Power Transformer Oil–paper Insulation Degradation Modeling

“

This study is for the case where the oil-paper insulation data available for power transformers are limited to a small amount of furfural data, to solve the problems existing in power transformer oil–paper insulation degradation process, such as few samples available，the unknown function form of the degradation process, differences among the degradation processes of individual transformers, and commonality of the degradation trends. From the perspective of building a degradation model to improve the accuracy of degradation estimation, a power transformer oil–paper insulation degradation modeling and prediction method based on FPCA was proposed. The effectiveness of the proposed method was verified using transformer oil–paper insulation accelerated degradation data and actual transformer oil–paper insulation degradation monitoring data. The conclusions are as follows: 1)Compared to other methods, the proposed method is less dependent on the amount of known degradation information, which ensures high prediction accuracy and reliability at any stage of the transformer oil–paper insulation degradation. It can be seen that the proposed method has good accuracy performance even in the case of less degraded data. 2) Because the functional form of the transformer oil–paper insulation degradation model established in this study does not need to be pre-set, the subjective factors based on traditional mathematical analytical modeling methods can be eliminated. Therefore, the original data characteristics of oil–paper insulation degradation can be retained to the maximum extent, which solves the unknown function form problem of transformer oil–paper insulation degradation well. 3)The proposed method is based on functional time warping technology and the FPCA theory, which can accurately extract the common characteristics and inter-individual difference information of the degradation process. 4)The proposed method has strong adaptability to the random volatility and uncertainty characteristics of the transformer oil–paper insulation degradation process. 5)The accuracy verification results based on actual transformer data show that the proposed method has high engineering application relevance, and its prediction results can provide reference suggestions for transformer maintenance work.

”