Pipe Stress Question 1

[dsg1985]

A grad engineer and I are just getting into piping and pipe supports. There's one issue we don't agree on so far

I believe a pipe elbow at a certain rating (eg200psi) does not require additional supports to relieve the tensile stresses in the pipe. My friend is of the opinion that an unrestrained elbow will fail, that the tensile stresses in the pipe are not designed to carry such loads.

Th right answer, if given here, could cost one of us a mars bar.

 

[sspeare]

What is your piping application? For example if you have a buried water line (potable, fire service) you have to restrain the thrust at changes in flow direction (fittings). A common way to restrain buried fitting thrust is with concrete "thrust blocks". For example, NFPA 24 "Private Fire Lines", in Annex A, shows how to calculate the thrust force reactions, and how to size the concrete thrust blocks.

On the other hand, if your piping is above ground, you need to design some flexibility into your layout. At the risk of oversimplification, I suggest that the flexibility of an above-ground system is generally obtained by NOT restraining elbows. For building services piping (hot water, low-pressure steam, domestic hot water, etc.) the ASHRAE Systems & Equipment Handbook, chapter titled "Pipes, Tubes and Fittings", has a nice primer on flexibility. My ASHRAE Handbook is the 2000 edition, and the "Pipe" chapter is number 41. The Handbook you use may be a different edition but the physics does not change.

Hope this helps.

 

[BigInch]

A "pipe elbow at a certain rating", hard to say what that actually means, but I'll put in my 2 centimos worth anyway, would probably imply only the allowable hoop stress produced by internal pressure has been considered, because the mfgr would have no idea what axial stresses were. If hoop stresses were limited out and, if the combined stresses were actually calculated under typical piping codes using their permitted combined stress theories, that would still leave 20%, or so, of unused longitudinal stress allowance, which would accomodate some degree of additonal bending moment stress being added to existing longitudinal stresses without reaching an over-stressed condition.

Hope Mars bars are allowed under your diet. You're probably going to have to eat a few of them.

"The top of the organisation doesn't listen sufficiently to what the bottom is saying." Tony Hayward X-CEO BP
"Being GREEN isn't easy." Kermit

http://www.youtube.com/watch?v=hpiIWMWWVco

http://virtualpipeline.spaces.live.com

 

[pipesnpumps]

I think what the OP is asking is much simpler and more elementary than that.

It seems to me the question is something more like: "do you need a support for each elbow to keep it from blowing apart".

It sounds like your friend is incorrectly thinking that an elbow cannot handle axial/longitudinal load due to pressure thrust, and therefore must have anchors for each elbow (such as underground piping where thrust blocks are used to hold elbows in place while the pipe slides through a gasket in the elbow and is free to move axially).

A pipe can have no supports (just laying on the ground) and the elbow will not fail. As biginch put it, there is enough allowable stress leftover to handle the axial load due to pressure thrust.

 

[dsg1985]

Sorry, I didn't phrase my question very well.

My question is - are pipes designed to accomodate longitudinal stresses that arise when elbows are not restrained? An infinitely long section of pipe (or one where the elbows were completely restrained) would have 0 longitudinal stresses due to pressure (ignoring any induced by thermal expansion). One without restraints at elbows, or a length of pipe capped at both ends and pressurised (a 'makeshift pressure vessel') would have longitudinal stresses. My friend believe(d/s, not sure) they are not designed this way, that elbows are but the lengths of pipe may not be (numerous warning messages on pipe catalogues have led him to suspect this).

It didn't make sense to me that these longitudinal stresses would not have been accounted for in design of the piping. As BigInch said, I would have thought pipe would be designed based on hoop stresses, and you would therefore have longitudinal stresses 'left over' to carry dead weight and such.

 

[rconner]

The part of the premise of your question, "...the tensile stresses in the pipe are not designed to carry such loads" was as it appears you now know arguably not real clear as to what exactly is debated. Of course in conventional engineering design, all manner of "loads" create stresses, and maybe even significantly rebating stresses in parts etc. For dependable service parts must in turn have adequate STRENGTH to withstand the loads applied (or in another way of saying, to withstand all manner of stresses created in the part by the loading).

