Bracing against LTB

[mastruc]

Hello, all:

I had a question about restraining flexural members against lateral-torsional buckling. I suspect that there's some misunderstanding on my end about beam theory that I'm not properly thinking through. When beams in frames are braced laterally to beams in an adjacent parallel frame, how does that result in a brace point for that beam? If one beam in a set of frames is loaded to the buckling point all on its own, I can understand how an adjacent unbuckled beam would provide lateral restraint. But if a floor or roof is loaded uniformly, wouldn't all the beams tend to buckle simultaneously? In that case, how is restraint against buckling applied? This is especially bothersome to me with stair stringers (where it's common to treat the unbraced length as the distance between treads), since the treads aren't even connected to the compression flange of the stringers. Can anyone help shed some light on what I'm not understanding about situations like this?

 

[JAE]

LTB bracing is primarily focused on resisting torsional twist of the member.

Joe Yura at UT Austin (Texas) has done a lot of research into this and much of his work is included in the latest AISC Specification (Appendix 6).

He has indicated that you can brace a beam against LTB either by resisting twist or by resisting lateral translation of the compression flange.

There are approaches for "relative" and "nodal" bracing in Appendix 6 where nodal braces provide rigid, non-moving restraints against LTB and the relative brace is more like what you describe in your parallel frame description.

For stair stringers, the treads are actually at an angle to the "vertical" webs such that rotation of the member is prohibited, or highly resisted. Therefore they act as a brace - almost like solid bridging between the two stringers.

 

[Hokie93]

The reference to the work of Joseph Yura is spot on. I highly recommend his paper "Fundamentals of Beam Bracing" from the AISC Engineering Journal, 1st quarter, 2001. The paper provides theoretical background as well as practical guidance for lateral and torsional beam bracing. The paper also includes several design examples. If you are a member of AISC (the American Institute of Steel Construction), the paper is available free of charge from the AISC website (

http://www.aisc.org).

If you are not a member, there is a nominal fee for Engineering Journal papers.

 

[mastruc]

Understood. Thanks for pointing me in the right direction! I had already read through Appendix 6, but it was rather brief and I was afraid there was something conceptual that I was missing. I'll snag the relevant paper from AISC -- much obliged!

 

[Ron]

The paper hokie66 references gives a good description of the differences between lateral bracing of beams and torsional bracing. Using your stair stringer example, the treads offer reasonable lateral bracing but offer little in the way of torsional bracing. Since you cannot reasonably brace the top flange of the stringer, you have to use the tension flange for torsional resistance or provide stiffeners between flanges (not always effective either) at appropriate intervals. The tension flange approach works but you have to consider that the lateral bracing actually helps the tension flange bracing to prevent torsion.

In stair sections that I have designed, I used cross bracing of the bottom flange to prevent torsion.

 

[hokie66]

Errata for Ron...it was Hokie93 who gave the good advice, not me. I know...it is hard to keep all of us Hokies straight.

 

[Ron]

Oops! Sorry to you both! Was watching the FSU-Miami game. At least I had the conference right!

 

[slickdeals]

I guess the original question on how the stair treads torsionally brace the stringers remain unanswered. Isn't it like two drunks leaning against each other? The top flanges could translate together and only if the treads/risers are connected to form a "diaphragm" could it be assumed to be laterally braced adequately.

 

[hokie66]

slickdeals,
The treads (and risers if present) do form a diaphragm of sorts. It is a Vierendeel truss. As long as the treads are adequately connected to the stringers, I don't think you need to worry about buckling or sway.

 

[slickdeals]

Right, of course. It's a Vierendeel.