Dear colleagues,
for simulating a rotating impeller I have used MRF approach,
Although I have changed all the walls to moving wall relative to adjacent cells with with zero velocity and same direction, fluent error remains,and continuity remains high.
I think it would be because of a mesh problem?
Do you have any suggestion?
Ur doing Mrf, then there should b two frames. One rotating and one stationary. Then the walls between two domains should b made interference.
You should make an interface because it needed to performing MRF study. Moreover, if you only consider the flow variation solely in the rotating part then instead of MRF, model the flow in a single reference frame. Beside that i suggest you to share your complete geometry so it will be easy to point out the problems.
In my experience, MRF is never a good way to simulate the turbomachinery. Sliding mesh of unsteady simulation is a better way. If you are only interested in time-averaged result, do the average on the instantaneous flow fields.
In this study, stationary and rotating blades of the mill is simulated with MRF. Moving and stationary zones are separated with interfaces.
http://www.powerengineeringint.com/articles/print/volume-17/issue-5/features/cfd-analysis-bowling-for-pulverized-coal.html
Dear yang
I thought that I maybe computationally cheap and have a better robustness to use MRF instead of sliding mesh for stady cases, while its accuracy is lower.
Does you valuable experience different with my attitude?
regards
Hi Mahdi,
If your inflow condition is axisymmetric, MRF still should be no problem. If you have non-uniform inflow, sliding mesh is better.
Regards!
- Badges
- Science topic