Doctorado en Ingeniería (tesis)

URI permanente para esta colección

https://hdl.handle.net/10784/81

Examinar

Mostrando 1 - 20 de 48

Metaheuristics for vehicle routing problems with data-driven methods
(Universidad EAFIT, 2024) Mesa López, Juan Pablo; Montoya Echeverri, José Alejandro; Ramos Pollán, Raúl; Toro Bermúdez, Mauricio
Rhizosphere bacterial community and its relationship with cadmium distribution in Theobroma cacao crops in Antioquia
(Universidad EAFIT, 2024) Jaramillo-Mazo, Claudia Yuliet; Correa Álvarez, Javier
Towards an understanding of particulate matter adhesion on electrospun nanofiber membranes to obtain high filtration performance
(Universidad EAFIT, 2024) Valencia Osorio, Laura Margarita; Álvarez Láinez, Mónica Lucía
Smart insect-pest management for cotton crops
(Universidad EAFIT, 2024) Toscano Miranda, Raúl Emiro; Aguilar Castro, José; Toro Bermúdez, Muricio; Caro Piñerez, Manuel
In this research, we address the problem of smart insect-pest management for cotton crops. For the study of this problem, we have positioned it in the framework of the paradigm of Smart agriculture. In this context, Smart agriculture, also known as precision agriculture or digital agriculture, involves the use of advanced technologies to improve agricultural productivity, efficiency, and sustainability. Its focus is to use data-driven and innovative approaches to optimize farming practices and reduce resource waste while ensuring food security. The development of approaches to aid in decision-making for smart insect-pest management for agriculture is necessary to avoid the massive spread of insect pests and the increase in environmental impact. Despite the existence of advances in smart agriculture, integrated management of insect pests remains a challenge. To address this problem, our objective was to develop methodologies, models, and approaches to support decision-making related to smart insect-pest management for cotton crops. To achieve this objective, several sub-objectives were raised, the first one was to design a metacognitive architecture for the smart management of cotton pests, the second was to implement knowledge models for the smart management of cotton pests, and the third was to implement novel AI concepts for the development of knowledge models. Particularly, several research articles were developed to meet the objectives proposed in this thesis. Initially, a review article on the latest trends in Smart agriculture using artificial intelligence and sensing techniques for the management of insect pests and diseases in cotton was carried out. On the other hand, for the first sub-objective, an article was conducted where a metacognitive architecture with metacognitive tasks (meta-memory, meta-learning, meta-reasoning, meta-comprehension, and meta-knowledge) was proposed for smart-pest management of cotton. To meet the second sub-objective, two articles were proposed. The first article is a classification model of the cotton boll-weevil population and the second article presented a fuzzy classification system to analyze the yield of cotton production. Regarding the third sub-objective, two articles were proposed. The first article is about a system with autonomous cycles of data analysis tasks for the integrated management of cotton. And the second article shows how to enhance the insect pest classification in cotton using Transfer Learning techniques. In each article, the strategies/models were evaluated using various datasets. The results showed the capacity of the developed methodologies and models for decision-making in smart insect-pest management for cotton crops. Specifically, our proposals allow the prediction of the boll-weevil behaviors, the diagnosis/prediction of cotton yield, and the prescription of strategies for cotton management into a framework of a meta-cognitive architecture, with good results in performance metrics.
