Jane,
I believe the formula is correct. the cubic term comes from p=z^3 being the
pressure manufactured solution. so in (pI-2e) you get a z^3 term and indeed a
linear term in the 2e portion.
I see. I missed that there is of course also a pressure.
This is how I have come to conclude that it is how i am implementing the
Neumann conditions that are causing the problem, since with Dirichlet
conditions the solver is good. All i can figure out is that perhaps we have
calculated the vector g_N (Neumann condition) incorrectly, but that has been
checked multiple times - if anyone notices something stupid we have done here
then do let us know too.
Well, then that is the point to debug the problem. You've established that the
solution is wrong. Now you need to figure out how exactly is it wrong. Is the
solution off by a constant? Is the profile wrong? If you look at the error as
a cellwise constant function, is there a pattern? What happens if you multiply
the prescribed boundary values by two -- is the error also multiplied by two?
Does the pressure stay correct in that case, but the velocity change in some
way? Is the sign of the velocity correct? If you collect enough clues, you can
typically guess where the problem must lie.
I will for now use vector extractors like you suggested and see if that helps
at all?
As you already found out, that's not helping with the bug. It just makes the
code simpler.
Best
W.
--
------------------------------------------------------------------------
Wolfgang Bangerth email: bange...@colostate.edu
www: http://www.math.colostate.edu/~bangerth/
--
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see
https://groups.google.com/d/forum/dealii?hl=en
---
You received this message because you are subscribed to the Google Groups "deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to dealii+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
