self supporting steel water pipe 1

[bjb]

I have been asked to look into designing a pressurized steel water pipe to span about 85 feet. The diameter of the waterline is 30 inches, I don't yet know the water pressure is. The diameter of the pipe can be increased if required for structural reasons. Can someone recommend a design reference for me? We are basically trying to emulate a similar design that someone did in the 1970's where they had a 36" dia steel sewer pipe that spans the same stream.

In particular, I am unsure how to consider the interaction of bending stress, shear stress, hoop stress, and localized stresses at the supports. Also, am unsure about allowable defflections. I'm in New York, so any references based on US standards would be appreciated.

 

[JoeTank]

Check out the publications of the Steel Plate fabricators Institute. They have one on welded steel penstocks.

Steve Braune
Tank Industry Consultants

http://www.tankindustry.com

 

[lilliput1]

Use the Mohr Circle to combine stresses. You must factor in 100 mph wind. You can use a suspension cable - catenary arrangement to support the pipe. However you have to make the system rigid enough to not collapse like the Tacoma Narrows Bridge. Bridge structural enginers can help to find the ridgidity required.

 

[bjb]

Thanks for the replies. I have talked to an engineer from a steel water pipe manufacturer, and he indicated to me that Manual M11 from the AWWA has all of the pertinent information required to design something like this.

 

[BarryEng]

As far as AWWA M11 is concerned, use the latest edition. The previous edition was suspect in some areas. In the latest edition, the saddle tip stresses are now reduced based on the work of Stokes (1965) from South Australia.

You can also refer to 'Steel Penstocks' an ASCE manual 79. Remember that power penstocks are used for power generation & hence they usually do not consider temperature stresses. These stresses are too large for both the design of the pipe shell & anchorages. The temperature loads are removed with the use of flexible joints. For most water utilities, temperature stresses have to be included in the analysis because the pipes are usually fully restrained.

If you want to consider temperature stresses, read Bednar 'Pressure Vessels' (I think that is the title). In Bednar, there is info on 'self limiting' stresses from temperature, that is based on ASME info & also included in the commentary of the Australian Standard for Pressure Vessels. These references include higher allowable stresses (up to 2Fy) based on combinations of stress. The design is 'strain based' - see Bednar for details.

For most above ground pipelines, the maximum design stress is usually the sum of the uniaxial stresses, in the circumferential direction (NOT longitudinal), at the saddle tip.

Longitudinal stresses (in a continuous & restrained pipeline) are the sum of bending stresses (wl^2/12 at the support), long. stress due to pressure (PD/4t), temperature stresses & poisson stresses (vPD/2t).

Circumferential stresses are the sum of hoop stress (PD/2t) & the saddle tip stresses (see Stokes or AWWA M11 above for the reduction factor). I usually use Fy as a MIN stress for this combination (see Bednar or ACSE 79).

Combined stresses are (Huber, Henky, von Mises): -
reference stress = sqrt(f1^2 + f2^2 - f1.f2)
See AWWA M11 for details.

If you want to increase the span (or reduce stresses) try using a ring girder as in AWWA M11 or ASCE 79. For a pipe of 900 mm dia, a ring girder could be used because a 'one off' cost would not be prohibitive in a pipeline. Remember that 'normal' beam formula are used for the analysis based on the pipe being kept circular with either a saddle or ring girder. I usually use a 150 angle saddle. Even though the formula for saddle tip stresses indicates a lower stress from a 180 angle saddle, this is impossible to achieve because the stiffening effect of a pressurised pipeline, will lift a pipe off the top edge of a 180 angle saddle & the effective angle will be about 150.

Deflections are not usually a problem for a span of 20 m & a 750 mm pipe with an 8 mm wall thickness (I am assuming a D/t ratio of about 100). Just remember that most utilities do not take the stiffening effect of the cement mortar lining into account when checking the deflection. I have test results that indicate for a reasonable cement mortar lining, the combined EI value can be as high as double the EI value for the steel shell only.

