Transformer Short-Circuit Withstand: What Evidence Should Buyers Request?

Quick Takeaway

• Transformer short-circuit withstand covers heating and mechanical stress from external faults. Check winding current and duration separately from the asymmetrical peak.
• Read the full test report against the project fault duty, and verify that the tested active-part revision covers the transformer being offered.
• Agree which evidence the contract accepts. Record the technical approver and procurement hold before release.

A transformer that has passed a short-circuit test has met the acceptance criteria for the physical design and test cases identified in that report. A program can include several shots and configurations. For procurement, check which current, winding, tap, duration and active-part revision the evidence covers, then compare them with the offered transformer and project duty.

Those four links — duty, report, design, revision — are what this review is built on. Any one of them breaking leaves the withstand claim open, and a passing certificate does not repair a broken link.

What transformer short-circuit withstand covers

The subject is the external fault, sometimes called a through-fault when current passes through the transformer. The fault is not inside the transformer, but one or more windings may carry a contribution to it. The sources and current paths depend on the network and transformer connection. Internal faults such as turn-to-turn or phase-to-earth faults inside the tank call for a separate protection and evidence review.

A through-fault affects the transformer in two distinct ways. The evidence for each uses different quantities.

Thermal duty is resistive heating. The power dissipated in a winding depends on I²R. For a short duration and under an applicable calculation method, the conductor temperature can be treated as rising approximately adiabatically, with little heat escaping. The result depends on the conductor mass, its specific heat, its resistivity at elevated temperature, the initial temperature it started from, and how long the current lasted. The governing quantities are a current and a duration, taken together. Request thermal evidence as a calculation for the stated current, duration and initial temperature under the governing method.

Dynamic duty is electromagnetic force. The current in one winding interacts with the leakage field. For fixed geometry and comparable current distribution, some force estimates scale roughly with the square of instantaneous current. In a typical concentric design, radial forces can push the outer winding outward in hoop tension and the inner winding inward, where buckling is a concern. Axial forces can load the end supports, clamping structures and lead restraints. At fault inception the current can contain a decaying DC component that makes the asymmetrical first peak higher than the symmetrical peak. That first peak is an important mechanical input. The governing stress and structural response also depend on the winding arrangement and fault case. Dynamic evidence can come from a special test or a theoretical evaluation where the governing standard and the contract accept that route.

The standard that frames this split, IEC 60076-5:2006 edition 3.0, addresses the effects of external-short-circuit overcurrents. Its public scope distinguishes calculation procedures used to demonstrate thermal ability from a special test and theoretical evaluation methods for the relevant dynamic effects. Use the applicable edition and project requirements to establish the duty, evidence route and acceptance criteria.

Read the evidence as a statement about a particular design, winding, connection and tap. Check the current and duration for thermal duty and the peak and fault configuration for dynamic duty against the accepted criteria.

Define the external fault duty before requesting evidence

Buyers routinely request “the short-circuit test report” before anyone has written down what the transformer has to survive. The result is a document review with no reference point, in which both sides negotiate over the report’s contents instead of over the project’s requirement. Fix the duty first.

The duty originates in the system study, which is a separate discipline with its own scope. IEC 60909-0:2026 edition 3.0 provides the system-study context for short-circuit current calculation in in-scope 50 Hz and 60 Hz three-phase LV and HV AC systems, with systems at highest voltages of 550 kV or above with long transmission lines excluded. Whether your project adopts that edition, an earlier one, or a utility’s own method is a project decision, not something to infer from a publication date. What matters at the transformer terminal is that the study’s outputs are translated into a duty statement the supplier can answer.

That statement needs to carry, explicitly:

• Reference side and voltage. A fault on the LV terminals and a fault on the HV terminals load different windings in different ways, and a current expressed “at 11 kV” means something different from the same number “at 0.4 kV”.
• Fault type. Three-phase, phase-to-phase and phase-to-earth produce different current magnitudes and different force distributions. State which cases the study found governing and which were checked.
• Network operating state. Record applicable source conditions, generation, parallel transformers and backfeed paths. The highest-current state and the case with the longest applicable clearing time can govern different checks.
• Impedance or X/R basis. The loop X/R affects how quickly the DC offset decays. The point on the voltage wave at fault inception also affects the first asymmetrical peak. Record the assumptions used to derive the reported peak.
• Winding and tap. The connection and tap affect the winding currents and impedance seen by the fault. Name the tap case the study uses and check any other relevant tap cases required by the project.
• Protection clearing basis, including backup. For each applicable primary or backup clearing path, record the total elapsed time from fault inception to current interruption. A backup case may pair a different current with a longer duration. Keep each current and clearing time together as a distinct study case.
• Units and conventions. kA RMS symmetrical, kA peak asymmetrical, seconds, and on which side of the transformer each number is stated.

