Investigation of flashover phenomena on high voltage insulators using artificial neural network

ABSTRACT PhD Thesis INVESTIGATION OF FLASHOVER PHENOMENA ON HIGH VOLTAGE INSULATORS USING ARTIFICIAL NEURAL NETWORKS Muhsin Tunay GENÇO?LU Fırat University Graduate School of Natural and Applied Sciences Department of Electrical and Electronics Engineering 2003, Page: 112 Pollution flashover, observed on insulators in high voltage transmission, is one of the most important problem for power transmission. Pollution flashover is a very complex problem, because of the modeling difficulties of the insulator complex shape, different pollution density at different regions, non-homogenous pollution distribution on the surface of insulator and unknown effect of humidity on the pollution. In literature, some static and dynamic models were developed by making some assumptions and neglections to predict the flashover voltages of polluted insulators. Static models, which take the partial arc observed after the dry band existence into account statically; are not suitable for exact prediction. Dynamic models, which take the effect of instant variations of discharge parameters on flashover voltage into account, represents flashover event better than static models. In this study after presenting the definition, physical mechanism and importance of pollution flashover, insulator equivalent models are investigated and the models are compared. Artificial neural networks are also given in detail. For this purpose, concerned insulators' scaled shapes are partitioned into triangular elements, then Finite Element Method is implemented and finally potential distribution on the surface of insulator, variation of pollution resistance and flashover voltage are determined with the help of developed new dynamic arc model. The computed flashover voltage characteristics of selected ten types of string insulators have been compared with the results obtained by other researchers for direct and alternative voltages Insulator pollution flashover behaviour has been investigated by using Artificial Neural Networks and the obtained results are verified by other researchers' experimental results. Keywords : Insulator, Pollution flashover, Pollution resistance, Flashover voltage, Dynamic arc model, Artificial neural networks. ^j^y-a XII