Thermal load and composite deck issue

[Anonymous]

I recently modeled a steel structure using RSA and have noticed that the axial thermal forces in the beams are excessively high where a composite deck is present.

The structure I modeled is split into two, a portion of it consists of a series of portal frames while the rest of it has an intermediate mezzanine level with a composite deck slab as shown in the figure below (the region with the large number of nodes is where the slab is located- it's just been hidden in the display for visual ease).

The composite deck has been modeled based on the suggestions in the following link:

https://forums.autodesk.com/t5/robot-structural-analysis-forum/composite-action/td-p/3381625

After applying a thermal load on the beams of +37 degrees Celsius and running the model, it produced very high axial forces in the beams to the magnitude of around 11000 kN. Please note that the beams in that area all had pinned-pinned releases.

After removing the composite deck and studying the effect of the thermal load on the beams, the model gave the much more realistic results below:

Even after attempting to apply the same thermal load to the composite deck panels (assuming that might be the cause), the same high values were given.

I can only assume that the cause of these high results is the fact that the nodes of the composite deck restrain the beams from movement in the axial direction, causing high internal forces. This is an unrealistic representation of the diaphragm action of the composite deck.

If that is the case, then I urge the developers to take it into account in the future Robot updates.

If not, then I would appreciate some enlightenment.

Thank you.

[Artur.Kosakowski · Accepted solution]

Hi @Anonymous

As rigid link are intended to keep the same distances among the nodes they connect they by definition do not allow for the beam elongation and the releases at the beam end have no influence on this situation at all. As the beam cannot 'expand' the temperature causes large axial force which doesn't exist in reality. 

The solution can be having phased model with and without the diaphragms so that you could make the superposition of the results for all but the temperature load for the one with diaphragms and the temperature load from the one without them.

If I managed to answer your question(s) press the Accept as Solution button please. This will help other users to find solution(s) much faster. Thank you.

[Anonymous]

Hi Artur,

Thanks for your response, however i believe that it should be addressed in future RSA updates in order to avoid the phasing process.

[Anonymous]

Hello Artur,

i have related issue regarding thermal loads and axial forces in slab surrounding beams. Even if i choose in "rigid connection of nodes for the slab" without stiffening, i still receive high axial forces in surrounding beams (the same temperature gradient is assigned for beams and slab). 

However results changed to acceptable, when i switched "transfer of loads" to Analytical.

Why it is related with transfer of loads?

I'm modeling one way deck slab and i would like to evaluate thermal loads, but i still didn't find a way to do it properly. Is there any way i could do it?

[Artur.Kosakowski]

Hi @Anonymous 

You cannot use the temperature load on panels along with the simplified load distribution as it is disregarded. In other words for the upper model you have the temperature load applied to beams only with slabs acting as restrains for their elongation.

If I managed to answer your question(s) press the Accept as Solution button please. This will help other users to find solution(s) much faster. Thank you.