Curved Beam - Residual Stress 1

[slickdeals]

Folks,
I have an interesting problem at hand. A canopy was designed too short and it needs to get extended. The original canopy was a cantilevered truss. We are now adding a support close to the 2/3 point of the truss and attaching/extending a curved W12 to form the new canopy. See attached section.

My questions are as follows:
1. I am going to calculate wind loads in a myriad of ways (open building, overhang, parapet etc etc) and determine an optimum number to use. Any suggestions?
2. The member is architecturally shown to be a bent the hard way to an outside radius of 7'-0". On first thoughts, it appears to be too tight and may not even be possible to be bent. Any additional info?
3. If the beam can be bent, how much residual stress will be built into it and how much will be available for design? The downforce will be opposite to the residual stress, but the uplift will induce more stress in the direction of the bending.

 

[Ron]

Will the bending be done cold or hot? Cold bending will induce more residual stress. You can estimate the residual stress in two ways...first you can determine the strain from the lengthening process of the outside radius. Another way, and generally more accurate is to test for residual stress using a special type of residual strain gage. This is a fairly easy process but it does require some equipment/instrumentation.

 

[rb1957]

obviously you'll heat treat after bending, right ? so the residual stresses would be relieved, no ?

 

[slickdeals]

I have not dealt with curved steel other than cambered beams. I would appreciate any literature/expertise you folks can provide in this regard.

 

[a2mfk]

slick, have you tried AISC.org and their FAQ and search their website...

 

[Ron]

rb1957...for most building construction applications, post-bend heat treatment is not used, thus the importance of trying to determine the stress state.

 

[rb1957]

so it'll be above yield ... i'd suggest non-linear FEA ... apply the load, develop some permanent set, release the load to get your deformed shape. i imagine that you'll be applying many small loads to bend the beam, say a load every 6"; you might be able to simplify the model by modelling one bay (one 6" length) loaded at the far end, cantilevered at the near end.

 

[XRS]

I wouldn't be to concerned about residual stresses since they have little effect on ultimate strength. A more important concern may be the thinning of the flange that occurs on the outside of the bend. We bent some 24" diameter pipe on a 12' radius for monotube overhead sign supports. Besides being expensive to fabricate, the pipes showed significant thinning of the walls at the outside of the bend. The walls at the inside of the radius thickened a similar amount. These were heated before bending.

I believe the W12's will be much easier to bend but flange thinning will probably occur. Depending on bend radius it could be significant enough to account for in the design. Some of the fabricators could advise you on what to expect. Advanced Bending Technologies, Chicago Metal Rolled Products Company, White Fab Inc, and Abina Pipe Bending Co are some fabricators that have advertized recently in Modern Steel Construction.

Here's two good sources of information on steel bending.

http://www.modernsteel.com/backissues

Free down loads of back issues. Look for "Bending Considerations in Steel Construction" Oct. 2009 MSC. There is another article on curved steel in the Oct. 2010 issue. Curved steel has been a popular topic recently.

http://www.aisc.org/benders

Provides links to bender/rolers and has free downloads of publications about curved steel.

 

[ash060]

With a curve that tight I would use two or three straight pieces and round the curve out with some light gauge framing. I think the fabricator would like that much better than doing some super tight bend.

 

[trainguy]

I would use a built up beam made of 2 laser cut plates for the flanges and 1 roll-formed web plate, with intermittent welding.

Seems to me this would be easier than bending.

tg

 

[rb1957]

wouldn't that be roll formed flanges and a cut (flat) web ?

 

[CTW]

I'd suggest contacting a member of the AISC bender-roller committee to get their input regarding the feasibility of the 7'-0" bend.

The list can be found at the link below along with some other useful information if you don't already have it.

http://www.aisc.org/content.aspx?id=18112

 

[csd72]

7 foot looks too tight to me.

 

[rb1957]

"we choose to ... do the(se) ... things not because they're easy, but because they're hard ..."

sure it's unusual and not efficient and maybe can't be formed as originally intended or needs some post-processing but that's the challenge (isn't it ?)

 

[Ron]

Here ya go...a little "residual" stress...

http://www.videobash.com/video_show/pipe-cutting-fail-1250