Lateral Torsional Buckling: bracing a steel W40x211 to a wood-framed roof. 1

[mnchamber]

The existing building is a 1979 tilt-up concrete warehouse will a wood-framed roof. The roof consists of 6-3/4" glulam beams at 40 feet spacing with TJL and TJ-50's between. The glulams vary in depth from 24" to 39". The maximum span of the glulams is 48 ft. My client wants to remove a post: this will create an 80 ft span. For the 80 ft span, the new beam is a W40x211 steel beam. My client wants to trim the glulam to leave the top 6" remaining, and then attach the new steel beam to the bottom of that remaining 6" of glulam beam....like a nailer.

My concern is lateral torsion buckling. I can connect the steel beam to the 6" of remaining glulam using vertical steel plates on top of the top flange and through-bolts. However, I am concerned that the connection of the glulam to the TJL and TJ-50 joists will not be strong enough to prevent buckling of the beam. I have reviewed Appendix 6 of the AISC 360 for Lateral Torsion Buckling: it only covers "Nodal Bracing" and "Relative Bracing" for LTB, but not "continuous bracing". I have also reviewed Yura's "Fundamentals of Beam Bracing" and cannot find an answer.

I would like to be able to calculate a force for the connection of the glulam to the TJL / roof sheathing. Please let me know if you have an equation or reference.

Thank you!

Here is a sketch of my proposed connection of the steel beam to the 'nailer'....but it is the connection of the existing TJL to the nailer that I am concerned about.

http://files.engineering.com/getfil...c901-407e-95bf-ba31d8ccacc8&file=03141601.PDF

 

[structSU10]

appendix 6 does cover continuous torsional bracing, 6.3.2b

This seems very difficult to build though. never mind the shoring of the roof, but cutting the glulam, and then installing the beam with that connection detail seems tough - there isn't much tolerance and you can bet the existing construction, plus the new beam, won't be perfectly straight.

Can you use 2 beams to support the trusses on each side of the glulam? you could attach the two together and perhaps increase buckling resistance.

 

[KootK]

Here's an alternative that would create a more robust LTB connection in my opinion. Kind depends on whether or not your worried about aesthetics I suppose.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[mnchamber]

@structSU10

Yes, Appendix 6 covers Continuous Torsional Bracing. However, it appears that that section only offers a recommended stiffness, rather than a force. Also, Torsional Bracing is different than Top Flange bracing.

 

[mnchamber]

@structSU10

We could add beams on each side of the steel beam to provide Nodal Bracing. I was just hoping that I would not have to go to that extreme. Also, as an engineer, I really feel like I should know how much force is needed for attached a beam's top flange to the diaphragm to prevent LTB.

 

[structSU10]

the torsional bracing is all you need though. if you read the opening paragraph of 6.3 "...lateral stability of beams shall be provided by lateral bracing, torsional bracing, or a combination of the two..." which I take to mean it can be either or - you are trying to prevent the relative displacement of the top and bottom flange and torsional bracing accomplishes that.

Also, the continuous torsional bracing section says you can use equations A-6-9 (for continuous bracing moment) so that is the force you need. the stiffness in that section is just for reference, to be used in lieu of the betasec in the nodal bracing section.

And the double beam approach I was getting at was to use a beam under each existing truss, and put lacing/bracing between the two members. A good general reference is the paper 'Fundamentals of beam bracing" by Joseph Yura, and for the two beam concept I am discussing, 'global lateral buckling of I-shaped girder system' also by Yura. The second article discusses the global buckling of a two beam system. The thing about that set up is the braced between each can act as a brace for each individual component, then you look at the two girder system for global stability, very similar to how a built up column is checked. It will be stiffer than a single beam, and you may get it to work.

 

[structSU10]

here is the second article attached

 

[KootK]

Lots of ground to cover here on the theoretical front. Consider all subsequent statements to be predicated with the "in my opinion" caveat. That said, I'm fairly confident in what follows.

1) The continuous bracing mentioned in the appendix is a form of torsional bracing and is a different animal from the lateral bracing being contemplated here. It doesn't apply to this situation.

2) To my knowledge, closed form, practical solutions for continuous lateral bracing only exist for situations where the smeared bracing spring constant may be considered uniform along the length of the braced member. There aren't many practical applications for that and this is definitely not one of them as the spring constant would diminish towards the interior of the diaphragm. In conclusion, your only practical way forward is to treat each truss as an individual brace.

3) Lateral bracing supplied by subframing fastened to a diaphragm is not nodal bracing. Rather, it is relative bracing because relative lateral movement of the braced flange is prevented between discrete brace points.

4) Counterintuitively, your required brace force is the same irrespective of the number of braces provided. Three braces or seventy six, the answer is the same.

5) Relative bracing required strength is half of that required for nodal bracing which is nice.

6) Relative bracing required stiffness does decrease with tighter brace spacings. That's helpful with wood bracing systems where stiffness can be an issue owing to the large disparity in elastic modulli between wood and steel.

7) Determining the requiremeets for the bracing is the easy part. Working out the brace stiffness supplied by the system is the tricky bit. You've got axial flexibility in the trusses, in plane flexibility in the diaphragm, nail slip all over the place, and possibly even take up in sloppy connections.

8) The truth is that most engineers, when presented with this situation, will check for brace strength, provide some manner of positive attachment between the top flange and the trusses, and call it a day. The detailed diaphragm brace stiffness evaluation doesn't happen much out in the wild unless the aspect ratio of the diaphragm is small enough to raise eyebrows.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.