The precision-complexity binomial to achieve the best mathematical-thermodynamic modelling on the Gibbs-function and its effect on the separation processes design

Abstract

A modelling methodology based on the precision-complexity binomial has been applied to a parametric model established for the Gibbs function, in order to achieve the best parameterization. This methodology is integrated into a multi-objective optimization method to describe, as accurately as possible, the real behaviour of a set of binary solutions of butyl butanoate-alkanes, for which several of their properties have been experimentally determined. The optimization is addressed with a genetic algorithm with elitism-control that favours the convergence, providing well distributed results along the efficient-fronts. A database of, approximately 1500 values, was generated, which used to check the numerical procedure. The representation of different properties is acceptably described with the multiproperty model constituted by the gE-model and its successive derivatives. In relation to other models, the approach carried out significantly improves the results because, in addition, it includes an estimation of the volumetric effects. A practical application was carried out with the simulation of a separation process for one of the systems that present a complex behaviour, providing solutions for its implementation in an industrial scale.