Deteksi Anomali Data Suhu Pada Sistem Turbin Untuk Meningkatkan Keandalan Operasioanal Pembangkit Listrik Tenaga Air (PLTA) Asahan Menggunakan Isolation Forest

Ali, Fahreza (2025) Deteksi Anomali Data Suhu Pada Sistem Turbin Untuk Meningkatkan Keandalan Operasioanal Pembangkit Listrik Tenaga Air (PLTA) Asahan Menggunakan Isolation Forest. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2040211006-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Pembangkit Listrik Tenaga Air (PLTA) Asahan merupakan sumber energi listrik berkelanjutan di Indonesia. Turbin berfungsi mengubah energi kinetik air menjadi energi listrik. Keandalan turbin, terutama komponen kritis seperti Turbine Guide Pad, sangat dipengaruhi oleh pemantauan suhu. Lonjakan suhu hingga 104,2°C pada turbine guide pad di PLTA Asahan menunjukkan perlunya metode deteksi anomali yang lebih adaptif dan akurat, serta kemampuan sistem untuk membedakan anomali suhu akibat gangguan operasional dari penyimpangan yang disebabkan oleh noise sensor atau kesalahan komunikasi data. Rumusan masalah dalam penelitian ini adalah bagaimana mendeteksi anomali suhu pada turbin PLTA Asahan secara real-time untuk mencegah kerusakan dini yang dapat mempengaruhi efisiensi operasional dan keselamatan sistem. Penelitian ini bertujuan mengembangkan sistem deteksi anomali suhu operasional turbin dengan menggunakan metode Isolation Forest, yang lebih cepat dan akurat dibandingkan metode konvensional. Metodologi penelitian mencakup akuisisi data suhu secara real-time menggunakan sensor RTD dengan frekuensi sampling 1 Hz satu data per detik dan protokol komunikasi MQTT dengan QoS level 2, memastikan pengiriman data dengan delay sistem yang sangat rendah, yaitu 150 milidetik per data point. Akurasi deteksi model mencapai 99,2% untuk rentang suhu 30-120°C, menjamin ketepatan dalam mendeteksi anomali. Proses penelitian terdiri dari beberapa tahapanyaitu, pengambilan data suhu real-time, penyimpanan di database MySQL dengan timestamp presisi, preprocessing data untuk ekstraksi fitur delta suhu, pelatihan model Isolation Forest untuk deteksi anomali, dan validasi hasil. Isolation Forest dipilih karena kemampuannya mendeteksi outlier dalam ruang fitur multidimensi dengan kompleksitas komputasi rendah. Hasil penelitian menunjukkan performa model dalam mendeteksi anomali pada Upper Guide Pad, dengan F1-score sebesar 0,94 dan AUC sebesar 0,859. Waktu respons peringatan tercatat 200 milidetik setelah deteksi anomali. Implementasi pada dashboard Grafana sebagai monitoring real-time dan memberikan peringatan dini sebelum suhu mencapai batas kritis. Penelitian ini berkontribusi pada pengembangan sistem pemantauan suhu turbin, khususnya pada komponen RTD, dengan kemampuan mendeteksi anomali secara real-time menggunakan algoritma Isolation Forest.
======================================================================================================================================
The Asahan Hydroelectric Power Plant (PLTA Asahan) is a sustainable source of electrical energy in Indonesia. Turbines function to convert the kinetic energy of water into electrical energy. The reliability of turbines, particularly critical components such as the Turbine Guide Pad, is significantly influenced by temperature monitoring. A temperature spike of up to 104.2°C on the turbine guide pad at PLTA Asahan underscores the need for more adaptive and accurate anomaly detection methods, along with the system's ability to distinguish temperature anomalies caused by operational disturbances from deviations due to sensor noise or data communication errors. The problem statement of this research is how to detect temperature anomalies in the turbines of PLTA Asahan in real-time to prevent early damage that could affect operational efficiency and system safety. This research aims to develop a real-time temperature anomaly detection system for operational turbines using the Isolation Forest method, which is faster and more accurate compared to conventional methods. The research methodology includes real-time temperature data acquisition using RTD sensors with a sampling frequency of 1 Hz (one data point per second) and the MQTT communication protocol with QoS level 2, ensuring very low system data transmission delays of 150 milliseconds per data point. The model achieves a detection accuracy of 99.2% for a temperature range of 30-120°C, ensuring precision in anomaly detection. The research process consists of several stages: real-time temperature data collection, storage in a MySQL database with precise timestamps, data preprocessing for delta temperature feature extraction, training of the Isolation Forest model for anomaly detection, and validation of results. Isolation Forest was chosen for its ability to detect outliers in a multidimensional feature space with low computational complexity. The results demonstrate the model's performance in detecting anomalies on the Upper Guide Pad, achieving an F1-score of 0.94 and an AUC of 0.859. The warning response time is recorded at 200 milliseconds after the detection of an anomaly. Implementation on the Grafana dashboard enables real-time monitoring and provides early warnings before temperatures reach critical limits. This research contributes to the development of temperature monitoring systems for turbines, particularly for RTD components, with the capability of real-time anomaly detection using the Isolation Forest algorithm.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	PLTA Asahan, Suhu Turbin, Deteksi Anomali, Isolation Forest, Monitoring Real-time, PLTA Asahan, Turbine Temperature, Anomaly Detection, Isolation Forest, Real-Time Monitoring.
Subjects:	Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science) Q Science > QC Physics > QC271 Temperature measurements T Technology > TJ Mechanical engineering and machinery > TJ266 Turbines. Turbomachines (General) T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK1007 Electric power systems control T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK1519.S68 Hydroelectric power plants
Divisions:	Faculty of Vocational > 36304-Automation Electronic Engineering
Depositing User:	Fahreza Ali
Date Deposited:	05 Aug 2025 06:54
Last Modified:	05 Aug 2025 06:54
URI:	http://repository.its.ac.id/id/eprint/127319

Actions (login required)

View Item