Cantilever attached on bottom 1

[Jazzy47]

Hello,

I am a structural EIT and am used to seeing cantilevers that attach to the side of the column etc., so it is just necessary to find the end reactions. Well, now I have a cantilever w beam which sits on top of a pipe, and I need to find whether the weld from the beam to the pile cap is sufficient. Is this more of a bearing stress problem? I attached a very very rough sketch. The pipe is 28" OD X 0.465" WT, the beam is a W24X104, the stiffeners are 3/8" plates and the cap plate is 1" thick and 20" diameter. Thank you.

 

[jayrod12]

you'd have to design the welds for the tension caused by the connection. the beam is going to try and rotate around the toe of the pipe causing a tension force to resist the loading. Then tension force is going to be in the range of 200 kips when only considering the 50 kip load.

 

[Jazzy47]

Thank you. I understand conceptually that the cantilever is going to rotate around the pipe and try to pull off of it, but I feel that towards the inner edge the weld might be in compression from the downward rotation and the outside will be in tension as the beam lifts off (is this correct?), and I am just trying to figure out the quantitative calcs for where this happens and the magnitude/distribution.

Thanks!

 

[Jazzy47]

Woops the 10' dimension is supposed to be to the centerline of the pipe, not the opposite end of the beam.

 

[TLHS]

These are really irritating connections.

Unless you start modelling this in some sort of finite element software, you're going to have to sketch out a series of assumptions and ensure they're all conservative.

Assuming you haven't actually prepared these things for bearing and don't have the ability to properly square everything I'd assume all the load transfer is via welds.

You have a few different things to check.

1. Whether you can get the load out of the beam - welds, stiffners, possible shear issues in the web
2. Whether your plate can bridge whatever areas it needs to so that you can get loads into the pile. Depending on the thickness of the plate this might be a bending issue or it might be a bearing/direct tension issue. You may have to sketch things out and decide if there are issues with variable stiffness.
3. Whether you can get the load into the pile - welds, there may be stress concentration issues you want to consider. Can you reasonably activate a weld around the whole distance of the pile?

The first item is fairly standard. The second two can get a little unusual but if you think about your load path and place your stiffeners and welds in places that you can ensure you have a stiff load path where you need it the whole thing can become a bit simpler.

I tend to make a series of overlapping conservative assumptions to simplify the problems Unfortunately, it's fairly difficult to explain them without making a few pages of sketches that I don't really have time for. If someone actually has a documented simple way to do this, I'd certainly love to hear it.

 

[Jazzy47]

Yes, I find these connections irritating as well haha. I find that here I am generally encouraged to make vastly simplifying assumptions and call everything rigid, but I'd like to understand the more accurate analysis. Although I'd love to hear your method with overlapping conservative assumptions as well. I find one of my biggest issues is coming up with those assumptions...I tend to conceptualize things exactly as they are and this can complicate things.

 

[Jazzy47]

Oh and yes, all of the load transfer would be through the welds.

 

[kingnero]

I understand conceptually that the cantilever is going to rotate around the pipe and try to pull off of it, but I feel that towards the inner edge the weld might be in compression from the downward rotation and the outside will be in tension as the beam lifts off (is this correct?), and I am just trying to figure out the quantitative calcs for where this happens and the magnitude/distribution.

similar question to this one:

http://www.eng-tips.com/viewthread.cfm?qid=347175

. see the attachment to the last reply. You'll need to add up the loads due to the moment (M) and due to the downwards force (F).

 

[RareBugTX]

Just a thought. Besides the moment connection, is column foundation fixity ok? Would it be possible to place a brace say midspan of beam to column say @45 or even 60 degrees? That way you would be able to analyze it as pin pin frame and be on the conservative side.

 

[Jazzy47]

Just a thought. Besides the moment connection, is column foundation fixity ok? Would it be possible to place a brace say midspan of beam to column say @45 or even 60 degrees? That way you would be able to analyze it as pin pin frame and be on the conservative side.

I wish! This structure is to be sort of a "Hang man" style support with a turnbuckle on the 50K end for a pipeline rack system which has a pile on one side settling a dangerous amount. It is supposed to provide tension to prevent further settlement, and reaches over half of the pipes, so I don't think there would really be room for such a thing. The foundation is pretty deep and has already been analyzed, however.

 

[dhengr]

Jazzy47:
Is the 28" pipe pile and its cap pl. already in place, or do you have some control over what it looks like and how it is welded? You have to know much more about these conditions before you can come up with meaningful details. And, it is not a particularly good condition to try to take your reactions into the pipe walls, through the cap pl. and its welds to the pipe. And, you should know more about the mating of the top of the pipe wall and the cap pl., is it a good bearing surface or cut free hand with a torch? You might be better off taking that 1" pile cap pl. off, to improve your detailing options. I would want to cut the cap pl. back to a 28" OD for about a 14" long arc under your canti. beam and rework the welds btwn. the cap pl. and the pipe wall as PJP welds. The compressive reaction column/point might be a 3' +/- long channel, with a cap pl. and toes welded to the pipe. The tension reaction point might be an end pl. on the beam, welded to the beam, with particular attention to the web welding to pick up the reaction. This end pl. would extend 2 or 3' down the pipe wall, and it would have a wide vert. slit centered below the beam, with a nice circular termination under the beam bot. flg. This slot would be welded to the pipe wall to pick up the tension reaction. These details have to be sized and cleaned up depending upon what’s actually there.

It seems to me that the basic math here is pretty straight forward, and I think it goes something like this: I’m using a 27" ID as a backspan, so you have a 11.125' beam (10' + 13.5"/12). The compressive reaction on the pipe wall is about 247.22k {(50k)(11.125')/2.25'} and the tension reaction on the pipe wall, at the back span is about 197.22k {(50k)(8.875')/2.25'}, and the sum of the verts. even add up. Don’t forget the DL’s and the lateral stability of that canti. beam. You said that the pipe pile has already been checked, but I would want to be darn sure of that. Then, your primary problem is getting your canti. reactions into the pipe wall without causing it to buckle or yield.

A statement like... “I find that here I am generally encouraged to make vastly simplifying assumptions and call everything rigid, but I'd like to understand the more accurate analysis.” is a rather inexperienced comment. We do this all the time to even begin the make our problems analyzable and solvable. Your assumptions should be conservative, not overly so, and they should reasonably represent the way the structure will actually act. You will find that most of our structures act the way we force them to act by the nature of the way we detail them. And, of course, these details should provide a clean/smooth load path from one element to the next. When we don’t provide this smooth load path in our details is when structures perform poorly or fail.

RareBugRA’s kicker idea probably won’t work because it would induce a punching/buckling reaction in the pipe wall, some feet below the canti. bm. If you have the headroom, I might be tempted to look at a truss form about 4 or 5' high over the compressive reaction, and with a turnbuckle and rods for a sloped top chord out to the 50k load.