Simulating large-scale biotech bubble columns: challenges and insights from the parcel approach

Area of Science:

• Biochemical Engineering
• Computational Fluid Dynamics
• Process Modeling

Background:

• Industrial demand for large-scale bioprocesses necessitates efficient modeling.
• Bubble-column bioreactors are crucial for gas fermentation, requiring accurate design.
• Computational Fluid Dynamics (CFD) is increasingly used for bioreactor design, with Finite-Volume Reynolds-Averaged Navier-Stokes (FV-RANS) being state-of-the-art.

Purpose of the Study:

• To investigate the consequences of the 'parcel size' approach in Euler-Lagrange (E-L) simulations for large-scale bioreactors.
• To compare Lattice Boltzmann Large Eddy Simulations (LB-LES) with FV-RANS for bioreactor modeling.
• To assess the impact of bubble behavior and distribution on key design parameters like the volumetric mass transfer coefficient (kLa).

Main Methods:

• Application of Lattice Boltzmann Large Eddy Simulations (LB-LES) to model gas fermentation in bubble-column bioreactors.
• Utilizing the 'parcel size' approach within Euler-Lagrange (E-L) simulations to represent multiple bubbles computationally.
• Comparison of LB-LES results with state-of-the-art Finite-Volume Reynolds-Averaged Navier-Stokes (FV-RANS) simulations.

Main Results:

• LB-LES revealed swarm-like bubble behavior leading to inhomogeneous bubble distributions.
• The 'parcel size' approach in E-L simulations can bias the volumetric mass transfer coefficient (kLa).
• Similarities in transient flow structures and averaged velocities were observed between LB-LES and FV-RANS, but local deviations exist.

Conclusions:

• Parcel-based modeling in E-L simulations has potential but also limitations for large-scale bioreactor design.
• Differences in turbulence representation and the 'parcel approach' impact local simulation results.
• Accurate prediction of bubble dynamics is crucial for reliable bioreactor design and optimization.