Intelligent model for monitoring, evaluating, and recommending strategies to improve the innovation processes of MSMEs
(Universidad EAFIT, 2024) Gutiérrez Buitrago, Ana Gissel; Aguilar Castro, José Lisandro; Montoya Múnera, Edwin Nelson; Ortega Álvarez, Ana María
The research focuses on how to improve the innovation process in micro, small and medium-sized enterprises (MSMEs). The study is framed within the Smart Innovation paradigm. In this context, innovation is considered a relevant factor for organizational performance that allows the creation and improvement of competitive advantages through the implementation of new ideas, products, concepts, and services to increase market positioning. For organizations aiming to enhance innovation performance, using intelligent systems and artificial intelligence to guide the innovation process poses a challenge. To address this problem, the goal was to develop methodologies, models and approaches to support decision-making related to the intelligent management of the innovation process. To achieve this, specific objectives were defined. The first one is to design an intelligent model to support innovation processes in MSMEs. The second objective is to apply Artificial Intelligence (AI) techniques to customer data sources in social networks and organizational data of MSMEs, aiming to enhance the innovation process; The third objective is to develop an intelligent system to evaluate the innovation levels in MSMEs. The fourth objective is to instantiate a case study in the fashion cluster of the department of Norte de Santander and in the national context, as part of the applied methodology. To fulfill these objectives, research articles were developed. The process began with a literature review article on the current challenges in applying AI techniques to improve innovation processes in MSMEs. A proposed innovation model was made based on the different innovation models that exist in the literature, and the four research articles were written in compliance with the scientific standards that accredit them, to meet the specific objectives outlined in this doctoral thesis. Each article evaluated the strategies/models using various data sets. The results demonstrated the capacity of the proposed methodologies and models for managing of innovation processes. For instance, the proposals enable the prediction of the level of innovation, and the definition of innovation problems, among other aspects, with positive results in performance metrics.
Desarrollo de un electrodo bioinspirado para generación de Hidrógeno
(Universidad EAFIT, 2024) Carmona Saldarriaga, Laura; Ossa Henao, Edgar Alexander
A Novel Injectable Piezoelectric Hydrogel for Periodontal Disease Treatment.
(Universidad EAFIT, 2023) Roldán Lopera, Lina María; Correa Vélez, Santiago Alberto
Periodontal disease is a multifactorial, bacterially induced inflammatory condition characterized by the progressive destruction of periodontal tissues. The successful nonsurgical treatment of periodontitis requires multifunctional technologies offering antibacterial therapies and promotion of bone regeneration simultaneously. For the first time, in this study, an injectable piezoelectric hydrogel (PiezoGEL) was developed after combining gelatin methacryloyl (GelMA) with biocompatible piezoelectric fillers of barium titanate (BTO) that produce electrical charges when stimulated by biomechanical vibrations (e.g., mastication, movements). We harnessed the benefits of hydrogels (injectable, light curable, conforms to pocket spaces, biocompatible) with the bioactive effects of piezoelectric charges. A thorough biomaterial characterization confirmed piezoelectric fillers' successful integration with the hydrogel, photopolymerizability, injectability for clinical use, and electrical charge generation to enable bioactive effects (antibacterial and bone tissue regeneration). PiezoGEL showed significant reductions in pathogenic biofilm biomass (∼41%), metabolic activity (∼75%), and the number of viable cells (∼2-3 log) compared to hydrogels without BTO fillers in vitro. Molecular analysis related the antibacterial effects to be associated with reduced cell adhesion (downregulation of porP and fimA) and increased oxidative stress (upregulation of oxyR) genes. Moreover, PiezoGEL significantly enhanced bone marrow stem cell (BMSC) viability and osteogenic differentiation by upregulating RUNX2, COL1A1, and ALP. In vivo, PiezoGEL effectively reduced periodontal inflammation and increased bone tissue regeneration compared to control groups in a mice model. Findings from this study suggest PiezoGEL to be a promising and novel therapeutic candidate for the treatment of periodontal disease nonsurgically.
