Pump Cavitation Remedy

[kryanl]

Hello,

We have a situation where a pump is cavitating. I think that some of teh probelem lies in the suction piping, since calculations show we have enough NPSH (although, admittedly, we could have more). We are pumping out of the bottom of a tank, where the water (200F) level is 15' above the suction of the pump. As the pipe comes down, there is a 90 degree bend, which goes directly into a 45 degree bend (the 45 is in the horizontal, ie. it changes the pipe direction from NW to due N). We think that this transition is causing some vortexing or flow separation in the line. However, the horizontal run to the pipe is about another 15', to a reducer at the suction. We had thought that this long run of pipe would be sufficient to straighten out the flow, however, we can hear what sounds like cavitation in the line about midway to the pump. Our thought was to put some straightening vanes in the line at this point, or maybe a little upstream. We have considered an inducer, however, looking at an adjusted curve reveals the inducer would not lower NPSH much at all (only about 2'). Any thoughts?

Thank you,
Kayla

 

[BobPE]

you didn't give us enough information to help you....fittings are just minor losses and should not be a big deal, but again, without all the infor, we cant help

BobPE

 

[scalleke]

As Bob mentioned: too little information.

What is the min level in the tank as oposed to the pump centerline?
What is the min. pressure in the tank?
As Bob was mentioning the friction losses should be negligeable but what is the diameter of the piping?
What is the flow that you will be pumping?

2' feet of NPSHR improvement is considerable. Might be the difference between cavitationg or not.

I recall an installation where we removed the pump bases to gain 1' of liquid head.

Best regards
Scalleke

 

[25362]

To kryanl, besides the right remarks and comments by scalleke and bobPE, vapour-filled bubbles can form when liquid flowing along a solid boundary separates due to a discontinuity: a cusp, a sharp edge, etc. When the liquid separates from the wall, a vacuum (empty space) is created in between. This causes some of the liquid to evaporate into that empty space and to be carried away as vapour-filled bubbles, which when entering zones of higher pressure cause cavitation by their vigorous collapse.

 

[Parado]

Something to look at to avoid cavitation from air trap & high turbulence. You have a reducer tie to your pump suction, is it a concentric reducer? What is the reduction size?

 

[kryanl]

Hello all, and thanks for the reply,

In answer to your questions:
- The piping as I described is all there is, with the addition of one 2" coupling drain on the horizontal run.
- The minimum level in the tank is 15' above the pump centerline. It can go up to 17 or 18', but for NPSH, I wanted to be conservative. Actually, the level in the tank is only 4', the other 11' is the straight vertical drop of pipe, leading to the elbows I mentioned.
- Normal pressure is atmospheric, but for NPSH calculations, I considered 3" H2O vacuum.
- Pipe diameter is 14", and the reducer is a top-level, 14x10
- Flow is between 1500 gpm and 2500 gpm,(typically, it will be about 2000)

Removing the pump and motor bases isn't really an option, as it is against our standards.

Thanks again for the assistance.

 

[Scipio]

Just to clarify, the problem isn't pump cavitation, but cavitation-like sounds coming from the suction piping between the 45° elbow and the pump suction? That being the case I think both you and 25362 have the right idea. Two elbows, one right after the other, especially a combination that translates the flow through three axis, are very hard on hydraulics. In fact, once you weld two elbows together fitting-to-fitting style you can pretty much forget using tabulated losses for fittings - they're based on laminar flow upstream and downstream, which you no longer have. I think you're probably inducing enough swirl into the flow to create some boundary layer separation in the suction piping and forming some bubbles just downstream of the 45° elbow. Once into the straight piping, the flow is stabilizing enough to collapse these again, producing your cavitation sounds. Straightening vanes downstream of an elbow are to prevent the prerotation from reaching the pump suction and causing problems there, but it doesn't sound like your problem is reaching the pump anyway.

If that's the case, there's two things you can try. The first is changing the piping to elimate the 45° elbow by installing the 90° elbow with a 45° rotation around the vertical axis. The second is replacing one or both elbows with elbows fitted with integral straightening vanes.

 

[lilliput1]

Put in a valved recirculation line from the pump discharge to the pump suction.

 

[kryanl]

Scipio, You brought up a good point that I forgot to clarify. The cavitation noise does start midway in the horizontal piping, but there is also cavitation in the pump. Whether it also exists between these two points, or goes away, we can't tell. I am leaning toward vanes in the straight pipe, but hadn't thought of in the elbows. Is there somewhere where I can find a good design guide on how to 'size' vanes, and determine where to put them?

 

[Scipio]

krynal,

Alright, if you're also cavitating in the pump then it's a bit different and straightening vanes in the last run may do the trick after all.

I must admit I've never seen straightening vanes in process piping elbows, just read books & papers which refer to their use but never seen any manufacturers who actually make them (except for things like air conditioner ductwork), so I can't be of much help there. I would expect they wouldn't be any different than the usual tube-bundle type straightening vanes installed in straight piping runs. All you're looking for is something that will break the secondary flow pattern that develops due to centrifugal force as the liquid swings around the bend, preventing the swirl from developing.

That 14" x 10" eccentric reducer may be causing problems too. An option recommended by Karassic is to look at a concentric reducer that vents back into the suction tank on larger transitions to prevent unstable flow, but if that vent point happens to be at a vena contracta point in the piping it can cause more problems than it solves.