NPSH and maximum vaccum of the pump 1

[adrich91]

Good afternoon,

I would like to ask you a question about the NPSHr of a pump and the NPSHd.

Let's imagine a pump that has a NPSHr of 1 m. This means that the difference between the pressure at the inlet of the pump and the vapour tension needs to be at least 1 m so that it does not cavitate. Ideally, it is designed so that NPSHd is greater than NPSHr.

However, if we assume the attached drawing, in which the tank is at atmospheric pressure and the difference in height between the pump axis and the level is 40 cm and assuming that the head loss in the entire suction line is 500 mbar, that the fluid has a density of 1 g/cm3 and a very low vapour pressure at that temperature (let us say 10 mbar) and that the NPSHr of the pump is 2.8 m, the calculation of the NPSHd would give as a result (approximate):

P atm = 1 bar --> 10 m
Pvap = 10 mbar --> 0.1 m
z = 40 cm = 0.4 m
F = 500 mbar --> 5 m

NPSHd = 10 - 0.4 - 5 - 0.1 = 4.5 m

As NPSHd > NPSHr, it will not cavitate, but my question is: Now I have to see what will motivate the fluid to move towards the pump inlet. The pump will create a vacuum that will cause the fluid to move, but.... What if the pump is not able to overcome the head loss despite having a NPSHd > NPSHr? How is the NPSHd and/or NPSHr and the maximum suction capacity of the pump linked?



I would like to analyse this situation and find out what I should look for when buying a pump to know if it will be able to overcome the friction and the height and be able to move the fluid to its suction.

Thank you!

Regards

 

[LittleInch]

Jeez, there are a few things here to unpack.

1) What is NPSHd? Nearly everyone I know uses NPSHA - Net pump suction head AVAILABLE. This is by convention always measured at the centreline of the pump inlet flange. Often the centre of the pump but not always.

2 "This means that the difference between the pressure at the inlet of the pump and the vapour tension needs to be at least 1 m so that it does not cavitate." NO. The NPSH figure / curve you get from a pump vendor is the point at which the differential head drops by 3%. It can be cavitating like mad, but still not lose performance by 3%. I've been there and done that and trust me - you need to add at least 1m to the NPSH curve to prevent cavitation and maybe more than that. The slower the impellor and the closer to BEP the lower this margin needs to be. Why 3% TDH drop to define NPSH? No idea, but that's what the pump codes say and that's what everyone uses.

This graph gives you an idea. The NPSHR 0% line is essentially the cavitation line. The figure and graph you get from a pump vendor is the NPSHR 3% TDH line.

The difference at low flow or high flow can be several metres head.



3) "What if the pump is not able to overcome the head loss despite having a NPSHd > NPSHr?" Then you've calculated it wrong.

4) "How is the NPSHd and/or NPSHr and the maximum suction capacity of the pump linked?" Well they are close to the same thing. Note that the pump is assumed here to be centrifugal and will only lift a fluid if it has a fully flooded suction before it starts or during operation. Most times this doesn't work is when the pump and inlet line are not primed and full of liquid when you start it. Nothing to do with NPSH.

The atmospheric pressure or the pressure above the liquid ( if the tank is pressurised or is below atmospheric pressure) is what gives the liquid the energy to be lifted into the pump inlet.

Does that make sense?

 

[pierreick]

Hi,
Are you a student?
Try to learn from the document attached.
To me your query makes no sense. Have you ever operated a pump?
Pierre

 

[innovativepumps]

Good afternoon,

I would like to ask you a question about the NPSHr of a pump and the NPSHd.

Let's imagine a pump that has a NPSHr of 1 m. This means that the difference between the pressure at the inlet of the pump and the vapour tension needs to be at least 1 m so that it does not cavitate. Ideally, it is designed so that NPSHd is greater than NPSHr.

However, if we assume the attached drawing, in which the tank is at atmospheric pressure and the difference in height between the pump axis and the level is 40 cm and assuming that the head loss in the entire suction line is 500 mbar, that the fluid has a density of 1 g/cm3 and a very low vapour pressure at that temperature (let us say 10 mbar) and that the NPSHr of the pump is 2.8 m, the calculation of the NPSHd would give as a result (approximate):

P atm = 1 bar --> 10 m
Pvap = 10 mbar --> 0.1 m
z = 40 cm = 0.4 m
F = 500 mbar --> 5 m

NPSHd = 10 - 0.4 - 5 - 0.1 = 4.5 m

As NPSHd > NPSHr, it will not cavitate, but my question is: Now I have to see what will motivate the fluid to move towards the pump inlet. The pump will create a vacuum that will cause the fluid to move, but.... What if the pump is not able to overcome the head loss despite having a NPSHd > NPSHr? How is the NPSHd and/or NPSHr and the maximum suction capacity of the pump linked?

View attachment 9455

I would like to analyse this situation and find out what I should look for when buying a pump to know if it will be able to overcome the friction and the height and be able to move the fluid to its suction.

Thank you!

