Bold/nut pulling through member

[bsimms89]

Ok so I have a very unique problem and would like some insight on how to go about it. We are currently creating a spreader that has to be capable of pivoting about its center. It will only be holding the load for a brief period and is not a permanent install. It will have to spread a 3' span in the upper section and 6' on the lower and pivot about the middle ( I have attached a sketch, I apologize that it is sideways, I cannot get it to save right side up). The lower section will have to hold 40K on each end. We have sized a HSS 10x10x5/8 for the lower section and a HSS 8x8x5/8 for the upper. To connect them and allow them to pivot we have decided to run a threaded rod through the middle of them with a nut on each end.

We have contacted our threaded fastener supplier and they said they could provide us with a 2.5" diameter threaded rod with a tensile strength of 134,900psi.
The given tensile area of this rod is 4in^2.
so Tensile strength * area = 539,600lbs
and with an omega of 2 it can handle 269,800lbs in tension (yes this is overkill since it will only have 80,000lbs but too much is better than too little for this)

Now I have 2 questions.

First in analyzing the nut bearing on the top of the HSS8x8x5/8 and the bottom of the HSS10x10x5/8 it is a legitimate concern that the nut could pull through the 5/8" thick walls of the HSS sections. We are planning on welding a plate on the top and bottom of these members to increase the thickness of the bearing surface so this will not occur. My question is how should i go about analyzing this. I have searched online to try to find something about bolt/nut pull through and most anything I find says it "typically isn't a concern", but in this situation it is. Should I analyze this as if it was a typical bolt in shear connection and check shear of the base metal and bearing on the connected metal, where:

Rn=1.2*Lc*t*Fu<=2.4*d*t*Fu.
t would be the thickness of the HSS section + the plate we weld to it
Fu is the ultimate tensile strength of the HSS/plate
d= diameter of the threaded rod
and for Lc would I want to take the area of the nut-area of the hole to find the total area bearing on the plate/HSS? or would i want to take the area of the washer that is bearing on them? I would like to know how this is typically analyzed.


My second question isn't as technical but more related to materials. We not have this threaded rod where the supplier provided us the tensile strength and the tensile area so i calculated it to be capableof holding 269,800 at stated with an omega of 2. Now we need to find a nut that is also capable of holding such a load. Our supplier sent us the specs for the nut that is typically used with that rod(it is the same grade) but I am not to familiar with nuts. The specification he provided only had chemical testing results. I have asked him is he can tell us what load this nut is capable of holding and he said he would get back to us. I would have assumed that this would have just been the proof load for the nut(which wasn't listed on the specs). In specifying a nut wouldn't I want to specify one that has a higher proof load than the rod is capable of holding?

The threaded rod was given on the mill test report as ASTM A193 Grade B7 with a tensile strength of 134,900psi.
The nut that they said they typically supply with it is an ASTM-A194 Grade 2H, described as 2.50"-4UNC-2B hvy hex nut, plain
I have looked online and found that ASTM-A194 Grade 2H nuts have a proof load stress of 175ksi, now in specifying a nut is it the proof load stress that i want to be greater than the tensile stnregth of the threaded rod, or am i supposed to multiply this by an area to find an actual proof load and compare that to what I and supporting?

Thank you in advance for any insight to my questions, a lot of what we do isn't typical so sometimes we have questions that aren't so common.

 

[JStephen]

I recall reading a while back that in that famous Hyatt-Regency walkway collapse, nut pull-through was one of the issues. I don't recall if that was actually the failure mechanism or not.

It seems to me that if nut pull-through was an issue, you may also have an issue with the whole top surface bending in as well.

If you look up the bolt specification, it will likely list acceptable nuts to be used with the bolt. I think generally, the nuts would be adequate for the strength of the bolt in that case, perhaps someone can confirm that.

Can you stack the members in the other direction where they just bear on one another instead of pulling on the bolt?

You describe this as a spreader. If in the US, it may need to comply with the ASME standard for below-the-hook lifting devices. In that case, there are some pretty generous factors of safety that get thrown in, considerably more conservative than typical structural steel.

 

[jayrod12]

In Hyatt regency nut pull through was the failure mechanism but it was compounded by the fact that the landing beams were double channels toe-toe with the threaded rod going directly through the splice. It also ended up having to resist double the intended load due to a fabrication/design error.

Either way, in this case I feel as though if you provide a plate washer thick enough that nut pull-through wasn't an issue then it would turn straight into a shear/bending problem for the hss.

I don't know how I'd go directly about checking that, but since plate is cheap I would likely just ask for 3/4" or 1" thick 8"x8" plates and wouldn't really concern myself with it again.

 

[bsimms89]

Yes you are correct that nut pull through was the failure mechanism for the Hyatt. The hanger rod was supposed to support two floors with one nut holding the each suspended floor, but the detail was changed and the nut was holding both floors (twice as much as intended). This cause the nut to pull through the built up box beam. Ive read in multiple places including the link I'm attaching that this problem could have been solved through the use of a bearing plate to prevent the nut from pulling through which is what we are trying to determine the size plate we would need for our application.

http://matdl.org/failurecases/Building_Collapse_Cases/Hyatt_Regency_Walkway.html

In my mind it seems like it could be analyzed as base material shear from a bolt bearing on it, I just wasn't sure what I should treat as the shear plane, or if this is a reasonable assumption.

