The loads are very large when rigid supports are assumed -- but this is often misleading. The magnitude of movement (thermal expansion/contraction) which we are translating to a force is often very small. So even normal bolted connections, flexible columns and foundation elements, etc will allow a lot of structures to adjust for climatic thermal loading, without specific detailing.

The usual rule of thumb for building structures in temperate climates is that thermal effects are often negligible for structures up to 300' in a given direction.

for heated building expansion joints... it depends. For exterior shelf angles supporting masonry, exposed to the cold I like to limit the lengths to about 20' (6m)... Our temperature range is from +30C to -40C.

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

Thanks for responding.
It is difficult to assess.
I haven't come across anything practical, in the form of a design guide.

Interesting to hear about the rule of thumb for 300' in a given direction, this is about 100m, which is quite a large dimension.

A lot of it depends on your configuration. If you have a series of portal frames or a long set of frames with braced frames at each end, you can develop some significant axial forces in your beams (that then get transferred to the frames resisting it). If you have a pair of braced frames in the middle and several 'leaning' columns on either side, you'll have minimal buildup of axial stresses in the beams as long as your columns are sufficiently flexible and can handle the displacement.

I think there was a recommendation/article by the AISC several years back that for steel buildings 300' was about max... I don't recall; it tucked back in the gray matter somewhere. This is predicated on the structure being inclosed within a heated building's envelope so there is not much thermal movement. With exposed structure, you have to look at thermal expansion, remembering that roof surfaces can be much hotter than ambient.

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

This was a building I did a while back. You can see the placement of the expansion joints on the roof.

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

Thanks again.

phamENG, as described it makes sense. But similar to your case with a single braced frame in the middle, for a single K-braced frame I recall getting a high reaction in each support, equal in magnitude with opposite directions. If released on one side, you lose the intended function of the structure.

dik, I'll search online based on this then.
Edit: Just saw the photo, interesting to see joint in building.

There's four or them... isolating the corner (a discontinuity) and one in each other part because of the length.

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

mte12 - based on my aerospace experience, thermal loads predicted by FEMs are often unrealistically high. This is typically due to various actual structural flexibilities (joints, etc) not being adequately modelled in the FEM. Rigid constraints just make it worse.

The forces that thermal expansion of large structural members can exert is huge. It is typically assumed that thermal expansion cannot be restrained or limited, and it must be accommodated in the design of the structure.

Rod Smith, P.E., The artist formerly known as HotRod10

Thanks for the additional comments.