Pump efficiency at same flow but lower RPM

[TimboGMO]

In the following thread: thread407-166705 the following quotes state that efficiency stays relatively unchanged at different rpms

"This is what Sulzer has to say on the subject of efficiency as function of speed of rotation:


Quote:
Where the system curve derives wholly or mainly from friction losses, speed control is best, because pump efficiency then remains prctically constant...With small speed changes (up to 10%) efficiency remains virtually unchanged. With bigger speed changes the velocity in the channel alters and with it the Reynolds number. The efficiency factors must be downgraded at lower speeds and upgraded at higher ones.

Sam Yedidiah:


Quote:
A comparison between the data calculated from tests performed at 1740 RPM and the actual data from the test at 3550 RPM leads to the following conclusions:

1. As far as the QH curves are concerned, the data calculated from tests at one speed seem to be reasonably accurate for the other speed. However, at the lower speed, the QH curve seems to be able to go out to a greater Q/Qd ratio than at the higher speed, possibly owing to cavitation in the casing throat.
2. The efficiency curves show certain differences, but they rarely seem to exceed ±3%.
3. The NPSH requirements at 3550 RPM, calculated from the results at 1740 RPM, vary considerably from the NPSH requirements determined from the tests performed at 3550 RPM. In terms of suction specific speed, the pump seems to have a different suction specific speed at each of the two tested operating speeds."

Yet in the following technical paper on page 4 when removing a control valve and installing a VFD to achieve the same flow rate but at a lower head, the technical paper says efficiency declines an appreciable amount.

http://ieeexplore.ieee.org/iel5/28/19171/00887227.pdf

Does anyone know where I could find an equation to estimate how efficiency would change with changes in RPM if not given on the pump curve? Thank you,

-Tim

 

[stanier]

Suggest you go to

http://www.powerqualityanddrives.com/energy_savings_pump_vfd_part_1/

for the whole paper. If you select Figure 4 in the text it will highlight part II of this paper.

 

[BigInch]

The supposition that efficiency does not change with a change in rpm is only valid when the system curve exactly tracks the pump's efficiency curves as does system curve 1 in the diagram.

When the system curve has significant static head, indicated by curve 2 crossing the ordinate axis well above zero, the flowrate will not track the same pump efficiency curve with change in rpm. The system curve crosses over the pump's efficiency curves.

Thus the efficiency at the full speed operating point may be 70%, but at the half speed operating point, may be much different, in the figure here shown at 63%.

http://virtualpipeline.spaces.msn.com

"What gets us into trouble is not what we don't know, its what we know for sure" - Mark Twain

 

[BigInch]

And with the same flowrate, drop straight down from full operating point to the intersection with the 1/2 speed pump curve and you have about 66%, but notice that your system curve indicates that you cannot flow there unless your system curve can somehow be changed. Changing the system curve to pass through that point might be accomplished by somehow reducing the static head and opening up flow in a parallel pipeline.

VFDs work when your flow requirements match the ratio of variable_rpm / full_speed_rpm * BEP_flow AND your system curve head matches the ratio of (variable_rpm^2 / full_speed_rpm^2 * BEP_head. Otherwise a constant speed pump with a control valve or some changes to your operating schedules or a larger tank may serve your purposes much better.

http://virtualpipeline.spaces.msn.com

"What gets us into trouble is not what we don't know, its what we know for sure" - Mark Twain

 

[quark]

A bit difficult to get in the absence of iso-efficiency curves of pump. If you see the compound efficiency(pump+motor) of pumping system, then you may find considerable difference. If you know motor efficiency at the given part load condition, then you can try calculating the efficiency of the pump, provided you exactly know the flowrate and differential head.

If we focus mainly on power consumption and not the efficiencies, speed control method is better over throttling, when the pressure drop is predominantly dynamic.

 

[TimboGMO]

BigInch,

You wrote:

"And with the same flowrate, drop straight down from full operating point to the intersection with the 1/2 speed pump curve and you have about 66%, but notice that your system curve indicates that you cannot flow there unless your system curve can somehow be changed. Changing the system curve to pass through that point might be accomplished by somehow reducing the static head and opening up flow in a parallel pipeline"

I am currently looking at lowering our boiler feed pump pressure to a certain pressure set point (to be determined) by means of a VFD. Our existing control valves would then have to open up wider, allowing less of a pressure drop to deliver the same flow rate at the proper pressure to our boilers. This would change the system curve. Using the pump curve you posted and assuming I was initially running at full speed and was at the 63% efficiency point. Lowering the speed to 1/2 speed and changing the system curve by opening up my control valve would actually increase pump efficiency to approximately 67% according to the efficiency curves (assuming I could lower resistance enough to drop straight down from full speed to the half speed level and keep the same flow). Correct?

