To address the difficulty in diagnosing mechanical faults of on-load tap changers in converter transformers of ±800kV converter stations

the author

as a frontline operation and maintenance personnel

proposes a fault diagnosis method based on multi-source information fusion

combining with the operation and maintenance practice of the ±800kV Qingyang Converter Station. By collecting multi-source data such as vibration signals

gas content in oil

and driving motor current

fault characteristic information is mined using data preprocessing and feature extraction technologies. Deep fusion of multi-source information and fault pattern recognition are realized by combining machine learning algorithms and evidence theory. Experimental verification shows that this method can effectively improve the accuracy and reliability of fault diagnosis

providing a new approach for condition monitoring and fault early warning of on-load tap changers in converter transformers.