The natural pollution which is mainly affected by the weather conditions are the main cause of flashovers on high voltage insulators leading to outages in power systems. In this work, characteristics of flashover for contaminated cup-pin insulators have been studied based on experiential test and a mathematical model. Information from laboratory test combined with new mathematical model results are used to define Artificial Neural Network (ANN) algorithm and Adaptive Neuro-fuzzy Inference System (ANFIS) for calculated the flashover characteristics (current IF and voltage UF). several of experiments and measurement are carried out for 1:1, 5:1, 10:1 and 15:1 ratios of bottom to top surface salt deposit density on contaminated samples (z). Dimensional Analysis Method (DAM) was used to derive new model for the variables which often effective in the flashover phenomenon of polluted insulators. The model was derived by establishment the relationship between flashover voltage UF and current IF, length of pollution layer LP, exposure time t, arc constant A and layer pollution conductivity of insulator σ. The both arc constants A and n is computed using genetic algorithm. Comparative investigates have clearly shown that the approach AI-based method gives the agreeable results compared to the mathematical model.
