Efficiency tests for pumps in operation 2

[tundelash]

Gents,
Does anyone have ideas on how to carry out pump efficiency tests for pumps in parallel and in operation, without having to stop any one of the other pumps that are not on test (centrigugal pumps and recips as well).

What are the consequences of "no-flow" tests while in operation. Can it be carried out while all pumps are in operation without affecting any of the pumps, and can it replace efficiency tests in a crude pumping station ?

I have also looked for International standards for procedures on efficiency tests. Does anyone know of any ?

 

[vanstoja]

I don't understand why anyone would want to conduct efficiency tests on operating centrifugal pumps, particularly when running in parallel with other centrifugal pumps. If you're trying to determine whether or not pump performance is degrading with time you should be measuring pump head at some specific flowrate and comparing the result to a baseline test of the performance of the pump when first run in your in your facility. If that was a single pump test and your condition monitoring test has to be with two or more pumps running in parallel you'll have to figure out what the decrement in head should be between single and multi-pump system resistance curve intersections with the head-flow curve from the original baseline facility test or else by using the pump manufacturers test loop curves which may not apply to your facility if the inlet piping configurations are significantly different. The test standards for efficiency like the Hydraulic Institute Standard and the ASME Power Test Code for Centrifugal Pumps apply to manufacturers' tests and do no cover in-service tests to my knowledge. The manufacturer and the Code people expect the pump user to provide the correct inlet piping to provide the "guaranteed" pump performance. I also don't understand why you are operating reciprocating pump in parallel with centrifugal pumps. If you hope to get more information then you'll have to provide input to make your question intelligible.

 

[bongo]

Good day all,
Well one could carry out a "Wire to Water" test but you should realise that running pumps in parrellel would not give you a proper reading for each pump as although the TDH would remain relaitivly the same, the flow does not double as most people seem to think. Another problem would be if the pumps in parrellel do not have the same duty.
I wish you luck in your struggle and hopfully some bright spark will guide us all to the path of enlightenment !

 

[bongo]

Good day all,
Well one could carry out a "Wire to Water" test but you should realise that running pumps in parrellel would not give you a proper reading for each pump as although the TDH would remain relaitivly the same, the flow does not double as most people seem to think. Another problem would be if the pumps in parrellel do not have the same duty.
I wish you luck in your struggle and hopfully some bright spark will guide us all to the path of enlightenment !

 

[electricpete]

The no-flow test I assume checks the pump performance by determining head at shutoff conditions.

I think this no-flow test would only be valuable if you can manipulate your system to isolate the discharge of the pump under test from other pumps (and still able to measure discharge pressure). The test would be limited if you tried to conduct it on multiple pumps all feeding in parallel to a shutoff header, since the performance of a weak pump would be masked by the performance of a strong pump.

If you have the ability to measure the individual flow (and dp and fluid temperature) on each pump, then you can locate the performance of that pump on its own curve regardless of what other pumps are in operation. Otherwise (if no individual flow info) you'll have to make some assumptions regarding the sharing of load which will decrease the ability to accurately evaluate each pump.

Note that flow measurements can often be made with temporary ultrasonic info outside of the pipe.

As it stands your question is pretty wide open. Can you clarify what you are trying to accomplish?

 

[electricpete]

next to last paragraph should have read "... with temporary ultrasonic test equipment outside of the pipe."

 

[sprintcar]

Running centrifugal pumps at shut-head (no flow) will build up significant heat very quickly within the pump cavity and turn the fluid to steam, usually followed by a catastrophic failure (ie explosion).
If you have the baseline information on your pumps from when they were installed, tracking flow vs current draw on your motors should give you an idea of system degradation as parts wear out. Keep the wheels on the ground
Bob
showshine@aol.com

 

[ShaunE]

I would do the following for check of efficiency:

1)with pumps running measure flow and outlet pressure where pump outlets combine. This will probably be the difficult bit if you dont have instrumentation!

2)Using SI units multipy pressure by flow i.e

m3/s * Pa = Watts (work being done)

NB 1Pa= 1N/m2, also 1Bar = 1 x 10^5 Pa

Read off the power consumption printed on pump motors.

Calculated value for work should be about 50% of the motor power consumption allowing for inefficiency of the motors, the pumps and any losses in the pipes / manifolds. If your in this ballpark all is well!

 

[panelman]

there is a method that involves inserting two temperature probes in the pipework(one before and one after)

It was a long time ago but i seem to remember it worked on multi pump sets but unfortunatly i can remember no more

 

[TD2K]

With Mikemk's method, you measure the temperature rise across the pump and back calculate the efficiency using:

tr = Head * (1 - E) / (778 * E)

E is the efficiency as a decimal
Head is in feet
tr is the temperature rise in deg F

 

[panelman]

thanks for the formula, the bit of kit i used had it all built in...insert the probes, read the results of the display

but i still can't remember the name

 

[electricpete]

If anyone has the time, I would be interested to know the rough derivation of that formula. Not with numerical values or units... just proportionalities.

 

[TD2K]

I suspect Pete it starts from the standard equation for pump power requirement Hp = H*Q/constant*e.

The remaining work that doesn't go into producing head goes into heating the liquid up. Throw in a heat capacity, flowrate (which would cancel out), you should wind up with the 778 factor though I haven't worked through it. Okay, who's bored on a Sunday night ;-)

 

[electricpete]

Thanks TD2K. I'm a little slow so I have to write it out in more detail so I can remember it:

Pout = Q * Head * rho (where rho = density)
SHP = Pin = Pout/E = Q * Head * rho / E
Losses = SHP * 1-E = Q * Head * rho * (1-E) / E
Also losses = Q Cv dt
equating two expressions for losses:
Q Cv dt = Q * Head * rho * (1-E) / E
dt = (rho/Cv) * Head (1-E) / E

 

[abeltio]

1 Btu = 778 lb-ft

The definition of Head is also energy by unit mass...so
1 ft = 1 lb-ft/lb = (1/778)Btu/lb

HTH
saludos.
a.

 

[abeltio]

Just a comment... to measure the flow of one pump before the tie-in with the others you could use one of those on-line ultrasonic flow measuring devices (you need about 6 feet of straight pipe to install the thing), with a differential press gage (not 2 separate gages!!!) before, after the pump... and the electrical data you could determine a rough efficiency estimate. This efficiency should match the one obtained from the pump's manufacturer.
HTH
Saludos.
a.

 

[panelman]

Having struggled to remember the details of the meter it's finally surfaced in my old brain.

Yates meter...

 

[panelman]

See

http://www.yatesmeter.co.uk/Products/PortableYM.htm

for more details

 

We have designed a pump efficiency and flow monitor which uses the thermodynamic method,and enables individual pumps to be tested. ISO standards apply. For more information, see

http://www.ltc.com.au