China applies AI technology to large-scale heating detection of transmission lines for the first time

The reporter learned from North China Electric Power University on August 11 that the infrared defect intelligent identification system for transmission lines developed by State Grid Power Space Technology Co., Ltd. in conjunction with the university and other units has recently been industrialized in the operation and maintenance of my country's major ultra-high voltage lines. This is the first time in China that artificial intelligence (AI) technology has been applied to transmission line heat detection on a large scale.

According to reports, during the peak summer period, the temperature across the country continues to rise and the power load increases sharply. In order to ensure the safe and stable operation of the power grid, hidden dangers in line defects must be discovered in a timely manner. However, in the past, the process of using artificial intelligence to identify infrared image data was relatively complicated, and required manual on-site identification of heating fault points in the image, which was easily affected by factors such as the experience and attention of maintenance personnel and caused omissions. In addition, the amount of infrared video data is very large. It is huge, the re-inspection work is extremely difficult and inefficient, and it can easily lead to dangerous incidents such as insulator disconnection. The newly developed intelligent identification system for infrared defects in transmission lines can quickly extract frames and intelligently identify heating defects by uploading inspection infrared videos with just one click, which can assist line operation and maintenance units in promptly eliminating hidden dangers of line tripping and power outages.

"This time, the technical research team combined the business scenarios and adopted the technical route of 'minimized annotation + stepped learning + interference point shielding' to achieve intelligent identification of hidden infrared defects, with a model identification accuracy of more than 90%." The system Guo Xiaobing, director of the Inspection Division of the Space Technology Application Center of State Grid Power Space Technology Co., Ltd., said.

According to reports, the system is currently deployed and applied at State Grid Power Space Technology Co., Ltd. It is the first time in China that artificial intelligence technology has been applied to transmission line heat detection on a large scale. Taking the infrared video of a 240-base tower as an example, traditional manual data review takes 5 hours. Now using this system, it only takes 2 hours from uploading the video to completing the analysis, and no manual intervention is required in the process.