(71c) Design of Reactive Distillation Column for Non-Parameterized Components - Which Experiments Are Essential

Design of reactive distillation column for
non-parameterized components - which experiments are essential

Litzmann, O., TU Berlin, Berlin; Lorenz, H.-M., TU Berlin, Berlin; Wozny, G., TU Berlin, Berlin; Grützner,
T., Lonza AG, Visp/CH; Repke, J.-U., TU Freiberg, Freiberg   

Designing reactive distillation processes
requires a great number of fitted parameters depending on the model used. These
parameters are known for test-systems like MTBE or TAME but not for systems
which have never before used in reactive distillation. As fitting of parameters
require a lot of time and money for trials it is crucial to know the impact of
an improved model to the prediction accuracy. If this is known it is possible
to carry out fewer experiments with and fit a set of sufficient parameters in
order to describe the RD process as good as needed to strive
the goal.

Application to a new industrial substance
system

The system analysed consists of an organic acid, an
alcohol, the corresponding ester and water whereupon some components cannot be
described by standard UNIFAC methods. The esterfication
takes place in a reactive distillation column designed for this system. The reaction
is catalysed by the heterogeneous catalyst Amberlyst 36.

In order to describe the results of the
column, models are used which take different phenomena into account and use
different sets of parameters which have been measured independently. 

For instance the reaction can be described either by
equilibrium or by kinetics ? which can be corrected by adsorption effects in addition.
This kinetic have been measured in a batch reactor and is described with significant
higher accuracy by an adsorption based kinetic model (see Fig.1) The
portability of these model improvement and similar effects to the reactive
column simulation will be presented and discussed by means of comparison with
the experimental results from the lab scale plant.