-
Notifications
You must be signed in to change notification settings - Fork 4
Mismatch between docu and implementation for bc in plate with a hole #68
Description
When sending out the problem to others we realized that there is a discrepancy in our model implementation of the plate with a hole compared to the documentation. In the documentation, both in [MathModDB ](https://portal.mardi4nfdi.de/wiki/Plate_with_a_hole_(linear_elasticity_model)_(linearly_elastic_model_of_a_concrete_plate_with_a_hole) and in our own docu we apply Neumann boundary conditions on the top and right boundary based on the analytical solution. However, in the implementation in all tools, we actually apply Dirichlet boundary conditions. That does not make a difference in terms of the maximum stresses that are computed, but it should be synchronized.
We have now three options:
- We apply Dirichlet boundary conditions (as we currently have, just change the documentation). Then a reasonable other metric (other than max Mises stress) should be the reaction forces at both boundaries.
- We apply Neumann boundaries at the top and right boundary based on the analytical solution, then the displacement of the top right node might be a good additional metric
- We do a mixture, so e.g. apply Dirichlet bc at that top and Neumann bc at the right boundary, and then compute the reaction force at the top and the horizontal displacement of the top right and bottom right nodes.
Technically, the third option is the most complex one, but it is maybe the most general test - any preferences from your side @jpthiele @srosenbu @div-tyg ?