Insulation Condition Prediction of Oil-Impregnated Paper Bushings Using a Novel Hybrid Geometric Approach

An expensive power system component, transformers play a crucial role in power systems. Among the parts of transformers whose health status may affect their performance is oil-impregnated paper (OIP) bushings. Therefore, OIP bushings must be monitored to ensure safe transformer operation. It is important to note that moisture is one of the factors that can damage the OIP bushing. Therefore, effective evaluation of moisture in OIP bushings is crucial to ensure the safe operation of power systems. Previous studies have revealed, however, that the insulation’s aging can cause inaccurate moisture level predictions. Consequently, OIP bushing aging should be considered when estimating moisture levels. In recent years, the frequency domain spectroscopy (FDS) technique has been widely used to evaluate the insulation conditions of OIPs. This study introduced a novel approach to evaluate the insulation conditions of OIP bushings based on the FDS results. Twenty-three sets of oil-paper samples with different insulation conditions were selected. Moreover, eight features are introduced to extract from
$\text {tan}\delta -{f}$
curves of OIP samples. Using the values of eight extracted features and the regression analysis, a database containing values of eight features and their corresponding insulation conditions was established. A novel technique was proposed based on the triangle’s area similarity-sector’s area similarity (TS-SS) to predict the moisture content and aging degree of OIP bushings. The verification results showed that the average relative error of moisture level and aging degree predictions is 5.57% and 7.32%, respectively. Therefore, the introduced method could be used to evaluate the insulation conditions.