Open YohannEude opened 5 years ago
Without the setup data, I cannot ensure this is not a bug, but is seems to be normal behavior for the current algorithm.
The "step" would be desired if the red area were for example the wall of a pie and the grey area the inlets/outlet. Zones outside the red (extruded) area should automatically be sliding sections (to avoid issue #8), which might also help here
To avoid this step, blocking the boundary layer in the gray zone parallel to the red zone (using a boundary layer insertion with zero thicknes) should work around this.
For better automation, some test based on vertex normals at insertion / no insertion boundaries (i.e. setting zero thickness when the vertex normal is too closely aligned with the non-extruded zone) might be feasible.
Your are right, the grey area near the red one is an inlet/outlet. I don't want to put the thickness to 0 on the red one because I have some pipe with an extrusion layer sliding on the inlet/outlet. For this case, maybe I would add first the extrusion layer around the pipes and then fix the thickness of the grey area (inlet/outlet) to 0 and add the extrusion layer on the red part.
In two steps, it works! Thx
Ok, automating this is tricky (it is easy to determine cases when the surfaces are on the same plane, as they are here, or orthogonal, such as in a duct inlet, but could be fragile for some geometries).
The chosen solution would be to add documentation on this feature explaining the usage an pitfalls, with a few screenshots similar to yours (possibly in a simpler, 2d illustration).
So I am not closing the issue yet, but moving it to "documentation" category. I'll close it when the documentation is up to the required level.
Here a configuration which doesn't work with CS 5.3.1. There are two images attached to this mail:
Thanks for your work,
Yohann