Moisture Evaluation of Oil-immersed Insulation in Bushings Based on Frequency Domain Spectroscopy and Grey Relational Analysis

Operation level and service life of oil-immersed bushings are related to the moisture inside theirs cellulose insulation. However, the parameters extracted by using existing frequency domain spectroscopy (FDS) based methods strongly depends on the data within a certain range or several sampling points, which could decrease the accuracy of extraction results if the measured FDS data is unreliable in the above regions. Given this issue, a universal and accessible tool is proposed to stimulate the change rule of FDS curves versus entire frequency regions, by which a set of feature parameters can be readily obtained by using the Taylor series model. Then, the database characterizing the moisture is established. An alternative method for evaluating the moisture in bushing cellulose insulation is proposed based on FDS and grey relational analysis (GRA). The feasibility and applicability of the proposed methods were later proved by the moisture evaluation of transformer oil-immersed bushings, in which the GRA is applied as a tool for moisture classification. The results showed that the average relative (absolute error) is less than 14.6% (0.17%). Thus, the proposed method can be preliminarily utilized for evaluating the moisture inside the bushing cellulose insulation.