Politecnio die Bari - Catalogo di prodotti della Ricerca
Not a member yet
36616 research outputs found
Sort by
Analysis of Model Parameters and Experimental Setup for Accurate Characterization of Ultrasound Phased Arrays Using a Low-Cost Sensor
No full textBackground: Ultrasound Phased Arrays (UPAs) are an emerging technology for contactless midair haptics and acoustic levitation, offering many advantages compared to other contactless methods. Accurate characterization of the airborne ultrasound field produced by UPAs is crucial for evaluating their performance. Objective: This paper introduces a novel UPA design and investigates the feasibility of using a low-cost commercial piezoelectric transducer, i.e., the Murata MA40S4R, for its characterization, in contrast to very expensive microphones commonly employed in literature for testing. Methods: The characterization is conducted both on the Murata MA40S4R transducer and on the proposed UPA, analyzing the experimental setup, the fitting of model parameters, the effect of wave reflection, duty-cycle dependency, and different directivity patterns on system performance. The architecture of the proposed UPA, the measurement chain, and the measurement protocol are described to ensure an accurate characterization procedure, and the influence of model parameters and experimental setup on the accuracy of measured data is investigated. Results: The performed tests show that experimental measurements closely match predicted data after optimizing the model parameters and setup, achieving a root-mean-square relative pressure error of 4.1% in the focal region. Conclusion: This research highlights the possibility of obtaining accurate ultrasound field measurements with a very low-cost sensor, paving the way for accessible UPAs characterization and development
Hydrogeological modelling of a coastal karst aquifer using an integrated SWAT-MODFLOW approach
No full textThe complexity of modelling in karst environments necessitates substantial adjustments to existing hydrogeological models, with particular emphasis on accurately representing surface and deep processes.
This study proposes an advanced methodology for modelling regional coastal karst aquifers using an integrated SWAT-MODFLOW approach. The focus is on the regional coastal karst aquifer of Salento (Italy), which is characterised by significant heterogeneity, anisotropy and data scarcity, such as limited discharge measurements and water levels over time.
The integrated SWAT - MODFLOW approach allows an accurate description of both surface and subsurface hydrological processes specific to karst environments and demonstrates the adaptability of the models to karst-specific features such as sinkholes, dolines and fault permeability. The study successfully addresses the challenges posed by the distinctive characteristics of karst systems through the integration of SWAT-MODFLOW. Additionally, incorporating of satellite data enhances the precision and dependability of the model by augmenting the traditional datasets.
The entire simulation period, which included both the calibration and validation phases, extended from 2008 to 2018. The calibration phase occurred between 2008 and 2011, followed by the validation phase between 2015 and 2018. The temporal choices were exclusively based on the availability of meteorological and hydrogeological data. During calibration, satellite data, previous study results, and groundwater level measurements were used to optimize the SWAT and MODFLOW models. Validation subsequently confirmed model accuracy by comparing simulated groundwater levels with observed data, demonstrating a satisfactory root mean square error (RMSE) of 0.22 m. Modelling results indicate that evapotranspiration is the predominant hydrological process, and excessive withdrawals could lead to a water deficit. Simulated piezometric maps provide crucial information on recharge areas and hydraulic compartments delineated by faults. The study not only advances the understanding of the hydrogeology of the specific case study but also provides a valuable reference for future modelling of karst aquifers. Additionally, it highlights the crucial need for ongoing enhancement in the management and monitoring of coastal karst aquifers
In-Process Detection of Defects on Parts Produced by Laser Metal Deposition Using Off-Axis Optical Monitoring
No full textLaser Metal Deposition (LMD) is emerging among metal Additive Manufacturing technologies due to its wide range of applications. This technique represents an evolution of laser cladding, currently used for fabricating and repairing complex metal components, promoting manufacturing sustainability. One of the main drawbacks hindering the widespread use of these technology is the complexity of implementing monitoring equipment on industrial LMD systems with limited modification setups. Therefore, it is essential to develop appropriate off-axis systems that allow effective monitoring of the deposition process. The present work proposes a prototype off-axis monitoring system consisting of a pair of specially set cameras capable of analyzing the evolution of the melt pool and discerning fundamental information on geometry, size and brightness intensity. By correlating this information with the process outcome, it could be possible to forecast the most frequent defects related to the deposition process. Experimental tests have been carried out, in which powder flow and laser alterations were specifically induced. The prototype system enabled the characterization of each type of process variation and the determination of specific indicators, serving as the basis for achieving a zero-waste sustainable manufacturing process
