Interpret tan delta (dissipation factor) test results for power transformers by first verifying the test mode, voltage and temperature, then comparing like-for-like readings against the transformer’s own baseline. Separate capacitance change from tan delta change, use trend direction rather than a single limit, and state uncertainty explicitly so the interpretation stays within the evidence.
Verify Test Mode, Voltage and Temperature Before Interpretation
The same insulation can produce different tan delta readings depending on how the test is connected. UST, GST and GSTg modes include or exclude different insulation sections, so a reading in one mode cannot be compared with a reading in another. Confirm the test mode, the energised terminals, the guarded terminals and the ground connections, and record them with the result.
Also confirm the test voltage and frequency. Tan delta is measured at a defined test voltage, and the reading can vary with voltage level on non-linear insulation systems. Temperature has the strongest influence: dissipation factor of oil-paper insulation generally rises with temperature and moisture content, so the interpretation must reference the winding or insulation temperature at the time of test.
Note which instrument family took the reading. Different test sets can differ in excitation frequency, voltage waveform and measurement algorithm, and readings from different instrument families are not automatically interchangeable for trend purposes. When an instrument is replaced, run both units on the same test object before the change so the difference between them is documented, or restart the baseline with the new unit.
Compare Like-for-Like Results Across Test Dates
A single tan delta value means little without context. The comparison must use the same test mode, voltage, frequency, temperature range and instrument family as the baseline. If the procedure requires temperature correction, apply the correction method defined for the insulation system and record the reference temperature; otherwise, compare readings taken under similar temperature conditions.
Build the comparison as a table across test dates so the reviewer sees the sequence, not just the latest value. Note any change in test conditions between dates in the same table, because a step caused by a different test mode or instrument is an artefact of the measurement, not a change in the insulation.
Where several similar units exist, a cross-unit comparison adds context: a reading that is consistent with the fleet and the unit’s own history is more reassuring than the same value on an isolated record. Use fleet comparison as supporting evidence, not as a substitute for the unit’s own baseline.
If no baseline exists, the first reading becomes the baseline. In that case, document the test conditions thoroughly, because the value of every future comparison depends on being able to reproduce them. Also record the units and sign convention the instrument uses, so a percentage expressed on a different basis is not mistaken for a change in condition.
Separate Capacitance Change from Tan Delta Change
Most tan delta test sets report capacitance as well as dissipation factor, and the two quantities carry different information. Capacitance reflects the geometry and the effective dielectric path of the insulation. A significant capacitance increase can indicate water ingress or a change in the insulation structure, while a tan delta increase with stable capacitance points more directly to increased dielectric loss in the existing insulation.
Track both quantities in the trend. A pattern where capacitance and tan delta rise together is associated with moisture-related changes in paper-oil insulation. A tan delta rise with stable capacitance may reflect ageing or contamination of the insulating material. These associations guide the investigation; they do not by themselves identify the defect mechanism, and the responsible engineer should combine them with oil analysis and other evidence.
At its simplest, tan delta is the ratio of the resistive (loss) current to the capacitive current through the insulation, expressed as a percentage. The interpretation gains precision when the capacitance component is tracked separately, because the two quantities respond differently to moisture, ageing and contamination. A change in only one of them narrows the possible mechanisms and makes the follow-up test selection more purposeful.
Use Trend Direction Before Relying on a Single Limit
Acceptance criteria for tan delta exist in factory test codes and field test guides, but their applicability depends on the insulation system, the test method and the asset’s history. IEC 60076-1 covers general requirements and test procedures for power transformers, and IEEE C57.12.90 describes factory test methods for liquid-immersed transformers, while IEEE C57.152 addresses diagnostic field testing. Confirm which document and section apply to your transformer and test before using any value as a limit.
For condition assessment, trend direction usually outweighs a single reading. A value that remains stable across several tests at the same conditions is more reassuring than a single reading that sits near a reference band without history. A rising trend, even from a low starting point, justifies closer monitoring and a defined follow-up interval.
