Surrogate-Assisted Model Untuk Prediksi Umur Kelelahan Pada Sambungan Tubular Multiplanar Jacket Platform Berbasis Mekanika Kepecahan

Hardian, Muhammad Akbar (2024) Surrogate-Assisted Model Untuk Prediksi Umur Kelelahan Pada Sambungan Tubular Multiplanar Jacket Platform Berbasis Mekanika Kepecahan. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6020231005-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (12MB) | Request a copy

Abstract

Berdasarkan informasi yang disampaikan dalam presentasi SKK Migas pada The 3rd Indo Decomm in Oil and Gas Conference, Indonesia telah mengoperasikan 613 unit anjungan lepas pantai terpancang sejak produksi komersial pertama di daerah lepas pantai. Dari total anjungan, sebanyak 54.65% telah berusia lebih dari 20 tahun, sementara 24.63% memiliki usia antara 16-20 tahun. Dengan perpanjangan waktu operasinya, integritas struktur platform tua tentu akan menurun. Salah satu akibatnya potensial terjadi kegagalan struktur karena beban siklis dengan adanya retak yang mayoritas bermula dari bagian las sambungan tubular pada strukturnya. Penelitian ini bertujuan untuk mengembangkan model baru dalam memprediksi umur lelah sambungan tubular multi-planar DKT berbasis Mekanika Kepecahan (Fracture Mechanics) dengan lebih efektif melalui surrogate model. Dengan menganalisis parameter geometri yang dapat mempengaruhi keretakan dan membangkitkan surrogate model yang lebih akurat untuk memprediksi umur kelelahan sambungan tubular multi-planar DKT kritis pada pembebanan aksial, IPB, OPB, dan gabungan. Analisis tersebut berhasil mendapatkan tegangan kritis yang terdapat pada brace 5. Umur lelah hingga kedalaman retak kritis juga didapatkan pada pembebanan aksial, IPB, OPB, dan gabungan untuk digunakan sebagai data training pada pembangkitan surrogate model dengan 2 variasi algoritma machine learning. Penelitian ini menyimpulkan bahwa algoritma radial basis function memberikan hasil yang lebih baik pada pembangkitan surrogate model daripada algoritma kriging. Hasil error yang diberikan pada algoritma radial basis function adalah 0.3%
====================================================================================================================================
Based on information presented in SKK Migas' presentation at The 3rd Indo Decomm in Oil and Gas Conference, Indonesia has operated 613 units of offshore platforms since the first commercial production in offshore areas. Of the total platforms, 54.65% are more than 20 years old, while 24.63% are between 16-20 years old. With the extension of its operation time, the structural integrity of the old platform will certainly decrease. One of the potential consequences is structural failure due to cyclical loading with cracks, the majority of which originate in the welds of tubular joints in the structure. This research aims to develop a new model to more effectively predict the fatigue life of DKT multi-planar tubular joints based on Fracture Mechanics through a surrogate model. By analysing geometry parameters that can affect cracking and generating a more accurate surrogate model to predict the fatigue life of critical DKT multi-planar tubular joints under axial, IPB, OPB and combined loading. The analysis successfully obtained the critical stress located at brace 5. The fatigue life to critical crack depth was also obtained under axial, IPB, OPB, and combined loading to be used as training data for surrogate model generation with 2 variations of machine learning algorithms. This study concluded that radial basis function algorithm gives better results in surrogate model generation than kriging algorithm. The error result given in the radial basis function algorithm is 0.3%.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Jacket Platform, Kelelahan, Mekanika Kepecahan, Surrogate Model, Sambungan Tubular, Jacket Platform, Fatigue, Fracture Mechanics, Surrogate Model, Tubular Joint
Subjects:	T Technology > T Technology (General) > T57.8 Nonlinear programming. Support vector machine. Wavelets. Hidden Markov models. T Technology > TC Hydraulic engineering. Ocean engineering > TC1665 Offshore structures--Materials. T Technology > TC Hydraulic engineering. Ocean engineering > TC1680 Offshore structures
Divisions:	Faculty of Marine Technology (MARTECH) > Ocean Engineering > 38101-(S2) Master Thesis
Depositing User:	Muhammad Akbar Hardian
Date Deposited:	11 Jul 2024 05:21
Last Modified:	11 Jul 2024 05:21
URI:	http://repository.its.ac.id/id/eprint/108251

Actions (login required)

View Item