Fauzi, Ammara Nayla (2026) Deteksi Busur Api Seri Pada Instalasi DC Fotovoltaik Menggunakan Kombinasi Wavelet Dan Artificial Neural Network. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5022221166_Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5022221166_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy |
Abstract
Sistem fotovoltaik (PV) semakin banyak dimanfaatkan sebagai sumber energi terbarukan karena ramah lingkungan dan mampu memenuhi kebutuhan energi listrik secara berkelanjutan. Namun, peningkatan penggunaannya juga diikuti oleh potensi gangguan pada sisi arus searah (DC), salah satunya busur api seri. Gangguan ini dapat disebabkan oleh koneksi yang longgar, kerusakan kabel, maupun korosi pada komponen instalasi. Busur api seri sulit dideteksi karena menghasilkan sinyal arus yang bersifat tidak stasioner dan menyerupai kondisi operasi normal. Oleh karena itu, diperlukan metode yang mampu mengidentifikasi karakteristik busur api secara akurat. Penelitian ini bertujuan mengembangkan metode deteksi busur api seri menggunakan kombinasi Discrete Wavelet Transform (DWT) dan Artificial Neural Network (ANN). Data penelitian berupa sinyal arus DC dengan frekuensi sampling 10 kHz yang diproses melalui tahap segmentasi, windowing, dan pelabelan menjadi empat kondisi, yaitu Normal, Starting Arc, During Arc, dan Ending Arc. Transformasi DWT menggunakan wavelet Daubechies orde 4 (db4) hingga level dekomposisi keempat. Koefisien detail level empat (D4) digunakan untuk mengekstraksi lima fitur, yaitu energi, nilai maksimum, nilai rata-rata, I_2^' dan ΔI. Seluruh fitur tersebut digunakan sebagai masukan pada ANN untuk mengklasifikasikan kondisi sinyal. Hasil pengujian menunjukkan bahwa metode yang diusulkan memperoleh akurasi klasifikasi sebesar 96,9%. Hasil tersebut menunjukkan bahwa kombinasi DWT dan ANN mampu mengidentifikasi karakteristik busur api seri dengan baik serta berpotensi diterapkan sebagai metode deteksi untuk meningkatkan keandalan dan keselamatan instalasi fotovoltaik.
=======================================================================================================================================
Photovoltaic (PV) systems are increasingly used as a renewable energy source because they are environmentally friendly and capable of providing sustainable electrical power. However, their widespread use also increases the risk of disturbances on the direct current (DC) side, one of which is a series arc fault. This fault may result from loose electrical connections, damaged cables, or corrosion of installation components. Series arc faults are difficult to detect because they generate non-stationary current signals that resemble normal operating conditions. Therefore, an accurate detection method is required. This study proposes a series arc fault detection method based on the combination of the Discrete Wavelet Transform (DWT) and an Artificial Neural Network (ANN). DC current signals with a sampling frequency of 10 kHz were processed through segmentation, windowing, and labeling into four conditions: Normal, Starting Arc, During Arc, and Ending Arc. The DWT employed the Daubechies 4 (db4) wavelet up to the fourth decomposition level. The fourth-level detail coefficients (D4) were used to extract five features: energy, maximum value, mean value, I_2' and ΔI. These features were used as inputs to the ANN for signal classification. The proposed method achieved a classification accuracy of 96.9%, demonstrating that the combination of DWT and ANN can effectively identify series arc faults and improve the reliability and safety of photovoltaic installations.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | Busur Api Seri, Sistem Fotovoltaik, Discrete Wavelet Transform, Artificial Neural Network, DC ============================================================ Series Arc Fault, Photovoltaic System, Discrete Wavelet Transform, Artificial Neural Network, DC |
| Subjects: | T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7882.P3 Pattern recognition systems |
| Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20201-(S1) Undergraduate Thesis |
| Depositing User: | Ammara Nayla Fauzi |
| Date Deposited: | 22 Jul 2026 06:17 |
| Last Modified: | 22 Jul 2026 06:17 |
| URI: | http://repository.its.ac.id/id/eprint/136445 |
Actions (login required)
 |
View Item |