LPG in Fluoride Optical Fiber via Micro-Tapering Technique: Mid-IR Mode Coupling
No full textFor the first time, micro-tapering of fluoride glass optical fibers is exploited as a feasible technique to obtain long period gratings. The spectral properties of the fabricated long period gratings are accurately predicted by means of finite element method modal analysis and coupled mode theory. An efficient optical grating inscription is performed by using a commercial glass processing system, based on graphite filament heating. The periodic tapering, obtained via heating and stretching technique, of a section of zirconium fluoride optical fiber, provides a measured transmission attenuation with a dip larger than 15 dB at the designed resonant wavelength. A good agreement between the simulation and the characterization is obtained. These results pave the way for a number of applications, among which notch filtering in fiber amplifiers and sensing in the Mid-Infrared spectral range
Graph Neural Networks for fluid mechanics: data-assimilation and optimization
Full text linkThis PhD thesis investigates the application of Graph Neural Networks (GNNs) in the field of Computational Fluid Dynamics (CFD), with a focus on data-assimilation and optimization. The work is structured into three main parts: data-assimilation for Reynolds-Averaged Navier-Stokes (RANS) equations based on GNN models; data-assimilation augmented by GNN and adjoint-based enforced physical constraint; fluid systems optimization by ML techniques.
In the first part, the thesis explores the potential of GNNs to bypass traditional closure models, which often require manual calibration and are prone to inaccuracies. By leveraging high-fidelity simulation data, GNNs are trained to directly learn the unresolved flow quantities, offering a more flexible framework for the RANS closure problem. This approach eliminates the need for manually tuned closure models, providing a generalized and data-driven alternative. Moreover, in this first part, a comprehensive study of the impact of data quantity on GNN performance is conducted, designing an Active Learning strategy to select the most informative data among those available.
Building on these results, the second part of the thesis addresses a critical challenge often faced by ML models: the lack of guaranteed physical consistency in their predictions. To ensure that the GNNs not only minimize errors but also produce physically valid results, this part integrates physical constraints directly into the GNN training process. By embedding key fluid mechanics principles into the machine learning framework, the model produces predictions that are both reliable and consistent with the underlying physical laws, enhancing its applicability to real-world problems.
In the third part, the thesis demonstrates the application of GNNs to optimize fluid dynamics systems, with a particular focus on wind turbine design. Here, GNNs are employed as surrogate models, enabling rapid predictions of various design configurations without the need for performing a full CFD simulation at each iteration. This approach significantly accelerates the design process and demonstrates the potential of ML-driven optimization in CFD workflows, allowing for more efficient exploration of design spaces and faster convergence toward optimal solutions.
On the methodology side, the thesis introduces a custom GNN architecture specifically tailored for CFD applications. Unlike traditional neural networks, GNNs are inherently capable of handling unstructured mesh data, which is common in fluid mechanics problems involving irregular geometries and complex flow domains. To this end, the thesis presents a two-fold interface between Finite Element Method (FEM) solvers and the GNN architecture. This interface transforms FEM vector fields into numerical tensors that can be efficiently processed by the neural network, allowing data exchange between the simulation environment and the learning model.Cette thèse de doctorat explore l'application des réseaux de neurones en graphes (GNN) dans le domaine de la dynamique des fluides numérique (CFD), avec un accent particulier sur l'assimilation de données et l'optimisation. Le travail est structuré en trois parties principales: assimilation de données pour les équations de Navier-Stokes moyennées à la Reynolds (RANS) basée sur des modèles GNN; assimilation de données augmentée par les GNN avec des contraintes physiques imposées par la méthode adjointe; optimisation des systèmes fluides par des techniques d'apprentissage automatique (ML).
