Aerodynamic drag dominates the resistive forces in many sports at racing speeds, and small athlete posture changes can produce practically meaningful changes in performance. Standard approaches like field testing, wind-tunnel testing and computational fluid dynamics (CFD) can provide accurate results, but are time-consuming. Recent advances in scientific machine learning have enabled surrogate models that predict flow quantities at a fraction of the computational cost of a full CFD simulation. However, most applications currently focus on industrial geometries, such as cars or aircraft, where the surface is morphed to generate large datasets for training. It remains unclear how well these methods transfer to athlete geometries where variability is dominated by articulated pose changes rather than smooth shape morphing. In this work, a dataset for cyclist aerodynamics was generated by combining 12 scanned athlete geometries with 20 postures per athlete. CFD simulations were performed for these 240 geometries, which were then used to train a state-of-the-art surrogate model. The generalization of the model to an unseen geometry was investigated, along with the balance between number of positions and number of unique geometries in the dataset. The surrogate model predicts drag area with a mean absolute percentage error of < 3

This article addresses the challenges of assessing pedestrian-level wind conditions in urban environments using a deep learning approach. The influence of large buildings on urban wind patterns has significant implications for thermal comfort, pollutant transport, pedestrian safety, and energy usage. Traditional methods, such as wind tunnel testing, are time-consuming and costly, leading to a growing interest in computational methods like computational fluid dynamics (CFD) simulations. However, CFD still requires a significant time investment for such studies, limiting the available time for design modification prior to lockdown. This study proposes a deep learning surrogate model based on a MLP-mixer architecture to predict mean flow conditions for complex arrays of buildings. The model is trained on a diverse dataset of synthetic geometries and corresponding CFD simulations, demonstrating its effectiveness in capturing intricate wind dynamics. The article discusses the model architecture and data preparation and evaluates its performance qualitatively and quantitatively. Results show promising capabilities in replicating key wind features with a mean error of 0.3 m/s and rarely exceeding 0.75 m/s, making the proposed model a valuable tool for early-stage urban wind modelling.

Pedestrian-level wind environments are strongly influenced by urban morphology, with large and tall buildings playing a significant role. City authorities increasingly mandate assessments of pedestrian wind conditions before approving new construction. Computational fluid dynamics (CFD) models can provide a detailed understanding of the aerodynamic environment; however, in early design stages, urban morphologies are subject to change, requiring multiple simulations, adding substantial financial and time burdens to projects. To address this challenge, we develop a deep learning approach for the rapid inference of pedestrian-level wind conditions using a multi-layer perceptron (MLP)-mixer architecture. By embedding 3D structural details into the training data, our model can infer wind conditions around complex structures such as lift-up designs and skyways while maintaining inference times on the order of fractions of a second. This extends the capabilities of deep learning models that typically reduce the problem to a 2D image-to-image translation task, omitting crucial structural details. We conduct an extensive evaluation of our model and compare its performance to the widely adopted UNet architecture, demonstrating that the MLP-mixer outperforms UNet across all evaluation metrics. Notably, the MLP-Mixer achieves a mean squared error approximately 2.6 times lower, a peak signal-to-noise ratio 3.7 dB higher and the highest recorded structural similarity index of 0.991. These results indicate improved agreement with the reference CFD data. We anticipate that the MLP-mixer model will serve as a valuable tool in early-stage urban design workflows, enabling faster and more efficient wind assessments.

An athlete's posture has a significant impact on aerodynamic drag. Although aerodynamic drag in different sports has been studied extensively, most studies have analysed only a limited number of positions, and no generalized methods for optimization are available. In this work, we present a methodology to perform athlete posture optimization with respect to aerodynamic drag reduction. The method combines the virtual skeleton methodology to adjust the athlete's posture, CFD simulations to evaluate the drag for a given posture, and efficient global optimization to find the optimum position. We demonstrate the method by optimizing the time trial position for a cyclist. The cyclist position was parameterized with 6 design parameters, and the optimization required 41 CFD simulations to converge. The optimal posture yielded a reduction in drag of 17 % compared to the initial posture (disregarding bicycle drag). The method has potential to make posture optimization more accessible across a wide range of sports, and lead to insight into the aerodynamic influence of posture in general.

Wind turbine sizes have grown rapidly in recent years with machine ratings of 15-16 MW available from multiple manufacturers for offshore wind turbines. While the industry advances large-scale turbine designs, offshore developers have initially focused on siting in shallow waters on bottom-fiixed foundations; however, eyes are also on deep-water locations where floating systems are required. The present study presents the definition for an initial baseline rotor design at 25 MW scale, designed for a floating offshore system. The purpose of this paper is to document this 25 MW blade definition for use as a reference design for future floating wind technology development by industrial and academic researchers and developers. As this is an initial reference design, some opportunities (and plans) for further mass or cost reduction are also noted. In summary, the paper documents the initial baseline rotor design including aerodynamic and structural design of the rotor blades, along with key details about the control system and floating system designs for the floating 25 MW wind turbine system.

