Transport, short circuit or earthquake? When an SFRA test is worth running

Three events subject a transformer to mechanical stress without leaving an external mark: transport, with its sustained vibration and acceleration; a through fault, whose electrodynamic forces act on the windings within milliseconds; and an earthquake. In all three the unit can go back to running normally and still have a displaced coil, a deformed winding or lost clamping pressure.

The difficulty is that the usual tests may not see it. Turns ratio and insulation resistance describe the electrical circuit and the state of the insulation, not the internal geometry. A deformed winding that has not yet failed electrically can pass both.

Frequency response works on something else. It injects a low-voltage signal across a wide frequency range and records how the winding responds. That response depends on the distributed capacitance and inductance of the assembly, which in turn depend on the physical position of the turns. The result is a curve — a fingerprint — that changes if the geometry changed.

Its usefulness is comparative: the curve is read against a reference. That can be a measurement of the same unit when it was sound, another phase of the same transformer, or a sister unit. Which is why a baseline taken while the unit is in good condition, before any event, is worth having: without it, interpretation rests on less direct comparisons.