Separate total bus fault current from the current through the transformer winding. An LV switchboard fed by two parallel transformers and motors can receive current from multiple sources. Use the study’s contribution through the specific winding or ask the system engineer and manufacturer to derive it on the required connection and current basis. An HV bus result also needs its own reference side and conversion before it can be compared with an LV winding value.

Two published utility approaches show how differently this can be handled, and how much of it is contract-specific. The Horizon Power substation transformer specification sets its own study and design basis, verifies the short-circuit design during purchaser review, reserves a batch-test option, describes a detailed design review after placement and before manufacture, separates routine impedance and load-loss measurement from short-circuit withstand testing, and asks for winding-specific withstand quantities in the technical schedule. The CEB Specification 098:2023 separates thermal calculations submitted with the offer from the dynamic test or waiver, the latter resting on a complete report for a similar transformer with corresponding documentation, and requires the proof at offer stage. Both are examples of a buyer who has decided where the duty lives and what proves it — not templates to copy.

Close the handover with the study revision number and the named owner. A duty statement without a revision is impossible to audit six months later when the offer, the design and the report have all moved on.

Read a dynamic short-circuit test report against the specified duty

A full dynamic test report gives direct evidence for the design, conditions and events it actually covers. Read it field by field. A certificate can identify the test and its reported conclusion. The report supplies the configuration, records and observations needed to assess that conclusion against your duty.

Identity and traceability. Report number and revision, issuing laboratory, and the laboratory’s accreditation scope for this type of test. If a certificate is offered, ask for the report behind it and check that the certificate number appears in the report rather than being a separate document with no cross-reference.

Object tested. Record the model, serial number and controlled build or active-part revision. Use them to identify the test object and compare it with the offered design.

Laboratory and date. Record where and when the test occurred. Read the actual standard edition from the report, and check any contractual evidence-age condition or design changes since the test.

Test standard and edition. Name and edition, because procedures and acceptance criteria have moved between editions, and the procurement pass criteria must be the ones your contract names.

Configuration. Connection of the winding under test, tap position, and the state of every other winding — shorted, open, or loaded. This is not a formality: a large three-winding transformer tested with the fault applied to one secondary and the other secondary open is a specific configuration, and a published laboratory account of a 220 kV, 400 MVA offshore transformer describes exactly that arrangement alongside before-and-after impedance behaviour and external observations. That is concrete evidence of one tested configuration, not a statement about the other winding, other taps, or another design.

Current and convention. Compare requested and recorded values, with each applicable phase and convention identified. A symmetrical peak calculated from the sinusoidal RMS current is different from the measured asymmetrical first peak. If the report leaves the basis or phase unclear, request the waveform or a documented clarification before accepting peak coverage.

Duration and sequence. Applied duration for each shot, number of shots, and the sequence of events — pre-test measurements, each current application, observations, post-test measurements, inspection.

Recordings and observations. Ask for the available current traces or oscillograms and recorded observations applicable to the test program. These are the records behind the summary values.

Before and after. Check the measurements and inspections specified by the accepted test method and contract. These can include reactance and other electrical measurements, condition observations or inspection findings, as applicable. The CPRI laboratory supply transformer specification treats routine impedance and load-loss measurement as distinct from the special test duty. For its own equipment, it specifies RMS current, peak basis, time and winding configuration separately, with before-and-after reactance and inspection requirements.

Acceptance basis and signatures. What the laboratory and the purchaser agreed constituted a pass, who signed, and any exceptions or deviations recorded.

Finally, keep two judgments apart. Achieving a test point means the recorded current and duration reached the stated values. The report’s pass decision also depends on the applicable observations, measurements and inspections. Meeting procurement acceptance criteria means the buyer has checked that the test outcome, covered duty and design applicability satisfy the contract for the offered transformer.

Show how the tested design covers the offered transformer

This is the section suppliers most often try to close with a sentence. “Same rating, same voltage ratio, similar impedance” is not a design equivalence argument, and a buyer who accepts it has bought a report about somebody else’s transformer.

