ESTiG - Dissertações de Mestrado Alunos
Permanent URI for this collection
Browse
Recent Submissions
- Solubility enhancement of antimalarial drugs through eutectic formationPublication . Araújo, Nathalie Ladares deMalaria is a potentially fatal disease transmitted by infected mosquitoes and remains a serious public health problem, especially in sub-Saharan Africa, where it causes hundreds of thousands of deaths annually. One of the main treatments for malaria involves artemisinin-based combination therapies, which have proven highly effective in controlling the disease. However, despite their crucial role in treatment, these therapies face a significant challenge due to the low water solubility of the most commonly used active pharmaceutical ingredients (APIs), such as artemisinin. This limits their absorption into the body, also reducing their therapeutic efficacy. The aim of this thesis is to explore greener solvents to enhance the solubility and bioavailability of antimalarial APIs through the formation of eutectic mixtures. Initially, the COSMO-RS predictive tool was employed to investigate the interactions between API molecules and a variety of terpenes and their mixtures. Thymol was selected based on the lower activity coefficient at infinite dilution, indicating strong interactions with the drugs. Subsequently, the solid–liquid equilibria of binary mixtures composed of thymol and artemisinin, artemether, artesunate, quinidine, tetracycline, quinine, dapsone, sulfadoxine or pyrimethamine were experimentally measured. The data revealed a significant reduction in the melting points of the APIs, particularly for artemisinin, quinidine, and quinine. Moreover, for these three systems, notable negative deviations from ideal behavior were observed, suggesting strong drug–thymol interactions. For these cases, the label "deep eutectic systems" may be appropriately applied. The measured phase diagrams were successfully modeled using the COSMO-RS approach, demonstrating excellent agreement with the experimental results and confirming the model's capability to screen and predict suitable solvent candidates for eutectic systems. The methodologies explored in this work are aligned with the principles of green chemistry, promoting more sustainable solutions for the pharmaceutical industry.
- Disease detection and mapping in olive groves using UAVs and deep learning for precision agriculturePublication . Morais, Maurício Herche Fófano de; Lima, José; Santos, Murillo Ferreira dos; Mendes, João CarlosThis dissertation presents the implementation and validation of a cost-effective, Unmanned Aerial Vehicle (UAV) based system for automated detection and spatial mapping of olive knot disease in olive groves. Addressing the need for accessible and efficient plant disease monitoring in Precision Agriculture (PA), the proposed methodology leverages existing UAV imagery capabilities with lightweight Deep Learning (DL) models, specifically the You Only Look Once (YOLO) object detection architecture, to enable scalable and accurate detection. The academic contributions presented in this work have resulted in two peer-reviewed publications related to the dissertation topic, as detailed at the end of this document. An annotated dataset of UAV-acquired images was compiled, and several state-of-the-art YOLO object detection models were trained and evaluated under identical conditions. The best-performing model achieved a strong F1-score, demonstrating good results in detecting olive knot disease and accurately mapping its spatial distribution within the plantation. The workflow integrates spatial cross-referencing of detections with UAV flight path data and proximity analysis, enabling the assignment of disease detections to individual trees. An interactive map interface, developed using the Folium Python library, provides visualization of the disease distribution and supports practical grove management. The experimental results indicate that cost-effective UAVs and lightweight DL models can be effectively combined for plant disease detection and spatial analysis, offering a robust and scalable approach for real-world agricultural applications. Limitations regarding early symptom detection and image quality are discussed, and directions for future work are proposed.
