This reader’s response to Dr. Nelik is from Global Pump Engineering Manager Ben Lindstrom
B arry Erickson’s response to your article (March 2019 Pumps & Systems) was well written. He notes that in many cases, gauges are not located on the pump flanges. This is significant, and not just because of the relation this has to the velocity head correction (and height correction as applicable). When the gauges are away from the pump, friction loss in the pipe, fittings and valves between the gauge can have a substantial impact on the pressure readings, and thus, any flow rate estimations or performance evaluations.
I encountered this with a user who was testing a pump for performance monitoring purposes. The user was having trouble matching the published performance curve.
When I visited the facility, I found that they were measuring discharge pressure not at the discharge of the pump, but after a check valve, two 90-degree elbows, a tee fitting (which the flow was directed at a 90-degree angle through), and a total of about 10 feet of pipe and hose.
I ran the pump we were testing through a full performance test, and the difference was, of course, most visible at high flow rates.
At about 3,000 gallons per minute (gpm) through this 8-inch discharge pipe section, we found the gauge reading at the discharge of the pump (before the check valve, and before accounting for the gauge height relative to pump centerline) to be 88.5 pounds per square inch (psi), and after the aforementioned assembly of fittings, valves, and pipe, we measured 72 psi. That is a 23 percent drop in discharge pressure: 16.5 psi (38 feet since the pumpage was water) of head loss that was unaccounted for as a result of the gauge location along the pipe. Additionally, the gauge on the pipe was also a couple of feet below the centerline of the pump, and this was not accounted for on the customer’s test results, resulting in further error.
Prerotation or preswirl is an unrelated error that can be typically seen at low flow rates at the suction of a centrifugal pump. This is the liquid column in the suction pipe spiraling in the same direction as the impeller rotation.
Aside from the effects this has on pump performance due to the change in the angle at which the water enters the impeller, this also affects the suction gauge reading. Due to the increased velocity head of a rotating water column, compared to a nonrotating column of water moving through the pipe at the same flow rate, the static suction pressure reading can be artificially low. This is, of course, a manifestation of Bernoulli’s equation.
The prerotation, and its negative impact on pump performance, can be minimized by adding a flow straightener at the suction of the pump (many pumps have a small one cast in).
However, in my experience, it is still visible in many cases when you measure the suction pressure right at the suction flange of the pump. Measuring farther away from the suction of the pump, especially when a flow straightener is installed, will reduce prerotation gauge error to a minimum.
We typically cannot directly measure what this prerotation is, but it can be estimated during a performance test. Since the friction loss—and other dynamic losses—in the suction pipe can be approximated to change by the square of the flow rate, we can plot the measured suction gauge reading at each flow rate against the square of the flow rate.
The resulting plot should be a straight line, with the line at zero flow terminating at the value of the static suction head. However, if you have prerotation, you will see the suction pressure decrease more than expected at low and zero flow rates. You can safely correct the suction gauge readings at low flow to match the expected values.
It is worth noting that the effects of this prerotation on gauge readings are, in my experience, measurable but typically small when a pump has even a small flow straightener installed. Without one, prerotation can have a large impact on shutoff head performance of a pump.
Finally, this prerotation or pre-swirl is not to be confused with the suction swirl you address in your December 2013 article in Pumps & Systems. That is a similar phenomenon that results in a different gauge error and with a slightly different solution regarding gauge correction.
Engineering Manager, Global Pump
Dr. Nelik’s Comment:
Thank you so much, Ben. Another good example of the importance of paying attention to details! Your observation regarding the effect of prerotation in the inlet on a pump head (making a pump curve “droop” near shutoff) is an excellent point.
As well documented by A.J. Stepanoff in his “Pump Bible,” inlet prerotation is not caused by the shear effect on fluid by the rotating impeller blade, but indeed by the inlet approach piping conditions—and, as you noted, correcting these inlet issues can also manifest in correcting the pump head curve shape.