Help needed on set a ALE simulation in Code_saturne4.0.2

Questions and remarks about code_saturne usage
Forum rules
Please read the forum usage recommendations before posting.
Post Reply
MartianDuan
Posts: 74
Joined: Fri Aug 02, 2013 4:26 pm

Help needed on set a ALE simulation in Code_saturne4.0.2

Post by MartianDuan »

Hello users,

I am practising on using the ALE method to do the FSI via code_saturne.

I genereated a 2d case which a movable cylinder in the channel. I defined the movemet of the cylinder in the subroutine usalcl.f90 and the mesh viscosity in the cs_user_physical_properties.f90.

But shortly after the FST triggered. The simulatio came across with an error :' Abort due to the detection of a negative control volume'.

The details of the simulation are included in the attachment. Please have a look and give me some advise.

Thanks in advance,

Yu.
Attachments
MovingCYL.tar.gz
(1.14 MiB) Downloaded 214 times
Martin FERRAND
Posts: 44
Joined: Wed Mar 14, 2012 10:06 am

Re: Help needed on set a ALE simulation in Code_saturne4.0.2

Post by Martin FERRAND »

Hello,
I cannot debug from my phone, but that it may happen that the cylinder is moving too much, and therefore the algorithm which is computing the mesh displacement (which is not really robust) make nodes crossing one to the other (and you obtain a negative cell volume).
What I would advice is:
- decrease the time step,
- decrease the displacement of the cylinder,

If you want to have your output on a moving mesh (and not the mesh at the initial time step), you have to specify "TRANSIENT_COORDINATE" in cs_user_postprocess.c.

Best regards
Martin
MartianDuan
Posts: 74
Joined: Fri Aug 02, 2013 4:26 pm

Re: Help needed on set a ALE simulation in Code_saturne4.0.2

Post by MartianDuan »

Hi Martin,

Thanks for your reply. I am have the moving mesh in the postprocess now. So i can investigate the simulation case a little more.

I find the simulation come across the error (negative volume) just after several iteration as the ALE was turned on.

I had tried you advise decreasing the displacement and time step. Both of the methods do help a little (making the simulation running several iteration more). But the case still diverge shortly. So I am wondering whether there is some general errors in my setting, i.e. the types of the boundary conidtions and numerical schemes.

Looking forward to hearing from you soon.

Best regards,

Yu.

Martin FERRAND wrote:Hello,
I cannot debug from my phone, but that it may happen that the cylinder is moving too much, and therefore the algorithm which is computing the mesh displacement (which is not really robust) make nodes crossing one to the other (and you obtain a negative cell volume).
What I would advice is:
- decrease the time step,
- decrease the displacement of the cylinder,

If you want to have your output on a moving mesh (and not the mesh at the initial time step), you have to specify "TRANSIENT_COORDINATE" in cs_user_postprocess.c.

Best regards
Martin
Post Reply