Moment forces in bar element vs shell element of bar

[trygve_kristoffer_lovli]

I am trying to understand the moment forces I get in a beam, depending on modelling type. I got a two-span one-way slab, with pinned supports. In the midspan, there is a crossing beam, supported on each side, like this: 

I have tried to model this with the shell element, as presented above, and with a beam element. 

Moment in bar-element:

Moment in shell-element reflecting the bar element(-1786.92*0.2m=357kNm), where 0.2m is the width of the bar.

When I model the bar like this:

I get 357 kNm in the bar - i.e. same as in the shell model. The red lines are rigid links, connecting the shell element nodes to the beam nodes. And the length of these represents the beam width. 

When I try to increase the stiffness of the slab, i.e. i went from 30mm steel to 60mm steel, I get this:

1691.9*0.2=338.24. 

So the tendency is, when the stiffness of the surrounding slab increases, the importance of the modelling type decreases. Can you explain this, and the importance of modelling a beam vs a shell? At this point, modelling a beam element with surrounding slab seems unconservative(for the beam), as the moments is smaller. 

Unfortunately, the file crashed at the end, but I can remake it, if this is necessary to respond.  

[Romanich · Accepted solution]

Hi @trygve_kristoffer_lovli ,

It seems that the results are logical. Please note that the part of the slab is also considered in the first case (when you have a beam).  Try to modify Iy of the beam and check the results. You can use this tool https://forums.autodesk.com/t5/robot-structural-analysis-forum/api-macro-creation-of-quasi-t-amp-l-r...

[trygve_kristoffer_lovli]

Thank you for your response!

Do you mean the extra stiffness we get from the slab because we do not "take up space" using a beam element? I have thought about that - this is why I thought that the error would be greater when I applied a higher stiffness in the slab - because the extra slab that do not exist in reality, give extra stiffness, resulting in a lower moment transfer by the beam itself. However, the result was opposite of this, as you can see in the last picture. Do you have any comments on this?

Nice tool btw!

[trygve_kristoffer_lovli · Accepted solution]

I tried again, with a longer span. Here are the results with a 50mm steelplate as the slab, and a 200mm steelplate:

50mm:

200mm:

Now the solution makes sense. I might have done something wrong in the previous attempt. Here I can see the extra stiffness contribution from the slab-beam overlap, and also that this stiffness contribution increases compared to the beam with rigid links (in the middle, links are not displayed) when we increase the slab stiffness.

(Note that the beam width was 0.8m, so you can multiply the moment by 0.8 to reflect the beam moment)