Venturi / Ejector Design to Boost Water Flow using a Pump

[rollingstein]

I have a water tank fed from a piping system that has very low head in it. As a result flow tends to be low. See sketch below. I am thinking of ways to boost the flow. I already have a centrifugal pump at the site. Could I use its flow to boost the flow from the incoming pipe via a venturi / ejector design?

As a DIY project what might be a good way of creating the venturi? Would a simple tee work? Or would I need a converging diverging nozzle (e.g. see the sketch below)? Any ideas how to chose the nozzle dimensions?

 

[Baluncore]

How long is the low pressure input line?
What is the height difference or the delivery head?
What is the diameter of the low pressure input line?
What is the peak flow now?
What peak flow do you require?

 

[ToolReview]

From out in the real world -

I played around with suction dredges for hobby gold mining for 30 years. Net water lift in that application is only about 12 inches but additional suction is needed to bring up beercan-sized rocks with enough velocity that they don't stall in the intake hose or stack up and jam in the area beyond the venturi.

For a suction dredge assuming a primary pipe diameter of 2 to 4 inches and a centrifugal pump of 1.5 to 5 hp, I found after a lot of experimentation that water pressure a little over 30 psi and a jet orifice of 5/8 inch is optimum. Keene Engineering sells 'Power Jets' for $100 that will save you a lot of cut-n-try engineering. Recommended, for what you are trying to accomplish.

http://www.keeneeng.com/mm5/merchant.mvc?Store_Code=KES&Screen=CTGY&Category_Code=PJ

Now I'm considering making a venturi system like you describe, to speed up the gravity-feed filling of the tank I tow around to water new orchard trees. Lift is negative, the trailer is slightly lower than the storage tank, but I want to minimize fill time.

Perhaps someone smarter than me can present the math needed for optimum design. Simply based on experience, it seems to me that you don't want a reduced diameter in the pipe beyond the jet. For gasses a reduction may improve the 'traction' that the jet applies to push the medium. But water is incompressible and has inertia, so a reduction in diameter seems to me to reduce efficiency. In fact the additional water added by the jet has to pass that area, another argument for not reducing the diameter. Undetermined is how long the pipe has to be downstream from the jet, for the energy from the jet to be imparted uniformly to the entire amount of water in the moving column. I think Keene's jets mentioned above are generally too short for maximum efficiency feeding the intake of a gold sluice box. At least in my own designs optimized for backpacking, there was sometimes water at greater velocity recognizable in the output to the sluice box, indicating that not all the force from the jet had been applied to the column of water.

When you figure out an optimum solution for your application I would love to see it, and apply it here to my watering rig!

 

[Baluncore]

Without some idea of the dimension of the problem it is hard to specify a venturi solution.

A venturi injector can generate a significant pressure step in a fluid. There is little advantage in boosting the pressure close to the outlet since suction is limited by atmospheric pressure. That is especially true when the line is long and the diameter small. It is therefore better to move the injector as far as possible towards the start of a long input line. To do that consider laying a larger diameter input line, if that alone does not resolve the problem when used in parallel with the original, use an injector at the larger pipe inlet driven by the centrifugal pump reversing the water flow along the original input line. That will maximise the advantages and economy of using a pump.

There is a common hidden problem with low head lines. They are often unable to purge themselves of air. That air forms a bubble that is effectively flowing in the opposite direction to the water and so restricts the flow. Avoid humps in a low flow line.

 

[insightful]

Schutte & Koerting are the "go-to" folks for this:

http://www.s-k.com/pages/pro_10.html

 

[rollingstein]

How long is the low pressure input line?
What is the height difference or the delivery head?
What is the diameter of the low pressure input line?
What is the peak flow now?
What peak flow do you require?

Length of Low Pressure Line = approx. 500 m.

Diameter = 1 inch

I don't have access to the upstream end of the line for pressure measurement etc. At the downstream end the current head is very low. Approx. 6 ft of water.

Desired flow is as much as reasonably possible. Right now the flow seems to be slugging. i.e. Instead of a continuous stream of low pressure it comes in slugs.

 

[rollingstein]

There is a common hidden problem with low head lines. They are often unable to purge themselves of air. That air forms a bubble that is effectively flowing in the opposite direction to the water and so restricts the flow. Avoid humps in a low flow line.

Thanks. Are there any other solutions for this? The line already exists and relaying it is not really an option. Most of the line is buried underground so not sure of the exact lay profile.

 

[rollingstein]

There is little advantage in boosting the pressure close to the outlet since suction is limited by atmospheric pressure.

Isn't there a strong suction at the throat of the venturi? Right now the downstream pressure would be atmospheric and I was hoping that if that dropped to a vacuum I could derive up to 1 atm of additional driving force.

How much this would boost flow I am not sure since I do not know what the exact upstream pressure is.

 

[Baluncore]

Right now the flow seems to be slugging. i.e. Instead of a continuous stream of low pressure it comes in slugs.

That is consistent with, and typical of, air in the line. The ideal situation is when the flow is continuous, the velocity builds up and the momentum keeps it running. Then air is unlikely to pool at a high point in the line.

Where does the feed water come from?
Is it permitted and possible to push water back up the inlet pipe with the pump to eliminate the air?

If the inlet pipe is fed from a pond then obstruction of the input can cause flow rate pulsing, as can a low water level but then air comes out also.

 

[Baluncore]

Isn't there a strong suction at the throat of the venturi? Right now the downstream pressure would be atmospheric and I was hoping that if that dropped to a vacuum I could derive up to 1 atm of additional driving force.

