Local Flange Bending due to Weld Along Flange

A sketch might help. It is too easy to misconstrue your meaning from a verbal description.

BA

Figured that was coming... And you're right BA. I'll try to post one to tonight. Thx

Sketch attached, hope this explains my rambling. Thanks all.

I don't see any prying action here with the welded connection as shown.

I think the applicable limits states for the temp beam are:
AISC J10.1 Flange Local Bending (make sure flange is thick enough)
AISC J10.2 Web Local Yielding (probably doesn't control)

and of course the typical beam bending, shear, and deflection checks for the temp beam.

Flange local bending, as defined by AISC, only applies when there is a plate welded transverse to the web of the beam (see the commentary).

You might be able to use part 15 of AISC about hangers - with the weld at the outside edge of the flange no prying action can occur, just normal bending of the flange. So table 15-1 may be applicable and conservative.

You might also consider just using a single hanger connection at the coped end of the existing beam rather than welding full length of the temporary beam - this might make disassembly simpler.

I probably shouldn't have mentioned prying at all; that was just an attempt to illustrate my level of knowledge on the subject. Enough to feel confident in the calcs that I do, but also enough to know there are issues I am not fully versed in.

I see the issue as a local web yielding rather that a prying action. I deal with it quite a bit with point loads being placed on the bottom flanges by the wheels of a beam trolley, and struct210, I was not aware that AISC so narrowly defined local flange yielding; thanks for pointing that out. The typical shape of flange deformation one would see in a bolted hanger still seems like the mode in which my upper temporary beam's lower flanges will deform. That said, I'd like to follow ASME NUM-1-2009 in which they outline a method for analyzing local stresses in beam flanges with concentrated loads. However, ASME does not define a length or effective length of the affected flange, though I assume they had something in mind that was sufficient to assume that the flange was fixed at the web. If that's reasonable, I believe that the effective flange length as outlined in AISC is reasonable, but I have a distributed load along the length of weld. You folks think an additional 10tf beyond the end of weld would be the effective length? Seems like the section modulus of the flange may end up a bit high, though I guess if you really are distributing the load along the web, you are gaining a lot of flange area.

I agree there is no prying action but the weld will be subjected to a tearing action caused by the combined rotation of two flanges, so it is not a simple shearing force.

The temporary beam must be capable of carrying the full load laterally unbraced, unless you propose to add bracing.

If there is sufficient clearance to use a pair of trapeze hangers about 14" apart, that would seem to be a better solution than welding. Each trapeze could consist of a pair of all-thread rods fitted tight against the flange of the existing beam with a strap plate under the existing beam and another over the temporary beam.

BA

Possibly. The effective length per AISC is based on a yield line analysis with a plate transverse to the web. A good conservative way to look at it may be to take an effctive length of the weld, coming up at a 30 or 45 from the point load, and analyzing from there. I think this system is complicated load distribution wise, since the beam is attached on one end, not on the other, then welded to a beam along its length. The combined stiffness of the beams work in bending with one another, as well as transfering the point load to the temp beam. A single connection point near the end of the beam would be easier to feel comfortable about analysis wise.

Of course you'll have bending on your flange. You just won't have the specific case of "flange local bending" defined (and capacity quantified) in section J10.1. The question is what effective length of flange can you take advantage of. You'll have to use engineering judgment here.

I would not use the bolted connection prying tables. The bending eccentricities are less in a bolted connection than what you'll have. I would think Table 15-1 is quite unconservative. Salman and Johnson has a good discussion on this (page 795 of the 4th edition).

Thanks all for the discussion... I agree struct210 that supporting at the "loose" and feels better analysis wise, and that's what I've been leaning towards this morning, but the question is still of interest to me. And nutte, I believe you're spot on that I have some kind of flange bending but not a real great way to quantify it.

And if I can ask a possibly stupid question, what is this Table 15-1? The 9th chapters aren't numbered that high and Table 15-1 in the 14th is about Crane Rail Splices. I do not have access to a 13th at this time.

Also, agree with BAretired - I'm assuming no contribution from whatever bracing effect may be seen by the temp beam and using L=Lu.

Unfortunately, the hanger clip and trapeze ideas, while valid, don't fit into the particulars of the situation. These, of course, are too numerous to show up on a simple sketch so none of you folks could have known about them. This all came about b/c of a ~80 kip piece of equipment was designed with lifting lugs that interfere by 5" with the steel required to support the thing. It's on us to fix it and hang the thing. Good times!

Prying action table is Table 15-2 in the 14th edition Manual.

I figured you would use a b equal to bf/2, but upon further examination the table doesnt accomadate a very large b value. Does the Salmon use a different approach than Fy*Z for flange bending strength? Sorry to go off topic, but is the Salmon & Johnson a good general reference for steel, incorporating some more complicated situations? Just curious - can never have to many good resources in the library.

The difference between the bolted and welded versions is more than just the "b" dimension. The bolted version provides clamping restraint from the bolts, so your model is a fixed-fixed beam. The welded version does not, so the model is a pinned-pinned beam. The moment in your flange will be higher with the welded version, even for the same "b" dimension.

I see your point there. Attached is a derivation - I think this is appropriate - showing T to be half as much as the table.

struct210, I'm assuming your (1/4)Fy*tf^2 is the (1/2) * base * height^2 of your stress on the section. Thanks for the jump start on that. I'm switching to an alternate method of supporting the ex W12x53, but I appreciate the discussion folks.

If I seen any of you folks out sometime, I'll buy you a beer.

Okay..Coors Light sounds good to me.

BA

Flange bending will result in concentrated stresses at the root of the fillet welds. This is not recommended. Similar to one sided fillet weld in tension.

Providing fabrication and erection efficient structural design of connections. Consulting services for structural welding and bolting.

Do something like structSU10 proposes. Take bending on flange and put a conservative safety factor on it. Otherwise FEA.

cannot you add some stiffeners in both beams in order to avoid flange bending?
I agree stress at the weld root is better to be avoided.