Water hammer mitigation in 13,000' 2" HDPE DR 17 water line

[Oasis Design]

Hello,
I hope this message finds you healthy and well.

I’m helping with the design of a long water line for an improverished non-profit in the high desert in Southern California.

I have good general water design skill but little specific water hammer expertise. Would you be willing to give the design a quick sniff test for them, and maybe point us in the right direction if we’re way off of the reservation?

If so…here you go—

• There’s 4300' of 2" buried DR-17 HDPE tubing, followed by 9000 of 1" buried DR 11 tubing.
• The max flow velocity in the 2 in is 4.2 ft/sec, in the 1" about half that, so the 2 in seems like the more critical application.
• The line runs down a 2%± continuously sloping valley. The max pressure in the 2” (which is PE4710 rated at 130 psi) is 79 psi, as it carries on in the 1” (PE4710 rated at 200 psi) it reaches 126 psi in the low end.
• Is there even a problem, or is the the nature of P44710 resin to just take sudden velocity changes in stride?? PE4710 resin in DR17 can take a 5.6 ft/ sec recurring sudden change in velocity, and 11.2 ft/sec occasional sudden change in velocity, according to

https://plasticpipe.org/municipal_pipe/advisory/hdpe/pressure-rating.html

On the other hand, online calculators suggest pressure spikes of >1000 psi and valve closing times of 10 seconds (screen shots below) what gives??---more detail:
• Most joints are to be fused, with some compression fittings at the transitions to galvanized above ground taps, and some fused transition nipples to pipe thread where water hammer is most likely.
• The system alternately freezes and fries, with 50°F temp swings every day.
• There's not much money for the install (it's only happening because of a grant) and maintenance can be iffy.
• All the regular operation valves are hose bibs where a handle needs several twists to stop the flow.

My concern is the drain valves on the 2", which are brass ball valves that will generate max, open discharge velocity when open, and could be closed in a fraction of a second by inexperienced operators.

I wonder too what happens when the drain at the very end of the 1" opens, and 4300' of 2" water battering ram adds to 9000' of 1" battering ram; can't quite picture this.


Online calculators suggest a valve close time of 10 seconds.

Here’s what I’m thinking of doing to mitigate:
• Put big signs on every valve saying to take 10 sec to close valve to prevent system explosion
• Pressure relief valves, possibly everywhere there’s a valve (there's about 15). A bit nervous about freezing, but this seems cheap and easy...is this effective?
• Air reservoirs in blind-ended uphill pipes, teed off the system at drain valves, of the same material in the same trench, sloped uphill so they’d catch air and hold it (and also drain when the valve was opened.) I’m guessing 20-50' of 2" in tubing at each of the two drain valves? This would be inexpensive and low maintenance. The cap at the end would be fused, so unlikely to leak. These could be spliced in anywhere there is a valve, so there would be shock absorbers everywhere. They would only be sloping about 2%, like the ground, though they could be sloped more in the trench as we’d not be so concerned about freezing or traffic. ANY GUIDANCE ON SIZING THESE? We could make them as long as we want. I realize these will slowly fill with water, but they can be made so they will drain when the system is drained every year or two.
• Bladder tanks—worry about freezing, maintenance, cost
• We could replace the ball valves with twist handle valves…but these would be especially problematic for the drains
• I wonder if the nature of HDPE will save us, by expanding as the pressure spikes then relaxing.
• Perhaps the compression fittings may pop off the ends of the HDPE at transition to galvanized, providing a softer failure mode than the pipe exploding (water is fairly abundant, so losing a tank of water would not be as bad as the line breaking)
• We're planning on going with DR 17 for the 2" line, rated at 130 psi, rather than DR 11, to lower cost and the amount of plastic (this saves $1300 and 2,000 lbs of plastic). This is a bit non-conventional, but fits the strong environmental ethics of the place. According to the online water calculators

Thoughts?

Thank you!!!

Art Ludwig

 

[TiCl4]

Change the 1/4" turn valves out with wheel-operated gate or plug valves. Even a very enthusiastic operator should take longer than 10 seconds to close a 2" gate valve.

