simple calculation fails

[Anonymous]

Hello everybody,

I've a problem with cfd simulation.

I designed a cylinder in inventor 2k13 which I imported into cfd.

In cfd I then put an external volumetric box around my cylinder.

Material wise I choose metal for my cylinder and air for my external vol-box.

That being done I choose my boundary conditions:

One side of my external vol-box got 10m/s for flow-speed.

On all the other 5 sides I set 0 pa overpreassure so it simulates an open area.

Last but not least a mesh, created automatically.

I'd say with this setting you should be ready to go and get a good result, right?

Well, not for me. You can see my result in the attachment.

Airflow comming from the top side, going down.

That shouldn't be the image for this case but more something like this:

cylinder-flow

Maybe somebody can helt me with my problem.

I would be very thankful since I broke my experiments down to the nearly most simple case and still don't get useable results.

Thanks in advance,

Pat

[Jon.Wilde]

Hi,

I would suggest a few things here:

1) Are you running this in 2D? This is the best approach as then you can achieve an excellent mesh and still have reasonable runtimes

2) The domain looks like it needs to be larger, you need to ensure really well developed flow before reaching the cylinder and then be sure to capture everything downstream. The walls need to be far enough away to have no effect 

3) Typically these are transient solutions - the eddies shedding over time and constantly changing, are you running as such?

4) For the Boudnary Conditions, is would probably be best to keep it simple, one inlet and one outlet on the opposite end

Give those changes a go and see how it look.

Kind regards,

Jon

[Anonymous]

Hello Jon,

thank you for your help!

I will try your suggestions and give it another go.

Could you please explain to me how I run the 2D?

Kind regards

Pat

[Jon.Wilde]

Hi Pat,

There is not much difference between 2D and 3D. Boundary conditions are applied to edges rather than surfaces, materials to surfaces rather than volumes.

CFD will assume a unit thickness, so if you are in mm, it will assume the whole thing is 1mm deep (just incase you are using a volumetric flow rate or similar).

The whole thing will be much faster!

Thanks,

Jon