Slip Critical Joint

[TravisMcC]

Hi, I'm working on some joints, an example of which is attached. I am looking for some calculation or formula for calculating the resistance or ability of these bolts to resist slip from a moment and shear load.
The plain shear load is fairly easy to figure out, but throw in the moment and I cannot for the life of me figure out how to relate the developed friction to the moment.
In the example the moment is centred within the tube welded to the flange. The flange is bolted to another flange which is firmly secured. Bolts are all the same size. Holes are laser cut with generous (.030 clearance). Bolts are pretensioned using a torque wrench.

 

[Bagman2524]

Sounds more like a case of tension and prying caused by the moment than slip. The following are reference sections from the AISC 9th edition on slip critical connections:
A3.4
J1.10, J1.11
J3.2, J3.4, J3.6
Table J3.2
J6
Commentary J1.10
Commentary J3.4

 

[TravisMcC]

Well, i suppose im lacking the information in the drawing. Also I may have posted this in an incorrect location. This is very much not a static structure. It is a mechanical component, and it is secured, very securely at either end. The moment appears to be generating enough force to cause slippage, which then lends itself to breaking the plate across the centroid of the bolts (we move from friction locked to a bearing type connection).
I also haven't any easy access to the AISC manual.

I do not believe prying or tension to be significant as the flange is quite thick. Also the plate to which it is bolted too is even thicker. Shear is also very small relative to the moment.

 

[BAretired]

You first need to find the shear in each bolt. There are two methods of doing this.

The elastic method assumes that all bolts carry an equal shear, i.e. V/n where V is the applied shear and n is the number of bolts. In addition, each bolt carries a shear normal to the radius between the bolt centroid which is proportional to its distance from the centroid.

The resultant shear on any bolt is the vector sum of the two shears calculated as above.

http://www.bgstructuralengineering.com/BGSCM/BGSCM004/BGSCM00403.htm

The other method is the method of Instantaneous Centers which is difficult to explain easily. The elastic method is easily understood and is conservative.

BA

 

[TravisMcC]

I understand how to do those calculations and they are present in my available texts (although that website just found a home in my favorites). What i need is a way to design or verify the design of bolted connection which will not slip. The bolts are not shearing, the plates are breaking.
I am of the mind at this point that if the bolts were able to generate enough friction to keep the flanges from rotating(slipping) relative to each other, then the plate would not break.
Basically i believe as this was originally designed it is a slip critical joint. Problem being that this is a mobile peice of equipment that has serious dynamic loading present. This makes stress calculation a guess at best.

 

[dhengr]

Travis:
You might be asking the wrong question. You shouldn’t be missing any info. in the drawing, you posted it, it’s your drawing isn’t it? But, there certainly seems to be plenty of pertinent info. which has not been brought forth yet. How does the moment load the plate, does it twist the plate loading it torsionally? How long is the hidden tube and does its far end move up or down, thus twisting the plate out of plane? Is the moment induced by a torsional loading from the hidden tube, thus causing the top edge of the plate to be in tension or compression and imparting an up or down shear force in the 3-bolt group to the right? Add the various loads to the sketch, what’s the plate thickness and material spec? What size and grade of bolts are you using? Show the failure line on the sketch and give some description of how and when it happens. How big is the other flange pl. and what does it look like in comparison to the one you show? On heavy mechanical equip. it might not be a good idea to count only on the bolts to take this constantly varying loading, since movement will inevitably start to occur. Why not add some strategically placed shear lugs around the flange you show, and welded to the other flange pl. to lock your detail in place rotationally. Maybe a thicker flange pl. is in order. I wonder if you have analyzed the loads which are causing the problem correctly.

 

[nutte]

Is this an existing piece that you have, that has broken?

AISC and RCSC tell you how much slip capacity a bolted connection has, based on the bolt pretension and faying surface conditions. You use these values in place of the bolt bearing values. If you can solve the problem for bearing bolts, it's the exact same process with slip critical bolts, just with a different (lower) bolt value.

 

[BAretired]

See 13.12.2 "Bolts in Slip-Critical Connections" in CAN/CSA S16-01.

BA

 

[TravisMcC]

dhengr,
The hidden tube is 85" long,
on the opposite end is an identical plate to what you see.
The moment is 27400 in-lbs generated by the force of a slurry on a large panel which is attached with structural uprights to the bottom of the cross-tube (uprights not shown).
I am considering that the moment is acting dead centre on the tube (not entirely true, but a good guess no?)
Other flange plate is a piece of 3/4 x 3 44w with 2 drilled holes and 3 tapped. bolts are grade 5 with hardened washers. Torqued to around 70% proof load.
I guarantee the loading analysis is little more than a good guess. What i'm dealing with basically defies calculation, what i'm trying to do more than anything is verify mathematically the source of failure so we can best learn going forward.

Nutte,
That is exactly what i'm looking for. Either an accurate estimation of slip capacity of this bolted connection, or a way to relate friction to the moment. Where would someone who doesn't regularly use AISC information find something like that?

BA, I guess that's a reference to the above, i will look for it. We don't have anything like that where I'm at.

 

[BAretired]

It is contained in the CISC Handbook of Steel Construction published by the CISC. You should have it if you are working in Ontario, Canada. It costs about $200 if I remember correctly.

BA

 

[TravisMcC]

Wow, I've been on this forum for less than a week and already 2 books i need. I like this forum. Only wish that CISC book was dual unit, but oh well. conversion time. or maybe i can convince my boss we should switch to metric.

