Research on Diagnosis Technology for Ultra High Voltage Gas Operations Based on Ensemble Learning

Gas insulated switchgear (GIS) is the core equipment of ultra-high voltage power grids. The control technology of GIS installation is currently a weak link in the full life cycle management of power assets. This article introduces the architecture of smart gas operations and analyzes the necessity of deploying intelligent dangers diagnosis algorithms at edge gateways. Analyzed he status monitoring methods and hidden dangers in smart gas operations, and fuse heterogeneous data such as pressure, micro water, and gas purity for feature construction. An intelligent diagnosis method for hidden dangers in gas operations based on ensemble learning has been proposed. The PSO-XGBoost model was used to achieve intelligent identification of seven types of safety dangers, including air chamber defects, non-standard operations, and machine failures. Finally, the effectiveness of the diagnostic method was verified through algorithm comparison. The diagnostic technology for ultra-high voltage gas operations proposed in this article provides important technical support for the construction of ultra-high voltage intelligent installation and control platforms.