How can I get rid of air buildup in a siphon system?

[OCR]

...have you eliminated the possibility of a small air leak into the hose?

He should have... ?

Also, if you can, check for any pin holes in the hose...

 

[Waterpro]

4yrs later what happened? Did we figure it out? A level hose without a smooth interior surface will trap air and over hours collect and stop flow. A smooth schedule 40 with 0 breaks fittings joints will solve it. Air is the worst enemy of water flow. Air is a Phantom plug that stumps most.

 

[KDunc]

On the original question - you could use a self acting pump to remove the trapped vapour/air .

Two possibilities :

(a) A venturi pump .

(b) A hydraulic motor driving a mechanical pump .

Edit : In fact a venturi pump fitted directly in the down flow pipe and a suction pipe going back to the highest point may be all that is needed .

Reviving this dead thread to ask about what you proposed here. I'm looking at a similar situation to the original post where I'll be using creek water with plenty of dissolved gasses, and the system will be located in a remote difficult to access area, so I want to make sure any accumulated gasses at the apex of the siphon are constantly removed so I don't need to regularly visit the site.

You suggested the use of a venturi at a point on the downward flowing pipe with the idea that it can pull the air from the apex of the siphon. My understanding of a venturi is that the pipe constriction causes a sharp increase in velocity through the constriction with an associated pressure drop. That pressure drop across the constriction is what would pull the air down through the suction pipe from the siphon's apex. But wouldn't the vacuum generated by the venturi have to be greater than the static vacuum that would exist at the top of the siphon in order to pull the air? Is this even possible in a system like a siphon? Since the action of the venturi is doing work in order to pull the air, what would the effective net head loss be across the entire siphon? That is, how much higher would the siphon intake need to be than the siphon outlet in order to have the net head needed in order for the venturi to do that work?

This all assumes that the flow through the siphon does not exceed the recharge rate of the source. If it did, then all if this is almost useless conjecture anyway, since the flow through the siphon would eventually slow to a stop as the source water was drawn down, preventing the venturi from being able to pull the air from the apex. There's still the potential for the gas to accumulate at the apex while the flow is low and the venturi is inoperable, eventually breaking the siphon. There might need to be the addition of a float and valve on the intake side that closes the intake when the source water level drops to a point where the net head difference across the siphon starts getting too small to effectively operate the venturi.

 

[Baluncore]

@KDunc Welcome to PF.
You seem to understand the requirements.
1. A valve that will close the input before air can enter the siphon tube.
2. A U-bend or water trap at the outlet that will prevent air running back up the siphon tube.
3. Some way to prime the siphon tube.
4. Some way of removing the gasses from the near vacuum at the top of the siphon tube.

A venturi is not a solution to 3 or 4.
The solution will be situation and detail dependent.

 

[KDunc]

@KDunc Welcome to PF.
You seem to understand the requirements.
1. A valve that will close the input before air can enter the siphon tube.
2. A U-bend or water trap at the outlet that will prevent air running back up the siphon tube.
3. Some way to prime the siphon tube.
4. Some way of removing the gasses from the near vacuum at the top of the siphon tube.

A venturi is not a solution to 3 or 4.
The solution will be situation and detail dependent.

Thanks for the reply. I thought as much - a venturi seems like it wouldn't work, and certainly wouldn't be able to compensate for the dynamic nature of such a system. Proven methods like a vacuum pump and/or some way to regularly re-prime the system are likely the best way to handle this situation.

To continue to explore out-of-the-box methods: a common method for priming a siphon is to close the inlet and outlet valves, then open a valve at the apex of the siphon and completely fill the siphon with water from this apex point. What if we left this apex valve open and connected it via a small diameter pipe to the bottom of an open-top tank that had a continuous flow of water coming into it? The siphon effect would also aggressively pull water down from the tank, but the small diameter of the intermediate pipe would restrict this flow of water. So long as the apex tank was recharged faster than the outflow into the siphon, then would this be an effective method to prevent the accumulation of gas at the apex of the siphon? The issue I see is that the intermediate tank-to-apex pipe would need to be small in order to decrease the tank recharge needs, but the low diameter of this pipe may also prevent the gas from flowing through it in the reverse direction, preventing its escape from the siphon. This idea would only be useful in the event that a small stream is available to supplement the siphoning of a much larger creek.

 

[anorlunda]

It seems ironic that your description of a small diameter tube at the apex is described as a way to prevent air buildup, but it is also the design of a siphon break device intended to let air in.

 

[Baluncore]

It seems ironic that your description of a small diameter tube at the apex is described as a way to prevent air buildup, but it is also the design of a siphon break device intended to let air in.

An automatic air bleed valve is placed at the crest of a pumped water pipeline to prevent the crest being blocked by compressed air. That air bleed valve is closed by the higher density of liquid water.

 

[Baluncore]

What if we left this apex valve open and connected it via a small diameter pipe to the bottom of an open-top tank that had a continuous flow of water coming into it?

