Common Pumping Mistakes
by Jim Elsey

Given a set of required hydraulic conditions (head and flow), a defined set of fluid properties and a specified duty cycle, many pump users would think it would be relatively easy to select the correct pump for an application. Sometimes it is simple, but most often it is not.

In reality, every centrifugal pump is designed for only one operating point of flow and head; all other points on the manufacturer's curve are, to some degree, simply a hydraulic compromise and departure from the best efficiency point (BEP). Good pump designers scrutinize common industrial processes and design their pumps to specifically match the corresponding head and flow rate as required. In rare cases, the user will have the manufacturer design a pump specifically for the application.

What selection criteria do you evaluate when buying a pump? Some people look first at the pump's initial cost, while others look at the operating conditions and the proximity to the pump's BEP. Some may also consider whether the impeller is at maximum or minimum trim. Giving major deliberation to the pump efficiency and expected energy consumption is always a good decision.

For simplicity, this article will discuss centrifugal pumps and fluid that exhibits Newtonian properties comparable to water. Assume the fluid is 68 F with a specific gravity of 1.0, a viscosity of less than 5 centipoise, and little to no suspended solids (SS), entrained air or non-condensable gases.

BEP Considerations

There is no set of universal rules for all pump selection cases except this: Be pump smart and informed. Analyze several pump selection aspects, not just one. Note that the right pump may not be a centrifugal pump. What may be the correct pump for one fluid may not be for a different fluid.

Selecting a pump with an operating point at or to the near left of the BEP is a commonly accepted and conservative decision, but do not be afraid of operating points to the right of the BEP. If the operating points are not too far to the right, there will still be plenty of margin before the end of the curve. Selecting a pump near the BEP will yield the most efficient pump with the least amount of vibration and radial forces acting on the shaft. Ensure that your calculations for the system design curve, also known as system resistance curve (SRC), are correct because the pump will operate where its curve intersects the system curve.

Do not expect to run near or at the end of the curve. This area is fraught with recirculation, cavitation and net positive suction head (NPSH) margin issues, regardless of the manufacturer. As I candidly state in my pump and fluid dynamics course, "There are demons and dragons in the area at the end of the curve."

Selecting a pump that will operate near the minimum flow limits is just as risky as choosing one that will operate at the end of the curve (see my November 2015 Pumps & Systems column, "Follow These Steps for a More Reliable Pump," for a detailed discussion on minimum flow). Be aware of when manufacturers claim a lower acceptable minimum flow than others for a similarly designed pump. The pump will operate in that minimum flow region, but the added costs for mechanical seals, bearings and unplanned downtime will exceed initial savings. Most manufacturers' warranties will not cover wear and/or failure of the mechanical seals and bearings.

As a general rule do not select a pump with the maximum size impeller. Selecting a maximum "wheel" (impeller) leaves no margin for mistakes in your calculations. Unacknowledged factors could increase system friction because of age, corrosion or other factors such as marine growth or fouled heat exchangers. Knowledgeable sources suggest selecting impellers that are no more than 90 to 95 percent of the maximum diameter. This decision allows for future growth in the system.

Many users may specify impellers that are less than the maximum diameter while sizing the driver to be non-overloading for the maximum impeller at or near the end of the curve. Problems with vane passing frequencies may arise; at maximum diameter, the impeller vanes are close to the volute cutwater (or diffusor vanes), creating pressure fluctuations that cause unacceptable vibrations. Pump manufacturers can incorporate good design practices to control the distance ratio from the impeller tip to the cutwater/diffusor edge.

Minimum impeller diameters are less efficient than larger impellers for a given pump. Be careful using the affinity laws for any approximations as you approach minimum diameter, because the calculation accuracy is significantly less as you go from maximum to minimum diameter. I recommend using the original equipment manufacturer's (OEM) published curves, if available. Do not trim the impeller less than the OEM's stated minimum, especially in a lift or self-priming application.

NPSH Margin

The difference between NPSH available (NPSHA) and NPSH required (NPSH R) must always be a major factor in the selection process. The higher the margin, the better. Different fluids will react to the margin in different ways. Water at 70 F will have far more deleterious cavitation effects than a hydrocarbon at the same or higher temperature. Also note that the NPSH R stated on the curve is really NPSH3, which means the pump was already cavitating with a 3 percent head loss when the OEM measured that point.

