149544 research outputs found
Sort by
Scaling mean velocity and Reynolds stress of a turbulent boundary layer submitted to an adverse pressure gradient
No full textDespite considerable progress in understanding zero pressure gradient boundary layers, turbulence in adverse pressure gradient (APG) boundary layers remains less well understood, particularly in high Reynolds number flows. Unfavorable pressure gradient regions are commonly encountered in industrial applications, but turbulence models often lack the physical basis necessary for reliable predictions in these flows. This study focuses on analyzing the effects of adverse pressure gradient on boundary layer scaling, essential for predicting flow characteristics and validating turbulence models. Building on recent advances in experimental methods and using large-scale particle image velocimetry (PIV), the research aims to provide an analysis of turbulent boundary layer flows in APG. Experiments have been carried out in a wind tunnel using inclined plates to induce pressure gradients at an angle of -8 deg, complementing an existing database obtained at -5 deg (see Cuvier et al., 2017) and offering new insights into flow behavior. An analysis of the literature has enabled the authors to compare various scaling approaches and to propose a scaling that is suitable for both mean velocity and Reynolds stress
Modelling cyclists‘ behaviour and interactions at urban intersections
No full textThe number of cyclists in Germany is increasing yearly, but infrastructure has failed to adapt to this growth. E-bikes are becoming more common, resulting in more speed differences on bicycle paths. In addition, manoeuvrable e-scooters and large cargo bikes also share the bicycle path with conventional bicycles and e-bikes. Bicycle paths are mostly narrow and it is difficult to manoeuvre without interaction with other cyclists. Little is known about the causes of bicycle-bicycle accidents. Due to the limited availability of accident data, real-life observational data can be used to gain a better understanding of bicycle-bicycle interactions.
This dissertation aims to provide a more comprehensive understanding of how cyclists interact with each other by presenting an overview of interaction patterns, descriptions of interactions, and initial modelling approaches. The data used for this thesis were collected over several weeks at different intersections in Braunschweig, Germany. At these intersections, the trajectories of road users were recorded by using camera systems. The thesis demonstrated that there are three primary categories of cycle interactions at intersections: overtaking, oncoming, and crossing. Surrogate Measures of Safety were utilised as criticality metrics to assess the degree of criticality of the interactions between cyclists in the scenarios. Route choice, speed during the interaction, lateral distances during the interaction and longitudinal distances before the interaction were analysed in order to subdivide the oncoming and crossing scenarios into further sub-scenarios. In addition, it was investigated how often cyclists adhere to the rules in the observation data (correct direction of travel, respecting right-of-way, bicycle path use) and how they themselves rated their cycling behaviour in a survey. The results of the analysis demonstrate that instances of deviant behaviour in particular can precipitate critical situations. The observed bicycle path is notably narrow, rendering it difficult for two cyclists to ride side-by-side. In the event that a cyclist travels in the incorrect direction on the bicycle path, there is an inherent risk of collision with another cyclist travelling in the wrong direction. Such an incident could result in injury. It is therefore of the utmost importance to educate children at an early age about cycling rules. Signs, pictograms or additional illuminations in critical areas of a bicycle path can alert cyclists that they are cycling in the wrong direction. In the crossing scenario, the installation of additional traffic lights could be considered to clarify the right-of-way. The study underlines the serious potential of conflicts of cyclists at intersections and the need for further research in this area. Parameter distributions can be used in the future to develop measures for mitigating cyclist conflicts or to simulate and plan infrastructure more effectively and safely
Round-robin study on the determination of weld geometry parameters-part B: analysis of welded specimen
No full textA Comprehensive Comparison of Federated Learning Frameworks
No full textFederated Learning is a Machine Learning paradigm that allows institutions to train models on distributed data while preserving data privacy. The selection of a suitable framework is a crucial step in designing a Federated Learning System, as the choice can influence the capabilities and limitations of the system. While previous research has mainly focused on comparing Federated Learning functionalities, other important criteria, such as usability, technical performance, and legal aspects have often been addressed only partially or overlooked entirely. To address this gap and support informed framework selection, we conducted a comprehensive comparison of Federated Learning frameworks. Frameworks were identified through a literature review and an analysis of GitHub repositories. The comparison considers five relevant comparison criteria: (1) Federated Learning functionalities, (2) user-friendliness, (3) technical aspects, (4) legal aspects and (5) performance evaluation
Advancement of SAR Imaging Techniques for the Observation of Terrestrial and Planetary Snow and Ice
No full textRadar remote sensing is an essential tool for observing Earth's cryosphere and has revolutionized our understanding of the state and dynamics of glaciers, ice sheets, and snow covers in the context of a changing climate. Beyond terrestrial snow and ice, radar imaging is a crucial technology for future exploration missions to the so-called icy moons of the giant planets, of which Saturn's moon Enceladus has recently been identified as the key target for investigating habitability on other worlds. Especially synthetic aperture radar (SAR) imaging has been extensively used for monitoring snow covers as well as the extent and dynamics of glaciers and ice sheets. Besides basic SAR imagery, SAR interferometry (InSAR) and tomography (TomoSAR) provide unparalleled measurement capabilities for the observation of Earth's cryosphere. Although modern radar remote sensing techniques like InSAR and TomoSAR are considered standard in Earth Observation (EO), they have not yet been adopted for the exploration of icy moons due to increased system complexity and strong orbit perturbations. However, these techniques have been recently identified as key developments for future exploration missions to Saturn's moon Enceladus. Upcoming Earth Observation (EO) SAR missions will acquire SAR, InSAR and TomoSAR data incorporating advanced capabilities, by: i) operating at lower frequencies (e.g., in the P- and L-band), resulting in considerable signal penetration into snow and ice covers, ii) providing very high spatial resolution, and/or iii) acquire as a satellite constellation to provide multi-aspect observations. The radar signal penetration capability at lower frequencies allows to image structures and processes within or underneath the snow and ice cover. Besides these opportunities, the penetration of the signals results in position ambiguities of imaged features, as well as biases and distortions in InSAR and TomoSAR products. An additional dimension of information in SAR observations of snow and ice that has not received much attention in the past is the propagation effect on the SAR signals when penetrating in the snow and ice volumes.
