Achim Hoffmann

Scale-up of reactive distillation columns with catalytic packings

The scale-up of reactive distillation columns with catalytic packings requires the knowledge of reaction kinetics, phase equilibrium and packing characteristics. Therefore, pressure drop, liquid holdup and separation efficiency have been determined for the catalytic packing MULTIPAK®. A new hydrodynamic model that describes the countercurrent gas-liquid flow for the whole loading range and considers the influence of the column diameter has been implemented into a rate-based column model. Simulation results for the methyl acetate synthesis are compared with pilot plant experiments that cover a wide range of different process conditions. The experiments are in good agreement with the simulation results and confirm the applicability of the modelling approach for reactive distillation processes with catalytic packings.

Reactive distillation: Non-ideal flow behaviour of the liquid phase in structured catalytic packings

Reactive distillation, the combination of chemical reaction and multistage distillation, is one of the most important industrial applications of the multifunctional reactor concept. The most promising column internals for reactive distillation are the so-called structured catalytic packings that combine favourable characteristics of traditional structured packings and heterogeneous catalysts. The non-ideal flow behaviour of the gas and the liquid phase is a fundamental aspect in multiphase reactor design since it has a strong influence on the reactor performance. In this study, liquid phase residence time distributions for the catalytic structured packing MULTIPAK® were measured by means of conductivity measurements under different liquid and gas flow rates and evaluated with differential models.

Computer-aided process design of affinity membrane adsorbers: a case study on antibodies capturing

Design of affinity membrane adsorbers for the purification of biomolecules requires a consideration of loading, washing, and elution. Modelling and simulation of membrane adsorbers in literature is, however, strongly focused on the loading step. Therefore, in this work, a complete process model which takes all the different steps into account was developed. Breakthrough experiments in which human IgG was captured onto and eluted from Sartobind Protein A downscale modules were used for model validation and for estimation of the required model parameters. The experimentally observed breakthrough curves were independent of the applied flow rate and from these results linear correlations between lumped kinetic parameters and linear velocity were determined. During elution, desorption was best described by an irreversible reaction of first order in H+ concentration. Applicability of the developed model to computer-aided design was illustrated through a process analysis study in which the influence of the amount of loaded protein per cycle on the process yield and productivity was investigated.

A knowledge-based system using reasoning with uncertainty in the VLSI chip synthesis domain

Abstract This paper presents a knowledge-based system which carries out a design step in VLSI chip synthesis—the VLSI chip-architecture selection. The system decides which VLSI chip-architecture scheme—a general architectural principle representing a class of chip-architectures—best fits a list of architectural properties given as input to the system. The system evaluates certain and complete as well as uncertain and/or incomplete input information. The paper shows how explicit reasoning with uncertainty, considering different points of view and both positive and negative evidence, supports the modelling of the selection process. The system uses a reasoning mechanism based on fuzzy logic.