Frequently Asked Questions

1. Q: Does excessive amount of air at the pump suction cause cavitations?
  A: No. Air has nothing to do with it. Cavitation is caused by the collapsing (imploding) vapor (not air) bubbles. These bubbles are simply a vaporized liquid in the region where static pressure dropped below vapor pressure. Air causes other problems, such as air locking, and even a very small amount of it causes significant loss of performance (head drops), but this is a different subject.

2. Q: Can a gear pump "lift" liquid? What is dry lift?
  A: Yes. Gear pumps have good lift characteristics, in the range of 5-20 feet, depending on the particular design. Lift characteristics of a gear pump improves significantly if even a minute amount of liquid is initially allowed to "wet" the internals, which is often the case if a pump was tested at the factory, and some residual liquid remains, or intentionally pre-lubed at the site. This minute amount of liquid acts as a capillary barrier in the clearances, preventing air from escaping back to suction during startup at lift. With no pre-lube, gear pump will still lift, but not as well.

3. Q: What determines the number of stages of the progressing cavity pumps?
  A: Number of stages depends on several factors; with the main one is total differential pressure. Typically, a stage is added for each 75-100 psi. For example, a 300 psi differential would require 4-5 stages. Manufacturers catalog provides different curves for different stage number designs.

4. Q: Is it true that if centrifugal pump runs in reverse, it will generate zero head?
  A: No. As a rule of thumb, a centrifugal pump running in reverse generates approximately half of its rated head. However, such operation is very inefficient, and motor horsepower would be much higher, as compared with half head operation of a pump running at the correct rotation.

5. Q: What is the effect of the degree of saturation of dissolved gasses on NPSH?
  A:
There is definitely an effect. The dissolved gas changes the molecular interaction of the liquid in which it is dissolved. Chemical engineers are familiar with this phenomenon via Henry’s Law, and Oswald coefficient, which relates the V/L (void fraction – the freed-up gas volume to liquid volume ratio) as function of saturation pressure and actual pressure of the mixture. This is not to be confused with the effect of free gas on pump suction performance, and neither it has anything to do, directly, with cavitation (which is caused by vaporization of liquid and subsequent collapse of vapor bubbles). The dissolved (not free) gas affects the “ability” of a liquid to become vapor when the pressure drops.

In practice, a good example are cooling water tower double-suction pumps, where the incoming water has been so well aerated when passing through the tower - that a significant amount of air stays dissolved, and reduces the NPSHA. The NPSH margin (NPSHA-NPSHR) for these pumps is not significant to begin with, and with air affecting the NPSHA, the propensity for these pumps to “get in NPSH trouble” is real. As an estimate, the reduction of NPSHA for these pumps is about 1-3 feet.

You should be OK if NPSH margin is good. Also, even if some nitrogen dissolved in water, it will probably stay dissolved and will not come out of the solution at the low-pressure inlet areas, because of the time delay – it flows through quickly. In the cooling tower example, the water stays well dispersed in order to get cooled, i.e. the surface area is extremely enlarged, and air can easily get in.

6. Q: For a backward curve vane centrifugal pump, what is the theoretically highest head achievable for a given impeller tip-diameter and rotating speed? How is the equation derived?
  A: As you can see from the velocity triangles shown above, head rises as flow is reduced. At zero flow, the meridional velocity vector is zero, and the tangential component of the absolute velocity is exactly is equal to the rotational velocity (more on nomenclature see Article #13 under Technical Papers section). This is the same regardless of the type of the blades or their number. Thus, the equation for the idea head becomes the following at zero flow (VTH=U):

H = (U x VTH)/g = (U x U)/g

For example, suppose you have a 6” impeller at 3600 RPM. Peripheral (tip) velocity of the impeller wheel is:

U = 6 x 3600 / 229 = 94.2 ft/sec, and H = 94.22 / 32.17 = 276 feet

(In above, we assumed to inlet prerotation, and 229 is a conversion constant for US units, and gravitational constant g=32.116 ft/s2)

Note also that the width of the impeller does not come into equation either! – but only the impeller OD. (You may want to check a few pump curves at a pump catalog for the value of head at shut-off, as a matter of interest).

