HDPE Piping Question

[Everynameistaken]

Hello Civil Friends,

I am a simple structural engineer and I am looking for some help in designing a HDPE pipe under pressure for a hydro power penstock.

We have designed these pipelines as a soil structure interaction using a 3D stick model with springs modelling the soil to determine the axial and bending stresses due to the pressure from the water column of the pipe and add them to the hoop stress. Typically these penstocks are from steel pipe where the interaction between the principal stresses in the pipe is very well defined and can be added with a Von mises formula.

Since for this project we are looking at the cheaper option of using HDPE pipe the process is a little different.

The initial modelling is the same to get the axial and bending stresses in the pipe and then there is the calculated hoop stress but the interaction between stress due to pressure (hoop) and axial (tension and bending) does not appear to be very well defined?

Looking through the PPI information they have a very simple equation where they limit the bending stresses to 1/8 of the hydrostatic design stress (HDS) but they do not show any interaction with the hoop stress?

Can I be at 99.9% efficient in hoop stress or operating pressure and still put bending into the HDPE up to the 1/8 HDS?

1/8 HDS seems pretty low, what about if the axial stress was fully in tension? Can I pressure the pipe to 99.9% operating pressure with a blind flange on the end so there is also quite a high tension load and how do these stresses interact?

Thanks for any info

 

[BigInch]

I would tend not to model HDPE pipe spans using soil springs as the bending limits are very low. It would be far better to assume full and continuous support is provided by the soil all along the pipe to keep bending stresses to the very minimum. Lay the pipe on 6" of compacted sand placed at the bottom of the trench to make it so.

Pipe embedded in soil when pressured will normally be in compression. It will normally only be in tension if it is not buried. End cap or blind ended tension stresses on unburied pipe can be up to roughly 1/2 the hoop stress, if pipe to soil friction is low. Tension stress from internal pressure is calculated roughly = Pressure /4 * pi() * D^2 / A
where A = Pipe's cross-sectional area

 

[Everynameistaken]

Thanks for the reply.

The pipe will be buried and fully supported however there are bends and we would like to have the bends also soil supported without a large concrete thrust block. To do this we would have a controlled backfill at the bends but there will certainly be some bending stresses in the pipe associated with this condition.

What I am getting from you is that there is quite a low capacity to have combined stresses in the HDPE?

 

[bimr]

You can download a design manual for HDPE pipe here:

http://plasticpipe.org/pdf/chapter07.pdf

Thrust blocks are not necessary on bends.

 

[BigInch]

HDPE is thermowelded, so no thrust block is needed, nor recommended. They will probably do more damage to the pipe than help it.

 

[Everynameistaken]

Hi Bimr,

I have read through the design guide and while it has some interesting information it is pretty simplified and looks at buried pipe in a trench only and does not really dress the actual soil structure interaction and the stresses in the pipe itself.

Consider the example in the attached sketch where the pipe is DR 7 (so quite stiff) and the embankment is of a quite low stiffness, perhaps an fill compacted around the pipe that is not able to support too much lateral load. If the DR pipe is at close to its maximum operating pressure and the pipe gets around an outside bend there is considerable thrust outwards and the soil structure interaction rely heavily;y on the pipe itself in bending to support this thrust and we would get a bending moment something like what is shown (there could also a net axial load which could be T or C dependant on the global configuration of the system).

In steel, we could add the hoop tension stress with the axial stress associated with the bending using the Von Mesis formula to ensure the overall stress is below the allowable steel limits. But with the HDPE i am not too sure how to consider the interaction between the different stress orientations. It is not enough to just say that the operating pressure is within the allowable range and since the pipe is buried there is not bending or axial stresses, there certainly could be and if not accounted for then we could get pipe failure.

 

[BigInch]

First of all we don't use von Mises for pipeline design. Use hoop stress, axial stress, bending and take them together to get an equivalent stress (different than von Mises). The allowable for the equivalent stress has an appropriate safety factor such that more complicated stress combination methods are not needed.

