The nameplate information for the 2B circulating water pump showed 4,000 V, 3,000 HP, 409 FLA, and 355 rpm. It had 240 slots and 196 bars. While this is not normal information to include on the nameplate, especially on older motors, it is good information for the specialist to have. It can be obtained from the motor shop or the OEM.

During routine testing using the MCEMax, the senior CBM specialist came across the following data: The in-rush test showed severe current fluctuations, which are highly abnormal for a circulating water pump (it could be normal for an application with frequent load changes).

An in-rush start-up tells a lot about the motor and the load. The next step was to look for data taken before this 2010 test. The in-rush test from 2008 showed a smooth current with no variations, ideal for a circulating water pump.

A demodulated spectrum can help identify electrical and mechanical faults. The demodulation from 2008 showed an overall noise level well below .1 dB. There was a lot of broadband energy, which is not unusual.

The 2010 demodulated spectrum showed a 10x increase in the noise level, with some peaks jumping out of the spectrum. The next step was to run the math and back-calculate it to some type of pump information, such as a blade pass frequency. Regardless of the cause of the peaks, a 10x increase in noise level is a big deal.

The technician included a number of variable tests that provided important data, such as current signature analysis. The test showed a lot of load changes occurring rapidly, throwing frequencies into different bands, causing elevated energy around the spectral peak, or smearing of the spectrum. It could also be an indication that they are using the wrong probes, but that was not the case here. This is a definite red flag.

Looking back at 2008, the current signature analysis shows a low noise level and healthy rotor bars.

Another great tool is a sister motor that fulfills the same function, allowing a comparison test to be performed. Often, there is no historical data, but it was available in this case. The other pump showed a normal current at 400 amps.

The culprit was the inlet valve, the part that is placed in a river or basin to pull water into the pump.

As you can see, the inlet bell was broken. That kind of crack can create turbulence in flow, which the diffuser may not be able to correct. This would result in differential pressures that cause fluctuations.

Where did the broken parts go? Are they inside the pump? This level of damage needs to be acted upon quickly.

The pump was repaired. It is always a good idea to perform a new test after repair to get a new baseline, as well as to make sure the repair was effective. The in-rush start-up test showed an amperage jump back to 420 (when the pump was not moving fluid effectively, it did not draw as much power). The current was smooth, with only minor fluctuations.

The demodulated spectrum showed an overall noise level well below .1 dB, which was the 2008 level.

The cost of this failure would have been substantial if it had occurred during peak season, which is June, July, and August, when people use the most electric power.