Julian Andrej

You are setting the dirichlet attribute to zero meaning they're now open boundaries aka you're implying "zero neumann"

It's relevant how you project your initial condition. Can you show a minimal example?

https://godbolt.org/z/8a3T6rEfa works as expected

> MFEM's C++11 requirement prevents us from using it. The project itself is not restricted to c++11 right? There is no requirement to use MFEM master during development and have...

See https://github.com/mfem/mfem/blob/navier-urans/miniapps/navier/examples/navier_fsi.cpp#L417

Here you go [fsi.msh.zip](https://github.com/mfem/mfem/files/14380506/fsi.msh.zip) [fsi.geo.zip](https://github.com/mfem/mfem/files/14380524/fsi.geo.zip)

The only fluid solid interface attribute for the mesh I posted is attribute 5 (array index 4). What am I missing?

Can you provide a small failing example code? You have to be careful how you project those functions.

You're on a recent master branch right?

See https://github.com/mfem/mfem/pull/4151