Gravity Fed Water Supply to Pump

In my opinion, for the scenario that you present, a short length of 50mm pipe will NOT drastically reduce flow.

Thanks for the quick responses.

I was thinking along those lines, but if I push this to extremes: imagine a column of water (a tank, really) 1m diameter 10m high. If you could suddenly remove the base, the water would come flooding out: basically zero friction loss. But if you instead opened a 50mm outlet at the bottom, with the length of the outlet 50mm, the tank would empty relatively slowly ie the effect of the short length of reduced diameter pipe would greatly reduce flow.

Going back to the original question about loss of flow by reducing from a 90mm pipe to a 50mm pipe, would the same rule not apply, suggesting that the loss of flow could be significant even with a relatively short length of 50mm pipe?

Perhaps there is only way to find out: go and set it up and measure it!

Yes, the loss wil be significant, but not drastic. The shorter the small pipe section is, the less significant the loss, also any elbows will cause flow losses as well.

Reduced diameter equates to increased velocity. Increased velocity equates to more friction.

Keep it short.

GA. The original poster also needs to appreciate that, just because these sorts of flows are coming out of an open-ended pipe, the figure will reduce when a hydraulic motor-pump is connected onto the end of it.

On the gravity powered pump you need to keep your losses to a minimum for the best performance. The more friction yo have the less head will be available to drive the pump.

Do your friction calculations at 20 l/s using the big pipe.

add the losses in the reducer and fittings (equate it to a length of pipe - found on the tables)

and add the losses through the nipple and the pump neck.

Subtract the losses from the static head .

and see what you have left to drive the pump.

Are you using a centrifugal pump in reverse?

Note that the outlet will be at zero head.

Pump suction lines are often one or two pipe sizes larger than the actual pump inlet connection. In SI sizing, the next two sizes above 50mm are 65mm (~2.5") and 80mm (~3"). I don't have my pipe sizing data here at home, but I would guess that 80mm would be your best choice, with a reducer and a 6D (300mm) straight length of of 50mm pipe entering the pump.

I would install a tank and run the pipe from there the tank will also collect trash that would cause problems for the pump if you tap the tank for the pipe 6 to 8 inches off the bottom.

What are you pumping ? Water ?

Think about how this sort of pump works.

A mass of water is flowing down a rigid pipe, a mechanism stops the flow, the energy stored by the moving mass is converted to a spike in water pressure, a one way valve allows the higher pressure to push water up hill.

So, since E = 1/2MV² you want maximum speed in the pipe and maximum mass of water moving, the largest rigid pipe possible is my recommendation, you can limit the total volume consumed by limiting how often the cycle runs.

Hi Ed, what you say seems reasonable so GA.

I think it's what we were brought up on that determines "easy", I always find the metric system clearer, there's no magic constants in the formulas like "/3960".

I don't think he should use thin walled pipe, true the static head is low but the system works by suddenly stopping the water flow causing a rapid, transient pressure surge.

ffej --- Good point if that's that is the type of "pump" that petecat wants to build. A tapered transition rather than a sudden change in pipe size is far harder to build and the efficiency gain due to recovering the kinetic energy (velocity head) is relatively small in this case. Also note the potential elasticity of a pipe that is only thick enough to contain the pressure may dampen the "water hammer" shock that makes this type of energy conversion work.

I'm swagging these comments having never analyzed or even closely studied hydraulic ram pumps. I Googled "hydraulic ram pump" and came up with a good number of links to study. Time is short for me this morning so I'll leave it up to the rest of you to get to know more about the subject before I do.

I do suspect that such a pumping method will produce usable water quantities regardless of the size of the pipe. This will likely be limited by how large a pipe diameter petecat can afford to buy. The one thing you can't cheat on is energy. He's got 373 watts (plus a bit more for velocity head) to work with in the stream flow. That's about 1.5 liters per second to 80 feet height. 1% of that will put over 300 gallons per day in a tank up there. That's enough for a family of 4 and from my own experience 80 feet of head will work fine with all modern household appliances. And it will leave plenty of outflow for irrigation if the garden is below the low end of the piping. Or he can just return it to the stream if there is a water rights ownership issue.

I think the important issue here is to have a more complete understanding of the configuration and necessary pipe sizes to do the job at hand. Again the question to petecat: How much water do you want to lift per hour, day or whatever?

Ed Weldon

GA for Ed, as his methodology and number crunching is sound. I concur with his presentation and conclusions.

If you're looking for an engineering formula to check your system capabilities, then utilize the Bernoulli Equation.

There is a Bernoulli Theorem Calculator that you could use located at:

http://www.ajdesigner.com/phpbernoulli/bernoulli_theorem_equation_velocity_v1.php

I would suggest that you use PVC pipe and fitting for your system, as the friction coefficient is very low compared to other materials, therefore you'll minimize head losses (velocity head) due to flow rate along both the suction and discharge lines. Also, try to minimize bends as much as possible throughout, especially fittings like Tees, 45 degree and 90 degree elbows. Use full port ball valves instead of regular ball valves and gate valves a well as using concentric reducers (if available) to minimize friction losses.

Since your "pump" is essentially acting like a water turbine, then I suggest you do research online regarding turbine mechanisms and their low efficiencies. Depending on what type you have chosen and employed, you may be lucky to get a turbine efficiency between 20 to 25% at best. You may also want to obtain the manufacturer's published pump curves for the specific pump you have chosen. And don't rule out cavitation problems, but then again, it all depends on the type of pump, it's size and it's pumping characteristics.

Assuming that you have a centrifugal pump, you may be better served by purchasing a true turbine pump specifically tailored for your needs....

Good luck with your project!