PSVs are not adequate for relieving shock-waves generated from water hammer. They do not open quickly enough to dissipate the energy. Rupture discs can be used, but carry the obvious headaches. If using wheel-operated valves is not an option, then you would need to install an expansion tank. Expansion tanks with no bladder (air is in contact with the water) can be a concern because the air will slowly dissolve into the water over time, lessening the capacity for shock absorption.

I would strongly recommend wheel-operated valves. It is the most sure solution (avoids the problem in the first place) and is also likely the cheapest.

 

[LittleInch]

Oasis,

I'm struggling here to work out your design and flow pattern - Can you draw a diagram?

How is the flow velocity in the 1" HALF of the 2"?? Doesn't make sense unless a lot of water goes off at the 2" to 1" connection

How many offtakes do you have ( the 15 you mention?) and will at least one of theme be open at any one time?

I'm struggling to see a problem here at that velocity. PE is remarkably flexible and attentuates pressure surges much better than steel. Also it's pressure rating is a long term 50 year creep rating. Burst is about 3 times.

You need to work out the max flow rate / velocity for when only ONE connection is in operation.

As soon as you get more than one connection open then your pressure pulse just disappears down the open route and doesn't reflect up.

The other way is therefore to always have a small flow leaving at the end of the pipe, in part to flush water out continuously and also act as a pressure relief. You loose a bit of water, but it could be a really small amount.



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

 

[TiCl4]

LI, I was under the understanding that these takeoff valves normally operated closed, and the scenario he is worried about is 1 end-user opening, then shutting the valve, resulting in a pressure spike. The flow velocity may be small, but the line length is huge, resulting in high pressure spikes if flow is suddenly halted. If multiple users are open at the same time, then the scenario becomes much less dangerous as pipe flow is not brought to a standstill.

I have no idea how well PE can attenuate shockwaves, but I'd rather the pipe not be exposed to a shockwave that exceeds its design pressure by >5x along ANY length.

 

[LittleInch]

There isn't the velocity to cause a problem.

I suspect that second table in the OP is based on steel pipe.

The other graph in the plastic pipe manual shows for 7fps in SDR 17 youre pressure rise is only 80 psi.

The OP has only 4.2 ft/sec.

There isn't a problem here to solve.
IMHO
Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[TiCl4]

LI,

It has always been my understanding that it isn't just the velocity that is an issue. If the entire column of water (all 4500' of it) is moving, then the entire column of water is suddenly stopped, all of that momentum has to be stopped. I.E. the length of fluid flow is important and is a factor in the pressure development. I my understanding of this incorrect?

 

[1503-44]

LittleInch
Where do you see 15?

Oasis

I think you have a problem, but waterhamer I think is not one of them.
A diagram would be of great help here, just to verify a few things.
We are trying to help you, so please help us do that.

"The max flow velocity in the 2 in is 4.2 fpss, in the 1" about half that, so the..."
Did you get the 4.2 fps from the pipe flow charts?
Did the pipe flow charts also say that the velocity in the 1" is about 1/2 that in the 2"?

As long as I'm asking questions,
What is your real sustained flowrate? 4.2 fps is 38 gpm. that may be a (very) temporary max?
What is the pressure at the inlet of the 2" Is it about 0 psig? Or do you have some pressure there?
What is the pressure at the inlet and at the outlet of the 1" when you have sustained flow?
What is the pressure at the inlet and at the outlet of the 1" when you have NO flow?

When you have NO flow, it could be around 79 at the 2" outlet (but I get less) and 126 psig at the 1" outlet as you say, but I think that those pressures may drop when you have a sustained constant flow rate, right?
I may be wrong, because you didn't give any good elevation data. I used a constant 2% slope as you said, but I think it is not 2%. Maybe 1.5%? Or maybe it is 2% for the 2" and 1% for the 1" ? 0.5% doesn't sound like a lot of difference, but for these small diameters, it could make a big difference in flow rates.


