anchor forces on long straight pipe

[swoosh172]

I have a 6000 foot above grade crossing for a new water line on a steel bridge. Due to tight constraints and aesthetics we are using bellows type joints. The pipe is 8 inches in diameter and is at 150 psi. I am wondering what the anchor forces will be if I have 2,000 feet between anchors with the expansion joint in the center. I have the force of the joint which is 12,000 pounds. To guide the pipe we are looking at pipe slides every 45 feet to act as a guide and a support in one. Roller supports will be used every 15 feet otherwise. The guides are PTFE. The water is at maybe 60 degrees. Do I need to account for the friction force of the pipe and the slides in the anchor load calculation, or is the temperature and pressure low enough that it is negligible? I don't have a modeling program to help.

Thanks for the help!

 

[LittleInch]

Yes you need to account for the friction. Even with PTFE you will still end up with a"virtual" anchor at some point. If this is less than 1000ft, your bellows will not do what you think it will. The real issue for this type of system is not full under pressure, rel low temp (I assume you mean 60F ) operation, but empty, hot summer day temperature when the expansion will be much higher. Your pipe is very likely to buckle, probably upwards. The issue is probably not the overall expansion (400mm for 40 degree C change), but more that the friction of the supports is too high and the pipe will "lock up" before it gets to the bellows. Any out of straightness of this pipe will quickly change into a buckle and I'm talking low nos of mm.

This could be a very expensive mistake not to get this analysed properly using stress analysis software.

Either that or many more bellows and anchors, say an anchor every 3 or 4 hundred feet with a bellows in the middle. As my strapline says, theta a good reason everyone else does it this way....

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[swoosh172]

Do you think the fact that the bridge is steel and therefore should expand at somewhat of the same rate that the pipe may actually slide very minimally over the guides and rollers so that friction may not be that great. The bridge would essentially move with the pipe and the support locations on the pipe would not change.

 

[LittleInch]

It would help a bit, but at some point the pipe is fixed to the ground?. Also your thin small pipe will heat up much more than the bridge unless it is all under cover. Is all 6000ft (1.8km) on the bridge?

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[BigInch]

You don't need a program to analyze a straight piece of pipe. An XL spreadsheet would do more than well.

I'm not clear what material the pipe is. You've only said the guides or slides are PTFE.

Thermal stresses are not dependent on length, only on temperature changes (positive, or negative) and Young's modulus.
Force = stress x cross sectional area of the pipe wall.

If everything is always 60F, there would be no stresses. Is there day, night, summer and winter temperature variations where you live. Bridge temperatures may vary considerably more than the water, thus expect to generate the maximum friction forces possible at slides. Water may remain above freezing, while bridge temperatures could be much lower.


Independent events are seldomly independent.

 

[LittleInch]

I agree with all of that. Basicallyyou need to work out a series of scenarios to check depending on where your anchors are (on the bridge?) and the relative expansion of bridge versus pipe, which might be positive or negative. First work out your max buckling load, subtract a suitable safety margin, then work out what your max length is that this could actually move your pipe using accumulation of static friction at all your supports. I'm still not sure where you get 12,000 lbs force from the bellows, but that's a secondary calculation.

It still just seems a long way between anchors and bellows to me...

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[swoosh172]

The pipe is 8" schedule 40 steel. Is there a chart for maximum buckling load for pipe? I googled and can't find one and don't have a structural engineer in house to ask. The bellows force came from the vendor catalog sheet for 8" expansion joint @ 150 psi. I am thinking of starting at 500 foot anchor spacing now instead of 2,000 feet. I was only doing 2000 feet because that is where the roadway expansion joints are but i'm sure the 8 foot girder and bridge bearings can handle much more stress than this tiny pipe.

 

[racookpe1978]

Remember that your bridge, although all steel all the way across, is not likely to be "one piece" of steel all the way across supporting your pipe. At EACH expansion joint (probably at each pier or truss joint) the isolated sections of the bridge steel and deck and handrails and footpath and what ever will be holding up your pipe supports are going to be moving independently of each other.

Have you tried suspending the pipe from hangers? Sounds like somebody has an "architectural" wet dream going if they demand you use bellows rather than expansion loops.

