Politecnio die Bari - Catalogo di prodotti della Ricerca
Not a member yet
36616 research outputs found
Sort by
Sustainable aviation finance: Integration of environmental impact mitigation and green investment strategies
No full textCan Dominant Runoff Generation Mechanisms Be Disentangled Through Hypothesis Testing? Insights From Integrated Hydrological‐Hydrodynamic Modeling
No full textIdentifying flood-inducing processes remains a challenge in catchment hydrology due to the complex runoff dynamics, particularly in semi-arid regions where surface and subsurface mechanisms alternatively drive streamflow across seasons. Tracer data can help identify hydrograph sources, but they are often unavailable or lack sufficient temporal resolution. To aid process identification at the event-scale, we developed an integrated hydrological-hydrodynamic framework and compared multiple model hypotheses informed by hydrological signatures. We systematically tested these hypotheses through falsification, meta-evaluation, spatial validation, and posterior diagnostics, using the semi-arid Salsola nested catchment in southern Italy as case study. While all model structures performed well on common calibration metrics, differences emerged in spatial transferability tests and alternative diagnostic assessments. Some models, despite strong performance, exhibited inconsistent representations of internal runoff mechanisms, indicating that they achieved good results for the wrong reasons. Furthermore, the choice of routing schemes significantly influenced high-peak estimations and overall model performance, particularly when Horton-type overland flow was considered. This underscores the need to treat routing methods as a key component in event-scale modeling. Our findings reveal that during consecutive storm events in the study catchment, surface processes dominate the initial stages, whereas subsurface processes become more influential in later events, providing valuable insights that may be applicable to similar semi-arid regions. Overall, we emphasize the importance of hypothesis testing in runoff process identification, which can compensate for the absence of hydrochemical data for hydrograph separation. Additionally, our results highlight the value of a landscape-based modeling approach for distinguishing alternative runoff generation processes
Generation of a digital structured database for the modelling of landslide mechanisms: two prototypes designed for different geo-mechanical contexts
No full textThis research presents a methodology for diagnosing landslides, demonstrated through two prototypes designed for distinct geo-mechanical environments. The study addresses both the highly complex, geo-hydro-mechanical context of the Pianello slope in Bovino, located in the Daunia Apennines, and the less complex, but potentially more disastrous, quick clay hazard case. The methodology integrates multidisciplinary data collection, advanced modelling techniques, and real-time monitoring to effectively diagnose landslide behaviour. It begins with the collection of critical data, such as borehole logs, geotechnical laboratory tests, in-situ tests, and geophysical surveys, which provide insights into underground conditions, including lithological heterogeneity, shear strength, and groundwater dynamics.
The methodology then incorporates field observations, structural damage surveys, and multi-temporal remote sensing data, using historical maps, aerial photographs, and LiDAR-derived digital elevation models (DEMs) to track the evolution of landslide bodies over time, while inclinometers and piezometers monitor ongoing displacements and porewater pressures. This combination of historical and real-time data is essential for identifying predisposing factors such as tectonic fracturing and anthropogenic activities, as well as triggering mechanisms like rainfall infiltration. These datasets are synthesized into models that include geological cross-sections, geotechnical models, and ERT profiles, which provide a detailed understanding of slope behaviour. In the case of Pianello, these models reveal deep-seated, roto-translational landslides and shallow earthflows, each with distinct movement patterns. DEM differencing and building inclination vectors are used to track surface deformations, while the GIS-based dashboard enables real-time visualization and analysis. The iterative process of feedback between data collection, modelling and monitoring is an essential aspect of the methodology, ensuring adaptive diagnostics as new data is acquired. Additionally, the quick clay hazard case where the main target is the mapping the occurrences of the target soil material, though less complex, demonstrates the versatility of the methodology in managing more catastrophic landslide risks. The approach proves scalable and adaptable to different geomechanical contexts, offering a framework for landslide risk management in diverse regions. In conclusion, the proposed methodology offers an advancement in landslide diagnostics, integrating geological, geotechnical, hydrogeological, and geomorphological data into a digital twin that serves as a digital structured database, for site-scale geotechnical soil characterisation
The multifunctional enhancement of sheeptracks networks. Policies, practices and perspectives for a reinterpretation of transhumance routes.