 

[wannabeSE]

The AISC and RCSC references that nutte mentioned are online:

http://www.aisc.org/content.aspx?id=2884

 

[paddingtongreen]

My guess is that the plate is breaking around the middle bolt, perhaps fatigue is involved. I don't think the plate is stiff enough to carry enough of the load to the three bolt group, so the two bolt group slips because it is carrying more than it's fair share.

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

 

[dhengr]

Travis:
The 27.4"-k is the torsion imparted to the flange plate from the tube? What are the fixed end moments and shears about the x & y axes at the joint btwn. the tube and the flange plate? How much does the tube deflect in the x & y directions under load? I suspect it’s a fatigue problem too, and I’m betting on a line through the middle bolt and the change in direction on the bottom edge of the flange. Show us a jagged line on the sketch representing the crack line. Are the holes beat up from bolt bearing when you take things apart? I think slip-critical bolted connections on a piece of mechanical equipment which sees hundreds of dynamic loadings per day or hour may be different, and act differently, than slip-critical joints on less dynamically loaded joints such as those on bridges and buildings. I don’t think I would depend only on 3 bolts in tapped holes the way these are being loaded in that joint.

 

[TravisMcC]

well, there are 5 bolts. Also hundreds of this exact design working in the field with no issues. Failures are not common, however have happened.
My understanding of slip critical connections is that when properly designed and assembled act to limit fatigue stresses by spreading the loading across a large clamped surface. This may be a misplaced understanding but hey, i am new.
Shear is simply the product of 3500lbs broken down into vector components. Also this weight is carried by 2 equal plates.
Remember that if the bolts are tightened and strong enough to resist slip between this connection then the strength of the plate is equal to the sum of both plates together as they are firmly fixed to each other. In order for one plate to break and not the other, slip must occur.

 

[BAretired]

Bolts are pretensioned using a torque wrench.

To what tension? Maybe the bolts were not tightened enough to get a slip-critical connection.

BA

 

[TravisMcC]

My fear was stripping the threads as the tapped holes are simply HR 50w (CAN GR 40.21 50W, i know i said 44w above, i always assume 44w when in fact our suppliers use 50w minimum).
The torque i believe relates to around 70% proof load.
It is in fact 200 ft-lbs. Its been a long time since i did those calculations, and i know i relaxed the torque just abit to prevent our guys from stripping the threaded holes.
We have tried to get them to lube the threads everytime with wd-40. Where i am ultimately going with this is that the 2 cases of plate failure i am aware of, both had loose bolts. I want some evidence that suggests if we maintain proper torque and faying surface prep that we can achieve satisfactory results. Also i believe there were assembly errors on these failed parts. It is odd that during both failures, the bolts are loose, and not the same bolts either, but not all bolts. On one the 3 bolts to the right in drawing were lose, and the other the two to the left were lose. Both failed in the same manner. At the same time I am redesigning part to accomodate 6 bolts and adding 1/4" to the thickness of the thinner plate. This all adds cost, but frankly the old addage is correct, an ounce of prevention is worth a pound of cure, every time. The added thickness also helps resist tearing the threads out of the plate.

 

[TravisMcC]

Oh and btw, bolts are 3/4" SAE Grade 5 UNC.
With hardened plate washers. The washers are stamped F436, and are assembled with chamfer facing bolt head.
Bolts are torqued using a 4' long torque wrench, so there is not jerking going on, its very smooth.

 

[dhengr]

Travis:
With the (your product) history, it does seem that very tight control during assembly is real important. Torque is not always a good indicator of bolt tension without this absolute consistent control of all the important factors associated with the bolting process. This can vary from batch to batch of bolts, finish on threads, thread tolerances, etc. Clean the treads with a wire brush first, and then lite lub. Maybe you should be using something like locktight. The exact torque is your best guess, then keep records of results for long term success, since most installations do seem to be working. These kinds of problems are usually one big, long term, experiment; lots of good educated judgement finally solving the problem. And, it seems you’re thinking and heading in the right direction.

But, you still aren’t admitting to, at least, the fixed end moment at the tube/flange pl. joint, due to tube bending and deflection in the horiz. plane, which causes prying in the flange pl. which shows up at the middle bolt and crack. These stresses are added to the torsionally induced stresses and tensile stress from the 3.5/2k shear. The cost of the .25" thicker flange plate is a pittance in comparison to one failure, but do a couple other things too. Make the change in direction (transition) at the bottom edge of the flange pl. a generous arc with tangent points an inch or two away from the current sharp change. Clean this edge up and even grind a small radius on the pl. corners in this area. Ream or grind the burrs (sharp corners) off that middle bolt hole too. If I could, I would make the flange pl. wider in this region, by making the bottom edge a straight line from the left corner to the lower right corner. Maybe move the top edge transition point further right also to improve plate strength (stiffness) in this region. Sounds like the two failures looked about the same, where did the crack start? I don’t know if an extra bolt will solve the problem, but I don’t know where you’re going to put the bolt either. I think you thinking about slip critical joints is about right, re: clamping force, faying surfaces, and bearing area distribution. The big problem with our type of dynamically loaded equip. is keeping the bolts tight under these kinds of loading, and that’s why I don’t like to count on them alone. You can’t pry on this type bolted joint hundreds of times a day without a few bolts eventually loosening.

 

[TravisMcC]

dhengr,
That is exactly my plan of attack for this going forward. I will take some time in the future to try and compare this to the AISC specified abilities for the connection. We do like many things about building this unit in this manner so I want to ensure that going forward with new designs I can make some more "scientific" evaluation of how the joint will perform before we build it.
Obviously the most efficient method is to ensure that the bolts do not enter bearing and remain non-slip, as that is the only way to ensure fully spreading the load across all bolts.