FCT: DEMI - Teses de Doutoramento
URI permanente para esta coleção:
Navegar
Entradas recentes
- Open and Sustainable Innovation strategy in the port ecosystem: A case study of the Portuguese Port of SinesPublication . Sabino, Maria Rosilene; Cabrita, Maria do Rosário; Pinho, Tiago; Mendes, AnaThis research examines the application of open and sustainable innovation strategies in the port sector, utilising the Port of Sines as a case study. Building on a conceptual framework structured around environmental, organisational, cultural, and technological dimensions, the study shows that sustainable port transformation relies on the alignment of multi-level governance, stakeholder engagement, and a participatory innovation culture. Findings indicate that, despite barriers such as regulatory fragmentation, cultural resistance, and uneven digital maturity, enabling factors are already present—including investment in digital platforms, collaborative governance practices, and adaptive leadership—which can accelerate the green and digital transition. The original scientific contribution of this thesis lies in the development of a strategic framework for open and sustainable innovation in port ecosystems, which integrates organisational, cultural, and technological dimensions with sustainability objectives. This framework fills a gap in the literature by proposing a systemic view of port innovation that goes beyond isolated practices, positioning co-creation and relational governance as foundations for sustainable transformation. Accordingly, the study advances theoretical understanding of sustainability-oriented open innovation and provides practical recommendations for port authorities and policymakers seeking to align competitiveness, digitalisation, and environmental responsibility.
- Enhancement of functionally graded material created by multi-feed wire and arc additive manufacturing using full-field identification methodPublication . Cunha, Filipa Alexandra Grifo; Xavier, José; Santos, TelmoMaterials and their importance to society have evolved over time. In today’s challenges, material innovation for economic and environmental sustainability has gained prominence. Additive manufacturing (AM) is a promising technology for producing 3D parts. Some AM processes allow for the production of functionally gradient materials (FGMs) in a single step, an advantage over conventional methods. Recently, FGM metal parts have been manufactured by direct arc and wire energy deposition (WAAM) to reduce material costs. Traditional tests are inadequate for evaluating the mechanical properties of FGMs due to their heterogeneity. Therefore, specialized approaches using full-field measurements and advanced inverse methods are required to accurately characterize the unique composition and structure of FGMs. This thesis presented a numerical analysis and experimental validation of the behaviour of a Cu-CuAl FGM. In the inverse identification, the Virtual Fields Method (VFM) was used to extract material parameters from full-field deformation measurements obtained by Digital Image Correlation (DIC). The FGMs were fabricated by WAAM, using gas tungsten arc welding (GTAW) and inert metal gas (MIG) welding as heat sources. The material parameter identification was carried out based on simulation first, as verification, then on experimental data. A library of Cu-CuAl material compositions was also created to evaluate their properties. This study evidenced the feasibility of using WAAM in the production of FGMs. The results show that Cu and Al have very different thermal conductivity coefficients and melting temperatures, which contribute to their metallurgical incompatibility. In this way, Cu and CuAl alloys were combined in an FGM which showed better metallurgical compatibility. It can be concluded that preheating the previous layers results in better cohesion between the different FGM materials.
- Methodologies for optimal design and additive manufacturing of metamaterials combining negative property indexesPublication . Almeida, Cláudia Sofia João de; Coelho, Pedro; Velhinho, AlexandreEngineering is evolving alongside technology, demanding new approaches to design materials capable of dealing with growing structural and functional challenges. Metamaterials are artificially engineered materials designed to exhibit unusual properties, often unattainable in natural materials. Among the most counterintuitive behaviors, Negative Poisson’s Ratio (NPR) and Negative Thermal Expansion (NTE) stands out, showing unique thermomechanical responses such as lateral expansion under tension and contraction upon heating. These unconventional behaviors are promising for a wide range of engineering applications, from aerospace to biomedical devices. This thesis presents the development of a computational framework for the systematic design of metamaterials exhibiting both NPR and NTE, incorporating advanced topology optimization techniques, such as multimaterial, multiobjective and multiscale optimization. Two complementary design strategies are explored, to tailor effective properties at the microscale and embed such behavior into structural applications: (1) a microscale approach, focused on investigating the intrinsic trade-off between NPR and NTE using multimaterial and multiobjective topology optimization and different types of microstructure discretizations (truss-like and continuum-like), obtaining Pareto sets; and (2) a multiscale approach, to investigate how the macrostructure geometry and loading conditions influence the performance and design of the microstructures. The proposed methodology incorporates homogenization techniques, material symmetry constraints and uniform and layer-wise material distribution in the structure to bridge the gap between microscale material behavior and macroscale functional requirements. Through a series of case studies, including multiobjective optimization, the influence of anisotropy, and applications under non-uniform loading such as pure bending, the developed methodology demonstrates its ability to generate functionally graded porous metamaterials with tailored auxetic responses. Regarding the competitive behavior of both indexes, the minimization of the Poisson’s ratio depends mostly on the architecture of the unitcell, while the NTE is due to both architecture and material distribution. So, both can be negative simultaneously, but extremizing one property results in worsening the other. In the multiscale framework, the concept of auxetic structure is introduced, allowing to formulate the optimization problem using the actual structural response, moving beyond the conventional material design approach. This methodology has proved successful in designing auxetic and thermoauxetic structures, while also evidencing the critical role of macroscale effects, such as loading and geometry, on the optimized microstructural layout. Additionally, a computational cost study is performed, providing insights into the scalability and efficiency of the proposed methods. The present work further explores the integration of optimized designs with additive manufacturing, aiming to bridge the gap between design and manufacture. Overall, this work contributes to the advancement of functional metamaterial design by enabling the concurrent tuning of mechanical and thermal properties and ensuring their effectiveness at both the material and structural levels.
