Using computational fluid dynamics and optical sensor technology to scale cell culture platforms
Date Degree Awarded
Restricted to Claremont Colleges Dissertation
PHD in Applied Life Sciences
First Thesis/Dissertation Advisor
Second Thesis/Dissertation Advisor
Third Thesis/Dissertation Advisor
Different cell culture vessels ranging from micro scale to laboratory scale to commercial scale play critical role in upstream process development for biologics manufacturing. Based on the mode of operation, cell culture vessels have different hydrodynamic environments, making it challenging to scale. Integrated approaches using computational tools supported by experimental studies can overcome these challenges. Computational Fluid Dynamics (CFD) is one such tool that can simulate hydrodynamics within the cell culture vessels and can provide insights at macro and micro-scale. Accuracy of a CFD model significantly depends on the fluid model and assumptions. Traditionally, simple two-equation fluid models were developed to simulate 2D flows. Unfortunately, these models are known to over predict hydrodynamic variables, such as shear and strain rates, in 3D complex flows, such as swirling and stirring, which occurs in these cell culture vessels. The lower cost and ease of operation makes shaken vessels a perfect candidate for high throughput process development platform. Due to inadequate characterization of shaken vessels, scaling with stirred vessels is challenging. Traditionally, eddy viscosity models (EVM), such as κ-ε and κ-ω models, have been implemented for flows with simple two-dimensional geometries. For three dimensional complex flows, such as fluid re-circulation, flow over curved surfaces and highly swirling flow Reynolds Stress Model (RSM) is hypothesized to predict flow properties with higher accuracy. The 1st chapter of this thesis focused on scaling deep well plates, unbaffled and baffled shake flasks using RSM. CFD modelling with RSM predicted volumetric mass transfer coefficient (kLa) within 10% of experimentally determined values and provided insights on additional hydrodynamic variables. This demonstrated the use of RSM in facilitating scale up of different shaken cell culture vessels.Bioreactor hydrodynamics differs from shaken vessels due to the stirred motion, sparged gases and bubble size distributions generated due to varied turbulence and sparge rates. To account for these additional forces and conditions, more complex fluid models need to be implemented. The 2nd chapter of the thesis focused on developing a robust bioreactor CFD model using Shear stress Transport (SST) - κ-ω model in combination with advanced Eulerian Multiphase Models, such as Population Balance Model (PBM). The proposed CFD model was able to estimate kLa within one standard deviation of experimental values making this CFD model an alternative and cost-effective way to scale-up or scale-down bioreactor processes. CFD models and data generated from the simulations provided significant insights into the hydrodynamic differences between shaken and stirred vessels. The 3rd chapter of the thesis focused on establishing cell culture scalability between vessels. To scale these vessel, two different strategies were tested to match cellular growth, productivity, and product quality across all scales. One involved keeping the volumetric mass transfer coefficient (kLa) constant, the second, preventing dissolved oxygen (DO) levels from falling below a minimum threshold value. The later method was found to accurately scale up two different cell culture processes and two separate cell lines. An approximate 10 % maximum variability in peak cell density and less than or equal to 5 % variability in productivity and product quality were observed with the later method. CFD variables, such as energy dissipation rate (EDR) and velocity λ2, were compared between different vessels and were observed to validate the later scaling strategy.
2023 Makwana Mandar Narayan
Makwana, Mandar. (2023). Using computational fluid dynamics and optical sensor technology to scale cell culture platforms. KGI Theses and Dissertations, 28. https://scholarship.claremont.edu/kgi__theses/28.