The fill in the trench should be adequately compacted, but even if not, it will usually compact itself quite well around the pipe in a short time. Use long natural bending radaii at bends to minimize concentrated outward thrusts. A long radius bend distributes outward thrusts along its entire length. I would be very surprised if the soil on the side of a trench gave way and the pipe climbed out of the trench. I suppose it could happen, but I've never seen that happen on any ambient temperature pipeline. At most you might get some ovaling of the pipe.

I've already said that, if the pipe is buried, you won't get much if any axial tension, at least as long as the temperature does not drop a lot. At normal or slightly elevated temperatures, you will have compression, if anything.

 

[Everynameistaken]

Hi BigInch,

Thanks again for more info.

Just to confirm, you would take the induces stresses in the pipe, hoop and axial (from both axial and flexure) and add them linearly and if this artificial total stress is under the equivalent allowable stress in the pipe based on HSD, temperature etc. then the pipe is acceptable? I am assuming that the different surges conditions can also be done this way using the modified limits?

I would like to ask you another question, Consider a hypothetical situation where the same HDPE pipe is on an outside bend but is not buried, now for sure there is both axial and hoops stresses. Under these conditions you would still use the same artificial total stress addition?

Thanks again for your help,

 

[BigInch]

Yes. That's the way it's done for ASME B31.4 and B31.8 pipeline designs. B31.1 and B31.3 are slightly different in the manner in which they calculate some stress components, especially how thermal stresses must be considered.

Although hydrodcarbon and natural gas pipeline surge stresses are treated differently. We (and maybe you) need to find out what design code you are using. Liquid hydrocarbon pipelines can have a 10% overstress for surge conditions, natural gas pipelines not. The design code might also give you some other pointers on how to do the stress analysis, define allowable stresses, etc. I don't do hydro penstock designs, so I don't know what is actually required there.

 

[bimr]

You should review some of the Corp of Engineer design manuals for penstocks.

http://www.publications.usace.army.mil/USACE-Publications/Engineer-Manuals/

https://www.usbr.gov/tsc/techreferences/hydraulics_lab/pubs/EM/EM03.pdf

http://www.publications.usace.army.mil/Portals/76/Publications/EngineerManuals/EM_1110-2-2400.pdf

http://www.publications.usace.army.mil/Portals/76/Publications/EngineerManuals/EM_1110-2-1602.pdf

With the HDPE pipe, one area that must be addressed is when the pipe exits/enters the ground. HDPE is probably not suited for an abovegrade penstock application. HDPE pipe is usually laid on top of the ground.

 

[LittleInch]

BI,

I assume you're still in New Year fug here "Natural gas pipelines can have a 10% overstress for surge conditions, liquid hydrocarbon pipelines not." - Surely the other way around??

Either way, every name, I suggest you look at all the other chapters in the plastic pipe link BIMR sent you. chapter 3 might answer some of your questions.

I think you're over complicating this, but to be a little fair, you appear to be in the twilight zone between a fully restrained system and your slightly odd surface laid with a bit of soil on top system with bends which is neither fully restrained nor unrestrained like a piping network. that's why you analyses it normally using Caesar or similar pipe stress analysis.

PE has quite a high Poissons ration (~0.45), but a large thermal co-efficient. It also exhibits elasto-plastic creep properties under long term stress, so it's not the same as steel.

I still don't get where your " pipe gets around an outside bend there is considerable thrust outwards" considerable thrust comes from where exactly??. Any force is resisted by the restrained part of the pipeline so the actual movement is low. This in one reason you don't really want or need thrust blocks because the movement is so low, but forces could be high if you try to prevent any movement.

It would be good to see some numbers here so we can figure out if you're exceeding the guidelines you mention.

I believe that yes, you can be at the max pressure rating of the pipe with a blank end. PE pipe has a substantial design factor built in.

Usually the biggest issue with PE pipe is thermal growth / stress between as-laid temperature and operating temperature - this is 10 X steel rates so can become quite significant, in either direction.

