code_saturne User's Forum

Hello,

I'm a new user of Code Saturne and I have some questions about the turbulence model Rij-EBRSM.
I am currently studying the flow in a wind tunnel (cf. picture). I tried to use this turbulence model for a steady state but the results always diverge. I have tried different meshes and different parameters but it seems that it doesn't change anything. I don't know what I am doing wrong.

Is there any tutorial or any guidelines for this model ?

Regards

Hello,

Your diagram has no axis legends, so I assume it's an inlet profile, but I can't be sure.

Please read the forum guidelines to see what Information to provide; with the litle you provided, I can only guess...

Your problem might be due to issues unrelated to EBRSM (bad mesh, too large time step, ..., anything not recommended in the best practice guideiines from the documentation section of this forum), or to EBRSM itself...

For EBRSM, using a non-flat initial profile may help avoid relaminarization (check the cs_user_boundary-conditions auto-inlet profile user example for this). Though in most cases with relaminarization, divergence is not immediate, but occurs after some number of time steps.

Regards,

Yvan

Hello,

I am working with code-saturne2.3.4.
Here are the mesh and the xml file, the listing and preprocessor.log Regards,

Myriam

Hello,

You computation does not crash immediately, but diverges progressively. This looks quite a bit like the relaminarization issue I mentioned in my previous e-mail, so using an non-flat inlet profile (using the automatic inlet profile example, or starting from it) might do the trick.

In any case, version 2.3.4 has not been deprecated since 3.0 was released more than one year ago, and is not maintained anymore. Many bugs have been fixed between 2.3.4 and 3.0, so upgrading to 3.0, 3.2, or 3.3 (if you wait a few more days) is strongly recommended.

Best regards,

Yvan

Hello

I am now working with version 3.2.2. I have done the same computation as before with (cf listing2) and without the automatic inlet profile example (cf listing1). The results are not better so maybe there is something else that I am doing wrong.
Am I using the automatic inlet profile correctly ?

Regards,

Myriam

Hello,

I'll check your logs later, but to check if you're using the auto-inlet profile correctly, you can simply test it with Rij-SSG, which is much less "fragile" than EBRSM.

If it works correctly, the profile starts ou flat (almost flat for EBRSM), then slowly converges to a profile matching the mesh and turbulence model. It assumes your mesh is orthogonal (ideally extruded) at the inlet.

Best regards.

Yvan