A comprehensive evaluation method for moisture and aging state of oil-immersed paper bushings based on dielectric modulus and K-nearest neighbor algorithm

The insulation performance of oil-immersed paper bushings is prone to deteriorate, primarily due to moisture intrusion and thermal aging. The frequency domain spectroscopy (FDS) method is commonly employed to assess the insulation condition of the bushing. However, identifying and extracting relaxation polarization information from the low-frequency region of the FDS curve can be challenging, and there is little research about the condition evaluation under the combined effects of aging states and moisture content. To address this issue, this article uses the Taylor formula mathematical model to extract characteristic parameters from the dielectric modulus curve of OIP bushings and uses the KNN algorithm to achieve the evaluation of aging and moisture status. Then, the effectiveness and accuracy of the proposed method are validated on three field OIP bushings. The results demonstrate that the evaluation accuracy of the proposed method exceeds 83%, which has significant advantages compared to other classification algorithms. The innovation of this article lies in extracting new feature parameters and combining them with intelligent classification algorithms to evaluate the moisture and aging state of the bushing.