 

[BigInch]

Bellows force that will require restraint is 12 Kips. Bellows force is calculated by multiplying x-sectional flow area of the pipe x pressure. for an 8" pipe it is appx 12,000 lbs TENSION at 200 psi internal pressure.
this is bellows, [<---12Kips BBBB 12 Kips--->]
That would require an anchor at each side of a bellows.
Anchor ----> [<---12Kips BBBB 12 Kips--->] <----12K Anchor

Normally you would need only one bellows. Why two? Unless the min/max expansion/contraction is greater than what one bellows can accept, use one.

if you had a figure like that above, with a bridge expansion joint at the center of the bridge, I'd try it like this,

===G====G===G====A== <-12K-> [BBBB] <-12K-> =A====G===G===G====
Guide Guide Guide Anchor Anchor Guide Guide Guide

each guide would only have a friction force = weight of pipe + contents * 0.3 friction factor, if steel on steel. PTFE slide plates have a friction factor of 0.1

Aside from the friction forces, the only real load in that arrangement above should be the 12 Kips compression (resisting the 12K tension force from internal pressure in the bellows) in the pipes connecting each end of the bellows to the anchor. That's what you get when you use only one bellows, with an anchor on each side. That way, thermal force is released by pushing all movement away from anchors and allowing it within the bellows. No restraint against movement = no thermal stress, or thermal force. Only friction forces remain that will try to buckle the pipe and they should be small enough to allow long distances between guides, especially if using TPFE slides.

If you have 1 bellows that will accept the full thermal expansion, then there will be no thermal load in the pipe from contraction during cold periods, or expansion when hot, except for what is caused by the friction forces of sliding at the guides. Very much less than pipe thermal loads in general.

Google the Euler buckling load formula, works for any section. Maximum compressive stresses are calculated by using that Euler buckling load divided by a safety factor, safety factor depending on what type of load caused the compressive stress. If thermal, I'd use sf=1.4 All you need to do is calculate radius of gyration of the cross section and know the length of the longest pipe between guides. Support adequately by brackets, or hangers, otherwise to avoid overstress from bending.

There is a difference in stress and force. Steel, bridge or pipe, if the same grade of steel, will both yield at the same tension stress. The bridge beam, having a larger x-sectional area, will probably be carrying many times the force in the pipe.



Independent events are seldomly independent.

 

[swoosh172]

Thank you everyone for your great input to this problem. In regards to BigInch's reply I think my problem may be in having the 2,000 foot of pipe between anchors. The friction force of this is 2000ftx50#/ft*.3= 30,000#. I used .3 because of steel on steel roller supports which are every 15 feet vs using the .1 of the guides. Is this correct? I take this to mean that it will take 30,000 pounds of force just to get the pipe to start moving toward the bellows joint. So by doing a thermal stress calculation I believe that the pipe won't even begin to move until there is at least a 20 degree temperature difference, so for the first 20 degrees all the stress is being transferred into the pipe section. Is the buckling a concern between guides only? So am I to find out whether a 45 foot section of pipe buckles at 30,000 pounds? 45 feet is the spacing of the guides based on 150 psi working pressure, this number is from the anvil website for slide guides at 150 psi. Again, thanks for the help and the great replys so far!

 

[BigInch]

Distribute the load to each slide
if you have 134 supports in 2000 ft of pipe
the load on each slide (with water in the pipe) is 50.2 lbs * 14.92 ft = 750 lbs
x 0.1 friction factor = 75 lbs per slide plate
The 75 lbs friction is generated in each slide
The pipe pushes out ,away from my first anchor above, to the left.
Each slide won't resist more than 75 lbs, so the load in each pipe segment between slides cannot possibly be more than 75 lbs.
Pipe load = 75 lbs <-, friction on slide-> 75 lbs, carried on to next pipe segment is +75-75=0
Next segment, 75 lbs<- add 75lbs ->, net load to next segment = 0
etc.
You can see that there is no load greater than 75 lbs using that method.
Force between two anchors, not having an included bellows is HIGH!
a = 0.0000065 in/in-F
dT = temp change 20 deg
E = 30,000,000 psi
stress = a * dT * E = 3900 psi
thermal force = 3900 * Asteel
Asteel = 8.4 in2 = 32,760 lbs

A.) Use one anchor, allow growth/contraction in each direction, do not restrict expansion/contraction, except on each immediate side of the bellows and your maximum load is 75 lbs in any segment. Bellows anchors have 12 kips.

