Electrical Signature Analysis for Final Testing at Repair Shops

1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)
Loading...

We were quite surprised to see a tech tip produced by an organization that stated that you required 50% load or more in order to perform motor current signature analysis (or, in our case, electrical signature analysis) and that you cannot use it for final testing in repair shops. Apparently, we have been doing the ‘impossible’ with all of the ESA/MCSA instruments for the past 20 years, or so.

Figure 1: 30hp no load with broken rotor bars – note long start time uncoupled

The 50% load statement from a few manufacturers in the 1990s, and a few that use ‘low resolution’ spectra now, has to do with the slip of an induction motor. In effect, if you have a four-pole, 1800 RPM induction motor, the motor actually runs close to the nameplate value under full load. Most newer motors will have a nameplate reading about 1785 RPM full load. At no load, uncoupled, a good 4-pole machine will run closer to 1799 RPM (low inertia motor) or, more often, at about 1795 RPM. In low resolution spectra, the peaks associated with motor defects would blend into the line frequency (power harmonic) peaks. However, most have resolutions within 0.001Hz+ at the lower frequency range (ie: <300 Hz) and about 0.01-0.1 Hz at higher frequencies (<5kHz).

Figure 2: 5hp motor with no load unbalance on test bench

When looking for issues such as broken rotor bars on an incoming motor to a repair shop or to verify the condition of the rotor and airgap following a repair, the fact that the motor is not connected to a load or coupling is better as it will eliminate outside influences on your tests. For rotor bar conditions we are looking for Pole Pass Frequency (PPF) sidebands around the line frequency, and harmonics if there are a large number of broken rotor bars with good bars in between, and the number of rotor bars times the RPM +/- 60Hz, 120Hz, 180Hz for static eccentricity (rotor off center) and sidebands of running speed around those for dynamic eccentricity (orbiting rotor). The PPF is calculated as twice the slip frequency, or 2*((1RPM/!800 RPM)*60Hz) = 0.0667 Hz, well within the 0.001Hz+ resolution available in most modern ESA devices available.

Figure 3: 1hp motor with unbalance and bad rotor bars

The peaks surrounding line frequency, and elsewhere in the spectra, are related to the peak voltage or current resulting in the ability to detect issues regardless of load. The peak values are counted down in dB from the peak current providing a value that does not vary significantly (ie: 2-3dB on average) regardless of load. In fact, the -dB (dB down) values developed by Oak Ridge National Labs, specifically when they developed ESA for the nuclear power industry, do not specify load as it was unnecessary.

In addition, one of the problems that does occur with a motor with broken rotor bars is a loss of torque and an increase in slip (multiple bars), making it even simpler to distinguish rotor bar issues. When we have run into instances where the findings are in question, we sometimes reduce the voltage to the motor in order to increase the slip to demonstrate that the PPF follows.

We perform this for quality assurance in repair shops from very small, low voltage motors to very large machines, 13.8kV loaded and unloaded, coupled and uncoupled, with great success, detecting defects before they become issues.  In most cases, however, we find the motors are in very good condition both with mechanical (vibration) and electrical (ESA) testing when following a good repair specification and processes.

At no load, current signature analysis with the correct instrumentation is far superior to vibration for rotor bar and rotor related fault detection.

Like This Post? Share It

Comments are closed.