The missing link in motor condition monitoring
Author - Mike Herring, Baker
Energy costs are a major part of the operating expenditure for any plant or facility, and in almost every case, electric motors are among the largest consumers of energy. Properly implemented monitoring of motor performance, which will help to improve reliability and extend the life of motors, is therefore an excellent investment, as it will reduce overall operating costs.
To be truly effective, condition monitoring must include not only tests like vibration analysis, oil analysis and thermography, which mainly detect mechanical problems, but also a structured testing regime for electrical faults. All too often, other than the basic tests, electrical testing is deemed unnecessary. This is unfortunate, as studies have shown time and again that after bearing failure, electrical winding faults are the most common mode of motor failure. A structured electrical testing regime is, therefore, not simply an optional extra – it’s a vital requirement for achieving plant reliability.
The insulation resistance test – which is often called a Megger test, after the trademarked name of one of the most popular instruments used to perform it – has long been the tool of choice for engineers, and this simple test is often the only electrical test performed on a motor. There’s no doubt that this test has a valid role, but it is simply not capable of detecting all of the faults that can occur within the windings of a motor.
Modern test equipment provides a much more comprehensive and revealing range of tests. Moreover, a modern test instrument typically uses a PC to provide automatic testing and fault diagnosis, removing the onus on the operator to interpret the results. The equipment is able to detect micro arcs and stop the test automatically if they occur. The associated database software stores asset identities along with the test results, so that the results can be trended over time, ideally starting when the asset is new.
Automated testing helps to eliminate operator error and inconsistencies between operators. In addition, the latest testers combine all key insulation tests in a single multi-function instrument, and they can also generate comprehensive and professionally presented reports automatically.
Let us now look in more detail at electrical tests that should form part of an effective motor condition monitoring program.
Static motor tests
Static, or offline, motor tests should ideally be carried out in the sequence shown below.
1. Winding resistance test
This test is used to detect loose connections and dead shorts. Tests must be performed with accurate equipment, capable of measuring down to 0.001 ohms. Resistance values must be corrected to a given temperature, typically 20 °C. The motor temperature must be measured accurately, and where possible, the copper temperature should be used. The motor should not be assumed to be at ambient temperature because if it has been operating recently, this is unlikely to be the case. When the resistance measurements are complete, the difference between the phase-to-phase readings is calculated. A delta of less than 1% is typical for a good winding. Trending the results is useful as, for example, a connection that is gradually working loose will show up as an imbalance that increases over time.
Figure 1: Example of current through a dielectric versus time when excited by a DC voltage
2. Insulation resistance (Megger) test
This test will show if the motor is wet or contaminated, or if it has a dead short to earth. An insulation resistance test cannot, however, confirm that a motor is in good condition, as it does not test the entire insulation system. It is vital to correct the megohm value to 40 ºC – the standards state recommended safe minmimum megohm values at 40 ºC, so unless you are using temperature correction, the values being used for trending over time can vary wildly, purely due to winding temperature changes.
3. Polarization Index (PI) test
The PI test is a 10-minute test that is used to measure/ quantify (interfacial) polarization effects occurring in the ground wall insulation of a motor or generator. The insulation resistance value after the test voltage has been applied continuously for 10 minutes is divided by the value after one minute. There are several types of polarization processes, the influence of which start to appear in a measurement immediately upon the onset of test voltage. This is illustrated in Figure 1, in which it can be seen that even when a non-time varying test source (i.e., a DC voltage) is applied to a dielectric, the resultant total current, IT, through the dielectric is initially time varying.
Absorption current, a.k.a. polarization current, which contributes to the total current measured, is drawn to support polarization processes. It is initially very large and progressively decreases as varying types of polarization processes complete. Typically, after 1 minute of DC voltage application, all polarization processes but interfacial polarization have completed; and after 10 minutes, for many test specimens, interfacial polarization has completed as well. This progressive drop in total measured current, which is a phenomenon mostly influenced by polarization effects, manifests as a progressive increase in an insulation’s measured DC resistance.
A PI test result of two or more is usually considered satisfactory. For some insulation classes, a lower PI value may be acceptable, and individual engineers may also set a different limit, depending on the application and location of the motor under test.
4. DC Step voltage test
This test is typically performed at twice line voltage plus 1000 volts. The voltage is increased in steps – ideally five steps or more – and the leakage current plotted. Good insulation to earth will show a linear plot (see Figure 2) whereas a non-linear plot (see Figure 3) suggests contamination or an insulation weakness at that voltage. A DC step voltage test provides considerably more information than a DC hipot test and is therefore recommended for in-service motors.
Figure 2: Linear leakage current plot
Figure 3: Non-linear leakage current plot
5. DC hipot test
The hipot test simply applies a voltage, measures leakage current and calculates insulation resistance. If the insulation resistance is higher than the accepted minimum, the motor passes. Even if there is an area of damaged insulation, as long as the insulation resistance is higher than the minimum value, the motor will still pass.
