Harokopio University
AI & ΙοΤ-enabled exergy optimization for carbon-neutral ship operation
Abstract
dc:descriptionIn this industrial PhD thesis a comprehensive holistic framework is introduced for the maritime sector that aims to enable stakeholders, shipowners and external vendors attached to the waterborne sector, to actively embrace the full spectrum of IoT (Internet Of Things), and AI (Artificial Intelligence) advancements and technologies to support smart green shipping for both existing ships as well as new buildings. This big data tool continuously harvesting data in real time related to the vessel’s operational state is realized through a set of state of the art streaming frameworks for distributed processing and edge computing methodologies supporting Continuous Integration/Deployment (CI/CD) of core services and Decision Support Systems (DSS) between the Headquarters (HQ) and the vessel. The realization and embodiment of a service like that, is of high importance for ship operations and maintenance procedures, as the maritime sector has witnessed an exponential growth in data availability over the past few years. Amongst other technological advancements related to ship operational optimization, this work presents the workflow and the main components of a multi-purpose, adaptive pipeline for data collection-"curation" and model deployment. Utilizing structured methodologies (Knowledge Graphs) to support semantic annotation between the acquired operational features, we aim to largely simplify and standardize procedures related to pattern recognition and reasoning in order to provide to the shipowners accurate insights regarding the operational state of the vessel. This generic-physics informed, cross vessel library (Knowledge Hub), consisting of data sources, relevant use case and appropriate simulation model, conceptualizes an approach inextricably linked with the emerging concept of the so called Digital Twin (DT) in the shipping industry. The aforementioned envisaged framework is supported by a variety of simulation models regarding emission control, safety and efficiency. These models correspond mainly to Analytical approaches utilizing Splines and standard marine engineering theory and knowledge, data driven models for Fuel Oil Consumption (FOC) approximation, as well as Routing Optimization techniques based on shortest path principles and dynamic programming to achieve efficiency in ship operations with tangible benefits in emission reduction, environmental compliance, and crew safety on board.
Degree
thesis:*- Grantor dc:publisher
- Harokopio University
- Year dc:date
- 2023
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Kaklis, Dimitrios
- Κακλής, Δημήτριος
Subjects
dc:subject × 21- Τεχνητή νοημοσύνη
- Ομαδοποίηση δεδομένων
- Εκμάθηση συνόλου εκτιμητών
- Τριγωνοποίηση επιφάνειας
- Τετραγωνικοί εκτιμητές
- Πολυπαραγοντική βελτιστοποίηση
- Εξόρυξη τροχιάς
- Deep learning
- Clustering
- Ensemble learning
- Quadratic estimators
- Multi-constraint optimization
- Trajectory mining
- Pattern recognition
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Φυσικές Επιστήμες
- Πληροφοριακά συστήματα
- Computer and Information Sciences
- Natural Sciences
- Artificial Intelligence
- Information Systems
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/54753