Developing a CFD-based axial compartment model for a lab scale bioreactor

Abstract

Processes involving biological reactions are essential for modern society. Food additives, pharmaceuticals, and commodity chemicals are all major industries that utilize cellular activity to generate desired materials. From penicillin to citric acid, products of bioreactors touch many aspects of day to day life. Accurate models are invaluable to minimize the time and cost associated with process design, scale-up, control, and optimization. However, bioreactors provide unique challenges for the development of accurate yet computationally efficient models due to sensitivity to small changes across different time and length scales. Ultra-high-fidelity modeling using computational flow dynamics (CFD) is too computationally expensive for versatile bioreactor modeling, while simple ideal models do not account for how the flow field affects reactor performance. A compartment model (CM) offers the best of both worlds; by sacrificing some resolution, flow dynamics can be incorporated into a fast-solving model. These models show promise in accurate prediction with short computation times, and the resolution can be adjusted depending on the desired modeling application. CMs are generated based on flow information derived either from experimentation or CFD modeling. A sampling of CM approaches is presented in this work. Additionally, a case study is presented, in which an empirical compartment modeling approach is adapted for use with CFD output. The developed one-dimensional axial CM showed good agreement with the CFD model for mixing times from a simulated concentration step change, although a loss of predictive ability was observed with rigorously modeled acid injection. The results indicate that, with minor modification to the compartmental framework, the developed CFD-CM can be made more accurate while remaining computationally inexpensive.