Determination of particle trajectories in an isolated conductor GIB using neural network approach

The superior dielectric properties of Sulphar Hexaflouride (SF₆) have long been recognized. Due to the high reliability of the equipment, Gas Insulated Substation can be used for longer time without any periodical inspections. Conducting contamination could, however, be seriously reduce the dielectric strength of a gas insulated system. These particles can either be free to move in the Gas Insulated Bus (GIB) or they may stick either to an energized electrode or to an insulated surface. If a metallic particle crosses the gap and comes into contact with the inner electrode or if a metallic particle adheres to the inner conductor, the particle will act as a protrusion on the surface of the electrode. Consequently, voltage required for breakdown of the GIS may be significantly decreased. Aluminium and Copper particle of size 10 mm/0.25 mm presented on the enclosure inner surface. Inner and outer diameters 55 mm and 152 mm respectively were considered for simulation. The distance traveled by the particle using appropriate equation, is found to be in good agreement with the published work for a given set of parameters. The results are also presented for Artificial Neural Network (ANN) approach and the results are compared with the published work.