simple shear connections

[RacingAZ]

Steel beam framing to another beam with a typical shear tab and 1 row of bolts in ssl holes. Where the end of the beam is at a concrete wall, it is connected to the wall with an embed plate and welded shear plate also with 1 row of bolt in ssl holes.

I have a reviewer telling me that these support connections should be considered fixed and the bolts designed for the reaction moment. I said I typically assume this as simply supported with a simple shear connection and I only design for the reaction shear and the eccentricity moment. I believe both conditions are more of a flexible type connection than rigid. Who is right?

 

[Ron]

In actuality, it is somewhere between; however, for modeling purposes, I would consider your connection to represent a pinned condition.

You could bracket the reaction by computing both fixed and pinned conditions.

 

[RacingAZ]

Thanks Ron, that's what I read here on some of the threads that I searched.

The problem is that the structure is already built and modeling the beams as fixed, the connections cannot accommodate the reaction moment without a substantial re-work. Even swapping the bolts with A490 would still not be enough and I cannot do a double shear connection due to the owner's preference for the structure to be removable.

One thing I was told when the structure is loaded, they would notice the deflection on the longest beam that approximates the result on my beam model for a simply supported beam. For me this is a confirmation that the beam is behaving closer to a pin connected beam rather than fixed.

I'm confident that the connections are OK, I really just need some ammunition to make the reviewer agree with me.

 

[BAretired]

Calculate the beam rotation with hinge assumption and demonstrate to the reviewer how that rotation can be accommodated by the support.

BA

 

[KootK]

I design these connection more or less as you do and that's been the standard practice everywhere that I've worked. Lately, I've been replacing the top row of embed studs with deformed bar anchors in an attempt to create a situation where I feel that I would retain shear capacity should the connection fail in bending or tension. Of course, since you're dealing with an as-built situation, that won't help you much.

How to deal with your reviewer will very much depend on what their particular concern is. Here are some ideas:

1) Support for a pinned connection.

- Flanges not connected.
- SSL holes might facilitate beam rotation.
- Your connection plate might yield in flexure before you develop the beam capacity.

2) Support for a fixed connection.

- If you have significant shear, your SSL holes might bind up (Link).
- The tolerance provided by the SSL bolt holes might actually be exhausted in situ.
- Pretensioned bolts may tend to act like slip critical bolts in service.

I'd have some ready answers to all of these questions before you make your pitch.


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.

 

[RacingAZ]

Thanks guys!

The reviewer's concern is with a fixed beam model, the reaction moment resolved into horizontal shear would over-stress the bolts to failure. About the ssl holes binding up, I think it's a non-issue here as the connections only have 2 bolts. I'm guessing that's more likely to occur in a long row of bolts.

 

[jdcollins]

It sounds like the reviewer is not concerned with the actual embed plate portion of the connection as much, but rather he is stating that the simple shear connection itself, due to being rigidly supported on one side by the wall, should be designed as a fixed connection. In my opinion this seems a bit of a stretch.

As for your ammunition, on page 10-4 of the 13th ed. it states that "The simple shear connections shown in this Manual are suitable to accommodate the end rotations required per AISC Specification Section J1.2."

I'm thinking that if your simple shear connection is comprised of the parts listed in the single-plate connections tables in the code then you've got what you need. Alternatively, if you want to get more theoretical you can calculate the rotation BAretired suggested and verify it matches with the code commentary required rotation for a simple shear connection.

 

[KootK]

If the bolts don't bind, then there's no moment transfer.

If your bolt failure mode is bearing rather than shear, as I would expect, you could also argue that mechanism is self limiting and would effectively create the required pin.

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.

 

[rb1957]

I think the reviewer is looking at the worst case ... a fully fixed end; and if you have more than one fastener than you've got the potential of a fixed end; but the real world is more flexible than that. for small loads it approximates a fixed end, but then yielding (of the fastener, of the hole sides), a small amount of slip, etc introduces rotation into the joint, relieving the moment but increasing the deflections and the springiness of the floor, no?

it's probably not the best time to review a design ... after it's been built !

another day in paradise, or is paradise one day closer ?

 

[RacingAZ]

He finally agreed with me, actually he just kind of conceded by not discussing the issue anymore. He raised a new one though, now questioning what is the guarantee that both bolts are engaged in shear, not just one bolt carrying the entire shear.

I'm thinking as long as the ssl hole tolerance per AISC is met, I don't have to worry about that scenario.

 

[hokie66]

Tell him the guarantee is in the inherent ductile properties of steel as a material...but maybe that's not a good tactic, as he may then focus on the non-ductile material, concrete. Sounds like this reviewer is trying to get an engineering education at your expense.

 

[TLHS]

There are all sorts of fun theoretical discussions that you can have. I often have them in the office. But realistically the real argument in support of these assumptions is that they're standard practice in the industry and if they didn't work buildings would be collapsing all over the place.

 

[canwesteng]

Is this reviewer a structural engineer? I would hope not. In any case there is plenty of literature and test data to support your positions in both cases.