NPSHA when taking in account the velocity head 1

[PalomaP]

Could you please, let me know how to calculate the NPSHA if I wish to take in account the velocity head in the suction tank?

- For suction lift:
NPSHA= ha - hvpa - hst - hfs -v2/2g

- For positive (flooded) suction:
NPSHA= ha - hvpa + hst - hfs +v2/2g

Are those calculations right? (+ or - v2/2g)? In which case should I consider velocity head different than zero?

Thanks a lot

 

[katmar]

You treat the velocity head the same, whether it is a suction lift or a flooded suction. In both cases pressure energy is consumed in accelerating the liquid into the pipe. However, the critical concept to grasp is that the velocity head is recovered as pressure when the liquid flows into the eye of the impeller.

So the short answer is that you can ignore the velocity head in both cases.

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[PalomaP]

However, I have just read in "Cameron Hydraulic Data", to calculate he NSPHA on an existing installation would be the reading of a gage at the suction flange converte to feet of liquid absolute and correcte to the pump centerline elevation less the vappor pressure plus the velocity head in feet of liquid at point of gage attachment.
Why do you think in this case is taking in account the velocity head?

thanks a lot

 

[katmar]

The Cameron book is saying the same thing I am. The pressure gauge reads the actual pressure at the suction. But the velocity head is lost at the start of the pipe where the acceleration takes place. So at the point where you are measuring the pressure the velocity head has already been lost. To calculate the NPSHa you then have to add it back in. By physically losing the velocity head and then adding it back in you are doing just what I said - ignoring it.

The situation you describe is where you have an actual installation you can go out and measure. If you were working off drawings and you wanted to calculate the NPSHa you could calculate the velocity head and deduct it along with the other losses (which would give you the pressure a gauge would see), and then add it back in as Cameron advise. I'm a lazy guy, so rather than deduct it and then add it back in I just ignore it.

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[BigInch]

So the real question is,

NPSHa = Total Head - vp, or Static Head only - vp?

NPSHa is Total Head- vp.

Many engineers simply neglect velocity head to arrive at a conservative calculation, however [COLOR=red white]IF[/color] velocity head was neglected in the calculation of NPSHa (only static head and fitting losses were considered), velocity head at the suction flange can/could be added into that result. Adding velocity head would give a less conservative result for the NPSHa calculation, but its theoretically valid to do so and you can include it, if you choose to do so.

Total head (HGL) at any point along the pipe is the sum of static head + velocity head. NPSHa is the Total Head at the suction flange - vapor pressure, so you don't subtract v^2/2g at the tank and then add it back in at the suction flange. Since NPSHa is really Total Head at the suction flange -vapor pressure, when calculating NPSHa from the tank level you can leave out all consideration of the velocity head term and subtract only pipe and fitting losses from the initial water level.

[/url]

NPSHa = s + v^2/2/g - vapor pressure

http://virtualpipeline.spaces.msn.com

 

[PalomaP]

BigInch thanks a lot for your reply. I am still a little confuse...Could I summarize as follow?

1) For a new installation, the NPSHA in the pump is:
- For suction lift:
NPSHA= (ha - hst - hfs)-v2/2g - hvp

- For positive (flooded) suction:
NPSHA= (ha + hst - hfs) -hvp (more conservative since I dont take v2/2g in account)

2) For an existing installation:
- For suction lift:
NPSHA= Pressure gage - v2/2g - hvp

- For positive (flooded) suction:
NPSHA = Pressure gage - hvp (more conservative without v2/2g)

Did I understand this right? This issue with the velocity head is giving me hard time...

 

[25362]

V[sup]2[/sup]/2g is always a component of the total head (expressed in meters or feet of liquid), thus it should be always added.

However, as explained by the experts, its value can be neglected in most cases. See, please, the following values of v[sup]2[/sup]/2g for three different linear velocities;

for 3 fps, v[sup]2[/sup]/2g [≈] 0.141 ft of liquid
for 2 fps, v[sup]2[/sup]/2g [≈] 0.062 ft of liquid
for 1 fps, v[sup]2[/sup]/2g [≈] 0.016 ft of liquid

 

[Artisi]

If the v2/2g is a real issue in the calculation of NPSHa then the NPSHa must be much too close to NPSHr, this being the case other aspects of the installation would require further investigation to increase the margin.

