Wind brace connections - using bolts and welding for vertical load transfer to column?

[abusementpark]

See the attached detail of a diagonal wind brace connection in which the gusset plate is directly welded to the column and beam and the beam is connected to the column used a bolted connection. I was recently wondering about the common assumptions regarding the transfer of the vertical component of the brace force to the column.

If you're using the Uniform Force Method (UFM), then some of the load will be transferred directly from the gusset plate to the column and the remainder will be transferred from the gusset plate to beam to the column. The actual force proportions are determined by geometry. However, I was thinking about how AISC has a general statement either prohibiting or limiting how much welds and bolts can be combined to transfer a force in a connection. I believe the concern is differing load-deformation relationships and ductility between the two transfer mechanisms. Wouldn't there be a similar concern in this type of connection? Seems to me like the welded connection of the gusset to the column offers a significantly stiffer path compared to the beam-to-column connection.

Any thoughts?

 

[ToadJones]

I would not use this type of connection for the reasons you have mentioned.
The field welded connection will take the majority of the load because of fit-up reasons on the bolted connection.
If the bolts were field reamed and used an interference fit you make get something close to theoretical load sharing but otherwise this is just a bad detail. I would think that the column to gusset weld will take the overwhelming majority of the live load shear forces on the beam after install.

 

[SteelPE]

The connection you show if very similar to figure 3-3 of the AISC seismic design manual. While I agree that there may be some issues it appears as if the AISC doesn't agree... or at least not for this example.

 

[paddingtongreen]

If the bolted connection is non-slip, I see no problem, but as bearing bolts I would not count them.

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

 

[aggman]

I am not sure of the exact section but you can make the assumption that the gusset to column connection is much stiffer for the vertical load than the gusset to beam connection (because of the bolted shear connection in the beam). This is one of the specific cases of the UFM. This will force the gusset to column connection to hog all the load and induce a moment to balance he connection. I would not count on the beam shear connection to carry much of the load myself. Any deformation would force the gusset connection to try and carry the load.

 

[abusementpark]

I would not use this type of connection for the reasons you have mentioned.

But isn't this connection commonly used? Do you think those structures are in danger?

I wonder if the load initially goes to the gusset weld to the column, but as that weld gets overloaded, there is some redistribution to the bolted connection. I believe welds loaded longitudinally have good ductility.

 

[nutte]

But isn't this connection commonly used?

I don't think I've ever seen this connection type used. Usually you have the same connection type at the beam as you do at the gusset: shear tabs, clip angles, end plates, etc. Sometimes a connection that was to be bolted winds up being welded in the field, due to fit-up problems. In this case, the capacity of the bolted connection has to be reduced in accordance with the AISC specification section concerning bolts and welds sharing load.

 

[port125]

The reference to the AISC seismic design guide bySteelPE got me curious. I had always thought that load sharing between bolts and welds is not allowed because of dissimilar stiffness.
It looks like the fig. 3-3 is showing the shear tab for erection only. After the connection is erected then the beam web is welded to the column using a CJP weld the shear tab becomes the backer bar. Basically the detail shown by Abusmentpark in the OP appears to be OK by the AISC if the beam web is CJP welded to the column after erection, but not as proposed with bolts and welds load sharing.