LRFD built-up beam questions 1

[DerChad]

No response so I'm trying a different title.

I have an HSS in a polygon shape which has a stitch bolted piece of reinforcement (channel) being added (built-up section). The section has an axial load but it is very small in magnitude when compared to the moment. Chapter F13 of AISC 13th Edition refers me to E6 and D4 for bolt spacing. Both of those sections refer me to Chapter J3.5. J3.5 instructs me to limit my bolt spacing for plate-plate and shape-plate sections but says nothing about shape-shape built-up sections. My understanding of the bolt spacing requirement is that it is there to ensure that the plate will not buckle between the bolts and the section will remain fully composite. If I use Chapter F13 for my nominal bending moment and limit my maximum stress to the maximum stress allowed by Chapter E, would that hold true to the intent of Chapter J3.5?

Some engineering firms are treating sections between the stitch bolts in the above mentioned problem as though they are fixed at one end (K=0.8). To achieve this K value, wouldn't we have to increase the lu to allow for a moment arm at one end (moment not possible without a moment arm)? The assumption is that the reinforcement will try to deflect in an S shape and the polygon surface keeps it from forming the S.

Thanks in advance for your input.

 

[BAretired]

I don't know about the AISC document because I don't have it.

It seems to me that the main polygonal section would prevent the channels from buckling inward. If the channel can only buckle outward from the main member, then it is fixed at each set of stitch bolts. The ends of the channel would need extra stitch bolts to develop fixity.

I am not following your argument about a moment arm. Aren't the channels being stressed axially?

BA

 

[DerChad]

Not having the document isn't such a bad thing.

My post seems like a lot of work for little result difference but the distance between the bolts is pretty small. In ASD design, I wouldn't worry about it but LRFD seems like it offers very little room to maneuver.

Yes, the channels are components of the beam that are being stressed axially. By moment arm, I was meaning that a single bolt doesn't offer fixity by itself so it would require the compression to resolve at the pole face at some point above the stitch bolt. The bolt spacing is fairly short (18") so it seemed to me that if we are going to count the fixity, we should probably increase the 18" to some higher number to show the fixity at one end. The figure I was considering was around 3". So if K=1, I would use Lu = 18" (KL = 18). If someone insists that I use K=0.8, I would increase Lu to 21" (KL = 16.8).

 

[dhengr]

DerChad:

I take it you are trying to reinforce a brake formed power transmission pole, or some such polygon shape. The term HSS probably throws people a bit of a curve ball since you probably won’t find your power pole section in the AISC tables of std. hollow structural sections. I don’t have the AISC document you are looking at either, so I don’t know exactly what it says and I live in the US. I assume the base plate and foundation are O.K.

I would be as worried about the possible buckling of the channel flange tips as about the buckling of the entire C section and, and at the moment, I don’t know which would go critical first. That being said, your bolting must be adequate to transfer the shear flow to make the C an integral part of the pole. Then the C section might buckle btwn. the bolts. The buckled shape would be something akin to a sinusoidal shape, but not a sine wave, because your pole prevents the buckling into the pole center, every other half cycle of the sine wave. I think that’s called a Cycloid? The bolts must be spaced closely enough to prevent this buckling at whatever stress level you are working the C section, and this is obviously not the first buckling mode of a pinned/pinned col. The lever arm or fixity term may be a bit of a misnomer, the fixity condition is actually more akin to a continuous beam, with pos. & neg. moment regions, over many reactions (bolts and washers). Thus, using something less than the bolt spacing for your buckling calcs. would seem quite appropriate, but I would need to think a bit about .8(bolt spacing) or something in that neighborhood. I don’t liken your K=.8 to the pinned/fixed col. which I think you are looking at, but rather as semi-rigid, btwn. fixed/fixed (K=.65) and pinned/pinned (K=1). Buckling might start, but that would start to cause a neg. moment under the bolt/washer, which would resist the buckling or shorten the buckling length. You should be able to calc. a buckling stress or load about the weak axis for a given unsupported length, then there will be a bunch of probabilistic multipliers and reduction factors which will get you near what you see in AISC. I can no longer keep up with all these multipliers and reductions placed on the original theory and testing, so I can’t help you with these.

Look at the thread “Local buckling of thin-wall tubes in bending” thread #507-268445, it’s another buckling problem, a bit like your’s. I’m not suggesting you have an easy problem, just suggesting ways to look at it which might help you understand what you see in AISC, or do some interpolating.

 

[DerChad]

The structure in question is a monopole for a telecommunications application.

I'll take a look at the post you suggested.

Thanks so much for the insight on the view of the K value. I'd feel much more comfortable with a K value of 1.0 but I can understand what you said about it being semi-rigid because of the continuous quality of the reinforcement element. A value of 1.0 for K would be conservative and perhaps I'm a little too conservative. The difference in the end results between the two K values for my problem is between one and two ksi.

Not to whine too much (this means that I'm going to whine far too much) but with AISC 13th Edition, I fight each and every attempt from manufacturers to "skinny up" their design. They're trying to be efficient (cheap) so they can get the bid. I understand their perspective but it feels that we've given up far too much already in this new code. And if you don't go along with the program, you lose the work to other engineers that might not be as interested in being conservative. I wonder if we're doing ourselves any favors by being so competitive.

 

[ishvaaag]

You can still model your outfit of plates (or as HSS member but this will reveal less what the channel does to the face of the member where bolted), the channel divided in member segments of lesser length than the separation between bolts (ensure then shear deformation is contemplated, because the segments will be far too short for the length), put links, rigid or whatever where the bolts, and use P-Delta with K=1 for every segment at its (segment) length. Verify the calculation converges (the structure is stable) and so you don't need to reduce the bending or axial stiffness due to the standing level of axial stress; if so, do the stiffness reduction and proceed again till the design stabilizes whilst the calculation converges. This should give a still conservative design or check for the channel.