The SFRA test is one of the few field tests that can determine if there is a problem with the geometry of the windings. The test set applies a 10v p-p sine wave to a selected winding, then measures the signal input/output from the other end of the winding as a function of frequency and using a network analyzer displays the response in graphical form on a laptop. The types of problems detected are:

  • Axial Collapse
  • Hoop Buckling
  • Bushing Failures
  • Movement of windings after relocation
  • Shorted turns
  • The test can be done with the unit full of oil or empty. Any before and after comparisons have to be made with the same conditions—Results from an empty tank SFRA should be compared to another empty tank result. This would be typical of a test run prior to transport and then run again upon arrival at its destination. The waveforms should be identical unless there was movement during shipping and handling. We've had experience with this test on furnace transformers and large power units sister transformers) with good repeatable results.

    Right–These are the waveforms for a furnace transformer that has a known bad winding with problems between H3-H1(red waveform on graph). A similar waveform was found on the secondary windings. The SFRA test was done as a confirmation that problems existed within the unit. An internal inspection was done after the SFRA but not much was visible from the top manholes and access was not possible down the sides of the windings.

    Left–Waveform from a typical high voltage winding. Delta connected–the 2 outside windings are almost identical and the center leg has a similar waveform with resonant points at the same frequencies.

    The set up for the furnace transformer tests required that the secondary windings be paralleled as they would typically be when the unit is in service. There are 6 coils per phase in the secondary windings. Special jumpers were made up to provide good connections. The secondary bus was cleaned at the connection points as well.

    We had a client with a transformer at a landfill site that uses methane to generate electricity. The transformer is located in the back end of a cargo container with the engine generator at the other end. There are 12 of these units at the site and they consume almost 100% of the methane from the de-composing garbage. This particular transformer had a history of problems when it was installed several years ago and was replaced. The transformer was sent to a service shop, repaired and left to sit for 2 years. When the unit was put into service, it acted the same way as it did prior to the repair. We took an oil sample and determined that the unit had 57ppm of acetylene dissolved in the oil with an equal amount of hydrogen and some other combustible gases. No base line DGA was performed so we assumed that the unit had generated the gas at the start of its most recent campaign. The unit has a high voltage winding at 24kv and a low voltage winding at 480/277vac. The way the operation runs, the transformer is back fed from the 24kv grid and then the generator is synchronized on the 480vac side. When this unit was energized, it was making a snapping and cracking noise internally as well as causing the top gas collection device to trip it off line. The transformer had never been loaded and had been energized a total of 2 hours with the longest run at 1 hour.

    A battery of tests were performed, the Megger, TTR, Power Factor which all had good results. The first test that showed a problem was the single phase excitation test on H2-H3 primary winding. The excitation test current was 10 times the other 2 windings. An SFRA test was done. This test showed a typical waveform for a condition with a shorted turn. The waveform for the H2-H3 winding had a resonant point missing in the 1kHz range. The transformer was sent to a service shop.

    Once the unit was untanked, we visited the shop to look at the core and coils. The iron had aluminum splatter on the lower frame and significant debris in the bottom of the tank. Another interesting item is the way the core was made. This unit has 5 iron legs and instead of being stacked as a normal transformer, the iron is wound similar to a current transformer.

    The transformer after testing but still resident in the generator container.

    The no load tap changer.

    The tank with debris in the residual oil.

    When the SFRA was done, there were several additional tests done. One shorting the secondary and testing the primary, this showed good waveforms. A second test was done shorting the primary and testing the secondary winding. This continued to show the winding short. The result of the tests led us to believe that the problem was resident in the secondary winding on the C phase. The shop confirmed this.

    The untanked core and coils.