[firedrake] Prescribe pressure at a corner for Darcy's equation

[David Ham]

Hi Justin,

This is expected behaviour. Your attempt to impose a pressure BC
using: dot(w,n)*patm*ds(6) will do nothing in the no-flow condition because
the no-flow condition restricts the test space to the space for which
dot(w, n) = 0. Your surface integral therefore contributes nothing to the
form. I'm not even sure that the Stokes equations admit simultaneously
specifying no flow and a fixed pressure. Intuitively, the no flow condition
implies that the pressure rises at the boundary to the level required such
that the pressure gradient normal to the boundary is zero and no flow
results.

Regards,

David


On Mon, 13 Feb 2017 at 07:25 Justin Chang <jychang48 at gmail.com> wrote:

> Hi again,
>
> So this may or may not be related to the original issue. I was able to
> define corner IDs needed for my pressure BCs but I cannot seem to enforce
> both no-flow and atmospheric pressure simultaneously. Enforcing no-flow
> seems to negate the weakly prescribed pressure flux and my pressure
> solution is complete wack. On the otherhand, not enforcing no flow gives me
> a pressure profile I expect, but I no longer have a no flow domain.
>
> Attached is my code, the custom boundary IDs begin at line 39, the weak
> formulations begin at line 164. Run the code as:
>
> python 2D_Darcy_Homo_Iso.py 600 150 2
>
> The code right now will generate a nonsensical pressure solution because I
> have no-flow enforced on all boundaries (including the corners).
>
> Any ideas on what to do here?
>
> Thanks,
> Justin
>
> On Tue, Feb 7, 2017 at 4:55 AM, Lawrence Mitchell <
> lawrence.mitchell at imperial.ac.uk> wrote:
>
>
> > On 7 Feb 2017, at 09:03, Lawrence Mitchell <
> lawrence.mitchell at imperial.ac.uk> wrote:
> >
> >
> > PETSc.Options("foo_").getAll()
> >
> > => {"bar": "-baz_bar"}
> >
> > I will push a fix for this one.
>
> Please try the branch proposed here:
>
> https://github.com/firedrakeproject/firedrake/pull/1004
>
> Lawrence
> _______________________________________________
> firedrake mailing list
> firedrake at imperial.ac.uk
> https://mailman.ic.ac.uk/mailman/listinfo/firedrake
>
>
> --
Dr David Ham
Department of Mathematics
Imperial College London
-------------- next part --------------
HTML attachment scrubbed and removed