Calculating crushing imposed by pipe supports 2

[Alex303]

Hi, I have a 600NB Std Wt pipe that is full with water and supported every 5m. My question is whether or not the pipe requires a seat or whether it can just lay flat on a concrete pedestal? Obviously if it is sitting in a rounded pipe seat it would be better due to greater contact support on the pipe, but I am wondering how much better it is, more so out of interest?

If the pipe is lying flat on a plate or concrete pedestal you are basically allowing the pipe to deform and increase it's contact area on the plate as it does so. This means that the initial stress would be quite high, but then it would peter out as the contact point flattens. I would like to know whether the pipe has yielded in this zone (ie dented the pipe). I would also be interested in whether it was possible for the pipe to buckle in this loading case?

I have looked at table 14.1 on Roarks formulas which kind of describes my loading case, however that is only for a cylinder, not a pipe. I know shell analysis can get quite complex, but I was wondering whether someone had experience on how to calculate such a thing?

Unfortunately I don't have experience using any stress analysis programs otherwise I would just do that

Thanks

 

[C2it]

I think you have pretty much got there already. Pipe support spans are usually governed by limitation of bending stress, mid-span deflection and local stress at supports. Local stresses are usually calculated via Roark's formulas. I don't see any difference between a cylinder and a pipe in this sense. To do better would, as you mention require finite element analysis, and a complex one too, since it would need to be non-linear to account for the contact area change.

 

[Alex303]

Do you think that a solid cylinder can approximate a pipe in this case though?

The Roarks formula indicates a 20mm contact zone due to deformation which seems rather high to me considering the wall thickness is 9.53mm thick (does it to you?), does this mean that the area would have yielded and bent the pipe? Although the stress calculation indicates 1856KPa, this would be after the deformation. I suppose my question is whether or not that 20mm deformation is elastic or permanent (perhaps damaging the coating for galvanised steel) and also whether or not I have to worry about buckling about the pipe wall.

 

[C2it]

Roark section 13 (Shells of revolution) is the normal source for basic equations.

 

[Alex303]

Can't find anything on section 13 dealing with this loading case.

If the loading is 1856Kpa what does that even mean? Surely the buckling load would be a lot less than gross yielding?

 

[rconner]

You may want to peruse the references per discussions in many past threads e.g.

http://www.eng-tips.com/viewthread.cfm?qid=133164

(I suspect you will want a 90-120 degree saddle with some meaningful width to avoid extreme localized stresses, that may occur either axially off the bottom support edge or some distance up/off the "saddle tips" for this/size pipe)

 

[racookpe1978]

Is the pipe sliding or moving at all under (how big?) a temperature difference. Even summer to winter is a measureable temperature change.?

Over time, the erosion (cutting) and corrosion in the grooves and scars will cause even greater loss of metal at the touch points. And the touch points are the greatest stress risers since they are at the edges of each support, where the pipe begins sagging back down.

 

[LittleInch]

The basic answer is no you shouldn't. The key issue isn't a uniform "crushing" of the pipe, it's the stress concentration and potential for denting and buckling at the end of the concrete pier. in many cases, the concrete itself can't stand the edge load and breaks and spalls.

To consider laying a 24" pipe on a concrete "plinth" is nonsensical when considering the miniscule cost of clamp curved supports.

The issue also arises from crevice corrosion and general corrosion on the underside of the pipe which you can't see until it leaks.

see attached if you really want to do some calculations and check it out.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Alex303]

Thanks for all the replies so far. I agree that the loading on the edge of the support should definitely be the critical factor, rather than the 'crushing' seen centrally on the support... At the edge you would have a stress resultant due to the bending moment and the axial force (ie mohrs circle).

The pipe will be fixed to the support with minimal movement and be wrapped in neoprene to protect from wear due to movement so my only concern regarding this is the loading stress.

 

[LittleInch]

Just build it with proper pipe supports/ shoes. Then all your problems go away. Neoprene is not a long term solution.

Something like this

http://www.aaatech.com/catalogs/pipe-shoes/5020/index.htm

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[NBrink]

Oof... Laying directly on concrete plinths and wrapping in neoprene are pretty much a sure path to contact point corrosion.

Shoes will be well worth the minimal investment.

As far as your concerns as to deforming the pipe, I suppose it's a possiblity, but I've certainly never seen it happen. It's fairly common practice to support piping on a steel or thermoplastic half round (maybe a half inch across), and I've never seen a failure or even visible deformation from those. Look into that if shoes are too costly...

 

[rconner]

Alex, not exactly sure what "Roark" formula you are looking at, but I would have expected maximum localized fiber stress at the flat (or for that matter even shaped)supports, even in steel pipe this thick, would have been FAR greater than 1856KPa(~269 psi)! have you looked e.g. at Roark & Young - Formulas for Stress & Strain (7th edition), Pipe on supports at intervals (Sec. 13.7 pg 591 etc)
While I guess I'll let others debate any practical issues of rust etc. of conventional well-shaped and padded vs half-round supports, I think particularly/near infinitely small contact areas along with other common factors inevitably breed higher levels of localized stress, particularly with weak or thin pipes or insulation etc.

