Pump Specific Speed is one of the several dimensionless parameters used in pumps. Pump designers like to use dimensionless numbers, because his allows them to analyze and compare pump performance regardless of size. Otherwise, how would you compare a 6" pump performance with a 10" pump?

What is a typical maximum efficiency is to be expected from a given pump? There is an infinite variation of pump sizes, flows, generated pressures, speeds, etc.- so, it would be impossible to compare a performance of the 6" pump running at 3200 rpm at 200 gpm, with a 10” pump running at 1400 rpm at, say, 310 gpm.

Besides, the same performance can be achieved by a wider impeller with a narrower volute, as by the narrower impeller with a wider volute, if their outside diameters are different. In other words, there is an infinite variation of operating parameters (flows, heads, speeds) that depend on an infinite variation of pump geometry internals.A specific speed NS reduces this infinite variation of performance parameters into a single number, and an infinite variation of geometries into a single characteristic non-dimensional impeller profile shape. Then, a given impeller shape would point at the unique performance factor, i.e. specific speed. Obviously, to eliminate dependency of the operating parameters specific speed must be dimensionless, otherwise the remaining dimension (gallons, pounds, feet, etc.) would affect the comparison. Originally, specific speed was defined as:

OmegaS = Rotating Speed x Flow^0.5 / Head^0.75 ,

The units had to be in any system of "consistent" units, for example:

Rotational Speed - 1/sec

Flow - ft³/sec,

Head in feet times the gravitational constant (which has units of ft/sec²):



This is how the initial definition of "specific speed" was meant to be "dimensionless". But, such set of units is awkward, and the gravitational constant is simply dropped, and flow is entered in gpm, and rotational speed in RPM.

Specific speed thus defined is denoted NS. Strictly speaking, such modified definition is not dimensionless, but the proportionality terms are constants (such as ft³/sec x 448.8 = gpm), which makes the NS directly proportional to "OmegaS". The non-dimensional nature of the NS (disregarding the dropped constant) makes it independent of rotational speed RPM. This is because of the affinity laws governing the operation of the centrifugal pump. These state that flow varies linear with speed, and head as a square:

Q ~ RPM

H ~ RPM²

If RPM is changed n-times, flow changes n-times also, and head changes n² times.



(Note that NS has not changed when RPM changed).

Incidentally, similar rule happen to apply to impeller OD cut: if impeller diameter changes n-times: flow changes linearly, and head as a square. The net result: no change in NS. Also, if all geometric linear dimensions of a pump are changed at the same ratio, its specific speed still remains constant. (This part, however, is more specialized, - usually of the realm of the pump hydraulics designers - see special discussion on that):



Same shape, but different size – i.e. Specific Speed is a measure of the impeller geometry similarity (i.e. “affinity”), - not the size of it. In other words, if you are looking at a cross-sectional drawing of a pump, there would be a certain specific speed NS, unique to the impeller shape - regardless whether you are looking at a cross-section of a 6" design, or a 60" design.

It was discovered that a pump efficiency at the best efficiency point (BEP) depends mainly on the Specific Speed, and a pump with Specific Speed of 1500 is more efficient then the one with specific speed of 1000. Charts and publications were developed, and are available as an estimate of a "reasonably achievable" efficiency for a given pump, versus actual.

Suction Specific Speed (NSS)

Suction Specific Speed is another dimensionless parameter used in centrifugal pumps. Much of the discussion on specific speed (NS) applies to suction specific speed (NSS). (See topic "Specific Speed (NS)- why is it dimensionless, or - is it?").

While specific speed (NS) is mostly related to the discharge side of the pump, the suction specific speed deals primarily with its suction (inlet) side. The head (H) term in the denominator of the defining formula for the NS is substituted by the NPSHR:

NSS = RPM x Q^0.5 / NPSHR^0.75, where

flow is in gpm, and NPSHR is in feet.

Values of NSS vary from about 6,000 to 15,000, and sometimes even higher for the specialized designs.

From the discussion of a pump suction performance (see topic "How does pump suction limit the flow?"), we know that conflicting demands are imposed on a pump system by the pump user and a pump manufacturer.
A user would prefer to provide as low NPSHA as possible, as it often relates to a system cost: for example, higher level of liquid in the basin of the cooling water pumps requires taller basin walls, or deeper excavation to lower a pump centerline below the liquid level. A pump manufacturer, on the other hand, wants to have more NPSHA, with an ample margin above the pump NPSHR, to avoid cavitation, damage, and similar problems.

In other words, a wider margin (M) can be achieved either by increasing the NPSHA, or decreasing the NPSHR, since

M = NPSHA - NPSHR

Thus it may appear that a lower NPSHR design is preferential, and a competing pump manufacturer might present a lower NPSHR design as automatically translated into construction cost savings - because of not having to increase the NPSHA. Since a lower NPSHR design means a higher value of NSS (according to the definition above), the highest NSS design might seem to look best. In reality, however, this is not so.

In a topic "How does pump suction limit the flow?" it was explained that higher flow velocities result in reduction of the static pressure, which may then become dangerously close to the fluid vapor pressure and cavitation. Thus, lower velocities result in higher localized static pressure, i.e. a safer margin from the cavitating (i.e. vaporization) regime. Since the velocity is equal to flow divided by the area, a larger area (for the given flow) reduces the velocity, - a desirable trend.

This is why a larger suction pipe is beneficial at the pump inlet. Cavitation usually occurs in the eye region of the impeller, and if the eye area is increased - velocities are decreased, and the resulting higher static pressure provides a better safeguard against vaporization (cavitation). So, a larger impeller eye seems like a way to lower the NPSHR:



Figure 3-1 Larger impeller eye results in lower NPSHR at BEP, but certain problems arise at off-peak operation

Unfortunately, the flow of liquid at the impeller eye region is not as simple and uniform as it is in a straight run of a suction pipe. Impeller eye has a curvature, which guides the turning fluid, like a car along the sharp curves of the road, into the blades and towards the discharge. If a pump operates very close to its BEP, the inlet velocity profile becomes proportionally smaller, but the fluid particles stay within the same paths:

If, however, a pump operates below its BEP, the velocity profile changes, and no longer can maintained its uniformity and order. Fluid particles then begin to separate from the path of the sharpest curvature (which is the impeller shroud area), and the resulting mixing and wakes produce a turbulent, disorderly flow regime, which makes matters difficult from the NPSHR standpoint.
 

Figure 3-2: Even thought large eye impeller has better NPSHR at BEP, it has flow separation problems at low flow

The upshot of all this is that a larger impeller eye does decrease the NPSHR at the BEP point, but causes flow separation problems at the off-peak low-flow conditions. In other words, a high Suction Specific Speed (NSS) design is better only if a pump does not operate significantly below its BEP point.

Interestingly, with very few exceptions, there is hardly a case where a centrifugal pump operates strictly at the BEP. The flow demands at the plants change constantly, and operators throttle the pump flow via the discharge valve. High NSS designs are known to result in reliability problems because of such frequent operation in the undesirable low flow region. Actual plant studies have shown, that above NSS of 8500 - 9000, pumps reliability begins to suffer - exponentially:



Figure 3-3 Plant experience shows that impellers designed with NSS greater than 9000 have poor reliability record

Realizing this, around mid-80s, users started to limit the value of the NSS, and a Hydraulic Institute uses NSS = 8500 as a typical guiding value. It might be of interest for you to calculate the values of NSS of your pumps, and find out if a correlation between those and reliability exists at your plant.

Source – Engineering Review
For any further queries with reference to the article write to shirish@engrreview.com