news
News Center
Providing high-quality power to the world
Search productSearch post
news
Providing high-quality power to the world
A transformer test report can contain an accurate number and still leave the purchase order unresolved. The missing item is often the rule used to turn that number into acceptance. When discussing a FAT scope with Zisheng Electric, EPC quality managers should establish that rule before the witness team travels, especially where guaranteed performance has commercial consequences.

This guide is for procurement and quality teams reviewing numerical acceptance results. Its purpose is to separate three questions: what requirement applies, how the result was obtained, and who may release equipment when the evidence is close to a limit. It does not create a new transformer tolerance or replace the agreed test standard.
Build a short acceptance register from the purchase order, approved data sheet and test procedure. Each row should identify a measured quantity, units, operating condition, applicable document revision, limit and decision authority. Keep guaranteed values separate from permitted tolerances and from any commercial loss-evaluation formula. Those items answer different questions and should not be merged into one spreadsheet cell.
For example, a loss value used to compare bids is not automatically the final FAT rejection threshold. A report corrected to a reference condition also needs to be compared with a requirement expressed on the same basis. The reviewer should be able to trace that comparison without reconstructing the contract from email attachments.
The transformer inspection and test plan is the appropriate place to assign review and witness responsibilities. Add the acceptance register as a controlled reference rather than leaving the witness engineer to negotiate the rule beside the test bench.
ILAC G8:09/2019 explains decision rules for conformity statements under ISO/IEC 17025:2017. Where a customer requests such a statement, the specification and rule need to be defined; a rule not inherent in the specification is communicated and agreed with the customer. This is a useful procurement discipline, but it does not mean every manufacturer FAT is an accredited laboratory service.
A decision rule states how uncertainty enters the conformity decision. Simple acceptance uses the specification limit as the acceptance boundary. A guard band can move the acceptance boundary inward. Neither approach should be imposed retrospectively because a particular result is inconvenient. The applicable transformer standard and contract remain controlling.
Ask the laboratory to explain its terminology, uncertainty basis and reporting convention. An instrument accuracy class, a calibration certificate and the uncertainty of a complete test result are not interchangeable deliverables. For a disputed measurement, request the test-specific explanation from the competent laboratory rather than assembling an uncertainty budget from unrelated catalogue figures.
Suppose an illustrative contract sets an upper limit of 100.0 arbitrary units. The reported result is 99.2 units and the expanded uncertainty is 1.5 units. Under an agreed simple-acceptance rule, 99.2 is below 100.0. If the agreed rule instead requires the result plus that expanded uncertainty to be no greater than the limit, the calculation is 99.2 + 1.5 = 100.7, and acceptance is not demonstrated by that rule.
These are hypothetical numbers, not transformer loss data or a Zisheng Electric FAT result. The example shows why the decision rule matters; it does not establish a universal guard band, coverage probability or retest requirement. A value that fails a guarded acceptance criterion is not automatically proof that the equipment’s true value exceeds the specification.

Keep the pack small enough to review before the witness date. The following table is a proposed EPC review tool, not a mandatory standard form.
| Review item | Evidence to request | Question it resolves |
|---|---|---|
| Requirement identity | Approved limit, units, condition and revision | Are both parties comparing the same quantity? |
| Measurement basis | Procedure, configuration and corrections used | Does the result match the contractual basis? |
| Traceability | Instrument identities and relevant calibration records | Can the reported result be traced to its measurement chain? |
| Uncertainty explanation | Test-specific statement and reporting convention | How should a value near the boundary be interpreted? |
| Decision rule | Agreed rule and worked comparison | What supports the acceptance statement? |
| Release authority | Named role and disposition record | Who can close a borderline or disputed result? |
Record missing evidence as an action with an owner and due date. A certificate supplied after shipment may close a document gap, but it cannot retrospectively prove that the correct setup was used. If a measurement is essential to release, make the evidence deadline part of the manufacturing schedule.
First freeze the original result, configuration and report revision. Preserve raw records through the agreed document channel. Then check for a clerical mismatch: units, reference conditions, tap position, arithmetic, rounding or an obsolete limit. Correcting a documented reporting error is different from changing a valid measurement.
If the measurement basis is challenged, ask the laboratory to assess the specific concern. Agree any repeat measurement in writing, including setup, stabilization, reporting, witness arrangements and how both results will be treated. Repeating a test until a favorable number appears creates a weak acceptance record. Keep superseded results traceable and explain the reason for the final disposition.
The procurement team should distinguish technical acceptance from commercial settlement. A negotiated concession may be valid within the contract, but it should be recorded as a concession with the appropriate approval. It should not be disguised as an unchanged test limit or an unconditional technical pass.
For a 35kV–46kV power transformer enquiry, send the guaranteed-performance schedule with the proposed FAT acceptance register. State which measurements require a conformity statement and which only require recorded results. This lets the supplier review evidence needs before committing to the witness date.
Share your transformer data sheet, project test standard, acceptance limits and reporting expectations with Zisheng Electric. The useful outcome is an agreed evidence package and a clear route for borderline results before the transformer reaches the test bay.
Technical reference: ILAC G8:09/2019, Guidelines on Decision Rules and Statements of Conformity. Confirm the applicable standard edition and contractual requirements for the individual order.
A practical EPC checklist for inspecting transformer coating repairs after delivery, including surface preparation, coating compatibility, dry-film thickness, records and release decisions.
View detailsReview transformer temperature-rise FAT evidence with a practical EPC checklist. Align the test duty, cooling mode, measurement channels, calculations, uncertainty and deviation closure before acceptance.
View detailsLearn practical transformer maintenance experience to extend service life. Cover oil‑immersed & dry‑type transformer inspection, DGA oil testing, temperature control, overload management, environment‑adapted checklists for Middle East & Africa pro...
View detailsSubstation local/remote control acceptance for EPC teams: verify command authority, rejected requests, mode transfers and traceable handover evidence.
View details