Pump capacity check

[C4Reactor]

Dear all,
Is there any step-by-step method for checking pump capacity?
I am looking into increasing a flow across a system. The volumetric flow is o.k with the pump but the density has increased. How do I translate this change into the pump curve?

Appreciate any ideas

 

[d23]

C4Reactor

I assume you are talking about centrifugal pumps:

In general terms the head and flow of the pump will remain constant due to minor changes in the fluid density. The BHP (Break Horse Power) will increase or decrease in a liner fashion with the specific gravity of the fluid. As an example if you assume fresh water to have a gravity of 1.00 and sea water to have a gravity of 1.03 I would expect about a 3 percent change in horsepower required.

This is general terms. If you have viscosity or vapor involved all of the calcs can make much more dramatic changes.

Hope this helps!

 

[ccfowler]

C4Reactor,

d23 is exactly correct. I've had some experience with some marginally sized pumps where the seasonal density changes in their fresh water pumpage caused some considerable concern. The original design presumed a temperature range that was not sufficiently realistic, and the lower than expected water temperatures increased the density enough to exceed not only the motors' ratings but even slightly more than their rated service factor.

A saving grace for the insallation was the combination of the characteristics of the pumps, the relatively simple circulating arrangement of the connected system, and the reduced actual shaft speed due to the "excessive" load on the motor. Since pumps ALWAYS work on their curves and everything is keyed to the actual shaft speed, the installations survived for a very long time (many years) with excellent reliability. At times, the motors needed some additional cooling assistance in the form of fans to provide more generous quantities of cooler air to the motors' cooling air fans.

I relate this experience because you may find that your installation may be able to "squeak by" at little additional cost in a similar fashion.

The all important affinity laws can probably be applied with relative simplicity to a situation such as you describe. The primary considerations for the pump are:

1- Centrifugal pumps produce differential "head" (not pressure) and volumetric flow rate. (For a constant head, the greater density will result in the pressure being increased in proportion to the ratio of the densities.)

2- The volumetric flow rate varies directly in proportion with the actual shaft speed.

3- The head varies directly with the square of the actual shaft speed.

4- The power (at constant fluid density) varies directly with the cube of the actual shaft speed.

Assuming that the only change in your system is an increase in the density of the flow, then you can use the above relationships to estimate the performance changes since the pump is probably working on just about the same points on its curves as it was before the density change. Since the pump was apparently operating satisfactorily prior to the density change, it would not be necessary to know the full details of the pump's curves. You can probably simply base your estimates on proportional relationships.

For estimating purposes, I've found it practical to estimate the torque output of an induction motor as being a linear function of its actual operating speed relative to its synchronous speed. The motor is presumed to produce zero torque at synchronous speed and "rated load torque" equivalent to rated power at rated speed. This relationship can be considered to apply for loads greater than the rated load within any range that the motor can sustain for an extended period of time (a realistically moderate overload condition).

By spending a little time with a calculator or a computer spreadsheet, you should be able to estimate the probable significance of your system's performance change due to the density change. You would surely be aided by getting an accurate measurement of the actual shaft speed of the pump to serve as a reference point for your calculations.

Good luck, and please let us know about the results of your evaluation.

 

[dadfap]

d23, I find your answer refreshingly simple, all too often people try and overcomplicate the problems associated with centrifugal pumps. The density / temperature relationship is often overlooked by those who specify especially submersible centrifugal pumps.

 

[d23]

C4Reactor

A couple other thoughts about density:

This will change your NPSHa at the pump intake. If going from a higher specific gravity fluid to lower gravity you will need to make sure you have enough intake pressure to prevent cavitation.

You need to check the pump seals, O Rings etc. for any compatibility issues.

Any radial or axial bearings need to be checked for the new pump load.

I used an example of fresh water and sea water stating the expected change in head and flow would be negligible. If you are making an extreme change say from saturated brine (1.17 SpGr) to xylene (0.87 SpGr) there will be a change in head and flow.

Hope this helps!

 

[Kawartha]

Hi D23,

My understanding is that centifugal pumps will provide a specific volume of liquid at a specific head. Changing the Specific Gravity of the liquid does not alter this. What can effect the pump output is a change in liquid viscosity, in which case capacity, head and bhp are all affected. Of course you would be right to expect some change in viscosity with major differences in S.G. Correction factors for viscosity are available in the Hydraulic Institute manual.

So I agree with your comments in your first post above, but it also applies to large differences in S.G.

Richard

 

[d23]

Kawartha:

You are correct. I should have explained my thought a little better!

In a lot of cases I think of pumps as being a dp device. Intake pressure plus dp equals discharge pressure. A very large change in the SpGr will affect the pump discharge pressure due to the intake pressure available. If the system curve is constant you can expect a change in both head and flow from the pump depending on where the pump curve and system curve intersect with the new gravity or density.

ccfowler

Your motor situation was informative. The extra fans are something I would not have though of. In that case did you look at power factor correcting capacitors? Lowering the system kW will help a motor. Maybe not enough for your problem.

dadfap

Thanks for the complement. I would like to know who you work for if you wouldn’t mind.

 

[ccfowler]

d23,

This was a very marginal but adequate installation. Most of the time, the motors were operating within their 1.15 service factor (admittedly not by much). The pumps were in a fairly small (but again adequate) space, and the simple ventilating fans were enough to assure that the warmed air in the room was not recirculated to the motors' fans (vertical motors) and that plenty of cooler outdoor air reached the motors' cooling fans to keep them sufficiently cooled.

Since these pumps and motors did work well and reliably, and any changes would have required great expense and disruption of plant operation, it made no sense to do anything but leave them be. The simple fans were quite enough, and the power supply was quite adequate.

I didn't get involved with these pumps much until they had been in operation for many years (complete with the improvised "supplemental cooling system&quot. Consideration was given to making some changes, but there really was no need to fix what wasn't broken. Their periodic maintenance and repair costs were quite reasonable and normal.

I've often thought of this as an example of why designs should include some reasonable margin. Obviously, this installation had just enough margin to not only avoid what could have been a very expensive and disruptive mess but actually provide many years of reliable service.