By E. M. Araza

CENTRIFUGAL PUMPS – Modern Design and Practices

Suction specific speed (Nss) is an index of suction hydraulic design for centrifugal pumps. Pumps that are in close range in Nss have similar suction characteristics. It indicates the pumps’ susceptibility to internal suction flow recirculation and cavitation – higher Nss values indicate higher susceptibility. The term suction specific speed is a misnomer because it is not a unit of speed and is considered dimensionless. It is calculated from the equation:

Nss = (N x Q^0.5) / (NPSHR3^0.75)

Where, in U.S. customary units:

N = speed of rotation in RPM

Q = capacity in GPM; (for double suction impeller, Q = GPM/2)

NPSHR3 = the impeller NPSHR in feet based on 3% head loss

Q and NPSHR3 are taken at the best efficiency point(BEP) - the flow with the highest efficiency on the pump curve, at maximum impeller diameter. NPSHR3 is commonly based on the impeller’s historical test records of 3% head loss but may also be estimated using empirical methods.

In two-stage or multistage pumps, the NPSHR3 is based on 3% loss on the first stage head. The NPSHR3 of any series stage impeller is ignored and has no effect on Nss. If the impellers are of dissimilar design the first stage head can be assumed to be equal to the total differential head divided by the number of stages because it is impractical to isolate and measure the actual first stage head.

The NPSHR may also be based on one percent (NPSHR1) or zero percent (NPSHR0) head loss in high energy or highly specialized services for added margin. Some users even call for 40,000 hours life or cavitation-free NPSHR. But NPSHR3 is the default value used in calculating Nss to get consistent comparison of the suction hydraulics of different pumps under evaluation for a particular service. The comparison is more accurate and valuable if they were done on pumps with the same configuration.

Example:

What is the Nss of a double suction pump running at 3560 RPM with 18 FT NPSHR3 at its BEP of 800 GPM?

Solution:

Nss = [3560 x (800/2)^0.50/ (18)^0.75] = 8,148

Typically, well-designed pumps have Nss in the range of 8,000 to 13,000. Pumps with Nss below 8,000 are found in pumps with exceptionally high speed, or exceptionally high HP with large shaft, that can cause them to have NPSHR higher than normal. But it can also indicate that the pumps may have poorly designed suction hydraulics.

The 13,000 Nss upper limit generally applies to end-suction overhang pumps. They have lower NPSHR because the liquid feeds directly to the impeller with minimal suction loss; they also have smaller shaft diameter resulting in lesser suction flow area blockage at the impeller eye.

On the other hand, top suction overhang pumps have circuitous flow path that results in more suction loss and higher NPSHR. Their flow has more pre-rotation, separation and vortices. They are more sensitive to suction piping and must have properly designed suction splitters, or vortex breakers. Using the same impeller, tests have shown that top suction pumps are about two feet higher in NPSHR than their end-suction equivalents.

Compared to overhang pumps, between-bearings pumps have bigger shaft diameters to fit the bearing housings on both shaft ends. Hence, to get the same annular eye area that an overhang impeller has, the impeller in between-bearings pump needs larger eye diameter to compensate for the bigger shaft blockage area - thus the impeller has higher NPSHR because of the faster peripheral speed if its eye diameter.

With 13,000 as baseline for end suction overhang pumps, some deducts may be applied to compensate for different pump configurations. A 500 incremental deduct for each and every differences in pump configuration may be used – for top suction overhang, for side suction between-bearings, for top suction between-bearings, for pump with open or semi-open impeller, for coke-crusher pump with cutter-screw or auger, etc. Example, for top-suction between bearings pump with semi-open impeller, a limit of 11,500 may be used.

These are not hard Nss numbers but are ballpark figures to use as general guideline in the absence of a more comprehensive methodology.

Some practical Nss applications

1. It is used  as basis for the recommended minimum continuous stable flow (MCSF) for the  pump, and its allowable operating region (AOR).
2. It can  be used as an index to compare competing pumps in a bidding process to  evaluate if a cost assessment should be made to level the playing field. Example,  would a pump’s higher cost be justified by its lower Nss implying that it  has better MTBF and, therefore, has less maintenance cost? 
3. It can  be used to determine if a hydraulic re-rate to reduce the NPSHR3 were      feasible. The reduction may be needed either (1) to increase its flow rate      due to higher product demand, or (2)  to meet a reduced NPSHA condition due to increased  friction loss as pipes and fittings aged over time. A low existing Nss may  indicate that an NPSHR3 reduction is feasible.
4. It can  be used as basis to proactively decide if some design upgrades are needed  upfront to improve the pump reliability (see further comments below).