BarryEng

 

[BarryEng]

I meant to mention, check the deflection & fabricate the pipe with a camber of twice (or more) of the deflection. The pipe will tend to settle but the main reason, is that a flat (horizontal) pipe will always look as if it is sagging.

BarryEng

 

[BarryEng]

Another thought for a starting point. In ASCE 79, the basic stress is 2/3 Fy or 1/2 Fu. Always start any design with use of the Barlow formula: -
stress = PD/2t.

Most utilities have pipe D/t ratios of about 100 to 120.
For a D/t ratio of 100, & an allowable stress of 2/3 Fy, the MIN allowable pressure is Fy/75.

In metric units, steel for pipes in Australia has an Fy of 300 MPa. So for a D/t ratio of 100, the minimum allowable pressure would be 300/75 = 4 MPa (or 400 m head).

This indicates (for a D/t ratio of 100) the MAX head as an upper limit (without any other considerations like saddle tip stresses etc) can be 400 m.

BarryEng

 

[JStephen]

You can also have vibration problems with a pipe in the wind- more common with stacks, but something to check on any long unsupported pipe.

 

[BarryEng]

If you have a concern with wind load, have a look at Tubular Steel Structures by Troitsky - one of the Lincoln Arc Welding series of books.

There is info on wind effects on both self supported & guyed stacks.

There are also good sections on above ground pipelines & horizontal storage tanks.

BarryEng

 

[JohnBreen]

Wow, lots of good stuff in this thread. The topic of very large diameter piping is not discussed much anywhere. BarryEng, you have obviously been down this road a few times before! Be careful of the Troitsky book - there are still lots of typos in there and some are in the equations. Over all though, the book is priceless.

Mostly (discounting gross material yielding) it is a structural instability problem (buckling). To collapse, the pipe would have to form three plastic hinges. That can only happen if the pipe loses its "roundness". The ring "stiffeners" will keep it round for some amazing spans. I would, however, feel much more comfortable with the cable bridge scheme.

Best regards, John.

 

[bjb]

Thanks to all for the replies. I hope to not use a ring girder and just have saddle supports. The pipe will be insulated and will have an aluminum jacket. The 36" dia. sewer pipe that we will share abutments with doesn't have ring girders, just steel saddles.

since my first post I have obtained a copy of AWWA M11 (most recent version) and is seems to have the info, that I need.

For projects in the USA, it there a grade of steel that is typically used for steel pipe for water mains?

 

[rconner]

I think steel water pipe properties/types etc. are discussed in AWWA C200, AWWA Manual M11 (Chapter 1...), and also many other publications such as ASCE MOP #79.

American SpiralWeld on page 10 of their manual at

http://www.acipco.com/aswp/pdfs/ASWP1-Introduction.pdf

suggests as capabilities, "Pipe meeting or exceeding the requirements of AWWA C200, ASTM A139, or
ASCE Manual No. 79".

 

[BarryEng]

To JohnBreen
Thanks for reminding me about Troitsky errors - it can be a VERY time consuming exercise to find (especially when you think that the printed text is correct & you don't even consider that there may be a mistake)! One of the VERY time consuming errors was in ASCE 79 for a formula that compared to Bednar, there is a '-' where it should be a '+,' in a two line formula. This error was a tad frustrating for me to find!

I agree about the buckling problem but for most 'normal' cases (& spans), the statement 'keep the pipe round with saddles or ring girders, & normal beam formluae apply' is OK. This statement has been supported by many authors on pipe design, dating back many years.

To BJB
I agree that you should try for a saddle FIRST for your pipe span & only go to a ring girder if you perceive that you may have problems.

Remember that M11 is a conservative approach for engineers who are not going to be above ground pipe design experts.

Also remember that the saddle tip stresses are local bending (within 15 deg of the saddle & extending for a very short distance in a longitudinal direction), bending stress varying from a max. on one face of the pipe wall, to zero & back to a max. on the other face, & so higher allowable stresses can be used based on 'strain design' as explained in M11. As long as the 'basic' stress (in hoop direction) does not exceed 2/3 Fy (as in ASCE 79), the addition of the 'local' saddle tip stress will only result in a plastic redistribution of stresses - justifying max. stress of Fy (or even up to 2Fy as explained in Bednar, ASCE boiler codes, ASCE 79 & the commentary of the Australian Code for pressure vessels).

