Analisis Perbandingan Performansi dan Fitur Signifikan Untuk Genset dengan Frekuensi 50hz dan 60hz dengan Pendekatan Machine Learning

Muhammad, Bendrad (2025) Analisis Perbandingan Performansi dan Fitur Signifikan Untuk Genset dengan Frekuensi 50hz dan 60hz dengan Pendekatan Machine Learning. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6010231013-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2027. Download (5MB) | Request a copy

Abstract

Penelitian ini bertujuan untuk menganalisis perbandingan performansi genset pada frekuensi 50 Hz dan 60 Hz serta mengidentifikasi fitur-fitur signifikan yang memengaruhi kinerja sistem menggunakan pendekatan machine learning. Fokus utama penelitian mencakup analisis efisiensi bahan bakar yang diukur melalui parameter heat rate serta evaluasi parameter kritis yang berpotensi menyebabkan kerusakan mekanis jangka panjang. Metode yang digunakan mencakup kombinasi algoritma seleksi fitur, yaitu Particle Swarm Optimization, Chi-Square, dan Optimized Selection, dengan algoritma klasifikasi, yaitu Decision Tree, Random Forest, Naive Bayes, dan Deep Learning. Hasil penelitian menunjukkan bahwa kombinasi algoritma Optimized Selection dan Decision Tree menghasilkan akurasi terbaik sebesar 93,89%. Fitur-fitur utama yang diidentifikasi sebagai signifikan meliputi Active Power, Turbocharger Bank A Speed, Exhaust Gas Temperature TC A Outlet, Exhaust Gas Temperature TC B Inlet, dan HT Water Temperature, Jacket Outlet. Selain itu, penelitian ini mengungkapkan bahwa perbedaan frekuensi 50 Hz dan 60 Hz tidak memberikan kontribusi signifikan terhadap efisiensi genset. Namun, instalasi pada frekuensi 60 Hz menunjukkan potensi risiko yang lebih besar terhadap kerusakan mekanis jangka panjang pada komponen utama genset. Sebagai langkah mitigasi, penelitian ini merekomendasikan optimalisasi tekanan udara pembakaran, rasio udara-bahan bakar, dan sistem pendinginan untuk meningkatkan keandalan serta efisiensi operasional genset. Hasil penelitian ini diharapkan dapat menjadi acuan dalam pengembangan sistem operasional dan perawatan genset yang diaplikasikan pada frekuensi yang berbeda, sehingga dapat mendukung keberlanjutan operasional yang lebih andal dan efisien.
==============================================================================================================================
This study aims to analyze the performance comparison of generators operating at 50 Hz and 60 Hz frequencies and to identify significant features influencing system performance using a machine learning approach. The primary focus of the research includes an analysis of fuel efficiency measured through the heat rate parameter and an evaluation of critical parameters that potentially lead to long-term mechanical damage. The methods utilized in this study involve a combination of feature selection algorithms, namely Particle Swarm Optimization, Chi-Square, and Optimized Selection, with classification algorithms, including Decision Tree, Random Forest, Naive Bayes, and Deep Learning. The results reveal that the combination of Optimized Selection and Decision Tree algorithms achieves the best accuracy, reaching 93.89%. The key features identified as significant include Active Power, Turbocharger Bank A Speed, Exhaust Gas Temperature TC A Outlet, Exhaust Gas Temperature TC B Inlet, and HT Water Temperature, Jacket Outlet. Furthermore, the study indicates that the difference between 50 Hz and 60 Hz frequencies does not significantly contribute to generator efficiency. However, installations operating at 60 Hz frequencies show a higher potential risk for long-term mechanical damage to key generator components. As a mitigation measure, the study recommends optimizing combustion air pressure, air-fuel ratio, and cooling systems to enhance the reliability and operational efficiency of generators. These findings are expected to serve as a reference for the development of operational systems and maintenance strategies for generators applied to different frequencies, thereby supporting more reliable and efficient operational sustainability.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Efisiensi, Frekuensi, Genset, Machine Learning, Feature Selection, Klasifikasi, Efficiency, Frequency, Genset, Machine Learning, Feature Selection, Classification.
Subjects:	Q Science Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > Q Science (General) > Q337.3 Swarm intelligence Q Science > Q Science (General) > Q337.5 Pattern recognition systems
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Industrial Engineering > 26101-(S2) Master Thesis
Depositing User:	Bendrad Muhammad
Date Deposited:	31 Jan 2025 02:29
Last Modified:	31 Jan 2025 02:29
URI:	http://repository.its.ac.id/id/eprint/117233

Actions (login required)

View Item