Research of GIS partial discharge type evaluation based on convolutional neural network

This paper presents a GIS (Gas Insulated Switchgear) partial discharge type evaluation based on the convolutional neural network. It is significant to study GIS partial discharge and effectively classify the types of GIS partial discharge. In order to effectively and efficiently distinguish the types of GIS partial discharge, phase resolved partial discharge is researched and evaluated based on the convolutional neural network. First, the hardware of the evaluation system is introduced in detail. Second, the partial discharge signal is analyzed. Third, the partial discharge signal evaluation based on the convolutional neural network is studied. Finally, the experiment of the evaluation is presented to verify that the evaluation system can effectively and accurately perform GIS partial discharge monitoring and improve the reliability of power grid operation.