The Method for Evaluation of Severity of Partial discharge by Power Equipment Based on Characteristic Quantity Forecast

In The Ubiquitous Electricity Internet of Things (UPIoT), the sensing layer is the basis of the power transmission and transformation IoT architecture. Achieving comprehensive intelligent sensing of the state of power transmission and transformation equipment is a prerequisite for the intelligent operation of the power grid. By assessing the severity of partial discharges in GIS equipment, the state of the equipment can be better grasped, thus ensuring safe and stable operation of the grid. In this paper, using the substation on-site detection data, using auto encoder, gated recurrent unit, convolutional neural network and other artificial intelligence techniques, a method for estimating the severity of GIS equipment based on feature quantity prediction is proposed. This method enables an effective assessment of the severity of the discharge of the device, allowing the device to operate safer and more reliably. Compared with other classification models, the calculation results of the partial discharge evaluation of this method have higher accuracy and can be better applied to the practice of IoT engineering of power transmission and transformation equipment.