Apply new conventions on Turek Hron FSI3 case

[uekerman]

This PR applies all main changes from @MakisH and my restructuring discussion on the FSI3 case. This case should serve as a role model to copy from when porting other cases.

Summary:

• I deleted the alternatives OpenFOAM-CalculiX and OpenFOAM-FEniCS and only kept OpenFOAM-dealii. The purpose of the FSI3 case is to have one proper verification case, not to show flexibility of preCICE. So, it's enough to have only one version. OpenFOAM-FEniCS never made really sense here as the structure model in FEniCS is only linear. OpenFOAM-CCX is complicated to port as the CCX adapter currently doesn't support stresses as coupling data. We wanted to keep one tutorial with stresses, however.
• The case is flattened.
• Both solvers are called from the subdirectories, i.e. exchange-directory="..". Run and clean scripts adjusted (I kept the clean script for the moment).
• Basic information is given in the README.
• Consistent names: Stress and Displacement.

@MakisH Please have a detailed look at the README.
@davidscn Please try to run the case and please add a displacement plot of the actual case (with this exact mesh). You can directly push to my branch.

[MakisH]

This adds the binary file nonlinear_elasticity. We need it in the .gitingore.

I also added a few more comments.

Overall, I like it, it only needs a bit of polishing.

[davidscn]

Please try to run the case and please add a displacement plot of the actual case (with this exact mesh). You can directly push to my branch.

I guess you mean the most refined mesh on the Fluid side? Or rather the exact given mesh?

[davidscn]

Another deal.II specific point: since we start the solver now from sub-directories, the log, result and parameter files are all in one directory. We should probably use a specific directory for the solution files.

[uekerman]

Please try to run the case and please add a displacement plot of the actual case (with this exact mesh). You can directly push to my branch.

I guess you mean the most refined mesh on the Fluid side? Or rather the exact given mesh?

The results we show here should coincide with the mesh that we provide.
Which mesh do we currently provide? Compared to the results in the OpenFOAM adapter paper. The mesh should be fine enough to give good results (compute time is not an issue for this case).

[davidscn]

Please try to run the case and please add a displacement plot of the actual case (with this exact mesh). You can directly push to my branch.

I guess you mean the most refined mesh on the Fluid side? Or rather the exact given mesh?

The results we show here should coincide with the mesh that we provide.
Which mesh do we currently provide? Compared to the results in the OpenFOAM adapter paper. The mesh should be fine enough to give good results (compute time is not an issue for this case).

We provide all meshes (refinement level one to three). The default mesh (20k cells) is the most coarse one, the most refined (45k cells) is called blockMesh_double_refined. If we just show a plot of the y displacement from 0 up to 15 sec, differences will probably only be noticeable in terms of the absolute time.

[uekerman]

We provide all meshes (refinement level one to three).

Ah ok. I'll add a note in the README.

The default mesh (20k cells) is the most coarse one, the most refined (45k cells) is called blockMesh_double_refined. If we just show a plot of the y displacement from 0 up to 15 sec, differences will probably only be noticeable in terms of the absolute time.

Yes, let's keep the 20k as default and let's then add all picture generated from that one.

[davidscn]

Perfect,
maybe also add a note for the mapping for the most refined case, where it is probably better to replace the global rbf by sth like:

        <mapping:rbf-compact-tps-c2 direction="read" from="Fluid-Mesh-Centers" to="Solid-Mesh" support-radius="0.011" constraint="consistent" />

[MakisH]

Some minor things to improve. Other than that, looks very good!