SYNTHESIS AND DESIGN OF HOMOGENEOUS AZEOTROPIC DISTILLATION SEQUENCES

An automated procedure to synthesize, design, and optimize homogeneous, azeotropic distillation sequences is presented. The procedure allows for the rapid screening of feasible separation trains and comparison of process alternatives. The synthesis step has been automated using techniques from differential topology and graph theory. The only physical property data needed to determine the physically permissible separation(s) for a given feed are the boiling point temperatures of all pure components and azeotropes. The procedure is well suited to screening candidate entrainers for separating binary azeotropic mixtures as well as determining viable sequences for many-component azeotropic mixtures. Once the distillation train has been synthesized, each column must be designed to meet the target product compositions. An algorithm for calculating the minimum reflux ratio for ideal, nonideal, and azeotropic distillation columns with significant heat effects is presented. The algorithm applies to single- and multiple-feed columns. An algorithm for determining the number of theoretical trays required for the column at operating conditions is also presented. The column sequences are optimized to ensure that comparisons among alternative entrainers and/or sequences are done on an equitable basis. The most significant optimization variable in the separation of binary azeotropic mixtures is the feed ratio, the dimensionless entrainer recycle flow rate. Sequences properly designed and optimized are found to be significantly cheaper than those previously reported in the literature