Estimation of k L a Values in Bench‐Scale Stirred Tank Reactors with Self‐Inducing Impeller by Multiphase CFD Simulations

Validation of computational fluid dynamics of shake flask experiments at moderate viscosity by liquid distributions and volumetric power inputs

Computational fluid dynamics (CFD) has recently become a pivotal tool in the design and scale-up of bioprocesses. While CFD has been extensively utilized for stirred tank reactors (STRs), there exists a relatively limited body of literature focusing on CFD applications for shake flasks, almost exclusively concentrated on fluids at waterlike viscosity. The importance of CFD model validation cannot be overstated. While techniques to elucidate the internal flow field are necessary for model validation in STRs, the liquid distribution, caused by the orbital shaking motion of shake flasks, can be exploited for model validation. An OpenFOAM CFD model for shake flasks has been established. Calculated liquid distributions were compared to suitable, previously published experimental data. Across a broad range of shaking conditions, at waterlike and moderate viscosity (16.7 mPa∙s), the CFD model's liquid distributions align excellently with the experimental data, in terms of overall shape and position of the liquid relative to the direction of the centrifugal force. Additionally, the CFD model was used to calculate the volumetric power input, based on the energy dissipation. Depending on the shaking conditions, the computed volumetric power inputs range from 0.1 to 7 kW/m3 and differed on average by 0.01 kW/m3 from measured literature data.

Influence of Interfacial Force Models and Population Balance Models on the kLa Value in Stirred Bioreactors

Optimal oxygen supply is vitally important for the cultivation of aerobically growing cells, as it has a direct influence on cell growth and product formation. A process engineering parameter directly related to oxygen supply is the volumetric oxygen mass transfer coefficient kLa. It is the influences on kLa and computing time of different interfacial force and population balance models in stirred bioreactors that have been evaluated in this study. For this investigation, the OpenFOAM 7 open-source toolbox was utilized. Firstly, the Euler–Euler model with a constant bubble diameter was applied to a 2L scale bioreactor to statistically examine the influence of different interfacial models on the kLa value. It was shown that the kL model and the constant bubble diameter have the greatest influence on the calculated kLa value. To eliminate the problem of a constant bubble diameter and to take effects such as bubble breakup and coalescence into account, the Euler–Euler model was coupled with population balance models (PBM). For this purpose, four coalescence and five bubble breakup models were examined. Ultimately, it was established that, for all of the models tested, coupling computational fluid dynamics (CFD) with PBM resulted in better agreement with the experimental data than using the Euler–Euler model. However, it should be noted that the higher accuracy of the PBM coupled models requires twice the computation time.