- Tensile behavior of FFF PLA: experimental characterization and numerical simulationPublication . Enriconi, Maria Eduarda Bordignon; Ribeiro, J.E.; Jocha, João; Araujo, Márcia de AraujoA Fabricação por Filamento Fundido (FFF) é um método amplamente utilizado de manufatura aditiva, conhecido por sua eficiência no uso de materiais e versatilidade geométrica. No entanto, o desempenho mecânico das peças impressas por FFF é influenciado pela variabi-lidade do processo, especialmente em polímeros frágeis como o ácido polilático (PLA). Esta dissertação combina uma revisão de abordagens de simulação numérica com um estudo expe-rimental-numérico para estabelecer uma metodologia de avaliação mecânica confiável para pe-ças de PLA impressas por FFF. A revisão categoriza os modelos existentes com base no esco-amento do material fundido, comportamento térmico e simulação estrutural, destacando limi-tações na integração multifísica e na validação experimental. Experimentalmente, foram reali-zados ensaios de tração utilizando Extensômetro de Vídeo Digital (DVE) e Correlação de Ima-gens Digitais bidimensional (2D-DIC), revelando desafios relacionados à fratura prematura, defeitos entre camadas e mapeamento de deformações. Um modelo de elementos finitos foi desenvolvido para simular o comportamento elástico em condições padronizadas, apresentando concordância com o ensaio mais representativo. Os resultados apoiam o desenvolvimento de procedimentos de metodologias de ensaios padronizados e de referenciais para simulação, con-tribuindo para os objetivos do projeto NaturFAB no avanço de trabalho sustentáveis e com origem local para manufatura aditiva.
- Impacto da IA no ambiente corporativo: desenvolvimento de um assistente com IA generativa para assessores de investimentosPublication . Ferreira, Lucas Scriptore Marques; Lopes, Isabel Maria; Souza, Wesley Angelino deEsta dissertação explora a aplicação prática de modelos de linguagem de grande escala (LLMs), com foco no ChatGPT, no apoio à rotina de trabalho de assessores de investimentos. O objetivo foi desenvolver um assistente virtual personalizado, treinado com engenharia de prompt e dados reais do escritório Wura Invest, credenciado à XP Investimentos. O sistema foi projetado para realizar tarefas como análise de carteiras de clientes, sugestões de alocação, preparação de reuniões e geração de relatórios de performance. A metodologia adotada envolveu o design iterativo do prompt mestre, a curadoria de materiais internos (carteiras, cenários econômicos, posicionamentos) e a realização de testes simulados para validar a eficácia da ferramenta. Os resultados indicaram que o modelo é capaz de realizar análises técnicas coerentes, alinhar recomendações ao perfil do cliente e gerar valor real à prática da assessoria. As limitações observadas estão relacionadas à extração de dados não estruturados e à ausência de conexão com APIs externas, apontando caminhos para futuras melhorias. O projeto contribui para o avanço do uso de Inteligência Artificial (IA) generativa no setor financeiro, ao demonstrar como soluções baseadas em Large Language Models (LLM) podem ser adaptadas com sucesso a contextos profissionais específicos.
- Exploring indicators to identify bias in artificial intelligence modelsPublication . Saadi, Omar; Pereira, Ana I.; Muñoz, Raquel ÁvilaThis thesis details the conception, development, and evaluation of the Bias Detector Tool, an open-source Python application specifically built in the context of this work to assess ethical aspects of artificial intelligence (AI) systems. The tool evaluates machine learning (ML) models and datasets across five key ethical dimensions: fairness, transparency, privacy, robustness, and accountability. It is intended to assist researchers, developers, and policy makers to identify ethical risks in AI systems. The tool employs a modular pipeline that processes CSV-format datasets, applies established metrics using specialized libraries such as Fairlearn, Diffprivlib, and LIME and generates comprehensive output reports in TXT, JSON, and CSV formats. It offers a scoring mechanism that classifies each ethical indicator on a scale from 0 to 1 and provides both technical results and simplified interpretations for non-expert users. To validate the tool’s functionality and reliability, a test scenario was conducted in collaboration with Professor Raquel Ávila Muñoz, an expert in Equality, Diversity, and Inclusion at the Complutense University of Madrid. The evaluation compared datasets with limited ethical integrity such as those lacking diversity or metadata with well-structured datasets showing inclusive data practices. The tool successfully reflected these differences through the scoring system confirming its efficacy in identifying ethically problematic datasets. In summary, this work contributes to the field of responsible AI by offering a practical, transparent, and user-friendly approach to ethical assessment. The tool is publicly available via GitHub, encouraging further adaptation and development by the research community.
