For reducing the number of sudden dielectric breakdown of the operating equipment due to the latent partial discharge defects inside electrical equipment

the dielectric discharge of one operating 126 kV GIS equipment is taken as the object in this paper and the research on fusion diagnosis and intelligent early warning system of partial discharge is performed. Based on the characteristics of acoustic signals and online monitoring methods characterized in the process of partial discharge

a set of partial discharge fusion diagnosis system using dynamic frequency acquisition method and cloud platform is adopted to complete the partial discharge signal acquisition and preliminary positioning of defects inside the equipment. The defect state is divided into the stable development stage and the high breakdown risk stage through the analysis of the state diagram variables such as average signal

standard deviation and maximum value inside the interval of partial discharge time

and the automatic assessment of severity and the intelligent early warning of the defect development state is achieved by the BP neutral network. The test results prove that the designed fusion diagnosis and intelligent early-warning system can complete the dynamic tracking and early warning of the defect at low-risk and high-risk stage.