Perancangan Sistem Klasifikasi Kondisi Operasi Feedwater Heater di PLTU Menggunakan Decision Tree Classifier

Robaitilah, Muhammad Chairusyakirin (2026) Perancangan Sistem Klasifikasi Kondisi Operasi Feedwater Heater di PLTU Menggunakan Decision Tree Classifier. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 02311940000162-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (9MB) | Request a copy

Abstract

Feedwater heater merupakan salah satu komponen penting pada Pembangkit Listrik Tenaga Uap (PLTU) yang berfungsi memanfaatkan uap ekstraksi turbin untuk meningkatkan temperatur air umpan sebelum memasuki boiler. Kinerja feedwater heater berpengaruh terhadap efisiensi termal dan keandalan operasi pembangkit. Penurunan kinerja yang disebabkan oleh perubahan kondisi operasi maupun degradasi peralatan terjadi secara bertahap sehingga tidak selalu dapat dideteksi secara langsung melalui pemantauan konvensional. Oleh karena itu, diperlukan suatu sistem klasifikasi berbasis data yang mampu mengidentifikasi kondisi operasi feedwater heater secara otomatis dan akurat. Penelitian ini bertujuan merancang sistem klasifikasi kondisi operasi feedwater heater di PLTU menggunakan metode Decision Tree Classifier serta mengevaluasi performa model klasifikasi yang dihasilkan. Perancangan sistem disimulasikan menggunakan perangkat lunak MATLAB dengan memanfaatkan data historis operasional feedwater heater tipe shell and tube pada PLTU Cirebon tahun 2024. Parameter masukan yang digunakan meliputi temperature water input, temperature water output, dan water level. Model yang dikembangkan menghasilkan empat kategori kondisi operasi, yaitu normal operation, schedule downtime, unschedule downtime, dan set up and adjustment. Hasil pemodelan menggunakan Decision Tree Classifier menunjukkan tingkat akurasi pada rentang 95,8% hingga 98,2% selama 1000 iterasi, dengan akurasi terbaik sebesar 98,2% yang diperoleh pada iterasi ke-327. Evaluasi menggunakan confusion matrix menghasilkan nilai accuracy sebesar 0,991, precision sebesar 0,982 (micro-average), recall sebesar 0,982 (micro-average), dan F1-score sebesar 0,982 (micro-average). Hasil tersebut menunjukkan bahwa sistem klasifikasi yang dirancang mampu mengidentifikasi kondisi operasi feedwater heater dengan tingkat performa yang sangat baik sehingga berpotensi menjadi dasar pengembangan sistem pemantauan kondisi berbasis data untuk mendukung strategi pemeliharaan preventif dan prediktif serta meningkatkan efisiensi dan keandalan operasi PLTU.
==================================================================================================================================
Feedwater heaters are essential components in steam power plant, functioning to utilize extraction steam from the turbine to increase the temperature of feedwater before it enters the boiler. The performance of a feedwater heater directly affects the thermal efficiency and operational reliability of the power plant. Performance degradation caused by changes in operating conditions and equipment deterioration occurs gradually, making it difficult to detect through conventional monitoring methods. Therefore, a data-driven classification system is required to automatically and accurately identify the operating condition of the feedwater heater. This study aims to design an operating condition classification system for a feedwater heater in a coal-fired power plant using the Decision Tree Classifier method and to evaluate the performance of the developed classification model. The proposed system was implemented and simulated using MATLAB based on historical operational data collected from a shell-and-tube feedwater heater at Cirebon Coal-Fired Power Plant in 2024. The input parameters consisted of water inlet temperature, water outlet temperature, and water level. The developed model classified operating conditions into four categories: normal operation, scheduled downtime, unscheduled downtime, and setup and adjustment. The Decision Tree Classifier achieved an accuracy ranging from 95.8% to 98.2% over 1,000 iterations, with the highest accuracy of 98.2% obtained at the 327th iteration. Performance evaluation using a confusion matrix yielded an accuracy of 0.991, a micro-average precision of 0.982, a micro-average recall of 0.982, and a micro-average F1-score of 0.982. These results demonstrate that the proposed classification system is capable of accurately identifying the operating conditions of a feedwater heater and has significant potential as the foundation for developing a data-driven condition monitoring system to support preventive and predictive maintenance strategies while improving the efficiency and reliability of coal-fired power plant operations.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Feedwater Heater, Decision Tree Classifier, Pembangkit Listrik Tenaga Uap (PLTU), MATLAB. Feedwater Heater, Decision Tree Classifier, Steam Power Plant, MATLAB.
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines.
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Physics Engineering > 30201-(S1) Undergraduate Thesis
Depositing User:	M Chairusyakirin Robaitilah
Date Deposited:	17 Sep 2026 06:01
Last Modified:	17 Sep 2026 06:01
URI:	http://repository.its.ac.id/id/eprint/144489

Actions (login required)

View Item