Modeling the kinetic behaviour, mixing and local transfer pheonmena and biologicial population heterogeneity effects in industrial fermenters

Simulations are becoming an essential tool to design and improve processes in the field of biotechnologies. They are especially relevant to facilitate the scale-up of biological cultures from laboratory to industrial scales which is a key difficulty as of now. This thesis focuses on developping a model structure for fermenters, which does not neglect either mixing issues known to occur in industrial bioreactor, nor biological complexity inherent to micro-organisms, while enabling fast and low-cost simulations. To account for all coupled and dynamic phenomena that occur in bioreactors, the developed approach couples (i) a dynamic metabolic model to describe cells behaviour, (ii) a population balance model tracking biological cell-to-cell diversity and (iii) a compartment model to account for fermenter hydrodynamics. A structure for low-cost dynamic metabolic model has been developed, applied to E. coli and S. cerevisiae and successfully challenged against experimental data. Among multiple numerical methods tackling population balance equations, the EQMOM method has been selected for its stability and precision, and its algorithm has been improved by reducing its cost by a factor 10. The gas-liquid hydrodynamics of an industrial fermenter has been obtained through CFD simulations, and tools have been developed to extract compartment model from these simulation results. Finally, the coupling between all these modeling blocks has been demonstrated by simulating an actual industrial culture. This work paves the way to the emergence of fast bioreactor simulation tools, which will then enable new enginnering studies for designing and optimising industrial bio-processes