-
Mixing & Cell Damage Optimisation Workflow
An optimization workflow was developed to find the optimal, or Pareto, trade-off between mixing effectiveness and cell damage in a conventional stirred tank bioreactor, through variations in Ruston turbine blade geometry and rotation speed. The multi-objective optimization was driven by BOTorch, the CFD simulations using OpenFOAM®, and parameterised blade deformation with PyGeM.
-
CFD Model Validation
The bioreactor geometry was constructed as used in a published work so as to compare to existing validation data. Radial and axial velocities and turbulence properties are compared to measured data over four sample lines for three different mesh resolutions, showing good agreement.
-
Optimisation Setup
The input parameters were the impeller rotation rate and the deformation of the outer corners of the Ruston paddles, in directions normal to the undeformed paddle. The performance metrics were the net flow through a plane above the impeller, to characterise mixing, and the volume average of the dissipation rate of turbulence kinetic energy to characteristics cell damage.
-
Optimisation Workflow
The fully automated workflow was built in Python and executed on Amazon Web Services, using high core count compute instances to run multiple simulations at the same time.
-
Optimal Mixing and Cell Damage Trade Off
The resulting map of mixing and cell damage performance shows the Pareto front - the optimal trade-off between the two metrics - allowing for bioreactor configuration to provide the best mixing performance for a limiting cell damage value, or the minimum cell damage for a specific mixing performance.
Get in touch:
ObscureMail
ObscureMobile
Get in touch:
ObscureMail
ObscureMobile
© 2022 Upstream Applied Science Ltd. Registered in England and Wales. Company No: 12118808. VAT No: 328275783
© 2022 Upstream Applied Science Ltd. Registered in England and Wales. Company No: 12118808. VAT No: 328275783