Fault Diagnostics of Oil-immersed Power Transformer via SMOTE and GWO-SVM

Abstract

Dissolved gas analysis (DGA) is an effective method for fault detection of power transformer. Transformer fault data are typically unbalanced, because the probabilities of different faults are different. This imbalance will cause a decrease in the recognition rate of minority class. In this paper, synthetic minority over-sampling technique (SMOTE) is used to balance the unbalanced fault samples set of power transformer, then Grey Wolf Optimization (GWO) is used to optimize the parameter of support vector machine (SVM). Incorporating above two procedures, the transformer fault diagnosis model is established. The case analysis shows that compared with the original model, the recognition rate of the model established is significantly improved in minority faults, and the overall recognition rate is increased by 7.5%, reaching 86.67%.