NPSH for Centrifugal Pumps

Pennpiper --Good link. Thanks, it's a keeper. Mohapatra will do well to spend a bit of study time there.

McNally's $99.95 CD or his book would be a good investment for any engineering office where the design, analysis, selection and troubleshooting of pumping systems are frequent tasks.

Ed Weldon

is it true that if the pump inlet pressure is above the fluid true vapor pressure (TVP) then no cavitation?

is there a typical/industry design margin (let say 3 psia) above the fluid TVP for pump inlet?

if bubbles do form at the pump inlet is there a method to design the impeller blade such that it will prevent sudden collapse of these bubbles?

Guest --

What is the definition of "true vapor pressure" I'm not familiar with that term (perhaps I'm dating my pump experience here).

With respect to design of impeller blades to prevent cavitation bubble collapse I know there are tricks like the use of inducers and special blade configurations for purpose designed condensate pumps. Having never worked on the design of such things I'm not going to attempt further comment. At the time I worked in the pump industry the design of such things like inducers was somewhat of a trade secret.

The definitive text on the subject of Centrifugal and Axial flow pumps by A.J. Stepanoff of Ingersol Rand, publisher John Wiley, 1957, discusses impeller inlet design, vane angles and such and the relationship to NPSHR. I'm sure subsequent texts on pump design treat the subject as well. But Stepanoff make no mention of inducers. Inducers are mentioned by several of the article authors in the McGraw-Hill Pump Handbook 2nd Edition but little specific detail of their design is provided. Possibly Sam Yedidiah's book covers the subject. I don't yet have a copy. Sam designed the inducers for Worthington's 2nd iteration of standard design back pullout chemical pumps around 1970 when I worked there.

Ed Weldon

Basically you can't calculate the NPSH a pump requires (NPSHR). That is, unless you first want to study one of the classic texts on centrifugal pump design.

But you can ask the manufacturer for an NPSHR rating curve for the pump. (usually presented as NPSHR in feet or meters vs. flow rate through the pump in gpm or similar metric units. All the large industrial pump (over about 2KW) manufacturers can usually provide that data unless the pump is very old. Pump rating curves tend to be a bit conservative to accommodate manufacturing variances and minor errors in system calculations.

Perhaps you mean to calculate the NPSH available (NPSHA) at the pump inlet. That would be the total usable mechanical energy content of the liquid = pressure head plus velocity head. Usable pressure head is calculated from gauge pressure + atmospheric pressure minus liquid vapor pressure. Velocity head is calculated from the flow rate and the density of the liquid. If you are simply drawing liquid out of a tank with negligible liquid velocity in the direction of the tank outlet then your energy content is only based on the gauge pressure in the tank (zero if an open/vented tank) + the atmospheric pressure minus the vapor pressure converted to energy units + the static elevation (in consistent units) of the tank liquid level above the centerline of the pump inlet. (this would be a minus number for liquid levels below the pump inlet centerline.

Ed Weldon

Hi Ed,

sorry but, you don´t consider the friction losses at the suction side of the pump in order to estimate the NPSHA? I usually do as a negative value.

Regards

Ernesto

Ernesto -- If pressure is measured directly at the pump inlet then the friction losses in the suction pipe need not be calculated. If you already know the liquid properties (density, viscosity and vapor pressure) at the pump inlet then only two additional measures are needed to be measured, pressure and flow rate, to determine the energy content of the liquid (NPSHA). The pressure measurement will reflect the upstream friction losses. And the flow rate will enable you to calculate the velocity head which must be accounted for since that form of energy is a contributor to NPSHA. This reflects the typical test stand environment for pumps where you are actually measuring.

If on the other hand the only known pressure is atmospheric pressure at some point in the inlet system such as the liquid level in a body of water feeding the pump then friction losses must be calculated. In this calculation pressure head and velocity head (colloquial terms for potential and kinetic energy of the liquid) need not be considered separately.

This is what usually faces the engineer in the real world when designing a new system or troubleshooting an existing system where it is not practical to insert a pressure gauge in the inlet line or accurately measure flow rate.

So yes, you do subtract friction losses from available total pressure head(which includes atmospheric pressure) in the inlet line to estimate NPSHA.

Ed Weldon