Steel Beam Connection Design for Horizontal Loading 1

[amain002]

Hello all,

I am seeking for some ideas for steel beam connection design. I have a condition where I have a steel beam connection into a embed plate on wall and I have three loading conditions: Shear, Axial and Horizontal Loading. I can design for shear and axial load for simultaneous loading however, I am not sure how to design for all three loading conditions. I was thinking, I use double angle connection for shear and axial loads, and adding a clip at top or bottom to account for the horizontal load. Please suggest if you have any other better ideas which you have used it in the past. The horizontal loading is about 15 kips. Any help will be appreciated.

 

[mike20793]

I'm not sure what you mean by horizontal loading. Do you mean shear in the weak axis direction? I'd use a seated connection for all three and keep the sum of the demand divided by capacities for all three loadings under 1.

 

[KootK]

That sounds fine. If you put the clip on top, consider only welding the toes of the clip angle so that the assembly won't offer too much restraint to strong axis beam bending. With a horizontal load as light as yours, you might be able to just push that through the double angles and forgo the clip.

This might be a good application for a plain old angle bearing seat welded to the embed. That, by itself should be able to transfer all of your primary loads. Then add a single clip angle high on the web with horizontally slotted holes for torsional stability. Your axial and horizontal loads would enter/exit the beam eccentrically but that should be easily designed for.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[KootK]

Great minds...

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[amain002]

Thank you all for your response.

@ Mike: Thank You for sharing your idea. In my case I have shear force (reaction) of 100 Kips in strong direction of W beam, I have an axial force tension/compression of 25 Kips and Horizontal force of 15 kips in the weak direction of the W beam. I will not be able to use a seated connection because I have a connection into a embed plate and I barely can fit anything in the given size of embed plate.

@KootK: Thank you for the concept. I think I might go with the first idea. Weld only the legs of the angle. I wanted to provide an angle with slotted vert holes, but the angle being connected to embed plate I wont be able to do so. Thus, as per your suggestion, I will only weld the legs of the clip angle, so that there will be room for flexibility required for an double angle connection. Let me know if I am missing anything.

Thank you all, rdally appreciate your help.

 

[KootK]

I will not be able to use a seated connection because I have a connection into a embed plate and I barely can fit anything in the given size of embed plate.

If you're still in a position to control the fabricated length of the beam, you might be able to use a seated connection with the vertical seat angle leg turned up such that it would be behind the beam web. It wold be best if an unstiffened seat could be made to work. However, depending on your exact geometry, you might even be able to slip an angle stiffener in either side of the beam.

I wanted to provide an angle with slotted vert holes, but the angle being connected to embed plate I wont be able to do so. Thus, as per your suggestion, I will only weld the legs of the clip angle, so that there will be room for flexibility required for an double angle connection. Let me know if I am missing anything .

If desired, I suppose that you could use slots in the horizontal leg oriented along the longitudinal beam axis. Tolerances required in the beam length may eat into the play in your slots however.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[mike20793]

Is the embed plate already installed; can you not just make it as deep as you need it? I think the seated connection would be preferred, but it's certainly not the only way. How is the beam "taking" the horizontal loading? Is it from studs attached to the bottom flange, etc? Could you design a shear tab connection to take all the loads and place a clip at the top and bottom like you mentioned to prevent twisting? How wide is the embed plate and how deep is the beam?

 

[DETstru]

This might be what was already suggested but see attached. You can handle the strong axis shear and axial load through the double angle connection at the web. The flange plates take care of the weak axis horizontal load.

Since it's an embed plate, resolving real moment into it would probably be a bad idea. Assuming it's a headed stud anchorage, they really lose capacity when you introduce significant tension, which is what you're doing if you restrain the flanges at all. The slots in the flange plates take care of that. You only have to design the embed for the tension/compression, shear, and moment induced by the eccentricity of the double angle connection.

 

[dozer]

Just to be clear, this horizontal loading produces bending about the minor axis of the beam, right? I've got a similar problem. I'm just looking at the horizontal force in the shear connection so as not to muddy up the water. First, I would be careful just using double angles. I did a FEA of a W12x30 beam with a more or less standard double angle connection. The end reaction is 15 kips (I changed my original number to match yours so you can see what happens) in the horizontal (normal to the web) direction. This doesn't even come close to working. I've attached the screen shot of the von Mises stress. I used a scale of 0 - 150 ksi even though I had peak stresses over 150 ksi so you can get a better feel of the stress distribution. I didn't plot the mesh because it is so fine.

I've searched the internet and I can't find a "classic" solution to this problem. It seems like textbooks just ignore this. In my case, the engineer is showing a 30 kip horizontal end reaction on his drawing. I personally don't think a shear connection is the right answer but I'm going to keep at it to see if I can make something work. I like the seated connection idea and will take a look at this.

