378467 Advances in Synthesis of Separation Systems for Azeotropic Mixtures

Wednesday, November 19, 2014: 1:30 PM
311 (Hilton Atlanta)
Stanislaw K. Wasylkiewicz1, Marco A. Satyro2, Yau Kun Li1 and Monika J. Wasylkiewicz3, (1)Virtual Materials Group Inc., Calgary, AB, Canada, (2)Virtual Materials Group, Inc., Calgary, AB, Canada, (3)Department of Chemical and Materials Engineering, University of Alberta, Edmonton, AB, Canada

The feasibility of distillation splits is the key for the sound design of non-ideal separation system for a novel chemical. So far the most popular approach is to use commercial process simulators to solve these design problems. The design-by-simulation approach requires repeated simulations for various design parameters usually selected by trial and error. Since not all desired specifications can be met, this approach can be frustrating and time-consuming.

Fortunately, in the last few decades, new conceptual design methods for distillation has been developed that allows designers to discover feasible separation schemes quicker, even for highly non-ideal azeotropic mixtures. These methods are based on the examination of the overall composition space, azeotropes, liquid-liquid immiscibility regions and distillation boundaries. Residue curves or column profile maps are useful tools that help to find out, through direct inspection, what separation can be achieved in a distillation column for ternary systems. The boundary value design method is at the core of the conceptual design methods, is intuitive and allows designer to see the interaction between design parameters. Several attempts have been made to extend the method to quaternary or more components’ mixtures for example through the use of nonlinear programming methods.

In our approach to the synthesis of separation systems, we start with a rigorous thermodynamic model of the mixture, determined from quality experimental data. Then all azeotropes are calculated using the homotopy continuation method for mixtures with any number of components. Based on this information, the problem of determination of what feasible products can be produced in individual distillation towers for a given feed composition, is solved by a rigorous feasibility test and split generation method.

The method requires the knowledge of all distillation regions and distillation boundaries in the mixture. While for ternary systems this information is easy available, for four or more components the calculation of adjacency and reachability matrices is necessary. To calculate adjacency matrices correctly all primary connections between adjacent stationary points must be found. They can be approximated by residue curves starting from saddle points in eigendirections that usually end at stable or unstable nodes.

Saddle – node connections are easy to find by forward or backward integration of the residue curve equations. Saddle – saddle connections are difficult to find and a special calculation procedure is necessary. In this presentation we show details of this procedure. Based on these results we can accurately generate distillation regions for continuous and batch distillation and identify respective boundaries. This information allows for a rapid generation of feasible distillation separation schemes, reliable screening for the preferred distillation regions for desired products and initial design of separation systems. A few examples of synthesis of separation systems will be presented for complex homogeneous and heterogeneous azeotropic mixtures.


Extended Abstract: File Uploaded
See more of this Session: Advances in Distillation
See more of this Group/Topical: Separations Division