7. Q: What is the permissible level of shear stress in pump shafts used for double suction split case pumps for ANSI 410, 431, 316 materials?
  A: Allowable stress levels, for static components, are usually related to the tensile stress of the material, per ASME rules. For cyclically loaded components, such as shafts, the endurance limits are applied, which produce values more conservative then for statically loaded members. For example, martensitic stainless steels, such as 400-series, often have a value of 7500psi allowable, but each manufacturer may have its own values and procedures.

The importance of knowing the allowable shaft stress is obviously to determine if a shaft diameter is sufficiently large. However, other factors need to be considered. For example, for the types of pumps you mentioned, the impeller can be positioned on a shaft axially in variety of ways. Some designs have threaded shafts, with locknuts against the impeller hub. Others may have lock-rings, positioned in a shaft groove. Threaded shafts present additional stress concentration - and the shaft diameter must be sized even more conservatively to accommodate that. For double suction pumps single-stage pumps, the impeller sits in the middle of a between-the-bearings span, and if the shaft is threaded, the threads end up also close to the middle of the shaft, where stresses and deflections are highest – not a very desirable situation.

If a sleeve positions an impeller axially, and threading of the shaft is thus eliminated, or removed further away from the center of the shaft, this would be a better design alternative. You may look up the cross-sectional illustrations from several manufacturers, to compare the designs, and how the impeller-to-shaft positioning is accommodated.

8. Q: When we calculate the dynamic head (v2/2g) is it in absolute pressure or gauge?
  A: If you substitute the units, you will get feet of head:

(Ft2/s2) /(Ft/s2) = FT (or meters if in English units)

Dividing by the appropriate constant gets us pressure units. For example, dividing by 2.31 x SG gets psi.

Typically, the gages on the suction and discharge side of a pump read pressure in psi (or bars if in non-US system) – and, these are usually in gage (e.g. psig) units. There are gages that read absolute pressure (psia)- they are more typically on a suction side, where the pressure is lower. The correlation is straightforward:

PSIG = PSIA + 14.7 (in US system)

The idea is that the gage shows 0 (zero) at normal atmospheric condition, - which is 14.7 psiA (absolute). This way, a 3 psi vacuum would technically be 14.7 –3 = 11.7 psiA, or – (minus) 3 psiG

So, the “gage” pressure gage dial show “0” for 14.7 psiA, and the “absolute” pressure gage dial shows “0” at 0 psia (absolute vacuum).

To calculate pump head, velocity heads at suction and discharge must be accounted for. These do have neither “gage”, nor “absolute” units. The total pump head is the difference between the discharge head and suction heads. Each of those has units – either gage, or absolute – depending how it was measured (what type of gage used) – but the TDH (differential head, or a pump head as it is called).

For example, say the discharge gage reads 100 psiG (100x2.31/1.0 = 231 feet (g), for water at SG=1.0), and velocity head at that location is 10 feet. Say, suction pressure is 20 psiG (20x2.31/1.0 = 46.1 feet(g), and suction velocity head is 5 feet.

Then, TDH = (231 + 10) – (46.1+5) = 199.9 feet

9. Q: What are about the effect of temperature on NPSHR?
  A: The cold water is a much “tougher” with regard to cavitation as compared to warm water. The explanation goes back to the fundamentals of thermodynamics of cavitation. The vaporization (boiling) of liquid in the process of cavitation is a thermal process, and depends on fluid properties, such as pressure, temperature, latent heat of vaporization and specific heat. To make vapor form, the latent heat of vaporization must be derived from the liquid flowing through the pump. This flow of heat can only be possible when the liquid temperature is above the saturation temperature at the main pressure in the low-pressure zone where cavitation is about to begin. In other words, the pressure in the cavitation region must fall below the saturation pressure corresponding to the liquid temperature.

As we know, pump head begins to drop when cavitation begins – as bubbles block the passages more and more. Keep in mind that the term “pump head” ultimately means “energy per unit of mass flowing through the pump”. This energy (i.e. enthalpy) is related to specific heat as:

Dhf = CL x DT

This heat transfers transforms some liquid into vapor. The ratio of the resulting vapor volume to the remaining liquid volume would determine the extent of blockage of the passage by vapor. “B” is a thermal criterion, defined as:

B = (Vvapor / Vliquid) x (Dhf / L)=(Vvapor / Vliquid) x CL x DT

As you can see, the “blockage of the impeller passage” is directly proportional to the latent heat, i.e. more pronounced for cold water then warm water.

