Transformer Partial Discharge FAT Review: Test Circuit, Background Noise and Release Evidence

A partial-discharge result is easy to misread when a FAT team sees a single number without the circuit, calibration and background conditions that produced it. For an EPC purchaser, a transformer partial discharge FAT review should therefore examine both the measured value and the evidence that makes the measurement meaningful. The test is not a substitute for the contract’s acceptance criteria, and an illustrative value must never be presented as a real factory result.

Agree the test boundary before the witness date

Start with the approved datasheet, purchase specification and applicable standard edition. IEC 60076-3 defines insulation requirements and dielectric tests for power transformers, while IEC 60270:2025 covers charge-based partial-discharge measurement, test circuits, calibration and interference discrimination. The apparatus-specific specification should state whether a PD measurement is required, during which dielectric test sequence, at what prescribed voltage levels and under what acceptance rule. Do not infer these from a generic FAT checklist.

Before the FAT, identify the test object and boundary: winding under test, terminals connected or grounded, neutral condition, unused windings, bushing test taps, measuring impedance, coupling capacitor, detector and the location of the calibration injection point. The test circuit drawing should match the actual arrangement on the floor. Changes to leads, earthing or auxiliary equipment after calibration can affect the measurement and should be assessed before proceeding.

Calibration and sensitivity are part of the result

Charge-based measurements are calibrated for the defined circuit and test object arrangement. The report should identify the calibration source, injected charge, injection location, instrument range, detector settings and the response observed at the measuring system. Calibration establishes a relationship for the specified setup; it does not prove that every possible defect signal will be detected or located equally well.

Ask the test team to record background interference before energization and explain how it was distinguished from discharge activity. A noisy environment may mask low-level signals or create apparent pulses. The team should document the source of interference, any mitigation, the remaining background level and whether the planned measurement sensitivity is still adequate. If the circuit or detector configuration changes, the calibration status must be reconsidered under the applicable procedure.

Interpret the pattern, not one isolated reading

Record the measured quantities in the form required by the procedure, together with voltage, time, phase information where applicable, tap position, temperature and test sequence. Preserve the raw data or instrument export if the contract requires it. The report should state whether readings were stable, intermittent or associated with a known interference event, rather than collapsing everything into one unexplained maximum.

Acceptance is specific to the transformer, contract and test specification. Avoid a universal pass/fail number copied from a different design or voltage class. If a reading approaches a contractual limit, first confirm the circuit, calibration and background; then follow the approved retest or investigation process. Do not change the acceptance criterion after seeing the result. A retest should have a recorded reason, defined setup and disposition of the earlier data.

Close the witness record for release

The signed FAT record should identify the unit serial number, winding and terminals tested, test equipment, calibration chain, procedure revision, ambient conditions, circuit arrangement, measured results, deviations and the authorized disposition. Attach calibration records and setup evidence so that the result remains interpretable after shipment. Link unresolved observations to the punch list, responsible party and release status.

Coordinate the PD portion with the overall transformer FAT witness plan and the approved bushing specification for EPC substations. For an equipment example, see the 35 kV–46 kV power transformer; its project-specific test scope must still be agreed in the purchase documents.

Work through an unexpected indication without losing evidence

When the trace changes during a FAT, pause long enough to preserve the original setup and event record. A purchaser can ask the test team to walk through a disciplined review without diagnosing a defect from one display alone:

1. Mark the event. Record the time, applied voltage, test phase or sequence, detector channel, apparent-charge reading and the operator’s observation. Save the native instrument file when the procedure and contract require it.
2. Check the boundary. Compare actual leads, grounding, terminal connections and measuring impedance with the approved circuit drawing. Note any equipment, routing or earthing change since the last calibration.
3. Reassess background. Compare the event with the pre-test background and contemporaneous interference notes. Ask what changed in the test bay, not just what changed on the instrument screen.
4. Follow the approved decision path. The responsible engineer should determine whether the procedure calls for repeat observation, recalibration, a controlled retest or technical review. The acceptance limit and governing sequence remain those already agreed in the contract.
5. Keep the disposition traceable. Retain the initial observation, any repeat run, the reason for each change and the signed decision. Do not overwrite a questionable trace or describe a retest as if the first indication never occurred.

Apparent charge is a calibrated electrical measurement quantity associated with the specified circuit response; it is not a direct measurement of a defect’s physical size or an automatic locator inside the winding. Interpretation depends on the complete test setup, signal behaviour and the applicable apparatus-specific criteria. If the purchaser needs localization or diagnostic conclusions beyond the contractual FAT measurement, define a separate method and decision authority rather than stretching the acceptance test beyond its scope.

Questions to settle before signing the FAT report

Does a calibrated value prove the entire transformer is equally sensitive?

No. Calibration documents the response for the defined test object and circuit arrangement. The FAT record should identify the injection point and setup, while the technical procedure explains the scope and limits of that calibration. Avoid claims of universal detection coverage unless the approved method establishes them.

Should the purchaser accept a number copied from another project?

Only if the contract, transformer design, test sequence and measurement basis make that criterion applicable. A value from another unit can be a comparison reference, but it does not replace the current project’s approved requirement or the agreed test procedure.

What is the minimum useful record for an event or retest?

Preserve the original indication, voltage and sequence, circuit state, calibration status, background notes, instrument settings, native record where required, reason for retest and authorized disposition. That allows a later reviewer to distinguish a test-system change from a genuine change in the observed response.

FAT release checklist

• Confirm the required standard edition and contract acceptance rule.
• Approve the circuit drawing, test sequence and measuring points before the witness date.
• Record calibration injection details and background interference.
• Preserve detector settings, instrument identification and raw data where required.
• Link deviations and retests to a documented technical disposition.
• Release only after the report, punch list and required evidence are complete.

A sound transformer partial discharge FAT review protects both technical acceptance and schedule by making the result traceable. Send Zisheng Electric the approved insulation specification, test procedure and witness checklist so the engineering team can flag missing boundaries before the unit reaches the test bay.