Bubble column - Eulerian Eulerian
I am a university student and working on my summer project which is about bubble column and I am using Ansys CFX 13.
Following are the details:
Axisymmetric, Hex elements
Eulerian Eulerian mono dispersed
Inlet: Gas - Mass flow rate of gas 28 kg/s, volume fraction 0.1
Liquid - Normal speed 0, volume fraction 0.9
Two symmetric boubdaries and tip nipped at axis where I specified free slip boundary
Initialization: Gas - Velocity automatic, volume fraction automatic
Liquid - Velocity automatic, volume fraction automatic
Monitors: Effective density, Gas volume fraction, Pressure at some
internal points and Maximum velocity at outlet
As a result, I get convergence as shown in figure 1 (Monitors except velocity flat), but volume fraction everywhere is 1 in the contour plot as seen in CFD post!!
I changed the initialization of liquid to volume fraction 1. But then pressure monitor oscillates with least and max value being -25 bar and 0 bar and won't go flat.
Both the scenarios are unphysical and I dont seem to know what is going wrong.
Have you considered this FAQ: http://www.cfd-online.com/Wiki/Ansys...gence_criteria
Thanks Glenn for quick reply.
Yes, I have tried all options for steady state simulation.
I found out approximate residence time and set value of physical timescale accordingly (I tried 3/4th, haldf, quarter of it). Since this was unsuccessful, I used local timescale factor as 5. Though the residuals are dropping and monitors seemed flat, still I get volume fraction of 1 everywhere in CFD post. Further, I found pressures at the bottom portion of the tank of the order of -300000 Pa. This can not be true given the tank is 15 m high and buoyancy is enabled, meaning there will be hydrostatic head.
Do you see I am missing something? Gas volume fraction of 1 everywhere doesnt make sense, as one of the tutorials in CFX has inlet air volume fraction more than my case, and yet the air volume fractions everywhere (except region just above sparger) are negligible.
I believe that some form of steady state (and physical) results should be possible for this simulation, before discarding SS approach entirely.
Rather than trying a smaller time step try a larger one. This is described on the FAQ.
|All times are GMT -4. The time now is 23:16.|