Implementasi Convolutional Neural Network Untuk Klasifikasi Hasil Inspeksi Isolator Kaca Tower Sutt 150 KV

Priambodo, Andhika Rizki (2025) Implementasi Convolutional Neural Network Untuk Klasifikasi Hasil Inspeksi Isolator Kaca Tower Sutt 150 KV. Masters thesis, INSTITUT TEKNOLOGI SEPULUH NOPEMBER.

Text 6022231003-Master_Thesis.pdf Restricted to Repository staff only Download (2MB) | Request a copy

Abstract

Untuk mendeteksi potensi gangguan pada jaringan transmisi 150 kV, PT. PLN (Persero) secara rutin melakukan pemeriksaan isolator menggunakan metode Climb Up Inspection (CUI). Seiring kemajuan teknologi, saat ini CUI dapat dilaksanakan menggunakan Unmanned Aerial Vehicle (UAV) atau drone, yang secara signifikan meningkatkan kecepatan inspeksi visual. Namun, peningkatan tersebut belum diimbangi oleh percepatan proses analisis dan pengolahan data gambar. Penelitian ini mengembangkan algoritma berbasis CNN yang mampu mengklasifikasi gambar isolator yang dipotret menggunakan Drone, kedalam dua kategori, yaitu normal dan pecah. Program yang dibuat mencapai akurasi validasi sebesar 99,09%, dengan precision sebesar 99,44%, recall sebesar 99,50%, dan F1-score sebesar 99,46%.
=================================================================================================================================
To detect potential faults in the 150 kV transmission network, PT. PLN (Persero) routinely conducts insulator inspections using the Climb Up Inspection (CUI) method. With advancements in technology, CUI can now be performed using Unmanned Aerial Vehicles (UAVs) or drones, significantly improving the efficiency of visual inspections. However, this improvement in data acquisition has not been matched by the speed of data analysis and processing. This study proposes a Convolutional Neural Network (CNN)-based algorithm capable of classifying insulator images captured by drones into two categories: normal and damaged. The developed system achieved a validation accuracy of 99.09%, with a precision of 99.44%, recall of 99.50%, and an F1-score of 99.46%. These results demonstrate the system's potential to accelerate the analysis of inspection images and support more efficient maintenance of transmission infrastructure.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Isolator, CNN, Drone, Climb Up Inspection, Insulator, CNN, Drone, Climb Up Inspection
Subjects:	T Technology > T Technology (General) > T57.5 Data Processing
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20101-(S2) Master Thesis
Depositing User:	Andhika Rizki Priambodo
Date Deposited:	28 Jul 2025 05:57
Last Modified:	28 Jul 2025 05:57
URI:	http://repository.its.ac.id/id/eprint/122013

Actions (login required)

View Item