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

 

[JedClampett]

I kind of scanned this exchange and I think there's a lot of valuable information. But my basic question is: What are you cheapening a penstock design? These are systems that need to last a thousand years with little or no attention. Failures are disastrous. Do you want to find out 10 years out that HDPE ages or deteriorates in an unexpected way? We have enough problems with concrete and steel pipes, why experiment? And how much are you going to save, especially considering the total cost of the project? One percent? Three percent?
There's a bunch of residential pipes in the Phoenix area called Orangeburg. I'm sure it was the latest, greatest (and cheapest) pipe you could find. Now it's being replaced, stick by stick. But think of the money saved!

 

[BigInch]

I wanted to say the same, "watch out for false economy", but I somehow feel that initial cost is a very important consideration here.

 

[fel3]

JedClampett…

Even with 36 years of experience in civil engineering (mostly infrastructure and mostly in California), I had never heard of Orangeburg pipe, so I looked it up. From

http://www.sewerhistory.org/articles/compon/orangeburg/orangeburg.htm,

about halfway down the page: "…for being a “coal tar impregnated toilet paper tube,” this product didn’t do so badly."

I also didn't know there was a website dedicated to the history of sanitary sewers, so I learned two somewhat useful things today. Thanks, I think.

Fred



==========
"Is it the only lesson of history that mankind is unteachable?"
--Winston S. Churchill

 

[Everynameistaken]

Hi All,

Thanks you for your comments.

I have been in contact with the PPI, and the rep I spoke to was yet to be able to confirm how they establish their combined stress criteria. They were very helpful and are in the process of digging a little deeper. Once I have an answer from them I will post here.

To answer a few of the question/comments above:

LittleInch: We are in preliminary design so I don't have the bending stress components but operating pressure is around 333 psi (so near the max for DR 7.3 pipe!). Yes we certainly are living in the gap between fully and fully un restrained pipe, thats why I have the questions! The trust comes from the water column pressure and the change in direction of the bend. I agree with a quite large bend radius this thrust will be spread over a large soil area, but as you could see we only have a small amount of soil on the edge of our mountain track where the pipe is being laid.

If i consider the basically the full rated pressure and then a blind plane I get something like (back of the envelope)

Hoop from pressure = 5.9 MPa (856 psi)
Axial from pressure on blind flange) = 2.4 MPa (348 psi)

So if i directly combine these stresses => 5.9 + 2.4 = 8.3 MPa (1200 psi)

So this is larger than the presumed allowable of 1000 psi or 6.9 MPa.

Jed and BigInch: I don't think we "cheapening" the penstock. If the HDPE can take the required loading and is a less expensive alternate then I think we are providing the best economical solution. This is not a major infrastructure project for a large municipality, it is a micro hydro looking to produce small amounts of off grid power, it needs to be low cost to be economical enough to build.
What I am trying to see is if I could use HDPE for the whole length vs using it for some and then switching to steel when the pressure increases. If we can stay with HDPE and back it with some sound engineering practices and judgements and we can same money then I think we have down our job. If not we will use the steel and the project cost go up, but nothing we can do!

Thanks again for all the comments, very helpful.

 

[JedClampett]

fel3, I'm no expert, but we have a lot of very good pipe engineers where I work. And one of them had to saw cut his patio to get to some root filled pipe. A plumber wouldn't rod it out because the Orangeburg pipe would fall apart with the roots. He's the kind of guy who believes in getting his hands dirty.
It was kind of ironic.

 

[BigInch]

I had a feeling it was a micro-hydro. It's common to use HDPE for maximum initial cost savings, which may not be the cheapest over the lifetime, but it is initial cost that is most important in these kinds of things.

How did you get the 333 psig max hoop stress? Is it static head only, or do you have some surge pressure included in that too?

 

[LittleInch]

Any thrust you get on a bend is resisted in the main by the axial tension in the rest of the pipe.

You also need to factor in the tensile stress from the hoop stress which cancels out a lot of "thrust" at the bends.

The allowable hoop stress is less than total allowable by some margin normally. I can't see your calcs as to how you get to the stresses quoted, but whilst you're close to the limit, I don't think you're far away. You can always add some restraints if you're worried about end loads, but I really wouldn't bother myself.

ditto the bends loads are not real discrete loads as they are resisted by things other than the soil load.



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