CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FLUENT (
-   -   3D mixing using moving reference frame (

maplebird August 13, 2010 20:41

3D mixing using moving reference frame
Hi all

I'm a new user of fluent and i'd like to know how to simulate a mixing tank with a single rotor.
I defined fluid zone, rotor surface wall and mixer wall. Then i run it in Fluent. i wasn't sure how to set boundry conditions. What i did is that
Mixer wall: stationary wall
rotor surface wall: moving wall; absolute velocity and i put in a speed 1000 rpm
fluid: i put stationary.

I'd like to know did i do somthing wrong there? Thank you very much


xrs333 August 14, 2010 11:11

unfortunately, you did.
The motion type of the fluid zone should be set as moving reference frame with rotational velocity equals that of the rotor, and the rotation-axis origin & the rotation-axis direction be define properly. Then go to the boundary condition panel and set the mixer wall as stational wall, the rotor wall as moving wall with a rotational velocity of zero raletive to adjacent cell zone. -- this is the single rotational reference frame model.
But you didn't tell us if there are baffles in the mixer. the SRF model is not valid for this case.
Good luck.

xrs333 August 14, 2010 11:23

the rule of a thumb is that no wall boundary should have a non-zero component of velocity in the reference frame for the adjacent cell zone unless a dynamic mesh is used.

maplebird August 17, 2010 13:30

Thank you very much xrs333 for your reply.

My model doesn't have any baffle for the moment (only wall and a rotor). When I applied what you wrote, I found that I got contradictory results when axis speed is the highest. I think SRF is probably not good for my case. And my wall is not surface of revolution neither, therefore Multiple reference frame seems not valid. Moreover. I also wanted to set up an interface to make moving/stationary fluid zone, but its not so convinient for a 3D and I don't know how this interface will be constructed like (compared to 2D, in which a simple circle can be used). I will try again what u said anyway.

maplebird August 17, 2010 15:10

I'm not conviced to set the fluid zone as rotating and then the rotor as moving to adjacent cell zone cause my result was that longer distance from the rotor, higher the speed is; which is contarary to the reality. COuld you explain to me?

xrs333 August 17, 2010 16:10

if you use moving reference frame, you may use relative velocity for postprocess.

maplebird August 18, 2010 12:47

Hello xrs333, thanks. actually later i tried to create interfaces in fluid zone and I divided the fluid zone into moving and stationary zone. Then i ran the MRF and it seems good. However i randomly divided the fluide zone using a face and I dont know if the choice of the face can give different result or not.

Mansureh August 18, 2010 14:05

Hi maplebird
can I ask in Which mesh generation software you made your geometry? If you made it in Gambit software , you can follow this procedure to have interface between fluid zones:
you should make both moving and stationary zones and splite the moving zone with stationary zone. In real, the stationary zone will divide to 2 domains, specify the faces between 2 zones as wall, after meshing the geometry and It's exporting to Fluent, you will see that wall and Its shadow in boundary condition dialog box. change It's type to interior; wall BC and it's shadow will disappear and fluid can move between 2 fluid zones.

cfd_confuse October 1, 2010 14:35

converging problem to mixingtank
Hi xrs333, i have tried your method of setting the boundary conditions for my mixing tank but the solution did not converge. i am currently using a 0.5m by 0.5m tank. there is 400,000 cells for the moving zone and 80,000 cells for the rest of the fluid in the tank. The rest of the boundary condition is use as the default given in fluent5/6. But solution starts to diverge at around 7 x 10^-3.. any possible reason for this?

All times are GMT -4. The time now is 20:22.