Predictive Maintenance That Works

This series of articles focuses on the nuts and bolts of predictive maintenance (PdM), also known as condition-based maintenance. A well-defined and well-executed PdM program saves time and money by reducing unneeded time-based maintenance tasks and by identifying and fixing problems before they cause equipment failure or plant shutdown. In this article, we begin introducing condition-monitoring techniques commonly in use at power plants. They are arranged alphabetically rather than in order of importance.

Electrical Surge Comparison

Monitoring the condition of an electric motor involves determining the extent of electrical insulation deterioration and failure. Traditional insulation tests have concentrated on the ground wall; a common test is insulation resistance. Less attention is paid to turn-to-turn or phase-to-phase insulation, yet there is evidence that deterioration of this thin film is also a major cause of motor failures.

Insulation failures are typically caused by one or more of the following five factors:

  • Contamination of the winding insulation caused by a chemical deposit.
  • Mechanical vibration of the windings or rotor, which wears the insulation system.
  • Slow deterioration of the windings due to normal thermal aging.
  • Premature failure of the winding, typically due to excessive winding temperatures.
  • Overvoltage spikes.

Surge comparison testing can be used to identify turn-to-turn and phase-to-phase insulation deterioration as well as a reversal or open circuit in the connection of one or more coils or coil groups. This test has been used for years as a quality-control measure in motor manufacturing plants, and development of portable test instruments now allows this testing to be conducted as part of troubleshooting a problem and as a routine PdM tool.

The surge comparison test can also identify excessive wear by applying a transient surge at high frequency to two separate but equal parts of a winding. The resulting voltage waveforms reflected from each part are displayed on an oscilloscope or computer screen. If the windings are identical, the two waveforms will be exactly superimposed on each other and a single trace will appear on the screen. If, however, one of the two winding segments contains a short circuit, or a reversed or open coil, the waveforms will be visibly different.

Once a problem is detected, the technician must determine in which of the segments the problem resides. This can be done by comparing each segment to a third segment and noting which combination produces the same waveform deflections. Generally, shorted or missing turns will cause fairly small differences in waveform amplitude. Misconnections, such as coil reversal or interphase shorts, tend to cause large differences or irregularities in waveform shape. An experienced test operator can gauge the type and severity of a fault.

With the surge comparison method it is often possible to determine the voltage at which turn-to-turn or phase-to-phase conduction begins. If this shorting is near operating voltage, then the motor has a serious insulation fault and should be replaced as soon as possible. If shorting is not detected up to twice operating voltage plus 1,000 V, the winding is considered good and the motor can be returned to service.

Unlike other PdM techniques, surge comparison produces no numbers that must be plotted or trended to identify a problem. Like a high-potential test, this is simply a pass/fail procedure. Spot checking (monitoring once a year or less) can be particularly cost-effective for critical or expensive electrical motors; however, it may also be effective for less-critical or balance-of-plant equipment. Because long-term surge comparison analysis and trending can be used to identify improper motor repair practices, either in-house or at a repair shop, surge testing can be an effective quality-assurance test prior to acceptance of motors that have just been repaired.

Surge comparison testing is a somewhat complex and expensive PdM technique. The test instrument, though very versatile, is moderately expensive, and it requires a trained and experienced operator if you want to obtain the most effective results. Although often used on stator windings of either induction or synchronous machines, the test is equally useful on DC armatures or synchronous field poles.

Most surge comparison test equipment is also capable of performing high-potential tests. The primary short-term economic benefits from this testing do not come from reducing the number of electric motor failures but from identifying problems early enough that maintenance can be efficiently planned and scheduled. These benefits can be substantial. Some companies pay for the investment in less than a year by identifying one or two critical electric motor problems.

Surge comparison testing cannot evaluate one coil by itself; because it is a comparison test, it requires careful repetition to determine the location and severity of an observed fault.

Motor-Current Signature Analysis

Motor-current signature analysis (MCSA) is a nonintrusive method for detecting mechanical and electrical problems in motor-driven rotating equipment. The system was developed by Oak Ridge National Laboratory as part of a study on the effects of aging and service degradation of nuclear plant components.

