Derating centrifugal pump performance for low NPSHa 2

[Gadsen]

Does anyone have a formula or method for predicting the performance for a single stage centrifugal pump based on lower than required NPSH?

I would like to be able to re-draw a pump curve based on the difference between the NPSHr denoted on a pump curve vs. the NPSHa I have in the real world.

This is for a small pump (15hp motor) pumping water.

Thanks!

 

[rmw]

There is a point on your pump curve where the Npshr exceeds the npsha. You can't re-draw the curve, because your pump curve is what it is, assuming that you don't do something like change speed or impeller geometry.

But, you can control the flow in some acceptable fashion so that the pump does not pump out far enough on its curve to get into the region where it doesn't have enough npsha.

You can do this with flow control devices, or you can design downstream piping with enough head loss so that the total head loss limits the pump.

I have done both with pumps that had npsha limitations.

You can also change your suction piping so as to minimize restrictions that lower npsha. Upsize the suction piping several sizes so that the head loss due to velocity is minimized. Don't exceed 6 fps, and get below that if practical. Get rid of restrictions like globe valves, (use ball valves if valves are necessary) piping size changes, elbows, etc. Move check valves to the downstream side of the pump if there are any upstream. Use an eccentric reducer at the pump suction to change from the upsized pipe to the pump suction size, and be sure to install it with the flat side up.

Get rid of vapor traps, (horizontal runs) in the suction piping.

In my experience, most npshr curves are relatively flat to a point, at which they begin to tail upward rapidly. The above suggestions are to keep you operating in that flat region. If you can't operate in the region where the npshr curve is flat, then you need a different pump.

rmw

 

[Gadsen]

Thank You rmw. I had already implemented all of your suggestions that were applicable, but was hoping for a magic bullet, which I probably really knew didn't exist. I've been handed a mess here (which they want to expand!), guess it's just time to recommend the drastic changes I've been trying to avoid.

thanks again for the prompt response.

 

[scalleke]

One a pump starts to cavitate things just start to go from bad to worse and one thing to certainly NOT do is to try to operate the pump with too little NPSHA.

Have you looked at all the possibilities to reduce the NPSH Required from the pump. There the manufacturer can maybe help. Installing an inducer, lowering the pump speed, increasing the inlet eye of the pump. Pillishing the inlet side of the pump. Lowering the pump position physically. Etc.

What NPSH Available and NPSH required do you have?

Best regards.

Scalleke

 

[Gadsen]

scalleke,

The one thing I had not considered was turning out the impeller eye.

The existing eye diameter is 2.293". If I leave .030 (per side) for material integrity (SWAG), I could increase the eye diameter to a max of 2.550.

Do you have a method for predicting the performance change, or is this a "try it and see" proposition? If nothing else, decreasing the angle of incidence at the leading edge of the eye would certainly relieve the low pressure area at the leading edge of the eye. Would you consider this amount of change worth the effort?

 

[scalleke]

I did not intend to suggest opening up the impeller inlet eye. Rather to do anything you can to reduce friction and obstruction in the pump inlet.

From the inlet flange to the time that the liquid enters the impeller there is quite a bit of cast surface on the sides that will cause quite a bit of friction. Polishing this will reduce friction and you will see a decrease in NPSH requirement of the pump.

Do not increase the impeller inlet eye. NPSH Requirement of the pump will go up.

Do you have the possibility of lowering the installed pump. It sounds like you have only a small cavitation problem as you still have some pump performance. Lowering the pump might give you what you need. I once requested to remove the 1 foot high cement base under a pump to solve a problem (I'm glad to say that it worked).

Can you share some more info on the installation.

Best Regards. scalleke

 

[Gadsen]

Actually I now have the problem solved. It was primarily political. I now have them, with your help, conviced that there is no other answer than to core through the wall and the back of the water fall/cooling tower in order to improve the suction side of the situation.

As it exists now I have about 100% excess hp online. As the project exists I have about a 10 ft NPSHR and have about 8 available. The pump is cavitating like crazy but it moved enough water to just get the job done. It is primarily a badly "engineered" suction side design. Once I am allowed to correct the suction situation I can actually more than double the capacity using the same pump. I will have a 16 ft NPSHR and will have 24' available (that's how bad the suction train was).

I basically knew better, but there is always a chance of learning something new, and your input helped me change the minds that needed changing. For that I am quite grateful.

Thanks guys.