Flexural Capacity of a thin walled round tube

[Everynameistaken]

Hello All,

I am looking for some information on determining both the flexural capacity and the shear capacity of a very thin walled round tube section (a pipe)

I know there is reasonably good information in the various steel codes in North America about the capacity of Class 1 , 2 and 3 sections (Compact etc etc in US). However I am look for information on a class 4 section that is not able to even develop the elastic yield stress and moment?

A little background, we are trying to develop a 'fuse' for a structural piping system. We need to ensure we have enough structural capacity but also able to reliably determine the maximum flexural and shear capacity in order to size the fuse to protect the larger primary more important structure. The intent is to select a thin walled section that will locally buckle and potentially remain in tact enough to pass water through but not transfer any more shear or flexure load.

Cheers,

 

[12345abc6ttyui67]

Isn't a Class 3 section what you describe as 'Class 4'? I've never come across anything referred to as Class 4...

Class 3 sections are typically defined as 'slender' - i.e. they buckle before reaching the elastic limit which is what you seem to be describing.

I have my doubts about the application you are talking about. Wouldn't a couple of slip couplings or similar do the same job with a lot more certainty?

 

[Everynameistaken]

Hi,

No we want a section that does not meet the D/t limits for a class 3. D/t < 66000/fy per CSA S16. This way we actually get the local buckling effects.

Unfortunately, a slip joint system is not possible as the pipe must be able to take the static hoop stress and the axial stress from the internal pressure. So the section needs to be sized just large enough to take these everyday forces and function normally just fine, but in the event of large abnormal conditions, settlements, earthquake etc. that impose additional lateral load to the system the pipe section is fails in a controlled and know manner.

But we need to be able to determine the load at which this local bucking will occur so we can determine if the "fuse" designed to meet the normal conditions is weak enough to protect the existing more import structure.

Definitely not a conventional approach and we are well aware of that, but many many other options have been looked at and seem to meet one objective of either meeting the normal conditions but to strong to protect the main structure or are to weak for normal conditions so they are no good!

 

[12345abc6ttyui67]

I am not familiar with CSA S16, but AISC 360-05 has a non-compact limit for CHS in flexure of D/t < 0.31E / Fy. Anything greater than this limit is classed as slender. Chapter F8 has a capacity equation for "sections with slender walls" - i.e. those with D/t > 0.31E / Fy.

You could design the section as utilisation < 1.0 for design, and adopt > 1.0 for the 'extreme' case. The section should therefore be over-utilised in the extreme case and buckle locally. Isn't this what you want, or do I still misunderstand your intention.

Do bear in mind you have 'hidden' safety factors in the code, and it will probably be hard to determine at exactly what load the section will fail and act as your 'fuse'. I don't know how accurate you need to be? You'd definitely want to test it to see how accurate you are. Also bear in mind steel strength may be stronger than the minimum specified etc. which may all play a part in what load the pipe buckles at.

I appreciate you say you have considered other options, but..... is this really the only way? Very flexible expansion loops? Supports designed to fail instead? Two totally isolated systems with a hose connection between them?

 

[JStephen]

Two tank standards, AWWA D100-11, and API-620, include allowable stresses for thin-walled cylinders. For that matter, I think ASME Section VIII Div 1 does as well.
One issue is that they may be conservative, so I wouldn't assume that they reliably estimated the buckling strength without further research.
One issue is that actual buckling strength can vary significantly due to imperfections. Usually addressed by specifying tolerances and using a large-enough safety factor in the design, but problematic if you're trying to maintain a maximum buckling strength.
Also be aware that buckling of a cylinder is affected by internal pressure in the cylinder.

 

[rb1957]

try this thread ...

http://www.eng-tips.com/viewthread.cfm?qid=268445#

"local buckling of thin walled tubes in bending"

another day in paradise, or is paradise one day closer ?

 

[SAIL3]

what are the maximum operating loads and the minimum loads that would trigger a collapse?...as others have mentioned, the OP is trying to anticipate/decipher the factors of safety, interaction of present forces on buckling, as built condition, etc...unless there is a significant difference between the operating loads and the trigger load, IMO, it is basically impossible to predict....if the consequences of getting it incorrect are not serious, then go ahead...

 

[MikeHalloran]

The intent is to select a thin walled section that will locally buckle and potentially remain in tact enough to pass water through but not transfer any more shear or flexure load.

... That is a functional description of a metal bellows, which is sort of a prebuckled tube.
Further, the circumferential buckles, usually called convolutions, keep the bore open,
whereas a simple tube with a flexure load applied may just fold at the maximum moment,
and folding tends to close the bore, which appears contrary to your design intent.

... About which, I am real fuzzy on what you are trying to do, and why, so please pardon the interruption.

[ I claim no expertise about anything, particularly metal bellows, which usually comprise multiple plies of thin material so as to provide flexibility. The extreme difficulty of forming a multi-ply bellows and joining its ends to ordinary pipe provides some justification for the price and limited availability. ]


Mike Halloran
Pembroke Pines, FL, USA

 

[rb1957]

and later ...
"So the section needs to be sized just large enough to take these everyday forces and function normally just fine, but in the event of large abnormal conditions, settlements, earthquake etc. that impose additional lateral load to the system the pipe section is fails in a controlled and know manner." This comment suggests more destruction than the previous one (ie after earthquake the pipe can be in two pieces).

there can be a reasonable difference between service loading, safe loading, and rupture loading. I guess the key point is that under severe loads the pipe is not so strong as to impart unacceptable loads to the supporting structure.

another day in paradise, or is paradise one day closer ?

