Fault diagnosis of power transformer based on tree ensemble model

Abstract

During the running operation of the oil-immersed transformer, some gases may be dissolved in the insulating oil which can be used to diagnose the incipient failure of the power transformer. This is the Dissolved Gas Analysis (DGA). This paper proposes a power transformer fault diagnosis method based on tree ensemble model (Extreme Gradient Boosting, XGBoost): constructing a large number of classification and regression trees (CART) to fit the residuals obtained by each learning. Compared with the commonly used SVM and BPNN methods, our method has a significant improvement in accuracy, F1-score, precision and recall.