The insulation degradation of a voltage transformer winding is not easy to find, but it may ultimately cause the transformer to explode, leading to accidents such as phase-to-phase short circuits. At present, power transformers, lightning arresters, and other equipment have on-line methods to detect the insulation state. However, there is no mature method for the on-line monitoring of voltage transformer winding insulation. In this study, a small current disturbance method based on a zero-sequence loop is proposed. The characteristic parameters of the low-frequency oscillation of zero-sequence voltage after a disturbance are used to evaluate the insulation state of the winding and detect faults. Theoretical modeling, simulation tests, and field tests show that when the insulation resistance of the voltage transformer winding is in the range 0–40 kΩ, a low frequency oscillation of about 10 Hz can be detected on the zero-sequence voltage, and its amplitude and duration are proportional to the degree of damage to the insulation. This can hence be used as a criterion for the on-line detection of voltage transformer winding insulation defects.
On-line identification for low frequency oscillation based on CCF-TLS-ESPRIT algorithm