Search

Electrical shorts in transformers on load can be diagnosed by frequency response analysis (FRA) using impulse or sinusoidal sweep signal. At HV laboratory of SASTRA, a low-voltage impulse based on-line on-load FRA (OLOL FRA) was carried out on a single phase 50 Hz, 1 kVA, 240 V/240 V small laboratory transformer while delivering power to lamp load, and the effectiveness in electrical short diagnosis was analysed. For this purpose, some artificial faults were created within one of the windings of the transformer investigated, and the responses of the winding to different frequency components of the impulse voltages were observed before and after introducing these faults. Comparison of the frequency response (impedance) of the winding under normal condition with frequency response under fault condition was made, which clearly indicated that faults had altered the winding impedance for some frequency components of the impulse. This reveals that fault within the winding can be identified and located, by the OLOL FRA technique, through careful comparisons of frequency response. Statistical analysis of transfer function magnitudes with correlation coefficient, standard deviation and absolute sum of logarithmic error has also validated the same. [ABSTRACT FROM AUTHOR]

Electrical shorts in transformers on load can be diagnosed by frequency response analysis (FRA) using impulse or sinusoidal sweep signal. At HV laboratory of SASTRA, a low‐voltage impulse based on‐line on‐load FRA (OLOL FRA) was carried out on a single phase 50 Hz, 1 kVA, 240 V/240 V small laboratory transformer while delivering power to lamp load, and the effectiveness in electrical short diagnosis was analysed. For this purpose, some artificial faults were created within one of the windings of the transformer investigated, and the responses of the winding to different frequency components of the impulse voltages were observed before and after introducing these faults. Comparison of the frequency response (impedance) of the winding under normal condition with frequency response under fault condition was made, which clearly indicated that faults had altered the winding impedance for some frequency components of the impulse. This reveals that fault within the winding can be identified and located, by the OLOL FRA technique, through careful comparisons of frequency response. Statistical analysis of transfer function magnitudes with correlation coefficient, standard deviation and absolute sum of logarithmic error has also validated the same. [ABSTRACT FROM AUTHOR]