Precision agriculture for grazing and animal health management : a case study in Colombia
(Universidad EAFIT, 2023) García Hoyos, Rodrigo Junior; Aguilar Castro, José Lisandro
In this research, we address the problem of fattening management and animal health in rotational grazing. For the study of this problem, we have positioned ourselves in the framework of the paradigm of precision farming, a technological approach that uses advanced information and communication tools and techniques to optimize agricultural and livestock production processes. In this context, precision livestock farming focuses on the use of technologies to improve grazing management and animal health on cattle farms. Some objectives of precision livestock farming are to increase farm efficiency and productivity, improve product quality and reduce production costs. In addition, it also contributes to environmental sustainability by enabling more efficient management of natural resources and reducing negative impacts on the environment. Although precision livestock farming offers many opportunities to improve efficiency and sustainability in livestock production, it also presents challenges that must be addressed for successful implementation. To address this problem, our objective was to develop methodologies, models, and approaches to support decision-making related to productivity management and animal health. To achieve this objective, several sub-objectives were raised, the first one was to develop a precision livestock farming architecture based on emerging technologies (Industry 4.0, artificial intelligence, etc.), the second on developing generic knowledge models of precision livestock farming for animal health and herding management and finally, in the third to develop meta-intelligent models for precision livestock farming in the context of autonomous grazing and animal health management. In general, several research articles were developed to meet the objectives proposed in this thesis. Initially, a review article on the latest trends in precision livestock farming using machine learning techniques was carried out. On the other hand, for the first specific objective, an article was conducted where three autonomous cycles of data analysis tasks based on autonomous computing were proposed for a beef production process for precision livestock farming. To meet the second specific objective, three articles were proposed. The first is a beef cattle weight identification model using machine learning techniques for anomaly detection, the second presented a system for monitoring the cattle fattening process in rotational grazing using fuzzy classification, in the third, a multi-objective optimization model was developed to maximize weight gain of cattle in rotational grazing. Regarding the third objective, three articles were developed, the first one proposed an autonomous cycle of data analysis tasks for the self-supervision of animal fattening in the context of precision livestock farming, and the second article presents a management system for the cattle fattening process in rotational grazing by means of diagnostic and recommendation systems. Finally, the last article proposed the use of the meta-learning paradigm in a cattle weight identification system for anomaly detection. In each article, we evaluated the strategies/models using various datasets. The results showed the capacity of the developed methodologies and models for decision-making in the management of the livestock production process. Specifically, our proposals allow the management of fattening and animal health in rotational grazing, considering, among other things, monitoring, diagnosis, and optimization of the productive process, with good results in performance metrics.
Design and analysis of photovoltaic surfaces based on a multiphysics approach
(Universidad EAFIT, 2023) Espitia Mesa, Gabriel Jaime; Mejía Gutiérrez, Ricardo
Predictive and prescriptive modeling for the clinical management of dengue: a case study in Colombia
(Universidad EAFIT, 2023) Hoyos Morales, William Segundo; Aguilar Castro, José Lisandro; Toro Bermúdez, Mauricio
In this research, we address the problem of clinical management of dengue, which is composed of diagnosis and treatment of the disease. Dengue is a vector-borne tropical disease that is widely distributed worldwide. The development of approaches to aid in decision-making for diseases of public health concern –such as dengue– are necessary to reduce morbidity and mortality rates. Despite the existence of clinical management guidelines, the diagnosis and treatment of dengue remains a challenge. To address this problem, our objective was to develop methodologies, models, and approaches to support decision-making regarding the clinical management of this infection. We developed several research articles to meet the proposed objectives of this thesis. The first article reviewed the latest trends in dengue modeling using machine learning (ML) techniques. The second article proposed a decision support system for the diagnosis of dengue using fuzzy cognitive maps (FCMs). The third article proposed an autonomous cycle of data analysis tasks to support both diagnosis and treatment of the disease. The fourth article presented a methodology based on FCMs and optimization algorithms to generate prescriptive models in clinical settings. The fifth article tested the previously mentioned methodology in other science domains such as, business and education. Finally, the last article proposed three federated learning approaches to guarantee the security and privacy of data related to the clinical management of dengue. In each article, we evaluated such strategies using diverse datasets with signs, symptoms, laboratory tests, and information related to the treatment of the disease. The results showed the ability of the developed methodologies and models to predict disease, classify patients according to severity, evaluate the behavior of severity-related variables, and recommend treatments based on World Health Organization (WHO) guidelines.
