Pipe expansion due to freezing

Date 12-10-2008 22:55

First of all, lowering the temperature will make the pipe to contract, not to expand.
Second, I don't understand your concern on O.D. contraction. The contraction coefficient of steel is roughly 1 hundredth of a millimeter per meter of length and per degree Celsius. It varies a little according to the type of steel (carbon, alloy or stainless) but the difference is small. You translate into thousandths of an inch per foot and per degree Farenheit, but even if you don't, you can see that the contraction of OD is very small. An O.D. of 40 inches (roughly one meter) will contract that amount.
More consideration deserves the longitudinal contraction of piping due to lowering of temperature. In this case, you'll have to make a piping flexibility analysis.
Giovanni S. Crisi
Sao Paulo - Brazil

By hogan

Date 12-10-2008 23:01

Crisi,
I believe he is asking in regard to expansion, due to product (water) freezing in the pipe.

By strat

Date 12-10-2008 23:44

We had a simuler situation happing in Mass.It was a new boiler and the eng. didnt have the drian lines in the proper place and after the unit was hydro there was water still in the bottom of the HRSG's and once thrus.there was a hard freeze and the expansion from the ice actually cracked some of the welds, it was small dia. tubeing (1 1/2" and 2 1/2" ) t91 grade material. We had to go in and cut out the welds and reweld, they were aproximently 40 joints

By JTMcC

Date 12-10-2008 23:59

I would think the welds would beat the pipe. Unless I'm really missing something the welds should survive the freeze and the pipe itself should fail.

JTMcC.

By raptor34 (**)

Date 12-12-2008 03:04

I was on a job last year were some 18 x .250w X52 pipe was on a hydro that froze and caused the pipe to swell by about an inch OD, at the welds there was no distortion. I didnt get any pictures but it was the coolest thing I ever saw.

By Arctic 510 (**)

Date 12-11-2008 00:53 Edited 12-11-2008 00:57

I'm speaking of expansion due to product or water freezing and expanding inside of a piping system (deadlegs, etc). Usually if the pipe fails, it is in a longitudinal direction away from a weld. If the piping doesn't fail, it can expand (bulge) dramatically. Happens quite a bit here in Alaska... :)

By js55

Date 12-11-2008 14:07

I've never seen 'Guidelines' per se. Depending upon the material in question you have essentially strain hardened it. Your tensile strength will be higher and your ductility will be lower. And impact strength will be reduced, of some importance in Alaska I would think. Not to mention that the pressure boundary may have been compromised though it isn't visually noticable yet.
This is a engineering decision. But personnaly I would think that if you can see it, you probably should replace it.

By Arctic 510 (**)

Date 12-11-2008 19:10

js55-
I agree with you, and am of the opinion that this type of damage is unacceptable. I wish there was a standard on this in black or white to move it out of the realm of judgement (should) and into the rule (shall) category. I've contacted API via e-mail about this but the response was simply that API does not address this in their codes or RPs.

By motgar

Date 12-11-2008 14:11

Similar to how an engine block, of a car, has freeze plugs. The freeze plug(s) will hopefully blow out, before the block cracks.

By G.S.Crisi (****)

Date 12-11-2008 19:57

You're speaking of water freezing and expanding inside the piping. There are no standards (ASME, API or else) covering this matter, because when water freezes in a close environment (a piping or a vessel, for example) it puts a tremendous amount of stress on the environment that contains it. The containing walls will bust, or at least crack, no matter their thickness, as the other gentlemen who answered your question have pointed out.
Gioanni S. Crisi

By jwright650 (*****)

Date 12-11-2008 20:32

I found this and thought it was an interesting question about how much force would frozen water excert...but the answer isn't as easy as it would seem. Here is where a kid in school asked this question.....the replies the kid received were very interesting.

http://www.newton.dep.anl.gov/askasci/eng99/eng99532.htm

By Arctic 510 (**)

Date 12-11-2008 21:06

The piping does not always fail (crack or burst). If the pipe is not full, or pressure is relieved somehow, it may expand but not burst. Of course, code or product carrying piping would be replaced but in the case of a low pressure water line or non-code piping, the pipe can be heated with a torch, hammered back into place, and welded shut with 6010/7018 combo. Again, that is more of a backyard emergency repair more suited to the mining industry (I did this several times at a mine I used to work at as a welder) than say, petrochemical refining.

