Impacts of moisture content and distribution on frequency domain spectroscopy of oil-paper insulation system by FEM approach

Moisture estimation by frequency domain spectroscopy (FDS) provides a novel way for condition diagnosis of oil-paper insulation system from the view point of dielectric response. We adopted the finite element method (FEM) to explore the influence of the moisture and its distribution in oil-paper insulation system on the frequency domain dielectric spectroscopy characteristics of transformer main insulation structure on the basis of analyzing the frequency domain characteristics of insulating paper in lab. Results show that the loss of oil-paper insulation system can be reflected in the curve of complex permittivity ε̇-f, which is closely related to moisture distribution. With the increase of oil conductivity, the conductivity loss extends from low frequency to high frequency, the real part ε' of complex permittivity increases in low frequency domain, and the imaginary part ε" increases in the whole frequency domain. The conductivity loss in low frequency domain and polarization loss in high frequency domain vary with the moisture content in insulating paper. With the increase of moisture in paper, ε' increases in the low frequency domain, ε" increases in the whole frequency domain, and the hump in middle and high frequency domain disappears gradually. As partial moisture in paper increases, the polarization loss becomes larger and ε" increases in the high frequency. When the non-uniformity of moisture distribution deteriorates, ε' remarkably varies a lot in low frequency domain as ε" remarkably changes in high frequency domain.