When choosing between two pumps that have otherwise equal selection factors, it is prudent to select the pump that is the most efficient. But if the pump will operate infrequently or for short periods, then allow some of the other selection parameters to weigh more. A simple operating cost versus reliability analysis may help you make that decision. A lower-speed pump that is a few points less efficient may yield 20 years of little to no issues as compared with the higher-speed pump that will wear eight times faster.

With the exception of very high-speed pump applications of more than 30,000 revolutions per minute (rpm) or multistage high-energy boiler feed pumps, I often select a slower-speed pump when compared with an otherwise equal but faster pump. While this choice initially involves a bigger pump, my preference is the result of many years in submarines and later at power plants and refineries; I want my pumps to be highly reliable for as long as possible.

If you change the fluid properties, many of these tips will change in their amplitude of importance. The intent is simply to make you think about the selection process. Note that pump boundaries with regard to critical speeds, temperature, pressure and materials were not discussed.

"How to Select the Right Centrifugal Pump." Robert X. Perez
"Influence of Impeller Suction Specific Speed on Vibration Performance." TAMU -29 2013 David Cowan, Simon Bradshaw, Thomas Liebner
"Centrifugal and Axial Flow Pumps." A.J. Stepanhoff
American Petroleum Institute (API) Standard 610 11th edition
Specifications and Standards ASME/ANSI B 73.1 M 2012

To read other articles in the 'Common Pumping Mistakes' column, go here.

Other Pump Selection Parameters

The following are other factors you may wish to add to your evaluation spreadsheet. They are listed in no particular order.
  • L-over-D ratio.—Also known as L3 ÷ D4, shaft slenderness or shaft deflection ratio, this is an index of how much, if any, a shaft will deflect when subjected to radial forces. Like golf scores, the lower the ratio, the better. The OEM will provide this ratio if it is not already in the literature. The radial forces causing the deflection are the result of operating away from the BEP.
  • Weight.—Pump weight often translates to reliability, casing thickness allowances and also allowable flange (nozzle) loading.
  • Impeller vanes.— While related to NS, you may also want to look at the number of impeller vanes and the slope of the pump curve. Does it continually rise as the flow approaches shutoff (zero flow)? Flat or drooping curves can be an issue.
  • Specific speed (NS) & suction specific speed (NSS).—These measurements will be the subject of a future article. For now, note that numerous technical articles and papers regarding NSS recommend keeping the values below 8,500, with 1,100 (USCU, or U.S. Customary Units) as an absolute maximum, but I postulate that those studies were based on erroneous data interpretation from the 1970s and '80s. If you strictly control where the pump is on the curve (near the BEP) or are selecting new pump designs, then this is not an issue. I recommend evaluating the ratio of the overall impeller diameter to the inlet (suction eye) diameter. This is also known as the D1-to-D2 ratio. Impellers with larger inlet eye areas can have issues with recirculation if operating away from the BEP (comes back to NS). One consideration to discern is if the impeller(s) is overhung or in a between-bearing arrangement.
  • Impeller tip speed.—This measurement becomes more important as the amount of solids increases in the fluid, but it is also an issue with regard to inertia (WR2), since slower pumps require larger-diameter impellers to develop the required head. Typically, in clean to slightly dirty water service, the tip speeds should be below 130 feet per second (ft/s); for medium to heavy slurries, keep it below 115 to 110 ft/s; and if the impeller is elastomeric, coated or lined, then keep the tip speed below 85 ft/s.
  • Piping size.— Some pipe sizes are drastically higher in price than others (even if a smaller size). Pipe size also translates directly to fluid velocity. While not as important on the discharge side of the pump, it can be the difference between success and failure on the suction side. A higher velocity at the suction flange will reduce the pressure and may create NPSHA issues. Higher velocities on the discharge side simply translate to higher pumping costs. I always recommend suction velocities below 10 ft/s, and 6 to 8 ft/s is even better, especially when faced with low submergence.
  • Pump operating speed. —I caution end users not to operate pumps faster or slower than the OEM recommends, because the impeller shrouds, and blades have torque and pressure limits. Low-speed operation may prevent hydrodynamic stabilizing forces from developing in critical clearance areas like the wear rings, so watch out for sustained operations below 600 rpm.