Closed Kevin-Mattheus-Moerman closed 8 months ago
Welcome to the community Kevin!
The "column" in u represets the x,y,z displacements, so in this point you are right. And indeed, the numbering of the dofs which you get out of the dof handler does not match the numbering of the nodes. Also note that visualization with this assumption does not work for sub- and superparametric elements.
What we do in FerriteViz is a bit less efficient in this regard. In a nutshell we generate a new mesh for the visualization, which only has the surface representation of the original problem and not all volume elements. For this surface representation we now evaluate the solution field for each node to "transfer" the solution between the original grid and the surface mesh.
Does this help?
Thanks for making this awesome looking library. It looks amazing.
To try out
Ferrite.jl
I started with thehyperelasticiy.jl
demo. I think it works fine, and if I add:I get:
However, I wanted to be stubborn first 😄 and attempt to cook up my own visualisation by using:
Now I get:
Since part of the model looks like, I think this is some kind of index issue, that the nodal indices for the points and the results (
u
) do not match.I think I might be wrong to assume
u
represents a column for the x, y, and z displacements, and that the order of these corresponds to the points (grid.nodes). Or is there some reordering happening due to the boundary conditions?Thanks for your help!