No full textThe present doctoral thesis explores the reinterpretation and multifunctional enhancement of transhumance routes, long-distance infrastructure historically used for seasonal livestock migration, which have undergone significant repurposing and transformation. In the context of increasing pressures of urbanisation, ecological fragmentation, environmental degradation and rural depopulation, the study argues that these inherited networks retain a unique potential to support contemporary strategies for sustainable territorial development. The work seeks to reposition
transhumance routes as multifunctional green infrastructures that can simultaneously address ecological, cultural, and socio-economic challenges.
In order to provide a comprehensive overview of the multifunctional potential of transhumance routes, the thesis employs three key concepts: the natural capital,
the ecosystem services and green infrastructure. A thorough analysis of the present ecosystem services provided by these routes is conducted, and their
potential to be recognised as green infrastructures is explored, in accordance with the extant definitions derived by scientific literature.
A central case study of the research is the Apulian transhumance routes (tratturi) network, which is one of the most extensive and historically rich in Europe. The
analysis demonstrates the potential for tratturi to be integrated into broader frameworks of territorial planning, ecological networks, slow mobility and cultural
development
A Comprehensive Overview of Heritage BIM Frameworks: Platforms and Technologies Integrating Multi-Scale Analyses, Data Repositories, and Sensor Systems
No full textThe concept of HBIM (Historic/Heritage Building Information Modeling) has attracted growing interest within research communities in recent years, as reflected in an expanding body of literature exploring its potential in data acquisition and modeling, historical evolution documentation, heritage management, and condition analysis. Yet, new challenges arise in extended HBIM capabilities by integration and interoperability with other technologies and environments for comprehensive heritage assessment. In this context, this paper presents a scoping review, based on the PRISMA protocol, of 60 publications from the Scopus database that document research frameworks and applications of IDPs (integrated digital platforms), where HBIM is combined with different systems to enhance data richness, functionality, and analytical evaluation, as well as to exchange, interpret, and use information effectively. The results show three major thematic areas, namely multi-scale analyses based on HBIM and GIS (geographic information systems); multi-source data repositories development; and sensor networks integration with advanced IoT (Internet of Things) systems. The overview outlines how these frameworks foster the development of interoperable, multi-layered, and data-driven ecosystems, advancing HBIM to an operational component in heritage management and enabling predictive diagnostics and real-time monitoring, while current limitations in semantic consistency, automation, and scalability still hinder full implementation
Advanced embedded systems for autonomous robots control
Get PDFIn recent years, Autonomous Underwater Vehicles (AUVs) and Remotely Operated Underwater Vehicles (ROUVs) have become essential for various underwater missions, including environmental monitoring, offshore resource exploration, structural inspection, and rescue operations. A primary challenge in these missions is precise underwater navigation, particularly during structure inspections, where acoustic and optical systems offer complementary benefits. These tasks require building a local 3D map of the object using inertial systems, a process complicated by the fact that global navigation methods like LBL (Long Base Line) or USBL (Ultra-Short Base Line) are often impractical due to acoustic signal transmission limitations from surface beacons. While optical systems deliver detailed object information, they are hindered by issues like stereo vision distortions, lighting variability, and noise from suspended particles. Conversely, Forward Looking Sonar (FLS) provides robustness against environmental noise but operates at lower resolution and in polar coordinates, complicating data alignment with optical data devices like stereo camera.
At the first part of our research we developed a nonlinear model of the ROUV in Computer-Aided Design (CAD) application and identified essential parameters, such as hydrostatic and hydrodynamic coefficients, through Computational Fluid Dynamics (CFD) simulations. These findings enabled accurate modeling of dynamic behaviors, including Coriolis effects and damping forces, forming the basis of the vehicle’s control equations. Additionally, a fractional-order PI controller was designed for yaw control, derived from a 6 DoF nonlinear model of the ROUV. The integration of Robot Operating System (ROS) and the Gazebo simulator allowed testing of control algorithms and sensor interactions within a digital twin environment, which included an Inertial Measurement Unit (IMU), FLS, Doppler Velocity Log (DVL) and stereo camera. This setup facilitated a smooth transition of navigation, computer vision, and path-planning functionalities from simulation to real-world deployment.
