Simon,
So far I managed to postprocess derived quantities from the displacement
solution like displacement gradients to get strains and eventually stress
for computation of the mises stress. However, when I view the contour plot
of the von mises stress, the picture looks rather ugly (see attached
Mises_quadrature_points.png).Mises_quadrature_points.png with the
discontinuities between the cells clearly visible. If I understand it
correct, dealii gives me the gradient values of the solution at quadrature
points. When I compute the mises stress and pass it to a DataOut-object
it attaches these values to the vertices of each cell. What happens if
there are more quadrature points than vertices in a cell? How are they then
mapped to the vertices.
How do you create this output? Via a DataPostprocessor object? In that case,
the gradients are computed in the vertices of cells, not quadrature points.
To me, the plot does not actually look particularly badly. Your mesh is just
too coarse :)
To polish the contour plot I tried to mimic the procedure of Step-18, where
the stress is averaged in each cell. To this end I used the following
piece of code: [...]
To receive the following picture: Screenshot from 2023-03-15 00-45-46.png
Not much of an improvement. Is there a possibility to get a smoother
output? I thought of averaging the individual cell contributions connected
to a vertex.
You can do that, and I believe that there is a function somewhere in either
namespace DoFTools or VectorTools that averages cell-based quantities to
node-based quantities.
In reality, though, if you *really* want a smooth field, then you want to use
the Zienkiewicz-Zhu "recovery" procedure. It is a particular kind of averaging
that produces a stress (gradient) field that can be shown to be
one-order-more-accurate than one would naively expect. Of course, the
alternative is to just do what you are already doing but on a finer mesh :-)
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.
To view this discussion on the web visit
https://groups.google.com/d/msgid/dealii/40dd551e-1960-43e9-307c-16b9e6b25ddc%40colostate.edu.
