Research On Insulator Pollution Detection Method Based On Terahertz Time-domain Spectroscopy

As a commonly used electrical insulation equipment,insulators have the role of supporting transmission lines and realizing insulation between transmission lines and towers and substations,and are widely used in power systems.Affected by the environment,pollutants inevitably accumulate on its surface,and when the pollution accumulation is serious,it is easy to lead to pollution flashover,and the impact range is wide and the loss is large.Therefore,it is of great significance to study the contamination detection of insulators.In this paper,the use of terahertz time domain spectroscopy to detect and analyze insulator contamination is studied,and the main work includes:1.The characteristics of insulator structure material and the pollution deposited on it are analyzed,and the experimental instrument selection and experimental planning were carried out.Silicone rubber insulation sheet was selected to simulate composite insulator umbrella skirt and pollution samples were made according to the standard to meet the detection requirements of transmissible terahertz time-domain spectroscopy system.In order to meet the requirements of reflection terahertz time-domain spectroscopy system,the ceramic alumina ceramic was used as the substrate to make porcelain insulator pollution samples.2.For the contamination detection of composite insulators,the transmissive THzTDS system was used to complete the detection of silicone rubber materials of different thicknesses and performed linear fitting.Starting from the presence or absence of pollution,the differences in refractive index and absorption coefficient were observed between the two types of samples,which were distinguished by PCA algorithm and classified by SVM model.Starting from the degree and composition of pollution,two types of soluble pollution samples,sodium chloride and calcium sulfate,and their pollution levels,were studied,and the PCA-SVM model was used to identify and classify pollution.And the experimental samples were added,and the multi-class salt-dense numerical samples made of sodium chloride contamination were analyzed,from PCA algorithms to a variety of classification models,and finally the optimization SVM classification model had the highest recognition accuracy.3.For the detection of porcelain insulator contamination,the reflective THz-TDS system is used to complete the detection of ceramic substrate contamination;Terahertz reflection signals were detected for six types of dirty samples and the refractive index was calculated separately;Sodium chloride and kaolin mixed sewage liquid to make a dirt sample for pollution thickness detection,the dirt thickness above 0.07mm can detect the result and the error is within 11%,with the increase of the thickness of the filth layer,the error of terahertz thickness gradually decreases.