Design of Moment Connections

[redhead]

As part of my small practice, I often design moment connections for steel fabricators, where the EOR has delegated that responsibilty to the steel fabricator. I am often faced with the problem that rather than give the design moments from the structural analysis, the EOR has a blanket statement requiring the full moment capacity of the beam to be developed, even though the connection is only for lateral load resistance.

Many of these situations would result in the moment capacity of the supporting column being greatly exceeded, especially for an exterior column connected on the y-y axis. It is my opinion that this is not an acceptable condition. I would appreciate your input.

Redhead

 

[SperlingPE]

I would agree.
Standard practice in my neck of the woods is to provide un-factored reactions at the joints to be designed. A detail showing the required/suggested type of moment connection is usually also included.

 

[SlideRuleEra]

Be careful when considering standard details; I am working with the Owner on a project where the EOR furnished a detail for moment connections, which the steel fabricator supplied - problem is that the detail did not meet the building code because of local seismic design requirements. All the steel connections had to be upgraded (at additional cost to the Owner), after erection.

 

[JedClampett]

I would be concerned that if the connection does not develop the full moment capacity of the member, there will be an impact on the analysis. Normal analysis (at least the ones I do) of moment frames assume a fixed connection between the members. If the connection is less than fully fixed, the analysis is not correct. There will be different moments and deflections in the frame. I doubt anything will fall down, but there's more to a moment connection than just the forces, there is also a stiffness component.

 

[UcfSE]

Even if the moment capacity of the beam is fully developed it will not behave as an ideally perfect fixed support, i.e. there will be some rotation however small it may be. I do agree that we should make the connection as close to our design assumptions as possible.

I tend to agree that these blanket statements for "developing the full moment capacity of the beam" are improper. Does that mean develop Mp necessarily? Here at my firm I've run into opposition with this when I say we should put reactions on our plans or design the connections ourselves, but the senior engineers here like to do it their way. When I ask about it, "that's just how we do it, we've always done that" and it isn't going to change. I just design the connections myself and that works for me. Typically we use a full penetration weld at the flange connection for moment, single plate for shear if possible and that takes care of it.

Generally I would expect that you would design for either the Mp of the beam or the Mp of the column, which ever is smaller while maintaining the fixity necessary not to alter the frame analysis assumptions too much. Afterall it seems that most beams are governed by deflection, or soemthing other than their moment strength so it doesn't make sense to have to design for the full moment capacity of the beam if you can provide the fixity required. Still, if you are not the EOR you have to follow his rules as best you can even if they are an ignorant way to do things, he is the one in responsible charge for the building.

 

[redhead]

Thanks for the responses so far. Let me clarify two things.

1. UcfSE: I am required to sign and seal the design of the moment connections, so I am not off the hook as far as liability is concerned.

2. JedClampett: The beams have been designed and sized as simply supported. The moment connections are purely for wind and seismic (in New Jersey). The seismic requirements are based on R=3.0, i.e. the special AISC requirements are not impacted. The lateral load frame analysis performed by the EOR should have produced actual design moments, which they are either reluctant to release, or can't bother to retrieve.

Redhead

 

[UcfSE]

It's my understanding that even though you S&S your engineering the EOR is still liable as well because he or she is the EOR. Is that correct?

 

[redhead]

UcfSE:

Yes, the EOR shares liability. But why should I incur it in the first place with a questionable design?

Redhead

 

[LPPE]

Just as you might put a reaction on a drawing at a beam for the fabricator to design the connection to resist that force, why not put the moment on the plan for moment connections?

 

[apsix]

Connection design using the moment capacity of the smaller member (column in this case) will always be safe. Designing for the larger is overkill in my opinion. The appropriate design shear will depend on the joint confdiguration.

The Australian Steel Code AS4100 stipulates a minimum connection moment capacity of 0.5 x member moment capacity, for rigid frames. There are similar rules for other connection types. These rules ensure a reasonable level of robustness.

As for the effect on frame behaviour if less than full capacity is provided at the connections: I analysed a portal frame with and without reduced stiffness at the joints (3 no. total), subject to horizontal and uplift forces. Reduced stiffness was modelled as 43% of member stiffness over a length of 10mm. This resulted in 0.5% greater sway deflection and 0.3% greater peak moment, negligible increases I think you would agree.

John

 

[UcfSE]

I would hope there isn't any questionable design on your part. I don't think I understand what you meant exactly. Either way, I agree with you that in many cases conections are not done correctly by the EOR and I have seen it first hand at the firm where I'm at. Basically it's lazy and ignorant to try to pull the whole scope of connection design, by those responsible of all people, into some simple phrase that scarcely covers a paragraph and is as clear as mud. There's a lot of attitude out there that claims "we've done it that for for 10^6 years therefore it must be right." I agree that designing for the moment cap. of the beam is often overkill and can make it appear as though the column is underdesigned and that the EOR who uses such unclear notes and details, if any at all, should wake up and find a better way to do things.