Hull fouling assessment through data-driven modeling
Loading...
URL
Journal Title
Journal ISSN
Volume Title
School of Electrical Engineering |
Master's thesis
Unless otherwise stated, all rights belong to the author. You may download, display and print this publication for Your own personal use. Commercial use is prohibited.
Authors
Date
Department
Major/Subject
Mcode
Degree programme
Language
en
Pages
130
Series
Abstract
The marine industry faces increasing pressure to reduce greenhouse gas emissions, with hull fouling—a buildup of marine organisms that increases resistance and fuel consumption—remaining a major challenge. This thesis develops a machine learning framework to monitor fouling-induced performance degradation using only operational ship data. Two approaches were investigated: a supervised residual-based method trained on clean-condition baselines, and an unsupervised autoencoder-based method with clustering. The supervised models, using convex hull and quantile filtering to define clean data, reliably detected fouling progression and showed residual drops after hull cleaning, enabling condition-based maintenance and even early detection weeks in advance. In contrast, the unsupervised models revealed operational patterns but required auxiliary signals to capture fouling trends. Overall, the framework demonstrates a practical, interpretable, and sensor-based solution that supports more efficient maintenance scheduling and contributes to sustainable maritime operations.Description
Supervisor
Manner, JukkaThesis advisor
Fodor, ViktoriaBourgerie, Rémi