Motor insulation degrades through thermal cycling, moisture ingress, contamination, mechanical vibration and, on variable-speed drives, through the additional stress that a fast-switching supply imposes on the winding. The tests are used to detect that degradation before it becomes a failure.

What insulation testing proves on a motor

An insulation resistance measurement establishes that a leakage path of a given resistance exists between the winding and earth at the applied voltage. It is a bulk measurement, so it responds to the general condition of the insulation rather than to a localised defect.

A timed absorption reading adds information about the condition of the insulation beyond the absolute resistance. The behaviour of the resistance over several minutes responds to moisture and to the state of the insulation system, which is why a measurement taken at one minute is less informative than a record of the rise.

Dielectric loss and partial discharge measurements add sensitivity to conditions that the resistance measurement misses, particularly on form-wound machines with a graded insulation system. Those measurements are more expensive and are normally reserved for machines whose failure matters enough to justify them.

Test selection by motor size and criticality

The scope should follow the consequence of failure rather than the size of the machine alone, although the two usually correlate. A small motor driving a critical pump is more important than a large motor on a duplicated duty.

Offline versus online assessment

Offline testing requires the machine to be isolated and the supply earthed. It produces a measurement under defined conditions and it can be performed at a defined test voltage, which makes the result comparable with the standard’s acceptance values.

Online assessment relies on permanently installed instrumentation, typically insulation monitoring or partial discharge sensors. It produces a continuous or frequent indication without stopping the machine, at the cost of a less controlled measurement and a result that is compared against a trend rather than against an absolute limit.

For a critical machine the two are complementary. The online system identifies that something has changed, and an offline test performed at the next opportunity establishes what the change is and how severe it is. The online system’s alarm threshold should be set so that it triggers an investigation rather than an immediate shutdown.

Reading results against winding type

Random-wound and form-wound machines are different test objects. A form-wound winding has a manufactured insulation system with defined grading and a defined insulation class, and its behaviour under test can be compared against the standard for that construction. A random-wound winding has a different insulation distribution and its acceptance values come from a different basis.

The temperature dependence of the measurement is the same in principle for both, but the magnitude differs with the insulation materials and the construction. Comparing a measurement on a form-wound machine with one on a random-wound machine of similar rating produces a difference that describes the construction rather than the condition.

The comparison population therefore has to be built from machines of the same type and, ideally, the same manufacturer and duty. Where the population is large enough, the behaviour of the group becomes the reference and an individual machine that departs from it is the finding.

Moisture, contamination and ageing

Moisture is the commonest cause of a falling insulation resistance in a plant environment. It enters through seals, through breathing as the machine cools, and through wash-down and process water. Its effect on the measurement is immediate and large, which makes the method sensitive to it.

Contamination acts more slowly and is more often found in machines in dirty or conductive environments. Dust, oil mist and conductive process material accumulate on the winding and provide leakage paths across the surface. The effect is similar to moisture in the resistance measurement but it does not disappear when the machine is dried.

Ageing acts slowly on the insulation itself. Thermal cycling causes the insulation to become brittle and to lose mechanical properties, and the electrical measurements respond later than the mechanical condition changes. That lag is why a machine with a clean insulation resistance can still be close to a failure, and why vibration and temperature monitoring complement the electrical tests.

When testing justifies a production stoppage

The justification is a comparison between the cost of the stoppage and the consequence of failure. A planned stoppage costs the production lost during the window and the labour to perform the work. An unplanned failure costs the same production lost, plus the damage to the machine, the collateral damage to the driven equipment, the emergency repair premium and, on a critical process, the cost of the lost batch.

Where the unplanned consequence is larger, the planned stoppage is justified. Where the machine can be replaced or bypassed quickly and the process can tolerate the interruption, testing can be scheduled into an existing window rather than justifying one of its own.

The decision should also account for whether the test will change anything. A test whose result would not alter the plan is not worth the stoppage. A test performed on a machine that is due for replacement within the planning horizon, with no prospect of an intervention, produces information that cannot be used.

Trending across a motor fleet

A fleet trend is more informative than a single measurement, because it establishes what normal looks like for the plant. Machines of the same type and duty normally show a similar baseline and a similar rate of change, and a machine that departs from that behaviour is the one to examine.

The trend requires consistency in the measurement conditions. The test voltage, the connection arrangement, the duration of the reading and the machine temperature all affect the result, and a programme that varies them produces a trend that describes the variation in method. Recording those conditions with every measurement is what makes the fleet comparison possible.

The trend also supports the interval decision. A population that is stable at a one-year interval does not need a shorter one, and a population that shows movement justifies examining the environment or the duty rather than simply shortening the interval for everyone.

Documentation for the maintenance file

The record should identify the machine, its rating, its winding type and its duty, and should state the test voltage, the duration, the values recorded at each time, the temperature and the connection arrangement used.

It should carry the comparison against the machine’s previous results and against the fleet, and it should state any follow-up action with its trigger. Where an online system is installed, the record should note its alarm settings and the offline measurement that confirmed or dismissed the alarm.

The requirements for rotating machine insulation testing are set out in IEC 60034-1, and the framework for field testing high-voltage equipment is published as IEC 60060-3. Instrument method documentation for the insulation resistance measurement is published by suppliers such as Hioki and Fluke, machine insulation practice is coordinated through CIGRE study committees, and the research background is published by EPRI. The instrument range is grouped on the generator detection testing hub and on the electrical resistance testing hub.