NPSHr (H2O vs Flowing Fluid) and Calculations 2

[dbanks]

My system is as such. I have a Depropanizer overhead reciever that operates at say 240 psig and delivers a pressue driven flow into an aborber tower a sub cooler and a knock out pot then into a sundyne pump. The pump curve shows that the NPSHr is in H2O feet and I have always interpreted feet as feet regardless of the fluid. Is that correct? In a system like this can one simply take the suction pressure and subtract from it the vapor pressure at suction condit for NPSHa (given that the entire system from the overhead accumulator is liquid filled). Is there something that I may be missing with this analysis? Note, the compositon of this stream is about 80%C3 and the balance is C2 and lighter.

 

[pardal]

The pump curve shows that the NPSHr is in H2O feet and I have always interpreted feet as feet regardless of the fluid. Is that correct?

Each liquid have its onw feet .

Water is about 30 feet , but Hg is 30" , it have to deal with specific weigth.

Water 1 kg/liter
Hg 13.5 Kg /liter


HIH Pardal

 

[zman22]

Feet is indeed feet regardless of fluid. However the NPSHr for hydrocarbons can be up to 50% lower than water.

For your NPSHa calculation you also need to take into account elevation and pressure loss in the suction line.

 

[pardal]

ZMAN22

Please apologize me for not agree with you.

Please supouse you have a "u" pipe.
At each U branch you will put Mercury Hg on left branch and water on rigth branch .
As the Hg is not missible in water they will not get mixed.

If you want to make a equilibriun , so not any branch move ,
and the union of the two liquid stay at the middle of the botoon of the U , you will need to put 13,5 feet of water to sustain 1 feet of HG.
So when talking about pressures , feet are not only feet they represent pressures.









Pardal

 

[Kawartha]

dbanks,

You are correct, if you have a reliable pressure reading (including the liquid velocity head) at the pump suction, subtract liquid vapour pressure for NPSHa.

Secondly, I agree with zman22 regarding NPSHr comments. Normally hydrocarbons have a reduced tendency to cavitation, when compared with water. However, it is recommended to use the water based NPSHr numbers when designing the system.

 

[zman22]

Pardal,

When translating feet to pressure the density of the fluid is indeed important (P=rho*g*h). However a centrifugal pump is a contant velocity device that will always deliver the same head regardless of the fluid. The power required to reach the same head will differ with the fluid though.

The NPSHr of a pump is determined by decreasing the suction pressure at constant flow during which the head is constantly measured. The suction pressure at which the head has dropped by 3% expressed in feet (meters) is the NPSHr for that flow.

The fluid used in the above experiment is normally water, so NPSHr is expressed in H2O feet. When the experiment is repeated with a hydrocarbon as fluid the NPSHr is usually lower. For an explanation see the following link:

http://www.pumps.org/public/pump_resources/discussion/NPSH_Standard/app_petroleum.htm

 

[SooCS]

To me the NPSHr, net positive suction head or the net positive suction pressure above vapor pressure that is required to push the liquid into the pump. The pump manufacturers have been using water head as the standard reference. Therefore, density of the liquid has to be taken into account.

Water at standard condition has 34 ft of NPSHa simply because the vapor pressure is almost zero. But for propane especially from the condenser, the vapor pressure in this case is the operating pressure, therefore the NPSHa is zero. That's why you will need to lower the pump below the lowest operating level by the NSPHr which is the static pressure created by the liquid minus the friction pressure lost througth the piping system.

 

[d23]

I hope this "adds" to this discussion.

NPSHr is the "required" pressure or feet to prevent cavitation. What you see on a chart is "typically" based on fresh water at a given temp. If you are using something other than what is specified by a manufacture you should get their input. Cavitation is a pressure / temperature (thermodynamics) calc. NPSHr is trying to prevent vaporization.

I may only need 10 feet for a pump when pumping fresh water. If I use that same pump in a CO2 application I will need a very high intake pressure to prevent vaporization.

Hope this helps!

