干旱区生态环境知识资源中心

Insulators plays a vital role in electrical transmission line. It gives a support and separate electrical conductors and these insulators are most commonly made up of porcelain. These insulators gets polluted when they are placed near different atmospheric conditionlike coastal, desert etc. Now-a-days porcelain insulators are replaced by the silicon rubber insulators due its hydrophobicity property. The design parameters of silicon rubber insulator such as creepage factor, profile factor also contributes the pollution flashover. In order to assess the pollution severity level and to prevent the unpredictable flashover voltage, it is important to predict the flashover before it takes place. In this paper, four different types of 11 kV silicon rubber insulators were taken and solid layer method was used in artificial pollution test, in which sodium chloride and kaolin are used to induce the conductivity. Using even rising method, the flashover voltage has been obtained for different equivalent salt deposit density (ESDD) values. By using artificial neural network(ANN) and least square support vector machine (LS-SVM) flashover voltage has been predicted with the input parameters such as height, creepage length, diameter and ESDD. Finally, it is concluded that the LS-SVM approach has better ability to assist the flashover voltage prediction than ANN.