Based on the designed nonlinear model of the ROUV and the developed control algorithm, we integrated a multimodal inertial-visual mapping and navigation system. This system enabled us to analyze the impact of the vehicle’s movement on the accuracy and stability of the visual-inertial system in a dynamically changing environment. Mapping, navigation, and control of the underwater vehicle equipped with a visual-inertial system that merges multimodal data from FLS, DVL, IMU, and a stereo camera devices through unsupervised deep-learning detector and descriptor algorithm is a core study of this research.
Our research focuses on the geometric and computational integration of optical and sonar images. Given that a classical computer vision and supervised deep-learning feature matching algorithms is inadequate for opti-acoustic data fusion, we introduce an unsupervised feature matching approach tailored to sonar and optical datasets. This novel method enhances motion estimation by utilizing a hybrid direct and indirect matching strategy together with opti-acoustic epipolar constrains, where sonar data refines the depth of optical visual features. Both Sonar-to-Optical and Optical-to-Sonar mappings are employed to improve feature matching reliability under varying lighting and turbidity conditions.
In our research, we introduced opti-acoustic image processing, including a calibration approach for each device and both devices the sonar and stereo camera, along with imaging enhancements to improve data fusion quality. To further enhance degraded optical images, we implemented a method combining stereo calibration in air and further data improving via physically guided underwater image enhancement framework based on synthetic and real images integration. By integrating unsupervised deep learning algorithm we was able to map sonar features onto optical images, enhancing image depth estimation while preserving key details. Testing across various scene geometries showed substantial improvements in visual odometry accuracy, which is crucial for effective navigation and inspection in challenging underwater conditions.
Lastly, this research presents a robust framework for simulating underwater missions by combining CFD-based dynamics modeling, control algorithms, and sensor fusion techniques with optical and acoustic data. This comprehensive platform addresses underwater navigation and 3D mapping complexities, providing a unified system for testing and optimizing ROUV operations in diverse environments
Off-Axis monitoring of the laser beam directed energy deposition process of metals (DED-LB/M)
No full textLa Fabbricazione Additiva (AM) ha rivoluzionato la progettazione e la produzione di componenti metallici
complessi, offrendo flessibilità e personalizzazione senza pari. Tra i vari processi di AM, il processo di
Deposizione Diretta di Metalli mediante Fascio Laser (DED-LB) eccelle nella costruzione, riparazione e
miglioramento di parti metalliche, apportando vantaggi in particolare ai settori ad alte prestazioni come
quello aerospaziale. Tuttavia, la sua complessità e variabilità portano spesso a difetti che compromettono
la qualità finale delle parti, sottolineando la necessità di strategie avanzate di monitoraggio e controllo.
Questa tesi indaga le tecniche di monitoraggio fuori-asse per il processo DED-LB/M focalizzandosi sul
miglioramento della rilevazione dei difetti e sull'ottimizzazione della qualità, efficienza e affidabilità del
processo attraverso nuove soluzioni. Il monitoraggio off-axis ha il vantaggio significativo di catturare dati
spaziali relativi al bagno di fusione e alla dinamica della deposizione che spesso non sono visibili dalle
prospettive dei dispositivi in asse, ed è anche relativamente più facile da implementare. La ricerca esplora
metodi di monitoraggio termico e ottico, insieme ad approcci guidati dall'intelligenza artificiale (IA), per
sviluppare un quadro completo di monitoraggio in-process per il processo DED-LB/M.
Lo studio esamina l'evoluzione, i vantaggi e le applicazioni della AM in vari settori, con un focus specifico
sul DED-LB/M. Vengono analizzati i componenti chiave del processo, come il sistema laser, il meccanismo
di alimentazione della polvere e i sistemi di controllo, evidenziandone le applicazioni, come la produzione
di parti 3D complesse e la riparazione di componenti meccanici, mediante l’uso di materiali comuni nel
settore aerospaziale, come acciai, superleghe a base di nichel e leghe di titanio.
