Suction Problem 1

[ziptron]

Hi Everyone,

I'm using a Dri-Prime Pump (close-coupled centrifugal pump with vacuum priming compressor) and I am using it to suck (and then pump) water out of a water storage pond at approximately 1000 GPM through a 12 inch PVC pipe. The pond is deep and has plenty of water, the pipe inlet within the pond is located approximately 250 ft away from where the pump is and I only have approximately 2 ft difference in elevation between the top of the pond and the connection point. So basically, I am trying to suck water up two feet and across about 250 ft. The pump is rated to lift 28' so even if its not 100%, I think I am in the clear.

It is an older system and it has worked in the past, but recently I am finding that even though my technicians can get it to prime and pump some water out, they suddenly lose prime and nothing seems to come out of the pump. They indicate that this happens at random. I've tried a different pump so the physical pump/priming system of the pump does not seem to be the issue. I cannot inspect the pipe as it is underground, however I cannot imagine anything happening within the underground pipe itself which will allow me to start pumping, but then suddenly lose prime. If I had a plug, I shouldn't be able to prime at all right? If I had a partial clog, I should be able to pump but at a lower GPM.. Yes, I get water out.. but then lose it.. Any ideas what is happening?

Any idea what I may be missing?

 

[DubMac]

250 feet sure seems like a long suction line. What is it's diameter in relation to the pump suction, what material, and what condition is it? Done your friction calcs?
The 28' rating is the very max a pump will do, in it's new condition. Under testing, that figure generally moves closer to 24-25 feet.

 

[Artisi]

Air leak, in pipework or pump, careful inspection if pump needed - look at gland / stuffing box to start with.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[bimr]

In addition to what the others have posted, one would expect a problem with a vortex at the pipe inlet. The calculation shows a minimum of 25.8-Inches of vortex. With approach-flow and wave action, there may not be enough submergence. That may allow air to air bound the pump on occasion.

Link

 

[ziptron]

Hi All!

Thanks for the replies. 250 ft is long, but I have a 12 inch PVC pipe and I'm pumping around 1000 GPM, so friction is quite low. Its a system that worked seamlessly before (apparently) but suddenly it stopped working. I did the calculations based on drawings provided and had them confirmed by a colleague. It seems to be less than 14 ft of loss when I account for length, elbows, fittings ect.., leaving me with a fair amount I would say even if my pump is not perfect.

I have the same issue with various pumps, so I'm thinking its more the piping than the pump.

Bimr, your idea of the vortex is definitely interesting and I must admit, I would not have thought of it myself. The depth of my pond is actually nearing the edge of this calculation and this is definitely something to consider! I'll have to give this some more thought and perhaps some testing, I'll post my findings within a few weeks hopefully!

Any/all other ideas are still welcome.

 

[ziptron]

Never mind, I thought I only had just under 3 ft of depth, but in reality I have 3 meters so I am almost at 9 ft of depth. Good ole imperial vs metric debate. Anyway, although definitely a good idea, I seem to be fine on the vortex aspect as well.

Looking for more ideas

 

[Artisi]

Check you don't have a dead or even alive crocodile stuck in the inlet pipe.


It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[LittleInch]

How long a time period / volume do you get tout of the pump after priming before it looses prime (i.e. you are presumably sucking vapour or a mixture of air and water?)

Is sounds suspiciously to me like you've got an accumulation of silt and dirt in a low spot in your line or possibly just the entire bottom half. At your flow rate you're doing less than 1m'sec, so you will get a lot of drop out. Hence your friction calcualtion is wrong as your net ID is now a lot less than 12".

The system will work for quite a while as you gradually eat into your margin on NPSH, but eventually it will get to critical point. I would like to bet your pump has actually been cavitating for some time and now your impellor is worse than it was and the NPSH is now lower and boom - no flow. You can prime as the the flow rate is much less. The randomness could either be related to water temperature, water level in the pond or simply a greater accumulation of dirt in a low spot causing a partial blockage which then goes away the next time.

What I would do is
- strip the pump and check the impellor for damage.
- Connect some other supply and at a calculated flow of 2-3m/sec flush at least 3 - 4 line fills of water back into the pond (if you're allowed to) and see the (massive?) amount of dirt that comes out.
- regularly back flush the line (say every month?

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

 

[TenPenny]

If it used to work, there are two things I can suggest.

Check the inlet to the pipe, there might be growth or debris causing problems
More likely, in my opinion, is a leak in the pipe that is allowing air in, once enough air builds up, you lose suction. How to find the leak, good question, if you can block the inlet, try pressurizing the suction pipe somehow.

 

[Artisi]

From memory the Dri-prime unit is an air ejector system used to prime the pump unit. Within the ejector unit there should be a non - return valve consisting of a ball - dismantle the ejector unit and check the condition of the ball and seat - if this is faulty you will have a problem maintaining prime.
As the unit was working previously and doesn't work now it is probably a minor fault, the secret is knowing what it is. [COLOR=][/color]

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[bimr]

How deep is the suction point?

Why is the suction so long?

If the suction line has a high spot in it, air will be trapped. You do not have enough velocity to flush out an air pocket. The pump will eventually starve.