- Human-like communication in conversational agents for delivering digital health interventionsPublication . Martins, Ana Catarina dos Santos; Gamboa, Ana; Nunes, Isabel; Lapão, LuísAs the global burden of non-communicable diseases (NCDs) continues to rise, healthcare systems are increasingly looking for adopting Digital Health Interventions (DHIs) to support behavior change and enhance chronic disease management. This thesis explores how human-like communication through Conversational Agents (CAs) can improve patient engagement, personalize care, and support sustainable self-management. Grounded in the Design Science Research Methodology (DSRM), this work integrates evidence from systematic and scoping reviews, user-centered design processes, and real- world pilot studies. It begins with a literature review identifying effective communication strategies and personalization mechanisms in DHIs. A scoping review follows, analyzing the automation techniques and human-like characteristics embedded in CAs across healthcare applications. These findings informed the development of a personalized, automated messaging platform aimed at supporting patients after cardiothoracic surgery. A solution was co-designed with clinicians, nurses, and patients and validated through iterative prototyping, a feasibility study, and a pilot trial at a public hospital in Lisbon. A pilot study was conducted with three patients for 1 month, which allowed assessing acceptability, system usability, perceived effectiveness, and improvements. The system achieved high levels of acceptability and usability and was considered a valuable tool to reinforce clinical recommendations during the post-discharge period by patients and the clinical team. Subsequently, a second iteration leveraged agentic LLMs to personalize behavioral interventions for NCDs in broader contexts, showcasing potential for generalizability. This solution was validated by physicians from a public health center. Key contributions include a framework for developing human-like CAs in healthcare, insights into personalization through behavioral and contextual factors, and the integration of AI for scalable communication. Limitations such as limited number of participants in the pilot studies and integration barriers with existing clinical systems are acknowledged, and future directions emphasize scalability, interoperability, and bidirectional communication. This work contributes to the design of secure, effective, and engaging digital interventions that enhance the continuity of care and empower patients in managing their health.
- A Computational and Deep Learning Framework for Fluid-Structure Interaction in Ascending Thoracic Aortic AneurysmsPublication . Mourato, André Filipe Geraldes; Xavier, José; Brito, Moisés; Fragata, JoséAccording to the World Health Organization (WHO), cardiovascular diseases are the leading cause of death worldwide. Among them, Ascending Thoracic Aortic Aneurysms (ATAAs) have an incidence of 6-10 per 100,000 people per year. Their progression is usually asymptomatic and may lead to fatal events, such as rupture or dissection. Clinical guidelines rely on a geometric criterion to assess risk, but this often fails to predict rupture accurately for all patients, highlighting the need for more robust diagnostic tools. Numerical models have emerged as a promising tool to support the diagnosis and treatment of ATAAs. Simulating the behaviour of complex systems enables the development of patient-specific virtualizations, providing additional information for clinical decision-making. The literature highlights several approaches, with Computational Fluid Dynamics (CFD), Computational Solid Mechanics (CSM), and Fluid-Structure Interaction (FSI) being the most common. Among these, FSI is considered the gold standard. However, its clinical use remains limited due to: (i) the difficulty of building fully patient-specific models; (ii) high computational cost; and (iii) lack of validation against robust datasets. This thesis addresses the second limitation by asking: “Which numerical approaches offer the best trade-off between accuracy and computational cost?”