Torsion on Steel Connections 1

[LeonhardEuler]

Can anyone point me to a good reference for find torsional force on steel connections.

I have bracket connections which will support longitudinal, lateral, and vertical forces. I am sure there will be torsion on my welded connection, but I am unsure how to find how much, because there are connections on either side of each and the load sharing, I would imagine, greatly reduces torsion.

I understand this is probably basic, but I am stuck.

Thank you

 

[WARose]

I've looked for that myself for years and not found much. Based on my own studies, I know that transferring any significant torque with pinned/shear connections is just about impossible. With flange restrained connections, you just have to break it into components and follow the load path.

 

[LeonhardEuler]

Pretty much the entire back side of that bracket is CJP welds. I know it is resisting torque. I just don't know how much.

 

[LeonhardEuler]

Is this as simple as drawing an FBD of the structure looking straight at the bracket and finding the moment at the point i'm looking for?

 

[canwesteng]

It looks like you're designed the runway beams for a bridge crane - typically these are connected at the bottom of the runway beam to take tractive loading along the length of the beam and vertical loading from the crane, and then at the top of the runway beam to take lateral loading occuring at the top flange of the runway beam. Torsion is not significant at either of these locations, but occurs in the column due to tractive loading not being centred on the shear centre, and depending on how you've analyzed the runway beam, may need to be considered there as well.

As a side note, I would not weld these connections if these are in fact runway beams. Typically bolts are used, with slots where the engineer wants to release the load. As the loads are large and cyclical, it's better to have a determinate load path. If the load path is indeterminate, the actual load path will vary based on stiffness. Typically this is not an issue, as steel has enough ductility to mold to our assumed load path, but overstressed welds cyclically overloaded are prone to cracking and brittle failure. Similarly, the runway beams are typically not made to be continuous, such that the structure is determinate and there is no worry about cyclically overloading members. I can share some typical details if you like.

 

[WARose]

Is this as simple as drawing an FBD of the structure looking straight at the bracket and finding the moment at the point i'm looking for?

Probably. But it depends how stiff some of those "supports" are and what is beyond. For example, you have something going to the top flange and it is pinned supported (which means it is modeled as infinitely stiff relative to the beam's weak axis). Well, is that accurate? Because if it isn't.....that could change things. Same deal with that haunch support on the bottom flange.

I know it is resisting torque. I just don't know how much.

That gets answered by considering what I discuss above.

 

[LeonhardEuler]

It looks like you're designed the runway beams for a bridge crane

It was designed in the 1960's and I am analyzing it for a 5 ton upgrade to the overhead crane. Do you have a reference I can quote to show that torsion isn't significant? This was my hunch at first, but I don't feel that I can leave the torsion force on the welds as 0 without a reputable source to back it up such as AISC.

 

[canwesteng]

Can you post some snapshots of the actual detail over the supports if it's existing? You won't find any AISC clause telling you to ignore torsion, you will need to work out solid load paths for wheels loads in each direction (vertical, tractive, sideload)

 

[LeonhardEuler]

I have an original detail, but I'm actually not sure if I'm allowed to post it. Everything is labeled "proprietary" on the originals

 

[LeonhardEuler]

After asking this question the answer has kind've bumped me in the head that I need to just make an FBD and solve for the connection moments.

 

[LeonhardEuler]

Would my fellow forum members consider this an accurate enough approach?

 

[canwesteng]

I'm not really sure what I'm looking at, I couldn't really say one way or the other. Can you make a simple hand sketch of the connection in question? That moment diagram looks like it's for the whole runway beam

 

[LeonhardEuler]

The moment diagram depicts two runway beams. I chose the largest moment at the support as my torsional moment for connection calcs. Although the runway beam itself is attached to the bracket with bolts, so I don't believe there will be any torsion in the connection (end moments 0); however for conservatism since my weld passes anyway I was going to model it as a full fixed connection and put that in my calc package.

