Aiming at the issue of abnormal vibration of gas insulated switchgear (GIS) during operation

the vibration defect detection method of GIS based on singular spectrum analysis (SSA) and gradient boosting decision tree (GBDT) is proposed in this paper. Firstly

the acceleration sensor is applied to acquire the vibration signals of the GIS operating in different states. The SSA method is adopted to proceed the obtained signals and extract the dominant components of the signal. Then

the characteristic parameters of the dominant components are extracted through time domain analysis and frequency domain analysis. Finally

based on the GBDT algorithm

the features are analyzed through recursion and classification

and the defect diagnosis model is thus set up to obtain the final vibration defect detection result of GIS. This method is applied to the laboratory GIS and successfully identifies such two defects as looseness of shield and anchor bolt. The described method and its application are mainly introduced

which provide a reference for the research on fault detection of electrical equipment.