I don't think there's any solid basis in specifically 1m, just ensuring that your NPSHA > NPSHR by some margin helps minimize cavitation and potential issues.

The question of NPSHa margin involves your skills as an experienced mechanical engineer. !!!!!.

https://pumpsdesign.com/npsh-margin-at-rated-flow/

My recollections are, when you get into NPSH margin specification, you must consider the specific pump service... No rule of thumb is suitable for all cases !!

Some particularly difficult pump services are:

- Cooling Tower Water Pumps (air entrainment and probable vortexing)

- Boiler feed water pumps (high energy pump, commonly with extreme convoluted suction conditions)

- Any, multi-staged pump that cycles frequently

This website is a good reference ........https://empoweringpumps.com/npsh-net-positive-suct...

also there is a fine discussion of this topic in the eternal "Pump Handbook"

MJCronin
Sr. Process Engineer

I think 1m is as said a general ROT to avoid cavitation which occurs at a higher point than the NPSH curve which is based on 3% head differential drop.

I don't know where this graph came from but they recommend 2 x the NPSH at the duty point. That would be my recommendation

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

Things that can mess up the NPSHa to a pump: These relate to liquid streams that are operating very close to or at bubble point
a)Poor degassing at source vessel, resulting in vapor carry under into the liquid stream, which in turn, raises the saturation vapor pressure of the stream
b)Even minor fluctuations in vapor space pressure control can result in vapor breakout of the liquid stream in transit to the pump. This is the reason for the recommendation to keep pump suction lines as short as possible, so that transit time from source vessel exit nozzle to the pump is minimised.

A 1m margin may be okay when the liquid is well away from bubble point conditions, but a larger margin would be preferable when dealing with bubble point liquids.

And while you're at it, stay well clear of Sundyne pumps and other makes operating at high speed or fitted with "NPSH inducers".

Hi,
To add to the discussion consider reading the document attached.
Pierre

• https://files.engineering.com/getfile.aspx?folder=dc378dfd-366e-4c60-95c2-5

@pierre,

Agreed, when pumping fluids close to bubble point, it would be best not to rely on the process fluid to cool bearings. Use some external cooling circuit with a barrier fluid in a API seal flush plan 52 or higher.