FEniCS OpenFOAM perp-flap and cyl-flap

[richahert]

I implemented the perpendicular flap with FEniCS and OpenFOAM.
It comes with a new Fluid Case that is almost the same as for OF-CcX but since FEniCS is in 2D I changed the thickness to 1.

There is still some code in perp-flap.py that is not relevant for a tutorial that can be deleted.

The displacements are smaller than expected for unknown reasons.

It makes sense to merge this PR together with precice/fenics-adapter#38, the PR brings FSI functionality to the fenics-adapter.

ToDos: perp-flap

• make fluid (in XZ plane) and solid case (in XY plane) consistent (solved by 2837129)
• fix VTK output (in the end we want to obtain identical number of samples for both simulation and a nice visualization in paraview), see FEniCS OpenFOAM perp-flap and cyl-flap #38 (comment), opened issue 0.orig messes up VTK output of foamToVTK openfoam-adapter#95

ToDos: cyl-flap

• test case runs with OpenFOAM 5 or older
• test case runs with OpenFOAM 6 or newer (see Some FSI tutorials are not working with OpenFOAM6 #40)

[BenjaminRodenberg]

With the version of the fenics-adapter on this branch I can run the example and the results look reasonable. There are some open points, though:

• The 2D-3D coupling is not handled properly. While the Fluid "lives" in the XZ plane, the Solid "lives" in XY plane. The coupling seems to work properly, therefore this is mainly a visualization issue.
• Some information in the README.md on expected output and hints for visualization would be good. You can also use the wiki for this. I already created a new page here.

[BenjaminRodenberg]

The 2D-3D coupling is not handled properly. While the Fluid "lives" in the XZ plane, the Solid "lives" in XY plane. The coupling seems to work properly, therefore this is mainly a visualization issue.

Regarding this I added richahert/fenics-adapter@2c9c2a0. This removes the entire swapping of coordinates from the adapter. I think we should try to fix the test cases instead. This means that quasi-2D OpenFOAM simulations should not live in the XZ plane, but in the XY plane. This makes the adapter much cleaner and also solves the visualization issue mentioned above.

[richahert]

Now the physics should be correct. The results for the FSI3 look very much like the deal.ii-OpenFOAM example. deformations are bigger than in the benchmark because of the linear elastic material instead of St. Venant Kirchhoff material in the original FSI3.

[BenjaminRodenberg]

* delete `Fluid_1.vtk`

see #38 (comment)

is still necessary. I think this is an issue of the adapter. Therefore, let's not worry anymore about this here. I opened an issue: precice/openfoam-adapter#95

The rest of the flap tutorial looks very nice!

[BenjaminRodenberg]

I cannot run the OpenFOAM Fluid case for the FSI3 benchmark using OpenFOAM 6. @richahert did you use OpenFOAM 5 or older? Did this work?

I will try to find out how to fix it for OpenFOAM 6.

[BenjaminRodenberg]

Both tutorials are now running on my machine. See #40 regarding compatibility problems with OpenFOAM6.