Model-based methods as tools for targeted process design

Abstract

Bioprocess engineering is an interdisciplinary field of engineering sciences, which is at the interface between biology and classical engineering sciences, such as measurement and control engineering, automation engineering, plant construction and all classical fields of process engineering. Bioengenineering focuses on processes including a material conversion by means of a biological catalyst. The development, design and implementation of these processes are primary tasks of bioprocess engineering. In order to develop an adequate process for a de ned task, countless methods are available to the executing process engineer. Depending on prior knowledge of the engineer these methods can be applied in dierent combinations and therefore lead to completely dierent results. One way to ensure more structure and understandable decisions are model-based methods. The advantage of these methods is the use of models, which should be statistically verified before application, and can also serve as a knowledge storage system for complex processes. This enables a transparent and ecient process development and design. Within the framework of the thesis, three superordinate tasks were dealt with: i) The description of the state of the art of model-based methods and automated process analysis technology. ii) A case study on the use of model-based methods for measuring, modeling and monitoring segregated biomass in a mammalian cell process. iii) The development of an automatable work ow for the set-up of kinetic process models. The thesis shows, that: model-based methods are essential for the progressive automation in biochemical engineering. combining measurement and modeling leads to an extended range of validity of mechanistic models shown for cell culture processes by considering lysed cells. by the application of model-based monitoring strategies (soft sensor), control, reactor and process quantifcations can be enabled, shown in a CHO process by taking into account cell damage the availability of mechanistic process models can be increased by the application of a developed "modeling work ow" and thus model-based methods become more applicable.