Question about Restraints

[vincenttt]

Hi,
I am just wondering if the restraint in 'restraints tab' in Seismostruct is about local coordination or global coordination?

[seismosoft]

The restraints are about the global coordinates system, which employs the X-Y-Z axes notation. The elements' local system that employ the 1-2-3 notation.

SeismoSoft Support

[vincenttt]

thank you for your reply.

If so, how could I define the boundary condition of a beam if it is not aligned to x or y axis.

Let's say the Beam AB, point A is at (0,0,0) and point B is at (1,1,0). now I want to define point A as a pinned support and releasing x,y,rx,rz.

[huffte]

Hi vincenttt.

To apply the pinned boundary condition to the beam you would simply specify M2 release at the start node ("A") when you define the element connectivity for the element running from "A" to "B". Note the importance of order here - Node "A" should be node "1" in the element connectivity specification. Otherwise, if you define node "B" as node "1", then you would need to apply the release at the second node instead of the first.

If node "A" is, in fact, a boundary, then it would receive a "restrain all" specification in the restraints dialogue.

Best of luck.

[vincenttt]

Thank you huffte.
So if input time-history in this case, i will 'restraint all' for Node A (node 1) and B (node 2)since they are the boundary of the beam. and to define M2 release at B, i can release the M2(B) in the element connectivity. and i will input the time-histories at node A and B. Is that correct?

[huffte]

Yes, vincenttt. If Nodes A and B are truly "attached" to the ground, you could restrain all DOF's (it is up to you to decide if this is appropriate for the structure you wish to model). Then you would define a moment release for the appropriate node and the appropriate local DOF. The time history would be applied to all nodes in the model which are "attached" to the ground.

[vincenttt]

Then if point A (0,0,0) point B (1,1,1). Now point A is on the ground and but point B is connecting to something else that i am not going to model. So I still restrain all DOF's, but don't applied the time history to B, is that right?

[huffte]

I think it is impossible for me to answer your question vincenttt.

Do you choose to assume that whatever B is attached to, it is so stiff that it will move along with the ground? If so, then it would seem reasonable to apply the accelerogram to B as well.

If, on the other hand, B is attached to framing which is flexible, I am unsure as to what you will actually accomplish by neglecting the structure to which B is connected.

So it would seem that your most recent question is one of modeling strategy and theory.

So you must ask yourself "Which points to which my structure is attached will move with the ground during the earthquake?" The answer to this will determine your support points and load points.

Best of luck vincenttt.