Estimasi Remaining Useful Life (RUL) pada Synchronous Hydro Generator Menggunakan Model Degradasi

Dirham, Adam (2026) Estimasi Remaining Useful Life (RUL) pada Synchronous Hydro Generator Menggunakan Model Degradasi. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5009201129-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (6MB) | Request a copy

Abstract

Pertumbuhan populasi dan perkembangan industri, kebutuhan akan listrik terus meningkat di Indonesia. Hydroelectric energy dianggap sebagai solusi yang andal dengan kelebihan kapasitas beban besar, fluktuasi daya minimal, dan dampak lingkungan rendah. Namun, untuk meningkatkan kinerja generator agar terus dapat beroperasi normal perlu meminimalkan terjadinya kesalahan. Sehingga diperlukan sebuah sistem deteksi kesalahan yang efektif. Kerusakan pada generator, seperti kumparan yang overheating dan berubahnya jarak pada sisi air gap, dapat mengganggu performa dan memicu ketidakseimbangan. Penelitian ini memiliki keterkaitan yang erat dengan beberapa tujuan dalam Sustainable Development Goals (SDGs). Kontribusi utama penelitian ini berada pada SDG 7: Affordable and Clean Energy, yaitu melalui upaya peningkatan keandalan pembangkit listrik tenaga air sebagai salah satu sumber energi terbarukan yang berperan penting dalam penyediaan energi bersih dan berkelanjutan. Penelitian ini juga mendukung SDG 9: Industry, Innovation and Infrastructure melalui pengembangan metode pemeliharaan prediktif berbasis model degradasi yang berkontribusi terhadap peningkatan keandalan infrastruktur energi, efisiensi operasional, serta inovasi dalam pengelolaan aset pembangkit listrik. Selain itu, penelitian ini berkontribusi pada SDG 12: Responsible Consumption and Production karena penerapan predictive maintenance membantu mengoptimalkan pemanfaatan peralatan, mengurangi kerusakan prematur, menekan kebutuhan penggantian komponen, dan meningkatkan efisiensi penggunaan sumber daya. Pada penelitian ini, peneliti membuat sistem deteksi kesalahan pada synchronous hydro generator yaitu kesalahan insulasi pada stator yang akan mengakibatkan pengurangan luas area penampang pada winding. Kesalahan akan disimulasikan dengan pemodelan state space untuk melihat perubahan respon dari sistem. Ekstraksi fitur akan digunakan untuk melakukan klasifikasi dari data arus keluaran dari generator akan dianalisa pada domain waktu yang kemudian didapatkan parameter health indicator. Kemudian model degradasi eksponensial akan digunakan sebagai model yang akan mengklasifikasikan pada kondisi yaitu normal dan failure berupa failure insulation. Didapatkan hasil bahwa model dapat memberikan nilai akurasi yang tinggi dengan nilai RMSE sebesar 0,873. Kesimpulan yang didapatkan adalah sistem deteksi kesalahan yang dibangun dengan menggunakan model degradasi eksponensial dapat mendeteksi kesalahan dengan baik.
======================================================================================================================================
The increasing population and rapid industrial development have led to a growing demand for electricity in Indonesia. Hydroelectric energy is considered a reliable solution due to its high load capacity, minimal power fluctuations, and low environmental impact. However, maintaining the reliable operation of hydro generators requires minimizing the occurrence of faults that may degrade their performance. Therefore, an effective fault detection system is essential. Generator faults, such as stator winding insulation degradation, can reduce the effective cross-sectional area of the winding, leading to increased resistance and degraded generator performance. This research is closely aligned with several Sustainable Development Goals (SDGs). Its primary contribution is to SDG 7 (Affordable and Clean Energy) by improving the reliability of hydropower generators as an important source of renewable energy. It also supports SDG 9 (Industry, Innovation and Infrastructure) through the development of a degradation model-based predictive maintenance approach that enhances energy infrastructure reliability, operational efficiency, and innovation in power plant asset management. Furthermore, this study contributes to SDG 12 (Responsible Consumption and Production) by promoting predictive maintenance strategies that optimize equipment utilization, reduce premature component failures, minimize replacement requirements, and improve resource efficiency. In this study, a fault detection system for a synchronous hydro generator was developed by modeling stator insulation failure using a state-space approach to evaluate changes in the system response. Feature extraction was performed on the stator armature current signals in the time domain to characterize the generator condition which health indicator parameters were subsequently obtained. An exponential degradation model was then employed to distinguish between normal and insulation fault conditions and to estimate the Remaining Useful Life (RUL) of the generator. The proposed model achieved satisfactory prediction performance with a Root Mean Square Error (RMSE) of 0.873. The results demonstrate that the exponential degradation model can effectively detect insulation faults and provide accurate Remaining Useful Life estimation for predictive maintenance applications in synchronous hydro generators.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	energi bersih dan terjangkau, health indicator, model degradasi eksponensial, remaining useful life, synchronous hydro generator, Affordable and clean energy, exponential degradation model, health indicators, remaining useful life, synchronous hydro generator
Subjects:	T Technology > T Technology (General) > T55 Industrial Safety
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Physics Engineering > 30201-(S1) Undergraduate Thesis
Depositing User:	Adam Dirham
Date Deposited:	01 Aug 2026 03:23
Last Modified:	01 Aug 2026 03:23
URI:	http://repository.its.ac.id/id/eprint/141212

Actions (login required)

View Item