Here is some info which might help.
I use FE_Q with GaussLobatto nodes.
I use DoFTools::map_dofs_to_support_points
I remember reading that Luca has recently done some sorting of support
points or quadrature nodes, not sure exactly. This might affect the
behavior of DoFTools::map_dofs_to_support
Dear all
I have included below an example code which converges with
commit ed1bdc6fc1931a664bc30920f66f054907a49cce
but fails to converge with
commit d1e3ffcaf6733e7d8509d93785b1f9e2b16a977e
In this code, I solve the Winslow equations using continuous finite
elements. I am using curved boundar
I implemented an exact solution for the mean curvature for ellipsoids
(it's a slightly ugly expression, I just took it from
here: http://math.stackexchange.com/a/540820), and I computed the L2 and
Linfty errors between my computed values and that expression. The
result: Convergence wrt mesh ref
On Tuesday, August 23, 2016 at 12:51:53 PM UTC-4, Wolfgang Bangerth wrote:
>
>
> Thomas,
>
> > I was able to solve for the vector mean curvature using the weak form of
> > the identity k_bar = laplacian_X id_X, where X is a codimension 1
> > manifold without boundary. The image above is a plo
Nice! Good luck with Gmsh!
On Tuesday, August 23, 2016 at 11:10:46 AM UTC-4, krei wrote:
>
> Thanks for the response. I have a more-or-less working version where the
> whole thing is solved together by Newton's method, but I bet solving
> electric field separately gives a good performance boost
So it looks like the problem is with
commit d1e3ffcaf6733e7d8509d93785b1f9e2b16a977e
Author: Luca Heltai
Date: Tue Jul 12 12:39:30 2016 +0200
Added get_new_point with two points and a weight.
I will try to send my code which is affected by the "bug", but it may not
be very simple.
Best
p
Thomas,
I was able to solve for the vector mean curvature using the weak form of
the identity k_bar = laplacian_X id_X, where X is a codimension 1
manifold without boundary. The image above is a plot of the square of
the mean curvature on an ellipsoid with semi principle axes a=1,b=2,c=3.
Gr
Thanks, that looks like it works. I don't have any problems left with
getting the output to work now, but for my own clarification, I'd like
to have a better understanding of how vectors work and what compress
actually does. I get the impression that modifying vector
elements results in some loc
Thanks for the response. I have a more-or-less working version where the
whole thing is solved together by Newton's method, but I bet solving
electric field separately gives a good performance boost. I will try to
fiddle with it, however, generating multiple meshes with matching boundary
for te
Daniel,
Thanks, that looks like it works. I don't have any problems left with
getting the output to work now, but for my own clarification, I'd like to
have a better understanding of how vectors work and what compress actually
does. I get the impression that modifying vector elements results i
Hey,
I had a similar problem: PDES in separate domains that are coupled through
an interface as a boundary condition. You can go about it using one
triangulation; I attempted to do this at first, but ended up using multiple
meshes. The fact you have matching meshes on the boundary is good. Wh
Dear Daniel and others,
Below is the part of my code that might explain better:
*//Below is the constructor for FEM class (a part of my own library).
The part of the code below in BOLD states that I assign different order
shape functions alternatively.*
template
FEM::FEM(
Triangulatio
12 matches
Mail list logo