The aim of this thesis is to improve SAR, InSAR and TomoSAR imaging techniques for snow and ice observation in the frame of future EO and planetary missions by developing novel approaches for exploiting and compensating SAR signal propagation effects, as well as enabling InSAR and TomoSAR for the exploration of icy moons.
This thesis presents several novel concepts grouped into four research objectives. First, it describes the information content in single SAR images regarding snow and ice volume properties and introduces new single-image retrieval approaches that can be applied independently of polarimetric, interferometric, or tomographic information. These single-image approaches are highly relevant in scenarios where interferometric or tomographic information is unavailable (e.g., in planetary exploration missions), as well as for calibrating interferometric and tomographic products over ice sheets and glaciers.
The remaining research objectives focus on advancing SAR interferometric and tomographic techniques for snow and ice observation. The second objective assesses the feasibility and potential of using repeat-pass InSAR and TomoSAR for exploring icy moons, particularly within the context of an Enceladus mission scenario. Despite the strong orbit perturbations around Enceladus, highly stable repeat-pass orbits are designed to meet the stringent conditions for InSAR and TomoSAR. This assessment is adopted in a mission proposal currently being developed at the Jet Propulsion Laboratory (JPL), targeting repeat-pass InSAR observations of Enceladus for deformation and topography mapping.
The third objective addresses the significance of commonly ignored propagation effects in elevation measurements of ice sheets and glaciers using InSAR. These propagation effects result in considerable geolocation errors of meters to tens of meters beyond the well-known penetration bias. Several adapted processing approaches are developed to accommodate the propagation effects in terms of range and phase offsets, representing an important step toward a robust penetration bias calibration in InSAR elevation products.
The final objective tackles the limitations of current differential InSAR (D-InSAR) techniques for retrieving snow parameters. A novel explanation of temporal decorrelation over snow-covered areas is provided, linking snow density changes to the decorrelation of SAR signals caused by changes in the wavenumber within the snow volume. Additionally, methods to mitigate the 2-pi phase ambiguity of the interferometric measurement are developed by exploiting multiple D-InSAR acquisitions with different squint angles, which can also serve as a direct measurement of snow density. The upcoming Harmony mission by the European Space Agency (ESA) is a suitable candidate to implement these developed concepts due to its large squint diversity among the satellite constellation.
This thesis demonstrates the significant potential of synergistically developing terrestrial and planetary radar remote sensing. It advances the state-of-the-art of SAR imaging techniques for observing glaciers, ice sheets, and snow covers, as well as for exploring icy moons
An Architecture: Solving The Lock Scheduling Problem At Artificial Waterways With Reinforcement Learning
No full textThis study proposes an architecture to address a critical challenge in inland waterway transport systems: the Lock Scheduling Problem (LSP). Locks and ship lifts are essential components of waterborne infrastructure, yet they frequently act as bottlenecks in the inland waterway network. These bottlenecks can lead to traffic congestion, increased waiting times, and economic losses for industries relying on waterborne transport. In practice, the First-Come-First-Served (FCFS) principle is often used for scheduling vessels. However, this approach often yields suboptimal and inefficient lock utilization. Alternative methods based on mathematical optimization, such as Mixed-Integer-Programming (MIP), offer improved scheduling capabilities but suffer from long computation times and limited adaptability to dynamic and uncertain operational conditions.
To overcome these limitations, this study introduces an architecture that applies Reinforcement Learning (RL) to the LSP
Integrating Earth observation into public services: Advanced machine learning techniques for flood-induced damage mapping
No full textRECONNAISSANCE OF IN-SITU RESOURCES FOR FUTURE CREWED MISSIONS TO MARS
No full textMars bears abundant raw materialsthat are of potential value for future human activities. Avariety of essential elements are readily accessible at theplanet's surface. However, the identification and thenature of these natural resource deposits, theirconcentration, and the practicality of extraction andprocessing remain open points that require furtherinvestigation. The inherent geological diversity of oredeposits can greatly influence the viability ofexploration sites and the engineering architecture. Thus,a thorough understanding of these characteristics is notonly essential for the selection of potential mining sites,but also for the choice and optimization of specifictechnical designs. This underlines the importance ofdedicated resource exploration missions. Additionally,the processes of refinement and restructuring must beexamined to establish reliable fabrication systems, as InSitu Resource Utilization (ISRU) has the potential tosignificantly reduce logistical reliance from Earth