Dans la première partie, la thèse examine le potentiel des GNN pour contourner les modèles de fermeture traditionnels, qui nécessitent souvent une calibration manuelle et sont sujets à des inexactitudes. En exploitant des données de simulation à haute fidélité, les GNN sont entraînés à apprendre directement les quantités non résolues de l'écoulement, offrant ainsi un cadre plus flexible pour le problème de fermeture des équations RANS. Cette approche élimine le besoin de modèles de fermeture calibrés manuellement, fournissant une alternative généralisée et basée sur les données. De plus, dans cette première partie, une étude approfondie de l'impact de la quantité de données sur les performances des GNN est réalisée, avec la conception d'une stratégie d'Active Learning pour sélectionner les données les plus informatives parmi celles disponibles.
Sur la base de ces résultats, la deuxième partie de la thèse aborde un défi critique souvent rencontré par les modèles d'apprentissage automatique: l'absence de garantie de cohérence physique dans leurs prédictions. Afin de garantir que les GNN non seulement minimisent les erreurs, mais produisent également des résultats physiquement valides, cette partie intègre des contraintes physiques directement dans le processus d'entraînement des GNN. En incorporant les principes clés de la mécanique des fluides dans le cadre de l'apprentissage automatique, le modèle produit des prédictions à la fois fiables et cohérentes avec les lois physiques sous-jacentes, améliorant ainsi son applicabilité aux problèmes réels.
Dans la troisième partie, la thèse démontre l'application des GNN pour optimiser les systèmes de dynamique des fluides, avec un accent particulier sur la conception des éoliennes. Ici, les GNN sont utilisés comme modèles de substitution, permettant des prédictions rapides de diverses configurations de conception sans avoir besoin de réaliser une simulation CFD complète à chaque itération. Cette approche accélère considérablement le processus de conception et montre le potentiel de l'optimisation basée sur l'apprentissage automatique dans le cadre de la CFD, permettant une exploration plus efficace des espaces de conception et une convergence plus rapide vers des solutions optimales.
Sur le plan méthodologique, la thèse introduit une architecture GNN sur mesure spécifiquement adaptée aux applications CFD. Contrairement aux réseaux de neurones traditionnels, les GNN sont intrinsèquement capables de gérer des données de maillage non structurées, ce qui est courant dans les problèmes de mécanique des fluides impliquant des géométries irrégulières et des domaines d'écoulement complexes. À cette fin, la thèse présente une interface en deux parties entre les solveurs de la méthode des éléments finis (FEM) et l'architecture GNN. Cette interface transforme les champs vectoriels FEM en tenseurs numériques pouvant être traités efficacement par le réseau neuronal, permettant ainsi l'échange de données entre l'environnement de simulation et le modèle d'apprentissage
Effect of wire feed rate on ER70S-6 microstructure of wire arc additive manufacturing process
No full textThis study investigates the manufacturing and characterization of ER70S-6 single line-wire arc additive manufacturing (WAAM) using a cobot to produce defect-free samples. Adjusting the wire feed rate aims to improve efficiency in fabricating steel frames for building structures. In this regard, 1-mm-diameter ER70S-6 wire samples were fabricated with feeder rates ranging from 4.5 to 6.5 m/min, maintaining a constant robot speed of 7 m/min. The voltage and current of the WAAM machine were controlled between 14.8–16 V and 125–159 A, respectively. Microhardness profiles and grain sizes at grain dilution of the area were systematically compared to monitor the solidification behavior after the process. Electron backscatter diffraction (EBSD) analysis assessed the crystallographic orientations and calculated the grain sizes. Optical microscopy and scanning electron microscopy (SEM) revealed uniform, defect-free surfaces of metal mixing during the WAAM process in the 100 ± 10 μm upper than dilution area. As a result of two types of cooling processes following WAAM, transferring heat into the substrate and environment, the formation of Widmanstätten ferrite on top of the beads’ peaks was more extensive. The grain size in the middle of the bead ranged from approximately 8.6 to 11.6 μm, while at the dilution area, it decreased to 3 to 5.6 μm. This variation influenced the microhardness, which reached 300 ± 15 HV
Raw Earth Buildings and Industry 4.0: An Overview of the Technology and Innovation of the MUD-MADE Project