Two transformers can share kVA, voltages and measured percent impedance while differing in features that determine mechanical withstand. The impedance is a measured electrical result at stated conditions. It does not identify conductor cross-section, spacer count, end insulation blocks, clamping ring stiffness, lead restraint or material.

What actually changes the outcome, and therefore must be compared:

• Winding arrangement and type. Layer versus disc, helical versus continuous disc, number of discs, transposition scheme, and the radial and axial sequence of windings.
• Conductor geometry and material. Compare the conductor dimensions, material, form and properties used in the force, stiffness and thermal evaluation. See the comparison of copper and aluminium transformer windings for the implications of changing an offered winding.
• Calculated short-circuit stress. Radial hoop and buckling stress, axial compressive and tensile stress between discs, and the end thrust on the clamping structure, with the allowable values they were checked against.
• End and radial supports. The blocks, rings and spacers that carry the axial force and prevent inner-winding buckling. A change here is a change to the load path, which is exactly why it cannot be waved through.
• Lead restraint and bracing. Leads carry fault current in their own magnetic fields and are a classic failure location.
• Clamping and manufacturing. Winding tension control, drying and the retention of clamping pressure in pressboard over time. A drawing does not show this; manufacturing evidence does.
• Known deviations. Any concession, non-conformance or design change recorded against the reference unit.

The CIGRE Science & Engineering CSE 042 guidance is organised around exactly this comparison problem: it distinguishes buyer decisions among design review, manufacture evidence and full-scale test, asks for the actual network sequence impedance or range, relevant faults, tap and winding cases and calculation basis, and its reference-transformer comparison covers application, insulation concept, core and coil construction, radial and axial winding sequence, conductor material and form per winding, winding type, relative duty and stresses, manufacturing processes, clamping and winding support. Its design-review discussion covers radial and axial forces, leads, clamping and manufacturing evidence. The point is not to import its tables, but to copy its discipline: every design feature that carries or generates force is a comparison item.

Run the comparison as a question per difference: does this difference change the force, the stress, or the strength? If it changes any of the three, it needs its own evidence. If it changes none, the supplier should still say so on paper, with reasoning, so the file records why it was judged neutral.

For proprietary drawings, ask for controlled review access or an engineering summary. Record the reference report number, tested and offered active-part revisions, the feature-by-feature comparison and the supplier engineer’s signed conclusion. CEB’s specification, for example, asks the bidder to show the similarity of the reference and offered transformer against Annex B of IEC 60076-5 and to furnish the stated evidence with the offer.

Review theoretical withstand evidence when the contract allows it

Where a full-scale test is not required or not available, the contract may accept theoretical evaluation for the dynamic effects, alongside the thermal calculation that is normally expected in any case. The two are separate deliverables with separate inputs and must not be reviewed as one package.

Thermal calculation. Ask for the exact revision it was performed against, the input current and duration on the stipulated basis, the initial temperature assumption, the conductor material properties used, the calculation method and its basis, and the resulting temperature against the applicable limit. A thermal calculation that cites the right standard but uses a different current basis, or silently assumes a shorter duration than your protection actually delivers, answers a question nobody asked.

Dynamic theoretical evaluation. This is where the review has to be more sceptical, because the output is only as good as the model behind it. Request:

• the exact revision and active-part drawing it applies to;
• the input currents, including the peak, and every assumption used to derive them;
• the calculation method and either its validation or a suitable reference accepted by the contract;
• the computed forces per winding and per region, the resulting stresses, the allowable values, and the safety factors applied;
• the construction and manufacturing assumptions the result depends on — clamping pressure retained, spacer count, drying process, material properties;
• lead, support and clamping checks, not only the winding conductor;
• engineering sign-off by a named, responsible engineer.

A calculation provides engineering evidence for its assumptions and design revision. A test records a physical result, and production verification checks that the actual build matches the design reviewed. Where the standard and contract accept theoretical evaluation, review it on that basis. If the contract requires a special test, record the procedure, notification and acceptance route in the contract file. Any proposed change to that requirement needs formal review and approval.

Keep routine impedance and switchgear ratings in their own roles

Three other pieces of paper sit near this review and are routinely confused with it.

