A Novel Model for Prediction of Flashover 150kV Polluted Insulator Based on Nonlinear Autoregressive External Input Neural Network

Abstract

This study aims to use an artificial neural network to forecast the flashover voltage of a polluted high-voltage insulator. Practical tests were conducted on a high-voltage insulator to gather data for the neural network. These tests were carried out with varying levels of real contaminants from used insulators, with each level of contamination measured in milliliters. The collected data provides flashover voltage values corresponding to different pollution amounts and their conductivity in each insulator zone. The Nonlinear Autoregressive External Input Neural Network (NarxNet) is employed to predict the flashover voltage and assess the pollution state of the insulator. The results demonstrate that the NarxNet method achieves a 93.74% accuracy rate in predicting the flashover voltage of high-voltage insulators, compared to the results from practical tests.

Keywords

• flashover, prediction, contamination, insulator, neural network, esdd, nsdd.