calculating pump efficiency 1

[pumpedup]

Does the efficiency of a pump change when the manufacturers pump curve is shifted upward as a result of a negative suction head? Ultimately I am trying to determine the predicted BHP of a pump operating at a given speed in order to predict the saving of implementing VSD.

 

[Scipio]

Yes and no, I'd expect efficiency to change, but not the pump curve - depends on what you mean by negative suction head. I'm not sure if you mean you plan on running the pump is a suction lift condition, or just operating at a lower-than-rated suction pressure.

Basically, if I assume you're talking about operating a centrifugal pump under a suction lift condition (that is, liquid level is below pump suction) the manufacturer's curve shouldn't change. I say shouldn't, because every manufacturer's curve I've seen describes flow vs. total developed head. It's a differential pressure curve that is only changed by changing pump speed or impeller geometry.

However, if you run a pump at a lower suction pressure and the same discharge pressure, it will increase the differential pressure (total developed head) that the pump sees, it will probably change the location of the operating point on the pump curve, and result in a reduced flowrate. Centrifugal pumps are head-hunters, they'll go to the point on the curve which matches the differential pressure they see, and deliver the flow corresponding to that point. As the operating point moves around on the curve, efficiency does change. It's impossible to predict without knowing the curve, but it should be easy to read off the manufacturer's pump curve once you know what you're looking for.

Sorry, I know that's a little vague, but if you can give me a little more detail I might be able to tighten that up a bit.

 

[chubPE]

If you are going to run a vacuum, then the pump mfg. can give you a performance curve for that. Positive displacement gear pumps can be designed to pull a vacuum while still maintaining 75% of rated flow, depending the the fluid.

 

[BUFFY]

That is wrong, I think…. We all know when we calculate efficiency we need head, flow SP gravity, 3960 and HP, so if one variable change your efficiency sum change. Just think of it as a ratio, if you change the denominator you must change the nominator or Vic versa to get the same sum. If you look on a good performance curve, like Taco’s you will see the drop in Eff.

 

[BUFFY]

One more thing use a VFD, it will solve all your problems…

 

[d23]

pumpedup:

Pumps add kinetic energy to the liquid produced. If your pump will produce a dp of 100 PSI at a given rate then the discharge pressure will be the intake pressure plus the dp at that rate. Fluid vortex or cavitation will change the operating characteristics of the pump. Outside of that the pump only efficiency is simple:

(GPM * Total head in feet * Ave SPGR) / (3960 * BHP)

Changing intake pressure changes the discharge pressure because they are additive. It does not change pump efficiency unless you get the intake pressure below the NPSHr of the pump.

Hope this helps!
D23

 

[ccfowler]

pumpedup,

The usual pump "curves" (head vs. flow, power vs. flow, NPSHr vs. flow) are based on tests with the pump operating at a constant shaft speed.

Within reason, these curves are translated according to the affinity laws as the shaft speed is changed from the test speed. Specifically, flow varies directly with shaft speed, head varies with the square of shaft speed, and power varies with the cube of shaft speed. These affinity relationships hold best at the BEP (best efficiency point), and other points of these curves vary proportionately.

Driving a pump with a VSD effectively translates the pump's curves, and the pump operates on its translated performance curves. The actual operating point is the intersection of the pump's translated head vs. flow curve and the corresponding system head vs. flow curve.

The simple answer to your question is that there is no simple answer. You must spend significant effort to determine whether the use of a VSD is appropriate to the particular application under consideration.

Buffy,

The use of a VSD is not always appropriate! In some cases, the benefits of a VSD can be astounding, but in other cases, a VSD can provide almost no benefit. Each case must be evaluated on its own merits.

 

[lilliput1]

Pump bhp = (GPM x Ft Head)/(3960 x Pump Eff)

Head 2 = Head 1 x (GPM2/GPM1)^2

Power hp input to motor = bhp/motor eff

1 hp = .746 kw

To get quick order of magnitude cost savings assume pump eff & motor efficiency as varying from at full to minimum at minimum speed. Use spreadsheet so you can easily vary your assumed efficiency & vary the calc for savings accordingly. You may even plug in a formula for variation of part load eff based on peak eff & % part load so the only input you have to put in is the peak eff.
You can then copy special - copy values the answere & put them in a block for the efficiency you used. Input the next efficiency & copy special - copy values. So you can generate blocks of savings for each efficiency figures you estimated.

 

[goutam_freelance]

The pump curves are understood now. I find VSDs are desirable for the pumps where there is a throttling at the pump discharge for control purposes and where pump operates at less than rated condition much of the time. Example-boiler feed pumps.

 

[canox23]

Pump Eff.% = Gpm X 8.33 X Ft.Hd / Bhp X 33000