Can You Spec A Pump with Negative NPSHA/NPSHR 3

[Odom71]

I'm working on specifying a pump to transfer LPG from storage offsite. The thing that I noticed is that the NPSHA that I'm coming up with is extremly low. Infact, it is negative. I've scanned a specsheet for another pump in similar service and I noticed that it was specified with a NPSHA of 0. The NPSHR under the performance data is -2. I've never heard of negative NPSHA or NPSHR before. I was hoping that someone could advise if it is acceptable to specify negative NPSHA which would in turn require a negative NPSHR. If so, what is the practical limit before I play games like placing the pump below grade etc. to overcome this issue?

Thanks,
Odom71

 

[RotaryG]

Odom,

The definition of NPSHA is simple: Static head + surface pressure head - the vapor pressure of your product - the friction losses in the piping, valves and fittings.

See for reference

http://www.mcnallyinstitute.com/11-html/11-12.html.

For NPSHA to become ZERO:

Static head + surface pressure head= the vapor pressure of your product + the friction losses in the piping, valves and fittings.

Assuming that in the worst case scenario, there is NO externally applied pressure in the suction tank, then

surface pressure head (in the suction tank) = the vapor pressure of your product


In that case:

For NPSHA to become ZERO:

Static head = the friction losses in the piping, valves and fittings.

For NPSHA to become negative

Static head< the friction losses in the piping, valves and fittings.

If Static head< the friction losses in the piping, valves and fittings that indicates that there is NO flow. If there is no flow there is no friction loss.

Hence, NPSHA can never become negative. Instead of going through all the above let us just see what the acronym NPSHA stands for in the first place: &quot;Net POSITIVE Suction Head&quot;.

You are getting negative value since you ignored the physical principle that without DP across a piece of pipe there will not be any friction loss. We all make mistakes. No one is immune. This forum helps us in understanding those so that we may not repeat them.



Regards,

Guru

 

[iainuts]

Guru,
I think you need to step back from the equation to understand what Odom is saying. Note first he's referring to a cryogenic liquid, but that's not terribly important as it applies to any fluid. One generally knows the conditions of a liquid in a tank (ie: pressure and amount of sub cooling or temperature) and one can generally calculate the amount of heat transferred to the product as it passes through the pipeline. One can also calculate the frictional losses in the pipe/fittings and effect of elevation change. So in addition to the parts of the equation you cite, you really need to consider the affect of heat transfer in this case, which changes the vapor pressure of the fluid as it passes down the pipeline.

The result of the fluid pressure dropping below it's vapor pressure is not that flow stops, the result is that the liquid begins to boil. As it does, the vapor fraction of the fluid increases. A vapor fraction of 0 means the fluid is 100% liquid. A vapor fraction of 1 means the fluid is 100% gas. So as the pressure of the fluid continues to decrease below the vapor pressure, the vapor fraction increases from 0 and rises depending on how low the pressure continues to drop. So the meaning of &quot;negative NPSH&quot; is generally understood to mean the fluid pressure has dropped below it's vapor pressure.

For example. If a fluid has a vapor pressure of 100 psig at ambient temperature, and the fluid sits in a tank at 110 psig, we have 10 psig to play with before the fluid starts to boil. If we put a long pipe on this tank and take it over to this pump, and the pressure decays below 100 psig because of flow losses (let's assume the pipe is horizontal so we can neglect elevation changes), then the end result is not that the flow will stop when it gets to 100 psig, the end result is the pressure continues to drop and the vapor fraction increases from 0 and goes up such that the vapor fraction stays in equilibrium with the pressure and temperature. Flow through a pipe is estimated as an isenthalpic process, with the exception of heat flux which changes the enthalpy. But for simplification, neglect heat transfer for the moment. If the enthalpy of the product is known in the tank (we know fluid pressure, and vapor pressure which gives us enthalpy) then we also know the enthalpy at 90 psig (10 psig below the vapor pressure) which gives us the vapor fraction when the pressure in the pipe decays 10 psi below the fluids vapor pressure. Hence the bubbles/gas in the fluid. That, I believe, is what Odom is referring to. Hope that clarifies things.

