Extended Endplate Shear Connections? 2

[JLNJ]

I have searched Google to no avail, so now I’m turning to the even more powerful eng-tips forum members.

We have been seeing a type of shear connection where the web of a tee is fitted perpendicular to a girder web and the connecting beam is bolted through the flange of the tee using an endplate. I guess the advantage to this is that the beam can be cut square and cut short. Additionally, no bolts are shared, so OSHA should be happy. Has AISC commented on these? The mechanism which provides end rotation flexibility is not intuitively obvious to me, and the apparent eccentricity hurts my eyes to look at.

I have used connections similar to these when have a column continuous through a girder (with their respective webs @ 90 degrees). I fit a tee to the girder web to provide a direct load path for the column flange load above into the column flange below through the outstanding tee flange. If a beam intersects at the same point, then it frames into the tee flange with standard clips.

These other connections I’m asking about are one-sided and are not at a column. The tee is simply to shorten the beam and to make the connection outside of the limits of the girder flanges. Has there been any discussion of these types of connections? I presume there has been. They seem to be somewhat common in parts of Europe.

 

[Brad805]

I think the reason is simpler myself. When I was looking into extended shear tabs I was surprised to read comments like, "we need further study". Simple shear tabs are very easier to create simple tables for load capacities so long as the basic criteria are followed. When we get into cases where the relative stiffness of the different members is more important that is much more complex for our litigious industry. Look at the pre-eng industry. Behind all those sheets we get they have very complex algorithms (most of the time) to calculate all of the connection details.

Out of interest I did a simple study to see the difference in the OP's case. I do not have much time right now to summarize this in more detail, but I thought I would post a few of the images for others and to see if there was any interest in looking at a few others. I know the von mises stress are not the best for the discussion, but they seem easier than a vast number of plots for the scope of this. I think there is merit to this connection detail for a fully automated plant. In addition to the savings in beam length it also seems very probable one will have more bolts than they would have with a shear tab. I would hope the steel companies will not suggest an asymmetric bolt pattern? For a plant that is not automated this detail seems more prone to fit problems.

 

[KootK]

In concrete, the compatibility torsion release valve is beam cracking & stirrup yielding. In steel, the release valve is connection flexibility. If you sacrifice your connection flexibility, you switch from having compatibility torsion to just... torsion. And trying to address torsion in steel is usually a much more expensive proposition than detailing to eliminate it.

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.

 

[hokie66]

KootK,
There are only two types of torsion: compatibility and stability. By "just...torsion" I assume you mean stability torsion. You don't have stability torsion in the OP's situation. The rotation doesn't have to be released by connection flexibility, but rather by supporting beam rotation. That is one of the most wonderful things about steel...it is ductile. Open steel shapes have little resistance to torsion, so they will rotate a lot with little change in stress state.

 

[KootK]

Yes, I mean stability torsion. Or equilibrium torsion in the nomenclature that I'm used to. Either way, I disagree with your interpretation.

Compatibility torsion is that which can be redistributed through yielding. The rest is equilibrium torsion. Steel is generally only ductile at the material level. At the member level, many members will fail via some form of non-ductile buckling. The primary story of steel is the story of buckling, not yielding.

Many steel section are indeed flexible in torsion. But that just changes the torsion that member will attract. It doesn't mean that redistribution will occur. You're unlikely to ever plastify a wide flange girder in torsion. Rather, you'll induce some complex mode of buckling or simply tear apart your hapless girder end connections.


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.

 

[hokie66]

I suppose we will just have to agree to disagree. You think I am wrong, and I think you are wrong. But if you can point to an actual failure as you described, I will listen.

 

[KootK]

I'm not going on a lord of the rings style quest to dig you up an example Hokie. AISC thinks that connection flexibility is important enough that they recommend it in the bible. Is that not a good enough reason for you to listen?

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.

 

[hokie66]

Buckling of the supporting member is a red herring. The rigid connections make buckling less likely, not more.

 

[KootK]

I think that's a complicated matter Hokie. I'll assume that we're talking about lateral torsional buckling:

1) The distance between LTB brace points remains unchanged as the distance between supported beams.

2) The fixity of the brace points improves due to the increased torsional restraint provided by the supported beams. This benefit is less significant than it might first appear however. You gain some rotational restraint but do not gain true torsional fixity as warping strain is not restrained.

3) To achieve true torsion redistribution, something has to plastify. In this case, I think that's the flanges yielding inwards from the tips under warping. As the flanges yield, Iy and Cw drop fast and the likelihood of LTB buckling between beams increases.

When you add all that up, I'm not really sure what you get.

I'm really not all that concerned about buckling although I always enjoy debating the finer points of theory. The primary point that I've been trying to convey is that there generally is not girder torsion redistribution at work in steel systems as there often is in concrete systems except, perhaps, in the connections when they are made rotationally flexible. You wanted to know why steel folks care so much about connection rotation capacity. That's why.

You do a funny version of agreeing to disagree. I like it.

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.

 

[Brad805]

Koot, I am curious if you would reject the proposed detail? I would not.

 

[KootK]

Nope, at that scale, I'm fine with it. As I mentioned at the top, I've done much worse with steel supported precast. With that, I ran into issues with precast being installed 100% on one side of the girder before the opposing side was installed. Beam rotation on the order of 30 degrees. Of course that truly was equilibrium torsion in the temporary condition.

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.

 

[Brad805]

Yes, I agree in a case like that it is much different. With steel beams supporting precast panels it gets much more difficult when one is dealing with longer spans and panel segments up until the welding is complete.