Model Probabilistik Berbasis Bayesian Network Untuk Pengelolaan Breakdown Cooling System Kapal Tanker

Maula, Ardan Afwal (2026) Model Probabilistik Berbasis Bayesian Network Untuk Pengelolaan Breakdown Cooling System Kapal Tanker. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6019241033-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (2MB) | Request a copy

Abstract

Keandalan main engine merupakan salah satu faktor utama dalam menjamin kelancaran operasional kapal tanker karena berpengaruh terhadap ketepatan waktu distribusi muatan, keselamatan pelayaran, dan kepatuhan terhadap aspek lingkungan. Pengelolaan perawatan main engine umumnya mengacu pada Planned Maintenance System (PMS) serta didukung pemantauan kondisi mesin melalui Engine Control Room (ECR). Namun, pendekatan tersebut masih memiliki keterbatasan dalam mengantisipasi risiko breakdown karena belum mampu memodelkan hubungan sebab-akibat antar komponen, memperbarui tingkat risiko berdasarkan informasi terbaru, maupun mengakomodasi ketidakpastian akibat perubahan kondisi operasional kapal. Oleh karena itu, penelitian ini bertujuan mengembangkan model probabilistik berbasis Bayesian Network (BN) sebagai alat bantu untuk memprediksi dan mengelola risiko breakdown main engine pada kapal tanker. Penelitian diawali dengan identifikasi faktor risiko dan hubungan kausal berdasarkan data PMS, data historis kegagalan, data ECR, serta expert judgment. Selanjutnya disusun struktur Bayesian Network, ditentukan prior probability menggunakan data historis dan penilaian pakar, kemudian dibangun Conditional Probability Table (CPT) untuk merepresentasikan hubungan probabilistik antarvariabel. Model yang dihasilkan digunakan untuk melakukan inferensi probabilistik sehingga tingkat risiko dapat diperbarui secara dinamis berdasarkan evidence terbaru. Hasil penelitian diharapkan menghasilkan model prediksi risiko breakdown main engine yang mampu menggambarkan pengaruh faktor perawatan, kondisi operasi, performa mesin, dan interaksi antar subsistem terhadap probabilitas kegagalan. Model BN yang dikembangkan diharapkan dapat melengkapi implementasi PMS dan monitoring ECR sebagai decision support tool bagi manajemen darat dalam menentukan prioritas pemeliharaan, mengoptimalkan strategi perawatan berbasis risiko, meminimalkan kejadian breakdown, serta mendukung pengambilan keputusan yang lebih objektif dan berbasis data.
====================================================================================================================================
The reliability of the main engine is one of the most critical factors in ensuring the smooth operation of tanker vessels, as it directly affects cargo delivery schedules, navigational safety, and compliance with environmental regulations. Main engine maintenance is generally managed through a Planned Maintenance System (PMS) and supported by engine condition monitoring via the Engine Control Room (ECR). However, these approaches still have limitations in anticipating breakdown risks because they are unable to adequately model causal relationships among components, update risk levels based on newly available information, or accommodate uncertainties arising from changes in vessel operating conditions. Therefore, this study aims to develop a probabilistic model based on a Bayesian Network (BN) to predict and manage the risk of main engine breakdowns on tanker vessels. The research begins with the identification of risk factors and causal relationships using PMS records, historical failure data, ECR data, and expert judgment. Subsequently, the Bayesian Network structure is constructed, prior probabilities are determined based on historical data and expert assessments, and Conditional Probability Tables (CPTs) are developed to represent the probabilistic relationships among variables. The resulting model is then employed to perform probabilistic inference, enabling dynamic updates of risk levels as new evidence becomes available. The expected outcome of this study is a predictive model capable of representing the influence of maintenance practices, operating conditions, engine performance, and interactions among main engine subsystems on the probability of failure. The proposed Bayesian Network model is expected to complement the implementation of the Planned Maintenance System and ECR monitoring by serving as a decision support tool for shore-based management in prioritizing maintenance activities, optimizing risk-based maintenance strategies, minimizing breakdown occurrences, and supporting more objective, data-driven decision-making.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Bayesian Network, Planned Maintenance System, Engine Control System, Risiko Breakdown, Kapal Tanker, Bayesian Network; Planned Maintenance System (PMS); Engine Control System (ECS); Main Engine Breakdown Risk; Tanker Ships.
Subjects:	V Naval Science > VM Naval architecture. Shipbuilding. Marine engineering > VM731 Marine Engines
Divisions:	Faculty of Marine Technology (MARTECH) > Marine Engineering > 36101-(S2) Master Theses
Depositing User:	Ardan Afwal Maula
Date Deposited:	05 Aug 2026 01:36
Last Modified:	05 Aug 2026 01:36
URI:	http://repository.its.ac.id/id/eprint/143541

Actions (login required)

View Item