Siphon 2

[stanier]

Hi,

I am seeking advice in respect of a planned syphon. The system has the following levels.

Dam level RL138m 30+m to dam wall
Dam wall 146m 15m over wall
Pipeline Outlet RL 122m Outlet is 130m from top of dam wall
Temperature 30C
Water
Pipeline DN710 PN 12.5 PE

Pipeline is relatively short. The system will be primed by a pump, isolated and then the siphon opened.

The software I am using predicts vacuum pressures greater than atmospheric at the high point through the dam. The gravity flow will be limited by the friction losses in the pipeline but velocities are still predicted at 6-7m/s.

My question is will the system be unstable as the water will boil at the high point? Or will the fluid move so fast as to drag any water vapour bubbles with it?

I have designed and built a similar system but it was connected to a pump suction that controlled the flow. It works perfectly. This system is just top drain a dam over the dam wall.

I am thinking a restriction at the outlet to provide some back pressure may keep the system pressure above the vapour pressure of the water.

ôThe beautiful thing about learning is that no one can take it away from you.ö
---B.B. King

http://waterhammer.hopout.com.au/

 

[chicopee]

I think that a sketch would help a lot. I would say keep the syphon outlet a few feet below the tailwater which is a similar concept to keeping the draft tube end under water with hydraulic Osberger and hydrolec turbines.

 

[racookpe1978]

Where is the control valve (emergency shutoff) for the siphon? The lower valve that is opened to start the siphon flowing after the vacuum is pulled? What time is expected to fully shut?

Yes, please confirm with a sketch what I think I (mis)understand from your description!

 

[katmar]

This is one of those problems that can be analyzed to death, but the real answer can only be obtained from experience, which unfortunately I do not have in this case. If the water does boil the bubbles will surely be swept along with the liquid. Because of the temperatures involved only a small percentage of the liquid will be vaporized. So the problem is not that the vapor would break the siphon. The potential problem I see is when the bubbles recondense as the pressure increases and you get cavitation. This is the bit that I believe can only be answered by experience - will the cavitation damage the pipe? A little bit of air in a siphon is not a problem because it stays as gas all the way through, and can be safely swept out. Water vapor formed in situ is a different matter.

But what we can say from pure theoretical analysis is that you will not get a velocity of 6 m/s. If my guess that you pipe ID is 530 mm and if my interpretation of your layout is correct then at 6 m/s the pressure drops from the inlet to the downstream end of the wall are

pipe friction - 15 kPa
inlet loss - 9 kpa if square or 1 kPa if bell mouth
acceleration loss - 18 kPa
static loss - 80 kPa
vapor pressure - 5 kPa

This gives a total loss of either 127 kPa or 119 kPa (depending on the inlet configuration).

The absolute pressure cannot be negative so this is an impossible situation. As a limiting case to avoid any vaporization we could specify that the centre line pressure as the pipe goes over the wall is 10 kPa - giving 90 kPa driving force. Unfortunately the static head does not decrease with decreasing flow so we really only have a small amount of head to play with and the flow rate (assuming ID = 530 mm and bell mouth inlet) has to be reduced to 2540 m3/h or 3.2 m/s.

In order to limit the flow to this rate you could introduce a valve or restriction orifice at the outlet, or if the pipeline has not been built yet then some money can be saved by installing a smaller diameter pipe downstream of the dam wall (maybe 350 mm?).

The difference between this scenario and the one where you allow the water to boil gives some additional driving force (95 kPa instead of 90 kPa) and the flow rate increases from 2540 m3/h to 3130 m3/h. The benefit of this extra flow has to be balanced against the potential risk of damaging the pipe with cavitation - and I do not have the experience to do that.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[chicopee]

Rereading your post "The software I am using predicts vacuum pressures greater than atmospheric at the high point..." I think that you meant "less" not "greater". You have not stated what that partial vacuum value is. Assuming a water temperature of 50 dF, the saturation pressure is .178 psia which is a low partial vacuum to get in your syphon for bubble formation.

 

[dcasto]

its hard to visualize the dimensions, but you are not ever going to syphon up over 10 meters.

 

[bimr]

It should start to cavitate at about 7 meters of siphon. One would doubt that you would be able to obtain 6-7 m/sec of velocity.

The cavitation would probably erode the piping. Cavitation is a problem at some dams.

 

[cvg]

approx 50 ft head and 24 inch pipe, seems like it will have at least 25 feet/sec or more. I would think that air accumulation would not be a problem. throttling the downstream valve will reduce your headloss a bit, but will also reduce your velocity. not sure if will be enough

 

[bimr]

The maximum practical limit of a siphon is about 8 meters for water (at sea level). You will have to put a restriction to keep the pressure from dropping below the vapor pressure of water, if you want to keep the siphon stable.

The water velocity will not work to flush out the air because the maximum driving force reached is a vacuum, no matter how large a drop there is over the dam.

