Analysis of Leakage Current to Predict Insulator Flashover Using Artificial Neural Network

Abstract

Problem statement: The phenomenon of flashover in polluted insulators has been continued by the study of the characteristics of contaminating layers deposited on the surface of insulators in high voltage laboratories. In the literature, Experimental investigations have been carried out on a real insulator or a flat plate model of insulators under high voltage application. This study proposed the Equivalent insulator flat plate model for studying the flashover phenomena due to pollution under wet conditions even at low voltage. Laboratory based tests were carried out on the model under AC voltage at different pollution levels. Different concentrations of salt solution has been prepared using sodium chloride, Kaolin and distilled water representing the various contaminations. Leakage current during the experimental studies were measured for various polluted conditions. Approach: A new model of Vc = f (V, I_initial, I_em, I_emax and I_σ) based on artificial neural network has been developed to predict flashover from the analysis of leakage current. The input variable to the artificial neural network are mean (I_mean), Maximum(Imax) and standard deviation(I_σ) of leakage current extracted along with the initial value of leakage current I_initial and the input voltage(V).The target obtained was used to evaluate the performance of the neural network model. Results: The optimum process has been carried out based on the training accuracy measured by RMSE, the network converged to a threshold of 0.0001.The trained model prediction is in good agreement with the actual results and the R² value of the developed model is 0.99996. Conclusion: The developed ANN model is well-suited for the analysis of leakage current to predict flashover on the insulator surface with high accuracy.