. First, the impact of different prestressing methods in FSI simulations was evaluated. Then, FSI results were compared with more straightforward approaches. Finally, a surrogate model of aortic wall mechanics was developed by combining Deep Neural Networks (DNNs) with CSM simulations. The aortic wall remains under stress throughout the cardiac cycle. Therefore, its reference configuration is not directly available in medical images. To overcome this, algorithms such as Zero Pressure Geometry (ZPG) and Prestress Tensor (PT) are commonly used. PT produced pressure fields consistent with ZPG but estimated a stiffer mechanical response. Incorporating regional material properties improved agreement with ZPG in terms of maximum stresses and strains. As ATAA progresses and stiffness increases, the relevance of FSI decreases. Simpler approaches, such as Reduced Order Models (ROM) and CSM, become suitable alternatives due to their lower computational cost. All methods produced similar pressure fields. However, for estimating Wall Shear Stress (WSS), FSI remained the most reliable. For wall mechanics, coupling CSM with boundary conditions derived from ROM yielded results close to FSI. Surrogate models have been proposed to reduce the computational burden of simulations. Artificial Intelligence (AI)-based methods have shown promising results in biomechanical modeling. In this work, a DNN was successfully trained using CSM simulation data. The model received as input information on aneurysm anatomy, material properties, and pressure fields, and was able to predict the spatial distribution of stresses and strains with an accuracy above 90 %. The results of this thesis contribute to bridging the gap between biomechanical modeling and clinical practice. By comparing numerical strategies and introducing efficient frameworks, this work identifies potential solutions for integrating patient-specific biomechanical analysis into clinical workflows. The proposed methodologies may improve risk stratification and treatment planning of ATAAs, supporting a more personalized cardiovascular approach.
- Additive Manufacturing of Polyetheretherketone (PEEK) for Patient-Specific Implant (PSI) Applications. Developments in the mechanical performance and biofunctionalization of 3D-printed PEEK for the production of load-bearing implantable medical devicesPublication . Rendas, Pedro Miguel Palma; Soares, Bruno; Vidal, Catarina; Figueiredo, LígiaPolyetheretherketone (PEEK) is a high-performance thermoplastic that has emerged as a bio- material alternative to metals like titanium for load-bearing implant manufacture due to its unique set of properties. Additionally, PEEK can be processed using extrusion-based additive manufacturing (AM) where the use of intricate geometries and patient-specific designs enable new possibilities for implant design and manufacturing. Despite this potential, 3D-printed PEEK's use in clinical practice remains limited due to the challenges associated with the control of additive manufacturing outcomes for demanding applications like implantable medical de- vices. In face of these challenges, this thesis explores developments in the mechanical perfor- mance and biofunctionalization of 3D-printed PEEK for patient-specific implant (PSI) manufac- ture. Regarding mechanical performance, PEEK specimens were tested under tensile, flexural, and compressive loads to identify effective printing strategies in the enhancement of strength and stiffness. Results showed that the best strength and stiffness outcomes are obtained with an effective reduction of the void contents of PEEK prints that can be achieved with a modified interlayer translation printing strategy paired with temperatures and deposition sequences that distribute the heat from the nozzle to the entire deposition region. Additionally, fatigue tests highlighted PEEK’s superior strength under high-cycle loading compared to other additively manufactured thermoplastics. Concerning PEEK's biofunctionalization, cell culture assays re- vealed that modifying rough-surface PEEK with sulfonation and hydroxyapatite (HA) created a better environment for fibroblast adhesion and proliferation, especially when compared with unmodified as-printed samples. Lastly, this work discusses the clinical use of 3D-printed PEEK, outlining regulatory requirements and design considerations for PSI production. Together, these developments aim to overcome the challenges posed by the performance and reliability of 3D-printed PEEK for implant manufacture, creating new opportunities for advanced medical solutions in surgical procedures for bone reconstruction.
