An Intelligent Transformer Warning Model based on Data-driven Bagging Decision Tree

Transformer is one of the important electrical equipment in power system. Its running state directly affects safe operation of the system. Transformer fault prediction and intelligent early warning are the basis for ensuring normal operation and maintenance of the system. In view of transformer differentiation and the scarcity of fault data, an intelligent transformer warning model based on data-driven bagging decision tree is proposed. A standardized early warning method is proposed based on different transformer equipment condition, and data-driven feature selection for tradeoff is carried out, which provides previous stage knowledge for fault diagnosis. Assessment results depending on fault datasets from State Grid Corporation of China demonstrate that the proposed model can provide a higher accuracy prediction compared to several other prediction models including neural network, KNN, SVM, Linear Discriminant and Logistic Regression, which could bring additional economic benefits and extensive social advantages.