Re: [Libmesh-users] Dirichlet boundary condition setting variable to zero?

[Roy S. <roy...@ic...>]

On Sun, 7 Jan 2018, Renato Poli wrote:

> I am struggling with a Dirichlet boundary condition setting the
> boundary to zero, instead of the value I impose (3.e7). Previously, I
> was using the penalty method, and it was working fine.
>
> Is there any known corner condition where the DirichletBoundary would
> drive the variable to zero? Any idea?

Are you using the homogeneous or the heterogeneous constraint
functions when constraining your system of equations?  Often we solve
for a delta between two solutions (Newton steps, time steps with fixed
BC values, etc) rather than directly solving for the constrained
solution, so we have different methods to handle constraints in the
former vs latter case.  If you've written your own assembly loop,
check it to make sure you're using the appropriate method?
---
Roy