Jannah, Namira Roudlotul Jannah (2025) Metode Deteksi Busur Api DC Seri Pada Sistem Fotovoltaik Menggunakan Artificial Neural Netwok. Masters thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 6022231053-Master_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
6022231053-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (8MB) | Request a copy |
Abstract
Perkembangan energi terbarukan, khususnya sistem fotovoltaik (PV), menawarkan solusi ramah lingkungan untuk memenuhi kebutuhan energi di masa depan. Namun, tantangan utama dalam penerapan sistem ini adalah adanya risiko gangguan yang dapat membahayakan keamanan, salah satunya adalah munculnya busur api seri pada rangkaian DC yang tidak terdeteksi oleh sistem pengaman konvensional seperti Miniature Circuit Breaker (MCB). Gangguan busur api ini dapat menyebabkan kebakaran, kerusakan komponen, dan menurunkan kestabilan sistem. Untuk mengatasi hal tersebut, diperlukan metode deteksi dini yang efektif dan akurat. Penelitian ini bertujuan untuk mengembangkan metode deteksi busur api seri pada sistem PV menggunakan pendekatan kecerdasan buatan, khususnya Artificial Neural Network (ANN). Karakteristik busur api diidentifikasi melalui analisis sinyal arus dan tegangan yang diambil dari simulasi sumber tegangan PSU dan PV. Data sinyal tersebut diproses menggunakan metode Fast Fourier Transform (FFT) untuk mengekstraksi fitur frekuensi yang khas, terutama pada rentang 0–800 Hz, dimana harmonisa ganjil umumnya muncul selama gangguan busur api. Selanjutnya, fitur tersebut diekstraksi menggunakan teknik cross correlation untuk menentukan sepuluh frekuensi terbaik yang akan digunakan sebagai input pada model ANN. Sistem ini kemudian dilatih dan diuji dengan data simulasi dan riil, menunjukkan kemampuan dalam mengenali kondisi busur api dengan akurasi sebesar 79,3%. Hasil penelitian ini menunjukkan bahwa kombinasi analisis FFT dan ANN dapat menjadi solusi efektif dalam deteksi gangguan busur api secara real-time, sehingga berpotensi meningkatkan keamanan dan keandalan sistem fotovoltaik.
=================================================================================================================================
The development of renewable energy, particularly photovoltaic (PV) systems, offers an environmentally friendly solution to meet future energy demands. However, a significant challenge in implementing these systems is the risk of disturbances that may compromise safety, one of which is the occurrence of series arc faults in DC circuits that are not detected by conventional protection systems such as Miniature Circuit Breakers (MCB). These arc faults can lead to fires, component damage, and reduced system stability. To address this issue, an effective and accurate early detection method is required. This research aims to develop a series arc fault detection method in PV systems using an artificial intelligence approach, specifically Artificial Neural Networks (ANN). The characteristics of arc faults are identified through the analysis of current and voltage signals obtained from PSU and PV voltage source simulations. These signal data are processed using the Fast Fourier Transform (FFT) method to extract distinctive frequency features, particularly within the 0–800 Hz range, where odd harmonics commonly appear during arc faults. Subsequently, these features are extracted using the crosscorrelation technique to determine the ten most significant frequencies that will be used as input for the ANN model. The system is then trained and tested with both simulated and real data, demonstrating its capability to recognize arc fault conditions with an accuracy of 79,3%. The findings of this study indicate that the combination of FFT analysis and ANN can serve as an effective real-time solution for arc fault detection, thereby enhancing the safety and reliability of photovoltaic systems.
Actions (login required)
 |
View Item |