Desain Adaptive Defense Scheme (ADS) Pada Sistem Kelistrikan Tambora Berbasis Artificial Neural Network

Utomo, Wahyu (2025) Desain Adaptive Defense Scheme (ADS) Pada Sistem Kelistrikan Tambora Berbasis Artificial Neural Network. Masters thesis, Institut Terknologi Sepuluh Nopember.

Text 6022231024-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (14MB) | Request a copy

Abstract

Sistem tenaga listik Tambora terdiri dari jaringan transmisi 70 kV dan 150 kV, jaringan distribusi 20 kV, serta unit pembangkit listrik yang beroperasi pada tegangan 20 kV, 70 kV, dan 150 kV. Keandalan dan kestabilan sistem sangat penting untuk mencegah terjadinya pemadaman listrik yang meluas dan memerlukan waktu pemulihan yang lama. Saat ini, skema pertahanan yang digunakan masih bersifat konvensional, yaitu dengan mengandalkan relay under-frequency untuk pelepasan beban (load shedding) yang sering kali tidak cukup efektif dalam mencegah terjadinya gangguan berantai (cascading failure). Untuk mengatasi permasalahan tersebut, penelitian ini mengusulkan perancangan Adaptive Defense Scheme (ADS) berbasis Artificial Neural Network (ANN) untuk sistem Tambora. Metodologi yang diusulkan meliputi simulasi berbagai skenario gangguan menggunakan perangkat lunak DIgSILENT PowerFactory-Python Interfacing, dilanjutkan dengan pelatihan ANN model Multi Layer Perceptron (MLP). Model ini mempertimbangkan parameter-parameter utama seperti kapasitas pembangkitan, beban sistem, penurunan frekuensi, dan cadangan putar (spinning reserve). Hasil penelitian menunjukkan bahwa metode yang diusulkan mampu menghasilkan nilai load shedding yang cukup akurat dibandingkan dengan metode konvensional dengan performa MSE 0,0338, MAE 0,1158 dan R² Score 0,9998 serta memiliki waktu eksekusi prediksi 0,1464 detik. Dengan ADS ANN MLP dapat membuat frekuensi steady state lebih cepat di 49,89 Hz pada waktu 5,326 s dibandingkan UFR eksisting dengan frekuensi steady state di 49,69 Hz pada waktu 6,023 detik.
========================================================================================================================
The Tambora power system consists of 70 kV and 150 kV transmission networks, a 20 kV distribution network, and power generation units operating at voltages of 20 kV, 70 kV, and 150 kV. The reliability and stability of the system are crucial to prevent widespread power outages that require lengthy recovery times. Currently, the defense scheme in use remains conventional, relying on under-frequency relays for load shedding, which is often insufficient to prevent cascading failures. To address this issue, this study proposes the design of an Adaptive Defense Scheme (ADS) based on Artificial Neural Networks (ANN) for the Tambora system. The proposed methodology involves simulating various disturbance scenarios using DIgSILENT PowerFactory integrated with Python, followed by training a Multilayer Perceptron (MLP) ANN model. This model takes into account key parameters such as generation capacity, system load, frequency drop, and spinning reserve. The results show that the proposed method can produce highly accurate load shedding values compared to conventional methods, achieving a performance of MSE 0.0338, MAE 0.1158, and an R² score of 0.9998, with an average prediction execution time of 0.1385 seconds. The ANN-MLP-based ADS is also capable of restoring the system to steady-state frequency faster, reaching 49.89 Hz at 5.326 seconds, compared to the existing UFR scheme which stabilizes at 49.69 Hz in 6.023 seconds.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Adaptive Defense Scheme, Artificial Neural Network, Kestabilan Sistem, Pelepasan Beban, Kestabilan Frekuensi, Power System Stability, Load Shedding, Frequency Stability
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK1010 Electric power system stability. Electric filters, Passive.
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20101-(S2) Master Thesis
Depositing User:	Wahyu Utomo
Date Deposited:	24 Jul 2025 04:33
Last Modified:	24 Jul 2025 04:33
URI:	http://repository.its.ac.id/id/eprint/121193

Actions (login required)

View Item