The basis for MCSA is the recognition that an electric motor driving a mechanical load acts as an efficient, continuously available transducer (the motor can be either AC or DC). The motor senses mechanical-load variations and converts them into electric current variations that are transmitted along the motor power cables.

These current variations, though very small in relation to the average current drawn by the motor, can be monitored and recorded at a convenient location away from the operating equipment. Analysis of these variations can provide an indication of machine condition, which may be trended over time to provide an early warning of machine deterioration or process alteration.

Although MCSA was developed for the specific task of determining the effects of aging and service wear on motor-operated valves used in nuclear plants, it has found application within a much broader range of plant machinery, such as electric motors. Motor-current signals can be obtained remotely using a current transformer, typically located in a motor control center, which may be located several hundred feet from the equipment being monitored. For temporary measurements, the signals can be obtained nonintrusively with a single split-jaw current probe placed on one of the power leads. Because no electrical connections need to be made or broken, the hazard of electrical shock is minimized.

The resulting raw data signal is amplified, filtered, and further processed (using fast Fourier transform digital signal processing techniques) to obtain a baseline signal pattern measurement of the instantaneous load variations within the drive train and ultimate load.

A comparison of motor-current and mechanical-vibration signatures obtained simultaneously from a motor-operated valve, for example, reveals similarities and distinctions. Both spectra contain frequency peaks corresponding to the motor speed, worm-gear tooth meshing, and its harmonics, though the amplitude relationships are different. Subsequent signal patterns can then be compared with the baseline to find faults and problem indicators.

One distinction is that there is a strong spectral component in the motor-current signature that is defined as the slip frequency. This signal is a general characteristic of AC induction motors and reflects the rate at which the spinning armature continually falls behind the rotating electrical field generated by the motor’s field windings. (No such peak appears on DC motor signatures.) Since this motor slip-frequency component is electrical rather than mechanical in origin, it has no vibration counterpart and it is not present in the vibration spectrum.

Tests on motor-operated valves indicate that MCSA is capable of detecting and tracking the progress of stem-packing degradation, incorrect torque-switch settings or varying switch trip points, degraded stem or gear-case lubrication, worm-gear tooth wear, restricted valve stem travel, obstructions in the valve seat area, and disengagement of the motor pinion gear.

Spot checking at intervals shorter than a year is of significant value only after a plant has developed its own baseline data. Once historical files have been developed, spot checking can be cost-effective for critical or expensive machines, as well as for less-critical equipment. It also can be used effectively as a troubleshooting tool. Information can be obtained by monitoring machinery once a month or once a quarter.

Periodic MCSA can provide a subtle indication of bearing, packing, coupling, or gear wear, allowing personnel to project acceptable machine performance into the future. Advance notice of developing problems gives technicians time to repair a component during normal, planned machine shutdowns, rather than allowing a serious machine failure to cause a plant forced outage. Because problems are detected when they are relatively minor, they are usually less expensive to repair.

The complexity of MCSA stems in large part from the relatively subjective nature of interpreting the spectra and the limited number of industry-wide historical and comparative spectra available for specific applications. In the past several years the technique has been simplified by several vibration data collector/analyzer manufacturers. This has improved the technique from a data-collection and data-analysis standpoint and significantly increased the amount of practical field experience.

The technique’s nonintrusive nature makes it particularly useful. Measurements can be taken without making or breaking electrical connections and without shutting down or opening up machinery. This eliminates equipment downtime for inspection and improves personnel safety.

In addition, because readings can be taken remotely, this technique can be more conveniently and safely performed on large, high-speed, or otherwise hazardous machines.

For those companies willing to commit resources to this PdM tool, the payback appears as attractive as for vibration analysis. And, for companies with limited budgets, there are several service companies that will perform motor current signature analysis on a contract basis.

More Coming

In the next segment of “Predictive Maintenance That Works,” we’ll continue our discussion of specific condition-monitoring techniques used at power plants and look at why each should be a part of your PdM program.

Dr. Robert Peltier, PE, is POWER’s editor-in-chief.