Take care,
Dave.

 

[iainuts]

Just a clarification to the above, as I don't see any way of editing these posts. Perhaps if there's a way of editing a post, someone can point it out for me.

For the case of a saturated fluid (ie: one at it's vapor pressure), as the fluid pressure drops traveling through the pipe, the vapor pressure (by definition) drops along with it.

I think that's where the confusion arrises. You can talk about NPSH assuming tank conditions and come up with a negative value, but calculating the NPSH at a pump, depending on your interpretation, may result in a value of zero since the two phase fluid at the suction of the pump is in fact saturated at it's vapor pressure. But this interpretation, IMHO, is very misleading. I've never liked the concept of NPSH, especially as it applies to cryogenic pumps, and perhaps this discussion helps to explain why. In the past, I've often wanted to take the acronym over to the cement wall out back and put a few bullets through it! Argh…

 

[25362]

iainuts, you got me totally confused when you say: &quot;...as the fluid pressure drops travelling through the pipe, the vapour pressure (by definition) drops along with it&quot;.

Do you mean to say the VP increases due to friction and heat, and there is some vaporization all along ?

There is no reason to get angry at a concept that was, and still is, very useful in the design and application of pumps in general, and centrifugal, in particular. Right ? Would you prefer changing it, for example, to MSCPC (Minimum Suction Conditions to Prevent Cavitation), thus deleting the word &quot;positive&quot; ?

 

[Scipio]

A similar liquid to LNG that's caused me headaches when it comes to NPSH calcs is liquid hydrocarbon condensates separated out of natural gas streams. Just try to make a sensible calculation with a room-temperature liquid at two or three hundred psi, that's 98% C5+, but is saturated with dissolved methane & ethane gas. I've never been able to figure out how to make a sensible estimation of vapour pressure with multiple-fraction liquids, so typically just to make life bearable assume vapour pressure = surface pressure. Not quite right, but I figure close enough. Always wondered though, if methane & ethane fractions bubble out of solution at the pump inlet because of reduced pressure (line losses, etc.), will they actually implode again causing cavitation? My experience, at least in recips, is it tends to stay out of solution and vapour lock the pump, but I've never been 100% sure on the mechanics.

Now, you really want to see a value for NPSHA (or NPIP, Net Positive Inlet Pressure, for the recip purists) that'll screw with your head? Take into account the acceleration head of a single-acting simplex plunger pump. I've got one vapour locking right now (until I get the stabilizers fitted) because vapour pressure assumed = surface pressure, static head is about 15 feet, and without even considering friction losses I've got a calculated acceleration head in excess of 50 feet

 

[iainuts]

Hi 25362. Got a laugh out of that response about one shouldn't get mad at a concept! That's a chuckle. Thanks, you got me smiling here.

Regarding &quot;...as the fluid pressure drops travelling through the pipe, the vapour pressure (by definition) drops along with it&quot;. Let's say the fluid is at it's vapor pressure at some point in the pipe as 100% liquid. This is the same as saying it's a saturated liquid. It's on the verge of boiling. Any additional heat will force it to boil. Similarly, any reduction in pressure will also force it to boil

If we reduce pressure on the fluid, the liquid boils and the 2 phase stuff that remains is still saturated somewhere below the dome. ie: if the fluid is at it's saturation point, then regardless of what quality or vapor fraction the fluid has, it is still in equilibrium at it's vapor pressure. Thus as the pressure drops, so does (by definition) it's vapor pressure.

