Eulerian Multiphase simulation, the simulation is simple a 2D tube filled with water (primary phase) and 5mm glass bead (2ndphase), the inlet boundary was set from bottom of the tube with constant velocity of 0.01 m/s and environment pressure at top of the tube, this is a simple fluidized bed setup. The illustration of the initial condition of simulation setup are shown in the attached file as "Illustration of Initial Model"
The setup detail of the model is shown below:
Simulation type : Eulerian Multiphase with VOF
Analysis type : Transient Implicit
Time Step : 0.001 s
Total Time : 60 second
Glass bead phase : granular 2ndphase with VOF phase localized discretization of 0 (because in future I need to include air bubble into the simulation, so for glass bead 2nd phase, I use VOF phase localized discretization of 0 which is similar to diffused phase setting)
Glass bead density : 2230kg/m3 (denser than water)
Glass bead diameter : 5mm
Drag Model : Wen Yu
Maximum Packing limit : 0.6
Initial volume fraction of glass bead : 0.6 (same as packing limit)
I found out if the inlet velocity is too low (let say 0.01m/s) and cannot achieve minimum fluidized bed (mean the glass bead fluidized bed is stationary at the bottom of the column), the model become unstable, the residual value will have sudden jump and failed to converge in the interval of around every 30 time steps (0.03 second), the residual plot are shown in the attachment "Residual Plot"
When the velocity is set to 0.04m/s and above (over the minimum fluidized velocity), the fluidized bed starts to increase in height, the volume fraction starts to reduce below 0.6, and the simulation do not have convergence problem anymore.
I suspect this phenomenon is due to the volume fraction of 2nd granular phase is close to 0.6 (maximum packing limit) and the 2nd phase is unstable.
In future I will migrate this fluidized bed from this simple tube to a stirred tank. I believe after the stirred tank flow achieved steady state (around 60 second), the velocity in this fluidized bed will over the minimum fluidized velocity and no convergence problem. However, the real case is, just after the stirrer of the tank start to rotate, the velocity in the fluidized bed is still close to zero, it will take time for that region to achieved certain velocity, it will unstable/fail to converge before the tank achieved steady state.
Anyone have idea of how to stabilize the granular 2nd phase (fluidized bed) even with very low inlet velocity?
Thank you for your time.
What relaxation factors are you using? Anyway, try applying underxation factors to the volume fraction
Hello
Make sure your mesh is fine enough.
Also the turbulence model used is it okay for this application.
And the solution method also.
Try to use non-iterative ' fractional step
Most of the times, stability can be achieved by negative feedback. That means you have an advanced sensor that tells the one at start to introduce small changes in the variables to regain stability. Most of the work then becomes finding the expression for this feedback. There is a huge literature on this subject.
Best wishes.
Dear Jian Ping Tan
Start with 1st-degree equations order first then when the solution is stable try to change it into 2nd order.
Regards
Dr. Ahmed
I refer you to two discussions which bear on the two problems of Eulerian two-fluid models closures: Turbulence and momentum interfacial transfer:
(6) Turbulence Closure for Two-Fluid RANS Modeling (researchgate.net)
(6) Interfacial Transfer Closure for Two-Fluid RANS Modeling (researchgate.net)
- Badges
- Science topic
- Similar topics
- Geoscience
- Atmospheric Sciences
- Atmosphere
- Air