 

[BigInch]

NPSHa Calculation (new or existing system, open or closed tank, including velocity head)

Here's a revised diagram with the NPSHa equation,

This equation adds the v^2/2/g term, however, as Artisi suggests the velocity head is seldom of any significance and most engineers safely ignore this term. I always ignore it myself, unless doing an accurate study and I feel I must incude it to accurately document the model.

I would also suggest that if you want to make such a model to a very high accuracy standard, and you want to include the velocity head, you should also consider other possible terms that are typically neglected, such as subtracting the distance from the suction flange to the top of the pump's impeller, for example, and to be sure to use the vapor pressure of the fluid at the temperature at which the fluid enters the pump, which would therefore also require some thermal considerations.

http://virtualpipeline.spaces.msn.com

 

[katmar]

I agree with 25362 and Artisi that in normal applications the velocity head is very small and can be disregarded completely. However, BigInch has drawn us such a great diagram I believe it is worth splitting a few hairs and getting the understanding correct. I would also like to ask a question to further my own understanding.

BigInch's diagram highlights the v[sup]2[/sup]/2g situation very nicely by showing that some pressure head is lost to velocity head as the liquid is accelerated into the pipe, but also that some velocity head is converted back into pressure head as the liquid slows down in the reducer (the green section gets narrower). This begs the question - if you are going to add the velocity head back in, which velocity do you use? At the start of the pipe or at the pump suction? If you were using BigInch's formula to do a rigorous calculation you would have to take the loss of pressure head to static head at the start of the pipe, plus the recovery of velocity head to pressure head at the reducer, into the HL (Pipe Friction) term. And then you would add back in the velocity head based on the suction size to get NPSHa. All this converting back and forth comes to a net zero in the end, so it is easiest to simply omit the velocity head from the friction loss calculation and then don't add it back in again at the end.

The question I would like to raise concerns the distance between the centerline of the suction and the top of the impeller. As far as I can recall, all the pump curves I have used have specified the NPSHr at the centerline of the suction. Why would you concern yourself with the distance to the top of the impeller? Surely NPSHr and NPSHa should be relative to the same elevation?

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[Artisi]

The height from suction flange to top of impeller (Hp) seems to be an orphan as it does not appear anywhere in the text or the formula and wouldn't be of use in any NPSHa cals' anyway. Possibly it is refered to in further discusion for gauge height correction or some such use.

 

[BigInch]

As katmar concludes, its not convenient to carry velocity head through a progressive calculation, and no need to do it either. In fact, when one ignores velocity head, its really equivalent to setting the frame of reference to HGL by default, meaning that velocity head should NEVER be added back into the calculations (unless it was subtracted by mistake somewhere else).

If the NPSHr curves are referenced to the CL, then there would be no reason to include the height to the top of the impeller. I doubt if I can find the reference for including that, as it has been at least 10 years since I first recall seeing it, however I'll make a few test trenches to look for it.

[COLOR=white red]So, when ignoring velocity head, HGL is set by default and I would have to revise the formulas below the diagram above. We must strike the v^2/2/g term completely.[/color]

It seems Mother Nature already had this figured out, since v^2 never results in a negative value!

http://virtualpipeline.spaces.msn.com

 

[BigInch]

Feeling guilty not cleaning this up, so...

and Leaving in Hp, to cover the case where NPSHr curve reference might be unknown(?). Or strike that too if you like.

Rev 2. HGL reference.
Including pump and discharge line to tank.

http://virtualpipeline.spaces.msn.com

 

[25362]

I think NPSH is one of the most discussed subjects in all forums. See, for example:

thread407-163007
thread407-171080
thread378-175054

and the references therein.

 

[sailoday28]

BigInch (Petroleum) 15 May 07 4:01
With reference to the above submittal
NPSHA= (Pt-Pv)(rho*g) + Hs - Hl (1)

Note: velocity head is already accounted for.

The above is based upon definition of NPSHA=
(P-Pv)/(rho*g) + V^2/(2g) (2)


But Pt/(rho*g)+ Hs = P/(rho*g) + V^2/(2g) +Hl (3)
Combine (2) and (3) to yield (1)

Hs is indicated in your first illustration.

Regards

 

[BigInch]

Yes, that's exactly why I proposed to totally ignore velocity head in NPSH calculations (done from a free liquid surface). It is automatically included.

If you started from some node where you only knew velocity and pressure, you could calculate the HGL at that point to begin a sequential NPSH calculation, or as suggested previously, v^2/2/g is normally so small it is not likely to have any significance and it would be conservative to ignore it anyway.

http://virtualpipeline.spaces.msn.com