Data-based dynamic compartment model: Modeling of E. coli fed-batch fermentation in a 600 m3 bubble column

Abstract Mathematical modelling is a powerful and inexpensive approach to provide a quantitative basis for improvements that minimize the negative effects of bioreactor heterogeneity. For a model to accurately represent a heterogeneous system, a flow model which describes how mass is channeled between different zones of the bioreactor volume is necessary. In this study, a previously developed compartment model approach based on data from flow-following sensor devices was further developed to account for dynamic changes in volume and flow rates, and thus enabling simulation of the widely used fed-batch process. The application of the dynamic compartment model was demonstrated in a study of an industrial fermentation process in a 600 m3 bubble column bioreactor. The flow model was used to evaluate the mixing performance by means of tracer simulations and was coupled with reaction kinetics to simulate concentration gradients in the process. The simulations showed that despite the presence of long mixing times and significant substrate gradients early in the process, improving the heterogeneity did not lead to overall improvements to the process. Improvements could however be achieved by modifying the dextrose feeding profile.