Hi petecat,

I have a ram based system supplying water to a property in the 'Glens'; it employs a ''vulcan'' ram which is fed through a 3'' main which falls roughly 10' from a small forebay in a fast flowing river. The ram consumes about 40 gpm and delivers 1.5 gpm to a storage tank 300' away and 60' above the ram.

I reckon that its worked continuously for 80 years and other than some repairs to the clack it has proved reliable, otherwise it wouldn't be there now!

In answer to your question the ram and the supply pipe should be identical in bore for the best results; as has already been stated the principle is based upon the mass flowing and the square of the velocity, therefore increasing the bore of the pipe will not increase the kinetic energy of the system.; you need to increase the throat diameter of the pumping device to do that.

As the conversion of kinetic energy into a pulse of discharged water depends on the robustness of the the ram and supply main you will appreciate the need for heavy grade pipe with everything well and truly anchored to some substantial rock structure or concrete foundation block, otherwise it will quite literally shake itself to pieces. In my case the ram is attached to an extremely large boulder set into the river bank.

So you will be unlikely to better the pumping rate by increasing the bore of the supply pipe, more 'drop' would help if you can get it but better still a larger bore pump would solve the problem.

Good luck,

Massey.

Thanks for the great answer, Massey. Nothing beats the "boots on the ground". Especially if they've been there for a fair lifetime. 1.5 gpm is 2160 gallons per day, enough to supply 7 American families without heavy garden irrigation. Given that the 40gpm consumption can usually be returned to the watershed that aspect is well covered. Your numbers indicate 22.5% efficiency for the combination of ram and pipe friction losses. This is excellent for such a small systems and probably beats the efficiency of the best water turbine and pump combo you can buy off the shelf in today's marketplace. And at a small fraction of the cost of the pump/turbine combo.

Building the system relies on availability of some quality steel pipe and the ability to fabricate parts primarily by welding. In North America these skills are fairly common among people who do farm and construction equipment maintenance, machinists, millwrights as well as hobbiests involved in construction of vehicle components.

My first post on sizing the pipe was appropriate for a conventional pump/turbine system to convert the energy. But it certainly seems that your ram is a more practical approach and the design is fundamentally diffferent from turbine or positive displacement devices. There is a problem with creating simple low cost water pumps that pump at very low specific speeds; i.e. conditions where the flow rate is very low compared the head they have to pump against.

Ed Weldon

Hi Ed Weldon,

Many thanks for your thoughtful comments. I have noted a number of these installations tucked away in river courses in the Highlands, their familiar 'heart beat' is usually their give-away and that usually sends me scrambling down precipitous banks in search of the beastie.

My own machine beats 35 times per minute and although I've messed about with the spring tensioner on the clack to see what happens to the delivered output I've found that increasing the stroke interval makes it prone to stalling ( clack remains open and water keeps running) or reducing the interval can make it prone to a form of 'fibrilation' where-in you get a virtually continuous stuttering with insufficient open time for the water column to acquire its optimum velocity before it picks up the clack and slams it shut.

Incidentally spun iron pipes are frequently employed in the supply line with bolted flanges.

Best wishes,

Massey.

" It is technically somewhat different to a ram pump." There are several different ways to produce power / deliver water / do something else. The efficiencies of these different methods are different, also. If you would take the time and effort to describe what the actual situation is, and what choices have already been made, there is a good chance this informed group of engineers could give you exactly what you want. But the greater their knowledge, the longer and more conditional their answer would be,and the harder it would be for you to use that knowledge. ?

I shall do my best to provide more detail, but please take into account my general lack of experience and knowhow with this stuff. The pump I am looking at installing is, from my understanding, a new style of water powered pump (for pumping water). It is a derivation from the Glockman pump. THe Glockman uses the water hammer in the drive pipe to drive a diaphragm which then pumps from a seperate inlet. (A classic ram pump uses the compression of the water hammer to push part of the water from the drive pipe through a one way valve-my interpretation) The glockman is good in low head situations.

This new pump works along a similar principle, but I'm afraid I don't know any more than that except that it is even more efficient. The supply from the weir drives the pump. The inlet for the water that is to be pumped is a separate port from the drive pipe. The main difference from the ram, and the glockman, is that in this case the inlet pipe is not part of the actual pumping mechanism, but simply delivers the means to drive the pump. That is why the connection at the pump is 50mm, rather than 90 or 100mm.

The table I have from the maker of the pump states that with a 2m drop, and 10L/sec flow, pumping up to an 80m head, the pump will deliver over 15000L/day. At 6L/sec, you get 9000L/day. That equates to a 70% efficiency rate. (FYI the formula is: 36*24*efficiency*supply flow rate*drop/delivery head). Don't ask me why they multiply by 36 and by 24. Something to do with changing L/sec in the intake, to L/day in the delivery?

So, I am working on the assumption that the pump can do what the maker says it can do. (I can hear the accusations of naievety, but the dealer is a guy I trust) With that in mind, my issue is how to deliver 10L/sec through a 50mm connection with a 2m drop, even if I run 90mm pipe most of the way (10-12m)to the pump and reduce to 50mm just at the end. If I can't do that, then there's no point buying the pump....Or building the weir, which my back would be very happy about actually.

I have found a pump that looks pretty similar to the idea I've been looking at, but a more polished looking product (not that that means everything)

Have a look at Dingo Pumps at planetsafepumps.com

Having trouble contacting the maker at present is the only problem.

I would add that this one does not make the claims of available delivery head that the other one does.

THere are youtube videos of this one at work.

If you could get 100% efficiency, 2 meters x 10 L/s = 80 meters x 0.25 L/s = 21,600 L/day. Your efficiency is likely to be considerably less, but may still be useful.

Do you have catalog data or a Web source for the pump in question?