25362

For what’s it is worth:

I personally don’t know any direct correlation between viscosity and NPSHr.  

Manufactures test NPSHr using fresh water at a fixed temperature.  Viscous liquids like oil have a different pressure/temperature relationship than water; therefore, the “required” NPSH will be lower for oil than fresh water.  

The other side is that viscous fluids don’t flow as well as water.  A manufacture may require you reconsider the intake piping or increase the NPSHr to accommodate the additional viscous drag.

Bottom line is you should contact the pump manufacture to obtain the NPSHr for a given application.  

Hope this helps some.

d23 !!!
There you are.
Where you been?

I was waiting on this thread for a guy like you to post because I never work with viscous sticky junk, but you do.
I like water, clean water in fact, I am no terd herder.

If no one posted I was going to say the same as you did, talk to the manufacturer each time.

I would guess that non-compressible newtonian fluids will be so similar to water that there is no difference, and then as viscosity increases at some point the flow regime within the pump will be affected enough to make a difference.

What do you know 25362?

PUMPDESIGNER

To PUMPDESIGNER, Sam Yedidiah in his Centrifugal Pump User's Handbook quotes one pump manufacturer that says that for viscosities higher than about 100 SSU a correction factor K should be applied to the NPSHr. This manufacturer presented data in the form of straight lines on logarithmic paper, following the formula below:

K = -0.4776[log(V/V₁)]⁴ + 3.8688[log(V/V₁)]³ - 9.875[log(V/V₁)]² + 8.1772[log(V/V₁)] + 1

where V₁ relates to a liquid with a viscosity of 100 SSU and V relates to the viscosity of the pumped liquid.

Using this formula for a liquid with viscosity 400 SSU (86 cSt) I found it gives a K=2.67 to increase the NPSHr originally intended for a fluid with 200 SSU (43 cSt).

Yedidiah adds, I quote: "At present I know of no experimental data that would confirm or deny the validity of this equation", unquote.

Anybody that could shed some light on this subject is invited to do so.

  

That is neat 25362,
I can see why you want to verify the formula with proof or work on that problem further, really neat.

I always have a few ongoing research problems like that it seems.  Have been very excited in the past when I solved one.

Am close to solving one now, how to calculate the amount of energy released from a pipe when pressure drops a specific amount (For PVC, HDPE, etc.).  I want that energy stated in ergs and I am very close, but now I realize I will have no experimental data to verify.  I have a few friends that could do the work in the field, so I will probably pay them to get it for me.

PUMPDESIGNER

   I ran an aol/Google search on viscosity effects on cavitation and didn't get much enlightenment except for an abstact of an ASME Paper FEDSM2001-18087 by F.C.Visser of Flowserve, Netherlands. He ran a CFD (TASCflow) model of an impeller to check effects of kinematic fluid viscosities about 100X that of water. It says "the influence of higher viscosity [is] boundary layer displacement in the impeller passageways yielding a shift of the Best Cavitation Point to a lower flowrate." Never have heard before of the term "Best Cavitation Point" (bcp), I am somewhat confounded by its implication. It's not clear whether Visser is referring to the incipient cavitation occurring at some specific chordwise length displacement in the boundary layer flow over an impeller blade or some state of cavitating flow that might be equivalent to the 3% headdrop criterion widely used for NPSHR determinations. Why does cavitation inception have to initiate in a blade boundary layer? It's possible that streaming mid-passage vortices in a suction side recirculating flow would be a more likely site for first cavitation inception and this can be more readily linked to some off-bep flowrate.
   23562's equation from Yedidiah is equally entralling. Sounds like that unknown manufacturer ran a multitude of tests at various fluid viscosities. Based on my experience with the variability and non-repeatability of NPSH test results, the coefficient of deterination (R^2) of their curve-fitted

equation needs to be known for a judgment of veracity.

   In Lobanoff,V.S & Ross,R.R., 1985, "Centrifugal Pumps - Design and Application", Gulf Publishing Co. it says in Chapter 9 on NPSH under the subheading, "Effect of Viscosity" that "Although the influence of viscosity is predictable on other hydraulic characteristics, particularly head, capacity and efficiency, little general information is available to indicate the effect on NPSHR. From experience we know that up to 2,000 SSU we are safe in using water NPSHR values. Above 2,000 SSU, it is necessary to use the pump designer's judgment or experimentally determine the cavitation characteristics."

I've been told that an Indian oil refiner has NPSH-related  problems in his visbreaker bottoms pump whenever changing over to more viscous (bitumen-like) feedstocks. I didn't hear more. If and when I get more information I'll let you know.