Date of Award
5-2025
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
Department
Ocean Engineering and Marine Sciences
First Advisor
Chungkuk Jin
Second Advisor
Seong Hyeon Hong
Third Advisor
J. Travis Hunsucker
Fourth Advisor
Stephen Wood
Abstract
This dissertation presents an integrated synthetic environment for simulating the motion and monitoring the integrity of a moored Floating Production Storage and Offloading (FPSO) vessel. The central hypothesis is that synthetic data-driven approaches can (1) accurately estimate spatio-temporal wave fields, (2) predict vessel motion, and (3) monitor the integrity of the mooring system—using stereo imagery of the wave field and structure motion, both of which can be acquired in real time. The first study develops a synthetic framework for estimating spatial wave fields using stereo images captured by dual cameras mounted on a moving FPSO. Dynamic sea states and vessel motion are simulated using Blender to generate realistic image sequences, which are processed through the RAFT-Stereo deep learning model to reconstruct wave elevations within a global reference frame. The second study builds upon this foundation by employing the estimated wave fields to predict the motion response of the FPSO using a Long Short-Term Memory (LSTM) encoder-decoder architecture. This approach enables short-horizon motion forecasting based on visual wave field data, supporting operational awareness and control. The third study introduces a two-step artificial neural network (ANN) model for mooring failure detection. Using synthetic platform motion data, the first ANN classifies whether a failure has occurred in any mooring group, while the second localizes the specific failed line. Feature selection and hyperparameter optimization further enhance classification accuracy.
Recommended Citation
Jebari, Omar, "FPSO Motion in Synthetic Environments: Stereo-Driven Deep Learning for Wavefield Reconstruction, Motion Analysis, and Mooring System Assessment" (2025). Theses and Dissertations. 1586.
https://repository.fit.edu/etd/1586