Water freezing in an Overland Pipeline

[ECD40]

We have a problem to solve with an overland water pipeline in very cold weather (minus 40 degrees). The pipe will be a 14" diameter DR11 HDPE material and 19,000 feet long. The flowrate will be 2,500 USgpm at a velocity of about 7.3 fps. The question is whether or not to insulate the pipe with the water flowing and how long will it take to freeze if the water slops flowing, with or without insulation. If we insulate the pipe, there will be a differential expansion between the HDPE pipe and the insulation material. The maximum summer outside temperature will average 75 degrees F. The pipe cannot be buried.
Can anyone help with this problem?
Thanks,
ECD40

 

[cvg]

thread378-446544

collect the data and run the calcs?

temperature of water
outside temperatures
length of time
sun angle and intensity
pipe wall thickness
thermal conductivity
wind speed
insulation r-value
etc.

however, without doing any analysis, I would bet money it will begin to freeze in minutes
have you ever been outside at 40 below?

 

[racookpe1978]

With no calc's, I am wrong to blindly assume too much; but initial fill of the pipe with fresh water, from empty, in cold conditions, will certainly freeze that first slug of water before it can get through to the tank. Insulation or not, the pipe walls, the insulation thermal mass, and the air inside the pipe will all have to come up from -40 deg to water flowing conditions.

That's a lot of thermal mass = will likely need eletrical trace heating before startup. Might need only "available" thermal trace heating after operation when the outlet temperature from the pipe is above freezing.

Your units are degrees (-40 = degree F and degree C), but US GPM and feet: What is lowest expected water inlet temperature going into the pipe? Deg F I assume.

 

[chicopee]

Now how is it that the common prevailing thought is that running water in a pipe does not freeze and you see that in snow making operations on ski slopes.

 

[georgeverghese]

Insulate and include self regulating type electrical heat tracing for low or no flow antifreeze protection.

 

[racookpe1978]

No, no. Running water in a pipe is less likely to freeze than still water.
But the dynamics make it much, much harder to freeze running water.

To freeze, the water MUST lose MORE than [mass x cp x (temp change of the water over freezing) + mass x energy_phase_change_of_ice ] in less time than the water is exposed to the cold air temperature. That's very, very difficult in a short time. We don't know travel time precisely, but have been told 7.3 fps for 19,000 feet. 2,602 second = 43 minutes. That's certainly a possible time to freeze some part of the 2,500 gallons each minute of pure water from ??? degrees to 32 degrees at -40 degrees air temperature, but isn't definitive.

Further, the water must "solidly" freeze - just a part of the water in the pipe "just beginning to change phase to crystals" isn't enough. Those early crystals will just get pushed out the end of the pipe if they don't stop the flow. Then again, as soon as the flow slows down, freezing becomes more probable.

 

[GD2]

Think about unplanned shutdowns, breakdowns and maintenance. No mechanical system can run 24/7. All risks and analysis must be taken into account in the Design. I support georgeverghese’s recommendation. It’s above ground piping at -40C/F.
For info, bare water supply water pipes in cold countries with same subzero temps are installed below the frost line.

GDD
Canada

 

[georgeverghese]

For 19000ft of pipeline, talk to a heat trace specialist firm to see it is better to go with self regulating or constant wattage parallel resistance or constant wattage series resistance style heat tracing - self regulating would be safest but may be more expensive that constant wattage type heat tracing. Suspect constant wattage series resistance would be the cheapest.

 

[LittleInch]

At that length (approx 5.5km) and time duration assuming some sort of wind I would be astonished if the water can get to the end without getting super cooled.

So it would gradually just get "slushy" and hence viscosity would increase, flow would slow down then viscosity increase a bit more....until it just freezes solid. You can run the numbers to see that but that's my instinctive response.

The issue of expansion and contraction you need to deal with anyway for an above ground PE pipe. PE is great stuff, but its thermal expansion is 10 times that of steel so needs proper consideration / expansion loops etc.

The insulation will just move with it.

Now if you insulate it and can keep some sort of flow going 24/7 you may not need trace heating, but it's a big risk. 14" is a relatively large surface are to volume ratio so heat loss even with insulation will be fairly rapid for a static system - probably 12 hours or less before it freezes. A 24" line might be quite different.

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

 

[TenPenny]

We have a bunch of pumps dewatering in the mines in Labrador, where it's -40 quite regularly.

If I recall correctly, none of the lines are insulated or traced, but they do have provisions for drainage when there is a failure or shut down.

I will do some checking to see.

 

[ECD40]

Thank you cvg, racookpe1978, Chicopee, GD2, georgerverghese and LittleInch for your comments. I'll add some additional information/clarifications as a result of the comments.

The project is a mine in Northern Canada. The pipeline will not be electrically traced, due to the shortage of power. The pipeline will not be buried due to the rocky terrain. The water temperature entering the pipeline will be about 50 degrees F (10 degrees C). The pipeline will be installed and commissioned in above freezing temperatures. The HDPE pipeline will expand about 120 feet between winter and summer average temperatures. Rivers still flow in the depth of winter in the minesite region. The lowest temperature on record at the minesite is -48.5 degrees C (-55 degrees F). Consider that there is no additional wind chill at the record low temperature.

So we are back to the question - will the water freeze, with or without insulation?

Your further comments will be appreciated.

ECD40

 

[TGS4]

I'm afraid that you are going to have to run the calculations yourself to find out. Either that or pay a specialist to do it for you. You're not going to find the definitive answer, for free, here.