6. Surge test
This final test is used to verify the turn-to-turn, coil-to-coil and phase-to-phase insulation condition and is typically performed at twice line voltage plus 1000 volts. It can detect weak insulation both between phases and within one phase, dead shorts, loose connections, and unbalances caused by incorrect winding.
The surge test works by injecting high voltage pulses into one phase at a time, with the other two phases being grounded via the tester. This creates a potential difference between one turn and the next. The pulses produce damped oscillations (ringing transients) in the motor windings. In the absence of winding defects, the wave shape of these oscillations should be very similar for each of the phases.
When motors start and stop, high voltage spikes are generated and, over time, these spikes can damage winding insulation, as can the effects of heat, contamination, and the movement of the windings due to the magnetic forces they experience during motor starting. Of all the tests described, the surge test is the only method of detecting the resulting weakness in turn-to-turn insulation. As studies have shown that 80% of electrical failures start as turn-to-turn weakness, it is clear that surge testing is essential if incipient winding faults are to be detected before they lead to complete failure of the motor.
Figure 4: Failed surge test
Figure 4 shows the result of a surge test on a 415 V motor. The motor had been subjected to severe overheating, but still ran normally. An insulation resistance test at 500 V gave an apparently satisfactory result of 1438 megohms. The surge test, however, failed at 1590 V on Phase 2. Phases 1 and 3 both passed at the full test voltage of 2000 V. Had the motor been in good condition, the waveforms for all three phases would have been almost identical.
Figure 5 shows the point at which Phase 2 failed. The pulse-to-pulse graph plots all pulses that are injected into the three phases and should show close balance between the three phases. The blue spike is the result of an arc within Phase 2. When the arc occurred, this caused the inductance of the phase winding to change momentarily. This can also be seen in Figure 4, where the white waveform is shifted to the left. Very early signs of turn-to-turn insulation weakness can be observed in pulse-to-pulse graphs, even when the surge waveforms show close balance between the phases.
Figure 5: Turn-to-turn arc detected
Why test at high voltages?
Questions are frequently asked about the voltages applied to motors during testing. Why, for example, should a 415 V motor be subjected to a 2000 V test? The answer relates to the large voltage spikes that motors see during starting and stopping. A typical 415 V motor, particularly if it is started direct on line, can see voltage spikes of up to 2000 V. To ensure motor reliability, therefore, testing needs to be carried out at similar voltage levels. International standards, including those from the IEC, IEEE, NEMA and EASA, apply to all of these tests.
Figure 6: Bad rotor bar graph
Dynamic motor testing
A more recent addition to motor testing technology is dynamic, or online testing which involves measuring the voltage and current in each of the motor’s three phases while it is operating in its normal environment. From these measurements, a host of information can be calculated, relating to the power supply, the motor and the driven load. Electrical and mechanical issues can be identified.
Power quality values, including voltage level, unbalance and distortion are determined and compared to industry standards. Poor power quality can lead to overheating within motors, and since heat is one of the biggest enemies of insulation, power quality issues should be identified and, where possible, corrected.
Current level and unbalance are used to determine overall electrical health of the motor, and as an aid to identifying connection issues, overloading, iron saturation and incorrectly wound motors.
Spectrum analysis allows the condition of the rotor bars to be determined (see Figure 6) as well as showing the voltage and current relationship to frequency. Saturation problems, inverter issues and mechanical defects can also be detected.
Torque monitoring can be used to identify a wide range of issues, including transient overloading, pump cavitation, mechanical imbalance and bearing faults. Figure 7 shows the instantaneous torque demands on a 2.2 MW motor at a cement factory over a period of 28 seconds. In such an application, the regular repeated peaks shown in the graph at four second intervals would not be expected. These peaks were, in fact, caused by a damaged tooth on a pinion gear.
Transient analysis plots rms motor voltage and current against time while the motor is starting, and also plots the torque profile during that time. Monitoring all three phases of voltage and current, plus torque, allows power, motor and load issues to be separated.
Figure 7: Torque ripple graph
In summary, dynamic electrical motor testing can be used to identify a wide range of electrical and mechanical issues simply by measuring the three phase voltages and currents at the motor control cabinet. Dynamic testing can also be used to provide comparisons with the results from other technologies, such as vibration analysis. It can be performed on submersible pumps, and motors in restricted areas, simply by having access to the electrical signals in the motor control panel. With other technologies where access to the motor itself is required, no tests would be possible.
Electric motors are the almost indefatigable workhorses of industry and modern motors are exceptionally reliable. Nevertheless, when a motor does fail without warning, the consequences can be costly and disruptive. Condition monitoring provides insurance against such failures, by giving advance warning of developing fault conditions. This gives time to plan the required maintenance or motor replacement. To be truly effective, however, condition monitoring must include ‘the missing link’ – electrical tests. Hopefully, this article has given a useful overview of the options available and the benefits they can be expected to provide.