Sistem Deteksi Anomali Data Suhu Dan Tekanan Pada Pembangkit Listrik Tenaga Mesin Gas (PLTMG) Menggunakan Metode LSTM Autoencoder

Naufaldhianto, Enggal Nur F (2025) Sistem Deteksi Anomali Data Suhu Dan Tekanan Pada Pembangkit Listrik Tenaga Mesin Gas (PLTMG) Menggunakan Metode LSTM Autoencoder. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2040211002-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (6MB)

Abstract

PLTMG Bima merupakan pembangkit listrik berkapasitas 50 MW yang dikelola oleh PT PLN Nusantara Power Services. Sistem monitoring pada pembangkit ini hanya menampilkan data sensor suhu dan tekanan tanpa analisis lebih lanjut, sehingga menyulitkan deteksi dini terhadap gangguan operasional. Salah satu gangguan yang pernah terjadi adalah shutdown engine akibat tertutupnya valve suplai bahan bakar berdasarkan perintah sistem alarm yang bekerja berdasarkan ambang batas tetap. Proyek akhir ini mengembangkan sistem deteksi anomali berbasis metode Long Short-Term Memory (LSTM) Autoencoder untuk mengenali pola data suhu dan tekanan serta mendeteksi penyimpangan meskipun nilai parameter suhu dan tekanan masih berada dalam batas normal. Model dilatih menggunakan data historis dari PLTMG Bima Unit 2 dan diuji dengan data real-time. Konfigurasi terbaik diperoleh pada model dengan 4 layer, 100 epoch, batch size 32, time step 10, dan fungsi aktivasi tanh, menghasilkan nilai Mean Squared Error (MSE) terkecil sebesar 0.00041. Threshold sebesar 0.001656 ditentukan berdasarkan distribusi error pada data validasi dan digunakan dalam proses deteksi anomali. Sistem ini telah diintegrasikan ke dalam dashboard berbasis web dan memberikan notifikasi otomatis saat terdeteksi anomali. Berdasarkan hasil pengujian, sistem mampu mengidentifikasi anomali secara tepat, salah satunya pada parameter Alternator Stator Winding Temperature Phase W. Sehingga sistem dapat berfungsi sebagai alat bantu informasi awal untuk mendeteksi potensi gangguan pada pembangkit.
==============================================================================================================================================
PLTMG Bima is a 50 MW power plant operated by PT PLN Nusantara Power Services. The monitoring system in this power plant only displays sensor data for temperature and pressure without further analysis, making it difficult to detect operational disturbances at an early stage. One of the disturbances that previously occurred was an engine shutdown caused by the closure of the fuel supply valve, triggered by an alarm system that operates based on fixed threshold values. This final project develops an anomaly detection system based on the Long Short-Term Memory (LSTM) Autoencoder method to learn the patterns of temperature and pressure data and detect deviations, even when the parameter values remain within normal limits. The model was trained using historical data from PLTMG Bima Unit 2 and tested with real-time data. The best configuration was obtained with a model consisting of 4 layers, 100 epochs, a batch size of 32, a time step of 10, and the tanh activation function, resulting in the lowest Mean Squared Error (MSE) value of 0.00041. A threshold of 0.001656 was determined based on the distribution of reconstruction errors in the validation data and was used in the anomaly detection process. This system has been integrated into a web-based dashboard and provides automatic notifications when anomalies are detected. Based on the testing results, the system was able to accurately identify anomalies, including one detected in the Alternator Stator Winding Temperature Phase W parameter. Thus, the system can serve as an early information tool to help detect potential operational disruptions in the power plant.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Deteksi Anomali, LSTM Autoencoder, Peringatan Dini, PLTMG
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5105.888 Web sites--Design. Web site development. T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5105 Data Transmission Systems
Divisions:	Faculty of Vocational > 36304-Automation Electronic Engineering
Depositing User:	Enggal Nur F. Naufaldhianto
Date Deposited:	04 Aug 2025 01:58
Last Modified:	04 Aug 2025 01:58
URI:	http://repository.its.ac.id/id/eprint/125930

Actions (login required)

View Item