No full textResearch on digital production technologies for the building sector, although several decades behind other sectors, is beginning to become more and more systematic. The use of natural materials such as raw earth makes the sustainability of such processes even more pronounced than current building solutions. Despite this, many limitations still prevent the use of digital technologies employing raw earth for construction from becoming current. The article investigates the state of research on the topic, identifying the reasons for current limitations. It also describes the MUD-MADE research project that aims to overcome these limitations and make the use of digitally fabricated raw earth components for the building sector a reality. This project proposes a novel artificial intelligence-supported workflow for designing raw earth building components produced with digital manufacturing technology. The workflow can support the designer in a multi-objective optimization involving different performances (e.g., thermal, structural, acoustic) by saving material and maintaining feasibility. The workflow exploits parametric design to set a predefined visual script able to support the user. Indeed, the predefined script will allow the user to design a building component by selecting (or creating) different possible external shapes and infill geometries. The designer can include information about the local material and the available technology to digitally manufacture the component directly in the predefined code. In addition, the predefined script sets the boundary conditions and priorities for the expected performances. Moreover, performance priorities are defined by the user based on the requirements of the component to be achieved. Finally, artificial intelligence, exploiting the artificial neural network (ANN) will support the designer by automatically identifying the optimal configuration among the possible combinations of parameters and generative algorithms
Search for light long-lived particles decaying to displaced jets in proton–proton collisions at s = 13.6 TeV
No full text: A search for light long-lived particles (LLPs) decaying to displaced jets is presented, using a data sample of proton-proton collisions at a center-of-mass energy of 13.6 TeV, corresponding to an integrated luminosity of 34.7 fb-1, collected with the CMS detector at the CERN LHC in 2022. Novel trigger, reconstruction, and machine-learning techniques were developed for and employed in this search. After all selections, the observations are consistent with the background predictions. Limits are presented on the branching fraction of the Higgs boson to LLPs that subsequently decay to quark pairs or tau lepton pairs. An improvement by up to a factor of 10 is achieved over previous limits for models with LLP masses smaller than 60 GeV and proper decay lengths smaller than 1 m. The first constraints are placed on the fraternal twin Higgs (FTH) and folded supersymmetry (FSUSY) models, where the lower bounds on the top quark partner mass reach up to 350 GeV for the FTH model and 250 GeV for the FSUSY model
Utile a cosa? Potenzialità trasformative e generative del piano paesaggistico della Puglia
No full textIl contributo propone una riflessione sulla messa in opera del Piano paesaggistico territoriale regionale della Puglia (Pptr), focalizzando l’attenzione sull’utilità della parte più originale del piano, quella relativa alla produzione sociale del piano e del paesaggio, e assumendo come caso di studio l’area rurale dei ‘Paduli’, nel cuore del Salento leccese, per evidenziare le potenzialità trasformative e generative del piano
Statistical Modeling and Characterization of Laser Marking on AISI 301LN Stainless Steel Using Short-Pulsed Fiber Laser
Full text linkThis study explores the effects of nanosecond short-pulsed fiber laser processing on AISI 301LN stainless steel, focusing on optimizing surface characteristics through precise parameter control. Using a Design of Experiments (DOE) approach combined with response surface methodology (RSM), the influence of laser power (30–60 W) and the number of laser passes (5–15 times) was systematically investigated. The results demonstrate that increasing the laser power and passes significantly affected the surface properties. The highest surface roughness of 16.8 μm and engraving width of 51 μm were achieved with 60 W power and 15 passes, whereas the lowest roughness of 13.8 μm and width of 35 μm were observed with 30 W power and 5 passes. Wettability measurements revealed an inverse correlation with roughness, with contact angles ranging from 86.4° for rougher surfaces to 92.4° for smoother textures. The findings demonstrate the capability of short-pulsed fiber laser processing to tailor surface properties effectively, with potential applications in manufacturing and surface engineering where controlled roughness and wettability are critical