This affects not only loss in performance, but also the damage to the pump. Vaporization causes performance drop, but their eventual collapse (implosions), as they move on to a higher-pressure zone, is likewise more violent for cold water, as compared to warm – for the same reason, back to enthalpy and specific and latent heat.

In fact, the analogy can also be extended to hydrocarbons. As you know, API allows NPSH corrections for hydrocarbons, versus tests on cold water. The reason – hydrocarbons are less damaging from the cavitation standpoint, as compared to water. In practice, however, this rule is not usually enforced, as most people prefer to rather have some safe margin of NPSH, instead of cutting “too close to the wire”.

10. Q: What happens if a pump starts with fully open valve (end of the curve)?
  A: For pumps with low to medium specific speed (NS < 3000), which is the majority of pumps at chemicals plants and refineries, etc., the pump horsepower rises with flow. The lowest horsepower is near the shut valve condition. This is why it is best to start pumps with discharge valve slightly cranked open, and open it up more, to the desired flow, after a pump has been started. This puts less stress and in-rush current on the motor, and it will last longer.

For high specific speed pumps, the shape of the power curve is different, and the lowest power may not be at the shut valve, i.e. starting of those is not as simple.

11. Q: What is the minimum continuous flow that must be guaranteed to a centrifugal pump? And how is it related with the viscosity?
  A: Hydraulic Institute addresses the general concept regarding the “..minimum and maximum flow, at which they should be operated continuously of for an extended period of time..”. The Standard continues to say that “..Operation of pumps at reduced capacities may lead to the following problems: temperature buildup, excessive radial thrust, suction recirculation, discharge recirculation, insufficient NPSH..”

Regarding viscosity, the usual pumping limitations of centrifugal pumps apply. Since pump energy level is an important consideration for the minimum flow, as the HI states, the corrections for flow, head, efficiency and power thus affect the MCSF.

12. Q: Why should we run an annual survey for pumps and what are its advantages and how is it useful?
  A: Installation Surveys could be very helpful for a plant. Typically, an “80/20-Rule” applies, e.g. 20% of pump “bad actors” cause 80% of problems. In other words, from a 100% pump population at a given plant, 80% usually work fine, 15% somewhat below the desirable, but with operators and maintenance people more or less learned to “live with a problem”. It is the remaining 5% that causes most headaches. The problem is that not always these are well documented, due to people turnover, lack of systematic follow-ups and updates, lost records, etc. Like everything else, it takes an investment of time and money, to save more time and money in the long term.

An Annual Survey is a good way to tack-in loose ends, and update records of a plant pump population. It is similar to annual parts inventory count, that purchasing and inventory control people do at the manufacturing plants. It is typically a 1-week (depending on a plant size, and for a large plant it could take longer) program to survey the machinery, fill-out the forms, to document operating conditions, list the number of failures or stoppages, with reasons behind those, as well as note criticality of each installation, and produce a survey report.

Some plants (usually larger ones) conduct much more comprehensive surveys, with a goal of identification, documentation and gathering of data for loading to their Computerized Maintenance Management system (CMMS). This type of survey is extremely involved since a database has to be created for all of the plant pumps. We know of a site with 1200 pumps, which took a team of 6 people about a year to complete the survey. Most of the pumps had no tags and the vendors had to be contacted to identify types and then the pumps disassembled to measure impeller diameters, etc. There are companies with existing databases that do this type of gathering and data management, and it is significantly more costly.

A Monthly Survey, and those by the Weekly and Daily records should follow annual Survey. If an Annual Survey is done thorough and properly, the Monthly surveys should not be too lengthy, perhaps taking one day; and the Weekly Recording could be a 1-hour exercise to note any abnormalities. The Daily Log is essentially a part of the on-going operating procedure that operators and maintenance departments go through routinely, as part of their job.

A plant can have the Surveys performed by their own personnel, such as Reliability Team, or to sub-contract services of a consulting agency. Each approach has benefits. Internal team, if established and active, may be more intimately familiar with plant issues, and have a quick access to other departments, if needed, in a course of their day-to-day presence. But it may not have sufficient time and resources to stay focused on the program, due to manpower limitations, emergencies, interruptions, and turnovers. Hiring consulting agency takes some time to become familiar with the plant specifics, but, in a long haul, maintains consistency and planned follow-up, as well as provides an independent and unbiased evaluation of the plant equipment operation and alternatives.