I guess the safest answer to most potentially unclear questions is, "It depends"! As others have noted, steel or ductile iron fittings (as many are familiar with, and with significant body thickness generally equal to or greater than the adjoining pipes) generally have adequate strength in all directions to handle the loads or rebating loads applied, whether or not the fitting is thrust blocked, of course as long as the system is suitably designed and installed. There have on the other hand apparently been many problems with weaker e.g. bend materials, or materials that have less/uniform material strengths or homogeneity in all directions, reported in at least some applications (an example might be observed in a detailed reading of the copy at

http://www.harcofittings.com/markets/irrigation/article-pvc-failures.html

).

"Mars bars, or Stars??"

 

[BigInch]

The combined stress equations prove that when you only have one directional stress (hoop) that is limited to around 70% of yield, the combined stress should not be close to its failure value, so axial stress from internal pressure, thermal, or bending cases can still be added.


"The top of the organisation doesn't listen sufficiently to what the bottom is saying." Tony Hayward X-CEO BP
"Being GREEN isn't easy." Kermit

http://www.youtube.com/watch?v=hpiIWMWWVco

http://virtualpipeline.spaces.live.com

 

[prex]

First point is that an infinitely long section of pipe cannot exist in our world, so all sections of pipes containing a pressurized fluid will have a longitudinal stress in their walls.
In trying to answer your rephrased, but still not well posed question: an unsupported section of pipe (of course with capped ends) containing one or more elbows, but no expansion joints of any kind, requires no additional supports to withstand the internal pressure, provided all components (pipe and elbows) are within their pressure ratings.
I would add that this is, under all codes of practice, even mandatory: you are not allowed to resist internal pressure by the pipe supports, unless this is required by the presence of expansion joints or by forces generated by fluid movement (pen stocks).

prex

http://www.xcalcs.com

: Online engineering calculations

http://www.megamag.it

: Magnetic brakes and launchers for fun rides

http://www.levitans.com

: Air bearing pads

 

[BigInch]

Prex. Lost me. Arn't you talking about holding non-welded end caps on the pipe by use of a restraint, whereas the OP is talking about resisting the forces-stresses in the pipe caused by internal pressure on welded end caps, rather than holding the internal pressure itself. For example, if you had a tube with welded end caps with the end of the pipe immediately adjacent to the caps rigidly anchored, end cap pressures convert to axial tension in the pipe cross-sectional area which would then transfer to the rigid anchor.

"The top of the organisation doesn't listen sufficiently to what the bottom is saying." Tony Hayward X-CEO BP
"Being GREEN isn't easy." Kermit

http://www.youtube.com/watch?v=hpiIWMWWVco

http://virtualpipeline.spaces.live.com

 

[prex]

BigInch, I was trying to interpret and answer the original question posed, the OP will possibly say a word on whether my interpretation is correct.
Also I was speaking about an unsupported section of piping (just pipe, a few elbows, two caps, all welded).
If you are questioning my statement that all sections of pipe will have a longitudinal stress, then I agree that you can set up a condition where the longitudinal stress is zero (apart from having an expansion joint of course). However this would be a quite theoretical situation, much like the infinitely long pipe: no code allows you to have a pipe thinner than the minimum required by pressure (so it's no use trying to resist it with an end block) and it would be quite difficult, in practice (though of course not impossible), to have end blocks that are stiffer than the pipe in tension.

prex

http://www.xcalcs.com

: Online engineering calculations

http://www.megamag.it

: Magnetic brakes and launchers for fun rides

http://www.levitans.com

: Air bearing pads

 

[BigInch]

No, more the part about resisting internal pressure with a pipe support. With some particular anchor arrangements I don't think there would be a choice, as in the example I mentioned with anchors near caps. It would seem that the anchors would be resisting much, if not all, the end cap's pressure forces. I don't understand how codes could prohibit that under those circumstances.

"The top of the organisation doesn't listen sufficiently to what the bottom is saying." Tony Hayward X-CEO BP
"Being GREEN isn't easy." Kermit

http://www.youtube.com/watch?v=hpiIWMWWVco

http://virtualpipeline.spaces.live.com

 

[prex]

A sketch of what you mean would be helpful.
The fact that a support able to resist the longitudinal thrust is built near the end cap, doesn't mean at all that the support will take the thrust in place of the pipe, if this is what you mean.
Suppose we have a straight pipe with two end caps and only one longitudinal support at one end. In this case of course the pipe takes the full thrust and none goes to the support.
If we now have two longitudinal supports, one at each end, then it is a matter of relative stiffness of the supports (and their foundations) to that of the pipe. However no one in practice would do that, as the pipe can withstand the full thrust, and there is no reason to transfer it to the supports. And even if the two supports were there for other reasons (can't see one though, also considering that there would be thermal expansion concerns), the longitudinal thrust taken by the supports could simply be neglected.
A different situation is if in the middle there was an expansion bellow: not only the thrust would be higher (because of the waves), but it would entirely go to the supports (and both of them are absolutely necessary here).

