Forbidding bar from being split by panel

[pnacht]

I am modeling the following bridge:

Note how the crossbeams do not bind with the slab. In my model, however, the crossbeams are being split by the panels and therefore acting as if they were bound to the slab, which is not correct.

 

Is it possible to define a bar which cannot be split? This would mean the crossbeams would remain where they are, but the slab would act as if the crossbeams weren't even there.

[Romanich · Accepted solution]

If I understood correctly, try to use this option:

[pnacht]

That almost did it. With that option the model became this:

The outside bars now are not considered in the meshing and therefore share no nodes with the panels. I don't believe the middle bar was considered either, but the mesh just coincidently landed on top of the bar.

I also know the bar itself was indeed broken by these nodes because its shear load diagram is

Clearly load is being transmitted by the panel nodes.

This solution almost worked, but by a simple coincidence the panel created nodes right on the bar, which was then broken and "bound" to the panel. I know I could simply get the panel nodes and move them off the bar, but is there a more elegant solution?

[Rafacascudo · Accepted solution]

Pedro ,

1 - Move the nodes is the fastest option

2- Looking at the picture, it  seems that the outer girders do share nodes with the mesh

3- Try another type of mesh for the central panel or maybe Create some nodes near the central beam ,Meshing process will use this nodes and the path will be clear for the central beam.

4- You can always locate the transversal beams just a litte(like 2cm) below the main girders and link them with rigid links.

[pnacht]

Thanks Rafa. The external girders are supposed to be bound to the mesh, its just the midspan ones that need to be separate.

And I'd thought of rigid links, but wouldn't they create binary forces due to the excentricity?

Regardless, my question has been answered.

[Rafacascudo]

Are the beams/Girders CGs really at the same height??

Are you modeling the main girders with offsets??

[pnacht]

No, but if I created the girders 2cm below the main girders and bound them with rigid links I would be, no?

Sure, the offset would be tiny and therefore so would the binary, but still.

[t.sautierr]

Search also for "T-beam" and "Offset beam" topics in the forum because these discussions is what you are looking for actually (disconnecting a beam from a pannel).

TS 

[Rafal.Gaweda]

To check whether bars are connected or not connected to mesh run Model Generation and then turn on display of "Numbers of calculation elements"