Full Moment Connection

[structuralex]

An I-beam requires a bolted full moment connection to an I-column. I told my friend it requires an end plate with bolts to the top and bottom flange, as well as angle cleats/bolts to the web.
Whereas my friend is insistent that it's fine to just cleat the beam off the column and fix with a couple of bolts, since the moment transfers through the bolts via push and pull of F=M/d where d=distance between the bolts.
How do I explain to him why he is wrong?

 

[jdgengineer]

Both connections will transfer "some" moment. It's all a matter of the stiffness of the connection and the strength of the connection. If your moment arm is very small, small amounts of bolt slip will result in large amounts of angular rotation which will limit the fixity the connection provides. Additionally, if the bolts are not attached to both flanges of the I-beam then the entire I beam is not effective is transferring the moment. Therefore, you would not be able to use the entire section modulus of the I beam for moment calculations.

Typically, in design practice a connection is only considered a moment connection if both flanges are attached. That's not to say that other connections don't have some moment strength/stiffness but instead that they are typically neglected. Even a simple shear tab has some fixity associated with it. In some seismic retrofits, this additional fixity is counted on to classify the connection as a "partially restrained connection"

 

[BenAustralia]

If the cleat is thick enough and the bolts have enough shear capacity, it could transfer a fair amount of moment.

It will, however, rotate a bit before it does this.

With an end plate setup, with bolts acting in tension, very very little rotation is allow in the joint.

It's all about rotations and displacements.

 

[JoshPlumSE]

moment reversal.... test results.

The bolted end plate moment connection is very well tested (see AISC Design guide 4 on the use of these moment connections for seismic applications and AISC 358 with authorizes its use for Special Moment Frames). I have not reviewed it personally, but there is supposed to be a lot of testing that goes into justifying moment connections for Special Moment Frames.

I have never used angle cleats to create a moment connection. Though it strikes me as being similar to some of the "partially restrained" moment connections described by AISC. If so, they don't exhibit very good moment-rotation behavior in general. Though it may be sufficient for low demand moment connections (like you use for non-seismic and non-hurricane regions), but the current testing does not allow them to be used for seismic applications here in the US.

 

[gobsmacked]

It's important to distinguish between flush end plates and extended end plates. Extended end plates can definitely be designed to have moment capacity equal to the beam, usually with a group of 4 bolts around the tension flange. You need to design it so that end plate plastic folding precedes bolt failure. Flush end plates (not extending beyond the flanges of the beam) CAN be designed to resist the full moment but the result is cramped and not good practice unless you have very good quality control all round, and space to operate torque wrenches. These connections have at least 20 different failure modes (See book 'steelwork connections' by Cheal and Owens for additional information and worked examples)

 

[jdgengineer]

And JohnPlum brings up a good point. If you are in a high seismic area (like me) all bets are off. You need to use a pre-qualified connection if this connection is part of the seismic force resisting system.

 

[ztengguy]

Ask him how much the beam needs to deflect before the bolts go into side bearing, if that is what you are talking about. I would imagine its quite significant deflection.