Connection Eccentricity

[slickdeals]

Folks,
I would appreciate some input regarding eccentricity considerations in this truss (see attachment).

The truss is comprised of Wide flange chords (web horizontal) and wide flange webs. The connection between the chord and the diagonal web consists of CJP welds between the chord flange and the web flange.

At locations between panel points, beams frame into the chord inducing bending in the weak axis of the chord. As a result, WT vertical members were added at the top chord to carry this load in compression to the bottom chord and distribute it to the diagonals. Obviously the WT is not concentric with the centroid of the truss members.

The connection between the chord and the WT is designed only for axial load. There is an axial connection at the bottom chord to prevent rotation and to brace the bottom flange at every panel point and half way between panel points.

The question is regarding the load path in this connection.

The way I see it is that there is an eccentricity between the reaction point (R) from the beam and the centroid of the WT producing a moment in the weld between the chord and the WT. However, this moment can be distributed back into the diaphragm at this location. Right? This would then leave the WT loaded with a compression = R.

Once the load gets into the WT in compression, the joint where the 2 diagonals and WT frame is subject to a torque(?) = (eccentricity between the WT and centroid of the truss * axial force). How does this force then get resolved between the 2 diagonals and the bracing member framing into the panel point?

I would appreciate a review of the assumed load path and tell me if there is an error in the assumptions.

Sorry for the long drawn post, but I can sure use the wisdom of you folks in this one.

 

[BAretired]

slickdeals,
I believe that the WT will have a moment varying from maximum at the top to zero at the bottom chord.

The only way to stress the bracing beam axially is for the WT to carry moment.

BA

 

[paddingtongreen]

Because of it's one sidedness, the truss is analogous to a channel, there will be shear lag out through the chords. If the load were applied at the beam connection, it would be approximately at the shear centre, but it looks as though it will be concentric on the beam.

That's a long winded look that leads to agreement with BA but that could be a nasty, out of plane, moment on the weld group.

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

 

[hokie66]

I agree with BA. The WT needs to be able to carry the moment at the top due to the reaction eccentricity. So I don't see the connection you show at the top as being adequate.

 

[slickdeals]

Sorry if this sounds stupid, but won't the composite slab and the chord help in taking this moment out of the system prior to the WT seeing it? There are studs welding to the web of the chord and reinforcing bars interfacing the slab and the chord.

Maybe I am not seeing/understanding the load path which is obvious to you guys.

 

[BAretired]

If you connect all of the WT members with a hinge top and bottom at the cg of the WT, they would each carry pure axial load. The eccentricity of R would need to be carried by torsion in the composite top chord.

The flexural capacity of the composite slab would reduce the torsion so slightly that most designers would ignore it.

BA

 

[slickdeals]

@BA:
Please review the attached sketch and let me know your thoughts. Thanks.

 

[BAretired]

You are showing a number of possible load paths.

1. Design shear connection for a moment of R*e1 (plus a shear of R). Ordinarily, the line of bolts is deemed to be a hinge even though it is not a true hinge. The bolts do not fit tightly in the holes and the outstanding legs of the angles are normally kept intentionally flexible, permitting an unknown rotation at that joint.

2. Design weld at WT and chord for P*e2 (plus a shear of P). This implies more accuracy than is warranted in my opinion. How thick is the flange of the top chord? Can it take the required moment without excessive bending?

You ask:

If rotation of top chord is prevented for live loads by slab framing into top chord, can the shear tab be designed for R(live)*e3?

I presume you mean R(dead)*e1 + R(live)*e3.

The slab cannot prevent rotation of the top chord...it can only reduce it. I don't think you would have a problem if you designed the shear tab for R*e3 with or without the slab. If the applied load to the truss is offset by (e1 - e3), this can be accommodated by a small horizontal reaction top and bottom.

The moment going into the WT will depend on the relative stiffnesses of the various parts. There are several possible solutions to the problem, none of which can be said to be "exact".

I am not sure that a WT is the best choice of member to use. You could place the stiffener plate below the web of the top chord and weld an angle each side.


BA

 

[slickdeals]

@BA:
Just curious if you were to design this connection as-is, how would you go about doing it?

 

[BAretired]

I would make the "shear" connection as stiff as possible so it could act as a moment connection too. This would mean selecting thick angles and designing the bolts through the beam web as "slip resistant".

I would extend the WT up to the web of the top chord, cope the flange as required to miss the fillet and weld all around. I would calculate the moment in the WT from a simple analysis considering continuity of it and the floor beam. I would neglect torsion in the chord and I would neglect any contribution from the composite slab.

BA

 

[Teguci]

Have you considered moving the Tee to outside of the truss?

The Tee could act as a short strut between the bracing beam and the floor beam. The vertical load can be placed at the bottom panel point of the truss with a large shear connection and any eccentricity can be accounted for as a lateral couple between the top and bottom chords of the truss.

 

[BAretired]

Teguci,

I like your idea. He could shop weld the flange to the top and bottom chords, then use the web as a shear connector for both the floor beam and the bracing beam. It would be a one-sided connection, so flanges would have to be coped on one side of each beam.

BA