And even if you think that you did get something that resembled a definitive answer - would you be willing to say to your boss - "Hey boss, I have a conclusion on that HDPE pipeline freezing issue. Some random/anonymous person on the internet website eng-tips.com said that we absolutely don't (or maybe you absolutely do...) need to insulate. I don't know if they're a professional, or even if they're an engineer, but they sure sounded certain of themselves. No, they're not going to take any liability for the potentially disastrous consequences (or costly consequences), because I don't know who they even are, or even what country they reside in. That's OK, I'm pretty sure that our professional liability insurance will cover this situation...".

Run the calcs yourself. You've been pointed in enough a right direction to figure out what calculations that you need to run.

 

[LittleInch]

OK,

curiosity got the better of me so I've done a very basis calculation.

I do it in metric because its easier for me, but feel free to poke holes or redo it in Imperial units

The thing that is probably saving you here is the PE....

Thermal conductivity PE is quoted on many locations as circa 0.4 W/m/K ( compared to steel which is 45), PE thickness at SDR 11 is 32mm for a 355 mm OD pipe

Plugging your values into a heat loss calculator - I used

http://tierling.home.texas.net/PipeHeatLossShell.htm

gets you about 400W/m (might be a bit less if you assume PE ID is OD in this calculator)
You've got 5750m so overall your water/pipe is losing 2.3 MW of heat along its length
Your water contents in a 14" pipe with 32mm PE wall is 384,000kg
Heat capacity of water is 4000 J/kg/C.

So total heat capacity in the pipeline per C is 1.53 x 10^9 J
Your water is in the pipeline for 2600 seconds so in total the heat loss during that time is 5.6 x 10^9 J. So you will lose between 3 to 4 C from one end to the other.

ON the same basis ( OK the heat loss will go down a bit as the temperature falls but the delta doesn't change too much) the time for a static water to cool from the now 6C is it at the end to zero is 66 minutes - basically an hour.

Do the maths yourself to make sure.

It does seem about right once you factor in 32mm of PE "insulation".

Make sure your incoming water is at or more than 10C... You don't have a lot to play with here, and keep the flow going or drain / clear it within an hour.

LI




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

 

[miningman]

The energy cost of heat tracing can become absolutely exhorbitant, especially if diesel generated. I dont see an allowance for wind chill cooling in Littleinch's calculations... again this can be high at minus 40 degrees and 40mph winds. Even if the line is partially buried by snow , there will be exposed areas. In my opinion , unless insulated, eventually the line will freeze. Perhaps due to abnormal / unforseen circumstances, but it will freeze. One technique that I have used successfuly in similiar circumstances, without heat trace or insulation is the provision and usage of "dump tees" in the line . Open all the dump Tees. Start the pump, and travel the line closing each Tee as the semifrozen ice and slush reaches that location. Travel along to the next TEE and repeatthe process. Might not be realistic over 19000 feet distance tho.

 

[LittleInch]

Wind speed is in the heat loss website calculation.

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

 

[cvg]

so in the case of shutdown (accidental or planned) you have about an hour at worst or maybe two or three at best to drain, at least partly before it freezes solid. since HDPE can expand without breaking, much of the pipeline could handle the expansion during freezing. however, good luck getting it thawed again until spring. so you would need to get on your horse and get out there in a hurry to open up the emergency tee drains and hope for the best. my risk assessment says that you need to plan for the worst and provide some sort of insulation and/or heat the pipeline

 

[Compositepro]

It seems that what is being overlooked in this discussion is that the latent heat of freezing of water is 80 times the heat capacity.

 

[LittleInch]

True, but what should happen is that the ice forms on the outside reducing the ID, hence reducing flow to the point where there isn't enough energy coming in to match the energy going out.

It does give you some more time OK, but will lead to "slushy" water.

If the coldest bit of the pipe takes an hour to drop 6 degrees on a stationary pipe = 24,000J/kg. Freezing energy is 333,000 J/kg. So about 13 hours for the whole pipe to freeze. Maybe a bit more as ice isn't as good a conductor as water. But whatever - it's less than a day.

Once it's frozen it needs that 80 times energy to turn it back into water. That would take a long time with your water at only 10 degrees.

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

 

[racookpe1978]

Hmmmn. The threat is obviously at the far end of the line, the section most distant from the origin.

So, if you cannot heat trace the entire line due to expense, heat trace the last 2000 feet. Heat trace that section actively with power ONLY when the temperature falls below -30 deg. In the low points of the line between 0.0 and 19,000, add a remote-operated drain valve to dump the water in case of a shutdown during low temperatures. The pipe will theoretically freeze solid ANYTIME the air temperature goes below freezing in static conditions, but you have time to restart, time to drain.

 

[miningman]

This is all getting a bit theoretical, and perhaps that should be expected in an engineering forum. However, before sticking his neck out and making a recommendation that might put his job and reputation on the line, the OP should consider the following.

How confident is he that the feed water will NEVER be below 10 degrees??. I'm making an educated guess that this is not a fresh water line, because if it is , the maximum temperature I would expect is 4 degrees. So, if it is "waste water" from a process plant ( mine tailings perhaps) what is the temperature variation in the plant when the ambient drops from say minus 35 (a normal cool day) to minus 45, as it almost certainly will at least twice every year. Similarily, how does the plant temperature change when the compressors etc that feed waste heat to the plant, go off line or least dramatically reduce the waste heat available between shifts?

Can he guarantee that the road access to the end of the line will never be blocked by snow drifts?? How often does the plant go down completely for maintenance , either scheduled or otherwise??.These and other practicalities need to be seriously considered.