(218a) Towards an Understanding of Fluid History in a Turbulent Mixing Tank Using Large Eddy Simulation Model | AIChE

(218a) Towards an Understanding of Fluid History in a Turbulent Mixing Tank Using Large Eddy Simulation Model

Authors 

Flamm, M. - Presenter, Merck & Co., Inc.
Raudenbush, K., Merck and Co., Inc.
Sirota, E., Merck & Co.
Cote, A., Merck & Co., Inc.
Many processes such as fermentation, crystallization, and chromatographic resin preparation require sufficient mixing to ensure production, homogeneity, and suspension but are sensitive to high levels of shear due to mechanical agitation. Process scaling can be particularly problematic due to differences in the distribution of shear and bulk mixing during scale up that is hard to replicate at smaller scales. Computational fluid dynamics can be utilized to understand the full 3D process behavior during scale up/down of such processes to debottleneck process development, improve scale up robustness, and develop engineering understanding of fundamental processes.

The typical analysis of strain or energy-dissipation utilizes histograms or percentiles of strain/energy-dissipation in the tank on a spatial basis, i.e. time-averaged. A Reynolds Averaged Navier-Stokes (RANS) model can only provide a time-averaged view, or a pseudo-time-averaged view when using unsteady-RANS, of the process. A large eddy simulation (LES) model provides information about the largest scales of motion and therefore can be utilized to start analyzing the time history of the fluid. This analysis has been difficult previously due to inadequate computational resources to fully sample fluid trajectories in a mixing tank.

We will present the time history of strain/energy-dissipation in mixing tanks across multiple scales using the LES model. First, the time-averaged strain/energy-dissipation predictions will be compared to literature reported values, and the accuracy and limitations of the LES modelling approach will be discussed. The strain/energy-dissipation history can be filtered through signal processing operations that leads to an understanding of power being applied to the fluid at different scales. We will show that there are fundamental relationships between impeller speed and scale that can be used to understand differences in local strain/energy-dissipation.

Topics