Study on SF6 Gas On-line Monitoring Method Based on Machine Learning

In order to improve the accuracy of fault detection in GIS gas chambers, a joint fault detection method based on semi-supervised deep learning network for GIS equipment was proposed. Firstly, the SF6 gas on-line monitoring sensors were used to obtain the GIS gas chamber physical index measurement values, and the semi-supervised deep learning network is used to detect equipment faults in the GIS air chamber, making the deep learning network model adaptive. Experimental results show that the semi-supervised deep learning network can effectively detect and analyze the fault types of GIS equipment.