Detection Method of Interturn Short-circuit Faults in Oil-immersed Transformers Based on Fusion Analysis of Electrothermal Characteristic

The existing methods of interturn fault diagnosis of a power transformer depend on the single signal indicator such as the winding current, the impendence or the dissolved gases etc., so that they can not accurately detect the exact location of the faults, especially the short circuit with a few turns. Taking the winding current, winding hot spot temperature and oil temperature into consideration, a new approach based on the fusion analysis of electrothermal characteristics is proposed to detect the interturn faults in the winding of an oil-immersed transformer. The main idea of the proposed method is to establish the digital space model of the physical transformer with the digital twin technology. Then, multi-physics simulation is used to derive the changing law of the electrothermal characteristics of the transformers under different operation conditions and interturn fault types in the digital space. With the main electrothermal characteristics parameters including winding current and hot spot temperature as the features, a twin fault sample based data-driven model is presented for the interturn fault detection. A case study is carried out on the 31.5MVA/110kV power transformer. The results show that the proposed detection method based on the fusion analysis of the electrothermal characteristic can effectively identify the early latent interturn fault of the power transformer. The overall accuracy of the proposed method can reach 94%.