By Andrey Reykherdt, Don Taylor, Junming Kuang, Mark Cotton, Adam Suleiman, Christopher Cox, Tim Pasanduka

Fault investigation on a 150 MVA 220/66/10.9 kV power transformer performed by Select Solutions utilising the great experience of condition monitoring engineers and asset management in finding a root cause of an electrical fault in the transformer.

A power transformer is a critical element of any transmission or distribution network. It is well-known that the transformers can develop a fault condition for a number of different reasons associated with external events such as lightning strikes, switching or transient over-voltages or due to internal electrical or mechanical faults in the transformer. In our case, a 150 MVA 220/66/10.9 kV transformer positioned in a sound enclosure (Figure 1) tripped on combustible gases generated inside the transformer the cause of which was unclear before the full investigation was carried out. The DGA analysis indicated a high level of acetylene (22 ppm) in the main tank which suggested a potential electrical fault as a cause for the transformer trip.

Figure 1. The tripped transformer in the sound enclosure


Initial investigation into the fault condition included a normal suite of condition monitoring tests which are periodically performed on the transmission transformers including, but not limited to, sweep frequency response analysis (SFRA), insulation resistance tests, dielectric dissipation factor (tan delta) & capacitance measurements on transformer windings & bushings and DC resistances through a tapping range. The test results were considered satisfactory and did not indicate any fault condition. This normal sequence of testing and the applied 10 kV test voltage did not identify a problem with the transformer, but at least confirmed that there was no mechanical deformation of the windings or short circuits. However a decision was made to further investigate the reason for the generated combustible gases in the main tank before the transformer could be switched back in service as it would compromise the integrity of the power network.

As a part of further investigation, the partial discharge (PD) measurements and their acoustic PD localisation were conducted by inducing the service voltages and above on the transformer windings using a HIGHVOLT test system which was specially designed for Select Solutions (Figure 2). Using this test set, the transformer can be energised in a single or three phase mode using a variable frequency up to 200 Hz to overcome transformer core magnetic saturation when inducing more than service voltages (>1.0pu.).

Figure 2. HIGHVOLT Power Unit with Select Solutions transportable test laboratory

Figure 3. Single phase diagram for the transformer under test using HIGHVOLT test set

The transformer tertiary (TV) winding was initially supplied in a single phase mode. Figure 3 shows the A phase excitation used for PD investigation on this transformer.

Electrically coupled PD quadruples were connected to all three phases of HV bushings for simultaneous PD measurements. By increasing the induced voltage above 10 kV, it was found that the A phase exhibits a relatively high PD activity with inconsistent PD inception voltage as shown in Figure 4.

Figure 4. Top – Replay overview of a single phase excitation test on a phase
(Red Line – HV Phase-to-Ground Voltage; Green Line – Partial Discharges);
Bottom – Phase – resolved PD patterns

The transformer was then excited by supplying all three phases as per service condition  which proved that partial discharges originated from the A phase were also cross-coupled PD to other phases in relatively small magnitude (Figure 5).

Acoustic sensors were positioned in various locations on the transformer tank during the three-phase excitation tests. Figure 6 shows the acquired acoustic signals with the sensors positioned 450 mm from the top lid of the tank on HV side against each phase. The calculated distances from a potential PD source for B & C phases in relation to the signal received by the sensor on the A phase were:

  • 43 m for B phase sensor and
  • 55 m for C phase sensor.

As one can see from Figure 7, the A phase signal has a sharper front rise in comparison with more attenuated signals coming from other phases. It means that the signals received by other phases are more remote and include both longitudinal and transverse waves propagated through transformer oil and tank.

Familiarity with internal structure of the transformer was presented by the asset management team to determine the most likely source of partial discharges. The investigation continued on the A phase assuming a direct acoustic path of PD signals coming from either a flux deflector on top of the A phase winding or from bottom of the HV bushing.

Figure 5. Three-phase induced test indicated significant PD activity on A phase

The sensors were repositioned many times on the tank with a final position of the sensors superimposed around a potential source of partial discharges pointing to the fault location within 5-10 cm.

Figure 6. Acoustic signals of PD (the sensors positioned 450mm from top lid of the tank against each phase)



Figure 7. Final position of acoustic sensors indicates equivalent distance from PD source

During visual inspection of the transformer active parts, it was noticed that one of the flux deflectors had moved towards HV A phase bushing approx. 100-120mm pushing hard against the pressboard of the bushing assembly (Figure 8). A by-product of electrical discharge and carbonisation of paper insulation between the flux deflector and pressboard of the bushing entry was also evident.

The A phase HV bushing was removed from the transformer for further investigation. The full set of HV tests were carried out on this bushing with the maximum test voltage of 190 kV (1.5pu.) at Select Solutions’ Yarraville facilities. No partial discharges were observed on the bushing during this investigation.

Following repair of the flux deflector, the transformer was re-tested at 1.2pu for 15 minutes and 1.5pu for 10 seconds at the elevated frequency of 100 Hz to avoid core saturation and overheating of core laminations, in consideration of the age and duty of the transformer. The transformer successfully passed over-voltage withstand test and all phases were clear of partial discharges.

Figure 8. One of the flux deflectors was found to be moved against pressboard assembly of HV Bushing entry
(highlighted by a white arrow), producing significant discharges in the area of HV bushing


Acoustic localisation of partial discharges in the 150 MVA transformer was successfully carried out using the HIGHVOLT variable frequency power unit designed for Select Solutions. This power unit is capable of inducing step-by-step high voltage on power transformers to a required level for fault investigation and PD analysis. Select Solutions is also involved in HV commissioning of different size power transformers on-site with a capability to perform induced high voltage withstand tests,  no-load loss at the rated voltage and high current impedance measurements along with a large number of standard condition monitoring tests.