 

[pumpingwell]

Pumpingwell (Mechanical)
March 4, 2003

Kawartha,

I'm confused about your reply to dbanks: " ...if you have a reliable pressure reading (including the liquid velocity head) at the pump suction, subtract liquid pressure for NPSHa." How can the 'liquid velocity head' be included in a 'pressure reading'?

"In the computation of NPSHa, including the velocity head, Vs^2/2g at the suction nozzle is somewhat illogical because it is not a pressure available to prevent vaporization of the liquid." quoted from Wastewater Engineering: Collection and Pumping of Wastewater, McGraw-Hill, 1981. I found that TD2K (Chemical) shared the same opinion in his reply in the thread407-43686 "NPSH in pipeline boosting pumps".

Hope you or anyone who is inersested in this NPSHa calculation could make it more clear.

 

[25362]

It is simply a mistake to express NPSH in feet of water. NPSH as well as pump heads are all expressed in ft. Not ft of water or ft of liquid, just ft. For each fluid it means ft of that fluid. Static pressures as well as vapour pressures (psi or atmospheres) are converted to "ft" or "m" by taking into account the density of the fluid in question.

 

[Kawartha]

pumpingwell,

Thanks for your comments.

I've seen the formula for calculation of NPSHa - including the velocity head in many publications, and also many without. Since in most low specific speed pumps the velocity head is rather small, I don't think it matters in many cases, if it's included or not.

I didn't write my previous post correctly, but I meant that velocity head should be calculated and added (not included) to the measured inlet pressure.

However, you make an interesting point (which I agree also makes sense) that it should not be included since the velocity head is not pressure available to prevent cavitation. For a conservative calculation, the velocity head should probably be omitted, resulting in a slightly lower NPSHa. In the case of high specific speed, high flow pumps such as C.W. Pumps for power plants, the velocity head can be a significant factor.

I'd like to review more comments and articles regarding this issue before eliminating velocity head across the board. Would you know where I can obtain a copy of the 1981 McGraw-Hill article?

 

[d23]

25362

I believe the reason most people refer to NPSHr (required) in terms of fresh water is because the test is typically conducted using fresh water at a given temp, usually 60 degree F. In most cases (all that I have seen) NPSHr on a pump curve is based on feet of fresh water. In my personal opinion I agree with your "PSI" verses feet. When working with liquids other than fresh water you need to consider vapor point which is pressure V temperature.

pumpingwell and Kawartha:
FYI
Being a pump "speed demon" you are both correct about velocity head. I don't have failed parts to prove it, but I do consider it with NPSH available. No termites may be a good test.

 

[25362]

One note on hydrocarbons in the range of gasoline to naphthas. When these are drawn from distillation towers, as an example, the water NPSHr can be reduced by 10%. Not so when they are taken from open vessels or tanks where they are exposed to air, CO2, etc. These gases are readily absorbed and can create erratic pump behaviour problems. In these cases for a start the NPSHr for water is commonly converted by multiplying by a factor of 2(!), until shown that anything less is satisfactory.

 

[pumpingwell]

Kawartha and d23,

Thanks for your quick response and/or comments.

"Wastewater Engineering: Collection and Pumping of Wastewater" is a book, written and edited by professor George Tchobanoglous.

I'm currently doing an investigation into a raw sewage pump cavitation-like problem. The velocity at suction piping or pump suction nozzle is quite high (4.6m/s currently and even higher while the impeller was new when operating at full speed). The velocity head could be a significant factor for identifying NPSHa may not high enough over NPSHr in this case.

The problem pump is of screw-centrifugal type. There are amount of grit, silt, fats and offal accumulated in the underground concrete suction channel observed when cleaning the wet well. The initial impeller lost a great amount of vane metal close to its discharge/outlet in 2 years service time. The replacement impeller already showed pitting all over its vane, and loss of metal especially oticeable on the pressure side of the vane close to its discharge in 5-6 months. One puzzle of identifying if it's a cavitation problem is no significant vibration/noise being observed up to date for the pump fitted with replacement impeller.

I'm sorry I may have talked too much about the pump problem and beyond the scope of this thread.