Yes. But suction only gives atmospheric pressure maximum.

A venturi injector can develop an output pressure well over 100 psi. But it can only generate 14.5 psi suction. Maybe 1 atm of suction from the centrifugal pump is sufficient to pull the air bubbles from the line? Recirculate some water around the pump to keep the pump filled with water to maintain suction.

Ideally an automatic air vent would be installed at the crests in the line where air bubbles settle.

 

[ToolReview]

The parameters are: 1 inch pipe at the outlet end and you don't have access to the inlet. Only 6 ft of drop in 500m. 'Slugging', surging, = flow intermittently stops.

Try a simple solution first. I would skip designing a venturi for now and try connecting that 1 inch pipe directly to your pump inlet.

Practical issues: Everything on the pump inlet side must be absolutely airtight.

And ... the friction losses in a quarter mile of 1 inch pipe will be great. I wouldn't expect a drastic increase of output.

 

[ToolReview]

Recirculate some water around the pump to keep the pump filled with water to maintain suction.

I don't see how to do this. When the pump pulls air for a moment then recirculating that air from its outlet would break the suction.

Maybe keep the pump primed using a second pump that draws from the destination reservoir?

I think in a practical application the inertia of that quarter-mile column of water arriving in the pipe would soon re-prime the pump. (Assuming a pump built to tolerate running dry).

 

[insightful]

From the flow tables: Water with a 30 ft head will give about 5 gpm in a 1 inch, 1640 ft long sch 40 pipe.
Do you know what your average flow is now?

 

[ToolReview]

From the flow tables: Water with a 30 ft head will give about 5 gpm in a 1 inch, 1640 ft long sch 40 pipe.
Do you know what your average flow is now?

At the downstream end the current head is very low. Approx. 6 ft of water

 

[insightful]

Sorry, I don't know what a "current head" is.

 

[ToolReview]

Water at each end of a pipe will settle to the same level. I think in this case he means water will flow out of a garden hose as he raises the tip until the tip is 6 ft higher than the outlet end of the pipe. At that level the flow stops. So by inference the source water level is 6 ft higher than his outlet.

 

[rollingstein]

Sorry, I don't know what a "current head" is.

Sorry, "current" as in, currently, right now, before any modifications. Status quo.

 

[ToolReview]

For rough estimating here: outlet pressure in psi is about half the head in ft.

More precisely:
Engineering Toolbox
1 ftH2O = 0.4335 psi

 

[insightful]

So, at 6 ft of head you should get about 2 gpm. Add 30 ft with an ejector (if possible), you still won't get much more than 5 gpm.

 

[ToolReview]

I don't think any sort of eductor and pump will pull water from that pipe at a velocity/pressure suitable for practical use.

Alternate solution: a large reservoir or low-profile tank at the outlet end filled by gravity (slowly) from the pipe. From there, a pressure pump to deliver water to your application.

 

[rollingstein]

Alternate solution: a large reservoir or low-profile tank at the outlet end filled by gravity (slowly) from the pipe. From there, a pressure pump to deliver water to your application.

How is that different from my present setup?

 

[ToolReview]

Not different.

In response to the original question, I don't think there is a practical way to increase the output volume significantly from the pipe.

A larger reservoir is the only change that I think might improve usability.

 

[Baluncore]

I don't see how to do this. When the pump pulls air for a moment then recirculating that air from its outlet would break the suction.

The pump output goes to below the surface of an open vented reservoir. The pump draws water from two sources,
1. from the input line and
2. from the bottom of the vented reservoir.
A choke valve in the line from the bottom of the reservoir can be adjusted to either circulate water freely or build up the required suction at a minimum RPM.

When a centrifugal pump runs with air input, it is unable to draw a good vacuum without a high RPM which may damage the shaft seals that need water lubrication and cooling.

When a centrifugal pump runs against a blocked output, it boils the contained water. The reservoir provides cooling water and an air escape route.

 

[ToolReview]

I like your solution of a feedback line from a higher elevation tank. I had visualized "Recirculate some water around the pump to keep the pump filled with water" as only a line from its output to its input.

 

[ToolReview]

When a centrifugal pump runs against a blocked output...

BTDT last week.

But 10 amp fuses averted further damage. (The problem was a kinked garden hose outlet from a tractor-carried watering tank with 12v pump).

Newbie here - are photos welcome that are peripheral to the original topic? I respect limiting replies to serious discussion only, if that's what the site prefers to stay focused on.

 

[rollingstein]

Newbie here - are photos welcome that are peripheral to the original topic? I respect limiting replies to serious discussion only, if that's what the site prefers to stay focused on.

I would love to see.

 

[ToolReview]

are photos welcome that are peripheral to the original topic?

I would love to see.

Practical application of an eductor: Hobby mining on our gold mining claim in the Northern Sierras 2008.

For 30 years we used larger gear I designed and cobbled together out of available materials. (Application of my interest in practical engineering). I finally decided I was too old to backpack a 5hp B&S engine (rototiller engine), my kids were grown and less interested in being my pack mules, and I could afford to buy the factory-built components shown here.

Stock photo, illustrating the 'suction nozzle' I am operating. The alternative, a 'power jet' installed at the intake to the sluice box, is shown in the Keene links I included in post #3 above. That alternative has greater lifting power, but is not suitable for shallow water since it loses prime if the intake is allowed to suck air for a moment.