Error Analysis of Extracting the Dielectric Spectroscopy of Oil-impregnated Paper from Oil-paper Insulation

Abstract

Frequency Dielectric Spectroscopy (FDS) is a kind of dielectric response diagnostic method which is performed in frequency domain and is the most widely used transformer insulation detection. An important procedure in the insulation diagnostic is extracting the dielectric spectroscopy of oil-immersed paper from oil-paper insulation based on the XY model. As in the on-site measurement, the parameters of the structure used in the calculation are often estimated, which will bring error to the calculation of the spectroscopy of paper. In this paper, firstly the theory of the XY model and the calculation of the spectroscopy of paper are introduced. Two solutions of the calculated spectroscopy of paper are shown and the selection of the two solutions is discussed. Then a representation method called 3D-dielectric spectrum is proposed for the error analysis of the calculated spectroscopy of paper. And the influences of the related variables are studied. The conclusions show that the structural parameters of insulation X and Y have few influences on the error of the calculated spectroscopy of paper. And the error of the structural parameters $\Delta X$ and $\Delta Y$ influence differently at different frequency bands.