Learning Forum

[luis.velasquez]

Hello,

I have a few general questions:

1. Remote Mass: are they fixed in space or can they move? If I define a remote mass to a geometry and then I use a intertial load (say gravity) that said PM will be there no matter what, and the moment/force transmitions comes from the equivalent F=m*a/M= F x r?
2. Shell Model: If I have two perpendicular thin geometries, meshed with shell elements and connected via mesh, does a stress concentrations arises from this set-up? Seen as solid geometry it´s a sharp corner with little contact area.

Thank you.

[peteroznewman]

Remote Mass is scoped to geometry that is flexible such as a face on a solid, or an edge on a surface body or a vertex on a line body.  Gravity or other acceleration loads pulls on the mass and the force is transferred into the geometry as you would expect creating shear and moment forces on the face, edge or vertex.

Two perpendicular thin bodies such as an L profile is meshed with shell elements has a row of nodes along the edge where the bodies meet.  Since each body has a shell thickness, the stress is spread over an area equal to the length of the edge times the thickness.  This avoids the stress singularity that a solid model of the same L profile meshed with solid elements creates because as the solid element size is reduced the stress keeps increasing.

[luis.velasquez]

Hello Peter,

Thank for the information.

Regarding the surface bodies, I understand what you said about the stress being spread over the length and thickness. Does the ratio of thickness to mesh size matters? Like, say you have a T joint, if the thickness of the central element is smaller than the mesh size of the horizontal element (6 times smaller) and using linear elements, the thickness won´t "reach" to the perpendicular face nodes and therefore does not spread (since stress is calculated at the integration points, derived from the node information).

 

[peteroznewman]

For two surface bodies that meet at a T joint, use Shared Topology so there is a row of shared nodes connecting the two bodies.

When you process stress results, plot Unaveraged Element Stress so that stress at the T has values from the thick plate separate from the thin plate. You don't want those stresses Averaged across the two bodies.