Oil-Filled Transformer Insulation Resistance Drop Before Return to Service
An annual outage is testing a 1500 kVA, 13.8 kV delta–480Y/277 V oil-filled transformer. Last year's 5 kV insulation-resistance test at 24°C gave 4.8 GΩ at 10 minutes and a polarization index of 2.9. This year, with winding temperature at 31°C and 78% relative humidity, a 5 kV test gives 300 MΩ at 30 seconds, 480 MΩ at 60 seconds, and 820 MΩ at 10 minutes. The guard terminal is connected to the transformer tank, and the bushings were cleaned with dry lint-free cloths. Oil sample results are pending. The transformer has a sudden-pressure relay and a conservator. The plant wants to re-energize the transformer tonight because the replacement unit is not available. A second test after 30 minutes gives 330 MΩ, 520 MΩ, and 860 MΩ at the same test intervals. The previous test was done with the same tester and leads.
1. Calculate the polarization index and temperature-corrected insulation resistance. What do the results indicate?
Read the worked answer
Using the 10-minute and 1-minute values from the repeat test, PI = 860 MΩ / 520 MΩ = 1.65. For the first test, PI = 820 / 480 = 1.71. Both are below the commonly used screening value of 2.0 for oil-filled transformer insulation; even against a 1.5 minimum, the absolute resistance and the downward trend are concerning.
Temperature correction is approximate but useful. Insulation resistance roughly doubles for each 10°C decrease in temperature. From 31°C to a 20°C reference, the factor is about 2^(11/10) = 2.14. The corrected 10-minute value is about 860 MΩ × 2.14 ≈ 1.84 GΩ, compared with 4.8 GΩ last year. The PI is not improved by correction, because both the 1-minute and 10-minute values are affected similarly. A low PI indicates moisture or contamination that prevents the absorption current from decaying normally; a low absolute value with a large year-to-year drop points to a real insulation problem rather than normal aging.
2. Can the transformer be energized tonight? Explain the judgement, including the pending oil sample and the repeat test.
Read the worked answer
The transformer should not be energized tonight on the basis of these results. The PI of 1.65 is below the usual 2.0 screening value, the temperature-corrected 10-minute resistance is roughly 1.8 GΩ versus 4.8 GΩ last year, and the repeat test did not show a meaningful recovery. The 78% relative humidity could depress surface resistance, but the guard connection should reduce surface leakage; humidity alone does not explain the large drop and low PI.
The pending oil sample is critical. Moisture, dielectric breakdown voltage, interfacial tension, and dissolved-gas results can confirm whether the problem is wet oil, contamination, or incipient fault. Energizing before those results are reviewed would be a judgement call made without the most relevant diagnostic. The plant should keep the unit de-energized, repeat the insulation test under better conditions if possible, and perform additional electrical tests. If the oil is wet or contaminated, drying or oil treatment is required before return to service.
3. What additional tests and corrective actions should be completed before return to service?
Read the worked answer
First, repeat the insulation-resistance test with the guard terminal verified and with individual windings tested separately: high-to-low, high-to-ground, and low-to-ground. A step-voltage or ramp test can help distinguish moisture from a localized defect, but it should be applied within the transformer's insulation rating. Perform power-factor or dissipation-factor tests on each winding and bushing, excitation current, turns ratio, and winding resistance. Compare all results with previous records and temperature-correct them.
Review the oil results as soon as available. If moisture exceeds limits, process the oil and possibly dry the winding. If dissolved-gas analysis shows arcing or overheating gases, perform a more detailed investigation before energizing. After any corrective work, repeat the insulation and power-factor tests and confirm that the PI has recovered to an acceptable level. Only then should the transformer be returned to service, and the sudden-pressure relay and conservator should be checked for correct operation.