Regards

Hi @adrich91, The NPSHr is different for each pump design and you did not advise if you were talking about a centrifugal or positive displacement pump; the difference in pumping action will also determine the effect cavitation has. My expertise is with the sealless positive displacement Hydra-Cell pumps. The suction stroke does essentially create a vacuum however if you use the NPSHr curves we provide or our excel based calculator, that aspect is baked into the NPSHr figures. The bottom line is you want NPSHa > NPSHr by some safety margin as its not all theory. There are line loses, turbulence and temperature fluctuations to consider and the accuracy of the instruments from which you obtain that data. I think you would be best served on a forum like this if you provided the design criteria and whatever limitations you might have (using an existing pump, pipeline size, etc...) and letting others chime in with comments. I wrote an article touching on the NPSHa calculation as pertains to my specific type of pump, but nonetheless it has all the applicable components of the NPSHa formula: https://innovativepumps.com/HYDRA-CELL/HYDRA-CELL-Installation-Guidelines.html, maybe this is helpful to you.

 

[Snickster]

As NPSHd > NPSHr, it will not cavitate, but my question is: Now I have to see what will motivate the fluid to move towards the pump inlet. The pump will create a vacuum that will cause the fluid to move, but.... What if the pump is not able to overcome the head loss despite having a NPSHd > NPSHr? How is the NPSHd and/or NPSHr and the maximum suction capacity of the pump linked?

I would like to analyse this situation and find out what I should look for when buying a pump to know if it will be able to overcome the friction and the height and be able to move the fluid to its suction.
_______________________________________________________

I think I know what you are asking, let me see if I can answer this and make sense.

The centrifugal pump imparts kinetic (velocity) energy into the fluid via centrifugal force created by rotation of the impeller. No matter what the inlet pressure to the impeller is, it imparts a velocity to the fluid to produce kinetic energy which is then converted to pressure head in the volute discharge. The pressure head causes the fluid to flow in the discharge pipe due to now having pressure head available to overcome friction in the piping. If you add the imparted kinetic energy of the pump to the suction head (pressure plus velocity) you get the discharge head/pressure plus velocity.

The only way you can get flow is if the pressure on the discharge is positive and above terminal pressure of the pipe to have some pressure available for friction loss. So if you have an open end discharge pipe where terminal pressure is atmospheric 0 psig at the end of the line, then the pump needs to produce at least some positive pressure to be able to overcome friction loss, the amount of positive pressure required is based on the actual flowrate desired. The higher the flowrate the more differential pressure you need. However even at very low positive pressure it still will have flow but very low.

So say you have 5 psia (absolute) pressure on the inlet of the pump with pump discharging to atmosphere. The 5 psia inlet pressure could be developed by having a suction tank at atmospheric pressure 14.7 psia (not considering static liquid level) and with pressure loss between the tank and the pump suction under flow, or it could be a pump that basically takes suction out of a close-coupled tank with tank pressure at 5 psia. The pump don't know the difference. In order for flow to exist the pump must produce enough differential head so that is at least some positive pressure on the discharge. To get positive pressure on the discharge the pump must develop at least over 14.7-5 = 9.7 psi differential pressure (= 22.4 feet of head) or flow will not occur.

If the pressure on the suction was 0 psia (absolute vacuum) the pump must able to produce above 14.7 psi of head (=33.9 feet) to get any positive discharge pressure and any flow. However in this case the suction pressure would be below the vaporization pressure of the water so you would get severe cavitation.

 

[georgeverghese]

As shown, this pump will work as intended only if the suction line is fully primed with liquid. Only then will it be able to lift this liquid from 40cm below. Some types of pumps in this service have auxiliary positive displacement pump that operate for a short while on startup to pull this liquid up the suction line, while others rely on a submerged foot valve in the suction line.

The other alternative is to use a fully submerged sump pump.

Each of these options has its pros and cons - talk to your Operations engineer for more details. Get Operations input and approval for the option selected.

 

[pierreick]

Hi,
To add to my previous reply 2 documents, one about centrifugal pump and one about immersed pump (suggested by George).
As an engineer you must provide relevant sketch with all the necessary pieces of equipment and meaningful calculations.
Note: I guess d stands for disponible in French, in English it's a for available.
Pierre

 

[seuenergy]

first impression is that there are two opitions for underground pit pump.
option 1: self priming pump
option 2: manual priming before start pump
detials please see the link:
https://chemicaltweak.com/pump-priming-in-detail/

 

[LittleInch]

Well the OP hasn't been seen since he or she posted the question on Tuesday so we're all shouting into the wind here....

 

[adrich91]

Hello again,

Sorry for the delay, I was travelling.

The pump is lobe, i.e. positive displacement.
Yes, when I say NPSHd I mean available (NPSHa).

I did the calculations to make an example to ask my doubts. It´s not a real problem so i cant give you more details. I was wondering if it could be the case of having a high NPSHa but a lot of head loss that makes the pump not able to make enough vacuum.

Imagine NPSHa is 4.5 m and is enough.

