The test itself is the oldest and cheapest in the transformer test set. A direct voltage is applied between a winding and ground, and the current that flows is measured over time. Because the insulation behaves as a resistance in parallel with capacitance, the current includes a slow absorption component that decays over minutes, and the way that component decays carries information about the condition of the insulation.

Insulation resistance as a screening test

Insulation resistance measures the leakage current path through and across the insulation system at a defined direct voltage. Its value is influenced by the volume resistivity of the oil and the paper, the moisture content of both, the temperature, the surface condition of bushings and insulators, and the geometry of the insulation path. It is a bulk measurement: a large insulation system behaves like many parallel paths, and a problem in one path is diluted by the health of the rest.

That dilution is why the test is properly described as a screening tool. It responds clearly to gross problems such as severe moisture ingress, a contaminated oil system, a bushing with a damaged surface, or an insulation path that has become conductive. It responds weakly to localised defects, which is exactly the class of defect that progresses into a failure.

Where the test earns its place is in the routine programme. It takes minutes, needs one instrument, requires no specialist analysis and produces a number that can be trended. As a first filter that decides whether further investigation is warranted, it is efficient. As a conclusion about insulation health, it is not sufficient on its own.

Polarisation index and the time it needs

The polarisation index is the ratio of the insulation resistance measured at ten minutes to the value measured at one minute. The ten minute duration exists because the slow absorption current in oil-paper insulation takes several minutes to settle; a ratio taken over shorter intervals reflects the faster processes and is less sensitive to moisture.

The physical reason the index responds to moisture is that water increases the conductivity of the insulation and shortens the time constant of the absorption process. A dry insulation system with a long absorption time constant continues to rise between one minute and ten minutes, producing a high ratio. A wet system reaches its steady value quickly, producing a ratio close to one.

These bands are a convention rather than a physics constant. They were established for rotating machine insulation and are applied to transformers by long practice, which is precisely why the same unit’s previous value is more informative than the band itself.

Absolute values versus trend

An absolute insulation resistance value has limited value because its magnitude depends on the size of the insulation system being measured. A large power transformer with a long insulation path can show a lower resistance than a small distribution unit in good condition, simply because more parallel leakage paths are being measured together. A minimum value is a useful floor for screening, not a measure of quality above that floor.

A trend on the same transformer removes the geometry from the comparison. When successive readings are taken at comparable temperature with the same voltage and the same connection configuration, the difference between them reflects a change in the insulation rather than a change in the measurement. A slow fall over several outages is more informative than any single reading, and it is the pattern that justifies an interval change or an investigation.

A trend also distinguishes a step from a slope. A sudden drop between two outages usually points to an event such as moisture ingress after a gasket failure or a bushing problem. A steady decline across several years points to ageing, oxidation or progressive contamination, which calls for a different response.

Temperature and humidity influences

Insulation resistance falls steeply as temperature rises, roughly halving for every ten degrees Celsius of temperature increase in oil-paper systems. A reading taken on a cold morning in winter and a reading taken in summer after a warm night can differ by a factor of two before any physical change has occurred. Without the temperature, the comparison is meaningless.

Humidity acts on the surface. Moisture films on bushings, on insulator surfaces and on the connections provide parallel leakage paths that sit outside the insulation being assessed but are measured as part of it. A test performed in fog or after rain can produce a reading several times lower than a test performed in dry conditions on the same healthy unit.

The practical controls are straightforward. Let the unit reach thermal equilibrium, record the top oil temperature and the ambient conditions, clean and dry the bushing surfaces before connecting, use the same test voltage as the reference measurement, and if the day is wet, either defer the test or note the condition so the reading is not used as a trend point. Where a reading has to be taken in adverse conditions, taking it again in dry conditions within a short interval separates the surface effect from the insulation result.

Interpreting a falling index

A falling index is a signal to check the conditions before checking the transformer. Temperature difference, surface moisture, a shorter stabilisation time than the previous test, a different test voltage and a winding that had not fully discharged from an earlier measurement all depress the index without any change in the insulation.

Once the conditions are matched, a falling index has a short list of physical causes. Moisture ingress from a failing gasket, a breather that has stopped working, or a conservator that has been run at low oil level are the commonest. Oil oxidation products that increase conductivity are the next. A bushing with reduced surface resistance contributes a surface path, and a partial discharge that has started inside the insulation adds a loss mechanism.

The index does not distinguish between those causes, and it should not be asked to. What it does is justify the next test. A falling index that is confirmed by direct water content measurement on an oil sample is a different problem from a falling index with normal water content and abnormal dielectric loss, and the two lead to different interventions.

When a second method is required

A second method is required whenever the decision has consequences. Interval extension, a return to service after a repair, a purchase decision on a used unit or a warranty discussion all justify more than a single resistance measurement.

The most useful companion is direct water content measurement on an oil sample, because it answers the question the polarisation index can only suggest. Dielectric loss measurement adds information about the condition of the paper insulation and the solid insulation system, and it does so at a frequency where moisture and ageing products have a measurable effect. Frequency domain dielectric response testing covers the same ground across a wide frequency range and is used where the moisture assessment has to be quantitative.

For units where the insulation condition is the central question, the tests are read together rather than in sequence. Resistance and polarisation index establish the bulk condition, water content establishes the moisture state, dielectric loss establishes the loss behaviour, and dissolved gas analysis establishes whether there is active degradation. The reference framework for the transformer characteristics these tests are checked against is IEC 60076-1, and the dielectric test definitions that govern how the unit is stressed during testing are in IEC 60076-3.

Recording conditions and instrument data

A useful record contains the test voltage, the one minute and ten minute resistance values, the calculated index, the oil temperature and how it was obtained, the ambient temperature and relative humidity, the connection configuration used and which windings were grounded, the discharge time allowed before the test, and the instrument model and serial number with its calibration status.

The connection configuration deserves particular attention, because insulation resistance can be measured between a winding and ground, between two windings, or with a guard connection that removes surface leakage. Those configurations produce different values on the same transformer, and mixing them in a trend produces a series that looks like a physical change and is not.

Automatic discharge and the guard terminal are the two instrument features that most affect the quality of the record. A tester that discharges the winding and reports the discharge time lets the next measurement start from a known state, and a guard terminal separates surface leakage from the bulk measurement in humid conditions.

What a passing result still does not prove

A passing polarisation index does not prove that the insulation is dry in a quantitative sense, because the index is a ratio and ratios can sit in the acceptable band for insulation systems with different moisture levels. It does not prove that the oil is in good condition, because the oil is one of several insulation paths and a bulk resistance measurement averages it with the paper.

It also does not prove that the solid insulation is undamaged. Localised mechanical damage, a partially degraded pressboard barrier or a developing partial discharge site can all sit inside a healthy bulk resistance reading. Detection of that class of defect depends on partial discharge measurement under IEC 60270 and on dielectric loss testing, not on insulation resistance.

What a pass does prove is valuable in its own right: the insulation system does not have a gross conduction path, the response is consistent with the previous measurement when the conditions are matched, and the asset can proceed to the next stage of testing or remain in service on its current interval. That is a screening result of real operational value, and it is the result the test should be reported as. Manufacturer method documentation from instrument suppliers such as OMICRON and Hioki states the discharge behaviour and guard arrangement of specific instruments, the general electrical safety context for the work is covered by the guidance published by the UK Health and Safety Executive, and the condition assessment framework that ties these measurements together is coordinated through CIGRE study committees.