Pump Replacement 1

[TomFitz]

I am not too knowledgable concerning pumps, and I have a question that perhaps someone could help me with...

I had a pump that went bad. :-(
It had put out:
21 GPM
56' Head
3" diameter impeller
Motor speed 3450

The contractor wants to replace it with a different pump with a larger impeller (6")
56' Head
a decreased motor speed to 1725
(and same size input/output)

He claims that it will still put out 21 GPM because of the decreased motor speed..

Would this be correct?

Thank you for any assistance... ;-)

TomFitz

 

[MortenA]

I theory this could be true yes - he halves the speed but doubles the diameter, these the cansells out. However different pump designs have different pump curves so there no gurantee (unless he has more information than us).

Best regards

Morten

 

[25362]

Let's start by saying I'm a pump user not a pump expert. Apart from that, you didn't explain what went bad with the original pump, although your post hints at problems related with rotating speed.
Anyway, let's try to concentrate on the query itself.

The contractor probably refers to a "different" pump having half the specific speed (aka "type number" in SI units) defined in the US as (RPM)(GPM)^0.5/(Head,ft)^0.75.

By looking at the largely quoted Worthington graph, showing a family of varying flow rate curves on an orthogonal plot of specific speeds vs efficiency, we can see that by halving the specific speed, with a different impeller geometry as for the case in hand, one can indeed get a pump doing the same service (GPM, ft head) at half the rotating speed but with an appreciable loss of efficiency.

For example, the graph shows, that for 100 GPM, pumps having 1000 and 500 specific head values, show about 60% and 47% efficiencies, respectively.

Experts are kindly invited to comment and correct my interpretation and add more information for my own education.

 

[PUMPDESIGNER]

25362 is as usual exactly correct.
I add the following:
If you trust the man making the suggestion, and he is willing to guarantee performance, perhaps not being an expert you can trust him, perhaps not.

Centrifugal Pumps do not "put out" or deliver an amount of fluid, but rather, they operate on their performance curve and the amount "delivered" is where the system resistance curve crosses the pump performance curve. In other words, system resistance determines flow rate.
Therefore, if the replacement pump curve is identical to the curve of the pump being replaced, performance will be identical. If however the curve is different, the performance will also be different.

Additionally,
There are vast differences between pumps with similar performance, including some extremely important differences, some of which are the following:
NPSHr - Cavitation problems
Specific Speed (Ns) - Size of operating window and performance out of that window, some pumps vibrate and cavitate severely out of the window, some do not, and the window is much larger for low specific speed pumps.

Suction Specific Speed (Nss) - Again, size of operating window and performance outside that window. But also, High Nss pumps require vastly larger margins of NPSHa over NPSHr to avoid cavitation damage.

Power Curve - Two pumps with identical performance curves can have opposite power curves, one draws more power as flow increases, while the other draws less power as flow increases. To be honest though, to be that different one would have to be a multistage while the other a single stage.



PUMPDESIGNER

 

[saxon]

Tomfitz, You can utilize the Pump affinity laws to check the comparison. Better idea, demand the Manu'f. pump curve for the pump in question. No curve, no problem, the pump doesn't come into the plant.

Hope this helps.
saxon

 

[PUMPDESIGNER]

saxon is very right about "no curve, no problem, the pump doesn't come into the plant".

That curve by the way should be the Full Pump Characteristic including: Power, NPSHr, Efficiency, and Operating Range or Window. If possible, get the efficiency info plotted as ISO type instead of line type.

PUMPDESIGNER