Purpose/Aim
The closed structure of GIS equipment makes it very difficult to detect its internal state. It is of great significance to invert and reconstruct the internal state of the equipment through external detection signals for equipment state evaluation and fault diagnosis. 
Experimental/Modeling methods
Research the internal temperature propagation model of the combined electrical appliance, study the temperature propagation and attenuation characteristics under different spatial structures, study the temperature propagation characteristics of different heating faults, and construct the inversion model of the internal temperature field of the combined electrical appliance. The influence of factors such as SF₆ gas pressure, gas chamber distribution, and busbar structure on the inversion model is studied, and a simplified inversion model under given boundary conditions is established.
GIS internal temperature distribution simulation model: First, simulate the temperature field distribution of the casing under different heating power consumption, and obtain the corresponding relationship between the heating power consumption and the casing temperature by analyzing the data. Then, the temperature field distribution result data under different power consumption is obtained through simulation, and the training data of the model reduction order is formed, and the temperature field reduction order model is generated by the method of machine learning. Finally, the reduced-order model of the temperature field is integrated into the GIS 3D twin model.
Results/discussion
A three-dimensional heat transfer model of the internal overheating of the three-phase sub-box GIS is established, and the corresponding relationship between the shell temperature and the conductor temperature is realized. The model parameters were revised based on the test data of the temperature distribution of the inner conductor and the shell under the condition of an ambient temperature of 25°C and an applied current of 2000A. Under the condition that the model parameters remain unchanged, the model is verified by the experimental data under other working conditions, and the simulation data is in good agreement with the experimental data.
Conclusions
The application of a new generation of information technology has triggered a revolution in the management of traditional electrical equipment. This paper studies the data modeling method and applies the "digital twin" to the field of fault diagnosing of substation equipment.
 

Digtal twin,fault diagnosis,inversion