Beam To Girder Analysis 2

[lindsbm]

I know you guys can help me on this. I'm having trouble getting something clear in my head on a beam to girder connection analysis. Hypothetical: Let's say there is a floor slab supported by steel floor beams. The floor beams sit atop steel girders, which sit atop the columns (typical situation). For load transfer analysis, I understand this situation. But, what if the beams are welded into the web of the girders, such that they do not sit atop the girders, but instead, the top of the flanges of both members are at a common elevation. Get the scenario? If so, is this connection looked at as pinned or fixed? Also, how does the load from the beam transfer to the girder? I would definitely appreciate the help (young engineer).

lindsbm

 

[Qshake]

Pinned! I don't know if this is your situation but this sounds an awful lot like many of the bridges constructed years ago. Rather than have a substructure unit that supported five girders with five expensive bearings, the girders framed into floorbeams which sat on two columns. Which eliminated 3 bearings. Another senario is where there are several girders (stringers) that sit atop floorbeams that are in turn framed into massive girders. Again, the system was founded on only two bearings and columns.

In general, these connections were shear connections only. Simple beam connections. In most cases, there is a lack of support to warrant fixity - no upper or lower plates on the flanges to prevent rotation. In most all cases, due to out-of-plane distortion, even if those plates existed it wouldn't do any good as the webs were large enough to rotate out of plane without the additional bracing. This bracing would have to have been top flange and bottom flange and may take the form of a space truss. To avoid that the floor beams were simply designed as simple beams with shear connections at the supports.

 

[dlew]

lindsbm,

The typical steel beam to girder connection is throught clip angles connecting both webs. The shear load is carried by the web of the beam to the welds (or bolts), to the clip angles, to the bolts, and to the web of the girder.
This is the most used type of connection in the industry.

This type of connection is considered "pinned". The clip angles are flexible enough, and they do not "fix" the end of the beam.

AEF

 

[lindsbm]

Thank you for your help. I was trying to rationalize how this situation could not be transfering a moment to the girder. I see now how using the clip angles prevents rotation, but if the beam were welded directly to the girder (without clip angles), wouldn't there be a moment transfered? As I understand it, welding creates a fixed location. Is that correct?

 

[Qshake]

You still have the rotational distortion from the out-of-plane forces in the web of the larger girder. So no matter how the beam is connected to the girder if the girder isn't braced against out-of-plane displacement the girder will not resist the moment. There are rules of thumb on the amount of rotation necessary to use the 'pinned' analysis and it isn't much. I suggest that you look at Salmon and Johnson's text "Design and Behavior of Steel Structures" or Gaylord and Gaylord's "Design of Steel Structures". The topic is brought to the readers attention in the design of moment connections.

 

[JSA]

lindsbm,

It's a very good question that you ask. It is difficult to visualize this. Let me tell you my way of understanding. I'm going to assume that you will utilize simple shear plate connections, which are the cheapest.

Engineer's loosely assume these connections are pinned. In practice, for buildings anyway, it's considered that for a standard, simple shear plate connection or any other connection only utilizing the web, you don't realy need to know the actual moment. This is because in most offices, the connections are "pre-designed" in the following manner: the simple shear connection is considered to relieve moment by shear yielding in the connection itself, softening the connection, but still supporting the load. Of course some things will need to be prevented, like breaking of the weld, but for the most part, the connection is expected to yield and rotate, relieveing moment, or at least, limiting moment. Visualize the shear yielding in the connection relieving moment.

You should research some AISC-sponsered papers on this at your local college engineering library. Look for back issues of Engineering Journal, published by AISC. This topic is actually very interesting and you will enjoy it. Good luck.

 

[chandr]

For single-plate shear connection, Appendix C of Manual of Steel Construction: Volume II (Connections) is worth reading. It shows that in the situation asked, even the connection is of simple shear type some moment must be considered in the design of welds and bolts. The amount of moment to be considered was experimentally determined as it is not easy to accurately determine by just doing math. models.

 

[Imagineer]

You could design the beam and the column for their respective worst-case scenarios. Design the beam as simply-support and the column as moment-connected. It's conservative but in reality there will be some fixity which is very hard to quantify.

Imagineer