Flow going backwards through a pump? Then the pump starts rotating in reverse? Where the heck is the check valve to stop the flow? These are all great questions (which I will answer later in this article).
Theodore von Kármán and Robert Knapp were working on these questions in the early 1930's. Most engineering students encounter von Kármán in their mathematics or fluid mechanics studies. Knapp was a lesser known colleague of von Kármán's at the California Institute of Technology. Knapp had a number of research interests, one of which was how pumps behaved when flowing and rotating backwards.
Knapp published several of the earliest papers on pumps under reverse flow. Together with von Kármán they created the Karman-Knapp circle still used today to represent pump behavior in all four quadrants.
Knapp was born in Colorado (my current home – for the last 24 years) but spent most of his professional life in Pasadena, California at CalTech (about an hour from where I was born and grew up). Knapp's data published in 1934 and 1937 is still used by waterhammer engineers around the world. It has been included in AFT Impulse since version 2 in 2002.
In 2018 I was interested in learning more about Knapp's pump research. On one of my frequent visits to California I made a visit to the CalTech University library archives. I made a reservation in advance to look through boxes of research notes from Knapp and von Kármán.
It was difficult not to get sidetracked that day. Especially by von Kármán's boxes. He had correspondence during the WWII years with Roosevelt, Eisenhower, Churchill and Einstein, among others. Personal letters between these great men on hydraulic research at CalTech were in the boxes I was searching. Crazy!
I was not successful in finding research original notes behind Knapp's paper. I did learn that Knapp's pumps were made by Byron-Jackson (now part of Flowserve) and Peerless Pumps (now part of Grundfos). I thus turned to some colleagues I know in the Hydraulic Institute to see if they could find any record of these pumps supplied to CalTech in their archives. That too was not successful.
I ended up authoring a paper for ASME this year with two esteemed colleagues, Tryg Dahl and Dave Rogers, that addresses some of the things I learned along the way. This paper was presented virtually at the PVP conference this summer: Pump Specific Speed And Four Quadrant Data In Waterhammer Simulation – Taking Another Look.
Today the ASME Fluid Engineering Division has an award named the Robert T. Knapp Award.
For those still reading and interested in the opening questions I posed, let's take them one at a time.
First, flow can go backwards through a pump in numerous situations. One example is when a pump is pumping uphill and loses power. Eventually gravity will pull the fluid back downhill which can cause the pump to have reverse flow.
Second, if the reverse flow lasts long enough the pump rotation can reverse direction and start spinning backwards.
Third, not all systems have check valves. In some cases a check valve is not suitable based on service conditions (e.g., an abrasive slurry flow, which will quickly damage any check valve installed) or size (large diameter pipes). Even if a check valve exists, there will usually be some reverse flow that occurs before the check valve can close all the way. This reverse flow is what leads to "check valve slam" which can be quite severe.
These kinds of things are important to fluid engineers who must design for waterhammer. I helped co-author a paper in 2018 on check valve slam. For those interested you can find it here: Surge Transients Due to Check Valve Closure in a Municipal Water Pumping Station along with a related blog I wrote Did you know that New Jersey is a barrier in Spain?.