 

[BigInch]

It should work like you describe, if the valves can give the lesser pressure drop you will need and can do it at the same flowrate. You'll have to check that, so find the valve's Cv at full open or so.

PD = (Cv/Q)^0.5
PD = new pressure drop,
Q = new flowrate,
Cv = valve coefficient at the new position (%open)
convert that to head and draw it on the system diagram.

Keeping the same flowrate, your system curve keeps the same curvature and just moves up or down on system diagram above, depending on the head drop at the control valve at any given time.

Head equivalent of Suction pressure +
pump differential head at reduced rpm -
control valve head loss at new %Open -
piping head loss (flow loss and elevation change if any)
= new Boiler inlet head

Convert that new Boiler inlet head to pressure and you have the new boiler inlet pressure.

Be carefull with pump speeds below 60%, as you may not develop enough of head for your flowrate. Lower than that you may also find your motor power costs increasing due to its lower efficiency at partial load.

http://virtualpipeline.spaces.msn.com

"What gets us into trouble is not what we don't know, its what we know for sure" - Mark Twain

 

[quark]

The head of boiler feed pump depends upon the boiler load and thus steam pressure.

During its operation the feed pump doesn't run at one fixed operating point. The operating point shifts along the new pump curve (at reduced speed) as per the steam pressure in the boiler. In this case, you have multiple efficiencies at various instants of time. This will vary the flowrate as well.

If you wish to provide a fixed flowrate, then pump speed should be manipulated continuously. Suppose, you want to maintain fixed flowrate. When the steam pressure increases, the operating points shifts to the left (on the new pump curve) and flow reduces. The flow sensor then indicates of a speed ramp up to maintain the flow. So, ultimately you will end up at a point where the flowrate and pump head matches and this may not be a point of your interest.

In a nutshell, thinking about efficiency in such systems is not practical.

 

[stanier]

When considering the efficiency you have to look at the over efficiency not just that of the pump.

The referenced paper needs to be read in its entirety. Points to consider:-
1) the energy losses in the VFD cubicle could amount to 5%. Thhis heat energy needs to be removed from buildings.
2) published motors efficiency data is based upon a perfect sine wave as in running DOL. Data is not produced for efficiency when the sine wave is regenerated as a chopped square wave and could also amount to 5%. There are many harmonics that also contribute to losses
3) running motors slower than full speed will change the power factor. Unless you invest in PF correction more energy is lost.
4) Typically slower running motors will need auxilliary cooling by additional motor driven fans.
5) The switch room may need to be forced ventilated or air conditioned to dissipate the heat generated by the VFD cubicle.
6) VFD cubicles take up room in the switchroom. Real estate has a value.
7) VFDs may last only 10 years before needing replacements. parts become harder to get.

Do you NPV calculations and VFDs are not a real good solution compared to control valves in may cases, particularly where the friction losses are <50% of the system losses.

The people selling VFDs are not the best people to listen to when they push the line on energy savings that can be made. They are there to sell their gear for their profit. They have little interest in your costs.

 

[BigInch]

Timbo, Stanier, Excellent advice. I was only intending to answer the questions asked, so my examples were meant simply to supplement the references given. As you note, and to which I fully agree, the suitability of VFD are highly dependent on the specific characteristics of each individual systems and the degree you are willing to accept both the possible advantages and disadvantages that go along with them.

http://virtualpipeline.spaces.msn.com

"What gets us into trouble is not what we don't know, its what we know for sure" - Mark Twain

 

[TimboGMO]

Thanks for all the replys, they will help me in my analysis. The VFD room cooling requirements and VFD maintenance reasons listed above have led us to use more Magnadrives around our plant which might be applicable in this application since feedwater from our deareator is only 220degF and demagnetization only occurs about 300 degF for Magnadrives.

 

[ScottyUK]

It is worth having a look over some previous threads regarding eddy current drives. They have their place in certain niche applications, but in general they're an outdated technology. If room cooling is an issue an eddy current drive is unlikely to be a wise choice. You may have to scroll through some of these threads to get to the relevant bits.

http://www.eng-tips.com/viewthread.cfm?qid=98040

http://www.eng-tips.com/viewthread.cfm?qid=107895

http://www.eng-tips.com/viewthread.cfm?qid=187627

----------------------------------

If we learn from our mistakes I'm getting a great education!