Ultraviolet Imaging Detection Of Composite Insulator Surface Discharge And Development Of Diagnostic System

As one of the most commonly used external insulation equipment in power transmission and transformation systems,composite insul ators have a direct impact on the safe operation of the power grid.The surface discharge characteristics of composite insulators can reflect their insulation state.The solar-blind ultraviolet imaging method has high detection sensitivity and good discha rge localization ability.Based on ultraviolet imaging,this paper studies the discharge of composite insulators under different pollution and wet conditions,and proposes an evaluation method for insulator discharge based on deep learning algorithm.The m ain research work is as follows :The UV imaging characteristics of the surface discharge of the composite insulator under different wet conditions were investigated experimentally,and the influence of water droplets on the electric field was analyzed.Wi th the increase of pollution level,the corona initiation voltage of larger water droplets decreases with the increase of conductivity,and the increase of water droplet conductivity will increase the corona initiation voltage and decrease the flashover vo ltage.When the conductivity of rainwater increases,the discharge intensity increases,and the discharge spot begins to develop from the high-voltage end to the low-voltage end.Using the finite element analysis software,the influence of water droplets o n the surface of the composite insulator on the electric field was simulated.Based on the visible light channel image and the YOLOv4 algorithm,the insulator image recognition is completed.The mosaic data expansion algorithm is improved,the accuracy of the model is improved,and the training error is reduced.Based on the self-built power equipment image database for training,the recognition effect is good.Based on quantitative parameters such as leakage current and UV imaging photon number,combined with the shape of discharge,the UV images of insulator fouling discharge are classified.An ultraviolet video statistical parameter diagnosis model is established to realize the evaluation of insulator discharge severity based on ultraviolet video,which provides a new research method for the insulation state detection of insulators.Based on the B/S architecture,a software system for UV imaging detection data management and fault diagnosis for external insulation equipment was developed.LAMP was used as a software development platform for integrating functional modules to realize the design and development of database,communication and control of shooting equipment;UV images were performed using OpenCV.The development of the quantitative analysis module realizes the analysis of the ultraviolet image of the outer insulation equipment and the diagnosis of the discharge degree;the development of the intelligent identification and diagnosis module using the deep learning algorithm realizes the recognition of the visible light image of the outer insulation equipment and the diagnosis of the insulation state.