(470b) Effective Scale-up of Aerated Fermentation Processes with Complex Rheology to Industrial Scale | AIChE

(470b) Effective Scale-up of Aerated Fermentation Processes with Complex Rheology to Industrial Scale

Authors 

Witz, C., Graz University of Technology
Eibl, P., Graz University of Technology
Khinast, J. G., Graz University of Technology
Schöpf, M., Novartis
Hardiman, T., CureVac AG
In the biopharmaceutical industry, process transfers from pilot scale into production scale are often carried out with large scaling steps, particularly when fast and efficient development cycles are desired and when targeting world-scale plants. An accurate scaling concept is therefore of utmost interest, as inaccuracies may lead to large changes in the process conditions.

One prominent example of a widely used, yet inadequate, scaling concept for aerated bioreactors is based on a constant specific power input and aeration rate. For large scaling factors, this concept leads to an undesired change of the impeller flow regime, e. g., flooding of the bottom impeller. The present contribution introduces a novel scaling approach derived from the full dimensional analysis and mechanistic equations. The concept is based on the adequate scaling of the impeller flow regime and enables scaling the rheology of non-Newtonian fluids through extending the concept with one additional dimensionless number.

A large experimental data set supports deriving reliable correlation equations for the Newton number, gas hold-up and volumetric oxygen mass transfer coefficient (kLa). The data covers a large range of process conditions typical for industrial-scale fermentations, which were scaled down to a Plexiglas 160-L pilot scale reactor equipped with four Rushton impellers. Multiphase GPU-based CFD simulations, as well as real measurements of the flow regime, gas hold-up, and power input at the individual impeller stages extend the new insights into the large-scale liquid and gas bubble flow. The study further verifies the concept of applying a separate data set from the pilot reactor and from two industrial reactors at the 110-m3 and 170-m3 scale. The resulting prediction quality is unparalleled, as illustrated through the critical analysis of existing correlations. Thus, the new concept provides adequate correlations for scaling the flow regime and rheology between pilot and industrial scale reactors, verified for a scaling factor of three orders of magnitude.