the BC's would need to be symmetric as well - try it with all the walls having an environemental BC. Even still the resulting flow maybe more or less symmetic but that won't mean it will necessarily be axis symmetric - non varying in azimuth. Depends on how the eddy's form up.
Dear Bill McEachern,
Thanks for your reply.
Followed your opinion, I replaced all wall’s BC with environmental pressure BC and re-ran the problem. The following figure is the result. It also looks quite strange.
I saw the similar air diffuser’s smoke test video before. The smoke distribution looked like a circle, not a star.
I didn’t have many experiences using Flow Simulation. I am not sure that the wrong results I got due to my wrong program setting. Or, the Flow Simulation’s character could not get symmetrical result. Would you please give me your suggestion!
(P.S. I also used a dense grid system to simulate the problem. The velocity contour still looked like a star.)
Thank you again for the answer!
While setting up the cut plots, turn off the check box for 'Interpolate Results'. This will show the digital nature of the mesh and the results.
You may have a dense grid through the room and still have a coarse mesh in the diffuser. It may not be able to resolve the fine details of the diffuser. Show a close up of the diffuser as a cross section, preferably with the mesh displayed.
Feel free to post your model.
Hi Rich, thanks for your reply.
Mesh (right side)
Mesh level is level 3.
I turned on the dialog box for 'Advanced narrow channel refinement' and 'Optimize thin walls resolution'.
I used local initial mesh to diffuser. Local initial mesh is level 3.
I turned on the dialog box for 'Manual specification of the minimum gap size' , 'Manual specification of the minimum wall thickness' and 'dvanced narrow channel refinement'.
The minimum gap size is 0.0215 m (diffuser's gap size), and the minimum wall thickness is 0.0007 m (diffuser's wall thickness).
Cut plot (no interpolate)
This figure is cut plot with no interpolate.
Thank you again for your reply!
based on that mesh, you're likely going to need a finer local mesh at the diffuser. Resolving a thin circular part in a large volume with rectangular cells will need a much finer mesh.
I'd also suggest using mesh level 6 - this is the same size to start as level 3, but it enables dynamic refinement. This allows Flow to refine the mesh in areas with high gradients (near the diffuser) and make it coarse in areas that do not require refinement (near the walls and floor).