Traditional computational fluid dynamics (CFD) methods used for wind field prediction can be time‐consuming, limiting architectural creativity in the early‐stage design process. Deep learning models have the potential to significantly speed up wind field prediction. This work introduces a convolutional neural network (CNN) approach based on the U‐Net architecture, to rapidly predict wind in simplified urban environments, representative of early‐stage design. The process of generating a wind field prediction at pedestrian level is reformulated from a 3D CFD simulation into a 2D image‐to‐image translation task, using the projected building heights as input. Testing on standard consumer hardware shows that our model can efficiently predict wind velocities in urban settings in less than 1 ms. Further tests on different configurations of the model, combined with a Pareto front analysis, helped identify the trade‐off between accuracy and computational efficiency. The fastest configuration is close to seven times faster, while having a relative loss, which is 1.8 times higher than the most accurate configuration. This CNN‐based approach provides a fast and efficient method for pedestrian wind comfort (PWC) analysis, potentially aiding in more efficient urban design processes.

Urbanization has underscored the importance of understanding the pedestrian wind environment in urban and architectural design contexts. Pedestrian Wind Comfort (PWC) focuses on the effects of wind on the safety and comfort of pedestrians and cyclists, given the influence of urban structures on the local microclimate. Traditional Computational Fluid Dynamics (CFD) methods used for PWC analysis have limitations in computation, cost, and time. Deep-learning models have the potential to significantly speed up this process. The prevailing state-of-the-art methodologies largely rely on GAN-based models, such as pix2pix, which have exhibited training instability issues. In contrast, our work introduces a convolutional neural network (CNN) approach based on the U-Net architecture, offering a more stable and streamlined solution. The process of generating a wind flow prediction at pedestrian level is reformulated from a 3D CFD simulation into a 2D image-to-image translation task, using the projected building heights as input. Testing on standard consumer hardware shows that our model can efficiently predict wind velocities in urban settings in real time. Further tests on different configurations of the model, combined with a Pareto front analysis, helped identify the trade-off between accuracy and computational efficiency. This CNN-based approach provides a fast and efficient method for PWC analysis, potentially aiding in more efficient urban design processes.

This study focuses on the aerodynamic influence of athlete posture in sports aerodynamics. To analyze a specific posture, wind tunnel measurements and computer simulations are commonly employed. For computer simulations, the growing trend is to use 3D scanning to create accurate representations of an athlete’s geometry. However, this process becomes cumbersome and time-consuming when multiple positions need to be scanned. This work presents a methodology to use a virtual skeleton to perform modifications of an athlete’s posture. This is an efficient approach that can be applied directly to a scanned geometry model, and that allows easy modification and use in optimization procedures. The methodology is applied to two different cases; small adjustment of arm position for a time-trial cyclist, and large alteration of a standing alpine skier into a tucked position. Computational fluid dynamics simulations show that similar results are obtained for aerodynamic drag using the proposed methodology as with geometry models obtained from 3D scanning. Less than 1% difference in drag area was found for the cyclist, and less than 2% difference for the skier. These findings show the method’s potential for efficient use in sports aerodynamics studies.

Several key development areas have been identified as having high potential for reducing the levelized cost of energy of offshore wind. Two of the most anticipated developments are future generation large wind turbines and the use of floating foundations. There is thus a need for developing large floating substructures that are capable of hosting future generation wind turbines. This work presents the preliminary sizing of two semi-submersible platforms for supporting a 25 MW turbine through a design space search using a simplified parametric analysis. Compared to simple theoretical upscaling, the substructures resulting from the proposed simplified parametric analysis have significantly lower steel mass and stiffer tower.

Proportional-integral controllers are extensively applied to the pitch control of wind turbines. Despite its simplicity, this control strategy achieves good performance in onshore applications. However, the application of proportional-integral controllers to floating wind turbines faces some challenges, such as negative feedback due to the platform motion. In this sense, the present work proposes a parametric study to assess the influence of tuning parameters of the pitch controller on the performance of a 25 MW floating wind turbine. Effects of including floating feedback in the control strategy are also investigated. Finally, optimum parameters for a proportional-integral pitch controller are defined for the 25 MW wind turbine.

