Optimal design of sieve trays in distillation columns

Abstract

The traditional approach to distillation column design is the heuristic procedure, based on the trial-and-error method. In recent decades, the literature presents several works on the optimization of distillation columns, mostly integrated with a commercial simulator. However, these works are only concerned with the sizing of the diameter and height of the column, which are part of the total cost of the column; the other important variables for the functionality of the column are not measured. In this context, this work presents the optimization of the optimal design of distillation column trays using two different techniques: mixed integer nonlinear programming and Set Trimming. In addition, it is proposed to refine the parameters of the flooding and entrainment correlations in the literature in order to improve the accuracy of the design formulation. The objective of this optimization is to size all the geometric variables of the tray while respecting its hydraulic constraints (flooding, entrainment, weeping, and downcomer backup). Two possible objective functions are tested, the column cost minimization and the column mass minimization. The proposed approach is applied to an example from the literature, the optimal result obtained by mathematical programming and Set Trimming are compared with the result obtained by the heuristic procedure (taken from the literature and described in the text), and the different techniques are compared with each other. Mathematical programming was used to estimate the parameters of the correlations. The new parameters are applied to three example column designs from the literature and compared with preexisting correlations. The results obtained showed a significant reduction in the mean and maximum error of the correlations compared to the original data.