problem with result, SolidWorks Flow simulation
 

Dear CFD Online Members,

I am a new user of the Flow Simulation. I tried to use it to simulate the flow field of a circular air diffuser.
Attachment 13720 Attachment 13721
The dimensions of the room are 12m*12m*3m. The position of the diffuser is in the center of the roof. The inlet velocity is 5m/s.
Attachment 13722

The following are the results (velocity contour, horizontal and vertical views).
Attachment 13723Attachment 13724
Since the shape of diffuser is circular and the shape of room is symmetric with diffuser’s axis, I think the results should be concentric circles around the center point. But it looks quite strange. Are these results right? Or I miss something when running Flow Simulation.

Hope someone can give me your opinions.

Thank you.

It is concentric....until you start to interact with the walls and distance to them.

Hi CCC,

could you please post the last two plots again with switched off lighting (the button in the Flow Simulation tools that has three colored dots) and with mesh displayed more colors and switched off interpolation.
Also please post an image of the flow trajectories with the reference as the inlet surface and a trajectroy length of 20 or 30 meters so the trajectories are not too long. Also did you use an outlet somewhere? Otherwise the flow is just flowing into a compressed room. You could test to apply an outlet surface pressure to the floor of the room and therefore see if the outflow is really circular as you will have no interference with the flow coming back due to the flow.

Regards,
Boris

Hi doviatt,
thanks for your reply.
I tried to decrease inlet velocity to avoid flow-wall interaction. But the graphics in the area of 0.111m/s ~ 0.222m/s also looks quite strange. The following figure is the result. The inlet velocity is 2m/s.
Attachment 13799

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!


Thank you again for the answer!

Hi Boris_M,
thanks for your reply.
Cut plot
Attachment 13800Attachment 13801

Cut plot (mesh)

Attachment 13802Attachment 13803

Cut plot (no interpolation)
Attachment 13805Attachment 13806

Flow trajectories (20m)
Attachment 13807

The following figure show my outlet position.

Attachment 13808

And I used an outlet to replace floor, the following figure is the result.
Attachment 13809
The figure also looks strange.

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!


Thank you again for the answer!

This is one reason I thought went wrong.
If you look at the image of your mesh with no interpolation, you can see that there are very high gradients between the cells right after you leave the vent. The mesh is changing rapidly and you have too high jumps of your velocity values between the mesh cells. You have two options:

1. Create a local mesh around the vent, about four cells in diameter that you can see as the largest cells in the plot and two of these cell sizes into the room measured from the ceiling.

2. Use the adaptive refinement to refine the mesh where you have too high gradients.

The latter version is an automatic refinement where you cannot really control how many cells it creates and in what region, so it does it globally, but it does it where it is needed. The first version you can specifically define a region and the mesh size and therefore better controll how many cells.

In general I would recommend a fluid mesh in the room that is about the size of the cells along the ceiling, so one level finer than you have in the center of the room. This should give you a little better resolution in the room.
If you can send some images afterwards, again the same you did, I can get a better idea of the flow and at least the resolution of the flow should be much better.

Boris