Robust Dynamic Simulation of Three-Phase Reactive Batch Distillation Columns

Abstract

The dynamic simulation of heterogeneous distillation columns is a demanding task because the number of liquid phases can vary over time on each tray. The success of the simulation depends on the rapid, reliable, and robust assessment of the number of phases and the correct initialization of the compositions of the liquid phases. In this work, a flexible simulation strategy is presented that decouples the phase split and flash algorithms from the process model. Therefore, no discontinuous model switching occurs in the process model, and the column can easily be simulated using standard tools employing a simultaneous equation-oriented simulation strategy. The phase state is assessed with a rapid and reliable homotopy−continuation algorithm that is well-suited for dynamic simulation because it decomposes the phase stability problem into two steps to minimize the necessary computational effort. The simulation strategy is applied to the production of butyl acetate from butanol and acetic acid in a three-phase reactive batch distillation column as an illustrative example.