For swelled but not cracked/split non-critical, non-hazardous, non-code piping, what do you think would be the best approach if replacement was not immediately an option? Rework the required thickness for internal pressure equation? (Radius and wall thickness would have changed to a measurable value, but what about the stress value?)

I guess this is more of a theoretical exercise than anything!

By 803056

Date 12-12-2008 04:38

I believe the expansion of water as it transforms from liquid to solid (ice) is on the order of 9 to 10 percent.

The ability of carbon steel to elongate is inhibited as the temperature decreases. The reduction in ductility (carbon steels become more brittle as the temperature drops) as indicated by notch testing, may be the reason the pipes burst at the low temperatures.

My conjecture is that carbon steels, at room temperatures, have the ability to stretch about 25%, but that decreases when the temperature drops below the ductile to brittle transition temperatures, typically below 40 degrees F. As a result, the 10% expansion that occurs as the water freezes can not be accomodated and cracking occurs.

Best regards - Al

By js55

Date 12-12-2008 13:41

If I may,
It would be interesting to determine elongation data in cold temp tensile test. Which would be a slow strain rate test as opposed to the Charpy test with a very high strain rate. Charpy ductile to brittles would not be equivelent to cold temp tensiles which is closer to the freezing of water phenomena. Its not just the notch in a Charpy test but the high strain rate as well. A material can demonstrate greater ductility at slow strain rates than high.
In fact, this high strain rate and notch sensitivity is the very thing that drove the invention of the Charpy type tests in the first place when the Liberty Ships started cracking even though their cold weather and tensile properties were acceptable.

By Joseph P. Kane (****)

Date 12-12-2008 14:56

Al,

If the expansion was ten percent across the cross section, then wouldn't the circumferential material stretch be multiplied by PI?. (31% stretch?)

By G.S.Crisi (****)

Date 12-12-2008 14:29

Obeamweldor,
we're talking about piping or vessels completely full of water, not just partially full. If they're partially full, there's an air cushion that will compress and accomodate the expansion of water. On the other hand, if they're completely full, as liquids are non compressible, freezing water will put all of its stress onto the piping walls and, unless they're awful thick, they'll crack.
Giovanni S. Crisi

By Arctic 510 (**)

Date 12-12-2008 20:12

I understand that freeze-ups do often result in bursting. My original post referred to those that expand or swell but do not burst.

Does anyone think that the use of a ring gauge, a simple go/no go similar to those used in heater tube inspections, would be a legitimate tool to use? Would anyone use an internal pressure calculation (ie ASME Sect. VIII) based on the new maximum radius and the wall thickness to derate the system? If one were to use this, what would one use for the stress value? Could a new MDMT be determined?

I have heard "cut it out since it has yielded" but I have to reject that because bent piping and tubing is allowed, even though the act of bending cold works and yields them.

Is anyone else having trouble posting?

By js55

Date 12-12-2008 21:12

I don't think a derate will do it for you. You have a metallurgical change. How do you calculate where to derate without destructive testing. So it not just a wall thickness issue.
Also, bends are qualified. You could certainly establish a test regime similar to a qualification but would still have to deal with the fact that the mechanicals may fall outside of the original material spec and it would therefore be invalid to claim as that material IMO.
In other words, if you start with SA-106 B and the min elongation is no longer there, you essentially don't have SA-106 B anymore. In which case you have no catagory in which to group materials under qualificaiton and it would be a case by case basis. You could easily move into a great number of necessary qualifications.

By ctacker (****)

Date 12-14-2008 19:18

I don't have alot of experiance in the piping industry, but not all pipe is perfectly round, and therefore I don't think a go-no go gauge would work, I have built some 12' and 22' pipelines and the larger one (22') had what we called spiders (spokes to hold it round) but only at the ends of 60' sections to help assist in fitting together, then they were removed and the pipe would sag. May not be as pronounced in small diameter pipe.

By Metarinka (****)

Date 12-12-2008 21:06

this is an interesting problem to think about. The coefficient of thermal expansion for water increases it's volume by about 9% at atmospheric pressure. Any increase in the liquid volume contained in a pipe would required a squared increase in the diameter of the pipe. So that 9% increase is closer to a 21% increase in the diameter of the pipe. However at higher pressures water expands more going up to 16-20% when crossing the freezing point. Additonally, the vessel would shrink, but as the professor pointed out it would equate to thousandths of an inch and therefore shouldn't be consequetial.