B.) Use two anchors, the compression load in ALL pipe segments in between is 12,000 lbs compression, the bellows resisting load, if there is a bellows in between the anchors.

C. No bellows, or if bellows max expansion is reached, pipe anchored both ends could be a high load of 32,760 lbs thermal + 12,000 lbs bellows resisting compression load (44,760 lbs compression total) in all pipe segments with only a 20 deg temp rise and the expansion of 0.26" was totally restrained. or 32,760 lbs tension -12,000 bellows load = 20,760 tension in all pipe segments with a 20 deg drop in temperature. If the bellows accepted some movement, the load would be proportionally reduced.



Independent events are seldomly independent.

 

[swoosh172]

BigInch, can you let me know what the forces are at the anchors if this is my arrangement? I am having trouble wrapping my brain around your proposed anchor system which appears to let the pipe free float with exception of directly below the roadway joint.

=A=======G====G===G====[BBBB] =====G===G===G=====A=====G====G===G====[BBBB] =====G===G===G=====A====

The bellows joint is 1,000 feet away from the anchors, which are 2,000 feet apart. I have 3 joints over the length of the 6,000 foot bridge.

change in temperature is assumed to be 100 degrees between summer and winter. Pipe will be dry in the winter and drained to prevent freezing.

Thanks for the help! Hopefully this is my last question

 

[BigInch]

Not quite what I envisioned, but anyway,

For pipe design purposes, it looks like 12,000 lbs in all pipes between anchors, provided the bellows accept all movement and don't bind up. Theoretically the center anchor will have no net force, since the +12Kip load balances the -12K load, but I'd design it for 12,000 lbs anyway.

I'd assume no friction at the guides for pipeline design purposes, but actually design them for at least 1000 lbs friction load (probably more than necessary, but a nice round number) and of course the vertical load of weight of pipe + contents.

Temperature will not make any difference, since supposedly the 2 bellows will accept all the thermal movement.



Independent events are seldomly independent.

 

[swoosh172]

Thank you BigInch. Everything I found so far leads me to believe I will have frictional forces also applied to the anchor consisting of weight of pipe incl water x length of pipe x friction factor. This would add another 50# x 2000 ft x .3 = 30,000# to the load of the anchor. The structural engineer's are going to kill me !!

 

[BigInch]

No I was giving you what I would use for design loads for the guides and anchors.
The friction loads are ACTUALLY resisted by and at each guide.
12000 lbs - 750 lbs/guide = 9750 lbs remaining load at the outside anchors.
The center anchor has no net force.
see diagram

Independent events are seldomly independent.

 

[StressGuy]

BigInch,

How are you not getting the pressure thrust and friction loads on your anchor to not be additive? I think you're signs are flipped on your friction.

Also, the joints themselves do have spring rates. If we're talking about a 100F temperature change over 2000' of carbon steel pipe, that's over 15" of compression into each joint....

...actually, I think you're going to have a tough time getting a single joint design to absorb that much compression without squirming at the design pressure the OP is looking for.

Edward L. Klein
Pipe Stress Engineer
Houston, Texas

"All the world is a Spring"

All opinions expressed here are my own and not my company's.

 

[KevinNZ]

When we heat up our steam lines we can have hot pipe expanding on one side of an anchor and cold pipe not moving on the other. We therefor do not get balanced friction loads, We get 100% friction on one side and nothing on the other.

Same could happen on a long cold water line. Empty and hot from the sun then cold water is flowed in from one end.

I have seen linear expansion bellows used on long lines but these have internal and external guides and bellows restraints.

Kevin

 

[DSB123]

KevinNZ,
It is physically impossible for a steam line to have no expansion on one side of an anchor point and full expansion on the other side unless you have a closed valve at the anchor. Even then there would be a temperature gradient. I agree that during warm-up there is a temperature gradient along the line but to consider full expansion on one side of an anchor and no expansion on the other is very conservative to say the least.

 

[KevinNZ]

DSB123

As you feed steam into a line the hot 100C steam moves up the pipe. Behind this front the line is hot and ahead is cold.

I have seen this many time times.

Kevin

 

[BigInch]

And when the hot part is expanding, it is also pushing the cold line ahead of it, thus producing friction at all the supports, even though half the line is still cold.

Independent events are seldomly independent.