NPSH vs. Pressure

[kryanl]

Hi all,
I am having trouble trying to distinguish the difference between NPSH and pressure. I know that NPSHr is what you must have in order for the pump to not cavitate.
It is equal to: (static height) + (pressure within vessel) - (vapor pressure) - (pipe friction losses).
For example: Let's say, in simplistic terms that:

There is a tank, 10' below which is the pump centerline. The tank is at atmospheric pressure, full of water at 80 degrees. And there is 1' friction loss in the suction piping.
NPSHa = 10' + 34' (atm. Pressure) - 1.16' (vap. pressure of water,abs. in feet) - 1' (friction)
Therefore, NPSHa = approx. 42'
Now, for arguments sake, let's say the tank is under a vacuum, say 3" Hg abs. (=1.47 psia = 3.4'). Because of the vacuum, the vapor pressure of the water will equal the vacuum, because the temperature of the water will lower.
NPSHa = 10' + 3.4' (the vacuum, when considered in absolute, and in feet) - 3.4' (vap. Pressure of water, abs. In feet) - 1' (friction)
Therefore, NPSHa = approx. 9'

Now, comes the troubling part. I must figure out what the pressure is at the suction of the pump for each case. I know that NPSH is not the actual pressure, but the more I think about it, I can not see how it isn't. Just multiply each NPSHa by 2.31 and you get the pressure at the pump inlet. Of course, this would be the pressure if the pump were off. I know there is a vacuum when the pump is running, isn't there?

Along the same lines, when your pump is rated for "50' discharge head", that means the pump is pushing the water at a pressure of 115.5 psi. Is that absolute pressure, or gage pressure? And, so, what exactly is the pressure
differential across the pump?

I used to know this stuff, but one of our new instrumentation people was asking me about it, and started confusing the heck out of me with her questions. I am relatively new myself, so am not used to answering
questions.

On an a different pumping note, I overheard two other engineers arguing over the calculation of static discharge head. They were discussing a pump which will be at elevation of 1150' pumping to a final destination of 1165'.
However, at some point, the line gets onto a bridge, which is at 1182'. One of the engineers argued that the static head was 1182' - 1150' = 32', which I agreed with. You have to be sure that the fluid can get to that height, and 'fall' down to the tank, to be on the safe side, especially during start-up. However, the other engineer said that the head was only 1165' - 1150 = 15'. Which is it? This would affect many pumps which the three of us have sized over the past couple of weeks.


Thank you for any help
Kayla

 

[TD2K]

Best thing with working this out is to always have a sketch.

For the second case, where the tank is under a vacuum, the pressure at the surface of the water is 3" Hg abs. You then have 10' of head on top of that minus the frictional losses, that gives you the suction pressure. The difference is that you don't subract the vapor pressure when calculating the suction pressure (but you do, as you did, when calculating NPSHR). Personally, I prefer working out NPSHR in feet as you do but work out the suction pressure in psig or psiA. You don't have to but I've found it confusing otherwise.

NPSHR is only equal to the suction pressure in very specific cases (eg. blind luck). If you look at the case where the pump is taking suction of a 100 psig bubblepoint liquid with no static head of liquid and no line losses, the NPHSR would be zero but the suction pressure would be 100 psig.

 

[TD2K]

Sorry, missed the other questions.

A centrifugal pump always produces the same head irregardless of the fluid's density. The head by a centrifugal is always the differential head across the pump. For your case, if the suction is at atmospheric pressure, the discharge pressure would be 0 psig + 50'. If the suction is at 100 psig, the discharge pressure would be 100 psig + 50'. The differential head is converted to a differential pressure as Head = 2.31*dP/sp.gr. (50' head of head on water will be 21.6 psi, you have the equation backwards).

For the last question, assuming the piping is totally full of liquid, the intermediate height does not affect the pump head calculation and 15' is the right answer. Although there is a pressure 'head' to get to that elevation, you get it all back when it falls back down. There are a couple of cautions here. First, though it's not at all likely in this case, check to make sure that the pump on startup can get the fluid (at a lower flowrate) all the way up the high point. Secondly, if the elevation is high enough or the fluid is close enough to its bubblepoint, you can have vapor breaking out at the high point. Once you have vapor and liquid, you don't get the pressure back when you come 'back down' AND because you have 2 phase flow, line losses can increase dramatically.

 

[MortenA]

I dont know if this is your question but: Pressure is pressure.

I assume its because a lot of mech. eng. have great difficulties understanding ABSOLUTE pressure (they always think in gauge) then the term Net Positive Suction Head came into existance.

Oh yes: and then its head instad of eg. bar or psi because the centrifugal pump has the same head where cavitation start no matter what the desity of the liquid is (assuming equal vapour pressure). This then makes it easier to use head instad of pressure.

Best Regrads

Morten

 

[tstead]

NPSHa is the absolute pressure of the fluid ABOVE the vapor pressure converted to feet of water.

Simple enough, eh?

As far as the hill is concerned, the Bernoulli equation is based on what happens between two points and any elevation changes between those two points is irrelevant. In other words, you can go up, go down, go back up, go sideways, etc. what matters is the differential height between the start and end points.

Now, if you go up very high and come back down, then the pump might not be able to get up that hill. You will have one system curve starting up and then another when it is full when the siphon effect kicks in. If you have open channel flow/partially-filled pipe then all bets are off.

Tim