Noise from Discharge Side of Centrifugal Pump 2

[SrinivasAluri]

I am trying to create a demo centrifugal pump to show increased vibration from cavitation. I have a pump running at 2900 RPM and a 2 inch suction side pipe and 2 inch discharge pipe with a increaser. The pump is a Kirloskar DB 32/13 End Suction Pump.

I can get cavitation by closing the suction side valve and my pressure sensor on the inlet side is confirming that -pressure is close to vacuum and I can see bubbles. The problem is the pump is too noisy without closing the suction valve i.e., without cavitation. As you can see in the picture I use transparent piping and when I run the pump, I see a vortex of air bubbles on the discharge pipe. The increaser (1 inch to 2 inch) is not smooth.

Ideally, I want the vibration to increase when cavitation is simulated but in my case the pump is too noisy without closing the suction valve and the noise is going down once I start closing my suction valve, which unfortunately is opposite of what I want to show. The broadband noise around 5000Hz is going down when I close the inlet valve.

Initially I had water from discharge pipe fall openly not submerged, but I fixed that. The noise on the discharge side sounds like gravel moving.

Appreciate any help.

 

[LittleInch]

There doesn't seem to be any discharge pressure / flow control valve??

I suspect you are cavitating due to too much flow with the inlet valve open.

A gravelly sound is typical of pump cavitiation.

A DB 32/13 delivers 5-6m head and up to 9 m3/hr

http://www.kirloskarpumps.com/download/prod_catalogue/04 DB.pdf

Without throttling the discharge down stream of your clear pipe you will be flowing more than that.

You need to measure and control flow.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[SrinivasAluri]

Yes. There is no discharge valve, I will add one. But I thought cavitation occurs on suction side, when pressure is close to zero? There are no bubbles in suction pipe and the suction pressure shows 20 psi with inlet fully open and close to 0 when almost fully closed. Also, Did I create the problem by expanding the size of discharge pipe to 2 inches?

Just to reiterate the pump is running at 2900 RPM

 

[LittleInch]

Yes but if you run the pump beyond max flow it can easily cavitate.

Yes the 2" pipe will have a lot less friction than the 1" pipe so will generate much more flow for the same length pipe.

Keep flow below 9 m3 / hr and preferably below 7 for this pump.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[SrinivasAluri]

As I understand if I keep the flow below 7 m3/hr by controlling the discharge valve then it would not cavitate. Then if I want to create cavitation, I can just close the inlet valve?

 

[Strong]

If you have the pump performance curve, then control discharge to reach BEP to minimize vane pass pulsation and Cavitation. You could then induce Cavitation by reducing suction pressure or reducing discharge pressure and running out on the curve (high flow). Note that Cavitation bubbles may collapse inside pump case and be heard but not visible, so you will not see them in the pipes. High flow is more likely to induce Cavitation on the discharge impeller vane tips, so bubbles may still be present in the discharge pipe before collapsing. Measure vibration (acceleration) near pump cut-water, or on the discharge flange or on the bearing housing to make a quantitative assessment of Cavitation rather relying on visible bubbles or audible sound.

Walt

 

[LittleInch]

Strong has it right.

You need to look at the pump curve for your pump. I took those numbers from the generic set of curves in the brochure.

Let us know if it works

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Valvecrazy]

You should also test how low the discharge flow can be restricted before cavitation becomes a problem. Most pumps can have the discharge restricted much more than you would think without causing any problems.

 

[SrinivasAluri]

Thanks for all the replies. I am using an accelerometer mounted in horizontal position on the outer casing of the impeller. So I can see the spectrum and it shows up as broadband noise if it cavitates.

@littleinch - The pump curve is same as shown in the catalog, running speed is 2900 RPM. Let me try controlling the discharge valve and I will post the results.

 

[Strong]

The vibration waveform can be a better indicator of cavitation by looking at the peak values and the shape. The vibration spectrum requires several averages for a steady or repeatable measurement. The spectrum shape typically has one or more rounded peaks that can be related to cavitation bubble size or usually the vibration response of the structure (pump case).

Walt

 

[SrinivasAluri]

I have reduced the discharge pipe dia from 50mm to 32mm and installed a discharge valve and had to close the valve by 80% to reduce cavitation (50mm dia), there is still some vibration in the 4000-6000 Hz range. The discharge valve is an electronic one and it can't be changed now and it is at 50mm internal dia. What else can I do to remove the remaining noise.

Please see the attached spectrum.

 

[LittleInch]

I suspect the noise is actually that of the pump itself with the vane pass. Is this frequency some fixed multiple of the RPM of the pump?

Are you able to measure flow?

You might need a pulsation damper, but it's difficult to see how much noise this is.



Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Strong]

"The discharge valve is an electronic one and it can't be changed now and it is at 50mm internal dia. What else can I do to remove the remaining noise."

You may have flow noise from the discharge valve. If your spectrum amplitude is scaled in acceleration-g, then vibrations are already low. If you really want to reduce valve noise, then you would have to look for a quiet valve. They are available, but expensive. Otherwise, you have to live with the residual noise/vibrations from the pump and discharge valve.

Walt

 

[LittleInch]

What sort of valve is your discharge valve?

If it's an isolation type valve ( ball, butterfly or gate) then you're going to get significant noise from it at 80% closed.

You really need some sort of control valve designed for low noise or maybe just get a piece of flexible pipe and add a clamp or two to it... and just don't shut it off completely...

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Artisi]

Run the pump at 4 pole speed and solve your noise problem, you can still demonstrate cavitation plus you might even get to hear it (cavitation) above the 2 pole operation.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[SrinivasAluri]

It is a broadband noise from 4000-6000Hz. There might be peak at vane pass frequency. I have a way to measure flow, what would that tell me?

 

[SrinivasAluri]

It is a Characteristic Control Valve (CCV)EP050R+MP -

https://www.belimo.eu/pdf/e/EP..R_MP_datasheet_en-gb.pdf

 

[LittleInch]

Flow would tell you whether you're in the normal range of the pump, i.e <8-9 m3/hr

Your valve is too big and you're working at below it's normal range ( 30% of 32 m2/hr for the 50mm size) and the data sheet states 2 bar max differential for low noise so that should be ok as you've only got 5-6m head.

The operating range is only to 20% closed so you're at max closure.

Get a smaller valve or introduce some lengths of smaller diam pipe somewhere.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[SrinivasAluri]

"Get a smaller valve or introduce some lengths of smaller diam pipe somewhere."

Just so that I understand are you suggesting increasing the length of pipe to increase the discharge head? Can I add couple of elbows to increase the head?

 

[Artisi]

Round and round in circles. Reduce the the flowrate, what's the point of running a pump at high flow and then reducing the flow by throttling the output, run at 4 pole, reduce the impeller diameter or put up with the noise from an extremely poorly selected pump / pump installation.

End of story.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)