Another comment - the formula in Roark & Young (Formula for Stress & Strain) for saddle tip stress, does not consider the stiffening effect due to pipe pressure. Use the M11 approach that includes the Roark stress with a reduction factor by Stokes (easiest option), or the ASCE 79 full formula approach.

The original analysis for saddle tip stresses, was developed by Zick. One of the readers of this thread may be aware of another thread where a contributor gave a web site that included a copy of the original Zick analysis (original published about 1950 I think).

BarryEng

 

[bjb]

BarryEng-Thanks for all the great info. that you have passed on. Regarding M11, in equation 7-1
k=0.02-0.00012(A-90), but in the example on the following page, k is taken as 0.02-0.0012(A-90). Which is correct?

Also, on page 69 under saddle supports it says that "If the pipe is only partially filled and the cross section between supports becomes out-of-round, the maximum fiber stress is considerably greater than indicated by the ordinary flexure formula, being highest for the half-filled condition". I don't see a method for calculating this stress in M11 for saddle supports. Is the half-filled condition something that really needs to be considered with saddle supports?

 

[BarryEng]

To bjb - Re M11

Look in Roark & Young - Formulas for Stress & Strain.
Art. 12.7 Pipe on supports at intervals (p509 in my edition).
k=0.02-0.00012(A-90)

If you do the calc, you get the following results
0.02-0.0012(120-90) = -0.0160 example on page 73 of M11
0.02-0.00012(120-90) = 0.0164 as per Roark & Young

The example (in M11) has used 0.0012 with the answer being the correct value with the wrong sign.

I would guess that the example has a misprint (0.0012 instead of 0.00012) & the answer is a misprint (-0.0164 should read 0.0164).

Look at the next line of the example, & the k value used is 0.0164. I guess there are two misprints in one line but the correct value was eventually used.

Partially filled pipes

The 'out of round' condition is the clue. This condition occurs with a relatively thin walled pipe & a partially filled pipe design condition. Water utility pipes designed for a reasonable pressure (giving a D/t ratio of 100 to 120) usually do not have this problem.

For a pressurised pipe, this is very rarely a design condition (see ASCE 79 for details of partially filled pipes). If the pipe wall thickness has been sized primarily on Barlow (PD/2t) with a resonable design pressure (such as a water utility pressures or wastewater pump station pressures), the partially filled pipe is rarely a design case.

However, if you are designing a low pressure pipe (a large drain or irrigation pipe, or a pipe at the top of a hill near the HGL) then this is definitely a design condition. I'm not sure what are the limits of D/t ratios for this to occur.

Have a look at JohnBreen thread above, & he raises this problem.

BarryEng

 

[BarryEng]

Partially filled pipes

If you look into ASCE 79, there is a graph indicating the maximum external pressure (internal vacuum) that can be taken by various D/t ratio pipes. From that the maximum D/t ratio to resist full vacuum, is about 140 to 150. If an air valve fails to operate, a D/t of 140 is OK, hence water utilities generally do not operate with pipes thin enough where the partially filled condition becomes a design criteria.

Power utilities that operate with large D/t ratios, also have large diameter anti-vacuum valves to guard against external instability. In this situation, partially filled pipes will have larger stresses than a full pipe (see ASCE 79 for details).

BarryEng

 

[bjb]

I have been trying to get a copy of ASCE 79, but so far have not been successful, I haven't been able to find it on their website.

 

[BarryEng]

ASCE 79 is a very large book & expensive. I would doubt that you could justify purchase, if you are only doing one design. Try your local Uni or perhaps a firm that does large diameter pipe designs. They may allow you to borrow a copy for a limited time.

See if you have a local office of USBR (United States Bureau of Reclamation) & contact them.

BarryEng