Analyses of the Morphometric Variation within Caiman crocodilus Species Complex in Colombia
(Universidad EAFIT, 2022) Angulo-Bedoya, Mónica; Webster, Mark; Benítez, Hugo A.; Balaguera-Reina, Sergio; Correa, Santiago; Roberto, Igor J.; Moncada-Jimenez, Juan F.; Mazzotti, Frank J.; Espinoza-Donoso, Sebastian; Lemic, Darija; Correa Vélez, Santiago Alberto; Pinel Peláez, Nicolás
Computational Geometry Contributions Applied to Additive Manufacturing
(Universidad EAFIT, 2022) Montoya Zapata, Diego Alejandro; Ruiz Salguero, Oscar Eduardo; Posada Velásquez, Jorge León
Engineering estimation of topographic effects in site response analysis
(Universidad EAFIT, 2021) Vergara Gallego, Juan Carlos; Gómez Cataño, Juan David
Despite the irrefutable amount of theoretical and field evidence of the impact of topographic effects on the local response at a site the engineering community still lacks practical methods for its consideration in a standard routinely basis. The incorporation of topographic effects into site response analysis has been a major challenge to engineers as it involves coupling between mechanical and geometric modifications to the incident seismic waves, which at the same time demands for field data that is rarely available to the practicing engineer. On the other hand, it has been observed that the main signature of topographic effects, besides the expected change in frequency con tent, is its spatial variation. These complexities combined to the field data required to build fully coupled mechanical-geometrical models have resulted in complete under consideration of these effects. In this work we follow a rational approach to study the effect of surface topography on the response at local sites after formulating the problem following a diffraction perspective. Since this work is framed in an engineering context our main result is a method to incorporate the effect of surface topography in ground response analysis. Although the method requires numerical simulations we show that if the analyst brings into the problem the dynamic properties of the structure for which the analysis is conducted in the first place very moderate models may be required. These idea leads to the concept of size conditioned response spectra which uses a target structural response spectra to fix the size of the computational model. The resulting numerical domain turns out to be of manageable size thus it can be handled with standard computing resources. Although our proposed approach is limited to surface topography the combination of the theory of diffracted waves and the consideration of the structural response creates new venues to advance in understanding the problem of topographic effects.
Estimación ingenieril de la intensidad sísmica en la vecindad de una formación morfológica
(Universidad EAFIT, 2021-11-13) Sierra Álvarez, Cesar Augusto; Jaramillo Fernandez, Juan Diego
Despite the relevance of topographic effects in seismic response by experimental and theoretical studies, these parameters is not widely used in practice engineering. Nowadays, just a few codes consider these effects, but they do not take into account the frequency of the movement and the relation between the dimensions of the topographic accident and wavelength of the incident field. A two-parameter expression for the quantification of topographic effects in ground response analysis at sites located in the vicinity of a topographic feature is presented. The expression is derived from parametric analysis of convex and concave canonical shapes after identifying that the diffracted field in these geometries remains relatively stable in the low dimensionless frequency regime. The proposed expression is used in the estimation of topographic effects at two sites located in realistic topographic scenarios over a given period range. To show the quality of the approximate formula results are compared with numerical simulation values.