Routine percent impedance is a factory measurement that serves two purposes: it can inform the fault study, and it provides serial-level measured data against the agreed impedance and tolerance. A measured value inside tolerance confirms that electrical parameter at the stated test conditions. It tells you nothing about whether the windings survive a fault — a transformer with excellent impedance conformity can still have inadequate end support. Measured impedance has a third, different role after a fault test, where a before-and-after change in reactance can indicate a change in condition. Interpret it alongside the other observations and inspections under the applicable criteria. The distinction between fault-study input and delivered records is set out in how to read a transformer routine test report before shipment.

BIL and dielectric tests concern insulation withstand. BIL refers to the basic lightning impulse insulation level. Other dielectric evidence can cover switching impulse and applied or induced power-frequency voltage, as applicable. They address electrical stress across insulation distances. External-short-circuit withstand concerns winding heating and electromechanical stress in conductors and their supports. A transformer can pass every dielectric test and still fail mechanically under a through-fault, and vice versa.

Switchgear short-circuit ratings — breaking capacity, making capacity and short-time withstand — describe what the circuit breaker or switchgear can safely interrupt, close onto, and carry for a stated short time. These are device capabilities. Separately, check that the protection and switching equipment can clear the fault within the total duration assumed for each applicable primary or backup path. That is the coordination link. The device ratings do not establish transformer withstand. The rating structure itself is explained in MV switchgear kA ratings: breaking, making and short-time withstand, and the choice of clearing device in circuit breaker versus switch-fuse protection for transformer feeders. Impedance as a fault-study input is covered in transformer impedance explained.

Worked example: a passing reference report leaves two open issues

The following comparison is constructed to show how a report can be genuinely passing and still leave the offer open. Fictional illustration only. Numbers are assumed findings from a project study and document extracts. They are not a rating, a general IEC threshold or a real test result.

Project case: One offered three-phase 11/0.4 kV transformer, offered active-part drawing Rev D. Specified external three-phase fault at LV terminals, principal tap. Project study supplies 12.0 kA symmetrical RMS through the LV winding for the thermal calculation basis of 0.40 s, and 30.0 kA required asymmetrical first peak for the selected dynamic duty, on the same LV current basis.

Contract position: The fictional contract allows a relevant dynamic special-test report or a separately reviewed theoretical evaluation, only if the offered revision and project duty are covered. The selected route needs buyer engineering signoff.

Read the two open issues separately, because they are separate and they are closed by different submissions.

The peak gap. The study requires a 30.0 kA asymmetrical first peak. The report records 28.0 kA. The report’s symmetrical current of 12.4 kA happens to exceed the 12.0 kA thermal basis, and it is tempting to treat that surplus as covering the deficit. It does not, and the reason is physical rather than procedural: the relationship between symmetrical current and first peak in a test is a property of that test circuit’s X/R ratio and the point on the voltage wave at which the fault was applied. There is no fixed multiplier you can carry from one report to another. The peak is its own quantity, compared on its own basis, and 28.0 kA does not cover 30.0 kA.

The applicability gap. The passing report belongs to reference Rev B. The offered Rev D changed the winding end support, which can alter how axial force transfers through the support and clamping structure. Ask for a Rev B-to-Rev D comparison with the effect on force, stress and support strength, or other evidence covering Rev D under the accepted route. The offered revision remains unverified until buyer engineering accepts that coverage.

What is already closed. The thermal calculation, on Rev D, for the stipulated 12.0 kA and 0.40 s, has been accepted by the buyer’s engineer. That closes thermal evidence for this scenario and only this scenario — a different tap, a slower backup stage or a revised study would reopen it. It does not touch the dynamic status at all, and treating a thermal acceptance as general reassurance is one of the most common ways a review stalls in the wrong place.

What the supplier can submit. A reviewed theoretical evaluation under the contract can address the project duty, including the 30.0 kA peak, and the Rev D end support with explicit force and stress checks and a named engineering sign-off. The other route is qualifying test evidence that covers the required conditions and applicable revision. Either way, buyer engineering reviews the evidence and procurement holds release under the stated contract until both gaps are closed.

Turn the short-circuit evidence into procurement hold points

Turn the technical review into procurement hold points with named owners.

Gate 1 — bid review. Before the offer is technically accepted, decide whether the evidence route itself is acceptable: which route the contract permits, what exactly will be submitted, by whom, against which revision, and who on the buyer’s side will review it. A bid that promises “test reports available on request” has not answered this. Record the decision, including any route the buyer accepts in place of a test.