Is the pump capable of move the fluid since it´s necessary a vaccum of more than 500 mbar? I am talking about the issue to move the fluid, not to cavitate.

If i understood u all, if a pump has a lower NPSHrequired, it means it is capable of doing higher vacuum, right? That would be the concept I am chasing

Sorry if i dont explain clearly.

Thanks in advance


Edit: With this kind of positive pumps, is it necessary to install a foot valve (In my scenario, at least)?

 

[Snickster]

Are you asking if the PD pump can pump out the air in the suction pipe to create a vacuum in order to get the liquid in the tank to move from the tank to the pump? Different PD pumps can pump different suction lifts with air in suction pipe. It depends on their construction. I think piston pumps are the best for this since there is no way the air can reverse flow while it is being pumped out of the suction to fill the pipe with liquid. With rotary lobe pumps the fluid pumped out of the chambers between lobes can reverse flow around the periphery of the lobes between the lobes and the casing when the suction to discharge differential pressure gets high enough so then air flow will cease and so will lift of the liquid. Consult with pump manufacturer for maximum suction lift capabilities of pump.

 

[georgeverghese]

Google says typical suction lift for a gear / lobe pump is about 8m, and you wont need a foot valve in the suction line. Ask the pump vendor if the pump can dry for say a minute or so until the pump casing is filled with liquid.

 

[LittleInch]

It's probably translation, but when you talk about "head loss", this normally refers to friction of a flowing fluid.

What you seem to be talking about is head lift or self priming.

This is a temporary issue when starting the pump and is very pump dependant.

Some will allow the pump to run dry for a short period and some won't. You risk destroying the seals or overheating the pump before liquid enters the pump.

NPSHR is a running flooded condition. So many centrifugal pumps have low NPSHR numbers, but can't generate a sufficient lower air pressure to lift the liquid into the pump.

It might not be necessary to install a foot valve, but as noted above, your particualr pump may not be capable of dry running for more than a few seconds so might not be able to create the lower pressure required to lift the liquid into the pump. So such a valve is always a good idea.

 

[adrich91]

Hello again,

I don't quite understand this, because on one occasion, a manufacturer of self-priming pumps told me that his pumps had an NPSHr of 0.25 m! but that the maximum head they could lift was 4 m on the suction side .

So I don't understand anything, given that with that NPSHr value, I could theoretically make an almost perfect vacuum! And on the contrary, it is telling me that I cannot have more than 4 m (0.4 bar) of height difference, that is to say, the maximum vacuum would be about - 0.4 bar g, when in theory according to that value of NPSHr I could reach -0.8 bar g more or less.

Can anyone tell me why in this case, the NPSHr is 0.25 m but the maximum suction height is (according to the manufacturer) 4 m?

Thanks

 

[LittleInch]

Because the NPSHr value is when you're pumping a liquid.

Suction lift is when the same pump is trying to pump air. Different fluids.

Small leakage of a liquid from one side to the other, not a problem.

Same leakage of air, big problem.

Did that help?

 

[seuenergy]

one picture may help you to understand.

this one is mannual priming vessel. The pump is one normal centrifuge pump. When priming liquid is pump out, then little vacuum would be created. Then the water inside underground pit will be pushed by ambient pressure.

For self-priming pump, please see my previous link, it is easy to understand its working principle when you understand the mannual priming pump.

 

[LittleInch]

one picture may help you to understand.

this one is mannual priming vessel. The pump is one normal centrifuge pump. When priming liquid is pump out, then little vacuum would be created. Then the water inside underground pit will be pushed by ambient pressure.

For self-priming pump, please see my previous link, it is easy to understand its working principle when you understand the mannual priming pump.

View attachment 9806

TBH that's probably confusing things. I've used vessels like that to allow centrifugal pumps to self prime and create a lower than atmospheric pressure in that vessel which can probably lift 4 to 5m. The OP seems to have a lobe pump or other PD pump which works differently.

Basically it's all about the seals on the rotating or sliding elements. They are designed for liquids which seal, lubricate and cool the seals. The leakage rate is probably only a few % from high pressure side to low pressure. They are not designed for air / less than atmospheric pressure air which will flow a lot easier that liquid past the seals, and does not cool or lubricate them, in fact the opposite. The leakage rate from higher pressure side to low side when in gas mode reaches 100% at some relatively low value. In the case being mentioned this value would appear to be about 0.4 bar differential pressure.

If you want to create a vacuum or low pressure like say 100mbara you need pumps designed specifically to do this like liquid ring or similar.

 

[pierreick]

Hi,
For priming, steam ejector is a good option.
I don't understand LI's reply about the pump's choice, the OP did not mention PD.
Here we are guessing. Should be a student.
Pierre

 

[LittleInch]

Hi,
For priming, steam ejector is a good option.
I don't understand LI's reply about the pump's choice, the OP did not mention PD.
Here we are guessing. Should be a student.
Pierre

post #10 - "The pump is lobe, i.e. positive displacement."

 

[pierreick]

post #10 - "The pump is lobe, i.e. positive displacement."

Should have been written in the first post. I'm out.
Pierre