Transformer Insulation Resistance: PI Ratio and DAR
A single megohm reading tells you almost nothing about insulation condition. The polarization index and dielectric absorption ratio — ratios of megohm readings taken at different times — tell you far more.
Insulation resistance testing with a megohmmeter (megger) is one of the oldest and most portable tests in the substation field technician's kit. Apply voltage, read the megohm value, compare to a threshold, pass or fail. The problem is that a single megohm reading is heavily influenced by temperature and by the size of the insulation being tested — a large transformer will read lower than a small one even with identical insulation quality, simply because there is more capacitance and more leakage path surface area. And a reading of several hundred megohms can look healthy on a wet or contaminated winding that would fail under load.
The polarization index (PI) and dielectric absorption ratio (DAR) address these limitations by measuring how the insulation resistance changes with time rather than its absolute value. The ratio of the 10-minute reading to the 1-minute reading (PI) or the 60-second reading to the 30-second reading (DAR) characterizes the insulation's absorption behavior — a property that does not change with equipment size and changes predictably with insulation condition.
The physics behind the ratio
When DC voltage is applied to insulation, several currents flow simultaneously. Capacitive charging current decays rapidly — within a few seconds. Absorption current, driven by dipole alignment in the insulation material, decays more slowly over minutes. And conduction current, which flows through actual leakage paths (moisture, contamination, degraded material), remains roughly constant. In good, dry insulation, the conduction current is very small and the absorption current dominates the early-time reading; as absorption current dies out, the measured resistance rises steadily over the test period. In wet or contaminated insulation, conduction current dominates from the beginning and does not decrease — the resistance stays flat or actually falls.
The PI and DAR ratios capture this behavior. A ratio significantly greater than 1.0 indicates absorption current dominates — characteristic of good insulation. A ratio near 1.0 or below indicates conduction current dominates — moisture, contamination, or extensive degradation.
Polarization index (PI): 10-minute to 1-minute ratio
The PI is the ratio of the 10-minute megohm reading to the 1-minute reading under the same applied voltage. IEEE C57.12.90 and IEEE 43 provide interpretation criteria. For oil-filled transformers and rotating machines:
A PI below 1.0 means resistance fell during the test — a definite problem. A PI of 1.0 to 2.0 is suspect; the insulation is wet, contaminated, or severely degraded and should not be returned to service without further investigation. A PI of 2.0 to 4.0 is acceptable — the insulation is reasonably dry and not heavily contaminated. A PI above 4.0 is good; the insulation is dry and clean. For oil-insulated power transformers, IEEE C57.12.90 sets an acceptable PI minimum of 1.0, but most utilities use 2.0 as their practical threshold because values between 1.0 and 2.0 have poor outcomes in service.
Dielectric absorption ratio (DAR): 60-second to 30-second
The DAR uses a shorter test window — the ratio of the 60-second reading to the 30-second reading. Because the test takes only one minute rather than ten, it is faster and more practical in some situations. It is less sensitive than the PI to subtle insulation degradation because the absorption current changes less over a 30-second span than over a 9-minute span. A DAR above 1.25 is generally considered good; 1.0 to 1.25 is questionable; below 1.0 indicates a problem. The DAR is most useful as a quick screening test; suspect results warrant a full PI measurement.
Temperature correction
Insulation resistance varies dramatically with temperature — roughly doubling for every 10°C drop. This means a transformer measured at 10°C will read twice as high as the same transformer at 20°C, with no change in actual insulation condition. Temperature correction to a standard reference temperature (20°C for power transformers per IEEE C57.12.90) is essential for trending results across different ambient conditions. The PI and DAR ratios are less sensitive to temperature than absolute megohm values because both the numerator and denominator are measured at the same temperature, but temperature correction of the absolute readings is still required for valid trending.
Published correction factors exist for oil-impregnated paper insulation. For dry-type transformers, temperature correction factors differ and the manufacturer's data should be referenced. Test results should always be recorded with the insulation temperature at the time of test so that corrected values can be calculated.
Test connections
Transformer insulation resistance is measured between three winding combinations: HV winding to LV winding plus ground (CH), LV winding to HV winding plus ground (CL), and HV plus LV windings together to ground (CHL). Each of these test configurations isolates different potential problems — the CHL test in particular is sensitive to core insulation and grounding issues that the individual winding tests miss. For three-winding transformers, additional combinations are possible and some utilities test all possible winding pairs.
Residual charge from prior tests must be discharged before each measurement. The standard is to short the windings under test for at least four times the test duration, or 10 minutes minimum, before the next reading. Failure to discharge completely produces falsely elevated readings on subsequent tests.
Using the results
A single PI or DAR reading is most useful at acceptance testing, where there is a clear pass/fail threshold and no prior baseline to compare against. For in-service transformers, trending is far more powerful than any single measurement. A transformer whose PI has dropped from 4.5 to 2.1 over three years is telling a story even if 2.1 is technically above the minimum threshold. The rate of change is often more actionable than the absolute value. Insulation resistance testing should be documented in a test log with date, ambient temperature, insulation temperature, applied voltage, and the timed readings at each interval so that trending can be performed across the life of the transformer.
Southern Switch performs transformer insulation resistance testing including PI ratio, DAR, and temperature-corrected megohm measurements per IEEE C57.12.90. Testing is performed as part of acceptance testing programs and periodic maintenance.
Transformer Field Testing →