Fault Diagnosis of Transformer Dissolved Gas Based on Random Forest

To improve the accuracy of transformer fault diagnosis, this paper proposes a transformer fault diagnosis model based on the Random Forest (RF) algorithm. First, the dissolved gas analysis (DGA) method is used to preprocess the concentration data of characteristic gases dissolved in transformer oil. The Random Forest model is then constructed using Bootstrap sampling and random feature selection, achieving high-precision classification of transformer fault types through the integration of multiple decision trees. To validate the effectiveness of the model, actual transformer fault operation data are selected for testing. Additionally, the Random Forest model is compared with five other models: GNB, LDA, KNN, SVM, and GBDT. The results show that the Random Forest model significantly outperforms the other models in terms of both training and testing accuracy, while also demonstrating higher efficiency in training time. The comprehensive performance of the Random Forest model is superior to the comparison models. This study demonstrates that the Random Forest-based transformer fault diagnosis method can effectively handle complex dissolved gas data, accurately identify multiple fault types, and exhibit high generalization ability and robustness, making it suitable for practical transformer fault diagnosis in power systems.