778 / 2022-03-31 21:49:00

Diagnosis Method for 220kV Transformer Bushing Based on Digital Twin Technology

digit twin,Transformer bushing,computer vision

Purpose/Aim

In order to determine the fault status information of the transformer bushing and improve the level of intelligent operation and maintenance of substations. Analyzing the temperature distribution of bushing can accurately judge the cause of fault. However, it is difficult to judge the fault category only by thermal imaging and temperature measurement of key parts. It is necessary to establish a simulation model to study the overall temperature distribution to find out the abnormal heating area and repair the fault.

Experimental/Modeling methods

Based on the finite element method (FEM) simulation software COMSOL, the electro-thermal  coupling calculation of thebushing model is proposed.

Results/discussion

By comparing the simulation model with the infrared image recognition results，the heating area of bushing and the rise of abnormal temperature under different working conditions are obtained. Based on this, the model’s accuracy can be verified by comparing the rise of abnormal temperature with the typical fault.

Conclusions

The digital twin model of transformer bushing is built, the temperature distribution of bushing under various fault categories is calculated and the temperature rise characteristics of bushing under different fault categories are obtained.

The infrared detection network of YOLOv4 based on PyTorch is built. The detection model of abnormal heating area in casing infrared image is trained. The location and identification of thermal fault area of bushing is obtained, which compared with the simulation model, the classification of fault category of transformer bushing is realized, which will provide effective support for on-site operation, maintenance and fault analysis of transformer outgoing bushing