prex

http://www.xcalcs.com

: Online engineering calculations

http://www.megamag.it

: Magnetic brakes and launchers for fun rides

http://www.levitans.com

: Air bearing pads

 

[BigInch]

That's what I meant, a pipe between two anchors. Whether anybody would do it knowingly and intentionally; well it seems to happen in any case between two vessels connected by a straight pipe, two pumps connected by a header, etc. But if it does or doesn't still isn't my question. I simply don't see where it's prohibited by the codes.

"The top of the organisation doesn't listen sufficiently to what the bottom is saying." Tony Hayward X-CEO BP
"Being GREEN isn't easy." Kermit

http://www.youtube.com/watch?v=hpiIWMWWVco

http://virtualpipeline.spaces.live.com

 

[prex]

What is prohibited is that a structure external to the pressure enclosure (support, stiffening ring, anything) would be required (necessary, indispensable, pipe failing without it) to resist pressure by a piping section, possibly including elbows (but not bellows).
Again, I stated this to support my answer to the original question.

prex

http://www.xcalcs.com

: Online engineering calculations

http://www.megamag.it

: Magnetic brakes and launchers for fun rides

http://www.levitans.com

: Air bearing pads

 

[howls88]

It sounds like you, prex and BigInch, are skirting a question that I deviates from this topic a little.

Background...I have a generic calculation that our company is wanting to use to determine allowable pipe spans depending on certain situations. Someone else started the calc and in the calculation it is said that the allowable stresses for stainless steel piping is 8350 psi (per ASME B31.3 it is 16700 psi). His reasoning is that he did not want to include pressure in the calculation so he took half the allowable stress and said the other half would be the maximum allowable stress due to pressure.

We need to have documentation or be able to prove something like this in our calcs for it to be used, but he has not given any reference to this or further explaination (He no longer works here so I cannot ask him). First off, is this correct? Next, if it is correct, how can I prove this, preferably using an ASME code?

 

[prex]

Well, may be acceptable, provided the allowable stress is at temperature.
I say acceptable, but it is not exact. This because in a pipe having exactly the required thickness per code, the longitudinal stress due to pressure is not exactly half the allowable stress. If I'm not in error, it should be a little less than 50% though, so you would be on safe side.
And there is no such procedure in the codes, they only tell you how to check the acceptability for a given calculated longitudinal stress. What you should do, to insure the conformance to code, is to check a posteriori that your result satisfies code formulae.
And of course, I and BigInch, we hijacked this thread, but you did it even worse...

prex

http://www.xcalcs.com

: Online engineering calculations

http://www.megamag.it

: Magnetic brakes and launchers for fun rides

http://www.levitans.com

: Air bearing pads

 

[sspeare]

Hey, I don't know if this will help or hurt the discussion, but the attached file is my attempt to get all this combined stress stuff onto 1 page.

howls88 said: "His reasoning is that he did not want to include pressure in the calculation so he took half the allowable stress and said the other half would be the maximum allowable stress due to pressure". Why would you "not want to include pressure in the calculation"? You just calculate the Hoop Stress from the pressure, pipe OD and wall thickness. Then plug the Hoop Stress into the Tresca formula at the top of my 1 page attachment.

I know this is tedious, but we're engineers - that's the way we are!

 

[howls88]

Sorry for the hijacking, but I wanted to get your input.

The attempt of this calculation is to develop standards for our company for allowable pipe spans for straight runs and distances between support and elbows for pipes under a certain temperature, say 100F. The reason we wanted to leave out pressure is this calc is to be used for all pipes (as specified in the calc) no matter the pressure. This should be conservative how it was described to me and how prex has described.

We are trying to keep it conservative and simple so it is easy to implement and safe to use.

Thanks for your help.

 

[IFRs]

No matter what the pressure? I'd be careful reserving only 50% for pressure stress. Some bonehead will use this for 3000 psi something or other. Why not include pressure since it is so easy to calculate?

 

[howls88]

Your right IFRs, it may be better to include the maximum pressure and temperature to better describe the systems that this calc can be used for.

There is no small task that I cannot over complicate.