Aging Condition Evaluation of Oil Paper Insulation Based on Adaboost Algorithm

Abstract

Aging is a critical element influencing the insulation Characteristics of transformers, and is a key link to maintain the balanced work of the power system. To more accurately and comprehensively assess the aging extent of the transformer, this paper deeply analyzed the change mechanism of FDS spectral lines with different aging degrees, and extracted three frequency domain characteristics significantly related to the aging degree of the transformer from the spectral lines for subsequent aging degree evaluation. Then based on adaboost algorithm, the aging degree of transformer was evaluated. The conclusions show that the root mean square error of the proposed model is significantly reduced compared with other models, which fully verifies the accuracy of the model built in this paper.