Implementasi Jaringan Syaraf Tiruan - Backpropagation dalam Pengenalan Flashover Isolator 150 kV

Studi Kasus : Payakumbuh – Koto Panjang

Abstract

The 150 kV Koto Panjang - Payakumbuh transmission line is a transmission line with a total length of 86 km. Environmental factors greatly influence the formation of a layer of contaminants on the surface of insulators, this is one of the causes of flashover in insulators. This research was carried out using the artificial neural network method - backpropagation in the introduction of 150 kV insulator flashover against temperature and humidity contaminated with moss and dust. So the flashover results obtained using artificial neural networks - backpropagation are compared with experimental flashover. Therefore, the implementation of an artificial neural network - backpropagation for the introduction of flashover of the 150 kV Payakumbuh - Koto Panjang insulator was successfully implemented by achieving flashover values with a small difference in the output so that the error obtained by training and testing the artificial neural network was also small. This means that the flashover voltage can be recognized by artificial neural network testing. The moss test parameter values are measured by the number of epochs with an average of 197.57, performance with an average of 17.8E-05 and gradient with an average of 27.38E-05. dust test parameters were measured by the number of epochs with an average of 101.85, performance with an average of 75.3E-06 and gradient with an average of 33.78E-05