We are essentially doing the same thing just with a HSS section instead of two channels welded together like at the hyatt.

You are also correct, before you posted I was looking up the specification for the rod and found ASTM recommended nuts for it, I'm assuming as well and would like confirmation that this means that the nuts meet the capacity for the rod.

 

[bsimms89]

Haha sorry jayrod u beat me to it about the Hyatt. But yeah no matter what we were going to go with an 8x8x2" thick plate on the HSS8x8 as a minimum, and a 10x10x2" thick plate on the HSS10x10 as a minimum no matter what any calcs told us(bigger is always better and we don't like to be on the upper end of any limits). I just wanted to know how one would analyze such a situation and I'd like to have some numbers to make myself a little more comfortable with it.

I know u said u don't know how u would go about checking that, neither do I, that is why I am here (I know that a 2" plate should be overkill but I'm also curious of the answer). So the question is still up for grabs.

 

[SAIL3]

this concept has so many problems associated with it that they are to numerous to list....also, unless the loads are perfectly balanced which would be difficult to achieve, one get all kinds of binding etc......maybe fabricating a sleeve connection for each member would bypass all the local failure problems with the HSS members,,,also a spacer between the members....maybe using some standard lift concepts would be more promising....lift lugs,threaded rods with some type of swival mechanism, etc...

 

[jayrod12]

I would be inclined to treat it the same as punching shear in concrete. get a shear perimeter (diameter of bolt would be conservative as the nut would extend it out) and multiply by the plate thickness to get a shear area. I'd then check that stress against allowable shear stress of the steel.

For your bolt size and plate dimensions I get a rough capacity of 300 kips. For the record this is totally a shot in the dark, but if another designer came to me with this calc I would likely accept it. I can't see anything being able to pull a nut through 2" of steel.

 

[MikeHalloran]

You can buy big nuts from Grainger for less than a fancy cup of coffee. Nice ornaments for your desk, and excellent paperweights.

Check local stock before designing in any particular size and wall of HSS.

Doublers are commonly used where a load is applied to the center of and normal to an HSS face, because the face effectively becomes a rectangular plate of indeterminate length supported on two parallel edges, and loaded in the worst possible way.

The Hyatt-Regency's original design used very long threaded rods, with HSS tubes spaced and supported by nuts (and I think big washers as doublers). It would have worked, but it would have been unwieldy to assemble. The design that was substituted without sufficient review used toe-to-toe channels as pseudo-tubes, which might have almost worked, but also segmented the hanger rods in such a way that the rod to sorta-tube connection for the uppermost deck also carried the entire load of the lower decks. It worked until a dancing load was applied. Then it didn't work.





Mike Halloran
Pembroke Pines, FL, USA

 

[jayrod12]

To clarify,

That 300 kips was based off of 2.5" diameter failure perimeter, 2" thick steel and an allowable shear stress of 22 ksi (which is a soft conversion from what I would use for steel here North of the 49)

 

[dhengr]

Bsimms89:
In today’s world, you guys don’t seem to read very carefully for the real, or full, meaning of a post. Jstephen has the right idea in his 4th para., in his 18:12 post. Stack the tubes so they are in bearing on each other, not in tension, and eliminate the thru bolt in tension holding them together and transmitting 80kips. Then, an improvement on his scheme, take 4 - 3x3 angles about 17" long and weld them to the web faces of the two tubes where they cross. This gets that load transfer down into the webs of the tubes where you want it and can rationalize it. You don’t have to try to chase those nasty loads around the corners of the tubes with bending, bearing and buckling problems. Then, you’ll still need thick plate washers at each of the four pick points out at the ends of the tubes, and you have to chase those tensile loads up into the tube webs. This latter is part of SAIL3's point about “so many problems.”

Better yet, why not one 6x6 tube (a piece of 4" pipe will likely work too), the length of the 10x10 tube, which I assume is predicated upon the distance btwn. the lift points on the load. On the ends of the pipe, you weld plates with a hole below the pipe for shackles down to the load; and above the pipe this plate is bent in slightly to conform to a sling angle, with another hole for a shackle to the sling and to the hook. Now the pipe is acting strictly as a spreader in compression. The load goes right through the end plates. Maybe you need to add double doughnut plates at shackle holes for pin bearing reasons and to match the shackle opening width at the pin.

 

[Tomfh]

Damn, I'm a lifetime too late to say "Hyatt Walkway!!"

A good check for pull-through is to do a yield line analysis of the tube wall.

Are you sure the loads will balance ok and that out of balance force can get through the detail?


Note that if you stack the lower tube on top, then it is more unstable, as the loading is above the support point, and thus it wants to fall over.

What exactly are you trying to do here? What are you lifting? What the pivoting for?