Stiffeners in torsion

[3doorsdwn]

Could someone direct me to a design example for a pair of stiffeners [full depth] on a beam subjected to torsion? Specifically, I am trying to check the weld of the stiffeners to the flange and the web. Thanks in advance.

 

[swearingen]

Using common assumptions in beam theory, there would be no load due in those welds due to beam torsion. In reality there is some, but it is almost certainly negligible.

What is your situation? Beam size, stiffener thickness, weld size, beam span, beam support, load application location and type?



If you "heard" it on the internet, it's guilty until proven innocent. - DCS

 

[3doorsdwn]

I have a w10x112 spanning 8'; it is subjected to 65 kips of shear (with an eccentricity of 2.5 inches) at two points (i.e. the quarter points: 2 feet from the beginning of the beam; and 6 feet from the beginning. I'm checking it with the beam modeled as either pinned or fixed (to be safe). It has 2 [full-depth] stiffeners (on each side) with 5/16" weld to the flanges and the web.

 

[Lion06]

The role of the stiffeners in this situation is for web crippling or web yielding under the concentrated load. I don't think they are contributing to any torsionaly restraint and as such there is no way to check the weld stress for it.

 

[3doorsdwn]

Obviously they don't provide much in the way of restraint, BUT when the beam is rotated they do "go along for the ride" [for the lack of a better way to put it]. I doubt there would be much stress due to warping torsion, but I would think there would be some due to St. Venant's.

 

[trainguy]

3doorsdwn,

If the stiffeners are at the load application points, helping to resist web crippling, the only calc that comes to mind is to size the vertical welds for transferring all the concentrated load into the beam's web.

Other than that, just use a weld that looks right based on thickness of joined parts.

tg

 

[csd72]

To be conservative:

1. convert the torsion into a force couple at the centerlines of the two flanges.
2. divide this force by the total length of weld between the top of the two stiffeners (1 each side) and the top flange.
3. Choose a fillet weld appropriate for the calculated k/in and apply it throughout.

 

[fasboater]

Wouldn’t you really have 2 different scenarios.

1) web be crushed or bent due to vertical load
2) web being bent due to eccentricity (ie moment through the web)

The moment is what I believe we are really looking to resist. So couldn’t you treat that section with the welded stiffener as a T section and find out your maximum allowable moment through the T?

 

[trainguy]

3doorsdwn.

Please let us know where along the length the stiffeners are, and how you plan on providing torsional fixity at the beam ends.


Also, how exactly is the 65kip load is applied?

tg

 

[3doorsdwn]

trainguy, the stiffeners are located where the loads are (at the quarter points). I am welding the flanges to the attachment. Since I want to look at it both ways (with both end conditions considered), the attachment really isn't important. The 65 kip load is applied by (eccentric) bearing on the flange.

 

[trainguy]

Ok,

then I would do 2 calcs:

1 - per my previous post, sizing the vertical weld for the full 65 kip load. I guess it's roughly 65k / 10 inches = 6.5 k/inch, into the 2 fillet welds that connect the bearing stiffener to the beam web. You may have to do a groove plus fillet weld. Asd you may know, bearing stiffeners have this sole function - to take a direct load from the flange and to carry it into the beam's web.

2- Check web crippling including the contribution of the stiffener, so a certain allowable stress (from your design code) over a portion of the web (including fillet) plus the cross sectional area of the stiffener, as if the stiffener is also a web that is resisting a concentrated load.
The effective area will be a (capital)T shape, unless your applied load bears wide enough to engage the stiffener on the opposite side of the web, in which case you'll use a t (cruciform) shape.

I hope my chicken scratch is clear enough...

tg