Robust and Quick Design and Synthesis Methods for Separation Process

  Libin Zhang and Andreas Linninger

Distillative separation belongs to the cheapest methods for separating mixtures that exhibit suitable volatility differences. Hence distillation columns, their optimal operation and heat integration constitute a major engineering activity in refineries and bulk commodity manufacturing firms. New business directions point towards future pharmaceutical manufacturing practices no longer limited to a single drug with many wastes, but to a PRODUCTS-ONLY process. The PRODUCTS-ONLY plant aims at delivering a principal high value added product alongside complimentary products each for direct sale to a market customer. PRODUCTS-ONLY manufacturing engages suitable by-product conditioning via advanced separations, mixing or further reactions steps as depicted in Fig. 1.  This flexible viewpoint required in future business scenarios poses a problem to traditional separation design methodologies based on repeated performance calculations with state-of-the-art simulation software. Novel more adaptable manufacturing requires fast and reliable design methodologies capable of determining feasibility or impossibility of desired separation targets. Ideally, fully automatic computer programs should perform optimal mapping of all process streams into the changing product portfolio.

The novel approach reduces the dimensionality of the design problem, eliminates all singularities encountered in the tray-by-tray approach, extends to any number of species and all customary vapor-liquid equilibrium solution models including constant relative volatility, ideal and non-ideal zeotropic and azeotropic mixtures and applies to both sharp and sloppy splits. A new minimum bubble point distance (MIDI) algorithm for assessing the feasibility of a desired distillation specification is presented. Our MIDI algorithm also extended successfully to compute the minimum and maximum reflux. We also found its potential application for the separation of quaternary mixtures and for optimal column sequencing. A computation experiment documented in Fig. 2 applied the feasibility test to 10,000 randomly chosen column design problems for a ternary mixture of constant relative volatility. All design problems converged within a few iterations; not a single convergence failure was reported. Fig. 2 also visualizes all feasible designs in a composition triangle and lists the number of infeasible designs.

Julka, V. and Doherty, M. F., Geometric Behavior and Minimum Flows for Nonideal Multicomponent Distillation, Chem. Eng. Sci., 1990, 45, 1801-1822.

Julka, V. and Doherty, M. F., Geometric Nonlinear Analysis of Multicomponent Nonideal Distillation: Simple Computer-Aided Design Procedure, Chem. Eng. Sci., 1993, 48, 1367-1391.