Performance Evaluation of Laboratory and Industrial Equipment in FRA Measurement

Abstract

Frequency response analysis (FRA) is the widely used condition monitoring technique for the power transformer. Test instruments performing FRA test can broadly be categorised as high-end equipment (HE) and laboratory equipment (LE), respectively. The accuracy of the generated data depends on the HE and LE to a greater extent. The proposed work includes absolute study of such test instrument in order to understand repeatability in FRA measurements. Moreover, natural frequencies and amplitudes are identified from the FRA plots in order to perform sensitivity analysis. This paper proposes, two non-parametric, statistical methods, namely the Mann–Whitney U-test and quantile–quantile plot for the comparison of the amplitude data. To validate the findings, experiments are performed on the model coil, the disc winding and the three-phase transformer.