Has NPSH Thrown You A Curve?

Have you installed a centrifugal pump with available NPSH (NPSHa) equal to or greater than the required NPSH (NPSHr)? And the pump still cavitated!

In the August 2000 issue of Pumps & Systems, Doug Kriebel discusses NPSHa and NPSHr. He states that NPSHa must exceed NPSHr for a pump to operate without cavitation. This only hints at what is needed. In the October 2000 issue of the same magazine a reader, Garr M. Jones, adds the following:

"NPSHr information published by the pump manufacturers is misleading owing to the definition adopted by the Hydraulic Institute decades ago. The engineer must understand that a pump can have NPSHa in excess of NPSHr and still be in a condition of heavy cavitation. The need for a more than adequate NPSHr margin is addressed somewhat in ANSI/HI 9.6.1, but not sufficiently for the average engineer to understand."

One petrochemical company requires a 3 ft. margin of NPSHa over NPSHr.

Mr. Kriebel responds to the letter of Mr. Jones as follows:

"The NPSHr by the pump is determined by the pump manufacturer by testing. We would assume the test data is what is plotted for the published performance curve. The NPSHr is determined by reducing the NPSHa to the pump (by suction valve throttling or by using a vacuum tank) until the differential head of the pump at the tested flow is reduced by 3% due to the induced cavitation.

Therefore, a pump running at an NPSHa equal to the NPSHr is already cavitating! This assumes an accurate test and plot of the data. Were multiple tests run? With repeatable results? How close were the results? What's plotted-max/min/average?"

The above discussion assumes that centrifugal pump curve is operating at a specified design point. Often, however, the operating point is to the right of the design point, because Total Dynamic Head is less than design. This may be due to a safety factor being added to the design calculation. What happens when the operating point is to the right of design? The NPSHr of a centrifugal pump goes up. That is not all. From this system standpoint, with increased flow, there may be more pressure drop on the suction side of the pump, which reduces the NPSHa. Thus the design margin of NPSHa over NPSHr can be pinched from both sides, in the not infrequent case where there is excess Total Dynamic Head specified.

Do you have a pump that operated satisfactorily for a period of time? And then the pump started to cavitate?

What if a pump operates for while and then cavitates? It may be that heat is building up inside the casing. This could result from a throttled discharge line. Mr. Kriebel refers to this as "low flow temperature rise." When heat is added to a liquid that is near its boiling point, vaporization and cavitation can result.

Is there a way to avoid these problems? Selecting a regenerative turbine pump can often do just that. A turbine pump will not destructively cavitate. The NPSH curve for a turbine pump is a straight, horizontal line. NPSHr is usually 3 ft. or 1 ft., depending on design. To learn how this pump may solve your cavitation problems, and obtain system design hints, please contact us at info@cjmichelson.com.

NOTE: The article by Mr. Kriebel is available in the archive section of www.pump-zone.com

© 2000 C.J. Michelson Company, Inc. All Rights Reserved.
20575 Center Ridge Road, Suite 110 • Cleveland, OH 44116 • Ph (440) 331-7181 • Fax (440) 331-7108
www.cjmichelson.com • info@cjmichelson.com