 

[eljauzy]

When the pump manufacturer uses feet, it is because they use water to test the pump and NPSHr 3% suppression test.
By rule of thumb you can take the suction pressure and subsract the fluid vapour pressur @ PTempt but the gauge location has to be as closed as possible to the pump suction flange.
To convert the NPSHr H2O to your product NPSHr just multiply by the actual S.G.

Please be aware that NPSHr of an individual pump like Sundyne increases as flow increase but the changes can be relatively small if you operate within +/- 10% of the rated flow.

Do you actually face any cavitation problem or something?

 

[vpl]

This thread has an interesting misconception running through it, one unfortunately caused by America's refusal to give up our beloved English units. NPSH is a pressure measurement and, as such, can indeed be expressed in feet of water ... or in of Hg or lb/in² or ergs or Pascals or atms...

If you have a Crane Flow of Fluids (and if you don't, order one), there is a table towards the back that gives the conversion factors between the various units.

And to answer the original question - if you're not pumping water, you need to take the density of your fluid into account in any calculations. Crane also has a pretty good explanation of this.

For those who disagree on this, may I recommend a basic review of fluid flow...

Patricia Lougheed

Please see FAQ731-376 for tips on how to make the best use of the Eng-Tips Forums.

 

[Scipio]

VPL

For the most part I agree with you, in the piping system outside of the pump. Inside the pump I'd have to attach a couple qualifiers to your statement. First, a quote from one of the pumper's bibles (Cameron Hydraulics)

"In a centrifugal pump the head developed (in feet) is dependent on the velocity of the liquid as it enters the impeller eye and as it leaves the impeller periphery and therefore is independent of the specific gravity of the liquid. The pressure head developed (in psi) will be directly proportional to the specific gravity.

In addition, even while the head developed (as opposed to pressure head) remains the same, power required is also directly proportional to the specific gravity.

As for applying specific gravity to directly factor the vendor's NPSHR in water to determine an effective NPSHR for the pump, I'd have to see proof this could be done before even considering something like that, as it goes against everything I've ever understood about NPSHR corrections.

While NPSHR corrections for lighter liquids exist, mainly hydrocarbons and hot water, based on test data available in a number of texts, including Karassik's Pump Handbook, Cameron Hydraulic Data, and more thoroughly by the Hydraulic Institute, these come with a variety of conditions and disclaimers. Note, this is information derived from actual testing, not basic fluid flow theory.

When calculating NPSHA, on the other hand, I agree completely. Actual fluid SG (or specific weight) has to be used when calculating the difference between the available suction pressure head and the vapour pressure, even though NPSHR is in terms of water, NPSHA can't be calculated as water unless you're actually pumping it and it's cold.

 

[ccfowler]

VPL,

Whether the English system, the Metric system, or any other coherent system of measure is used, the use of pressure to express the pump's characteristics is wholly unnatural (with respect to the pump) and necessitates needless conversion computations.

The wonderfully simple reality is that a particular pump's performance characteristics are all tied together by relationships based on only three parameters: (1) actual shaft rotational speed, (2) volumetric flow rate, and (3) head. The actual measurement units used for these parameters may be any coherent system whether English system, Metric system, or whatever other system one may wish to construct. I freely admit a strong preference for the English system, but we should recognize that pumps really don't know or care what system we happen to use to evaluate their characteristics.

Most conveniently, the pressure drop due to friction related to the volumetric flow rate in the attached piping system can also be expressed in terms of head of the flowing liquid thereby providing a pleasantly direct means of relating the pump's performance characteristics to the piping sytem's characteristics without the need for burdensome measuring unit conversion computations.

Pumps always operate on their curves (head vs. flow, NPSHr vs. flow, power vs. flow--all linked to the impeller's actual rotation speed). For Newtonian fluids, the pressure drops (conveniently expressed in terms of head of the flowing liquid) in the connected piping systems (suction and discharge) vary with the square of the volumetric flow rate.

Why complicate life (and engineering) with needless (and potentially obscuring) conversion computations to express the "natural" unit of head (in whatever system of measure) in the "unnatural" unit of pressure?