Hydrostatic pressure issue

Hi,

On shell elements, the pressure is supposed to follow the normal elements. In the current study

  • all normals are “inside” oriented (see gmsh screenshots)
  • the mesh has been exported from gmsh to inp format
  • if i want to inflate the part, the applied pressure is theoretically negative
  • it’s (more or less) the case for the cone, but not for the cylinder in which a positive pressure should be applied to have “outside” arrows (a bit disturbing)
  • furthermore, one can see in the cone screenshot that some arrows remain “inside” oriented, leading to inconsistent results

Note i may have made a specific “lighter” model and check if the same behaviour occurs :slight_smile:

I do not think having done something wrong

Paul

screenshots from gmsh:

screenshots from prepomax:

The pmx file:
test_normals.zip (5.8 MB)

The gmsh mesh:
cylinder_cone_gmsh.zip (3.1 MB)

Erratum:

  • I reopened th pmx file i gave and all normals seems correctly oriented unless we can see in the screenshot
  • i noticed the pmx file size grows if we rerun the FEA several times: maybe something went wrong at this stage (???)

Yeah, I opened your .pmx file and I don’t see any issue with normals. Btw. you can also visualize them in PrePoMax using Color Annotations:

Do you agree that the sign of the pressure should be the same for all parts to inflat it?

The thing is that it also matters on which side of the surface you click. If you choose the inner side of the cone’s surface, you won’t have to change the sign:

1 Like

ok i can confirm the orientation depends on which side you click on (“internal” surface or “external” one; it’s a bit confusing regarding the common definition of the shell (lower face / upper face … see offset for instance). But ok, the pressure is applied following the arrows orientation whatever are the normals.

PrePoMax knows how the faces are oriented, so the direction of the applied pressure depends on the surface side the user selects. This is a much more user friendly way of defining the pressure direction for new users. The user does not have to care about the face orientations only selects the surface the pressure acts upon. As for solids.

Can we close this bug report?

1 Like

Actually it’s not a bug, so yes (and thanks for the support)