Relationship between enamel mechanical, chemical, ultrastructural properties and mammalian bite force
(Universidad EAFIT, 2021) Fernández Arteaga, Juliana María; Ossa Henao, Edgar Alexánder
Mammalian enamel is one of the hardest and most mineralized tissues in the body. Its main function is to support the loads generated during the chewing process. Mechanical, chemical and ultrastructural properties are responsible for providing it with the high resistance necessary to withstand constant loads and for making the animal’s tooth functional through its life. Animal bite forces as well as their feeding patterns can influence enamel ultrastructure, improving its behavior when facing chewing loads. A brief review of enamel mechanical and chemical properties as well as a brief review on mammalian enamel decussation characteristics are presented in chapter 2. The methodology used in this study is shown in chapter 3, experimental results in terms of mechanical, chemical and ultrastructural properties are presented in chapter 4. In Chapter 5 the results of the experimentation are analyzed in terms of their statistical correlations and the relationship between enamel properties, bite force, and feeding patterns of the analyzed specimens. Finally the conclusions of this investigation are shown in chapter 6. The bite force of the analyzed animals (BFQ) is related to the elastic modulus of the enamel and that the enamel of the analyzed species shows similar characteristics to human enamel in terms of the variations in mechanical and chemical properties. The properties analyzed were compared in terms of the taxonomic classification or the feeding patterns of the analyzed specimens. The mechanical variables (elastic modulus and hardness) do not seem to be related to taxonomic classification or feeding patterns. The decussation fraction is greater than 0:8 regardless of the taxonomic classification or feeding patterns, enamel thickness and decussated thickness are statistically correlated with decussated band thickness, this could indicate that these variables are important in delaying crack growth. Ultrastructural variables do not seem to depend on taxonomic classification or feeding patterns.
Wave propagation in periodic materials with generalized continua
(Universidad EAFIT, 2021) Guarín-Zapata, Nicolás; Gómez Cataño, Juan David
The unrestrained loss of social traits of Bacillus tequilensis EA-CB0015 compromises its success as a biological control agent
(Universidad EAFIT, 2021) Cuellar Gaviria, Tatiana Zazini; Villegas Escobar, Valeska
Integrated operation and maintenance policies for the economic, social and environmental optimization of interconnected urban transport systems
(Universidad EAFIT, 2021) Martinod Restrepo, Ronald Mauricio; Castañeda Heredia, Leonel Francisco
Topographic effects in seismic engineering based upon a rational approach
(Universidad EAFIT, 2019) Sáenz Castillo, Mario Andrés; Jaramillo Fernández, Juan Diego
Improving DFT-based approaches to study CO2 electroreduction on transition metals
(Universidad EAFIT, 2021) Rendón Calle, Jessica Alejandra; Calle Vallejo, Federico; Builes Toro, Santiago
The industrial-scale conversion of electricity obtained from renewable sources is crucial to achieve an economy based on renewable energy. In that scenario, the electrochemical reduction of CO2, offers the possibility of producing some of the most demanded fuels and chemicals in a sustainable way. However, its efficient implementation on industrial scale is limited by factors as the high energy requirements for the product formation, the low selectivity and efficiency of electrolyzers, and the long-term deactivation of the catalysts. Understanding the many aspects that influence the reaction behavior is a challenging task because, apart from solvent and electrolyte effects, there are multiple intermediates, pathways, and products possible under similar operating conditions. In the recent decades this research field has been highly active in theory and experiments, and many studies have focused on finding the main factors that enhance the reaction performance. In this thesis, the electrochemical CO2 reduction is studied using state-of-the-art density functional theory (DFT) simulations, incorporating solvation effects as a crucial factor for improving thermodynamic predictions. To this end, a systematic micro-solvation method was developed to determine the number of hydrogen-bonded water molecules in the first solvation shell and the energetic stabilization granted by those hydrogen bonds. The reduction of CO2 to CO, CH4 and CH3OH on Cu, was considered to test this method, finding very good agreement with experiments without the need to include calculations of reaction kinetics. The estimation of solvation contributions for the CO2 reduction to CO has been extended to other transition metals such as Ag, Au, and Zn, finding significant variations between solvation corrections for the same adsorbates on different metals and finding very good agreement with experimental results. The increase in accuracy of the predictions make possible the development of a semiempirical method to explain the deactivation evidenced experimentally on Cu electrodes during CO2RR to CH4.

Examinar

Envíos recientes