Problems during the induction test of an HKSSPZ-6300/110 type furnace transformer

A HKSSPZ-6300/110 electric arc furnace transformer has the following basic parameters:

Rated capacity S = 6300 kVA, primary voltage U₁ = 110 kV, secondary voltage U₂ = 110–160 V, vector group YNd11, with both low-voltage winding ends (start and finish) brought out, and equipped with 13-step on-load tap changing. Insulation levels: HV/HV neutral/LV, LI480AC200 / LI325AC140 / AC5.

The transformer uses a dual-core series voltage regulation design, with an "8"-shaped low-voltage winding configuration. The schematic for the induced voltage test is shown in Figure 1.

Test conditions: tap changer set at position 13; 10 kV applied to tertiary windings Am, Bm, Cm; with K = 2, only phase A is illustrated (phases B and C are identical). Calculated values: UZA = K × 10 = 20 kV, UG₀ = K × 110 / √3 ≈ 63.509 kV, UGA = 3 × 63.509 = 190.5 kV (95% of rated), UAB = 190.5 kV, frequency = 200 Hz.

After completing the test connections per the diagram, the induced voltage test began. When UZA was raised to 4000–5000 V, distinct "crackling" corona discharge sounds were observed near the low-voltage terminal bushings, accompanied by the odor of ozone. Simultaneously, the partial discharge (PD) detector indicated PD levels exceeding 1400 pC. However, the measured voltage between low-voltage terminals remained correct. Initially, we suspected potential issues with the low-voltage terminal material or the effect of the 200 Hz test frequency on the resin terminal. In a second test using a 50 Hz power source at the same voltage (4000–5000 V), the same phenomena were observed, thereby ruling out the influence of the 200 Hz frequency.

We then carefully reviewed the test circuit diagram and actual connections. It was noted that the low-voltage winding ends (start and finish) are both externally brought out and are normally connected externally into delta or star configuration when connected to the furnace. During the induced voltage test, however, the low-voltage terminals were neither connected in star nor in delta, nor grounded—leaving them in a floating potential state. Could this floating potential be the cause?

To test this hypothesis, we temporarily connected the x, y, and z terminals together and reliably grounded them before re-running the test. The aforementioned discharge phenomena disappeared completely. When the voltage was increased to 1.5 times the level, PD was only about 20 pC. The test voltage was further increased to 2 times, and the transformer successfully passed the induced voltage withstand test.

Conclusion: For this type of dual-core series voltage-regulated furnace transformer with both low-voltage winding ends brought out, although the voltage between terminals (e.g., a and x) is low, the absence of a reliable ground connection can create a floating potential, leading to the observed partial discharge. Therefore, during induced voltage testing, the x, y, and z terminals should be shorted together and reliably grounded to eliminate such anomalies.