 

[Alex303]

Okay, i'm hearing what you all are saying but every calculation I do I don't see any of these high local stresses.

With my 600NB pipe I have 1.410KN/m of steel and 2.690KN/m for water.

Depending on the length of the line and therefore the number of supports I can assume that if the spacing is 4m then the reaction at the support shouldn't be much more than 20KN.

Even if the pipe contact is 0.5mm over a 0.5m support the stress is only 80 MPa and as the pipe flexes this would be reduced even further so I don't see any possibility of undesirable deformation.

Regarding the contact loads at the edges of the support rconner mentioned, that calc gives 18.3 MPa for a 90 deg saddle, well below anything even worth worrying about. So I can't really see the stress suddenly jumping to ~300Mpa just because it's sitting on a flat support...

After just googling 'pipework' in google image search it's not that unusual to see pipes sitting on flat plates or flanges so I'm not sure why everyone is convinced i need them...

Regarding corrosion, I don't see how sitting a pipe on a flat plate rather than a seat causes more corrosive conditions if my stresses are still so low.

Everyone on this thread seems so adamant that this is not the thing to do but nothing so far has convinced me. I suspect saddles are only necessary for cross country \ long length lines where you are placing your supports at maximum lengths.

 

[LittleInch]

Some details of your pipe would be good to see how thick it is etc and how you worked out your figures, especially 80 MPa. Actually just seen it is 9.3mm, which doesn't seem that big.

It causes more corrosion because there is a small gap where water gathers and where small movements create erosion of the steel.

Feel free to ignore literally hundreds of years of experience from the posters on this thread and build your system as you want, but it will probably take several years to show you why you're wrong. Not everything can be calculated.

Good luck.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[rconner]

It is not necessarily just the contact stress that is of most concern but instead also the maximum localized stress in the wall/membrane, that I believe may be located several degrees above the saddle tips, or instead axially off the very bottom of the pipe where it contacts the pier or support. In any case, it sounds like the proper formula for maximum localized stress near supports, at least for common application to larger and more flexible pipes on shaped supports, has now been located (and it has been found that the stress off a 90 degree saddle or shaped support is in at least that case certainly acceptable as I would expect, though I see it is actually about 10 times what was originally reported here at a flat support for 5m span).
I believe a whole lot of factors may enter into the design of some piping on supports systems, and maybe even some e.g. like seismic, vibration and modal effects etc perhaps not real obvious upfront. While I would not state "adamantly" that flat supports cannot be utilized in this particular application involving pretty good-sized and 3/8 inch thick steel pipes, and who knows may even be desirable for some reason(s), and I for one am not going to "worry" about this, one should however be aware for this or future reference that much higher and perhaps even more variable stress concentrations (in the order of a unit of magnitude or higher) and deformation can result between the results obtained at any shaped saddle and those actually near a flat support.
While there is indeed some record of suitable performance of at least legacy steel and iron piping systems in at least the smaller diameter sizes (with substantial ring and beam stiffness) on essentially “flat” supports, larger diameter pipe systems often involve shaped saddles, cradles, or braced hangers that better distribute various loadings around the bottom portion of the pipe. The exact amount of stress in any particular application is dependent on pipe size, pipe wall thickness, type of determinate or indeterminate support scheme, distance between supports, if and how pipes are strapped down, location of supports along the pipe length, loading (including anything inherent to the system and environment, seismic factors etc) and probably many other factors. Readers of this thread unfortunately don't know as much about all this as the OP.
Though I think some more involved formulas addressing these high stress concentrations for cylindrical shells and pipes based on pure point or line localized loads for lesser support schemes have even been published in some technical literature, I suspect one may find more variability and less confidence in same as pipe sizes get larger. Some designers of steel piping systems may go with shaped supports or related schemes like ring girders and others mentioned on this thread etc in some cases as pipe sizes get larger for greater confidence in results or to optimize span lengths etc (therefore building no more piers or pedestals to than necessary for dependable service, and maybe even minimizing total installed cost, minimizing construction time by building less piers or pedestals, and also minimizing clutter to the facility or landscape in the process?)
All have a good weekend.

 

[Alex303]

Thanks for the thoughtful reply rconner.

I've been reading some of the Australian standards and there's no requirement for seats detailed there but it does say:

3.28.1.2 Design loads
Pipe supports shall be designed to withstand the most adverse combination of the following
loads:
(a) Piping expansion and contraction.
(b) Reaction of piping that discharges to atmosphere.
(c) Snow and ice.
(d) Mass of equipment installed to counteract or control expansion, contraction, and
associated reactions.
(e) Mass of insulation.
(f) Mass of the operating, cleaning or test fluid, whichever is heaviest, except that where
the pipe is to be held up with additional supports during testing, the mass of the test
fluid is disregarded.
(g) Mass of the pipes and associated fittings.
(h) Wind or earthquake, whichever is greater.
Where imposed vibration or shock is expected during operation, suitable anchors, dampers,
or restraints shall be provided to remove or reduce any adverse effects.
The calculation of loads for variable and constant supports shall be in accordance with
Clause 3.27.
The design of anchors and guides shall take into account additional forces to overcome
friction in other supports.