Vengono esaminate le tecnologie di monitoraggio all'avanguardia esistenti, seguite dalla presentazione di
diversi casi studio sul monitoraggio termico e ottico utilizzando telecamere CCD/CMOS. Questi sistemi
analizzano l'evoluzione della pozza fusa, consentendo di rilevare, durante il processo di deposizione,
distorsioni geometriche e anomalie come spegnimenti del laser e interruzioni del flusso di polvere. Viene
inoltre introdotto un sistema di monitoraggio multi-vista, che offre soluzioni flessibili adatte a varie
configurazioni di DED-LB/M, in grado di catturare con maggior accuratezza l'evoluzione spaziale della pozza
fusa.
Viene presentato uno studio comparativo tra nuovi metodi di monitoraggio mediante scansione laser e
quelli basati sull’uso di telecamere CMOS ad alta risoluzione che mette in evidenza i rispettivi punti di forza
e ne suggerisce le relative applicazioni ottimali. Sono stati sviluppati indicatori chiave di prestazione (KPIs)
per la rilevazione di difetti geometrici e sub-superficiali attraverso algoritmi personalizzati di elaborazione
delle immagini, validati da analisi metallografiche. Viene proposto un approccio innovativo per la
valutazione in-process dell'efficienza di deposizione, utilizzando i dati di uno scanner laser prototipale per
identificare deviazioni del volume delle tracce depositate rispetto a quello previsto.
Sono state applicate tecniche di IA, per la prima volta nella letteratura scientifica, per monitorare in modo
innovativo la distanza di standoff e rilevare l'eventuale otturazione dell'ugello. In particolare, un modello
di rilevamento degli oggetti, appositamente addestrato, migliora la sicurezza e il controllo del processo
avvisando gli utenti su condizioni critiche come l'otturazione dell'ugello o la diminuzione della distanza di
standoff. D’altro canto, un modello di machine learning (ML) getta le basi per poter prevedere le modifiche
ottimali da apportare a parametri di processo chiave, come la distanza di standoff o la potenza laser, per
migliorare la precisione dimensionale dei prodotti attraverso un approccio “feed forward”.
Attraverso questi casi studio, questa tesi fornisce preziose informazioni sulle tecniche di monitoraggio
fuori-asse, contribuendo a nuove soluzioni di controllo qualità nella fabbricazione additiva dei metalli e al
progresso del processo DED-LB/M.Additive Manufacturing (AM) has revolutionized the design and production of complex metal components,
offering unparalleled flexibility and customization. Among the various AM processes, the Laser Beam
Directed Energy Deposition (DED-LB) process excels in the construction, repair, and enhancement of metal
parts, particularly benefiting high-performance industries such as aerospace. However, its complexity and
variability often result in defects that compromise the quality of the final parts, underscoring the need for
advanced monitoring and control strategies.
This thesis investigates off-axis monitoring techniques for DED-LB/M, focusing on improving defect
detection and optimizing process efficiency and reliability through new solutions. Off-axis monitoring has
the significant advantage of capturing spatial data related to the melt pool and deposition dynamics, which
are often not visible from on-axis perspectives, and is also relatively easier to implement. The research
explores thermal and optical monitoring methods, along with artificial intelligence (AI) driven approaches,
to develop a comprehensive in-process monitoring framework for DED-LB/M.
The study examines the evolution, benefits, and applications of AM across various industries, with a
specific focus on DED-LB/M. Key process components, such as the laser system, powder feeding
mechanism, and control systems, are analyzed, highlighting applications like the production of complex
3D parts and the repair of mechanical components using materials crucial in the aerospace sector, such as
steels, nickel-based superalloys, and titanium alloys.
Existing state-of-the-art monitoring technologies are reviewed, followed by the presentation of several
case studies on thermal and optical monitoring using CCD/CMOS cameras. These systems analyze the
evolution of the melt pool, allowing for in-process detection of geometric distortions and anomalies such
as laser shutdowns and powder flow interruptions. A multi-view monitoring system is also introduced,
offering flexible solutions suitable for various DED-LB/M configurations, capable of more accurately
capturing the spatial evolution of the melt pool.
A comparative study between new laser scanning monitoring methods and those based on high-resolution
CMOS cameras is presented, highlighting their respective strengths and suggesting optimal applications.