- Development of the Wire Arc Additive Manufacturing process monitoring and inspection system using Non-Destructive TestingPublication . Ramalho, André Filipe Gaisita; Oliveira, João; Santos, TelmoWire Arc Additive Manufacturing (WAAM) is an Additive Manufacturing (AM) process that enables the production of large parts with complex geometries in reduced production times. Relying on its intrinsically high heat input, WAAM achieves significantly higher deposition rates than laser-based AM processes. However, the same high heat input yields the process susceptible to the formation of defects such as porosity or cracking, geometrical distortions, and inconsistent mechanical properties. To ensure the industrialization of WAAM, reliable in- situ monitoring methods are essential to guarantee the production of high quality, flaw free parts. Accordingly, this work aims to characterize the WAAM process using multi-sensor data and to develop monitoring strategies that enhance the reliability of the process. The potential of acoustic sensing for WAAM monitoring was assessed through the deliberate introduction of contaminations and the identification using frequency and time-frequency analysis on the acoustic data. The acoustic analysis was successfully validated, however it also highlighted the limitations of the application of the Fourier based time-frequency analysis in the detection of subtle variations in the process. In contrast, localized frequency analysis is revealed to be more sensible to process variations. A multi-sensor monitoring approach was developed to detect the onset of process instability by using acoustic and power data from the process. High-speed imaging of the meltpool region was used for ground truth of process stability. A segmentation algorithm was developed to separate the data into three key moments and allow for the extrac- tion of physically meaningful data: arcing, material transfer, and arc ignition. The extraction of data from these intervals allowed for the training of a machine learning model to classify pro- cess stability. This model achieved an F1-Score of 96.2% on thin wall samples and maintained reliable performance when tested on previously unseen geometries and deposition strategies. To complement the previous approach, an adaptative algorithm for extracting meltpool key characteristics from high-speed process imaging was developed. This algorithm was effective in isolating the meltpool and electric arc from the surrounding objects, allowing for the extrac- tion of physically relevant features from the resulting shape which were used to detect the onset of humping and humping-induced porosity. Together, these monitoring methods offer a contri- bution to the development of intelligent WAAM monitoring systems capable of improving the process consistency and part quality.
- MICRO GAS METAL ARC (μ-GMA) DIRECTED ENERGY DEPOSITION: A NOVEL REDUCED-SCALE GMA APPROACH FOR THIN WALL PRINTINGPublication . Dornelas, Paulo Henrique Grossi; Santos, Telmo; Oliveira, JoãoIn recent years, significant efforts have been directed toward advancing metal additive manufacturing (AM) processes to meet the growing miniaturization demands in sectors such as aerospace and electronics. Consequently, new downscaled direct energy deposition (DED) technologies, termed μ-DED, have emerged. This thesis presents the development of a micro gas metal arc (μ-GMA) DED prototype — a downscaled approach utilizing a micrometric wire (ø ≤ 300 μm) for deposition — and evaluates its technical feasibility. This work includes the development of the μ-GMA prototype, characterization of the metallic transfer mode, analysis of deposition parameters, and assessment of the microstructure and mechanical properties of deposited low-alloy and stainless steels. A comparative analysis between μ-GMA and standard GMA-based DED highlights their fundamental differences. Results indicate that the prototype is capable of depositing layers approximately 1 mm wide at a build rate of 30 cm³/h, positioning it between GMA-based DED and other μ-DED methods in terms of dimensional accuracy and build rate. The μ-GMA approach has shown its capacity to deposit low-alloy and stainless-steel walls with mechanical properties comparable to those achieved by conventional GMA-based DED, underscoring its potential as a viable and effective deposition method for these materials.
- Desenvolvimento do Processo Friction Stir Channeling à MicroescalaPublication . Sabor, Wagner de Campos; Vidal, Catarina; Machado, MiguelA literatura existente revela uma lacuna na pesquisa abrangente sobre os limites de processamento do processo Friction Stir Channeling (FSC), especialmente na criação dos menores canais contínuos e integrais usando ferramentas com pinos roscados de 2 mm de diâmetro ou menos. Este estudo está na vanguarda da investigação dos limites extremos do processo FSC em microescala, que representa promessas significativas para o desenvolvimento de trocado- res de calor ultracompactos visando melhorar sua eficiência e sustentabilidade. Para atingir isso, ferramentas personalizadas com dimensões e geometrias específicas foram desenvolvidas para definir parâmetros que produzem microcanais contínuos de forma confiável, otimizando o diâmetro hidráulico dentro dos limites do projeto e da geometria de cada ferramenta. Avaliações rigorosas, incluindo testes de continuidade, estanqueidade, micro tomografia computadorizada, tomografia computadorizada de nêutrons, microdureza e desempenho térmico, foram realizadas para verificar a integridade estrutural e a aplicabilidade dos canais para sistemas de aquecimento e resfriamento supercompactos. Ensaios experimentais foram realizados utilizando ferramentas com diâmetros de pino de 0,2, 0,3, 0,5, 1,0 e 2,0 mm. A fabricação bem sucedida de canais internos foi obtida com pinos de 0,5, 1,0 e 2,0 mm de diâmetro, combinados com diâmetros de shoulder correspondentes de 3,5, 4,0 e 5,0 mm, respectivamente. Os experimentos foram conduzidos em placas da liga de alumínio AW1050-H111 com espessura de 5mm. Notavelmente, o menor canal atingiu um diâmetro hidráulico de 191 μm com um pino roscado de 0,5 mm de diâmetro, classificando-o como um microcanal. Além disso, a eficiência térmica de um modelo compacto de trocador de calor foi avaliada, indicando que, apesar dos altos custos associados à produção das ferramentas especializadas, o processo FSC é um método viável, confiável e repetível para a fabricação de mini e microcanais.