Do not transfer an acceptance value from a different asset type, insulation system or test method. The same numerical reading can be unremarkable on one transformer design and significant on another, so the reference must come from the document that applies to your unit and test, and the comparison must be documented in the record.
Phase-to-phase comparison is a useful second dimension of the trend. Phases of the same transformer are expected to behave alike, so a deviation that appears in one phase only, when the mode and conditions are identical, is a stronger signal than a uniform shift across all phases. Compare the phase pattern with the baseline, and treat a phase-specific change as a candidate for targeted follow-up on that phase.
Patterns That May Justify Additional Investigation
Certain patterns raise the priority for investigation: a tan delta that rises consistently with temperature more than expected, a capacitance step between test dates, a difference between phases that did not exist before, or a large difference between similar units. Each pattern changes the follow-up, from a repeat measurement to oil sampling or a full diagnostic programme.
Do not escalate from a single value. Confirm repeatability with a second measurement under controlled conditions, review the test record for setup changes, and compare with the baseline before deciding that the insulation condition has changed. The pattern that survives these checks is the evidence that justifies additional testing such as dissolved gas analysis, moisture-in-paper estimation and, where indicated, frequency response analysis.
Keep the conclusion within the method’s limits. A tan delta trend indicates that the dielectric loss of the measured section is changing; it does not locate the defect within the section, identify the exact mechanism, or predict failure time. The mode-based measurements separate which section is involved, and the complementary tests narrow the mechanism. Writing the interpretation inside these limits is what keeps the assessment defensible.
How to Write the Interpretation Without Overclaiming
The written interpretation should distinguish the measurement from the inference. State what was measured, the test mode, voltage, frequency and temperature, the baseline being compared, and the computed change. Then state what the pattern suggests, with the uncertainty attached, and what follow-up is recommended.
| Test date | Mode | Tan delta | Capacitance | Temperature | Uncertainty flags |
|---|---|---|---|---|---|
| 2025-04-12 | UST | 0.32% | 11.2 nF | 24 °C | Baseline established |
| 2025-10-08 | UST | 0.34% | 11.3 nF | 26 °C | Within temperature range |
| 2026-04-20 | UST | 0.47% | 11.6 nF | 25 °C | Rising trend; capacitance step |
The illustrative record above shows how a trend table with uncertainty flags supports a defensible interpretation: the third row would justify a repeat measurement and oil analysis, while the conclusion would remain limited to what the pattern suggests until the additional evidence arrives.
Finish the report with the asset identification, the instrument and calibration data, the test mode and voltage, the raw and corrected values, the baseline source, the interpretation with its confidence, and the recommended follow-up with timing. A report structured this way allows the next engineer to reproduce the comparison without re-running the test, which is the standard an audit-ready condition record should meet.
Frequently Asked Questions
Does tan delta need temperature correction?
Yes, because dissipation factor of oil-paper insulation is temperature-sensitive. Use the correction method defined by your procedure or standard, record the reference temperature, and compare readings taken under similar conditions so the trend reflects the insulation rather than the weather.
Why track capacitance together with tan delta?
Because the two quantities separate different mechanisms. Capacitance reflects the dielectric path and structure, while tan delta reflects dielectric loss in that path. Tracking both helps distinguish moisture-related changes from ageing and contamination patterns, and makes the interpretation more defensible.
How should the trend be used instead of a single limit?
Compare like-for-like readings over time and judge the direction and rate of change against the transformer’s own baseline. A stable or slowly drifting trend is managed with routine monitoring, while a rising trend with a capacitance step justifies additional investigation even if no single reading has crossed a reference band.
For the framework that places tan delta testing in the wider insulation assessment, see the tan delta testing guide. When you need to qualify a tan delta test set for transformer measurements, review tan delta test set options and request a technical proposal with your test procedure and asset list.