Understanding how postural changes in alpine skiing affect the overall aerodynamic drag is highly important for enhancing performance. Although the arm configuration of the athlete can have a significant impact on the overall drag force, this effect is currently less understood. The purpose of this investigation was to examine how the arms of an alpine skier influence the overall drag. Experiments were performed in a wind tunnel for a male and female athlete, and computational fluid dynamics simulations were performed on 3D scans of the athletes. The influence of the arm configurations in three different scenarios are considered; low-tucked, high-tucked, and flight postures. Consistent trends are found for both athletes and between the experiments and simulations. In general, the arms were found to be highly influential of the overall drag, and hence also performance in alpine skiing. For the low-tucked posture, the maximum variation in total drag area depending upon the angle of the underarms is 2.8%, with the lowest drag found with a medium angle of 20 $$^\circ $$ to 25 $$^\circ $$ . For the high-tuck posture, the difference in drag area between a closed and open posture was found to be 17% to 21%. The flight postures showed the highest influence of arm configurations, with a maximum difference in drag area of 64% between the considered postures. These results contribute to the understanding of aerodynamics in alpine skiing, and could be implemented directly in the training of athletes to improve their aerodynamic performance.

An approach to aerodynamically optimizing cycling posture and reducing drag in an Ironman (IM) event was elaborated. Therefore, four commonly used positions in cycling were investigated and simulated for a flow velocity of 10 m/s and yaw angles of 0–20° using OpenFoam-based Nabla Flow CFD simulation software software. A cyclist was scanned using an IPhone 12, and a special-purpose meshing software BLENDER was used. Significant differences were observed by changing and optimizing the cyclist’s posture. Aerodynamic drag coefficient (CdA) varies by more than a factor of 2, ranging from 0.214 to 0.450. Within a position, the CdA tends to increase slightly at yaw angles of 5–10° and decrease at higher yaw angles compared to a straight head wind, except for the time trial (TT) position. The results were applied to the IM Hawaii bike course (180 km), estimating a constant power output of 300 W. Including the wind distributions, two different bike split models for performance prediction were applied. Significant time saving of roughly 1 h was found. Finally, a machine learning approach to deduce 3D triangulation for specific body shapes from 2D pictures was tested.

The purpose of this investigation was to examine the aerodynamics of tucked positions in competitive alpine skiing. To further our understanding of how a skier?s position affects the air flow and the resulting aerodynamic drag, a combination of both experimental and simulation methods was used. This study focused in particular on the effect of skier torso and thigh angles relative to the air flow direction, as these two angles have been previously found to be important determinants of aerodynamic performance in tucked positions. Two top 30 world-ranked skiers were investigated in two different wind tunnels, and the results were compared with Computational Fluid Dynamics (CFD) simulations performed using a 3D scan of one of the athlete. To quantify the effect of torso and thigh angles on skier drag, changes in drag were measured relative to baseline positions. Skier drag area increased by approximately 0.8 and 1.2% per degree increase in torso and thigh angles relative to the baseline position, respectively. This trend was consistent between both of the experimental wind tunnel tests as well as the CFD simulations, indicating good agreement between methods. The CFD simulations further indicated that the air flow about the lower legs made the largest contribution to skier drag, accounting for as much as 40?50% of the total drag area in low tuck positions. Based on these findings, a low tuck position where the torso angle approaches 0? and the knees help to fill the gap behind the armpits will minimize skier aerodynamic drag. ? 2021 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Approximating wind flows using computational fluid dynamics (CFD) methods can be time-consuming. Creating a tool for interactively designing prototypes while observing the wind flow change requires simpler models to simulate faster. Instead of running numerical approximations resulting in detailed calculations, data-driven methods and deep learning might be able to give similar results in a fraction of the time. This work rephrases the problem from computing 3D flow fields using CFD to a 2D image-to-image translation-based problem on the building footprints to predict the flow field at pedestrian height level. We investigate the use of generative adversarial networks (GAN), such as Pix2Pix [1] and CycleGAN [2] representing state-of-the-art for image-to-image translation task in various domains as well as U-Net autoencoder [3]. The models can learn the underlying distribution of a dataset in a data-driven manner, which we argue can help the model learn the underlying Reynolds-averaged Navier-Stokes (RANS) equations from CFD. We experiment on novel simulated datasets on various three-dimensional bluff-shaped buildings with and without height information. Moreover, we present an extensive qualitative and quantitative evaluation of the generated images for a selection of models and compare their performance with the simulations delivered by CFD. We then show that adding positional data to the input can produce more accurate results by proposing a general framework for injecting such information on the different architectures. Furthermore, we show that the models performances improve by applying attention mechanisms and spectral normalization to facilitate stable training.