- Development of a fuzzy model to support ergonomic analysis of lower limbs and define requirements for an exoskeletonPublication . Santos, Catarina Dias dos; Nunes, Isabel; Gabriel, Ana Teresa; Quaresma, CláudiaOrthopedic surgery is generally a physically demanding activity in which surgeons are subject to prolonged standing, awkward and sustained body postures, and repetitive and/or forceful movements. Being repeatedly exposed to these risk factors can contribute to the development of work-related musculoskeletal disorders (WRMSD). Given the lower limb’s role in maintaining body stability and balance, WRMSD affecting lower limbs could significantly compromise motor skills. Therefore, performing ergonomic risk assessments of orthopedic surgeries can identify risk factors and prevent lower limb WRMSD among surgeons. However, existing assessment tools often overlook lower limbs and rely on subjective data. In addition, it is also imperative to seek solutions that mitigate or alleviate the Physical risk factors surgeons face. Lower limb exoskeletons, while used in the industrial sector, currently lack research regarding their application in healthcare, particularly in surgical settings. To ensure adoption, such devices must be safe, comfortable, effective in reducing Physical risk factors and acceptable to surgeons. This thesis presents the development of a Fuzzy Decision Support System (FDSS) model aimed to aid ergonomic analysis associated with orthopedic surgical procedures and identifying the design requirements for a lower limb exoskeleton tailored to orthopedic surgeons. The FDSS model is an adaptation of ERGO-X, a FDSS designed to support ergonomic auditing activities related to upper limb WRMSD; the tool helps with the identification, assessment, and provides recommendations regarding the risk factors present at work- stations. In this adaptation, the ERGO-X has been modified to determine the possibility of orthopedic surgeons developing lower limb WRMSD for each surgical task, analyzing each limb independently. The surgical tasks were predetermined in collaboration with the surgeons prior to data acquisition. The FDSS model evaluates the surgery tasks by assessing Physical, Individual, and Psychosocial risk factors associated with the development of lower limb WRMSD. Each risk factor is assessed based on attributes indicating risk factor’s severity. Data from questionnaires characterize Individual and Psychosocial risk factors, while objective data obtained from inertial and EMG sensors characterize Physical risk factors. Questionnaire responses are converted into risk factor’s inadequacy degrees using fuzzy sets. Similarly, objective data are converted into attribute’s inadequacy degrees and aggregated using union fuzzy operators into Physical risk factor’s inadequacy degrees. Membership func- tions are based on literature and occupational physicians input. Inadequacy results are expressed as membership degrees to an inadequacy fuzzy set, defined in the range [0, 1]. The possibility of developing lower limb WRMSD is assessed by aggregating the inadequacy degrees of each risk factor, weighted by their relative importance. A defuzzi- fication process translates the result into a linguistic variable, indicating the likelihood of developing lower limb WRMSD per task. For the relevant tasks, the computed risk factors and attribute’s inadequacy degrees are analyzed to identify the Physical risk factors and respective attributes contributing the most to the overall result. The FDSS model was tested using real surgical data collected at Curry Cabral Hospital. Based on the outputs of the FDSS model, namely the most mechanically stressed body parts, and the analysis of existing lower limb exoskeletons, a set of design requirements was proposed. The design requirements were assessed through the application of the Mudge diagram and Kano model to orthopedic surgeons of Curry Cabral Hospital. Applying the FDSS model to 27 knee arthroplasties revealed consistent biomechanical patterns across the explanations of relevant tasks. These included sustained activation of the gastrocnemius, rectus femoris, and biceps femoris muscles, and significant strain on the hip and ankle joints. These findings shaped the design requirements for the lower limb exoskeleton, emphasizing adaptive actuation, multi-axial joint stabilization, and real-time sensor feedback for responsive support. The FDSS model was validated against LEAT and PSSI tools, and consultations with an occupational physician. The model successfully identified the relevant surgical tasks and the key risk factors and attributes contributing to lower limb WRMSD, confirming its reliability and practical relevance. Regarding the design requirements for the lower limb exoskeleton, surgeon feed- back emphasized stability, freedom of movement, and long-term comfort as top design requirements. This work contributes to filling a critical gap in the ergonomic analysis of surgical practice and offers a foundation for developing assistive technologies that enhance surgeon health and performance.