Vertical Pump question
A question here..I have 5 vertical pumpsets in parallel and they all discharge into a common header. If 3 or more pumps are online, the system functions well, however, if only one pump is in operation, that one pump becomes very noisy as if it's cavitating and its discharge pressure becomes almost 0 psi. However, the sump water level is up fairly high..what could be causing this?? Any ideas?
My immediate reaction is that it sounds like the piping system head
curve (with piping providing the resistance to flow) is gradual so that a
single pump is operating at run-out. In other words, the piping is
oversized relative to the flow / capacity performance of a single pump -
or conversely, one pump is undersized relative to the piping system.
If
you have inverter duty rated motors and MCC + IO space, maybe try
VFD(s), otherwise, install some means of discharge
throttling. Alternatively, if your process allows it, raise the
downstream delivery pressure when you expect to have one pump running.
Cavitation in a vertical pump such as this depends on more than just
adequate submergence. It is likely that when one pump is running alone,
you are running out to the end of the curve (run-out as SNORGY pointed
out). At a flow rate this high the Net Positive Suction Head required
could very easily be greater than the NPSH available, even with a full
sump. Do the math. Calculate your NPSH available at the end of the
curve using a realistic value for the vapor pressure of your
water. Compare that to the NSHR required as shown on the test curve (or
book curves) for this pump.
The simplest solution would be to
pinch back on the discharge to limit the flow to a point where you would
not require more suction head than you have available.
I'm taking a couple of guesses here. Let's do a mind experiment.
Let's
assume the pumps are sized correctly as a system. Let's also assume
that your system curve is mostly friction (not a high static head
rise). What would happen if the flow were reduced significantly.
The
system resistance would decrease by the square of the system flow --
that is, the head the pump(s) would see would be reduced.
Let's
say one pump is operating. What would it see? It would see the reduced
system head. What would it do? It would move way out on the
curve. Head would reduce to near zero (your resistance is down from the
reduced to 1/9th it's value -- remember resistance goes as the square
of the flow).
What happens as it goes out on the curve? You
could look at your NPSH requirements, but that may not tell the
story. What you would really like to see is the incipient cavitation
curve. Such curves (NPSH vs. Flow) for incipient cavitation, are often
"U" shaped, where the NPSH head falloff curve is flatter. The bottom of
the "U" is normally around the best efficiency point.
What would
the pump do, then as it's NPSH requirements suddenly got higher? It
would make a lot of noise, head may fall off, and flow drop because the
suction was starved.
Could this be your problem? Take a pump curve and estimate a system curve, and find out.
If
you wanted to fix this problem, you would need to put flow restriction
in the discharge line. This would fool the pump into thinking that the
system resistance just got higher. Yes, flow would be reduced, and
you'd need to turn on a second or third pump sooner, but you'd be better
off.
Run a test. Scrunch down on a system discharge valve. See if the noise goes away.
MORE NEWS