水电站GIS设备内部异物识别技术研究

Research on recognition technology of foreign matters in GIS equipment of hydropower station

GIS设备作为水电站内电力送出的关键设备,其内部异物引起绝缘击穿是GIS设备占比最大的故障,直接影响着水电站的效益。由于GIS设备筒体较长,在设备安装及开展检修工作时,其内部毫米级异物难以通过经验判断。为了更加有效地检查识别GIS管道内的异物,对比分析了超声波、电磁波、光学不同类型的识别方法对GIS内部异物的有效性。研究了将光学敏感法和光影法相结合的异物识别方法,并设计了基于双光源补光的异物识别算法,将其搭载在GIS设备内部异物检查机器人上实现了应用。应用结果表明:采用双光源补光的异物识别方法,可有效准确地识别出GIS腔体场景中的毫米级异物,准确率可达95%以上。

GIS equipment is the key equipment for power transmission in hydropower stations.The insulation breakdown caused by internal foreign matters is the most common fault in GIS equipment,which directly affects the benefit of hydropower station.Due to the long barrel of GIS equipment,it is difficult to judge the internal millimeter-level foreign matters by experiences during the equipment installation and maintenance.Therefore,in order to check and identify the foreign matters in GIS pipelines more effectively,we analyzed the effectiveness of different types of identification methods such as ultrasonic wave,electromagnetic and optics wave on the internal foreign matters in GIS equipment.Then we proposed a method of foreign matters recognition combining optical sensitive method and light shadow method,designed a algorithm of foreign matters recognition based on dual light source filling light,and equipped with foreign matters detection robot in GIS equipment to realize the application,which solved the problem that it is difficult to identify and find the foreign matters in GIS equipment.The results showed that the foreign matters recognition method based on dual light source filling light can effectively and accurately identified the millimeter-level foreign matters in the GIS cavity scene,and the accuracy rate can reach more than 95%.