I recently send an inquiry about using temperature loading with stiffened plates and found that the stiffened plates were being restrained more than they should be. On the same project, I am having an issue with applying the beam section temperature to a beam which is supporting plate elements. The plates seem to be restraining the main beams, inducing a very large compression force into the beams. I tried to include element temperature on the plates to model the temperature on the top fibre of the beam. I have trying to model the Eurocode Approach 2, vertical temperature gradient model (see below).
The following shows the axial forces I am getting in the beams.
The reason for the large axial force here is that your Beam Section Temperature loads are not applied correctly. You have applied a single temperature layer with a width of 267mm, however, you need to discretise this into two layers to capture the change of the width of the section from the flange to the web:
When you do this, you will obtain a much smaller axial force in the beams, which is due to the temperature load in the plate elements. As the beams are under non-uniform temperature load, the top fibre of the beams is heated to 21C but it does not develop the full deformation of the 21C strain due to the bottom of the section not being under temperature load. This creates internally balanced stresses in the section which restrain the full deformation of the section, i.e. top flange does not experience free 21C deformation. But the plate is a single element and when heated wants to expand to the full 21C strain. Due to the interaction between the beam and plate elements, the difference between the full deformation of the plates and the partial one of the beams a balancing force and moment is generated in the elements which you can see in the model.