Single plate moment connection 1

[druminman]

I am reviewing some moment connections that were submitted to me and there is an unusual moment connection that I have essentially rejected, but wanted to get the opinion of you guys on.

The connection is a WF beam to an HSS column, and the designers used what is essentially a large shear plate with no flange plates at all. The large shear plate is slotted through the HSS and bolted with two vertical rows of bolts to the beam web. The bolt rows are spaced up to 9" apart.

My biggest issue is the transfer of the bending stress in the flanges into the web.

I've never seen such a connection before, and could find no reference to one anywhere.

What do you guys think?

Thanks,
Patrick

 

[Lion06]

That's not a moment connection. Unless your end moments are very small and they provided calcs showing it works, I'd reject it, too.

 

[rscassar]

The majority of the stiffness of a W-beam is derived from it's flanges. Unless you can engage the flanges in the connection then it won't be able to transfer the moment. Good call in rejecting it.

 

[ToadJones]

sounds like an extended shear tab connection, not a moment connection.
I can see where they are trying to develop some moment capacity by spreading the bolts out, but I still think the moment transferring mechanism is not there.

 

[nutte]

How big is the moment? If it's very small, the connection you describe could be made to work.

 

[druminman]

Moments are up to 30 k-ft on 12-16" deep beams -- not very small by my estimation

Thanks for your input. I'm not even sure how one would go about verifying the stress transfer from the flanges to this plate short of a solid model.

 

[steellion]

The sanity check for moment connections is to make sure that both flanges and the web are engaged to the column. If not, it's probably not a moment connection.

You can direct him to the AISC Manual 13th Ed, which has several examples of how to detail proper moment connections to HSS members.

 

[nutte]

30 k-ft is a 36-kip shear load with 10" of eccentricity. This is not beyond the realm of possibility for a shear tab.

 

[ToadJones]

Shear tab might work, but is it providing the proper rotational stiffness to the joint?
I am assuming that the connection is part of a moment frame.

 

[apsix]

I would be wanting the web to have sufficient capacity for practically all of the moment (and shear), a web doubler plate could make it work.
The bolts should be friction grip.

 

[frv]

Without getting in to actual capacities, I think it's just plain bad detailing.

 

[apsix]

If it's adequate and simple shouldn't we call it innovative?

 

[ToadJones]

I don't think so.
It is misguided, poor detailing. It ignores the most basic concepts central to a moment connection.

 

[Lion06]

I agree with the idea that it's just bad. As Toad notes above, the rotational stiffness is not there. It doesn't much matter if the bolts and connection have the capacity if it allows so much rotation that all of the moment is dumped back as positive moment in the span because of the rotational stiffness (or lack of) at this connection. If the rotation required to get to the design moment is more than the simple beam end rotation then the moment isn't going to develop no matter what the actual capacity is.

 

[Rancorinco]

Wow! You guys! C'mon, get it together!! Of course, from simple statics obviously it is a moment connection! Consider that each bolt can take a shear, so a vertical column of bolts can be summed up (sum of forces in Y) to give a total shear. If the columns are spread a distance apart, than obviously taking sum-of-the-moments around one of the bolt columns gives shear x arm = moment. A configuration of eccentric bolts used to create moment is called a "torsional bolt group" and can actually be derived from the stiffness distribution method used to analyze shear walls.

Then, having transfered the moment into the plate, the next step is to check the plate stress. If the plate is "t" thick and "h" tall, I = th3/12, and stress = Mc/I = Mh/2I.
Obviously, it can work, and should work well, especially if the moments are low and the arm is wide.

Calculating rotation is more complicated. First, the shear diagram between the bolts (say "s" apart) is a horizontal line, then zeros between the column and the vertical line of bolts closest to the column (say "s0" apart). Then, the moment diagram starts 0 on the outer bolt line, tapers diagonally up to M at the inner bolt line, then is contant to the face of column. Since the rotation is the area under the moment diagram divided by EI, we have theta = M(s0 + 0.5s)/EI.

See page 12-2 of AISC #13 Code, which shows a graph of rotations. This single-plate moment connection is not "simple" nor is it "fixed", but is closer to fixed than simple. Generally, it rotates more than the end-plate connection resisting the similar moment.

We can evaluate how stiff it is as follows. Here the stiffness K = M/rotation. By AISC #13, "stiff" is K = 20EI/L, while "not-so-stiff" is 2EI/L. The stiffness of the connection can be assessed by seeing where the it falls between these two extremes.

In fact, torsional bolt groups are often used to make moment-resisting beam splices. One slaps on a channel each side of beam webs, extending equally onto each beams, then configures a group of bolts to resist the splice moment. I know of several engineering firms who use this detail.

 

[apsix]

We don't know enough about the proposed situation to rule out the proposed detail.
- It requires a very significant connection stiffness reduction to get a significant reduction in the attracted moment.
- Any connection moment due to lateral loading will never be significantly reduced.
- It's a non-issue if the connection stiffness is matched to the column stiffness.