Cálculo de mapas de curvas residuais aplicando modelos de equilíbrio e correção por eficiência

Abstract

Distillation is the most applied unit operation in industrial plants due to its large application for ideal and non-ideal mixtures. Residue curve map is a tool used when one wants to test the feasibility of separation of a given mixture in a distillation column. This tool can be used as a preliminary estimate of probable products will be obtained after distillation of a mixture. Residue curve maps are usually calculated assuming equilibrium between liquid and vapor phases. However, it is known that the inclusion of mass transfer parameters can correct the assumption of equilibrium data presented closer to real distillation. An important property of residue curve maps is the presence of the distillation boundaries. A distillation boundary breaks the residue curve map in different regions of distillation, making it impossible to obtain any pure components of the mixture. In the 90 decade, few scientific studies have been published indicating the possibility of crossing distillation boundaries when non-equilibrium models were applied to calculate the residue curves. This verification contests the concept of distillation boundaries and should be analyzed in the further. The main purpose of this work was to develop software suitable for the construction of residue curve maps for ideal and non-ideal mixtures applying equilibrium models with and without correction for efficiency. Conventional activity models (NRTL, UNIQUAC and Wilson) were evaluated to describe the non-ideality liquid phase. Was considered the correlation of BARROS and WOLF (1997) for the calculation of efficiency in order to correct the assumption of the equilibrium phase. For this study, we evaluated the sensitivity of the residual curves when this correlation was employed, and also to examine the possibility of crossing the distillation boundaries. Using case study the ideal mixture of pentane-hexane-heptane and the mixture is non-ideal ethanol-water-acetone for validation of the developed software. The use of different activity models did not result in significantly different of residue curve map, showing that, for the case evaluated formed by acetone-chloroform-benzene three activity models behavior were quite similar and efficient displaying clearly the residue curve of the mixture evaluated. To study the sensitivity of residual curves as well as the possibility of crossing the distillation boundary five different cases were studied. The first case of study involved again the ideal mixture formed by pentane-hexane-heptane with relative deviation between the modeling of 3.83%. To a mixture of ethanol-water-acetone the relative deviation was approximately 3.67% were not observed the crossing of distillation boundary. For methanol-isopropanol-water were performed a comparison of data from this study with the literature pointing the model corrected by efficiency was closest to the real values of a distillation process. For the mixture formed by chloroform-acetone-benzene the relative deviation was 5.21% and was not found crossing the distillation boundary. Finally, we analyzed the mixture formed by acetone-chloroform-methanol where the relative deviation was 4.82%. Thus, it becomes apparent that the use of different models can create different regions of distillation due the distillation boundaries sensibility, so it is recommended to use non-equilibrium modeling or correlation for efficiency, because these models take into account parameters of mass transfer and energy generating more accurate data for the distillation process, particularly in regions near the distillation boundaries.