
Artificial Intelligence (AI) has become an increasingly relevant technological component in the digital transformation of ports, supporting improvements in operational efficiency, safety and environmental performance. In recent years, several literature reviews have examined AI applications in maritime and port-related contexts; however, these studies typically focus on specific dimensions, such as safety and risk management, smart port development, or individual Machine Learning techniques. Consequently, a consolidated and application-oriented overview explicitly centered on port management and operations remains limited. This study addresses this gap by conducting a systematic literature review aimed at synthesizing current scientific evidence on the application of AI in port management and operational activities, while identifying emerging research gaps. The review follows the PRISMA protocol and analyses 30 peer-reviewed journal articles published between 2021 and early 2025. The results indicate that AI is predominantly applied in operational areas such as vessel arrival forecasting and scheduling, berth and quay crane allocation, container handling and predictive maintenance. Machine Learning and Deep Learning approaches are the most frequently adopted and exhibit higher levels of empirical validation. Reported outcomes include improvements in operational efficiency, enhanced decision-support accuracy and reductions in environmental impact. Despite these advances, the literature highlights persistent challenges related to data availability and quality, system interoperability, cybersecurity and the adaptation of AI solutions to heterogeneous port environments. The study concludes that, while AI adoption in ports is progressing beyond experimental stages, further empirical research and structured implementation strategies are required to support scalable and secure deployment across different port contexts.
In Indonesia, ocean currents represent a promising resource and one technology that can be used to harness them is the Savonius turbine. Although the vertical-axis Savonius turbine has been widely explored due to its ability to operate at low flow velocities, research on the cross-flow type remains limited. This study analyzed the influence of adding a deflector and placement variations on the performance of a cross-flow Savonius turbine. The turbine was tested with several deflector configurations, including different distances 0.22, 0.23, 0.24, and 0.33 m from the turbine, to evaluate their influence on performance. The deflector was designed to redirect flow to avoid the returning blade while strengthening the advancing blade, thereby enhancing energy capture. A numerical investigation was conducted using Computational Fluid Dynamics (CFD). The results of adding the deflector significantly improved the turbine performance in terms of its torque (Ct) and power coefficients (Cp) by 44%. The optimal configuration was obtained at a deflector distance of 0.23 m with a Tip Speed Ratio (TSR) of 0.739, yielding a coefficient of power of 0.289. This study indicates that proper deflector positioning can effectively improve the performance of a cross-flow type Savonius turbine.
This paper analyses the EU regulatory framework governing scheduled coastal maritime passenger services, focusing on the design and implementation of public service obligations and public service contracts in maritime cabotage. This paper adopts a normative and comparative legal approach by outlining the relevant EU legal context and examining Croatia, Greece and Italy – selected for their pronounced island dependency and distinct governance arrangements. The analysis compares competent authorities, public service instruments, tendering logic and transparency requirements, and key parameters relevant to service continuity and public compensation. The findings confirm that a common EU baseline permits differentiated national solutions: Croatia relies on a centralised model with a specialised public body for tendering and contract supervision, Greece applies a tender-based public service obligations system administered by the competent ministry, and Italy operates within a multi-level framework in which regional contracting practice is complemented by stronger regulatory oversight, including methodological standards relevant to compensation and reasonable profit. It is concluded that safe and legally regulated connectivity between islands and the mainland is a prerequisite for sustainable island development.
This research analyzes in detail the reception of NAVTEX signals in the northern Adriatic, revealing significant discrepancies between empirical observations and predictions of standard groundwave propagation models. The NAVTEX receiver at a fixed location on shore consistently failed to receive NAVTEX messages during daytime broadcasts from the Split and Piombino stations, despite being within their coverage area. At the same time, signals and messages from the nearest Mondolfo station were reliably received. Signal strength measurements confirmed that the signals from Split and Piombino stations were weak and unstable, while Mondolfo station provided a strong and stable signal. Comparison of these observations with the predictions of propagation models showed that the models predict signal strengths which should be sufficient for detection from all stations, including Split and Piombino. These findings indicate that current propagation models fail to accurately capture complex environmental factors or potential transmitter performance issues. It is concluded that these models cannot be reliably used to predict NAVTEX coverage in such complex maritime environments, which has critical implications for maritime safety due to the inability to receive Maritime Safety Information (MSI) via the NAVTEX system.
Jack-up Rigs (JuRs) are integral to offshore resource extraction, mainly due to their need for robust structural stability in challenging marine environments. This study introduces a Multi-Cascade Neural Network (MCNN) framework designed to evaluate the structural integrity of JuRs. We implement a Fast-Marching Algorithm (FMA) to process the vibration data obtained from a Multi-Sensor Network (MSN) strategically installed at critical locations on the rig. The FMA facilitates the construction of an overall state ridge from the vibration datasets, which, over time, informs the MCNNbased assessment model. During an 8-week evaluation period, the assessment model was rigorously tested, resulting in a structural state classification value, which effectively encapsulated the rig's overall condition. The application of our proposed methodology in various test scenarios demonstrated promising outcomes, validating the efficacy of our approach in structural assessment for JuRs.