Put some pressure gauges on the pump to determine what is happening.

 

[zdas04]

It does sound like something mechanical like Artisi suggested. It could easily be a ball check that seals at max vacuum, but doesn't quite seal at lower dP. What you are describing sounds very much like air ingestion instead of accumulation of solids (or crocodiles) in the suction.

David Simpson, PE
MuleShoe Engineering

In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual. Galileo Galilei, Italian Physicist

 

[TenPenny]

If he switched out the pump assembly (assuming this includes the priming system, mostly they do) and the same thing happened, I'd say that rules out anything mechanical with the pump/priming assembly.

 

[LittleInch]

I thought it might be air ingress, but if the pump can prime, usually at a much lower flow in than in operation then it implies the pipe is sound. The sudden ingestion of a lot of air makes some sense, but if this is replaced then maybe not. Partial blockage would allow priming, but once you hit somewhere around NPSHR you will struggle snd then stop once the prime is broken. A high point air bubble makes sense also, but how did it work for a long time before?

you need some gauges to see what is happening.

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

 

[ziptron]

Hi everyone, sorry for my lack of replies. I was on site all day away from a computer. Thank you for your replies however!

Yes, as TenPenny said, I think I can rule out the pump being an issue because I've switched pumps all together (have access to large portable pumps) and I still had problems.

I do like the theory of silt/accumulation within the pipe and I have considered as it does make some sense. At low flow, I can maintain NPSH.. but, once I increase flow, the friction kicks in and suddenly it requires too much force to draw the water making it impossible to do so. This makes sense intuitively, however, I would think that I shouldn't lose prime completely, but instead just suddenly decrease my water output to a rate that is dictated by the silt/accumulation. Instead of drawing 1000 GPM for example, due to the blockage, I would draw 500 GPM or even 100 GPM.. yet I'm being told that the pumps are completely losing prime. Can someone help me understand this aspect via calculations? Should I lose prime all together, or should I just decrease down to a lower flow rate?

As to the question of why this pipe is so long.. I have no idea haha.

 

[Artisi]

No, the pipeline is not too long provided the friction losses don't exceed NPSHr.
As it worked before there is no reason why it can't work now - so what has changed.

You need a vacuum gauge on the inlet system to monitor what is happening to give some hard facts:
when fully primed what is the vacuum gauge reading.
does it stay constant or does it surge
when does flow start to diminish (time) and what is vacuum gauge reading
what is the vacuum gauge reading once flow stops completely.

If you lose prime completely where will be no vacuum reading, whereas if you have a high vac reading then good chance you have a blockage.

Can you send a ferret through the line with a CCTV camera attached to have a look at what is going on.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[LittleInch]

The pump is working on the basis of flooded suction. Flow is fixed by what you do on the outlet side to increase the head and decrease the flow. For a pump like this the inlet pressure can only vary from 0.9 bara to about 0.3 bara, not enough to vary the discharge head much. A pump which has a positive pressure might show a reduction in flow, but not this one.

What might be happening is that as the pump approaches the NPSH limit, you would or should start to get cavitation, but if you have what seems to be a very low NPSHR pump, you might just go the wrong side of the line and suddenly have large vapour pockets develop. When this happens, especially at a relatively low velocity, the incoming fluid stops and actually reverses down the line. your pump is left trying to pump vapour which doesn't work so well.

As said it's only one possibility and a pressure guage reading in bara should show you what is happening.

Also maybe try restricting the flow to say 500 GPM by throttling the discharge and see if the same thing happens. Then there will be much less friction drop. if it still happens at that flow or even 250 then something else (air entrainment) is a more likely probable cause.

Let us know how it goes - this is an interesting one.

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

 

[bimr]

The pipe is too long if there is silt in the pond. When the flow stops, the silt will build up. You do not have enough velocity to flush the pipe out.

The pipeline is too long is there is a high spot that will trap air. You don't have enough velocity to flush out an air pocket.

In either of these scenarios, the pipe will be restricted and the pump will starve.

As mentioned earlier, buy a few cheap pressure gauges. Monitor what is occurring. If you have a strap on flow meter, it would also be helpful.

 

[ziptron]

Sooo.. I have the answer.

I went to site myself and brought with me a pump. I could not get the line to prime for 10-20 minutes even though I had a good vacuum so I assumed there was no chance. I stopped and basically gave up, though for some reason decided to give it one last try and fired it up again. Pure chance, suddenly a trickle of water started flowing, then more, then more and then more.. Suddenly after about a minute of the trickle, the water turned muddy (really dark almost black) for almost a minute of 1000 GPM flow but after that it became clean again and flow continued. I suppose the problem was a clog all along which I successfully persuaded to enter the pump. The system works as intended now. I still don't know why they would have lost prime while it was working.. but perhaps they got a trickle and once they increased RPM and tried to draw more, they would cavitate the pump and this is what they called a 'loss of prime'.

Anyway, I think my job on this one is done! I thought I'd let everyone know the result. Thanks for your generous input!

 

[Artisi]

A vacuum gauge fitted to the pump would have sorted the problem from day one.
But good result and a lesson learned.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)