ANALYSIS OF BREAKDOWN IN POLLUTED, INSULATOR USING NEURAL. NETWORK MODEL

Abstract

Electrical energy is generated at a distance from load centre and to minimize losses on transmission line; it has to be transmitted at high voltages. One of the faults that exist in high voltage insulator is the pollution flashover and subsequent outage of transmission lines. Since outage of EHV lines is a serious matter, research on pollution flashover invites concern. Flashover on polluted insulator can occur when the surface is wet due to fog, dew or rain. Most commonly seen pollution related problem to flashover exist in coastal areas (sea salt), industrial areas (chemical pollution), and other areas (desert sands, etc.) In the present dissertation, an attempt is made to develop suitable ANN models for predicting the flashover voltage (FOV) of contaminated insulators. The ANN models developed make use of the above three control (input) namely: salinity of contaminated salt, solution current and resistivity of salt solution The output variable is the Flash over voltage (FOV). Since different architectures are possible, it is a voluminous task to explore all possible structures. Therefore, the study is restricted to the investigation of a few selected architectures, and the best ANN model from these is selected for subsequent simulation studies