- Estudo e análise de um sistema de transmissão utilizando a norma ANSI/AGMA2101-d04 e o método de elementos finitosPublication . Verbaneck, Nicollas Hamon Trevisan; Mesquita, L.M.R.; Nuñez, David LiraDevido sua eficiêcia e confiabilidade, as engrenagens são elementos essenciais em sistemas mecânicos na engenharia, sendo usadas nas mais diversas máquinas. Devido sua alta aplicação, faz-se necessário aprofundar estudos sobre seus modos de falhas, melhorias e otimizações nos sistemas de engrenamento. Com isso, a Associoação Americana de Fabricantes de Engrenagens, AGMA (do inglês – American Gear Manufacturers Association) descreveu a norma ANSI/AGMA 2101-D04 (2016) onde baseia-se em padrões desenvolvidos a partir de pesquisas e testes empíricos em que busca pontos de melhoria e redução de falhas em projetos. Definindo tensões de contato entre os dentes das engrenagens devido ao engrenamento e tensões de flexão. O presente trabalho buscou avaliar e comparar os resultados de tensão de um sistema de transmissão de uma máquina de pitting do tipo FZG, fabricado com Aço SAE 4118, por meio de duas abordagens distintas: Uma análise baseada nas diretrizes da norma ANSI/AGMA 2101-D04 (2016) e uma análise numérica de simulação utilizando o Método de Elementos Finitos no software ANSYS®. O método analítico baseia-se em equações e fórmulas que resultam de pesquisas, experimentos e testes desenvolvidos ao longo do tempo. Por outro lado, o método numérico emprega técnicas computacioanis para resolver problemas práticos de simulação na engenharia e em outras áreas, utilizando a discretização de equações matemáticas. Também, entender como a mudança de alguns parâmetros geométricos e condições de contorno podem interferir na resistência do sistema. Avaliando as vantagens e limitações de cada abordagem, concluindo na implicação da escolha do método apropriado em diferentes cenários de engenharia.
- The application of lean construction solutions in site works of residential refurbishment projects: an overviewPublication . Missaoui, Ahmed; Oliveira, RuiThis master's thesis critically examines the enduring challenges of low productivity, high costs, and extensive waste that plague the construction industry. It focuses on Lean Construction, an increasingly popular methodology, as a potential solution to these issues, with a special emphasis on its application in building refurbishment. The thesis provides an in-depth exploration of Lean Construction's core principles and tools, such as Value Stream Mapping, Pull Planning, Just-in-Time Delivery, the Last Planner System, and Visual Management. It assesses their effectiveness in diminishing waste, enhancing productivity, and reducing refurbishment project costs. A distinctive feature of this thesis is the empirical research undertaken, which includes surveys and interviews with two companies in the French and Tunisian building refurbishment sectors. This research aims to assess the practical implementation of Lean Construction techniques in these regions. The results from this study underscore the ability of Lean Construction methodologies to improve project delivery, cut down costs, and elevate the overall success of refurbishment projects. Drawing from these findings, the thesis presents actionable recommendations for the successful integration of Lean Construction strategies in the realm of building refurbishment. It underscores the importance of effective leadership, the engagement of all stakeholders, the adoption of continuous improvement processes, and the utilization of data and metrics for progress measurement. This research not only enriches the academic and pragmatic understanding of Lean Construction but also acts as a strategic guide for upcoming building refurbishment projects and lays the groundwork for future research in this area.
- Development of a database migration and compatibility systemPublication . Pereira, Marcos Vinícius Alves; Alves, Paulo; Cadavez, Vasco; Leite, MateusBiological hazards, such as bacteria, viruses, and parasites, pose significant threats to human health when encountered through inhalation, ingestion, or skin contact. The Pathogens-in-Foods (PIF) database was established to address these concerns, serving as a centralized resource for accessing information about food borne pathogens in Europe. The European Food Safety Authority (EFSA) plays a pivotal role in ensuring food and feed safety within the European Union. However, the EFSA and PIF databases employ different data structures, making compatibility and data migration essential to enhance the relevance of PIF. This thesis focuses on the development of a data migration and compatibility system, allowing seamless communication between the PIF and EFSA databases. Through this system, the databases can align their information more effectively, ultimately improving food safety and public health. Using Python data migration scripts, it was possible to migrate the old PIF database to EFSA standards, making the two databases compatible. The script was developed following a system of classes, making it possible to use these same classes for future migrations if necessary or even migrations to other databases. The compatibility system developed offers the possibility of converting variables between different databases without the need for data migration. This system is especially useful for variables or entities that cannot or should not undergo the data migration process. In this way, the system operated as a conversion system, making PIF entities compatible with EFSA variables.