“What I told you was true ... from a certain point of view.” - Obi-Wan Kenobi, "Return of the Jedi"

 

[LittleInch]

ax1e,

From the OP - "...possibly everywhere there’s a valve (there's about 15)."

TiCL4 - Not mistaken exactly, but often a longer length has a benficial effect on the pressure rise. Veleocity is the key factor due to the energy equation (1/2 mV^2).

Whilst water is "nearly" non compressible, t does compress and has a bulk modulus. Similalry pipe is not super rigid, but expands very slightly with pressure. PE has a much lower Youngs modulus so expands more than say steel or copper for the same pressure/ stress. PE abut 0.8, steel about 200 GPa. OK, PE is thicker for the same pressure rating, but only by a factor of 15 to 20, not the 200 for YM.

Also think of it this way. The column of water is moving. The molecules at the back of this column don't know what is happening at the front. The only way they find out is a pressure wave sent by the action of the valve closing. This travels at the speed of sound in the fluid - in water it's about 1400 m/sec.

That pressure pulse reflects off the start point and if the reflection makes it back to the valve being closed before it is fully closed then this acts to reduce the pressure wave. Pressure at the end point will climb in order to slow the column of water and also settle out the pressures int he pipe back to their static pressure. As it does so the water compresses a little bit and the pipe expands a little bit. Both of these reduce the pressure rise. Hence why valve closure speed is important.

In long pipes / pipelines ( say> 20-30km) the pressure wave is attenuated as it travels due in part to the small energy taken to do these things and bends and other changes in direction also reduce the pressure pulse.

But it is the velocity which kills you. Much above about 3.5 to 4 m/sec in short lines you have an issue you actually need to solve. Below that and especially in a flexible line like PE, I really don't think there is a problem.



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

 

[1503-44]

I just couldn't see that 15 anywhere, but yeah, there it is. Today I see it. I was looking at it far too late last night. Thanks.

Oasis,

Close all valves branching off from the 2" and 1", except for the inlet into the 2" and outlet valve on the 1" (if they exist).

“What I told you was true ... from a certain point of view.” - Obi-Wan Kenobi, "Return of the Jedi"

 

[bimr]

OK, had my coffee and revised this.

I agree with ax1e (Petroleum) and think that you have problems, but water hammer is not one of them.

The title of your post states "13,000' 2" HDPE DR 17 water line" feet, but the text states "4300' of 2" buried DR-17".

The headloss in 4,300 feet of plastic is 60.8 psi. However, you state the maximum pressure in the 2-Inch line is 79 psi at the downstream end. That would equate to 79 psi -37.3 (elevation head) + 60.8 psi (headloss) = 102.5 psi at the upstream end. Suggest that you revisit your calculations.

The problems that you will encounter are related to the startup and shutdown of this pipeline.

Link

 

[KernOily]

I would do as the others suggest and just put in slow-closing valves such as gate valves and also put the sign at each valve.

I think you may have other problems too. I didn't see any data on elevation change, but you may have a slack flow problem if there is enough elevation change between two points anywhere along the line. Check your hydraulic gradeline and if the pressure drops below the vapor pressure for water at the temperature in question, you will have vapor pockets form and slack flow. Say on a hot sunny day, and if the water temperature increases due to low flowrate, if the charge pumps trip out (power failure for example) or if the source pressure suddenly drops, you may get slack flow with the resulting large pressure spike if the vapor bubble collapses. Might be enough to rupture the HDPE. Also, as HDPE pipe gets hot, it loses strength FAST. It has a pretty step derating curve vs. temperature. How is the line charged: Pumps? Pressure vessel charged by well pump? Atmospheric tank and elevation head?

I've been where you're at, doing projects just such as this with nearly zero budget, donated materials, and volunteer labor. Very rewarding but very challenging. Best of luck to you.

 

[btrueblood]

Buried pipeline will swing 50 deg. F? Not sure about that value, the temperature of the soil 2-3 ft. below grade might vary a bit seasonally, but won't swing that much on a day/night basis. If this is at a high enough altitude that the buried pipe may freeze, then are you planning to drain the system and shut down during the winter?