Combined Welded and Bolted Steel Connection 2

[abusementpark]

If I am making a wide flange beam to column connection and I provided a CJP weld at the flanges to develop moment, am I allowed to provided a bolted web plate connection as well to take the shear force?

I remember reading somewhere that you can't do that, because you may have to fail the welds in flanges to engage the bolts in shear at web.

 

[KootK]

I've read the same thing in the context of seismic design. I'm not sure that it's codified anywhere however.

 

[abusementpark]

KootenayKid,

I thought it was just a seismic thing as well. However, I want to make sure it does not apply to all types of loading.

 

[KootK]

The trouble is, if the stiffness argument holds up for a seismically induced shear load, the same logic would apply equally well to any other post installed shear load.

Perhaps in a static load application, you would fail the flange weld locally at the center of the flange, distribute the shear load to the bolted plate, and still have capacity left in the remaining flange weld? Obviously, under dynamic loading, this would be a prime location to initiate some kind of terrible unzipping failure mode.

Right or wrong, that's still a very common detail in many regions of North America. Could you start off with a bolted plate for erection and then have it welded up after the fact?

 

[KootK]

This article is interesting wrt bolts + welds. It's obviously not intended to cover this particular case of course. Particularly interesting is the quoted AWS clause.

 

[nutte]

Y'all are thinking about this too hard. Yes, you can do it. Yes, it is very common. Refer to the AISC speicifcation, section J1.8. The commentary explicitly states that the prohibition of bolts and welds sharing load does not apply to these types of connections.

I'm not sure if you can do it for seismic connections. I know the FEMA documents show the web bolted, but the more current AISC prequalified connections don't appear to show the web bolted. Regardless, for non-seismic connections, bolt away.

 

[FranzR]

For seismic connections, the bolts have to be slip critical. Check out AISC 341 seismic provisions.

 

[paddingtongreen]

As far as I can remember, friction bolts have been permitted in combination with welds since the beginning of time, or since friction bolts have been on the matket, but only friction bolts. Be clear, welded flanges taking tension or compression from bending are not considered as combining with the web connection in shear.

Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[slickdeals]

Could you not provide a slotted hole in the direction of the axial force for the web? That way it is not taking any axial load (not in combination with the weld) and it can be used to resist shear.

 

[abusementpark]

paddingtongreen,

Do the bolts have to be slip-critical, even for non-seismic loads? If so, where in the code does it say this?

 

[ToadJones]

AISC Spec & Commentary J1.8

 

[KBVT]

J1.8 does not require the bolts to be slip-critical in a typical welded-flange moment connection. I believe this provision statse that bolts can be combined with a welded connection between the beam web and the column/girder with a 50% reduction in bolt capacity (bearing).

 

[nutte]

Abusementpark, nowhere does the specification say that they have to be slip critical for non-seismic connections. So, they can be bearing connections.

Paddingtongreen, the 2005 specification changed the long-standing rule that slip critical bolts in the same faying surface can share the load with weld. This has been acceptable to the code forever, but recent research has shown that it might yield unconservative results. The 2005 specification says that the bolts must be checked in bearing, with only half of their capacity used for resistance, when combined with welds. (Again, this does not apply to Abusementpark's connection.)

 

[paddingtongreen]

nutte, thanks. I would like to have stayed up-to-date with the manuals, but they cost too much to have just for the sake of having them.




Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[nutte]

If you're trying to stay current, the specification is a free download.

 

[iponom]

if you have a Full-Pen weld at the flanges, you are not sharing the load with the bolts. Same as welds aren't sharing the shear from the bolts. Very common detail.

 

[structuresguy]

iponom: I don't think I agree with your statement. There is deffinitely some sharing going on. It may not be the type of sharing intended by teh Code. But take out the bolts, and I garantee the welds will carry the load (up to a point). Typical sequence of construction would be fit the beam in place with 1 or 2 bolts each end. Finish erecting other steel in the area. Plumb up columns, if required. Finish installing remaining bolts. Weld flanges. Install deck and floor slabs.

So at the point just before the welds are made, the bolts are in bearing contact with the shear plate / clip angles. Then the welds are made. From that point on, the welds will share the load with the bolts. We just assume that they do not, hoping that the shear connector is stiff enough to no deform under the shear load imposed after welding. If there is any deformation of the shear connector, then the welds will definitely start sharing a significant portion of the shear load.

 

[iponom]

structuresguy,

Well technically you are right and I agree with you 100%. However, when we detail the connection we design the shear plate/dbl angle for the shear loading (w. bolts), and the flange for the flange force due to moment, and that's that, no combined loading anywhere... I know that most of the stress due to shear and moment combination is closer at the flanges and the Full Pen welding and it bothers me to know that some engineers ask us to take the moment into the beam web when the flanges are failing at the full-pen weld location due to flange forces alone... I can't wait for more technical papers to come out with a design simple enough to make it practical...

 

[iponom]

structuresguy,

And what I mentioned above is the topic of the discussion I opened earlier in the topics. do you have any ideas?

 

[structuresguy]

I should have started with the fact that I agree with you, and others above, that welded flanges and bolted web is very common connection, and does not require slip critical for non-seismic. My bad.

Generally speaking, I prefer to think of structural behavior in terms of "what is really happening" rather than "what we assume is happening". Some goes for pin vs fix connections, particularly column base connections. I would much rather create a 3D FEA model to see what is really happening, than just assume my column base is pinned (or fixed). Of course, time does not always allow me this luxury. But I see too many younger engineers just going on the assumption that their college professor (or text book) told them to use "all the time".