PREDIKSI SUHU OUTLET PADA SISTEM FLUIDA KERJA REAKTOR NUKLIR DENGAN PENDEKATAN TIME SERIES MENGGUNAKAN METODE GATED RECURRENT UNIT (GRU)

Muhammad Zidane Zukhrufa, . (2026) PREDIKSI SUHU OUTLET PADA SISTEM FLUIDA KERJA REAKTOR NUKLIR DENGAN PENDEKATAN TIME SERIES MENGGUNAKAN METODE GATED RECURRENT UNIT (GRU). Skripsi thesis, Universitas Pembangunan Nasional Veteran Jakarta.

[img] Text
ABSTRAK.pdf
Download (309kB)
[img] Text
AWAL.pdf
Download (638kB)
[img] Text
BAB 1.pdf
Restricted to Repository UPNVJ Only
Download (279kB)
[img] Text
BAB 2.pdf
Restricted to Repository UPNVJ Only
Download (572kB)
[img] Text
BAB 3.pdf
Restricted to Repository UPNVJ Only
Download (504kB)
[img] Text
BAB 4.pdf
Restricted to Repository UPNVJ Only
Download (1MB)
[img] Text
BAB 5.pdf
Download (310kB)
[img] Text
DAFTAR PUSTAKA.pdf
Download (234kB)
[img] Text
RIWAYAT HIDUP.pdf
Restricted to Repository UPNVJ Only
Download (143kB)
[img] Text
LAMPIRAN.pdf
Restricted to Repository UPNVJ Only
Download (4MB)
[img] Text
HASIL PLAGIARISME.pdf
Restricted to Repository staff only
Download (8MB)
[img] Text
ARTIKEL KI.pdf
Restricted to Repository staff only
Download (713kB)

Abstract

Operational safety of nuclear reactors relies heavily on the stability of thermal parameters, particularly within the secondary cooling system. This study aims to develop a prediction model for the outlet temperature (T-CTB OUT) of a nuclear reactor's working fluid system using the Gated Recurrent Unit (GRU) method with a time series approach. The research data consists of 11,001 real-time data points obtained from a water heater prototype at the National Nuclear Energy Agency (BATAN). Through preprocessing and windowing experiments, it was found that a window size of 24 seconds yielded the most optimal performance. The model was trained using five selected input features based on correlation analysis and tested for robustness using anomaly injection scenarios involving sudden spikes and temperature drops. Evaluation results demonstrated very high accuracy with a Coefficient of Determination (R²) of 0.991, along with low error rates indicated by an MAE of 0.025 and RMSE of 0.033. In robustness testing, the model proved responsive in detecting extreme fluctuation patterns. Furthermore, the model was implemented into a web-based dashboard acting as an early warning system capable of forecasting temperature trends for the next 3,600 seconds. This study proves that the GRU method is effective as a decision support tool for reactor safety monitoring.

Item Type: Thesis (Skripsi)
Additional Information: [No.Panggil: 2110511112] [Pembimbing 1: Didit Widiyanto] [Pembimbing 2: Hamonangan Kinantan Prabu] [Penguji 1: Supriyanto] [Penguji 2: Ichsan Mardani]
Uncontrolled Keywords: Temperature Prediction, Nuclear Reactor, Gated Recurrent Unit, Time Series, Early Warning System.
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Divisions: Fakultas Ilmu Komputer > Program Studi Informatika (S1)
Depositing User: MUHAMMAD ZIDANE ZUKHRUFA
Date Deposited: 17 Mar 2026 02:26
Last Modified: 17 Mar 2026 02:26
URI: http://repository.upnvj.ac.id/id/eprint/42452

Actions (login required)

View Item View Item