It also says this

3.11.3 Shear stress
The value of any primary shear stress shall be appropriate for the material and design
temperature. It shall be 60% of the tensile design strength.

So that means I have a maximum shear capacity of 150MPa.

If my supports were at the specified 12m spacing I think I would be worried due to the combination of bending stress and shear stress at the corners of a flat support but at 4-5m I'm not.

I would have thought that having a pipe on a flat support would actually provide better ventilation to the contact area compared to a saddle. That coupled with the fact I am using neoprene for vibration dampening and abrasion protection I am just not seeing a corrosion risk that wouldn't be there anyway. That said if my stresses were higher because I was using a longer spacing between my supports then I could definitely see the reasoning behind using a saddle to reduce the contact stress.

 

[LittleInch]

Alex,

I can see where you're coming from, but this requires the concrete supports to be initially built perfectly flat and remain that way for the life of the pipe.

You said earlier "the pipe will be fixed to the support worth minimal movement...." How exactly?
How many supports are there?,
what is the plan view?
Will the concrete peers settle over time?
Why don't you want to make the supports farther apart and use a shoe system?

If it looks justified then maybe it will work, but it doesn't look like at the moment.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Alex303]

So I have gone away for a while and analysed this problem to death and have an answer, so for those interested I will share my results.

Using hertzian stress calculations from Shigley's Mechanical Engineering Design which can be found here

http://www.amesweb.info/HertzianContact/HertzianContactCalculationSteps.aspx

it is possible to find the principal stresses developed by the localised contact stress. These principal stresses can be analysed using mohrs circle and are acting in the y and x direction when looking at a cross section through the pipe. The highest stresses induced from contact are developed a few millimetres higher than the contact zone, however when looking at 3D principal stresses, the flexural stress developed due to the bending moment on the pipe is nearly always critical. When this is considered, the highest stress zone is always a shear that is localised to the very bottom of the pipe, at the edges of the support.

However even if the pipe yields, the pipe will self relieve this stress, so buckling is practically impossible provided the support structure is sensible. Having said that, micro cracks will develop in the yielded pipe and the risk of corrosion becomes quite high.

The contact stresses developed are not only dependent on the material of the pipe, but also the material of the support. It is important to not only check the shear capacity of the pipe, but also the support.

Using a design yield capacity of 138MPa which is appropriate for API 5L Grade A steel typically used in my area, the maximum allowable shear is 0.6 times that (82.8MPa). This is very different to what I thought earlier in this thread as I was still wrapping my head around the problem so bear with me. My calculations are based on the UDL generated from the self weight of pipe and water, with no other components attached to the pipe. I found some typical maximum flat lengths with practical support intervals for use on a steel support. The stress found is shown below.

So for example a support length of only 210mm for 12.8m support lengths of DN600 pipe is actually fine (81.5MPa contact stress). When I switch the support material to concrete the stress on the pipe drops off quite a lot due to the lower elastic modulus of concrete. However, note the stress on the concrete is well in excess of allowable shear stress...

When I make the support into a shoe of 615 ID and return it to a steel support the stresses drop off quite a lot.

When I return it to a concrete support the shear stress noted on support for the DN600 pipe is acceptable due to the 615 ID shaped slot.

So in my mind I think that flat supports are okay if you keep the support interval rather short and an appropriate support length is selected. If using concrete supports then shaping the concrete pedestal into a shoe a few mm larger than the OD of the pipe is desirable to stop the concrete from cracking at the edges. If the pipe lengths are supporting valving or other equipment, or if the pipe configuration is not a typical run of pipe (tee pieces etc) then the moments and reactions of the support must be calculated in order to find the contact stress.

That said, it is always better to use a shoe (as expected), but not always necessary.

 

[KevinNZ]

"flat supports are okay if you keep the support interval rather short"

Nothing wrong with this conclusion. Starting with an empty pipe resting on a flat floor, the stresses are low.
Them remove parts of the floor and add content weight the stresses go up until the pipe deforms and sags over the remains of the floor.

In practice steel resting on concrete is going to suffer damage from corrosion and/or wear.

 

[rconner]

In retrospect, as saddle angles reduce say from 30 degrees to 0 degrees (i.e. to the "flat" support apparently wanted in the OP), I believe the maximum localized or principal stresses will become more primarily "circumferential" in nature (and also will more diverge from/and be less conservatively estimated by application of conventional pipe-on-supports localized stress formulae as per Hartenburg, Roark, and that with subsequent basic adoption into such AWWA manuals as M11 and M41 etc).
If your application happens to involve pressurized piping, as of course many do, you therefore may wish to also consider any additive effects of e.g. hoop stress due to at least very high internal pressure, that are of course also mostly circumferential, also tensile in nature and also located at the inside bottom surface of the piping. [Internal pressure will also tend to "re-round" the piping, reducing the contact area you mention early in this thread, but in so doing perhaps obviating additive effect to maybe some extent.]