- Nunca vou pisar numa loja dessas! Desconstruindo os tabus com a SafadeusaPublication . Rocha, Gleidcy Helle dos Reis; Costa, CarlosA presente dissertação aborda a Safadeusa, uma ideia de negócio inovadora no mercado de produtos eróticos, que busca romper com os tabus existentes e democratizar o uso desses produtos. O estudo examina a evolução do mercado de sextoys, contextualizando-o dentro das transformações sociais, culturais e tecnológicas que têm moldado as perceções sobre sexualidade ao longo das últimas décadas. Ao longo do trabalho, são discutidos aspetos como o impacto da tecnologia no mercado dos produtos eróticos na experiência dos consumidores, as barreiras morais ainda enfrentadas por produtos eróticos e as tendências emergentes que prometem redefinir o setor. A Safadeusa se posiciona como uma resposta às lacunas do mercado, oferecendo soluções que vão além do prazer físico, abordando também questões de educação sexual, saúde e bem-estar. Com uma proposta diferenciada, a marca visa atender a públicos diversos, incluindo mulheres, homens, casais, pessoas idosas e a comunidade LGBTQIA+, promovendo uma sexualidade mais inclusiva e menos estigmatizada. A dissertação analisa como a Safadeusa pretende explorar o uso de tecnologias avançadas para oferecer produtos personalizados e conectados, e como a interseção entre prazer e saúde sexual pode contribuir para uma mudança cultural em direção à normalização do uso de sextoys. Além disso, o trabalho investiga os desafios do mercado, ressaltando a importância de garantir a segurança e a confiança dos consumidores. A Safadeusa propõe-se a ser uma marca que, ao mesmo tempo em que se adapta às mudanças nas preferências do mercado, também lidera o debate sobre a liberdade sexual e a importância do prazer como componente fundamental do bem-estar humano. Com uma abordagem baseada na inovação e no rompimento de barreiras, a empresa pretende redefinir o papel dos produtos eróticos na sociedade contemporânea, transformando-os em instrumentos de autoconhecimento e empoderamento.
- Path planning and optimization of autonomous mobile robots in a logistics warehouse scenarioPublication . Cangussu, Gabriel Henrique Silva; Lima, José; Pereira, Ana I.The use of robots in manufacturing and storage environments is increasing every day. Their benefits within a company’s logistics and their speed in performing tasks make them more productive in terms of carrying out tasks directed towards a common goal. With this, the main objective of this work is to optimize the path planning of autonomous mobile robots (AMRs) for later use in various applications, focusing on logistics warehouse scenarios. The work involves the analysis and improvement of algorithms (and their variants) for control and movement of AMRs, focusing on reducing the execution time of the algorithm itself. Variants of the A* algorithm were implemented and used in path planning problems, aiming to find the shortest path for robots, even in environments with obstacles. The adopted methodology includes a literature review of algorithm models for use in the context of this work, as well as the analysis of algorithm variants found for adaptation and use in the proposed scenario. After analyzing them, the obtained results were checked and compared for two different map models. The performance of the algorithm depends on the functions programmed into it and the steps it must go through to get from the start point to the end point. The greater the number of functions in the algorithm, the longer its execution time will be. The generation of graphs to visualize the algorithm’s result requires most of this execution time, as it uses visual resources to show the user the proper functioning of the tested algorithm. The analyses and tests conducted revealed that, although both algorithms have quick response times suitable for use in scenarios with entry warehouses, exit warehouses, and obstacles along the route, the Sakai A* algorithm variant proved to be more efficient due to its simple structure, fewer steps within the algorithm, and agile response time in generating optimal solutions. This had a shorter execution time compared to the Mehdi A* algorithm variant, which, although good, proved to be slower in obtaining the optimal path.