Gate 2 — design approval hold. Before manufacture release, the active-part revision must be frozen and the evidence matched to that frozen revision. This is where the design-comparison document belongs. Procurement does not release the manufacturing order until electrical engineering has accepted the match. Where a batch-test option exists, this is also where the batch rule and its applicability to your units are fixed.

Gate 3 — inspection and test plan and delivery documentation. Where required, record special tests, inspections and witness or hold points in the ITP, with notification and acceptance criteria. Trace the delivered serial to the accepted active-part drawing revision and manufacturing records. If the built revision differs from the accepted evidence, reopen technical applicability and document any approved deviation before release.

The responsibility split matters more than the paperwork. The supplier produces the evidence: reports, calculations, comparisons, manufacturing records, signed and dated, against identified revisions. Buyer electrical engineering judges whether that evidence covers the duty and the offered design and records the acceptance with reasoning. Procurement holds and releases, and does not make technical judgments on the supplier’s behalf. No single party closes a gate alone.

Corrections take one of three forms, and the contract decides which is available: fuller documentation, when the substance exists but the file does not — a missing design comparison, an unsigned calculation, a report without its configuration page; accepted additional verification, such as a reviewed theoretical evaluation under the contract’s evaluation route; or new or repeat testing, when nothing short of a physical result will close the gap. Choosing the cheapest of the three without checking the contract is how a release gets signed against evidence that does not meet the specification.

A reusable request list for the next RFQ:

• the project duty statement, with study revision, reference side, fault types, X/R or asymmetry basis, winding and tap case, current conventions, and total clearing time for each applicable primary or backup path;
• for a test route: the full report behind any certificate, including build revision, configuration, per-phase RMS and peak conventions, durations, recordings, before-and-after measurements and the acceptance basis;
• for a design-similarity argument: reference report number, reference active-part revision, offered revision, and a feature-by-feature comparison with engineering reasoning and a named signatory;
• for a calculation route: exact revision, inputs and assumptions, method and validation, force and stress results against allowable values, construction assumptions and engineering sign-off;
• the proposed ITP hold points, the batch or unit-specific test rule, and the documentation package that will trace as-built serial to accepted revision;
• the supplier evidence owner, the buyer engineering approver and the procurement release authority.

Related reading on the adjacent steps: transformer impedance explained for the fault-study input, how to read a transformer routine test report before shipment for the delivery records, and breaker versus switch-fuse protection for the clearing-time side of the thermal duty.

FAQ

Is a short-circuit test certificate enough to accept an offer?

Request the report behind the certificate. Check the build revision, winding and tap, RMS and peak conventions, durations, before-and-after measurements and acceptance basis. Then verify that the recorded conditions cover your duty and that the tested design applies to the offered revision. A missing report or comparison leaves that part of the evidence open.

Why can a report with a higher RMS current still fail to cover the required peak?

A symmetrical RMS value and an asymmetrical first peak describe different parts of the fault current. The peak depends on the circuit and fault inception, so a higher RMS entry does not establish a higher peak. Thermal acceptance also needs the applicable current, duration and thermal calculation. Compare the report’s recorded asymmetrical peak with the requirement on the same basis.

Can a thermal calculation take the place of a dynamic test?

A thermal calculation addresses winding heating under the stated current and duration. Dynamic evidence addresses mechanical forces, deformation and support strength. Use the dynamic test or theoretical evaluation route the contract accepts. Where a special test is required, close it through the agreed test and acceptance procedure.

Does a transformer with higher impedance withstand faults better?

Under otherwise comparable network and voltage conditions, higher transformer impedance can reduce the prospective current contributed through that transformer to a downstream fault. Winding geometry, materials, support and clamping still determine how the offered design responds. Treat impedance as a fault-study input and measured electrical check. Verify withstand for the actual duty and active-part revision.

What if the supplier’s design details are proprietary?

Ask for controlled access under review, or for a reviewed engineering summary that names the reference report number, the reference active-part revision, the offered revision, and a feature-by-feature comparison covering winding arrangement, conductor geometry and material, supports, clamping and lead restraint, with the supplier’s engineering conclusion and a named signatory. A refusal to provide any comparable information is itself a finding. What you cannot do is accept “same rating, same impedance, therefore equivalent” as a substitute for the comparison.