Module 2 — Factory acceptance testing

Self-study notes. This page is static reading. It does not ask you to sign in. It does not certify a product, a factory, or a person. Use the edition of each standard named in the contract. Questions: hello@transformerpath.com.

Module 1 fixed the rating, the vector group, and the tap changer. This module is about the factory tests that show whether this serial number meets that rating. The words routine, type, and special are different promises. An invitation that says "full factory tests" does not say which list applies.

Three lists, not one stamp

IEC 60076-1 divides tests into routine tests, done on every unit, type tests, done to prove a design, and special tests, done only when the contract asks. IEEE C57.12.00 and the test code IEEE C57.12.90 make the same kind of split. Name the document, then name any test you want on this serial number that the document would otherwise treat as a type test or a special test.

Routine measurements on an IEC unit include winding resistance, voltage ratio and phase displacement, no-load loss and no-load current, load loss and short-circuit impedance, and the dielectric routine tests in IEC 60076-3. Resistance is measured with the oil temperature recorded, because copper resistance changes with temperature, and it is the figure used to correct load loss to the reference temperature from Module 1. Ratio and phase displacement confirm the turns and the vector group on the taps that will be used. A ratio outside the IEC 60076-1 tolerance, or the IEEE limit you named, is a reason not to ship.

No-load loss and no-load current are measured at rated voltage and rated frequency. Load loss and short-circuit impedance are measured with one winding shorted, at rated current, or at a reduced current corrected to rated current by the method in the test code. The report has to show the correction to the reference temperature, not only the raw meter reading. The tolerance table from Module 1 is the contractual reference unless the contract replaced it.

A temperature-rise test, a sound-level test, and a short-circuit withstand test are the items most often assumed to be included and most often not included. Temperature rise is a type test under IEC 60076-2. Sound level under IEC 60076-10 is a type or special test unless the contract makes it routine. Short-circuit withstand under IEC 60076-5 is a special test. A calculation of short-circuit ability is not a test. If you need the test, the contract has to name it.

Dielectric tests have an order

IEC 60076-3 covers the separate-source AC test, the induced-voltage test, lightning impulse, and switching impulse where it applies. IEEE C57.12.90 is the corresponding IEEE code. Partial-discharge measurement during the induced test is routine above the voltage threshold in the IEC edition you named. The acceptance level is the level in that edition, or a tighter level the contract wrote down.

IEC 60076-3:2013 sorts the tests by Um. Check Table 1 of the edition in the contract before you treat a sister unit's test as covering this serial number. In that edition, lightning impulse is a type test when Um is 72.5 kV or below, and a routine test when Um is above 72.5 kV. Above 170 kV the chopped-wave lightning impulse is the routine impulse test, and a full-wave test still applies to line terminals at or below 170 kV on the same transformer. Switching impulse levels are defined for Um above 72.5 kV. Switching impulse is a special test up to 170 kV and a routine test above 170 kV. A chopped wave on a lower-voltage unit is a special test unless the contract asks for it.

When several dielectric tests are performed on the same unit, IEC 60076-3:2013 puts them in this order: lightning-impulse tests; switching impulse; the separate-source applied-voltage test; the line-terminal AC withstand test, when that test applies; the induced-voltage withstand test; and the induced-voltage test with partial-discharge measurement. By agreement, switching impulse may be done before lightning impulse. When the partial-discharge induced test is performed, it is last. Temperature rise is not in that sequence. It is a thermal type test under IEC 60076-2. Short-circuit withstand is not in that sequence. It is a special test under IEC 60076-5. A list that runs temperature rise, then impulse, then short circuit is not the dielectric order.

What to watch on an impulse test

A witness does not repeat the measurement. A witness checks that the transformer on the floor is the serial number in the contract, that the tap is the tap in the procedure, that the instruments are in calibration, and that the connection matches the diagram. For a lightning-impulse test, the witness checks that the crest voltage and the waveform meet IEC 60076-3 or IEEE C57.12.90, and that a comparison between the reduced wave and the full wave was kept.

The comparison is the discriminator. A collapse, a growing difference between the reduced-wave and full-wave current oscillograms, or a chop that was not specified, is the thing to stop on. A single "pass" stamp without the oscillograms is not a record you can defend later. Each line terminal is tested in the connection the standard requires, with the tap position recorded. The neutral, if it is brought out and the insulation level applies to it, is a separate test. Do not treat a test on one phase as a test of the other two unless the procedure says the connection covers them.

What to watch on a short-circuit test

IEC 60076-5 is the withstand standard. The test, when the contract requires it, demonstrates the ability to withstand the dynamic effects of short circuit. The current is set from the measured impedance, not from a catalogue percent. The test includes the asymmetrical peak, which depends on the instant of closure and on the ratio of reactance to resistance. The laboratory record should show the current traces, the tap, and the number of tests the standard or the contract requires.

After the shots, the short-circuit reactance is measured again. IEC 60076-5 limits how far that reactance may move, and the limit depends on the winding shape. Use the limit in the edition you named. Do not substitute a single percent remembered from a slide. The other evidence is the visual and diagnostic checks the standard names, which can include an inspection of the active part and, for a liquid-filled unit, a look at the oil. A unit that still has the right ratio and the right reactance change can still have displaced clamping. The standard's pass criteria are the criteria. A quiet test bay is not one of them.

Leave with records the site can open

The report should contain the serial number, the standards, the measured values, the guarantees, the corrections, the calibration status the quality system requires, and the connection diagram for the dielectric tests. A one-page pass stamp without numbers cannot be compared with site results.

Ask for a frequency-response measurement to IEC 60076-18 on the taps and the connections the site will repeat, as an electronic record and not only as a picture. IEEE C57.149 is the guide used to compare those traces. The factory trace is the baseline for Module 3. A trace taken for the first time after a transport incident has no baseline. On-load tap-changer operation across the range, the ratio and polarity of bushing current transformers, and the rotation of cooler fans belong on the factory schedule as well. They are a common source of delay on site when they were skipped. Liquid filled at the factory should meet the unused-liquid specification: mineral oil to IEC 60296, natural ester to IEC 62770, or synthetic ester to IEC 61099. Breakdown voltage is measured by IEC 60156.

Read the report against the guarantees before the unit is released for transport. A loss overrun, an impedance outside tolerance, or a dielectric test that was repeated after a repair needs a written disposition. The report should say what was corrected. Copy the frequency-response file and the ratio sheet in a format the site instrument can open.

Check your reading

Is a short-circuit withstand test a routine test under IEC 60076-1?
What is the dielectric order in IEC 60076-3:2013 when several of those tests are done on one unit, and which of them is last when partial discharge is measured?
What makes an impulse record usable after you leave the factory?
Why does the short-circuit current have to be calculated from the measured impedance?
Which document defines the frequency-response measurement you need before shipment?

Answers. Short-circuit withstand is a special test under IEC 60076-5, not a routine test. The dielectric order is lightning impulse, then switching impulse, then the applied-voltage test, then the line-terminal AC test when it applies, then the induced-voltage withstand test, then the induced test with partial-discharge measurement. That last test stays last. Switching impulse may be moved ahead of lightning impulse by agreement. Temperature rise and short circuit are outside that order. An impulse record needs the crest, the waveform, and the comparison of reduced and full waves, tied to the serial number and the tap. The test current comes from the measured impedance because the catalogue percent is not the tested value. The measurement standard for the baseline trace is IEC 60076-18.