Fluids

Fluids

Topics related to Fluent, CFX, Turbogrid and more.

Bioreactor Mixing time CFX analysis (Workbench)

    • Pau Cunillera Bori
      Subscriber

      Hello, I am working on WorkBench CFX Analsys workflow, trying to analysis the mixing performance of a Bioreactor stirred with a Rushton Impeller. Once the single-phase and multiphase sim become available, I would like to perform a parametric analysis of the system to optimize it using surface plots. My inputs should be a range of stirring speeds (rpm) and a range of aerations rates (VVM). Thus, my desired response is the mixing time of the system after a tracer injection. I know how to use the VVM and rpm as a input variables but I am completely lost on how to generate a output variable of the mixing time. Is there any equations, expressions... that I could use?

      I did a mixing time calculation once manually (no parametric study) by solving the system two time with two different stirring speeds, adding a source point (pre) and three monitoring points (post). Then, the tracer concentration evolution in each MP was exported to Excel and I manually calculated the mixing time in each point and condition. This method took me days (solving time, checkin it...) Is there a way to speed up the mixing time calculations with the parametric tool? So, I can calculate a range of rpm and VVMs without simulating each DoE conditions for hours manually.

      Many thanks! ????

    • Rob
      Forum Moderator

      I'm not a CFX user, so more generic comments. 

      If you're using DPM can you run steady state and then run the transient tracer on a frozen flow field? That'll also work with a full multiphase calculation if you don't mind a small error margin as the flow does change with time. 

      In terms of monitoring mixedness. You can't use mean/min/max concentrations as such, as that will be global and there's always a region with zero concentration or similar. In Fluent there is a uniformity index, maybe review that for CFD Post?  I'd not try and parameterise this, it's more a case of running the model(s) and then doing the last bit manually. It will take run time (transient always takes many hours/days) but shouldn't use too much engineer effort. 

Viewing 1 reply thread
  • The topic ‘Bioreactor Mixing time CFX analysis (Workbench)’ is closed to new replies.