The impact of flow-normal ribs and small-scale surface roughness on the drag and vortex shedding of a circular cylinder was investigated. Three rib heights, four relative rib spacings and three different forms of micro-roughness were combined to produce 28 unique surface coatings for the cylinder. The drag was measured in a wind tunnel for Reynolds numbers in the range 20,000<Re<160,000, representing nearly a decade change centred about the drag crisis. The drag measurements were complemented by hot-wire measurements in the wake to investigate the vortex shedding frequency. The results show significant average drag reduction, up to 23%, for most of the ribbed geometries compared to a smooth cylinder for Re<160,000. Increasing the rib height was found to reduce the critical Reynolds number and increase the minimum drag coefficient. Varying the rib spacing resulted in an ‘‘optimal” spacing, approximately five times the rib height, that caused the lowest critical Reynolds number. Increasing the micro-roughness resulted in a reduction in the critical Reynolds number and an increase in the minimum drag coefficient.

Compared to other Olympic sports, little research exists on competitive shortboard surfing—especially research comparing field and numerical data. In this paper, GPS sensors with 9-axis motion sensors were used to collect data on nearly 2000 surfed waves. Data were collected from four surfers of differing skill levels, ranging from intermediate/advanced (Level 6) to top-ranked professional (Level 9). The results revealed a positive correlation between surfer skill level and roll/pitch/yaw rates during a cutback. Some surfers used two different fin types: a standard commercial fin (C), and a 3D-printed, humpback whale-inspired fin (RW). Statistically significant cutback performance improvements were seen when surfers used the RW fin. Because of the skill level differences suggested by the field data, dynamic computational fluid dynamics (CFD) analysis was performed to simulate cutback maneuvers at three different rotation rates (roll/pitch/yaw). Sustained resultant forces relative to the rider direction were lower for RW fins during the turn, suggesting a less-skilled surfer could generate faster and more powerful turns using RW fins. Field results also confirmed that a skill Level 8 surfer performed closer to skill Level 9 when using RW fins, but not control fins. Surfers experienced more stability using RW fins, and CFD results confirmed RW’s ability to dampen the effects of turbulent flow.

The surfing performance of two shortboard fin types with surface features were compared to a standard (control) fin with a smooth surface using dynamic computational fluid dynamics (CFD) simulations. The fins with surface features included designs with a partially grooved and serrated surface (CR), and humpback whale-inspired fins with tubercles and other features (RW). Surfboard roll, pitch and yaw during cutback maneuvers were simulated based on field data from surfers of intermediate, expert and professional (WCT) skill level surfing on ocean waves. Sustained resultant forces relative to the rider direction were significantly different between fin types, and lowest for RW at WCT-level rotations. CFD results also revealed RW’s ability to dampen effects of turbulent flow. RW fins were always the last to stall during a turn, and always exhibited the most gradual stall. CR fins had significantly lower pre-turn drag, and the highest mean resultant forces during the turn. Overall, CR fins appear best for forward acceleration and hold on the wave, while RW fins appear best for maneuverability and stability.

In professional cycling, even small adjustments in position could mean that valuable seconds are gained over the course of a time-trial race. This study investigates the influence of arm position on the aerodynamic drag of a cyclist. Based on a 3D scanned model of a professional cyclist, 64 alternate positions are generated. The parameters that are investigated are the distance between elbows, elbow extension, and distance between hands. Computational fluid dynamics (CFD) simulations of all positions are performed, and a regression model is built from the results. The results indicate that the optimal posture is achieved for a minimum in all investigated parameters, which means that the hands and elbows should be kept together with hands up towards the face. Furthermore, elbow extension seems to be the most crucial parameter, followed by the distance between elbows, and then by the distance between the hands. The presented methodology can be applied to study other parameters relevant to cycling aerodynamics or be applied to other sport activities as well.

Recently, manufactures who produce surfboards using computer aided design and robotic (CNC) shaping tools have gained a larger share of the surfing market, allowing board producers and shapers to produce lighter and more durable boards. The improvement in design and production process of surfboards has been directly linked with the adaption of technologies tested and validated in other industries. However, the surfing industry still lacks methodologies, standards and testing facilities to scientifically investigate and quantify the structural and hydrodynamic properties of surfboards. It is widely accepted that distributed rigidity, damping and stiffness in the three directions play a huge role in defining surfboard performance. These properties are rarely stated by producers and never measured. The present paper compares two boards with equal outline and geometric shape but built with two different technologies: one board built with a traditional high density polyurethane (PU) blank, polyester resin and a stringer to improve longitudinal stiffness and one board built with bio epoxy resin, expanded polystyrene (EPS) and two side inserts made of cork to improve longitudinal stiffness and damping. Different tests were carried out to compare the two boards and a testing methodology with the aim to evaluate board structural properties and identify key parameters that will influence the board performances. Accelerometers were used to characterize stiffness and damping while a custom-made flex machine was used to evaluate the torsional stiffness. Differences between the two boards are highlighted, leading to the fact that the two boards, even being identical in size, will behave differently in different wave conditions and for different surfers.