Understanding Centrifugal Pumps, head and flow rates

1) The gauge should be on the pump side of the discharge valve.

2) A gauge tap usually present.

3) You are only reading the discharge pressure - you need to measure or compute the suction head and add it to obtain the comparable total pump working head.

4) remember to add the difference in height of the gauges.

5) A control valve on the suction side is not advisable. It may cause caviitation.. An isolating valve is in order.

Hendrik cited the most important issues. Some other possible causes for the variation you see are as follows:

1. Pressure gauges should be 5-10 pipe diameters away from any fitting to insure straight flow past the gauge tap.

2. Velocity head corrections will have to be included in the calculation of head if the suction and discharge gauges are located in pipes of different diameters.

3. Make sure the actual diameter of the pump impeller corresponds to the diameter for the pump rating curve. However, it is unlikely that impeller diameter would be purposely cut down on such a small pump.

4. Heavy internal wear of casing or impeller can reduce performance to the extent you observe. But again this is unlikely in small clean water pumps such as yours.

5. Be certain of your flowmeter and pressure gauge calibration.

6. Make sure that your electric motor performance is adequate. Pump head changes with the square of the rotational speed.

7. Note that centrifugal pump rating curves show head, not pressure unless the specific liquid is stated on the rating curve. If your liquid is not specific gravity = 1.00 then you must make a correction.

8. Make sure there is no entrained air in the liquid or cavitation taking place. If you are operating with a substantial suction lift then you will need to determine if the NPSH available is sufficient for the pump at the flow rate you are running.

Ed Weldon

Hendrik and Ed have given good advice.

I would try to plot an additional point on your manufacturer's curve and that is the power absorbed by the motor. Having Power, differential Head and Flow all plotted on the original curve can be very useful, as it gives some indication of which measurements could be wrong or what is wrong with your pump. For the motor power, take ampere measurement with a Tong Tester in the sub-station and calculate power, do not believe the local ammeter.

Dear Rick

1-The manufacturer NEVER gives the pressure in the specs.of the pump.but only the head.you have to use the equition:

H(ft)= P(psi)x2.31/sp.gr

H(m)=P(bar) /10xsp.gr

Most of curves are given for clean , fresh water at 68deg.F

2-Visit this site ,please:

www.lightmypump.com/pump_book.htm

you will have afree ebook(pump system analysis and centrifugal pump sizing),

the chapter3(pages 17 to 28) explains the effect of Reynolds number.

greetings

Aly

Dear Aly,

Thank you very much for that interesting site.

I would not attempt this with a larger, higher pressure pump but consider this quick test:

Close in the discharge valve of the pump momentarily and see what the pressure on the gage is. (Open the valve as soon as you are done). Add the suction head to this pressure. Compare this pressure to the dead head pressure of the pump curve. They should be identical if the pump impeller is the right size, the pump is not worn down, etc. I find if the deadhead pressure matches the pump curve, usually, all else is very close also.

The reason for this test is that it gives you a cleaner test without all the errors from:

1) Flow rate measurement (you know the answer is zero),

2) Errors in calculating the head loss from friction in the suction (again..zero flow means zero frictional suction head loss)

3) It PROBABLY negates the concern of electrical motor performance but you can check this with an rpm gage if you want.

Note that often, I see the RPMs listed as 1800, 3600, etc for the pump curve but the actual electric motors cannot physically operate at nominal speed (slippage is necessary--unless you have a VFD that is capable of allowing nomnal speeds to be attained). As such, you may have to correct the pump curve for the lower, actual motor RPM.

I welcome comments on this test--I have been doing it for years and find it quite useful.

performance curves or characteristic curves of pumps as given by any reputed manufacturer are usually correct with a small margin of error. i agree with several reasons offered by fellow bloggers for flow reduction

an unlikely possibility could be clogging of impeller eye on the suction side which adds resistance and reduces flow. this should be reflected in the vacuum guage reading.

once the reduced discharge from the pump in my farm was ultimately traced to a 1 foot long snake(dead of course) neatly coiled at the impeller inlet.

shankar, india

Dear Mr. PuttyTor,

You have not mentioned about the speed of the pump, during your testing of the pump, which has a direct link to discharge (proportional to the speed) and head ( proportional to sqyare of the speed of pump)

Rajeswari.