Study On The Discharge Ultraviolet Image Characteristics And Flashover Pre-warning Of Contaminated And Wet Porcelain Insulator

Contaminated and wet suspension porcelain insulator will radiate ultraviolet(UV) signal, which is an important sign to assess the insulation condition of ceramic insulator. The insulation condition assessment and flashover pre-warning were studied in this paper, and the main work and achievements are as follows.The UV band spectrum distribution of discharge was studied with a spectrum analyser, research shows that the UV radiation intensity increases with the increase of discharge intensity. Discharge spot area was extracted with the algorithm of image transform, threshold segmentation, morphological filtering,small area elimination algorithm and multiple region contour tracking; and the spot area parameters was defined; the continuous ultraviolet images are extracted from the UV video, and constitute the spot area sequence. The image processing and parameter extraction software is develop with VC++ software and Open CV.In the artificial climate chamber, the UV image characters of single, three and seven pieces insulators were analyzed under different humidity and pollution. The average spot area, variance and the frequency of large spot area are extracted to quantify the discharge, the correlation of the three parameters with the humidity and ESDD was analyzed. The relationship between the electrical signals and the discharge spot area was analyzed in corona, spark and arc barrier discharge stages;the condition of insulation was divided into 4 kind conditions based on the electrical signals, then the correlation of the spot area with the distance was studied, and a correction formula was put forward.According to the UV imaging characteristics of the contamination discharge,the average spot area, variance and the frequency of large spot area are chosen as character parameters, and the cluster centers of “relative better”,” relative poor”,”poor”,”very poor” four grade were obtained with FCM fuzzy clustering algorithm. The insulation condition of insulator was assessed by the minimum distance criterion. The insulation state of “very poor” was chosen as the flashover pre-warning criterion, and the practical application shows the effectiveness of this method.