Some questions catch you by surprise. I certainly was caught off guard when someone at a plant asked: “Are centrifugal pump inlet velocity limits calculated based on the inlet piping or the inlet nozzle?” Quite honestly, I couldn’t ever remember paying any special attention to a pump inlet velocity when sizing lines or selecting pumps.
In many cases, inlet velocity to a pumping system is important. However, that’s because the velocity affects inlet head losses, nozzle sizes or strainer requirements — not because it’s a specific requirement for pump selection.
The concept of a velocity limit on centrifugal pump suction lines has some merit, though. High velocities can lead to pronounced flow patterns at the impeller inlet. This may reduce pump performance and reliability.
So, I decided to explore the issue by looking at installed pumps and reviewing engineering standards, major reference works and publications.
I checked pumps installed during the last five years. Their inlet velocities ranged from 2 ft/sec to 12 ft/sec. The only pumps with known problems had low inlet velocities, not high ones, and were in high-suction-specific-speed services. None of the pumps with velocities greater than 6 ft/sec had any problems.
I then reviewed engineering standards from six major companies. Three had no inlet velocity requirements. Of the three that did, one specified an inlet velocity maximum of 9 ft/sec in the suction line, another referred to the Hydraulic Institute standards, and the last gave a typical range but allowed for an unspecified higher limit if appropriate.
None of the three major reference works on pumps had explicit limits on centrifugal pump inlet velocities.
I also checked roughly 100 publications on pumps and pump inlet piping. All of them evaluated inlet velocity at the piping size upstream of the last reducer to the pump. All but one either ignored the velocity or gave a limit of 8 ft/sec. The exception, a paper from 1975, included an example in which 10.2 ft/sec velocity in the inlet line was acceptable for the specific situation examined; the publication appeared to be extremely conservative in many other aspects.
As my final literature check, I searched the proceedings of the Texas A&M Turbomachinery Laboratory symposia. These are premier pump conferences; all the proceedings are available online and fully searchable. I found no mention of pump inlet velocities as a topic of discussion from 1984 to the present. If it hasn’t been mentioned here at least once, the topic probably isn’t worth worrying about.
The final conclusions for conventional centrifugal pumps are:
• The inlet velocity is evaluated in the dominant line size upstream of the pump, not the pump suction nozzle.
• The velocity, unless it’s extremely high, isn’t very important. Don’t spend a lot of money getting the velocity down.
ANDREW SLOLEY is a contributing editor to Chemical Processing. You can email him at ASloley@putman.net