The following Q&A is provided by the Hydraulic Institute (HI). HI is a leading developer of pump standards, technical reports and guidelines, data books, and educational material. As an ANSI-accredited standards developer, the Hydraulic Institute continuously reviews and updates its pump standards and adds new subjects as requested by the pump community. The organization encourages users to submit questions, and HI will respond based on existing HI standards, technical documents, and educational products. Please direct your inquires to

Q: How can we get longer life from the ball bearings in our pump? Our process uses an end-suction centrifugal pump to circulate heat transfer liquid at 750 F. The bearing housing is cooled, and the bearings look clean after they fail. What else should we look for?

A: Pump bearings are most likely to fail due to contamination and poor lubrication. However, that does not seem to be the case since the bearings are clean. In addition, the bearing cooling is a big help. That leaves the matter of excess bearing load. This is also consistent with the regularity of the failures.

One common cause of high bearing load is coupling misalignment. If the coupling halves are aligned when the pump is cold, they will usually move out of alignment as the pump heats up. Even with the pump supports at the shaft centerline, some movement is still possible. To avoid this problem, shut down the pump when it reaches operating temperature, and recheck the coupling alignment.

Another cause of high bearing load is operation below 50 percent of the pump”s best efficiency rate of flow. At low flowrates, all of the flow from the impeller cannot exit from the pump casing so it recirculates in the impeller. This causes an uneven distribution of pressure around the impeller resulting in a high radial force on the impeller, which must be supported by the bearings. To guard against this, determine the actual rate of flow through the pump, and make sure it is close to the pump”s best efficiency flow.

Another problem, which is unique to high-temperature applications, is excessive expansion on the bearing inner race. Even with bearing cooling, heat is transmitted to the bearing through the pump shaft. This heat causes the shaft diameter and the inner raceway of the bearing to expand, which in turn squeezes the rotating balls in the bearing. The result is shorter bearing life, which can be corrected by using bearings with greater internal clearance.

Q: I have been told that centrifugal pumps handling hydrocarbon liquids can operate safely with less NPSH Available than the NPSH Required by the pump rating curve. Is this true? If so, how much reduction in the NPSHR can be allowed?

A: Yes. The NPSH (net positive suction head) requirements of centrifugal pumps are normally determined on the basis of handling water at or near normal room temperature. Operating experience in the field has indicated, and a limited number of carefully controlled laboratory tests have confirmed, that pumps handling certain hydrocarbon liquids, or water at significantly higher than room temperatures, will operate satisfactorily with less NPSH available than would be required for cold water.

The thermodynamic properties of boiling liquids at high vapor pressures are such that the vapors produce fewer blockages between the impeller vanes and less reduction in head produced by the impeller. The following table is an approximate guide to how much NPSHR can be reduced as a function of vapor pressure and fluid temperature.

Liquid Vapor Pressure, PSIA
NPSHR Reduction, Feet

Q: We are designing a circular sump for the installation of three sewage pumps. Because of local restrictions on the size, we are concerned the pump intake structure may not provide a good hydraulic design and are considering whether to build and test a model of the intake. How do we select the appropriate rate of flow for the model test?

A: Models involving a free surface are operated using Froude similarity since the flow process is controlled by gravity and inertial forces. The Froude number, representing the ratio of inertial to gravitational forces, can be defined for pump intakes as:

F = u/(gL) 0.5

u = average axial velocity (such as in the suction bell entrance) in ft/sec
g = gravitational acceleration, 32.2 ft/sec 2
L = a characteristic length (usually bell diameter or submergence) in ft.

The choice of the parameter that is used for velocity and length is not critical, but the same parameter must be used for the model and prototype when determining the Froude number. For similarity of flow patterns, the Froude number shall be equal in both the model and prototype, and solving for the velocity in the model will answer your question.

In modeling a pump intake to study the potential formation of vortices, it is important to select a reasonably large geometric scale to minimize viscous and surface tension scale effects and to reproduce the flow pattern in the vicinity of the intake. In addition, the model shall be large enough to allow visual observations of flow patterns, accurate measurements of swirl and velocity distribution, and sufficient dimensional control. Realizing that larger models, though more accurate and reliable, are more expensive, a balancing of these factors is used in selecting a reasonable model scale. However, the scale selection based on vortex similitude considerations is a requirement to avoid scale effects and unreliable test results. ANSI/HI 9.8 Pump Intake Design includes additional information on this subject.

For More Information: