Session: 15-05-01 Mooring, Riser and Pipelines

Submission Number: 180802

A Data-Driven Framework With Bayesian Optimisation for Mooring Fatigue Prediction of Hydroelastic VLFS Using Neural Network Approach 

This study aims to develop a data-driven framework for predicting mooring line fatigue by incorporating the hydroelastic behaviour of a Very Large Floating Structures (VLFS). The framework integrates hydroelastic model with a Bayesian-optimised Artificial Neural Network (ANN). Since the hydroelastic response of a VLFS significantly influences mooring fatigue performance, it is essential to develop reliable fatigue prediction methods. To address the high computational cost of hydroelastic fatigue analysis, a neural network surrogate model is trained to reproduce numerical responses efficiently. The Discrete-Module-Beam (DMB) method, based on multibody hydrodynamics with Euler-Bernoulli beam assumption, directly solves the coupled time-domain hydroelastic–mooring simulations, and fatigue damage is evaluated using rainflow counting method and Palmgren-Miner’s rule. Simulation data covering various environmental parameters are generated and divided into training, validation, and testing data sets. Prediction accuracy is evaluated by calculating the Mean Absolute Percentage Error (MAPE) and the coefficient of determination (R²). To further validate the performance of the trained ANN model, mooring fatigue damage predictions are carried out for a set of new environmental cases. The results indicate that the ANN model reproduces numerical fatigue responses with high accuracy while substantially reducing computation time, providing an efficient tool for evaluating mooring performance under hydroelastic effects.

Presenting Author: Priscilla Yola Aulia Loe Seoul National University

Presenting Author Biography: Priscilla Yola Aulia graduated cum laude with a Bachelor of Science in Ocean Engineering from the Bandung Institute of Technology, Indonesia. She is currently pursuing her Master’s degree in the Department of Naval Architecture and Ocean Engineering at Seoul National University, South Korea, and is expected to earn her M.S. in 2026. Her research focuses on multi-body floating structural safety assessment using the Discrete-Module-Beam (DMB) method with potential theory and high-order beam formulations in the time domain. Since joining Seoul National University, she has actively participated in several national conferences, including the Korean Society of Ocean Engineers (KSOE) and the Society of Naval Architects of Korea (SNAK). She is also the first author of a paper published in Ocean Engineering, titled “Recent Advances in Discrete-Module-Beam-Based Hydroelasticity Method as an Efficient Tool Approach for Continuous Very Large Floating Structures.”

Authors:

Priscilla Yola Aulia Loe Seoul National University
Chungkuk Jin Florida Institute of Technology
Farid Putra Bakti Bandung Institute of Technology
Moo Hyun Kim Texas A&M University
Malakonda Reddy Lekkala University of Massachusetts Amherst
Do Kyun Kim Seoul National University