Accepting a power transformer is not a single event but a structured sequence of verifications spanning factory and site. The stakes are asymmetric: a missed dielectric defect or undetected winding movement can remain latent for years before failing catastrophically in service. International practice therefore converges on a two-gate model—Factory Acceptance Testing (FAT) and Site Acceptance Testing (SAT)—governed by a coherent set of standards and bounded by explicit acceptance criteria.

The standards framework. Three normative families dominate. IEC 60076-1 classifies tests into three tiers: routine tests performed on every unit, type tests proving the design on a representative unit, and special tests specified by the purchaser. IEEE C57.12.90 provides the parallel test code for liquid-immersed transformers, distinguishing routine, design, and “other” tests. China’s GB/T 1094 series mirrors this architecture and adds explicit handover test provisions, while GB 50150 governs field acceptance. The practical implication is that a purchasing specification must not merely cite a standard; it must state which tests from each tier are mandatory, witnessed, and subject to which acceptance limits.

Factory acceptance: proving design and workmanship. FAT must be performed on the fully assembled transformer, with bushings, coolers, and tap changers in service configuration—not on a stripped-down test article. The core witnessed set should include winding resistance, ratio and vector-group verification, insulation resistance, no-load and load losses, applied and induced-voltage dielectric tests, and, for units at 72.5 kV and above, partial discharge and swept-frequency response analysis (SFRA). SFRA deserves particular emphasis: it is increasingly treated not as a “special” test but as a routine mechanical-integrity baseline, because it detects core or winding displacement that electrical tests may miss. Dissolved gas analysis of insulating oil is also essential for oil-filled units, establishing the chemical fingerprint against which future site samples will be compared. Table 1 summarizes the principal FAT and SAT obligations.

Table 1. Core acceptance tests and their standard basis

Test	Stage	Primary standard basis	What it verifies
Winding resistance	FAT, SAT	IEEE C57.12.90 §6; GB/T 1094.1	Joint integrity; loose connections
Ratio and vector group	FAT, SAT	IEEE C57.12.90 §5; IEC 60076-1	Correct turns ratio and phase relation at all taps
Insulation resistance / PI	FAT, SAT	IEEE 43; GB 50150	Moisture and contamination baseline
No-load and load losses	FAT	IEEE C57.12.90; IEC 60076-1	Performance against guaranteed values
Applied and induced voltage	FAT	IEC 60076-3; IEEE C57.12.90	Dielectric withstand
Partial discharge	FAT	IEC 60076-3; IEEE C57.12.90	Insulation defects (Class II limits)
SFRA	FAT, SAT	IEEE C57.12.90; IEC 60076-18	Core/winding mechanical integrity
Oil DGA and quality	FAT, SAT	IEC 60599; GB/T 7252	Incipient faults; oil condition

Site acceptance: confirming integrity after transport and installation. SAT repeats a subset of FAT measurements and adds checks that only make sense on site. The transformer must be inspected for shipping damage, impact-recorder data reviewed, and insulation resistance, ratio, and SFRA repeated to compare against factory baselines. Core and frame insulation resistance must be verified, bushing power factor measured, and oil samples retested for moisture and dielectric strength before energization. The principle is continuity: without a factory baseline, a site measurement has no interpretive context.

Acceptance criteria must be numeric. A test without a limit is merely a measurement. Table 2 consolidates representative limits drawn from standards and utility practice. These are voltage-class dependent and must be tightened or relaxed in the purchasing specification, but they illustrate the order of magnitude of acceptance decisions.

Table 2. Representative acceptance limits

Parameter	Normal / accept	Investigate	Reject or remediate	Basis
Winding resistance imbalance	≤ 2% (phase), ≤ 1% (line)	2–4%	> 4%	GB/T 1094.1; GB 50150
Ratio error	≤ 0.1%	0.2–0.5%	> 0.5%	IEC 60076-1; utility practice
Insulation resistance (new unit, 20 °C)	≥ 500 MΩ; PI > 1.5	100–500 MΩ	< 100 MΩ	IEEE 43; GB 50150
SFRA correlation with factory baseline	High correlation across all phases	Local deviation	Gross deviation or phase-to-phase mismatch	IEC 60076-18
Oil dielectric strength	≥ 60 kV/2.5 mm (mineral oil)	50–60 kV	< 50 kV	IEC 60156; GB/T 507

Two further requirements complete the acceptance package. First, witnessing: routine tests should be witnessed by the purchaser or an independent inspector, with written notice before testing and full access to calibrated instrumentation. Second, documentation: the FAT report, type-test evidence (no older than five years for a representative design), calibration certificates, and the SAT report together constitute the evidentiary chain. Acceptance is not granted on the basis of a single passing measurement but on a coherent, comparable, and documented set of results that establishes the transformer’s condition at the moment it enters service.