LTB Bracing Using Web Stiffeners?

[kpkoma]

I am designing a building that needs to meet the requirements of UFC 4-023-03 "Design of Buildings to Resist Progressive Collapse".

One of the provisions in this manual requires that the floor framing be designed for a net uplift load equal to the self wt. DL + 0.5LL.

This results in huge beam sizes due to the unbraced length of the bottom flange, unless some type of bottom flange bracing is used.

I really don't want to require the installation of a gazillion angle braces. This would be labor-intensive as well as a major headache for the HVAC boys.

The floors are metal deck and concrete on steel beams with studs for composite action. I figure that if I use full-height web stiffener periodically along the length of the beams, I can assume these as brace points for LTB.

My theory is that the top flange is well-restrained against torsion due to the studs extending into the slab, and I can provide lateral support to the bottom flange by "cantilevering" off the top flange with stiffener plates.

Anybody have any comments or words of wisdom regarding this issue?

Thanks in advance.

Ken Kilzer, P.E.

 

[JAE]

Ken - just throwing some initial thoughts your way - I would say your concept is appropriate...I just know that for bracing against LTB, you need sufficient strength and sufficient stiffness.

The strength is probably OK but you should check the added tension in the studs due to a lateral load (this is provided in the Third Edition AISC LRFD Manual for nodal bracing).

As the beam tries to bend sideways (bottom flange kicking out) the top flange will be in bending unless the stiffener and the stud are in close proximity - which I would think could be specified on your plans. You might consider first calculating the studs required for normal composite action, and then ADDING more studs - in line with the sitffener plates, to get the desired behavior.

As for stiffness - I think that, again, using the AISC LRFD provisions - Section C3.4a - would give you what you need. You might want to model this in a detailed way to determine a "typical" condition and ensure that you have adequate strength and stiffness for the general system.

 

[SlideRuleEra]

Sounds like a reasonable approach to me. I see the requirement in paragraph 2-2.2 of the document. Looks like you have other factors that are working for you:

1. Load applied to one bay at a time (depending on floor continuity, could the dead load from adjacent bays help to "hold things down"?)

2. Allowable strength reduction & over-strength factors.

3. Steel connections are probably much "better" than simple supports. If you have not already done so, perhaps you can assume that the beams have partially or totally fixed ends, that would help.

I would put in far more that the minimum number of stiffeners that you calculate, since the assumption that they are "true" braced points is probably only partly correct.

http://www.SlideRuleEra.net

 

[haynewp]

There is an interesting statement "A load path from the slab to the foundation for this upward load does not need to be defined" in 2-2.2

In the OK City bombing I thought the biggest reason for the collapse was the perimeter beam rolled off of the column when the slab was lifted in the blast. Is this type of failure considered taken care of in the tie requirements?

I haven't used this document before, I know the TI-809 makes me pull my hair out for essential buildings in high seismic areas.

 

[UcfSE]

You will need to check the web distorsional stiffness in the AISC LRFD 3rd edition to use the concept you propose. It definitely can be done.

 

[kpkoma]

Thank you all for the valuable input.

I am going to add studs above that required for composite, specify their location adjacent to the stiffener, and do a detailed analysis of the required load and stiffness.

Haynewp, my project is classified as IIIE, so I am slogging my way through the TI 809. Not a lot of fun. Interestingly, I am in SDC "C", and so hoped to be able to use an R-value of 3 and dispense with alot of the required seismic detailing. But since I have a IIIE building, all that detailing is basically required, so I might as well detail the thing for seismic and take advantage of the higher R-value.

Thank you all again,

Ken Kilzer, P.E.

 

[miecz]

Don't weld web stifferners to a bottom flange. This is your tension flange under gravity loads. Welding to a tension flange increases the susceptibility to fatigue or brittle fracture.

 

[JAE]

jmiec - fatigue is not an issue in buildings.

 

[miecz]

JAE - I guess that brittle fracture isn't a design issue either. Truth is, I lifted the reasoning out of a Salmon & Johnson textbook. Many years ago, when I was right out of school, I was told by my supervisor that welding across a tension flange was a bad practice. The rule stuck, but it was so long ago, I've forgotton his rational. So, I would check with a welding expert.

 

[JAE]

Well, welding across a tension flange, say across the bottom, when its under load, is truly a bad idea in that the weld heat could initiate a semi-molten/brittle fracture during the weld. But a vertical stiffner, such as that discussed in this thread, poses no such risk as the weld is up the web, and across the top half of one flange only.

 

[miecz]

Though we never said so, I was assuming we were considering stiffeners on both sides of the web. If stiffeners are on one side only, then the stiffness analysis will have to consider flange bending, even if the stud and stiffener are in close proximity.

 

[SacreBleu]

I did my Master's degree thesis on progressive collapse in 1975. I haven't had occasion to design anything using the concept since. Admittedly, not being aware of the Design Code Criteria you are using, I am no expert. But in my opinion, using stiffeners as a justification to eliminate angled bracing is stepping out on very thin ice. Is there any way you can stitch weld continuous angles to the bottom flange to increase the beam's lateral-torsional stability?

 

[SlideRuleEra]

Would a composite section (say a W shape & channel - see page 1-84 of AISC, 9th Ed., ASD) that was installed "upside-down" provide the necessary properties?

The composite steel/concrete provides plenty of bracing for the top flange (everyday loading). The channel would provide increased bottom flange bracing for the uplift situation. As a bonus, the channel would increase I and S for the entire beam, perhaps allowing the W section to be a little lighter.

Just thinking "out loud".

http://www.SlideRuleEra.net

 

[VoyageofDiscovery]

Hi kpkoma,

If you're concerned about welding to the bottom flange then bolt to the bottom flange with a clip angle, accounting for the loss in section. Just something I have done in the past for certain bridges.

Regards

VOD

 

[VoyageofDiscovery]

Hi SacreBleu,

If the design incorporates the stiffness of the top flange to concrete connection, his idea may well work. Perhaps testing is a good idea to gain faith in it. In railway bridges we use a similar concept, an old concept, called u-frames which provide LTB and torsional stiffness for the girders when the train travels longitudinally between two girders and thus no bracing can be placed.

Regards

VOD

 

[JAE]

VOD - I agree - the RR "U" bridge is exactly what is suggested here - in fact, I once designed a pedestrian bridge using wide flanges. However, there was limited space below the walkway of the ped bridge so we raised the WF shapes so that the bottom flange was near the bottom of the floor framing. This placed the top flanges high above the floor - unbraced laterally - and the bridge was spanning some 80 feet. We used vertical web stiffners to tie the top flange in laterally to the flexural stiffness of the floor structure and this served as a deterent to LTB in the top flanges.

 

[SacreBleu]

Since this is a situation where progressive collapse is to be avoided, I still have a feeling that the use of stiffeners is questionable.
My $0.02

 

[miecz]

We were faced with this problem in the boiler industry in the design of air and gas ducts. These steel ducts could become quite large and were stiffened with channels. Under negative duct pressure the outer flange of the stiffener was in compression.

I only have xerox copies from old textbooks that reference work by Muller-Breslau and Zimmerman. The solution to the required stiffness of intermediate lateral supports to an unsupported flange was a theory developed in 1893 by a German engineer F. Engesser. The solution was referred to as the "Engesser Formula".

 

[miecz]

After a little searching, I found that Engesser's solution is published in the Structural Stability research Council's "Guide to Stabilty Design Criteria...", The chapter is titled "Members with Elastic Lateral Restraints." There's even a paragraph devoted to applying the method to web stiffeners of plate girders.