Research on Transformer Fault Identification Based on Multiphysics Coupling Simulation

In order to study the noise signal characteristics of the transformer under different working conditions, it is helpful to diagnose the operating state of the transformer. Using finite element analysis software, a multi-physics finite element model of transformer electrical-magnetic-structural force-sound field is established, and multi-physics transient simulation analysis is carried out for three working conditions of transformer normal state, winding deformation and iron core deformation. And the transformer fault diagnosis algorithm based on the similarity of feature matrix is used to analyze the similarity of the noise signal between the fault and the normal working condition of the transformer. The research results show that there are obvious differences in the noise distribution of the noise signal in the solution domain of the transformer under different working conditions. The noise signal generated by the vibration of the transformer during normal operation presents a symmetrical distribution trend, and the noise radiated when the winding and the iron core are deformed is obvious. Asymmetric distribution; the proposed transformer fault diagnosis algorithm can comprehensively realize transformer fault identification of multiple groups of noise signals and different measuring points, and identify transformer faults combined with noise distribution map.