plotting pump curves

[199292]

I have been asked to plot pump curves for three 125 horsepower pumps that are presently installed at our water treatment facility. The pumps are 1-stage VARIABLE FREQUENCY DRIVE vertical turbine pumps. We have the pump curves from when they were installed but would like to plot the present day condition of the pumps. I would like a methodology to performing the test. I've always had to interpret the data. Now I am being asked to create the data. Some guidance and direction would be appreciated

 

[hydrae]

199292
To derive the pump curves for current performance you will need
outlet pressure minus inlet pressure (H)
flow (Q)
speed (N)
for each operating condition
then using the speed unity equations

Q1/Q2 = N1/N2
H1/H2 = (N1/N2)^2

Adjust any data points that are taken at lower speeds and plot as if at full speed
Hydrae

 

[BobPE]

199292:

What are the flow rates for the pumps? Head? What are you using them for? Location in the plant? Age? There are a lot of ways to go about what you need to do. One way would be to get the help of your engineers since you may have to figure a way to isolate the pumps from the plant and you will want help for that so as not tu upset the plant.

BobPE

 

[Ballintoy]

You should be able to generate the curves rather easily. This is assuming you have decent flow meters, pressure gauges, and isolation valves on the outlet side of the pumps. Allow the pumps to run full blast at some moderate speed. Record the speed, flow, and pressure. Using the isolation valve, slowly throttle the pump to maybe 4/5 of the maximum flow and again record the pressure and flow. Continue throttling at 1/5 points until you reach shutoff head or get nervous about blowing something up. Convert the pressures to feet. With that information you will be able to plot a curve. Use the pump affinity relationships mentioned above to develop curves for other speeds.

 

[Ballintoy]

I forgot to mention something in my previous response. Unless you can pump directly to atmosphere somewhere near the pump, you will only get part of the curve; however, the new curve should be similar to the orignal curve so you should be able to extrapolate. Also, the pressure gauge needs to be between the pump and the throttling valve.

 

[199292]

The comment about getting nervous about blowing something up I'm sure is in jest but it doesn't make me warm and fuzzy. The low lift pumps operate under an existing head that is reasonably easy to figure out. The high lifts pump to a distribution system with tanks and pressure zones attributed to it. Hydrae mentioned the difference in pressure betweeen the outlet and inlet. Mow do you apply this to a vertical open shaft pump? Is the inlet pressure the static pressure at the inlet of the impeller. I have been looking at the original data ( computer printout) and the pressure on the discharge guage seems to be the key pressure.

 

[hydrae]

199292

If you are drawing from an atmosperic tank, your suction pressure is the negative value in feet between the discharge pressure gauge and the open surface of the drawn water. The diffence in pressure is to be taken a the same elevation, if you use atmosheric pressure then the discharge pressure will also need the same adjustment. Whatever units you use the suction and discharge pressure need to be the same and adjusted for any elevation difference.

As for getting nervous, has the pump been dead headed at full speed recently? There is concern about degradation of the pump in terms of performance, is there also degradation of the discharge pipe, that is where things may blow up. Particulary if you close down a discharge valve at full speed on a thin wall pipe with lots of internal corrosion.

The other thing to be concerned about while performing these tests, is if you run the pump where it will cavitate, it will get noisy and do all sorts of damage, something to avoid.

Hydrae