Optimal Design of Medium Voltage Composite Insulator Profiles Using Finite Element Method and Artificial Neural Networks

Abstract

Polymeric insulators are extensively used in high-voltage applications due to their performance hydrophobic properties. Electrical field distribution along the insulator surface is considered one of the important parameters for the performance evaluation of outdoor insulators. Reliability of medium voltage composite insulators depends on achieving an optimal electrical and geometrical design that produces the lowest electric field strength. This paper proposes a method to find the optimal geometrical design of composite insulators for different voltage levels. This method combined Finite Element Method (FEM) and Artificial Neural Networks (ANN).First, it was used the FEM to evaluate the influence of insulator geometrical parameters on the electric field strength such as shed radius, core diameter, shed inclination angle and distance between the first shed and end fitting. The electric field strength was calculated by using the commercial software package COMSOL Multiphysics. Then, the ANN algorithm has been implemented using the MATLAB program to improve the electric field distribution of composite insulators by reducing the value of maximum electric field strength in order to optimize the original profile that produces the lowest electric field strength. The results of this study are able to provide theoretical support to design and select the suitable profile of composite insulators in order to obtain better performance.