Replace deal.II cylinderFlap tutorial with fsi3

[davidscn]

Replaces #105.

Resolves #94 and (at least partially) #77. This PR only updates the real 2D case of deal.II and not our pseudo 2D case. Potential tests will fail since this tutorial requires groovyBC.

There is one potential incompatibiliy with newer .org versions due to the residual setting in Fluid/system/fvSolution.

[uekerman]

Looks good. I was able to run (the first 100 timesteps of) the case and the results look meaningful. Installing SWAK worked out of the box for me.

Minor things:

• Please format your preCICE config with https://github.com/precice/precice/tree/develop/tools/formatting
• Could you please add a plotDisplacement (gnuplot) script?

@MakisH Could you please comment on the incompatibility with newer .org versions? How do we treat this for other cases?

[MakisH]

There is one potential incompatibiliy with newer .org versions due to the residual setting in Fluid/system/fvSolution.

@MakisH Could you please comment on the incompatibility with newer .org versions? How do we treat this for other cases?

This refers to this issue: #40

I fixed that in #78, essentially by adding a commented-out version for OpenFOAM 6 or newer in fvSolution:

https://github.com/precice/tutorials/pull/78/files#diff-47adf48426d6c46fe2b3a89be87fe0a253244777e6e10c7b40f3d03e578339ca

We already have this patch here, so I am not sure what else we should be looking for, @davidscn.

[davidscn]

We already have this patch here, so I am not sure what else we should be looking for, @davidscn.

I guess there is currently an incompatibility with newer org versions. I would nevertheless keep the setting as they are. As you wrote in #78

This may affect convergence, but the purpose of the tutorial is to demonstrate how to configure preCICE and the adapters.

but in this case, we really care about convergence and the total number of iterations. The settings are related to the OF paper as well. In case anyone really wants to run this cases with the org versions, the commented out version is still there and comments are sufficiently given. So, I would just keep it this way.