility in general, does not mean that it cannot additionally present certain specific limitations. [26][27][28] On the other hand, it is always necessary to double-check if the solution obtained in the experimental data correlation process is totally consistent with the experimental behavior of the system under study, analyzing the entire range of compositions. [23][24][25]29,35 In this sense, Figure 2 shows two examples of Gibbs mixing energy surfaces calculated from recently published NRTL binary interaction parameters obtained by correlation of ternary LL experimental data, which, however, are inconsistent with the experimental behavior of the studied system when the whole range of compositions is analyzed (not only the tie lines used in the correlation process).…”

“…However, despite the flexibility shown to reproduce complex behaviors, as mentioned earlier in the introduction section, certain precautions must be taken when using the NRTL model, as with any other model, to correlate experimental equilibrium data. On the one hand, that a model presents great flexibility in general, does not mean that it cannot additionally present certain specific limitations 26–28 . On the other hand, it is always necessary to double‐check if the solution obtained in the experimental data correlation process is totally consistent with the experimental behavior of the system under study, analyzing the entire range of compositions 23–25,29,35 .…”

“…To avoid all these difficulties when correlating experimental data from different kinds of phase equilibria, including LLE, VLE, LLLE, LSE, LLSE, etc., different approaches have been published over the years. For instance, analyzing the own limitations of the models such as the existence of gaps in the NRTL model where solutions for binary homogeneous liquid behavior do not exist 26 or considering higher temperature or pressure dependence in the binary interactio n parameters when it is required [27][28][29] . It is also possible to find different mathematical strategies [30][31][32][33][34][35] and algorithms based on the Gibbs energy of mixing (G M ) [36][37][38][39][40][41][42] , for improving the convergence and robustness of phase equilibria calculations in complex systems.…”

“…For instance, analyzing the own limitations of the models such as the existence of gaps in the nonrandom twoliquids mode (NRTL) model where solutions for binary homogeneous liquid behavior do not exist 26 or considering higher temperature or pressure dependence in the binary interaction parameters when it is required. [27][28][29] It is also possible to find different mathematical strategies [30][31][32][33][34][35] and algorithms based on the Gibbs energy of mixing (G M ), [36][37][38][39][40][41][42] for improving the convergence and robustness of phase equilibria calculations in complex systems. These alternatives normally include topological strategies for the selection of more adequate supposed phase equilibrium compositions (to start the correlation process) or for the confirmation of the consistency of the correlation results, for example, in the case of LLE for ternary systems with three partially miscible binary subsystems (type 3) including a three-liquidphase region.…”

Analysis of Vapor–Liquid Equilibrium in Ternary Systems for an Adequate Planning of Their Experimental Determination and Correlation

“…In previous works, the lack of flexibility to correlate certain types of phase equilibrium behaviors, including both LLE and VLE, has been discussed, − and also the lack of reliability of certain published parameters that may lead to inexistant equilibrium regions, false solutions, and other problems of a different nature. , Since commercial databases of model parameters used by the simulation software include many of these problematic or inconsistent parameters, severe problems may arise when working with the corresponding systems. Consequently, it would be very convenient to use additional tools to validate the equilibrium data, and the model parameters obtained, prior to entering them into any database.…”