This study presents the Life Cycle Assessment of a High-Temperature Methanol Fuel Cell for a passenger vessel. The proposed propulsion system is based on onboard hydrogen production via methanol steam reforming and subsequent power generation through a high-temperature PEM fuel cell. A cradle-to-wake LCA analysis was carried out using openLCA, employing as a functional unit one day of vessel operation (8 h, 1920 kWh of mechanical work at the propeller shaft, supplied by 2000 kWh of fuel-cell electrical output). A 20-year service life, covering the 16 impact categories of ILCD 2011 Midpoint+, ten of which are discussed in detail (climate change, particulate matter, acidification, human toxicity non-cancer effects, freshwater ecotoxicity and mineral resource depletion among them), was considered. The system was evaluated against conventional diesel propulsion in terms of fuel consumption, operational cost and life-cycle impacts. The results indicate that the diesel-based case shows a GWP of 1458 kg CO2-eq per day. The fuel cell system is reported for all four methanol sources: bio-methanol gives −82% and wind-based e-methanol −86% against the diesel reference, while fossil methanol gives +36% and grid-based e-methanol +115%. A parameter sensitivity analysis and a Monte Carlo simulation (1000 paired iterations) show that the fuel cell system has lower impacts than diesel in eight of ten categories in all iterations, whereas mineral resource depletion is 26–63 times higher, a gap that end-of-life platinum recycling reduces to 5–14 times. This work supports methanol as a hydrogen carrier for decarbonising maritime transport where a direct electrification scenario is still not applicable.
Attention is a process that guides organisms to engage in goal-directed behavior, such as seeking food when hungry. However, attention is subject to distraction, which can impact the execution of these behaviors. The ocean is full of auditory stimuli that are potentially distracting. Bottlenose dolphins (Tursiops spp.), who rely on sound as their primary sensory and communicative modality, have been shown to be impacted by noise, especially anthropogenic noise. In our study, we assessed whether louder background noise, including both natural and anthropogenic noise, was associated with worse performance on two common cognitive tasks. Using previously recorded cognitive task data, we collected information on how long it took dolphins to make a choice, how many times they considered each alternative, whether they succeeded at the trial, and the root-mean-squared amplitude of the background noise across three frequency bands while they were making a choice. In both datasets, longer choice times and more oscillations between alternatives were associated with louder noise in the highest frequency band, which mostly consists of social sounds produced by nearby dolphins. Worse performance was also associated with this higher frequency band in one of the datasets. Results were mixed for low and middle frequency bands. Our results suggest that social sounds may be more distracting than vessel noise for dolphins in San Diego Bay, though these data are correlational and should be further investigated. Future experiments should manipulate the types and magnitudes of noise that are most distracting across additional types and difficulties of cognitive tasks. This would further shed light on both welfare considerations and comparative cognition.
Damage identification in offshore jacket structures requires state screening when damaged samples are scarce and local damage classification under measurement noise. We propose a framework that combines unsupervised screening with a lightweight bimodal classifier. An autoencoder learns undamaged response features from undamaged data alone and provides initial damage screening. The bimodal classifier combines acceleration time histories and mode-shape curvature difference (MSCD) images through bidirectional cross-modal attention and gated fusion. Beam simulations achieved 92.26 ± 1.41% accuracy at 15% Gaussian noise, exceeding direct concatenation by 3.63 percentage points. In scaled jacket tests, the autoencoder detected 98.58% of damaged samples. With 5% low-frequency colored perturbation, the bimodal classifier achieved 100% accuracy across seven structural states on a bootstrap-resampled test set. The classifier contains 3.44 million parameters and required approximately 12 ms per sample on the tested CPU.
Automatic recognition of maritime signal flags requires not only identifying individual flags but also reconstructing their ordered combinations and interpreting the corresponding maritime meanings. This study develops a sequence-aware framework that integrates visual perception, ordered sequence reconstruction, reliability validation, and International Code of Signals (ICS)-based semantic interpretation. A maritime signal flag dataset containing 7865 images from 40 ICS categories was collected using a single small model vessel in sunny weather. The visual perception module employs Flag-YOLO, an enhanced YOLO11-based detector integrating DualConv, GSConv, and Dynamic Head to improve the recognition of small, distant, blurred, and wind-deformed flags. Detected flags are subsequently arranged in top-to-bottom order, validated using confidence and spatial-consistency constraints, and interpreted using a 313-entry ICS subset restricted to single-group hoists of up to three flags without substitute pennants. Flag-YOLO achieves 93.1% mAP@0.5 and 61.9% mAP@0.5:0.95 with 2.1 M parameters. In a sequence-level evaluation of 1500 samples, the proposed approach achieves 92.9% accuracy, compared with 88.5% for YOLO11n, with the largest improvement of 7.2 percentage points observed for three-flag sequences. The results demonstrate the effectiveness of integrating visual perception, structured sequence recognition, and domain knowledge for automated maritime visual communication understanding, providing a potential perception component for intelligent maritime systems.
Discrete floe ice increases the propulsion resistance of small unmanned surface vehicles (USVs) and generates impulsive bow and side loads. For high-density polyethylene (HDPE) USVs, these loads govern propulsion demand and provide inputs for structural assessment. Existing studies address floe-ice resistance of conventional ships and HDPE material behaviour separately. Using the geometry of an HDPE USV, this study combined the Reynolds-averaged Navier–Stokes (RANS) equations, volume-of-fluid (VOF) method, and discrete element method (DEM). The model computed open-water resistance and longitudinal loads under varying speed, ice concentration, and ice thickness. The numerical method was assessed against KRISO Container Ship (KCS) experiments in open water and floe ice. Open-water resistance increased from 49.2 to 230.7 N between 2 and 4 kn. At a 60% ice concentration and 0.02 m ice thickness, the mean total load increased from 125.6 to 428.9 N over this range. Increasing speed raised both fluid resistance and contact force, whereas higher concentration and thickness primarily increased contact force; the peak load reached 929.2 N at 0.04 m thickness. These fixed-attitude predictions provide comparative load estimates for propulsion assessment and subsequent structural analysis of the HDPE USV.
