Perancangan Sistem Pengendalian Daya Reaktor Nuklir Menggunakan Jaringan Syaraf Tiruan (JST) Levenberg Marquardt Di Pusat Teknologi Nuklir Bahan Dan Radiometri Badan Tenaga Nuklir Nasional (PTNBR Batan) Bandung - Design Of Nuclear Power Control System Based On Artificial Neural Network Levenberg Marquardt At The Nuclear Technology Center For Materials And Radiometry - National Nuclear Energy Agency (PTNBR Batan) Bandung

Kloatubun, Margaretha Maria Lucia (2011) Perancangan Sistem Pengendalian Daya Reaktor Nuklir Menggunakan Jaringan Syaraf Tiruan (JST) Levenberg Marquardt Di Pusat Teknologi Nuklir Bahan Dan Radiometri Badan Tenaga Nuklir Nasional (PTNBR Batan) Bandung - Design Of Nuclear Power Control System Based On Artificial Neural Network Levenberg Marquardt At The Nuclear Technology Center For Materials And Radiometry - National Nuclear Energy Agency (PTNBR Batan) Bandung. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 2406100092-Undergraduate Thesis.pdf]
Preview
Text
2406100092-Undergraduate Thesis.pdf - Published Version

Download (1MB) | Preview
[thumbnail of 2406100092-Paper.pdf]
Preview
Text
2406100092-Paper.pdf - Published Version

Download (319kB) | Preview
[thumbnail of 2406100092-Presentation.pdf]
Preview
Text
2406100092-Presentation.pdf - Published Version

Download (825kB) | Preview

Abstract

Di dalam tanki reaktor terjadi reaksi fisi antara bahan bakar nuklir dengan neutron. Reaksi fisi inilah yang menyebabkan naik-turunnya daya reaktor nuklir. Untuk mengendalikan daya reaktor nuklir digunakan 5 (lima) batang kendali yang digerakkan naik-turun di dalam tanki reaktor. Oleh karena di dalam tanki reaktor terdapat proses yang kompleks dan non-linear maka pemodelan dilakukan dengan Jaringan Syaraf Tiruan dengan struktur Multi Layer Perceptron (MLP). Struktur model yang diturunkan adalah struktur Nonlinear Auto Regressive with eXternal input (NARX). Pengaturan bobot jaringan syaraf tiruan dilakukan menggunakan algoritma Levenberg-Marquardt yang mampu memberikan hasil dengan RMSE dan VAF cukup baik, yakni 0.0209 dan sebesar 98.8682. Setelah didapatkan model tanki reaktor, selanjutnya dirancang sistem kontrol daya reaktor dengan metode Direct Invers control berbasis Jaringan Syaraf Tiruan. Dari hasil simulasi menunjukkan sistem direct invers control berbasis Jaringan Syaraf Tiruan memiliki respon yang sangat baik. Pengendalian mengikuti set point dengan maximum overshoot 0 % untuk semua set point.
========================================================================================================================
Fission reaction occurs in the inside of reactor tank between the nuclear fuel and neutrons. Fission reaction causes rise and fall of nuclear power reactors. In order to control the nuclear power reactor used five control rods which is moved up and down inside the reactor tank. There is a complex process and non-linear modeling in the reactor tank, conducted modeling reactor tank with Artificial Neural Networks with the structure of the Multi Layer Perceptron (MLP). The structure is a structure model derived from Nonlinear Auto Regressive with external input (NARX). Weights controlling of neural networks was carried out using Levenberg-Marquardt algorithm, which could give good results with RMSE and VAF good enough, ie 0.0209 and 98.8682. After getting the tank reactor model, then the reactor power control system designed using the Direct Inverse Control based on Artificial Neural Networks. The simulation results show a direct inverse control system based on neural networks have a good response. Direct inverse control follow the set point with maximum overshoot 0% for all set point.

Item Type: Thesis (Undergraduate)
Additional Information: RSF 006.32 Klo p
Uncontrolled Keywords: nuclear power, reactor tank, artificial neural networks, daya reaktor, tanki reaktor, jaringan syaraf tiruan
Subjects: Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions: Faculty of Industrial Technology > Physics Engineering
Depositing User: ansi aflacha
Date Deposited: 27 Nov 2019 08:30
Last Modified: 27 Nov 2019 08:30
URI: http://repository.its.ac.id/id/eprint/72091

Actions (login required)

View Item View Item