Does reduced speed testing have an effect? 1

[testeng001]

Recently a vertical pump was tested on water at 53% speed and everything was calculated up to full speed with a S.G of 0.440. The efficiency was ridiculously high. My question is does reduced speed testing have an effect, e.g. how do you know that when the units at full speed, theres no power losses on seals, bearings, recirculation, or losses between bowls?

 

[dcasto]

Nope, the fans laws all work.

 

[dickon17]

You can test at 53% of duty conditions and calculate flow and head quite accurately. However, I would not trust the efficiency to be exactly the same. How much margin of error do you have?

 

[d23]

testeng001

Depending on the pump design you could have some re-circulation that affects the test at low speeds. The pump affinity laws offer guidance, but are not absolute. Without knowing any of the specifics I would guess the affinity laws could be anywhere from 3 to 15 percent off at 50% speed.

I assume the test was performed at low speed due to the pump load and available power in the test facility. This would indicate you have a high flow pump, possibly high head too. Reducing flow by 50% could allow reverse flow inside the vanes. That depends on the impeller design.

Several years ago I saw a paper posted on this forum named “Stop Falling Off The Curve.” It detailed test that were performed at several speeds to confirm the accuracy of the affinity laws. I have not been able to find that paper, but it did prove affinity laws were not 100% accurate.

D23

 

[Artisi]

What is full speed flow and head (this can give us some idea of what the efficiency could be) and what is the calculated efficiency from the reduced speed testing.

 

[ccfowler]

The validity and accuracy of the affinity laws all depend upon the real, known BEP at rated speed as the reference point. 53% of rated speed is already getting a bit shakey for great accuracy, and your calculations are apparently based on the presumed BEP performance at rated speed. I presume that this is necessitated by limitations of the test facility.

I would take the indication of an absurd efficiency at rated speed by the scale-up calculations to be a sufficient indication that the scale-up is not sufficiently valid.

 

[Artisi]

" ridiculously high " is a relative term based on a perceived expectation. If this is a "large" pump high 90's are achievable and a half speed test for large pumps is an acceptable test speed.

 

[Artisi]

Just to add a little more to my previous post - large pumps, say 2-3000hp + when designed and tested have to be done on scaled down units and some times at reduced speed, the major worlds manufacturers, Weir, Sulzer, Allis- Chalmers, Worthington etc etc are able to calculate/ predict full size / speed results with great accuracy as most applications / sales have to carry performance guarantees including efficiency which reflects in power consumption.

 

[thrusterpump]

Artisi is correct in the level of accuracy in reduced speed tests when lab calibrated drivers are used with known calibration curves (speed vs efficiency). If your using a job driver with only a guaranteed efficiency, then there is some uncertainty in the efficiency of the pump at reduced speed when your using a VFD. Also affinity exponents for the pump have been determined through testing. From one pump to the next, the exponents are close but not right on. my 2 cents