There are several similar possibilities that require a supply of water at the crest.
I gave some solutions way-back in posts #2 and #7, so will not repeat them.

I discussed foams in post #22.
I still wonder if it would be possible to install a foam generator at the siphon crest. A protein foam could help carry the gasses away in the flow. It would be activated by the presence of gas. Obviously, the end-use of the water would determine the choice of foaming agent.

 

[Baluncore]

Ideally, avoid power operated pumps, especially those pulling a vacuum.

There is a way to vent the crest of the siphon by pushing water only. Start with a flap valve at the input that will prevent flow back to the source. At the crest, provide an automatic air bleed valve that will release gas whenever the pressure in the siphon rises above atmospheric. Now, at the outlet put a stop valve, with a plunger in a cylinder connected above the stop valve to the falling siphon tube.

Turn off the outlet. Maybe the sudden stop will cause the water in the rising tube to ram gas at the crest out through the automatic valve. If not, cycle the plunger. Since the water is not compressible, the water will rise at the crest, pushing gas out through the valve. Open the outlet valve to restore flow.

Given a little more fall at the outlet, it may be possible to arrange the outlet valve to cycle like a ram pump, ejecting gas from the crest of the siphon.

A plunger in a cylinder after the input flap valve could be used to prime the siphon.

 

[sophiecentaur]

It seams as if a conventional dosing siphon would maintain your levels.

Would a dosing syphon work in this case? For a dosing syphon to work (look at the above link), the supply level needs to be above the high point in the inverted U at some time. Can that happen in a simple system where the upper reservoir level is always lower than the top of the U? I remember the standard system in ancient men's urinals (the urinals not the men) was a dosing syphon and the tank was constantly filled from a mains pressure supply. That doesn't seem to be the case here.
In fact, if this happened in the OP's arrangement then wouldn't there always be a flow?

 

[sophiecentaur]

It seems ironic that your description of a small diameter tube at the apex is described as a way to prevent air buildup,

Bearing in mind that the pressure at the top of a syphon can be almost zero then I can't see how, without some 'special help' that can ever fill with water.

I did wonder whether, if the water flow is fairly fast, then a Venturi pipe, lower down the hill , could 'suck' air out of the U. There would need to be a one way valve to stop air getting back up the tube and making the problem worse.

Or perhaps a ram-type of mechanism could act as an air pump.

 

[Baluncore]

I did wonder whether, if the water flow is fairly fast, then a Venturi pipe, lower down the hill , could 'suck' air out of the U. There would need to be a one way valve to stop air getting back up the tube and making the problem worse.

In a steady state flowing system, there is a logical problem with duplicating the outlet tube from the crest with a longer tube to a venturi. Rather than solving it, that duplicates the problem. If modification of an outlet tube could cure the problem, then one should modify the outlet of the existing tube.
Remember how, unlike water, the hydrostatic pressure does not rise in a tube filled by partial vacuum as it falls towards the outlet.

Or perhaps a ram-type of mechanism could act as an air pump.

As I suggested in post #39, the cyclic pulsing of a ram pump could be used.

Given a little more fall at the outlet, it may be possible to arrange the outlet valve to cycle like a ram pump, ejecting gas from the crest of the siphon.

The two effective columns of water, separated with the partial vacuum at the crest, could generate a cyclic positive pressure pulse at the crest, blowing the partial vacuum and a little water out of the air bleed valve at the crest each time. In effect, the siphon tube forms the drive tube of the ram pump, with the ram pump outlet valve at the crest. While generating a one bar pulse would be a trivial job for a ram pump, the system must still be primed somehow to start it operating.

The siphon tube would need to be straightened, with a more rigid tube towards the outlet valve, where the initial pressure pulse will appear each time the spring-loaded outlet valve slams shut. I would expect it to cycle between five and ten times per minute. Perhaps waste water ejected at the crest could be accumulated to aid in priming the siphon next time it is needed.

I think the self-resonant ram-siphon is the best candidate yet for ejecting the partial vacuum gasses from the crest of a siphon. Notice that it operates only by liquid-water pistons against valves. It avoids any attempt to pump or pull a deeper vacuum at the crest.

 

[sophiecentaur]

In a steady state flowing system, there is a logical problem with duplicating the outlet tube from the crest with a longer tube to a venturi.

Is that true? The two pipes are not equivalent. One has a large volume of water, traveling fast and the other will initially have a small volume of water (initially), traveling slowly. I can't see why water will not be drawn into the Venturi section. In a static situation then obviously there will be no flow but is there there not a difference in the available power which can support the pumping action? Are you saying that the pressure difference in the two pipes will not be enough and that the level in the small tube will never reach the Venturi section? That could make sense but the Venturi section need only be a small distance below the crest. Also, the bubbles in the main pipe will reduce the water density but they could be arbitrarily few - just enough to balance the rate of formation.
I do realize that this is not a self priming system.

I missed the Ram suggestion but it sounds like a better idea - although it still wouldn't be self priming. I think there will be no way round that problem.