(97b) Synthesis and Quick Identification of Low-Energy Sequences for Multicomponent Distillation

Shah, V. H., Purdue University
Agrawal, R., Purdue University

Industrially, distillation is usually the preferred method for separation of multicomponent mixtures. To separate an N-component feed into essentially N pure product streams (N > 3), we can choose from several distillation configurations. These configurations carry out the same overall process; however they differ significantly in total cost. Choosing the optimal distillation sequence is thus an important and challenging problem.

The first step for solving this challenge is to generate a useful search space of distillation configurations. It has been shown that distillation configurations that use more than (N-1) columns for an N-component separation should be omitted from the search space because they are never optimal. We have thus developed a computationally efficient method to generate a search space of exclusively the (N-1) column configurations. We have also extended this method to include configurations with thermal coupling in the search space.

We then focus on minimizing energy consumption since operating cost usually dominates capital cost for large industrial applications. However, since the size of the search space grows exponentially as number of components in the feed increases, we need a quick screening method to identify attractive low-energy sequences because we cannot evaluate each configuration rigorously.

Therefore, the goal of this work is to describe a procedure to identify low-energy sequences by a short-cut screening of the search space. Our formulation enables us to automatically apply Underwood's equations for all the configurations, instead of formulating different problems for each configuration. In this way, we can quickly identify optimal distillation configurations. We can then evaluate the configurations obtained by the quick-screening method more rigorously. The power of our method is reflected in the fact that when applied to mature commercial applications such as crude distillation and ethylene separation, we have been able to identify a number of distillation configurations with potential energy savings in excess of 10%.