Closed mschreter closed 3 years ago
Many thanks Peter, looks good to me.
Exactly, the velocity rate is then proportional to the pressure gradient:
From my side, this PR is ready.
One question: the velocity in the 1D case does not seem to converge to a fixed value.
In addition to what @mschreter said, you need to consider that incompressible flow in 1D needs to be constant throughout the domain (due to div u = u_x = 0
), so the only driving force is the pressure gradient. If that gets excited as in the case here, we must have an increase in velocity.
@kronbichler: Thank you for the review, I have included your suggestions. I suggest that we wait with merging until #dealii/11952 is closed.
This PR enables to solve 1d-Navier-Stokes problems, which makes the integration of Adaflo into
MeltPoolDG
also fordim==1
straightforward. It depends on PR #11914 of deal.II. I added two test cases1d_flow.cc
for testing the 1d-implementation (although it is very boring :-))couette.cc
due to personal interest