search

precice / calculix-adapter

preCICE-adapter for the CSM code CalculiX
GNU General Public License v3.0
52 stars 20 forks source link

No support for Stress data #111

Open MakisH opened 1 year ago

MakisH commented 1 year ago

Our Turek Hron FSI3 tutorial (OpenFOAM-deal.II) exchanges stress and displacement data. However, CalculiX does not currently know how to handle stress data. Supporting stresses would make adding a CalculiX case to this tutorial much easier.

boris-martin commented 1 year ago

Do you want to make the adapter able to write stress ? Because reading Pressure is already possible since https://github.com/precice/calculix-adapter/pull/91 But AFAIK solids write displacements and read stresses/pressure, so this should be fine

MakisH commented 1 year ago

Oh, then we should probably add the word "stress" somewhere in this page: https://precice.org/adapter-calculix-config.html

We only need to read stress. I assume that pressure is only a scalar with the normal component. Is that correct?

boris-martin commented 1 year ago

Yes. So no tangential traction is possible here. I don't know if it's a CalculiX limitation or something else. Maybe it's useful to have a wrapper that reads a vector input and takes the normal part ? (But it's actually "reading the full stress tensor" I think... but preCICE doesn't support tensor data)

MakisH commented 1 year ago

Yes. So no tangential traction is possible here. I don't know if it's a CalculiX limitation or something else.

I don't know if CalculiX has any such limitation. Maybe @MatthiasFreimuth has some clue here.

Maybe it's useful to have a wrapper that reads a vector input and takes the normal part ?

What do you mean by "wrapper"? This already sounds doable with preCICE actions.

(But it's actually "reading the full stress tensor" I think... but preCICE doesn't support tensor data)

But then how do we do it in deal.II? @davidscn

boris-martin commented 1 year ago

Maybe it's useful to have a wrapper that reads a vector input and takes the normal part ?

What do you mean by "wrapper"? This already sounds doable with preCICE actions.

Didn't think of actions, I guess it could work! By wrapper I meant updating the adapter to add a conversion layer, but actions are probably cleaner

MatthiasFreimuth commented 1 year ago

As far as I know, there should be no limitations from CalculiX side regarding tangential traction.

davidscn commented 1 year ago

But then how do we do it in deal.II? @davidscn

We just read a Stress vector. On the OpenFOAM side we also compute it via normal vector x stress tensor.

precice-bot commented 6 months ago

This issue has been mentioned on preCICE Forum on Discourse. There might be relevant details there:

https://precice.discourse.group/t/can-stress-be-transmitted-to-calculix-instead-of-force-in-fsi/1836/3