As the word suction implies, Nss is affected by the pump’s suction hydraulic design  parameters such as suction nozzle diameter, casing suction bay development, splitter or vortex breaker, impeller eye area and inlet diameter, impeller suction vane angle, impeller annular eye area and inlet area between vanes, etc. 

On the other hand, the pump discharge parameters such as discharge nozzle, impeller discharge width and discharge angle, volute throat area and cutwater diameter, etc., have no effect on Nss. Their effect is on the pump specific speed (Ns). 

The myth of 11,000 Nss limit

Years ago, many end users specified that acceptable Nss should be less than 11,000. This practice came about after studies by some refineries showed that many bad actor pumps have Nss of 11,000 or over. 

The practice was misguided and have fallen out of favor. Pumps became bad actors not because of their high Nss but because of other contributing detrimental factors such as the pumps being too large for the flow rate, oversized impeller eye area, running at low flow, long slender shaft, poor suction hydraulics, high thrust loads, etc. In most cases, the impellers also have a high incidence angle between the suction flow angle and the actual impeller inlet vanes. But it is undeniable that there countless other pumps with Nss over 11,000 that operate reliably and with high efficiency because they are of excellent design.

Additionally, any concern about the reliability of pumps with Nss of 11,000 and above can be proactively addressed by including the additional requirements that the pumps:

1. have consistent history of  operating reliably at the given Nss 
2. be run within its preferred operating region (POR)
3. have low shaft flexibility factor (SFF), or low L^3/D^4
4. meet the specified vibration and noise levels
5. may be upgraded to more cavitation-resistant material

Compliance with the Nss limit has become a numbers game

The old practice to limit Nss to below 11,000 can result in high cost to end-users, and possibly inferior pump selection. Consider this example:

A double suction pump in hydrocarbon service was specified for these conditions: 3000 GPM, 750 FT head, 29 FT NPSHA, Nss below 11,000. These conditions can be met nicely by an 8x10x15 BB1 at 3560 RPM but the Nss will exceed 11,000 even if all the NPSHA is used without any NPSH margin (NPSHR=NPSHA). 

There are pumps that will meet these conditions with 26 FT NPSHR but their Nss are about 12,000. But because of the Nss limit of 11,000 a big, more expensive, and less efficient, two-stage pump at 1780 RPM was selected. The arbitrary Nss limit of 11,000 has put the end user in a big technical and financial disadvantage.

In one example, an end user bought identical OH1 pumps with the requirement that their Nss shall be less than 11,000. The pumps meet all the technical requirements but the Nss tested slightly above 11,000. The end user asked that they be modified to reduce the Nss below 11,000. The manufacturer made the impeller nuts larger to increase the swirl and suction blockage thereby increasing the NPSHR by 0.5 foot, just enough to bring the Nss below 11,000.

In another example, BB1 pumps were bought by the end user with the requirement that the Nss shall not exceed 11,000. The test results were excellent, except that the Nss was above 11,000. The end user also asked that they be modified to reduce their Nss below 11,000. The manufacturer cut back the volute lips to move the BEP 5% to the right where the NPSHR curve rises faster. The higher NPSHR at the new BEP was enough to reduce the Nss below 11,000.

In these cases, the pumps were modified to bring their Nss below 11,000 to comply with the specifications. But did the end users get better pumps? No! On the contrary, they may have become inferior. These show how the arbitrary limit, and with no consideration to the pump configuration, has become just a numbers game that can be manipulated with no resulting benefit to the end users. A better approach is to frame the 11,000 Nss, if a limit were to be specified, as a preferred value but is not mandatory. Another approach is to set an acceptable variance on any Nss limit, such as +/- 3% of the specified value.

Indeed, there are several factors that can cause the actual Nss to vary from the initial estimate. These include variations in the surface finish of the impeller and suction bay of the casing, casting variations of the impeller inlet vane thickness, variations in the in-between inlet vanes area due to pattern set-up, etc.

In the full version of this article the author discusses the suction hydraulics design flaws that can cause pumps to have higher NPSHR and lower Nss. It explains why a lower Nss is not always preferable, and why it is not always an indicator of a good pump hydraulic design.