Transformer Temperature-Rise FAT: EPC Witness Checks Before Acceptance

At a transformer factory acceptance test (FAT), a temperature-rise result is only useful when the test duty, cooling arrangement, measurement points and calculation basis match the purchase specification. Zisheng Electric works with EPC teams to freeze those review points before the witness date, so a borderline reading can be discussed against an agreed basis rather than an assumption made after the test.

This guide is for EPC electrical engineers, owner’s engineers and quality representatives reviewing a liquid-immersed power transformer. Its purpose is not to replace the contract or the applicable standard. It gives a practical way to check whether the thermal evidence is complete and comparable with the guaranteed design.

What the temperature-rise test demonstrates

Temperature-rise testing evaluates the thermal performance of a transformer under a specified test arrangement and duty. The review should distinguish the measured quantities from values derived by calculation. Depending on the agreed procedure, records can include ambient conditions, top-liquid temperature, winding temperature by resistance method, cooling equipment status, applied losses and test duration. The exact channels and method must follow the project specification and applicable standard.

IEC 60076-2:2011 applies to liquid-immersed transformers, identifies transformer cooling methods, defines temperature-rise limits and gives temperature-rise test methods. It also introduced a winding hot-spot temperature-rise limit in that edition. Do not lift a limit from an old template: confirm the contract edition, transformer insulation system, cooling designation and purchaser requirements before FAT.

Freeze the test basis before witnessing

Start with the approved datasheet, guaranteed loss schedule, cooling designation and FAT procedure. Confirm which operating configuration is being demonstrated: natural cooling, forced-air stages, or another specified arrangement. A test performed with one fan stage available cannot automatically validate a different guaranteed operating mode. The witness plan should show which pumps, fans, relays and temperature indicators are energized, and how their status is recorded.

Also agree the measurement boundary. Identify the ambient sensors, oil-temperature points, winding-resistance connections, current and voltage channels, time stamps and data logger. Ask for instrument identification, calibration status and stated accuracy. Where the method relies on resistance measurements, the procedure should state winding temperature correction assumptions and how the reading is stabilized. Do not infer a pass or fail from a single display value without the calculation sheet.

Decision table for an EPC witness

Read the trend, not only the final number

Ask for the time series that supports the final report. A trace lets the witness see ambient drift, cooler transitions, sensor interruptions and the basis for any steady-state decision. It also helps explain why two apparently similar tests may not be directly comparable. If a fan bank starts late, a sensor is replaced, or the test arrangement changes, the event should be recorded with its time and technical disposition.

For winding temperature derived from resistance, retain the before-and-after resistance readings, winding identity, tap position, measured winding temperature basis and calculation. The report should identify whether the value is measured directly or derived. These are not interchangeable statements. If the project uses a calculated hot-spot value, the model or correction route should be identified and tied to the agreed method; do not silently substitute a generic factor.

Handling a borderline or disputed result

Before calling a result nonconforming, check that both parties are using the same reference conditions, cooling mode, guarantee, correction convention and acceptance rule. Then review measurement uncertainty and the stated decision rule. A result near a contractual limit may require an agreed treatment of uncertainty, repeat measurement or engineering disposition. The rule should be established before the test rather than invented after a value is known.

If a repeat test is proposed, state what will change: test connections, sensor placement, cooling stage, instrument channel, stabilization period or calculation input. Preserve the original data. The report should retain both runs and explain which result governs under the contract. This protects the EPC record while allowing a legitimate technical correction to be made.

FAT records to close before release

• Approved procedure and applicable standard edition are identified.
• Transformer rating, cooling mode and test connections match the approved design.
• Raw readings, time stamps, instrument IDs and calibration evidence are attached.
• Calculated quantities can be traced to recorded inputs and equations.
• Any interruptions, deviations, retests and engineering dispositions are closed.
• The signed FAT report states the contractual acceptance basis and final status.

For related EPC review context, see how to agree transformer FAT measurement uncertainty and the bushing field-test release guide. For product configuration context, review the 2500 kVA oil-immersed transformer and the 35/46 kV power transformer.

Questions EPC teams should settle early

Does the test prove every cooling stage?

Only the configurations actually tested and accepted under the agreed procedure are demonstrated. The witness should compare the test mode with the guaranteed cooling designation and contract scope.

Is one final temperature reading enough?

No. The final report needs traceable raw data, time history and calculation inputs so the result can be reviewed and reproduced.

Can the same numerical limit be used for every transformer?

No. Acceptance depends on the applicable standard edition, design, cooling method, insulation system and project guarantee. Use the approved project criteria.

When planning a transformer FAT, ask Zisheng Electric to align the test procedure, measurement list and acceptance record with your project specification. Share the rating, cooling designation and required standard edition so the engineering team can review the intended test basis before the witness date.