Regarding DETstru detail, that looks an awful lot like a moment connection to me. I guess the theory is the slots will let the beam flange slide relative to the plate causing it not to transfer major axis bending into the embed plate. I'm concerned though that the bolts shouldn't be pretensioned so they are really able to slide. I get nervous when I tell the field to pretension all the bolts EXCEPT these. Just snug tight and damage the threads (or whatever). Exceptions are a recipe for mistakes.

I sure would like to hear from someone who has had to carry a significant horizontal force in a shear connection because I'm starting to see why the "experts" always dodge this subject. It's hard stuff!

 

[Teguci]

http://www.eng-tips.com/viewthread.cfm?qid=403879

We just had this discussion on Axial load through a double angle shear connection (doesn't even include lateral shear). In short, Thornton, (one of the many authors listed for the AISC manual) has specifically advised against just this type of connection (use a shear plate). You may disagree with his answer (I do in parts) but it is worth looking into and treading carefully.

 

[dozer]

Teguci, yeah, but he is talking about axial load. I'm asking about a transverse horizontal load. This would produce minor axis bending in a single plate connection which would limit it's load carrying ability. Of course, it produces minor axis bending in the legs of a double angle but at least you have two legs to resist it. What I'm finding is having the shear transfer from the beam flanges down through the web and then into to angles is a real problem. You get high bending forces in the web. Not a very efficient way to do it.

 

[KootK]

What I'm finding is having the shear transfer from the beam flanges down through the web and then into to angles is a real problem. You get high bending forces in the web. Not a very efficient way to do it.

I've looked into this a bit myself. I've found an appealing antidote to be the use of a full height vertical stiffener on one side of the beam, 4"-6" in from the end. Granted, stiffeners aren't free.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[dhengr]

Amain002:
I think that Bill Thornton’s dislike of the double angle shear connection, which also see significant axial tension down the length of the beam is as follows.... When the axial load is applied and resisted, and the angles flex due to this load, there is considerable prying or a tensile stress field applied right across the weld root (perpendicular to the weld length axis) on the angle legs, and also at the terminations on any weld returns on the legs. This is a serious tensile stress across unprotected, and inferior, weld roots. Draw a FBD, a horiz. section through the connection and the beam web, and show how the angles flex and rip at the weld roots. We just don’t want to do that. That rationale for allowing this on a simple vert. shear connection is that the beam end rotation and movement is small enough, so that with weld returns, we fairly well protect the weld roots.

I have no trouble imagining Dozer’s screen shot of the stresses at that connection, and those are only due to the horiz. lateral force of 15k. You really should take a look at some combined stress conditions if the 100k vert. shear, the 25k axial tension along the beam axis, and the 15k horiz. lateral load can happen at the same time. I would take a look at a single shear plate on one side of the beam web, bolted to the beam web. This should be able to take the vert. shear and the axial load. Someone designed you into this pickle, and it leads to a pretty complicated connection, thank them for that. And, this all assumes that the embedded pl. can actually take these loads and reactions. This will likely be a thicker plate than you would see just for normal shear. I would bevel the two vert. edges on this shear pl. so I could get a good weld throat, both sides, without too large a fillet weld out onto the embedded pl., a PJP with a fillet weld reinforcement, both sides. I would stop these welds short of the t&b edges of the shear pl. and not weld around the t&b. This weld around is much too likely to leave a nick on the shear pl. outstanding edge (corners?), another stress raiser, we don’t want. This welding of the shear pl. provides a superior root, and a protected root, as the vert. shear and axial loads (stresses) move through the connection. Now, for the lateral horiz. load, still finally taken by the single shear pl., I would apply a single fitted, web stiffener pl., flg. to flg., clipped at the web/flg. radii, on the side of the beam away from the single shear pl. I would apply this stiff. pl. as near to the end of the beam as my welding would allow, so as to minimize the eccentricity “e”of this stiff. pl. line from the embedded pl. This web stiff. pl. will more likely apply a 15k line load to the vert. shear pl., as a moment on the shear pl. and its welds [(15k) / (vert. shear pl. height, less 1" or so in height)][ e inches to the embedded pl.]. This moment, over the vert. shear pl. height is resolved as a tension and compression stress, separated by the distance btwn. the center axis of these two vert. welds. Now, what do my combined weld and shear pl. stress look like, and I would be conservative.

 

[Teguci]

I'll add that, for sizeable welding, you will want to provide instructions(and require a review of the welding process) to the contractor to minimize heat from the welding process (stich welding and multiple passes). Unless you have a gap around the embedded plate, you will probably see some chipping of the concrete when the plate heats up. What won't be visible will be any damage that occurs to the headed anchors and embedded rebar on the back side of that embedded plate.

 

[KootK]

@Teguci: I've encountered such problems in the past on some large scale connections. Would you be able to supply any more detail with regard to the instructions that should be provided to the contractor? In keeping with how things seem to work around here these days, got a reference?

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[Teguci]

@KootK: Sorry, no reference. Just part of an old company standard. The note on the detail read something like - Contractor to minimize heat by stitch welding with multiple passes.