Novel Characteristic Quantities for Determining the Moisture State of Oil-Impregnated Cellulose Insulation Using the Extended Debye Model

Abstract

Despite the technical popularity of the dielectric response (DR) methods used for the nonintrusive diagnosis of the cellulose insulation condition, it is still considered a challenging and tough task so far to trustworthily interpret the details of the measurement results for further understanding and tracing the insulation changes behind. From a novel perspective of parametric study on the circuital DR equivalence of cellulosic insulation—extended Debye model (EDM), the new dielectric characteristic quantities and the quantification approach to determine the cellulosic moisture condition are investigated in the present study. First, a kind of tailor-made scale-down bushing model is prepared, with its cellulosic condenser body absorbing differently controlled quantities of water to simulate the dampness of real-life insulation, of which the frequency-domain spectroscopy (FDS) is measured under controlled temperatures. Then, the EDM is employed for a parameterized study on the FDS details with the moisture changes, by identifying the model parameters against the FDS curves, whereby the two time-domain dielectric spectra—polarization/depolarization current and return voltage—are transferred from the FDS so that a set of hybrid time–frequency characteristic quantities is proposed, which is revealed to be indicative of the dampness variations of cellulosic insulation. Finally, the principal component analysis and the feature recognition algorithm based on minimum distance are applied to determine the moisture content of the condenser body, which verifies the holistic effectiveness of the proposed method.