Boundary Conditions in tubular space frame structure

[matcunni]

Hi,
I'm still pretty new to Abaqus. I'm having some trouble with the boundary conditions when modelling the following structure:

Model of an isolated node:

Complete structure:

I need to know the tensional state in the plate. I managed to get the model running and the analysis completed without errors, I've already defined materials, sections, contact interactions, loads, etc,, and now I'm wondering if I'm using the right boundary conditions.

I've defined rigid body constraints using "tie node", I've selected the shell's perimeter and a reference point in each tube end to then apply the boundary conditions to the reference points.

I'm not restricting rotations in the nodes. (Should I?)I've tried using simmetry boundary conditoins but I think I'm not getting them.

Anyone can help?

If the question isn't clear above, what I need are some thoughts about the right boundary conditions for this type of structure, to have only axial loads in each tube..

Thanks in advance!

 

[DrBwts]

Not knowing what the structure is or what sort of loading it would experience in reality makes giving advice about boundary conditions kinda problematic.

 

[matcunni]

My mistake..
I thought drawing the big picture would deviate the attention or disencourage possible answers ..
The structure will serve as a building's slab. It's something like this:

But instead of the trusses, we have something like this:

I don't know the english name of this type of structure. (Knowing that would be useful too) I've found some of these searching "Tubular space frame slab"
The loads are Selfweight, Live load due to occupancy and Wind. With those loads in the big model under SAP2000, we obtained the axial forces on the bars.
For the plate desing, I'm loading the bars of the isolated node with those axial forces obtained from SAP.

If I restrain, for example, the displacements dx,dy,dz in the extreme nodes of the oblique bars. (Yeah, pin supports) everything runs fine. But the reactions I obtain in those nodes that way aren't in the local x direction of the oblique bars. (The resultant is). I need those bars working JUST under axial loads..

If I restrain the local dx displacement of each oblique bar, and nothing else, the analysis won't run. If I add some restraints in the extreme nodes of the horizontal bars (in their axial direction), it runs, but then I'm out of nodes to apply the axial loads..

I understand that I need just axial loads and reactions in the model for it to work as an ideal truss.

Any thoughts are really welcomed..

Thanks for the interest