 

[katmar]

A velocity of 3.2 m/s in a 530 mm ID pipe gives a Froude Number ( = V / (gD)[sup]0.5[/sup] ) of 1.4. This is plenty to ensure that any air is flushed out of the pipe.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[stanier]

Here is a sketch of the system

ôThe beautiful thing about learning is that no one can take it away from you.ö
---B.B. King

http://waterhammer.hopout.com.au/

 

[bimr]

Have to disagree as the velocity of 3.2 m/s is too slow to flush air bubbles out of a vertical pipe. Yes, if the pipe was horizontal it would flush, but not vertical. It takes much more velocity to push the bubbles downward. See the attachment.

 

[bimr]

It appears that you are missing the elevation of the top of the siphon where it passes through the dam.

 

[stanier]

Stage 1 is RL 148m and stage 2 RL 146m.

I have remodelled this system using a flow control valve to keep a back pressure at the high point. It can keep the vapour pressure in the high section below that for water at that temperature. This will maintain stability. When built we may try and open the valve and see what happens in respect of it getting air locks.

Thanks for all the input.

ôThe beautiful thing about learning is that no one can take it away from you.ö
---B.B. King

http://waterhammer.hopout.com.au/

 

[katmar]

stanier - you had not mentioned the non-return valves on the inlets before. I had not made any allowance for this in my earlier post, but I guess they are necessary for priming and I should have foreseen them. Pressure drops vary greatly between different types of NRV - choose carefully.

bimr - I do not recognize the reference you posted, please can you give the author and title. My "go to" reference when it comes to siphons is the article "Sizing Piping for Process Plants" by Larry L. Simpson (Chem Eng, June 17, 1968, pg 192-214). This is getting a bit old and does not cover sloped (non-vertical) piping, so if you have something more recent that would be useful. Simpson's values for vertical pipe give very similar values to your Table B-9, i.e. a velocity of about 2.5 m/s in vertical pipe of 525 mm ID. As this is less than my calculated velocity of 3.2 m/s I am confident the air would be flushed out. Especially now that we have seen the layout sketch and there is no vertical downward-flowing piping.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[stanier]

The check valve has a CV 19240 and head loss would be minimal.

The solution I have come up with uses a DN400 Crane Flow Seal butterfly control valve operating at 30 to 60 degrees open to control flow from 2500 to 4100m3/hr in 200m3/hr increments based upon increments of dam level changes of 1m.

Velocity ranges from 2.5m/s to 4m/s. ie As the dam level increases the flow is allowed to increase. This is in line with the project objective of emptying the dam as it takes run off.

Commissioning will be used to determine if greater flows can be achieved without instability.

I recall now doing a similar installation for a dam supply to a water treatment plant some 25 years ago.


ôThe beautiful thing about learning is that no one can take it away from you.ö
---B.B. King

http://waterhammer.hopout.com.au/

 

[LittleInch]

Just a couple of questions/ comments from me:

1) What is RL? Relative level? What is the actual elevation above sea level - This is critical for these applications - if 123 s mASL then you're OK, but a dam at a few hundred metres ASL is a different matter.

2) I'm kind of intrigued by your priming system - the same pipe having air in one direction and water in the other will need to be fairly big - two separate ones would look better to me

3) With PE I would be less worried about cavitation than other materials - it is quite resilient to this

4) If 138m s your low level, then if it only works well to 139m is that really an issue?

5) Moving flow I think will be less of an issue than starting a flow at a level of around 138m. It think that is when you could struggle to develop a flow.

6) Do everything you can to reduce flowing losses on the dam side of the system, every 1kPA will make a difference.

My motto: Learn something new every day

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

 

[stanier]

LittleInch,

RL is "reduced level" and is a local datum. AHD is Australian height datum but is not used in engineering often. The dams are not very high above sea level so elevation is not a concern.

The priming system works on other similar systems installed but they are used for priming pumps rather than purely a gravity system. The idea is to fill the pipeline to the high point then open the FCV and system will prime.

The system will work down to 138m then stop. Flow will not be started at an RL of 138m, rather 148m. The point is the system is to empty a dam not to provide water to a resource. Hence when it has reached 138m the system has done its job and the valve will be closed. As the dam fills to near RL148m the system will be started again.

Inlet system has been designed for low losses. The inlet is a device like a mussel. Two domes separated so flow is low velocity in a horizontal direction then accelerated vertically. The aim is to keep from disturbing any solids on the dam floor and to eliminate vortexing.

ôThe beautiful thing about learning is that no one can take it away from you.ö
---B.B. King

http://waterhammer.hopout.com.au/

 

[LittleInch]

OK, thanks for info. I think you'll be OK with that design and operating system.

I know it's off topic a bit, but my point about the priming system was about the connection shown on the diagram at the top of the dam which has a single pipe and an AV and CV (air out, water in at the same time?) and if the pipe isn't big enough you'll get some interesting effects... Just a thought.

My motto: Learn something new every day

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

 

[bimr]

katmar, the reference is Pumping Station Design by Garr M. Jones.