Nuari, M. Farhan (2023) Model Bayesian Network Untuk Estimasi Causation Probability Tubrukan Kapal: Studi Kasus Analisis Frekuensi Tubrukan Kapal Selat Sunda Dan Selat Lombok. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 04211940000043-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
04211940000043-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (8MB) | Request a copy |
Abstract
Berdasarkan data dari Komite Nasional Keselamatan Transportasi (KNKT), tubrukan kapal merupakan tipe kecelakaan dengan persentasi terbesar diinvestigasi sehingga menjadikannya jenis kecelakaan dengan tingkat variasi penyebab yang tinggi. Selain itu tubrukan kapal juga menimbulkan ancaman yang serius karena terjadi pada dua kapal yang berbeda yang akan menimbulkan kerugian material hingga korban jiwa. Kondisi tersebut menjadikan tubrukan kapal sebagai masalah yang serius dan harus dilakukan upaya minimalisir terkait pencegahan dan penyesuaian kondisi yang ada. Penelitian ini akan melakukan pemodelan penyebab tubrukan kapal di Indonesia dengan menggunakan metode Bayesian Network hingga mendapatkan nilai peluang tubrukan kapal mengalami tubrukan atau selamat dari tubrukan (near miss) serta mendapatkan nilai Causation Probability. Metode Bayesian Network merupakan metode pemodelan peluang yang memperlihatkan secara jelas hubungan sebab-akibat antara variabel sehingga dapat memenuhi tujuan dari studi ini yaitu mengetahui faktor penyebab terjadinya tubrukan kapal, mengetahui nilai probabilitas kapal mengalami tubrukan, mengetahui nilai Causation Probability setiap tipe tubrukan kapal (Head-On, Overtaking, dan Crossing) di Perairan Indonesia, dan mengetahui faktor – faktor penyebab yang memiliki kontribusi terbesar pada probabilitas tubrukan kapal di Indonesia dengan analisis sensitivitas. Dalam meraih tujuan tersebut, juga dilakukan validasi model dengan menggunakan metode Hold-Out. Hal tersebut dilakukan untuk memastikan tingkat akurasi model dan nilai probabilitas yang didapatkan. Hasil yang didapatkan dari model dengan bobot 70% untuk training data yang dibangun adalah probabilitas kapal mengalami tubrukan kapal pada saat kondisi bahaya tubrukan adalah 63% dengan akurasi dan sensitivitas sebesar 93.75% dan 100%. nilai causation probability tubrukan kapal di Selat Sunda untuk Head-On Collision sebesar 2,74 x 10-4, Overtaking Collision 9,84 x 10-6, dan Crossing Collision 8,41 x 10-5, sedangkan untuk Selat Lombok adalah 2,79 x 10-4, 1,59 x 10-5, dan 6,24 x 10-5. Nilai causation probability tersebut diaplikasikan ke dalam perhitungan frekuensi tubrukan kapal di Selat Sunda menghasilkan frekuensi tubrukan head-on, overtaking, dan crossing di Selat Sunda berturut – turut sebesar 0.0216, 0,000264, 0,004, sedangkan di Selat Lombok senilai 0,000026, 0,0000048, 0,0000022. Berdasarkan model tersebut didapatkan faktor penyebab dengan pengaruh terbesar yaitu “Crew Competence”, “Decision Making”, dan “Ship Communication.
=================================================================================================================================
Based on data from the National Transportation Safety Committee (NTSC), ship collisions are the type of accident with the largest percentage being investigated, making it a type of accident with a high degree of variation in causes. In addition, ship collisions also pose a serious threat because they occur on two different ships which will cause material losses and even fatalities. These conditions make ship collisions a serious problem and efforts must be made to minimize related to prevention and adjustment of existing conditions. This study will model the causes of ship collisions in Indonesia using the Bayesian Network method to obtain a near miss and collision probability of a ship crash and obtain a Causation Probability value. The Bayesian Network method is a probability modeling method that clearly shows causal relationships between variables so that it can fulfill the objectives of this study, namely knowing the factors that cause ship collisions, knowing the probability values of ships experiencing collisions, knowing the Causation Probability values for each type of ship collision (Head-On, Overtaking, and Crossing) in Indonesian waters, and knowing the causal factors that have the largest contribution to the probability of ship collisions in Indonesia with sensitivity analysis. In achieving this goal, model validation was also carried out using the Hold-Out method. This is done to ensure the level of accuracy of the model and the probability values obtained. The results obtained from the model with a weight of 70% for the training data that was built is that the probability of a ship experiencing a ship collision during a collision hazard is 63% with an accuracy and sensitivity of 93.75% and 100%. the causation probability value of ship collisions in the Sunda Strait for Head-On Collision is 2,74 x 10-4, Overtaking Collision is 9,84 x 10-6, and Crossing Collision is 8,41 x 10-5, while for the Lombok Strait it is 2,79 x 10-4, 1,59 x 10-5, and 6,24 x 10-5. The causation probability value is applied to the calculation of the frequency of ship collisions in the Sunda Strait resulting in the frequency of head-on collisions, overtaking and crossing in the Sunda Strait respectively 0,0216, 0,000264, 0,004, while in the Lombok Strait is 0,000026, 0,0000048, 0,0000022. Based on this model, the causal factors with the greatest influence are "Crew Competence", "Decision Making", and "Ship Communication".
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | Kecelakaan Kapal, Tubrukan Kapal, Causation Probability, Bayesian Network, Pemodelan Kecelakaan Kapal, Selat Sunda, Selat Lombok, Ship Accidents, Ship Collisions, Causation Probability, Bayesian Network, Indonesian Waters, Ship Accident Modeling, Sunda Strait, Lombok Strait |
| Subjects: | V Naval Science > V Naval Science (General) |
| Divisions: | Faculty of Marine Technology (MARTECH) > Marine Engineering > 36202-(S1) Undergraduate Thesis |
| Depositing User: | M. Farhan Nuari |
| Date Deposited: | 04 Aug 2023 03:22 |
| Last Modified: | 05 Sep 2023 03:15 |
| URI: | http://repository.its.ac.id/id/eprint/101441 |
Actions (login required)
 |
View Item |