Structured Kinetic Modeling for Rational Scale-down and Design Optimization of Industrial Fermentations

Abstract

Bioprocesses exploit the versatility of microorganisms to produce bio-products from renewable feedstocks. However, industrial-scale implementation often suffers from the “scale-up effect,” manifesting as reduced yield or productivity due to environmental heterogeneity in large bioreactors. This thesis presents a model-based approach to quantitatively reproduce these heterogeneous conditions and predict their impact on microbial performance.

For Penicillium chrysogenum, a 9-pool metabolically structured kinetic model was developed and integrated with an Euler-Lagrange computational fluid dynamics (CFD) simulation of a 54 m³ industrial bioreactor. This coupled model captured glucose gradients, penicillin productivity loss, and enabled the design of scale-down systems to mitigate mixing limitations, achieving predicted reductions of productivity loss by up to 50%.

For Saccharomyces cerevisiae, a 7-pool kinetic model was extended to include glucose uptake mechanisms, storage carbohydrate dynamics, and ethanol/glycerol re-consumption. The model reproduced Crabtree and Pasteur effects and demonstrated stability under highly dynamic pilot-scale conditions. The updated model provides a compact yet predictive framework for full-scale CFD integration.

Finally, this work outlines the foundation for implementing the digital twin concept in bioprocessing, emphasizing model simplification, fitness-for-purpose, and integration with real-time simulation for smart biomanufacturing. The results demonstrate the potential of combined kinetic-CFD models to optimize industrial fermentations, predict scale-up effects, and guide future bioprocess development.