Análise global da estabilidade termodinâmica de misturas: um estudo com o método do conjunto gerador

• Autor(a) principal: Souza, Joviana Sartori de
• Data de Publicação: 2010
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações da UERJ
• Texto Completo: http://www.bdtd.uerj.br/handle/1/13641
• topic: Chemical equilibrium; Phase stability; Gibbs free energy; Mathematical optimization; Methaheuristic; Equilíbrio químico; Estabilidade de fases; Energia livre de Gibbs; Otimização matemática; Metaheurística; CNPQ::CIENCIAS EXATAS E DA TERRA::MATEMATICA::MATEMATICA APLICADA

The phase equilibrium calculation is an important problem in engineering processes, for example, in the separation by distillation, in extraction processes, and simulation of the tertiary petroleum recovery, among others applications. But, before this calculation is advisable a phase stability analysis, which will determine if a given mixture presents one or more phases. Such stability analysis can be approached as an optimization problem, called the minimization of the tangent plane distance function to molar Gibbs free energy, where thermodynamic models, with a non-convex and also non-linear nature, are used to describe this problem. Such fact has been motivating a large interest in robust and efficient techniques for optimization of problems related with the phase equilibrium thermodynamics. As has been highlighted in the literature, in order to provide a complete prediction of the phase equilibrium, the determination of the global minimizer of the objective function related with the stability test is not enough. Is also necessary the obtaining of all its stationary points. Thus, the development of methodologies for this challenging task became a new research area of global optimization applied to the phase equilibrium thermodynamics, with common interests in chemical and petroleum engineering. The focus of the present work is a new methodology to solve the stability test problem. For such, here it is used the called Generating Set Search method, which accomplishes local searches in points of a net previously generated by an evolutionary methaheuristic, the Particle Swarm algorithm. To obtain more than one stationary point, polarized merit functions are minimized, where the poles of such merit functions are the points previously found. The methodology proposal was tested in the analysis of fourteen polar mixtures, previously considered in the literature. The results showed that the proposed method is robust and efficient, being capable to find all stationary points previously indicated in the literature, besides the global minimizers. For two ternary mixtures, it was noticed that this method detected some stationary points not obtained before by the interval analysis method, a reliable technique very diffused in the literature. The stability analysis using only the Particle Swarm algorithm together with the polarization technique (without the local search done by the Generating Set Search method) constitutes other methodology for the resolution of the problem considered here. That use is a secondary novelty of this work. This simplified methodology exhibited also a significant robustness, being capable to find all the stationary points. However, when compared with the more general approach proposed here, in some cases, where the merit function presents a more complex geometry, it was observed that such simplification can consume a relatively large computational time. Therefore, the simplified methodology is less efficient.