Artificial Neural Network Approach for Prediction of Leakage Current of polymeric insulator under Non-Uniform Fan-shaped Contamination

Abstract

The accumulation of environmental pollution in symmetrical and asymmetrical categories conditions can lead to the aging of polymeric insulators through leakage current enhancement. In this paper, real asymmetric contamination, namely fan-shaped, is investigated accurately. However, the main approaches are that the real configuration of fan-shaped pollution is not normally uniform, and asymmetrically non-uniform fan-shaped pollution is detected. In particular, the segment angle, thickness, and electrical conductivity of the contamination layer are three major effective parameters in the insulation analysis of a fan-shaped polluted insulator. Finite element method (FEM) simulation and the proposed hybrid FEM combined with the method of moment (MOM) of a 20 kV polymeric insulator under asymmetric non-uniform fan-shaped contamination through leakage current calculations illustrate that variation in non-uniformity of related parameters of contamination layer are so crucial in the insulation situation of insulator. Also, leakage current is predicted using an artificial neural network (ANN) in processing obtained data from proposed methods, by applying effective parameters of non-uniform fan-shaped contamination. Experimental tests have been performed to validate the results and evaluate the leakage current of the fan-shaped polluted insulator. The results of the predicted values show that the difference is less than 5% compared to the actual values.