DYNAMIC CATALYTIC DISTILLATION - ADVANCED SIMULATION AND EXPERIMENTALVALIDATION

Reactive distillation offers a number of potential advantages, so that many traditional operations are currently being investigated in order to discover further applications of this technology. Increasingly, it is performed in columns with catalytic packings that combine the adva ntages of normal structured packings and heterogeneous catalysts. Anal ysis of reactive distillation is difficult due to strong physico-chemi cal interactions, and it is even more complicated for catalytic distil lation (reactive distillation in catalytic packings), where the knowle dge of column hydraulics, as e.g. hold-ups, pressure drop, and liquid distribution, is more important than in the case of traditional unit o perations. In this paper, a detailed rate-based approach for modeling and simulation of catalytic distillation is presented, including all m ajor aspects of the description of column hydraulics, mass and energy transfer, chemical reactions and thermodynamic non-idealities. The mod el equations have been implemented into the ABACUSS large scale modeli ng environment. Even in the frame of such sophisticated modeling, a va riety of experimental parameters have to be determined. It is shown ho w these parameters appear in the model equations and bow experiments a nd computer simulation interact. For a completely new class of catalyt ic packings all experimental parameters have been derived in accordanc e with the model assumptions in the form of correlations. All models h ave been formulated for dynamic operation and numerical implications o f the dynamic modeling ale addressed. For the semicontinuous catalytic distillation of the quaternary reactive mixture of acetic acid, metha nol, methyl acetate, and water, simulated and experimental results are presented and compared. Based on the integration of detailed modeling , experimental parameter determination and a modern simulation platfor m it is possible to predict the dynamic, non-linear and non-ideal proc ess behavior successfully. (C) 1998 Elsevier Science Ltd. All rights r eserved.