CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   solid-liquid suspension with multi-reference frame (http://www.cfd-online.com/Forums/fluent/29021-solid-liquid-suspension-multi-reference-frame.html)

Min-Hua Wang October 25, 2001 14:47

solid-liquid suspension with multi-reference frame
 
Hello, All:

Does anybody have experience in the simulation of the stirred tank for solid-liquid suspension with multi-reference frame.

Thanks

MHW

Alain October 29, 2001 05:00

Re: solid-liquid suspension with multi-reference f
 
I do,

what is your question ?


Min-Hua Wang October 29, 2001 12:39

Re: solid-liquid suspension with multi-reference f
 
Thanks.

I am working on CFD simulation for a stirred tank without baffle blade. I was using Mixsim to setup multi-reference frame and to start single phase simulation. The working fluid is water. The convergency is good. Then I run two-phase simulation withing fluent 4 environement. The solid has 200 micro meter in size and the solid density is 1200 kg/m^3. Volume fraction f solid is about 15%.

I run simulation in time transient flow with Eulerian (Granular) model. Also I selected Gidaspow model for granular viscosity. Schuh model is selected for Drag force. The rest I left tehm by default data.

How do I justify the final solution. I always observed that the mass residual is very hard to be reduced. The suspension of solid seems to me not developed. How long should I run the simulation.

Can you suggest me some experience? Many thanks. MHW

Alain October 29, 2001 14:59

Re: solid-liquid suspension with multi-reference f
 
Hi MHW

I did a simulation very similar to yours a couple years ago. Multiphase model can be very difficult to handle, depending on your flow condition.

First of all, if you don't have any baffle you can run your problem as a single rotating frame. This may help your convergence.

You should underelax your interphase momentum exchange term even if you run a transient simulation. it may help to suppress this momentum exchange at the begining of the calculation.

In some case you'll need to deactivate multigrid for pressure. Convergence is longer but the space marching solver is more stable than the MG solver for multiphase calculation.

At last, I will suggest you to lower your time step. if your calculation is too long and you are looking for the steady state solution you can also make fewer iteration in each time step.

I hope this could help you.

Alain

Lanre October 30, 2001 12:54

Re: solid-liquid suspension with multi-reference f
 
MRF is a steady-state model ONLY, hence, is incompatible with the Eulerian-Granular multiphase model in FLUENT 4.

Since you have no baffles, delete the MRF zone and convert to a single frame of reference (See FLUENT 4.4 Users' Guide). Now you can solve the transient mixing and include the Eulerian-Granular model for your multiphase calculations.

Min-Hua Wang October 30, 2001 21:56

Re: solid-liquid suspension with multi-reference f
 
Now can i use MRF model and Eulerian-Granular model in steady state simulation?

Thanks.


Alain October 31, 2001 05:28

Re: solid-liquid suspension with multi-reference f
 
Yes you can use these models in steady state simulation but convergence can be very difficult to reach.

As Lanre said, if your tank has no baffles, you do not need a MRF approach, a single rotating frame will be more correct as it can be use in both transient and steady state.

A point about MRF and transient simulation : MRF is a steady state approach so it is true that you may have no significant results if you use it for a transient simulation. Nevertheless you can use it in a transient simulation if you are looking for steady state solution : the intermediate, time dependent, solutions may be not significant, but the converged, steady state solution does.

Best regards

Alain

tantular April 30, 2010 18:22

Please observe the solid volume fraction during the iterations. I have problem that the solid volume fractions decreased during the iterations. After the iterations converged, the solid volume fractions is different with the initial value when I patched in the beginning of simulation.


All times are GMT -4. The time now is 14:48.