In this work, a methodology is established
to manage and use, in
a more rigorous way, the experimental information that reflects the
thermodynamic–mathematical behavior of dissolutions. The management
of experimental information is carried out with an application on
binaries of esters and alkanes which is useful in any other case.
Specifically, for this work a new real database (of several properties
under different conditions) is generated for eight binaries formed
by four alkanoates, with a carbon number number ≥ 4, and two
alkanes C6 and C8. A sequence of operations
is proposed, ranging from experimentation to simulation, with two
highly relevant intermediate stages, modeling↔verification
of the quality of data, whose impact on the simulation is evaluated.
The experimental contribution of some properties v
E, c
P
E, h
E, g
E, gives rise to two very important operations, such as
the combined modeling of the properties, taking into account the thermodynamic
formalism, and the verification of the vapor–liquid equilibrium
(VLE) data. For the latter process, the methodology designed in a
previous work (J. Chem.
Thermodyn.2017105385) is put into practice, as well as a new method,
rigorous under a thermodynamic–mathematical point of view,
in which the modeling of properties is considered. The binomial model-consistency test is generated as a strategic stage
to define the quality of the data. To achieve an accurate modeling
in the multifunctional correlation that is proposed, two procedures
are adopted: (a), step-by-step (SSO), according to the inverse order
of the derivation of the Gibbs function, and another (b), by multiobjective
optimization (MOO). The parametrization obtained by the latter is
implemented in the commercial software of Aspen-Plus to design a rectification
operation to purify the compounds of one of the studied systems, comparing
the results with those that the simulator emits with the information
estimated by UNIFAC.