 

[LeonhardEuler]

Here is the connection info. I believe I could leave torsion as 0, since the bolts can be ideally modeled as pinned connections. The runway beam is connected at top and bottom using vertical bolts.

 

[JAE]

With the (4) 1 inch diameter bolts connecting the crane beam to the bracket, with a continuous runway beam over the bracket, you have a level of fixity there such that when the crane is on one span, the beam will have some level of rotation that will occur at the bracket.

This will cause torsion in the bracket about the axis that is perpendicular to the runway beam and column.

For bridge crane design, you also have longitudinal forces and these will cause lateral bending in the bracket at the same time. This will cause torsion in the column.

A lot going on here with moving loads, etc. The connection looks hell-for-stout but with a crane upgrade beyond its original intent you have to check it.

Also - with all those welds, and an "old" bracket, you might have a possibility of fatigue issues present - have you had the plates and welds inspected for cracking?



Check out Eng-Tips Forum's Policies here:
faq731-376

 

[LeonhardEuler]

I have only visually inspected the welds myself and did not see any obvious cracks forming. This crane runs about once per month and does not operate near max capacity.

It's an interesting thing to think about that the connection can offer some moment resistance, but how much cannot be readily determined. Unless there is a method I'm notaware of, which is very possible. If there is I would love to read some material on the topic

Surely it will be conservative if I assume a fully fixed connection and calculate the bracket welds based on the full torsion present. It seems that my FBD should be sufficient without drawing the whole structure, because the only thing that attaches to these brackets are the runway beams. Correct me if I'm wrong in this thinking.

 

[JAE]

You can’t always see cracks. Sometimes you have to use a dye penetrant to find them.

I think you are correct that an exact solution would take time in that the torsional stiffness of the bracket gets weird when you torque it by only a portion of its section ( ie the top flange).

And that torque load application suggests also that cracks may form in the web just below the seat plate.

One load application a month isn’t much but that frequency may have been quite different back in the 1960’s.

Check out Eng-Tips Forum's Policies here:
faq731-376

 

[WARose]

Ok so we are talking about torsion on the bracket/haunch from strong axis bending of the runway support beam? And you are wondering how much torque will develop because of the end connections involved?

If that is the case, I misunderstood what you said: I thought you were asking about torsion developed from a lateral load (applied to the top flange of the runway beam & perpendicular to the strong axis of the runway beam).

But to answer your question (under the assumption that is what you are after): in calculations I've typically seen that treated like a simple support. However, there will be some fixity so you may want to check it out. What I would do is figure the rotational stiffness of that support (taking into account the stiffness of the bracket, flange of the bracket where the bolts connect [maybe the weakest link], etc), and use that fixity in your model of the runway support girder.

Chances are you aren't going to get much restraint (and ergo, little torque transferred to the bracket from those forces).

 

[WARose]

It's an interesting thing to think about that the connection can offer some moment resistance, but how much cannot be readily determined. Unless there is a method I'm notaware of, which is very possible. If there is I would love to read some material on the topic

There is a lot of stuff out there (i.e. AISC Journal articles, AISC Design Guide 8, etc) that deal with PR type connections. (Although my money is on this thing behaving closer to a simple connection than anything else. A lot of people don't realize that your average simple shear connection has some degree of fixity as well.)

I've dealt with similar situations before myself......not just with brackets but (for example) running a beam over the top of another. I dealt with it as I outlined it my post above.

 

[LeonhardEuler]

To build on this post I also have a concern with the strength of the bracket plate. I have analyzed it as a stiffened seat connection using calculation procedures from blodgett and salmon. However, it is "failing" due to buckling.

The problem is this isn't truly a stiffened seat connection due to the web and flanges of the beam being welded to the column. Could I justify reducing the eccentricity to the middle of the w18 instead of using 26 in and assuming load is transferred to the plate as a uniformly distributed load?