Dead headed runing of centrifugal pumps-SOS

[ivoryhorseman]

Hi All,


setup:
I have two centrifugal pumps, pumping against each other. The pumps are placed in separate tanks which contains water. The two pumps are connected to each other using a flexible pipe. There are non return valves sitting outside both the pumps, which ensures that neither of the pumps doesnot pump water back into the other tank.There is a flow meter in the flexible pipe which is used to measure any cross flow which take place due to one of the non return valves opening up. This is whole purpose of the setup , i.e. to find out any cross flow taking place which is not suppose to happen by making artifical scenarios. I have a pressure reading from both the pumps as well pressure reading at the exits of NRV. The aim of the setup is to find at what condition this cross flow would occur between the pumps. I have managed to understand that when both the pumps are producing same head or pressure there is a possibility that the NRV could open up due poppet opening up intermittently for few seconds and then closing, this opening of the poppet is due a either vibration/ resonance since the spring which holds the popet is very weak(say arnd 5 psi of strength would open up the popett). The vibration could actually be produced by the pumps operating in dead headed condition since the liquid in the flexible pipe is static and ideally not expected to flow.

I have been plagued with problems eversince i started the testing , i.e.
1. when I switch on both the pumps the pressure of both the pumps never equalise, i have been told that the fluctuation between both the pumps is because of them being dead headed. is this true???
2. I have been told to attach a volume of cyclinder right after the NRV to dampen out the pressure variation( can this used ) or can anyone suggest a better way to stop the fluctuation produced between the pump(please bear in mind the the pumps are of the same category and does produce same pressure out when run individually).
3. Once I have sorted the above mentioned problem of equalising the pressure between the pump, I have to create a vibration enough to open the small spring inside NRV to replicate the liquid flow between the tanks, which is the main purpose of this experiment.
4. The flowmeter placed in the center of the flexitube gives very funny output , i.e at one second the output is 0litres then the next second it would show 100 litres and then 75 and some random number( the numbers are just examples). i have been told that the flowmeter adapter could have been picking up noise from the nearby electronics of the pump or is this due to water pulsating in the tube.the flowmeter is a magnetic pickup type.


can any of you kindly provide some insights into problems I have mentioned above

 

[LittleInch]

I have read this a few times now and have absolutely no idea why you are doing this and any results I suspect will be meaningless due to the arificial conditions you are running at. Apart from probably destroying two pumps, unless they are quite small units, you really need to explain a bit more why you are doing this. However, to answer your questions see below (sigh):

1) The fluctuation between both pumps at dead head is simply a matter of margins of manufacture. No two "identical" pumps are truely identical in flow charactristics unless you really mesuered every micron of them to ensure duplicity. Just like two "identical" cars off the same production line when run flat out, one will actually go a few MPH more than the other, then two pumps, especially when operated at dead head, where they are not supposed to be, will produce different results. They should be within about 10% to 15% of each other, but that's about as good as you get for what I guess are quite small pumps.

2) The best way to operate this is to allow some flow through the pumps by means of a bypass line and needle or globe valve or orifice plate. This will stop the variations and provide much smoother presure / lack of "swirl" which you are probably getting by dead heading these things and should allow you to exactly equalise prssure from your slightly different pumps. Otherwise yes, a pressure damper or water hammer arrestor would do a simialr, though less effective job. Aim for 20% of max flow through the bypass line. Operating any centrifugal, no matter how small, at no flow for lnong periods will lead to its premature destruction.

3) Why??. You said at the start that the NRVs are there to stop flow back through the tank. Spring loaded NRVs are there to stop the "chattering" you sometimes get at low differential back presusres on some NRVs. This is the thing I can't get my head round.

4) In this sort of scenario, you will get "swirl" or random flow with some flow going one way and some the other, confusing the hell out of the flow meter which is only designed to handle proper flow. YOu probably need some sort of PD type meter to stop tis. Any sort of mag flow, UT or turbine meter is going to give you garbage results.



My motto: Learn something new every day

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

 

[Artisi]

"the pressure of both the pumps never equalise" -- of course - why would you expect the 2 pumps to be equal. The superior pump will shut the NRV and the second pump will only achieve its CV pressure maybe something less than the other unit. As discussed by LittleInch - no 2 pumps will be identical in terms of head or flow - in this arrangement only head is the consideration.


If you have two pumps interconnected in this manner the pressure in the cross connecting pipe will reach a pressure equal to max. pressure at CV of the superior pump, once the cross connection is pressurised it cannot change, unless of course 1 or both NRV's are not sealing correctly.

Now for discussion, let us assume that 1 or both NRV and not sealing ---- as a pump nears CV there is a lot of discharge re-circulation and possible cavitation -- who knows what is really happening hydraulically within the pump at any one instant in time - pressure could reduce slightly and if a NVR is leaking a hunting effect between the two pumps is possible as the pressure in the cross connection may momentarily reduce due to the leaky NVR.

Something to think about??





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.)

 

[MikeHalloran]

If someone is paying you to conduct this experiment, I have some investment opportunities that will be of interest to them.


Mike Halloran
Pembroke Pines, FL, USA

 

[Artisi]

Mike, had similar thoughts about this "project" and deleted a cryptic comment at the very last minute giving the benefit of the doubt -maybe we are missing something.

Guess your investment opportunities are along the lines of double you money in a short time etc. no-risk involved, please advise your bank details and forward your investment to the Nigerian Farmers Bank etc.

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.)

 

[ivoryhorseman]

Little inch, Artisi---sorted the problem.. thank you guys for the advice

MikeHalloran---I have passed on ur message.

 

[Artisi]

So what was the problem?

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.)

 

[DubMac]

Is your project funded by Stimulus money???