Mengoptimalkan Efisiensi Performa dengan Menggunakan Artificial Neural Network dan Sensor Berdasarkan Data Operasional Full-Scale Studi Kasus Crew Boat Tipe Sea Area 2

Riyadi, Soegeng (2025) Mengoptimalkan Efisiensi Performa dengan Menggunakan Artificial Neural Network dan Sensor Berdasarkan Data Operasional Full-Scale Studi Kasus Crew Boat Tipe Sea Area 2. Doctoral thesis, Institut Teknologi Sepuluh Nopember.

Text 04111960010008-Dissertation.pdf Restricted to Repository staff only until 1 April 2027. Download (19MB) | Request a copy

Abstract

Saat ini, dunia pelayaran melalui IMO (International Maritime Organization) memiliki target secara bertahap mengurangi emisi karbon. Upaya ini dilakukan dengan mengoptimalkan penggunaan bahan bakar selama operasional. Dalam beberapa tahun terakhir, penerapan Machine Learning telah banyak diaplikasikan untuk mendukung efisiensi operasional kapal. Namun, di sisi galangan kapal, prosedur standard menggunakan Towing Tank Test dan simulasi CFD (Computational Fluid Dynamics) masih digunakan. Penentuan tenaga mesin yang dibutuhkan dalam tahap desain juga masih bergantung pada penerapan Sea Margin (SM) yang berkisar antara 20% hingga 30%. Penggunaan sensor pada kapal telah memberikan kontribusi signifikan terhadap peningkatan efisiensi energi, dengan menyediakan data operasional secara real-time. Data tersebut, ditambah dengan data lingkungan dari BMKG (Badan Meteorologi, Klimatologi, dan Geofisika), merepresentasikan kondisi operasional dan lingkungan secara aktual. Dengan memanfaatkan Machine Learning, prediksi total tenaga mesin (total power engine) untuk Crew boat dapat diperoleh berdasarkan data operasional yang dihasilkan oleh sensor. Metode Machine Learning yang digunakan meliputi Artificial Neural Network (ANN), Decision Tree (DT), Random Forest (RF), dan K-Nearest Neighbor (KNN). Hasil deviasi relatif Root Mean Square Error (RMSE) yang diperoleh masing-masing adalah 7,60%, 6,13%, 4,82%, dan 5,30%. Berdasarkan hasil tersebut, evaluasi desain SM menunjukkan bahwa tahap displacement sesuai dengan nilai standard sebesar 20%, tahap pre-planing mencapai 10%, dan fase planing mencapai 5%. Dengan demikian, penerapan Machine Learning berbasis data operasional mampu memberikan prediksi yang lebih akurat, mendukung optimasi energi, serta mendukung upaya pengurangan emisi karbon secara global. ===================================================================================================================================
Currently in the shipping industry, through the IMO (International Maritime Organization), has set a target to gradually reduce carbon dioxide emissions. This effort is being carried out by optimizing fuel usage during ship operations. In recent years, the application of Machine Learning has been widely implemented to support operational efficiency. However, on the shipyard side, standard procedures such as Towing Tank Testing and CFD (Computational Fluid Dynamics) simulations are still used. The determination of engine power required during the design phase also still relies on the application of a Sea Margin (SM), which ranges from 20% to 30%. The use of sensors on ships has contributed significantly to increasing energy efficiency by providing real-time operational data. This data, combined with environmental data such as those from BMKG (Indonesian Meteorological, Climatological, and Geophysical Agency), represents actual operational and environmental conditions. By utilizing Machine Learning, predictions of total engine power for Crew Boats can be obtained based on operational data generated by the sensors. The Machine Learning methods used include Artificial Neural Network (ANN), Decision Tree (DT), Random Forest (RF), and K-Nearest Neighbor (K-NN). The relative Root Mean Square Error (RMSE) deviations obtained are 7.60%, 6.13%, 4.82%, and 5.30%. Based on these results, the evaluation of Sea Margin (SM) design shows that the displacement phase aligns with the standard value of 20%, the pre-planing phase achieves 15%, and the planing phase reaches 5%. Thus, the application of Machine Learning based on operational data can provide more accurate predictions, support energy optimization, and contribute to global efforts to reduce carbon emissions.

Item Type:	Thesis (Doctoral)
Uncontrolled Keywords:	Emisi karbon, Efisiensi energi, Crew boat, Data operasional, Machine Learning, Sea Margin. Carbon emissions, Energy efficiency, Crew boat, Operational data, Machine Learning, Sea Margin
Subjects:	V Naval Science > VM Naval architecture. Shipbuilding. Marine engineering > VM156 Naval architecture
Divisions:	Faculty of Marine Technology (MARTECH) > Ocean Engineering > 38001-(S3) PhD Thesis
Depositing User:	Soegeng Riyadi
Date Deposited:	07 Feb 2025 02:29
Last Modified:	07 Feb 2025 02:29
URI:	http://repository.its.ac.id/id/eprint/118520

Actions (login required)

View Item