Reliable estimates of dolphin abundance are essential for conservation and impact assessment, yet manual analysis of aerial surveys is time-consuming and difficult to scale. This paper presents an end-to-end pipeline for automatic dolphin counting from unmanned aerial vehicle (UAV) video that combines modern object detection and multi-object tracking. We construct a large detection dataset of 64,705 images with 225,305 dolphin bounding boxes and a tracking dataset of 54,274 frames with 207,850 boxes and 603 unique tracks, derived from UAV line-transect surveys. Using these data, we train a YOLO11-based detector that achieves a precision of approximately 0.93 across a range of sea states. For tracking, we adopt BoT-SORT and tune its parameters with a genetic algorithm using a multi-metric objective, reducing ID fragmentation by about 29% relative to default settings. Recent YOLO-based cetacean detectors trained on UAV imagery of beluga whales report precision/recall around 0.92/0.92 for adults and 0.94/0.89 for calves, but rely on DeepSORT tracking whose MOTA remains below 0.5 and must be boosted to roughly 0.7 with post-hoc trajectory post-processing. In this context, our pipeline offers competitive detection performance, substantially larger and fully documented detection and tracking benchmarks, and GA-optimized tracking without manual post-processing. Applied to dolphin group counting, the full pipeline attains a mean absolute error of 1.24 on a held-out validation set, demonstrating that UAV-based automated counting can support robust, scalable monitoring of coastal dolphin populations.
Abstract This paper presents a realistic cost reduction method for a reference semi-submersible platform supporting a 15MW floating offshore wind turbine. A range of design load conditions that represent key European areas for floating wind development are applied to the system using a numerical model and a structural design optimization. Results aim to show the realistic material savings through structural design changes that maintain good stability in the offshore environment. A total saving of 16% was possible under the most severe load case and a 2% drop in typical 1GW farm LCoE, which shows potential for future savings especially under milder metocean conditions with lower extreme values.
Abstract Snap loads generate sharp discontinuities in the axial tension distribution along the mooring line, thereby posing a numerical challenge for conventional methods. This study presents the comparison between an in-house high-order accurate nodal Discontinuous Galerkin (DG) method with the industry-standard OrcaFlex lumped-mass (LM) model. The DG method employs a conservative shock-capturing formulation to resolve tension discontinuities and wave propagation – an approach with limited application in literature. Two cases are examined: an experimental slack-taut catenary chain test and a numerical benchmark test using the OC4-DeepCWind platform. For mild environmental loading, both methods show excellent agreement in fairlead tension (RMSE within 0.4%). Under slack conditions prone to snap loads, however, the DG method accurately captures tension-shock propagation and reflection, while the LM model exhibits significant numerical smearing, mispredicting the tension peak location observed experimentally. These results motivate further investigation into the conditions under which lumped-mass models remain adequate, and where higher-fidelity approaches become necessary for reliable mooring design.
Abstract Two RANS-based CFD approaches were assessed for predicting calm-water towing resistance of the VolturnUS 1:8 semi-submersible floating wind platform. Simulations were performed in FINE™/Marine (ISIS-CFD) using a rigid-lid (RL) single-phase formulation and a free-surface-resolving (FSR) two-phase VOF formulation with quasi-static heave and pitch and prescribed surge. Grid-refinement studies with six near-geometrically similar meshes were conducted with Spalart–Allmaras and SST k – ω models. Pressure resistance contributed 98–99% of the total, confirming bluff-body behaviour. Turbulence-model sensitivity was modest (≤2.5% for RL and ≤0.9% for FSR on the finest meshes). SA provided resistance close to SST while yielding lower discretisation uncertainty, whereas SST exhibited large residuals for the specific dissipation rate. Additional towing-speed cases were compared with published sea-trial data and a DNV-RP-C205 estimate: RL underpredicted resistance, FSR reduced the bias and remained slightly conservative, and DNV-RP-C205 overpredicted even without modelling wave-making. CFD and DNV-RP-C205 followed an approximately quadratic speed trend, whereas the sea-trial data were non-monotonic, suggesting unmodelled external disturbances. Overall, the results support the FSR approach as a practical basis for routine engineering towing-force assessment.
Abstract The integration of large-scale offshore wind generation introduces challenges related to electricity price volatility and economically driven curtailment. This study presents a techno-economic assessment of different energy storage systems (ESS) for a floating offshore wind farm (OWF) located in Portugal. Three storage scenarios are evaluated: a Hydrogen ESS (HESS), a Battery ESS (BESS) and a Hybrid ESS (HybESS). A dispatch model allocates wind power between grid injection, battery storage, and Hydrogen (H 2 ) production based on electricity market prices. The results show that HESS provides the most favorable economic performance, achieving positive Net Present Value (NPV) at H 2 selling prices ( p H2 ) above approximately 6.5 €/kg while significantly reducing curtailment. In contrast, BESS remains economically challenging under the considered market conditions and hybrid configurations only achieve limited improvements if enhanced control strategies are applied. These findings highlight the potential of H2-based storage for monetizing low-price electricity periods in large offshore wind projects operating under wholesale market exposure.
