Lateral torsional buckling in steel beams 1

[rte4563]

I have a steel beam which fails in lateral torsional buckling at top and bottom flange when I run my calculations. However, the steel beam is kept in place at the top flange by a concrete deck on top of it - can I then look away from LTB for the top flange atleast? Any sources on this?

 

[kostast88]

Buckling is happening on the compression flange. A beam is unrestrained if the compression flange is free.

If it's connected to something (eg concrete deck) you have to explain how stiff the restraint is and recalculate. If you want to develop a solid practical restraint, you need to properly secure it to the concrete deck. This can happen with bolts/anchors or other connector means.

If it's concrete slab on metal deck, welded studs on the beam top flange is the obvious answer. If you pre-weld them, that's the case. If you shot-fire them though, I wouldn't say it's 100% restrained but rather a 50%. In any case, it's a bit open to interpretation.

At the supports you have options like e.g. add stiffeners to deal with bottom flange compression. Is it a continuous beam over the support, or a fixed-fixed beam?

 

[rte4563]

This is how the steel beam is; pinned at the two ends and in the middle. Then concrete slab (orange) at the top. The steel profile is a HEB beam.

Edit: what would be the compression flange in this case?

 

[phamENG]

The internal moment of a beam is comprised of compression and tension stresses. In a wide flange beam (which I believe is equivalent to an HEB), those stresses are predominantly in the flanges. So you end up with a tension flange and a compression flange. You should be able to determine which is which based on the moment at that location in the beam.

Checks for the beam occur all along the length, not just at one point. In many cases we can simplify our designs down to a simple span beam and we know, by experience, where the worst case or "controlling" condition will be and we can just check that one. But in continuous beams or more complex setups, we sometimes have to check the capacity of the section at several points along the beam and compare it to the moments, shears, and axial forces that occur there.

 

[DaveAtkins]

The unbraced length for a continuous beam such as you illustrate is equal to the length of one span. Decades ago, the inflection points could be considered as braces, but AISC did away with that. Now you assume the unbraced length is the entire span length, and include the Cb factor in your calculations.

DaveAtkins

 

[KootK]

I would strongly encourage you to seek advice from someone local-ish (Norway/Europe it appears). One of the interesting things that I've learned here on Eng-tips is that lateral torsional buckling design practice can be quite different in different parts of the world.

In North America, this would be my answer:

1) Make sure that the slab and it's attachments to the top flange do, in fact, provide competent lateral support to the top flange (strength and stiffness).

2) Make sure that you have beam torsional support at all of your supports. This often takes the form of lateral restraint provided to both flanges at the supports.

3) With #1 and #2 in play, I think that your situation is this:

a) You've eliminated the possibility of top flange buckling anywhere.

b) You have two spans to check for bottom flange buckling that you can check independently.

I don't personally favor describing LTB as "top flange buckling" or "bottom flange buckling" but, at the same time, I got your gist so I just rolled with that.

 

[KootK]

Edit: what would be the compression flange in this case?

With respect to lateral torsional buckling, your critical "compression flange" will be the bottom flange. That, because:

1) It sounds as though the top flange will be well braced so we're not concerned about it buckling even where it will experience compression, such as over the supports.

2) Wherever your bottom flange has compression anywhere between points of lateral support, it will become a "compression flange" for checking lateral torsional buckling.