Analisa Karakterisasi Pola Partial Discharge Pada Kubikel 20 KV Menggunakan Kamera Akustik

Kuncoro, G. Satrio (2026) Analisa Karakterisasi Pola Partial Discharge Pada Kubikel 20 KV Menggunakan Kamera Akustik. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6022241039-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (10MB) | Request a copy

Abstract

Partial Discharge (PD) merupakan pelepasan muatan lokal yang dapat menyebabkan degradasi sistem isolasi pada kubikel tegangan menengah. Penelitian ini bertujuan mengkarakterisasi pola PD pada kubikel 20 kV menggunakan kamera akustik melalui analisis Short-Time Fourier Transform (STFT) dan Mel-Spektrogram. Pengujian dilakukan pada lima kondisi, yaitu Normal, Void, Corona, Surface, dan Arc, dengan masing-masing 10 sampel audio. Sinyal dipraproses menggunakan bandpass filter 5–20 kHz, kemudian ditransformasikan ke STFT dan Mel-Spektrogram serta diekstraksi menggunakan fitur RMS Audio, Spectral Centroid, Band Energy Ratio (BER), Spectral Entropy, Mel Energy, dan Mel Entropy.
Hasil penelitian menunjukkan bahwa setiap jenis PD memiliki karakteristik spektral yang berbeda. Arc menghasilkan energi akustik tertinggi dengan RMS95 sebesar 0,020975, sedangkan Corona memiliki nilai Spectral Centroid (≈11,1 kHz), BER (0,509), dan Mel Energy (45,09) tertinggi. Void memiliki Spectral Entropy tertinggi, sementara Surface menunjukkan Mel Entropy tertinggi. Hasil tersebut menunjukkan bahwa kombinasi fitur berbasis STFT dan Mel-Spektrogram mampu mengkarakterisasi berbagai jenis Partial Discharge secara efektif serta berpotensi mendukung deteksi dini degradasi isolasi dan pemeliharaan prediktif pada kubikel 20 kV.
======================================================================================================================================
Partial discharge (PD) is a localized electrical discharge that can degrade the insulation system of medium-voltage switchgear. This study aims to characterize PD patterns in 20 kV switchgear using an acoustic camera based on Short-Time Fourier Transform (STFT) and Mel-spectrogram analysis. Five conditions (Normal, Void, Corona, Surface, and Arc) were investigated using 10 audio samples for each category. The recorded signals were preprocessed with a 5–20 kHz band-pass filter, transformed into STFT and Mel-spectrogram representations, and analyzed using RMS Audio, Spectral Centroid, Band Energy Ratio (BER), Spectral Entropy, Mel Energy, and Mel Entropy.
The results show that each PD type exhibits distinct spectral characteristics. Arc produced the highest acoustic energy (RMS95 = 0.020975), while Corona achieved the highest Spectral Centroid (≈11.1 kHz), BER (0.509), and Mel Energy (45.09). Void exhibited the highest Spectral Entropy, whereas Surface showed the highest Mel Entropy. These findings demonstrate that the combined use of STFT- and Mel-based features effectively characterizes different PD types and provides a promising approach for early insulation degradation detection and predictive maintenance of 20 kV switchgear.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Partial Discharge, kamera akustik, STFT, Mel-Spektrogram, Ekstraksi Fitur, Kubikel 20 kV, Partial Discharge, Acoustic Camera, STFT, Mel-Spectrogram, Feature Extraction, 20 kV Switchgear
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK201 Electric Power Transmission T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK3401 Insulation and insulating materials
Divisions:	Faculty of Electrical Technology > Electrical Engineering > 20101-(S2) Master Thesis
Depositing User:	G. Satrio Kuncoro
Date Deposited:	31 Jul 2026 06:35
Last Modified:	31 Jul 2026 06:35
URI:	http://repository.its.ac.id/id/eprint/140543

Actions (login required)

View Item