Pinned support

roberto_p

Hi all,

attached a very simple model with two configurations concerning supported nodes.
In the first config (pinned) the constraints are applied to a selection of 4 nodes, in this way the supported node is the only one in the gc of the 4 nodes.

Victor, it is possible to have an option to force the pinned constraint remaining on every single node separately like in the config "node displ"?

Many thanks
Roberto

Hengre

I am also a bit wondering about Roberto's example.
I would have expected the same behaviour for "pinned" and "node-displ" case, but pinned faild. I tried to make a calculation with each of the 4 Nodes pinned and got a different result compared to "node-displ"...

How works pinned support applied to a named selection

Regards
Henning

JohnM

The pinned support holds the beam at 1 location in X,Y,Z. The beam can still rotate in RX,RY,RZ direction. The force that is applied puts a moment on the beam, rotating about the pin joint which has no torsional stiffness, and it blows up.

The node_displ case has 6dof held, it is statically determinate.

To make these two behave the same, I did the following:

Change pinned support to node-surface coupling. This uses a master node that I held in X,Y,Z,RX,RY,RZ.

Hold all DOF in node_displ case. Since the dependent nodes in the pinned support case create a rigid region, the node_displ case was modified to hold all six nodes in XYZ. If you keep it hold the way it was, you get subtle effects that make big difference in the results.

Victor

This feature often causes confusion and I'm not sure what to do about it.

The pinned support is distributed over the whole selection using constraint equations and does what JohnM said. It only reduces to ux=uy=uz=0 for a single node since that has the same meaning.

The constraints in "node displ" look too complicated to have a single object for. But you could at least make it from only 3 displacement constraints (one in each direction), instead of 6.

Perhaps an option to apply it to each node separately would make it clear and allow what people often expect it to do. Though then it becomes redundant with 3 displacement constraints.

roberto_p

Hi Victor,
many thanks for your reply

even if I understand your position, I personally prefer to have a switch option that leave me the possibility to apply constraints to each node separately or in one node like as the actual situation.

Bye
Roberto