In the case of a liquid flowing through a pipe with pressure dropping as it goes, this process is governed by the 1'st law of thermodynamics. One also needs to add the affects of elevation change and velocity change. This is all 1'st law energy balance with the difficulty being calculating pressure drop due to frictional flow of a fluid who's composition keeps changing (vapor fraction continues to increase as pressure drops). This all makes for a relatively difficult 2 phase flow through a pipe problem that I prefer to leave to a computer program.

Hope that helps.
Dave.

 

[iainuts]

Yet another clarification (wish I could edit posts here!)

The above statements I made are only true for a pure component, ie: pure LNG, pure liquid hydrogen, pure liquid nitrogen, etc... I'll differ to a ChemE to answer the question proposed by Scipio where the fluid is a mixture of multiple components.

Thanks,
Dave.

 

[25362]

Dissolved gases have a half-life of evolution shorter than for solution. That's one reason for their incomplete collapse. The remaining bubbles are recirculated to the impeller inlet passages to join with freshly-cavitated bubbles. This leads to an accumulation of gas blocking part of the impeller inlet.

Many pumps, especialy those working on level control, operate with flow rates well below their BEPs. These flow reductions promote recirculation.

Besides, an impeller can act as a centrifugal separator, evolving gas and holding it near the eye periphery. The accumulated gas may even start backing up into the top of the horizontal section of the suction pipe.

Venting the pump at the top of the volute while running is usually useless. Venting near the suction nozzle from its top, may do some good while running to avoid vapour lock. Anyhow, the normal procedure is to stop the pump and vent it.

In order to estimate the &quot;right&quot; NPSHa for cases like those mentioned by Scipio, I think one could get help from an article in Chem. Engineering of July 26, 1982 by Mao J. Tsai, titled Accounting for dissolved gases in pump design

 

[16347]

I'm not a mechanic engineer, but an operation engineer and I was wondering if someone considered that,even when NPSHa<NPSHr, at the pump suction you cannot vaporize all the liquid even if it is LPG . Simply, because there is not enough energy to vaporize it .Once reached the liquid boiling temp at given suction pressure, it will no more vaporize so you can have negative suction head and still have a lot of liquid and a bi-phasic flow in the pump eye. At that low temp, the gas bubbles will easily collapse in the impeller causing cavitation.

I didn't find a detail explanation of the dynamics of this process, but I believe that the vaporization enthalpy is reconverted thereafter, when the gas is re-liquefied but this happens somewhere after impeller, so that this energy is transferred only partially, by conductance, back to the suction nozzle, providing some additional energy for vaporizing...

 

[PUMPDESIGNER]

16347,
You bring up a valid point.
As with many natural processes, a momentary equilibrium is achieved. Some of the liquid goes gaseous, which momentarily and in small localities raises pressure back above vapor pressure so that not all liquid is phase changed.

Pressure varies throughout the liquid column, which means that not all liquid &quot;sees&quot; vapor pressure. Once an area falls below vapor pressure, some fluid in that area vaporizes, which then raises pressure in that area and thereby prevents furhter phase changing, but only in that small area.

PUMPDESIGNER

 

[25362]

A visit to

http://www.cheresources.com/centrifugalpumps2b.shtml

may be of help.

 

[smckennz]

This thread is unusual.Where is the definition of NPSHR?
I thought it was the point where there is a 2 or 3% drop in TDH (I forget which) over the expected value if the fluid pumped remains liquid.

The original question said the NPSHR was -2.

Would someone be so kind as to refer me to the pump manufacturer who provides pumps with negative NPSHR ratings?
After all the question is about specifying.

Cheers

Steve McKenzie

 

[sb00]

just wanna add that in the refinery i worked for, we used to store LPG in horton spheres at around 12 barg. It avoids two phase flow thru pumps and keeping LPG at high pressure also minimizes loss due to vaporization.

 

[25362]

To sb00, one is inclined to think that the 12 barg were cancelled out by the vapour pressure of the LPG (=12 barg) thus leaving your NPSHA equal to the elevation difference less the friction drop in the suction lines. Am I right?