Parameter identification based on vector fitting method for extended debye model of oil-paper insulation system

In order to obtain more specific information of insulation states of oil-paper insulation system from time-domain data, we proposed a novel method for parameter identification of extended Debye model for oil-paper insulation system. In the method, the depolarization current spectrum of polarization and depolarization current (PDC) in experiments was obtained by the fast Fourier transform (FFT). Then combining the vector fitting method (VFM) and nonlinear fitting algorithm (FNA), we derived the distribution of the poles and residues of the obtained spectrum, and calculated the branch parameters of extended Debye models. The PDC and frequency domain spectrum (FDS) measurements were performed on some oil-impregnated pressboard samples, which were in various aging states and under different temperatures. The corresponding extended Debye models to the samples were obtained from the PDC measurement, based on which curves of tanδ varying with frequency were figured out and then compared with those from FDS measurements. The tanδ obtained from PDC agrees well with that from FDS under low frequencies, which verifies the proposed method. With further analysis, it is suggested that, for oil-paper insulation system, FDS measurement should be conducted in high frequency range but replaced by PDC measurement in low frequency range, for saving time, i. e. the combination of FDS and PDC methods could help to improve the efficiency of measurement of on-site oil-paper equipment.