Transformer State Awareness Technology Based on Multi-Kernel SVM

Power transformer is an important part of the power system. The rapid awareness of transformer state is directly related to the power supply stability of the power system. Therefore, a transformer fault types identification method based on multi-source feature pattern recognition is proposed in this paper. Firstly, the electrical test data of the transformer and the dissolved gas analysis (DGA) related data are combined to form the fault feature space. Then the radius information is introduced into the multi-kernel support vector machine (SVM) to select the kernel function weight. Finally, the multi-kernel SVM with integrated radius information is used to identify the transformer fault type. The simulation results show that the proposed identification method can effectively combine the information from different sources, and the algorithm has high identification accuracy and good robustness.