Optimal Synthesis of Pressure Swing Distillation Columns for the Separation of Azeotropic Mixtures

In this work, we present a methodology for the global optimization of Pressure Swing Distillation Columns to separate and purify azeotropic mixtures through the Infinite DimensionAl State-space (IDEAS) approach. The IDEAS conceptual framework gives rise to an Infinite linear program (ILP). The solution of a series of finite dimensional linear programs is the values that converge to the infinite program's infimum. The proposed optimal design methodology is demonstrated in a case study involving the separation of the binary azeotrope formed by Methyl-Acetate and Methanol. A detailed flash distillation model that accounts for vapor/liquid phase equilibrium (VLE) is used in the IDEAS context. It employs a VLE thermodynamic model based on modified Raoult's law, with Wilson equation to model the non-ideality of the liquid phase. Despite the complex and the nonlinear nature of the underlying thermodynamic model the IDEAS approach yields nevertheless a linear program formulation of the global minimum capital cost. Capital cost is considered proportional to a weighted linear combination of the total flow in the network. An optimal IDEAS generated design is finally presented and compared to an optimized standard double-column pressure swing system for the separation of the mixture.