Thesis

Data-driven prognostics and health management for maritime systems employing trustworthy digital twins

Downloadable Content

Download PDF
Creator
Rights statement
Awarding institution
  • University of Strathclyde
Date of award
  • 2025
Thesis identifier
  • T17212
Person Identifier (Local)
  • 202170327
Qualification Level
Qualification Name
Department, School or Faculty
Abstract
  • Maritime Autonomous Surface Ships (MASS) are expected to revolutionise the maritime and shipping industries benefiting the operations sustainability, whilst enhancing safety and resilience. However, the unmanned operations of MASS present immense challenges pertaining to emergency responses, including shipboard corrective maintenance. Prognostics and Health Management (PHM) systems and their applications to ship machinery have received increasing attention as means of addressing these challenges. PHM systems require data-driven models to estimate machinery health status, the development of which, however, is impeded by the scarcity of data representing anomaly conditions and stochastic degradation of machinery components. This study aims at advancing PHM in maritime systems by developing digital tools, namely physics-based digital twin and data-driven PHM model, as well as frameworks to evaluate and manage their trustworthiness. The first digital tool is the digital twin that integrates thermodynamic models, component degradation models, and sensor models, while ensuring its trustworthiness by using a framework that uses the steps of validation, verification, and robustness. The second digital tool is the data-driven models for fault diagnosis and health prognosis, including Health Indicator (HI) construction and forecast sub-tasks. The data-driven methods for each sub-model are systematically selected and integrated to develop a comprehensive PHM model, which subsequently supports maintenance decision making by estimating the Remaining Useful Life (RUL) of the components. The trustworthiness of data-driven models is managed throughout the engine lifetime considering accuracy and robustness metrics. A marine four-stroke engine is used as a reference system, assuming that its cylinder valves degrade according to different stochastic degradation patterns. The novelty in this research stems from the integration of methods to develop digital tools and frameworks that ensure their trustworthiness while addressing real-world challenges in the maritime domain.
Advisor / supervisor
  • Theotokatos, Gerasimos
Resource Type
Note
  • This thesis was previously held under moratorium from 5th February 2025 until 5th February 2026.
DOI
Date Created
  • 2024
Funder

Relations

Items

Thumbnail Title Date Uploaded Visibility Actions
file details PDF of thesis T17212 2025-02-05 Public Download