Low carbon steels have an elastic limit between 20-30% off the top of my head and that would be lowered at O celsius for most metals. Therefore it's very easy to deduce what we all know that ice can and will burst pipes especially in a completely filled pipe. If we consider water to be inconpressible than a near unlimited amount of stress will be exerted on the pipe walls during freezing. If a pipe doesn't burst than this means either the strain was never great enough to leave the elastic limit of the metal. OR it did but stopped before it reached the ultimate tensile strength. In the first case no permanent damage has taken place. In the second case Cold working (no pun intended) and residual stress now remain in the swelled pipe and I'm guessing impact and hardness tests would confirm this

I think ultimately if the pipe has swelled but did not burst then it's apparent it has gone past it's elastic limit and has been plastically deformed. Above that it no longer has it's original shape and thus may have new stress risers and would no longer carry it's intended load or would have strength deductions. I would think any material that has been plastically deformed and is not repairable would need to be removed and placed with sound material in critical applications. In a backyard repair sort of sense. Annealing or stress relief could return the ductility, If it's not in service above 15 psi there probably wouldn't be any harm in continuing to use it.

By js55

Date 12-12-2008 21:28

If its changed a dimension AT ALL, it has exceeded the elastic limit. And in evaluating a stress strain curve for carbon steel strength will increase with strain up to a point. That point being fracture, as demonstrated by a true stress strain curve corrected for necking, not the engineering stress strain curve.
The problem will most certainly be ductility, and most likely impact toughness as well, probably better demonstrated through lateral expansion testing.

By 803056

Date 12-14-2008 00:20 Edited 12-15-2008 14:14

Here's a quote from a website I found. Sorry that I didn't copy the URL, but it was one of the first of several that popped up when I did a web search on the subject.

"Question - When water expands as it freeze, how much force (ie. psi) does it exert.
Application: if a say a crack in a foundation wall freezes with water in it, How much
force would be exerted in that crack.
---------------------------------------------------------------------
It depends on the temperature. At 0 degrees Centigrade, a very small pressure increase
will melt ice. As the temperature drops, the pressure required to melt ice increases
rapidly, by approximately 145 atmospheres per degree Centigrade. At -22 degrees
Centigrade, ice can exert 2,700 atmospheres (~40,000 lbs. per square inch) of pressure
without melting. (Data are from the web page http://www.benbest.com/cryonics/pressure.html)
Beyond this temperature and pressure, things get complicated because there are many
different forms of ice."

This web site provided some useful information, but the other did little to provide the type of information I was looking for, so I did the following analysis.

The elongation required to accommodate the expansion that takes place when water freezes isn't that dramatic. It appears the properties of elongation may decrease with a drop in temperature. The elongation appears to be on the order of 3 to 5% if the change in volume (water to ice) is about 10%. The elongation of carbon steel should be on the order of 25% at room temperature.

The following is my solution to the problem. Assume you have a sphere with a volume of water equal to 10 cubic units. Upon freezing the volume would increase to about 11 cubic units.

The volume of the sphere would be V=4/3 X Pi X r cubed (I wish there was a way to write formulas easily)
Circumference is = 2 X r X Pi
Elongation is = ((C2-C1)/C1) X 100%

The radius would increase from 1.33 units to 1.38 units upon freezing (assuming there is no restraint). All units are rounded off

The circumference would increase from 8.4 to 8.7 units

Elongation would only be 3.2% to accommodate the increase in volume of the sphere.

Initial Volume 10 = 1.33 3.14 r^3
2.394521335 = r^3
Radius1 1.337807408 = r
Circumference 8.40143052

Final Volume 11 = 1.33 3.14 r^3
2.633973469 = r^3
Radius2 1.380987598 = r
Circumference 8.672602114

Elongation 3.22768359

Instead of a sphere, if you use the area of a circle, that is a section of pipe with a length infinitely small and consider it to be a constant, , the elongation increases, but not close to what you would expect the elongation of steel to be. It worked out to be something less than 5% if I remember correctly.

If the thought process is correct and if I didn't make any blunders, it would appear the ductility of the carbon steel drops considerably as the temperature decreases. It would be interesting to see if it coincides with the data for the transition temperature (ductile to brittle transition) obtained by the Charpy Impact text.

Best regards - Al