Internal Temperature Prediction Method of Transformer Based on Multi-Point Inverse Transformation

As the core equipment of the power system, the internal temperature monitoring of the transformer is of great significance to the safe operation of the power grid. Aiming at the difficulty to arrange distributed sensors inside the transformer for direct temperature monitoring, this paper proposes an inverse method for internal temperature based on multi-point inverse transformation mapping of transformer surface temperature. This method constructs the matrix inverse problem model of the internal and external temperature mapping of the transformer, solves the inverse matrix through the electromagnetic heat flow coupling finite element simulation data of the transformer, and then implements the internal temperature inversion by combining the temperature data of the outer measuring points. In this paper, the validity of the multi-point inverse transformation model is verified by multi-physical field simulation and the D-800/35 scaled transformer model. The results show that the average error of the internal temperature inversion error of the transformer is less than 1 K under the condition of less temperature input of the measuring point on the surface of transformer, which has higher accuracy and lower computational complexity. This method can inverse the internal temperature of the transformer quickly without sensor intrusion, which can provide theoretical support and technical support for temperature monitoring and fault diagnosis of transformer.