Key performance indicators (KPIs) for detecting geometric and subsurface defects were developed through
custom image processing algorithms, validated against metallographic analyses. An innovative approach
for in-process evaluation of deposition efficiency is proposed, utilizing prototype laser scanner data to
identify deviations in the deposited track volume.
AI techniques have been applied for the first time in the scientific literature to monitor standoff distance
and detect potential nozzle clogging. Specifically, a custom-trained object detection model enhances
process safety and control by alerting users to critical conditions such as nozzle clogging or reduced
standoff distance. On the other hand, a machine learning (ML) model lays the groundwork for predicting
optimal adjustments to key process parameters, such as standoff distance or laser power, to improve the
dimensional accuracy of the products through a feed forward approach.
Through these case studies, this thesis provides valuable insights into off-axis monitoring techniques,
contributing to new quality control solutions in metal additive manufacturing and advancing the DED
LB/M process
Modeling the spatial resolution of magnetic solitons in magnetic force microscopy and the effect on their sizes
Get PDFIn this work, we theoretically investigated the spatial resolution of magnetic solitons and the variations of their sizes when subjected to magnetic force microscopy (MFM) measurement. In addition to tip-sample separation, we considered tip magnetization reversal and showed that the magnetic soliton size measurement can be strongly affected by the magnetization direction of the tip. In addition to previous studies that only consider thermal fluctuations, we developed a theoretical method to obtain the minimum observable length of a magnetic soliton and its length variation due to the influence of the MFM tip by minimizing the magnetic energy of the soliton. We show that a simple spherical model for the MFM tip can capture most of the physics underlying the tip-sample interactions, with the key requirement being an estimate of the magnetization field within the sample. Our model uses analytical and numerical calculations and prevents overestimation of the characteristic length scales from MFM images. We compared our method with available data from MFM measurements of domain wall widths and performed micromagnetic simulations of a skyrmion-tip system, finding good agreement for both attractive and repulsive domain wall profile signals and for the skyrmion diameter in the presence of the magnetic tip. In addition, the theoretically calculated frequency shift shows good qualitative agreement with experimental measurements. Our results provide significant insights for a better interpretation of MFM measurements of different magnetic solitons and will be helpful in the design of potential reading devices based on magnetic solitons as information carriers
Explainable Intelligent Inspection of Solar Photovoltaic Systems with Deep Transfer Learning: Considering Warmer Weather Effects Using Aerial Radiometric Infrared Thermography
Get PDFSolar photovoltaic (SPV) arrays play a pivotal role in advancing clean and sustainable energy systems, with a worldwide total installed capacity of 1.6 terawatts and annual investments reaching USD 480 billion in 2023. However, climate disaster effects, particularly extremely hot weather events, can compromise the performance and resilience of SPV panels through thermal deterioration and degradation, which may lead to lessened operational life and potential failure. These heatwave-related consequences highlight the need for timely inspection and precise anomaly diagnosis of SPV panels to ensure optimal energy production. This case study focuses on intelligent remote inspection by employing aerial radiometric infrared thermography within a predictive maintenance framework to enhance diagnostic monitoring and early scrutiny capabilities for SPV power plant sites. The proposed methodology leverages pre-trained deep learning (DL) algorithms, enabling a deep transfer learning approach, to test the effectiveness of multiclass classification (or diagnosis) of various thermal anomalies of the SPV panel. This case study adopted a highly imbalanced 6-class thermographic radiometric dataset (floating-point temperature numerical values in degrees Celsius) for training and validating the pre-trained DL predictive classification models and comparing them with a customized convolutional neural network (CNN) ensembled model. The performance metrics demonstrate that among selected pre-trained DL models, the MobileNetV2 exhibits the highest F1 score (0.998) and accuracy (0.998), followed by InceptionV3 and VGG16, which recorded an F1 score of 0.997 and an accuracy of 0.998 in performing the smart inspection of 6-class thermal anomalies, whereas the customized CNN ensembled model achieved both a perfect F1 score (1.000) and accuracy (1.000). Furthermore, to create trust in the intelligent inspection system, we investigated the pre-trained DL predictive classification models using perceptive explainability to display the most discriminative data features, and mathematical-structure-based interpretability to portray multiclass feature clustering