Weld web stiffeners to a loaded beam

[ugandabob]

I am designing web stiffeners for an existing roof beam to accommodate a new rooftop platform. The existing web cannot support the full factored reaction related to the future platform. The intent is to construct the platform prior to welding the web stiffeners. The full factored load will not be there, so it will be ok, but the web will still be stressed by the dead weight and snow (Ottawa, Canada).

If the web is loaded to, say, 75% of its capacity prior to adding stiffeners, how does this affect the reinforced load-carrying capacity? The beam web and stiffener plates will be stressed at different amounts; I expect the web will yield much sooner. Should I just ignore the web and design the stiffener plates to resist 100% of the load?

Thank you.

 

[dik]

Provide temporary support if necessary.

Rather than think climate change and the corona virus as science, think of it as the wrath of God. Feel any better?

-Dik

 

[XR250]

I imagine the failure mode is buckling so it should not matter when you weld them on as long as it has not buckled yet. Someone please correct me if I am wrong.

 

[bugbus]

Without doing any numbers, my gut feeling is it would be OK to ignore the fact that the web and stiffeners are stressed to different amounts. I always imagine that (ductile) structures tend to be highly forgiving with regard to pre-existing loads and stresses. If your bearing capacity is governed by yielding, then only a small amount of initial yielding in the web would be needed to redistribute the remaining load into the stiffeners. If the bearing capacity is governed by buckling, I believe that the code provisions probably already provide some allowance for residual stresses in the various elements resulting from welding, which may be near yield anyway. At least in the Australian codes, the stiffened web is treated as a notional column with alpha_b = 0.5, which normally would apply to welded built-up sections that may have residual stresses near the yield stress of the steel.

That's just my first instinct, happy to be corrected though.

 

[dhengr]

Ugandabob:
I would want to shore the beam before I did any welding on it, take most of the load/stress off of it, and I would want to do that reinforcement and welding before the new loads are applied. It is reasonable to assume that the existing beam is fairly highly stressed due to the preexisting loads, and now you are going to add significant new concentrated loads. With a concentrated pipe leg, or some such, coming down from a loaded platform above, you are (should be) worried about the potential of high bearing stresses in the web, immediately below the radius btwn. the flg. and the web, at the ‘k’ dimension, and of course, this load can corbel out (wider distrib. length on the web) at about 45° as it moves down to the web, or you are worried about web crippling/buckling as this concentrated load distributes down into the web, thus the want for the stiffeners. If it is primarily a bearing stress issue, they can be run right up to the limit, they are quite forgiving and are well confined, by surrounding mat’rl., in terms of causing any problems. Finally, remember that whatever the existing stress level in the beam flg. and web, when you start welding on it, you essentially reduce the material’s “E” to a very low value, it’s a molten puddle over a small area, which tries to form a hinge right there, and thus high potential for large rotations. There are already high residual stresses in the beam from rolling and cooling at the radius btwn. the flg.. and the web, so clip the inner corners of the web stiffeners so you don’t/can’t weld into that corner and cause an added nasty triaxial stress condition. You are trying to get that concentrated load down into the web of the beam, but you can’t do that properly when the beam is under high bending stress.

For the most part, all we do with rolled beams and welding assumes that there well be some residual stresses do to the processes, and these have not particularly come back to haunt us, and have been taken into account in our analysis. But, in the extreme, when you have a beam highly stressed and then melt parts of it in a welding process, what prevents it from just caving in, particularly when you weld across the flanges, the most highly stressed part of the beam?

 

[KootK]

It's long been my understanding that one ought not weld across the width of the flange of a loaded beam. Similar to dhengr's comment about the elastic modulus, it's simply dangerous to melt something non-redundantly required for global member stability, even temporarily. For a lightly stressed beam with a wide flange, perhaps you could use intermittent welding on the flange or do that welding in phases. I've not yet attempted those strategies myself. Put me in the jack & shore camp for now.

Again similar to dhengr's comments, many of the failure modes for which you would add a stiffener could also be addressed with a wider, and perhaps stiffer, incoming base plate.

 

[dik]

Just a caution... Es diminishes quickly at elevated temperatures.

Rather than think climate change and the corona virus as science, think of it as the wrath of God. Feel any better?

-Dik

 

[ugandabob]

Thanks everyone. I will push for shoring during welding to take out the risk