Rigid body constraints in SAP2000

[bigrod223399]

Hi,

I am working on a problem where I am using rigid body constraints on shell element nodes in SAP2000 (to simulate rigidity without increasing stiffness modifiers and diverging the analysis). However, when I run the analysis and look at the deformed shape, it seems the shell is only translating when it really should be translating and rotating to find equilibrium. My question is two-fold:

1) Constraints simply allow for rigid body motion of a particular set of nodes, why would it be restraining against rigid body rotation?
2) If this is not the case, how can I get this to work?

 

[rb1957]

I don't know SAP but I do know what I mean by "rigid body constraints".

It sounds to me as though you have a model maybe with balanced applied loads and reactions and so you need to constrain rigid body motion of the model. This means to constrain the 6 dof which you can do by ...
1) constraining 1 node in all 6 dof,
2) constraint three nodes (the 3-2-1 dof constraint ... yes?)

another day in paradise, or is paradise one day closer ?

 

[bigrod223399]

I don't know SAP but I do know what I mean by "rigid body constraints".

It sounds to me as though you have a model maybe with balanced applied loads and reactions and so you need to constrain rigid body motion of the model. This means to constrain the 6 dof which you can do by ...
1) constraining 1 node in all 6 dof,
2) constraint three nodes (the 3-2-1 dof constraint ... yes?)

Perhaps I didn't explain the problem well enough. I am looking to find the equilibrium position of the suspended shell elements, but when I run the nonlinear analysis, the elements are translating when I would expect them to translate and also rotate as well to find equilibrium. I'm wondering now how I can model this rigid body motion behavior correctly because when I increase the stiffness modifiers of my shells, the analysis blows up, so I figured constraining my nodes may be the only way.

 

[rb1957]

ok, you have a model which is supported on "cables", a set of generalised directions which balance the applied loads.
So you have a starting point for these cables, and the applied loads should move the structure so that the cables react the loads.
I think you need to run "large displacement".
I'd run some test cases, with loads in different directions that should push the model differently.

another day in paradise, or is paradise one day closer ?

 

[bigrod223399]

ok, you have a model which is supported on "cables", a set of generalised directions which balance the applied loads.
So you have a starting point for these cables, and the applied loads should move the structure so that the cables react the loads.
I think you need to run "large displacement".
I'd run some test cases, with loads in different directions that should push the model differently.

I applied an SDL load on the corner of a shell element that should, in theory, rotate the piece. It still only translated when it should be rotating + translating, which is the same problem I've been describing. I am beginning to think nodal constraints are not the way to go about this.

 

[ProgrammingPE]

It should be possible to get the nodes to rotate together. Is it possible that you are using the wrong constraint type? This example shows that an "equal constraint" can translate, but not rotate, while a "body constraint" does both.

Structural Engineering Software:

http://www.structuralcentral.com

Structural Engineering Videos:

http://www.youtube.com/channel/UCOj2mXXf3ZbZtgxTc6BtkGg

 

[bigrod223399]

It should be possible to get the nodes to rotate together. Is it possible that you are using the wrong constraint type? This example shows that an "equal constraint" can translate, but not rotate, while a "body constraint" does both.

I am indeed using rigid body constraints on the nodes, which is what I'm finding weird about this whole thing. The way I thought those worked was that it simply grouped the nodes and allows them to move as "one", but it seems to lock/couple certain DOFs together. The shells are sort of behaving like frame elements, which perhaps is exactly what is happening here. Maybe that is the cost of using constraints (i.e. losing certain DOFs).