Insulation problems are to blame for around 30% or more of motor failures – meaning that electrical motor testing should be a larger part of every condition monitoring regime. Mike Teska, Product Line Manager at SKF, explains.
Motors are being asked to work with less margin and cost these days, so it’s no surprise when they occasionally fail. Around 40% of these failures are caused by problems with the bearings. For this reason, it makes perfect sense that many companies use predictive maintenance (PdM) techniques to keep a constant check on bearing health.
Techniques such as vibration monitoring and thermal monitoring help to spot problems at the earliest opportunity, preventing small problems turning into large ones: better to prevent motor bearing failure than be forced to fix it later.
A further 40% of motor failures – even higher for motors operating above 4kV – are caused by insulation problems such as coil windings in the motor or loose connections. Despite this, PdM to detect insulation weakness is not as commonly used. A ground wall insulation test using a ‘megger’ test is quite common, but is not a very good or complete test.
The problems caused by insulation weakness – such as catastrophic motor failure, or even plant fires – can be just as serious as those caused by worn bearings or overheating. For this reason, it is important to have a way of assessing insulation strength – and take action in good time.
Motor Insulation
There are two types of insulation in an electric motor, groundwall and winding. Groundwall insulation is found between the motor stator and the electrical windings. The insulation strength of new groundwall insulation is very high, often 40 times operating voltage. Winding insulation is the thin insulation on the wires used in the motor windings. The insulation strength of new winding insulation is about 15 times operating voltage.
Most motor insulation failures start as winding insulation failures since that insulation strength is so much weaker. When a winding insulation failure occurs, the motor can fail quickly, often becoming so hot as to also damage the groundwall insulation and create a groundwall insulation failure.
Automatic testing
Automated motor insulation testing, using a device like the SKF Baker AWA-IV, makes insulation testing “push button” easy, helps to remove operator error and inconsistency.
Static insulation testing is carried out with the motor disconnected from the power supply, and is typically performed from the motor control cabinet. Testing from the MCC also allows detection of electrical faults outside the motor itself – such as in junction boxes or feeder cables.
The typical tests are:
• Coil resistance tests: this low voltage test can find shorted coils, poor connections, and rewind errors such as the wrong number of turns or wrong wire material
• Meg-ohm test: this test is run below the operating voltage of the motor and shows if the motor groundwall insulation has failed, or is wet or contaminated.
• Polarization Index (PI) test: this is a 10-minute test of the ground wall insulation. The test voltage is held steady for 10 minutes, and the ratio of the insulation resistance at 10 minutes and a 1 minute is calculated. Strong groundwall insulation will have a PI value greater than 2.
• DC step voltage test: in this groundwall insulation test, the voltage begins at or below the operating voltage and then is increased in several steps to double the line voltage, plus 1000V. Voltage is increased in five or more steps – dwelling at each voltage step for 1 minute. Good insulation will show a linear increase of current with each voltage step, while insulation weakness is revealed by non-linearity in the current increases.
• Hipot test: This is the groundwall insulation resistance in meg ohms at 2 x line voltage + 1000V.
The preceding tests have all tested the groundwall insulation, but the winding insulation is the most prone to failure. The surge test focuses on the winding insulation.
• Surge test: determines turn-to-turn, coil-to-coil and phase-to-phase insulation condition. It injects fast rise time voltage pulses into each phase and analyses the resultant sine waves. The surge test results can identify winding shorts, or high resistance connections or imbalances. If the waveform shifts frequency during the surge test, this is an indication of winding insulation weakness. SKF Static Motor Analyzers automatically analyse the surge waveform.
Static test results are clear and unambiguous, and produce a specific result that requires little interpretation.
As well as giving a clear output, these results can be trended over time. This allows operators or maintenance managers to assess the progress of a potential condition over time. For instance, if a step voltage shows increasing non-linearity, this suggests a weakening of the groundwall insulation.
Critical effect
These tests can have a critical effect on productivity and profits. At a leading pulp and paper company, technicians began using SKF Baker AWA-IV testers to identify problems in about 800 motor systems.
Among the many problems that it found – and solved – were:
• Blown holes in insulating boots that covered cable lugs in junction boxes (identified via step voltage test);
• A bad lug connection in the motor junction box (found after failing resistance test);
• A stator coil turn-to-turn short on a booster fan (identified by surge test);
• A cable shorted to ground in starter, and a pinhole in the cable (found after failing surge test and step voltage test).
In all, the company cut its annual motor costs by nearly one-third after it started using SKF Baker testing rigs.
Importance of testing
A number of other real-life examples illustrate the financial benefit of assessing insulation strength properly.
A steel mill in Australiahad a 6.6kV pump motor that required maintenance – including a rewind. The motor was rewound at a motor shop, then transferred to a second facility 800km away, for vacuum pressure impregnation (VPI). A surge test indicated that insulation strength was still not right. The motor was then put onto a test stand and run. Motor currents and vibration tests were OK so – because the mill was in a hurry to resume production – the motor was put back into service. It failed catastrophically after three days, causing a fire. The reason was traced to the windings – which had probably been damaged in transit to or from the VPI facility. The surge test picked up on the problem, but had been discounted.
This shows that tests for insulation strength are ‘real’ – and are ignored at the user’s peril. If vibration monitoring indicates that a bearing is starting to fail, it is replaced – and the same should happen with insulation.
Dodging mistakes
These are all examples of static testing of motor insulation – where the motor is disconnected from its power source and tested separately. This identifies problems within the motor. However, motors can also be electrically tested in situ – so-called dynamic testing – which can reveal problems in the wider power-machine-load system.
US-based utility company Pacific Gas & Electric was facing a potential $23,000 bill to replace a motor on a 125hp screen refuse pump, which was overheating and drawing excessive current. Rather than simply replacing the motor, the company looked into the reason for the high current – as there were no signs of bearing problems, current imbalance, excessive harmonics or rotor bar problems.
An SKF Baker Exp3000 revealed that the load was running higher than the rated value. Looking back through the maintenance history, the company found that a 15.75in impeller on the pump had been replaced with a 17in impeller. Once the correct-sized impeller was installed, the current returned to normal values.
The testing helped to prevent a costly mistake – as a new motor would also have been overloaded by the oversized impeller.
Also in the energy industry, on-line testing helped prevent a huge loss at a Progress Energy power plant in the US. Technicians were investigating why one of three submerged circulating water pumps was requesting less input power and – as a result – running faster. Using an SKF Exp3000 it captured the torque signature of all three motors, giving a snapshot of the load demands of each one. The pump under investigation had a torque of about 75% that of a healthy pump. The torque band was also too wide, and varied dramatically. The underwater pump was examined by a diver – who discovered that the end bell had fallen off, causing the problem.
The pump was quickly repaired – which helped to maintain output when one of the other pumps failed soon afterwards. The company estimates that it would have lost $3.5m in revenue if output levels had fallen. In each of these cases – and many others – a cheap, simple testing regime has saved huge amounts of money, while ignoring test results can cost a company dearly.
In a recent survey, plant managers were asked what types of condition monitoring they carried out. Vibration and thermal analysis were both commonly used, but insulation testing was not. While it may not be as prevalent, it is just as necessary – especially as insulation weakness causes nearly as many problems as bearing failure.