 

[JoshPlumSE]

My thoughts:
1) AISI cold formed code is probably a good place to look.
2) The Canadian code explicitly allows "de-rating" of the material to make it a class 3 section. Meaning if it's 50 ksi and Class 4 then if you pretend that it's 42.5 ksi and that just barely makes it a class 3 section then you can calculate the capacity based on 42.5 ksi and you'll be good. To me this option is suitable (via engineering judgment) for just about any code.
3) The TIA tower code (I believe) does something similar to item 2. They have some fairly large diameter monopoles where this may be applicable as well. So, you might check out that standard.

 

[Everynameistaken]

Hi All,

Some interesting inputs.

The problem we have in a little more detail is there is an existing structure that is designed at 100% efficiency and now a new buried pipeline must be attached to it. The pipe is under pressure +/-300 psi and needs to function, statically, in normal condition when no other loads are applied, just pressure and since the pipe is continuous welded steel and buried it functions autstatically to deal with hydraulic thrust. Under seismic there is a large differential moment in the ground, the original 100% efficent structure is rigid and while it can "resist" the movement it cannot except anymore load from this new pipe line, which is connected to it and also subjected to the large movement. So the "fuse" must be able to fail under this condition such that no more load is applied to the main structure. If it can still convey water that would be great, but we understand if the load and movement are high enough it will pinch like a straw as oppose to crumple like a can. But we need to know what these loads might be to see if the fuse is "weak enough" to even protect the structure at all. The flexural capacity should be based on the local buckling effects and i will look through some of the suggestions above. The shear load that will get transfered is just as important and I have not been able to find to much on the shear capacity.

We will have tight control over the "fuse" section, cupon testing and QA for initial imperfections, as best as possible.

 

[12345abc6ttyui67]

It sounds to me like you want to design the new structure / pipeline independently from the existing structure and use a proper isolation between the two. Mike's suggestion of a bellows sounds like exactly what you need.

Not quite the same situation, but we use bellows to 'isolate' equipment nozzles from loads applied from piping systems and to protect the nozzle from overloading. Just add a big stiff anchor support on the 'new' side of the bellows and treat the two systems independently. You will maintain fluid transfer and not have significant forces transferred.

The various different bellows manufacturers will be able to size the bellows to suit the applied loads, pressure and required movement.

 

[MikeHalloran]

Wait; What?

The 'fuse' is supposed to bend without rupturing in the event of an earthquake, so as to 'protect' the 'main structure', which is expected to be deformed or displaced by said earthquake.

But if the main structure is deformed or displaced, can it still perform its intended function, with or without water?

Given that the main structure moves or bends in an earthquake, is it reasonable to assume that the water supply facility and the distribution piping will survive said earthquake, and hence there will still be pressure available to drive water into the protected structure through the flexible pipe?

Okay, not my problem.

Consider an alternative: the water supply pipe is made to normal standards, whatever that means wherever you are, but bent into a big helix and supported at the top end only, so it forms a coil spring. Figure out the spring rate, and design the support to withstand the force applied by the spring, given some arbitrary deformation or displacement of the supporting structure. Then the only problem that remains is finding space for the helix.

To my mind, thin wall pipe/tube is a nonstarter, because it's a bitch to bend without buckling and/or fracturing, in a shop, with fancy tooling to support the walls while it's being plastically deformed, never mind unsupported free bending in air with uncontrolled temperature and strain rate.





Mike Halloran
Pembroke Pines, FL, USA

 

[dik]

Does the attached help?

 

[Everynameistaken]

I will read through that document but the heading sounds good! thanks

MikeH. The whole system is buried and the movement is due to soil liquefaction. The pipe is large and very ductile and will move with the ground in the event of the liquefaction event. This event moves the ground say 15 ft. The main structure is a huge beast, very very still and it does not move the same 15 ft but say 3 ft. Now we have a pipe that wants to move 15ft and is connected to a structure (designed at 100% capacity) that will not move the 15ft. So due to the soils relative movement the pipe is held in place by the main structure and the soil is trying to flow past the pipe imparting huge huge loads, basically full passive soil load. The main structure has been designed for these huge loads from its sail area but not the additional loading from the new piping. Here is our problem!

We need to have a system that under normal conditions can allow water to flow, then when the big one hits the soil moves and the pipe must fail in a way that we can try and quantify such that it does not impart more load on the main structure than it can handle (it will put some load on it in order to fail, how much load is what we are trying to figure out!). The idea is to make a pipe as thin as possible, this pipe takes hoop stress and axial stress for the normal conditions by von mesus we can tune the pipe to be at say 95% under these conditions. Then when the big one hits the soil movement start to occur, the main structure stays put and the pipe starts to move (relative) and starts to impart additional load (the arrangement is such that this movement and load on the pipe wrenches it in flexure), we then want the pipe to fail, in a controlled manner (locally buckling is OK) such that the load imparted on the main structure is as small as possible. We can then fix the fuse section of the pipe and restore flow in say a week, two weeks, if water can still flow through a buckled pipe great, if not then at least we have saved the main structure.

A loop will have too much flexural stiffness (3-4ft diameter pipe) as a spring and impart too much load, and hurt the main structure.

 

[271828]

The SSRC Guide has info on this.

 

[Tmoose]

I'm unclear if there are one or two structures.

Sections of pipe with a coupling at each end can "toggle". Thrust restraints can keep things from going too far. If the pipe, with a couple of toggle sections, was outside the building and centered in 6/10/20 foot wide concrete pit or bunker butted against the building, a fair amount of lateral motion of the a pit could easily be accommodated by the toggling pipe sections. If the pit had a removable cover the pipe could be inspected or repaired much easier.

If you are sure of the direction of soil motion six 90 degree elbows and a few sections of pipe in the pit can accomodate pretty significant motions.

Next to last page here -

https://www.victaulic.com/assets/uploads/literature/26.01.pdf