VFD pumps and efficiency

[smbunn]

We have been trying to determine the impact using a VFD drive has on the efficiency of large water pumps used by municipalities. Let’s say in a typical 110 kW pump driving a single stage double suction split case centrifugal pump, that we add a VSD.
The fixed speed pump was operating at a BEP of 90% at 1470 RPM. If I reduce the RPM to 1200 RPM I know from the affinity laws how to shift the head/flow curve and the power curve. However what happens to the efficiency curve? Does it retain the same shape and peak and just move to the left on the flow axis? Does it move left and drop down on the efficiency axis? Is there a general formula to calculate the new efficiency curve based on the speed ratio of N2 / N1?

 

[Artisi]

usually the pump manufacturer can supply curves for lower speed operation, however in your case for the speed change you have I would use the same efficiency at 1200 rpm as shown at the equivalent 1470 rpm.

 

[smbunn]

This is a general question as I want to apply the correct policy to all municipal water pumps. Some are 4 MW and above. At 80% speed we are only using 51% of the power, and therefore not doing a huge amount of work with the pump. At this point I would assume some of the VFD losses and motor losses start to become more important, but they are well known. Its getting an accurate picture of efficiency that is important as even 1 to 2% is important when these pumps run for long periods of time.

 

[Artisi]

Contact the pump designer / manufacturers, that are the only ones to give you accurate information unless of course you have the knowledge / skills to undertake on-site performance testing of these units and the reduced speeds, which could be a good idea anyway as you reduce all guess work.

There are no rules for efficiency change due to speed change - only assumptions based on knowing the particular pump design etc.

 

[smbunn]

Thanks Artisi. I had assumed this might be the case. For example there is an assumption, derived probably from two point curves in hydraulic models, that centrifugal pumps for potable water have quadratic head/flow curves. Of course this is not the case, the curve can only be obtained from empirical measurement, from the pump manufacturer, and it's shape is a complex result of manufacturing materials, surface roughness, pump volute and impellor shape and a raft of chaotic influences. The good thing is that given an empirically measured head/flow curve at least the affinity laws allow us to scale it for different speeds and impellor sizes.
However if I am given a pump efficiency curve by a manufacturer, can I not use the same affinity laws to scale the efficiency curve? Has anyone done this from first principles and compared it to measured data? can anyone point me to a manufacturers web site for water pumps where various speed curves are shown for a pump? I have found no end of sites where different impellor curves are shown but none with speed.

In fact given that;
R is impellor radius, d is speed

Capacity varies directly as the speed or diameter:
G2 = G1 (R2/R1) or G2 = G1 (d2/d1)
Head varies as the square of the speed or diameter:
H2 = H1 (R2/R1)^2 or H2 = H1 (d2/d1)^2
Horsepower varies as the cube of the speed or diameter:
P2 = P1 (R2/R1)^3 or P2 = P1 (d2/d1)^3

So the impellor curve relationship appears to be the same as the speed relationship. I have impellor size graphs that clearly show the peak efficiency drops as the impellor gets smaller, so can I make the reasonable extrapolation that EXACTLY the same drop occurs when I vary speed?

 

[TenPenny]

If you want to play around with on-line pump curves, go to

http://www.gouldspumps.com

and use the PSS Pump Selection On Line system...once you've picked a pump, you can use the Line Curve tab to generate multi speed curves.

The link below is to a scanned copy of the variable speed curve for a 3196 6x8-13 - you can see how the efficiency varies with speed as an example.

 

[BigInch]

The best guess you can make, without having the mfgr's exact curve at different speeds, is to assume the pummp efficiency remains the same as rpm varies, BUT, and this is an extremely important "BUT", it also depends on the system's static head at the reduced head. When speed is reduced and there is significant static head, the efficiency does NOT track directly. It shifts farther to the left and in greater proportions each time speed is reduced. So, knowing exactly what your system curve looks like at lower flows becomes very important.

Another warning! Some manufacturer's programs do not ask for system curve information, only pump information. That is a signal that they DO NOT take into consideraton the effects of the system curve in their calculations. Therefore, if you have static head, be extremely careful with any results you get from them until you are positively sure that they do consider system curves too.

Other important factors that enter into the calculation of overall efficiency are the efficiencies of both the motor and the VFD when they are operating at reduced load. It is another thing that some pump manufacturer's programs do not consider. Under low power loads, those efficiencies reduce very fast. Be sure they consider those reduced efficiencies before believing any predictions of energy savings at the lower power loads.

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/