Maestría en Ingeniería (tesis)
URI permanente para esta colección
Examinar
Envíos recientes
Publicación Designing Short-Format Learning Experiences through Play and Storytelling : Teaching Flourishing Business Models in Colombia(Universidad EAFIT, 2025) Gil Pérez, María Alejandra; Vidal Ramírez, Alejandra; Vidal Ramírez, AlejandraPublicación Desarrollo de una metodología para la elaboración de modelos de exposición con fines de riesgo sísmico a partir de información censal. Caso de aplicación : Santiago de Cali, Colombia(Universidad EAFIT, 2025) Arenas Guzmán, Juliana; Acevedo Jaramillo, Ana BeatrizThis work presents a methodology to develop exposure models for residential buildings using data from the 2018 Colombian Population and Housing Census. An exposure model compiles information about the infrastructure of a location and the parameters that define its behavior under external loads; it is used together with hazard and vulnerability models in catastrophic risk analysis. Census data are valuable because they are public, cover the entire country, and allow estimation of relevant information such as number of buildings, number of floors, construction materials, and uses. The methodology uses the census level of disaggregation in urban areas (block level) and variables from housing surveys, such as wall material, socioeconomic level, and number of inhabitants. Based on these surveys, the number of residential buildings is estimated. In Cali, 326,382 buildings were identified with more than 1.7 million inhabitants. To assign structural typologies, three variables were used: number of housing units per building, socioeconomic level, and wall material; the latter correlates with structural typology, although it may be associated with multiple categories. For this study, information from 24,499 building surveys conducted by the Cali Mayor’s Office was used. The developed model made it possible to create an exposure scenario at the block level that includes number of buildings, structural typology, and inhabitants. Validation against existing models showed differences between 6–11%, values considered acceptable and consistent. The methodology is easy to implement and can be used by government entities to develop exposure models and support seismic risk assessment across the national territory.Publicación Estandarización y automatización del ciclo de vida de modelos de Machine learning en entornos empresariales mediante prácticas de MLOps(Universidad EAFIT, 2025) Escobar Diaz, Juan Esteban; Montoya Múnera, Edwin NelsonCurrently, the development and deployment of Machine learning (ML) models in enterprise environments face significant challenges related to reproducibility, scalability, and process automation. The absence of clear standards for managing the ML model lifecycle can lead to operational inefficiencies, data governance issues, and integration difficulties with production systems. This work proposes the implementation of an architecture based on a structured methodology following MLOps practices, aiming to standardize the development, deployment, and monitoring of ML models within the organization. The proposed methodology and architecture integrate modern tools and approaches to automate the ML pipeline, manage model versioning, enable production monitoring, and optimize computational resource usage.Publicación Evaluation of economic and environmental impacts of green H2 integration in biosyngas methanation(Universidad EAFIT, 2026) Rojas-Valderamma, Daniel Ignacio; García-Freite, Samira; SanJuan, Marco; Acosta, David; Aristizabal, Adriana; Botero, María Luisa; Builes, Santiago; Builes, Santiago; Botero, María Luisa; Aristizabal, Adriana; Maestría y trabajo de grado financiados por la alianza GreenGas entre la Universidad EAFIT, Promigas S.A. y su subsidiaria Transmetano S.A. (Acuerdo P2.0.002-2022).Publicación Reducción de sobreespesores de pavimento mediante procedimiento de superficie suavizada entre las desviaciones constructivas y la rasante de diseño en la superestructura de un puente(Universidad EAFIT, 2025) García Giraldo, Mateo; Agudelo Ospina, John JairoInfrastructure works are vital for the connectivity of regions and territories; therefore, it is essential that both their design and construction are safe, functional, comfortable, and aesthetically pleasing for users. However, construction deviations from the original design generally occur, inherent to human error or other factors such as climatic, site-specific, or technological constraints. While design codes account for tolerances regarding these deviations, they seldom provide solutions for cases where the permissible error is exceeded. In bridge engineering, these scenarios are particularly common, as these structures possess high degrees of flexibility, leading to greater susceptibility to deformations. For instances involving aesthetic deviations in the superstructure that do not result in functional detriment, this study proposes the generation of a smoothed surface. This "envelope-type" surface adapts to the as-built finish while aligning with the design profile, thereby preserving aesthetics and transit comfort. This approach aims to minimize asphalt overlay over-thickness where not required, preventing excessive dead loads and cost overruns, while harmonizing complementary elements such as traffic barriers or railings.Publicación Metodología gráfica para priorizar activos y definir estrategias de mantenimiento óptimas. Caso de estudio : flotas mineras(Universidad EAFIT, 2025) Trujillo-Peña, Juan Esteban; Martinod Restrepo, Ronald MauricioThe purpose of this paper is to present an integrated methodology to support strategic decision-making in maintenance management. This methodology is based on graphical analytical tools that use failure rate and mean time to repair as key indicators. Its main focus is the identification and prioritization of critical assets to simultaneously define the most appropriate maintenance strategies. To this end, it integrates four graphical analysis tools: first, the Total Risk Criticality Matrix; second, the Jack-Knife Diagram; third, the Cost Scatter Diagram; and fourth, the Decision-Making Grid. The development of the methodology is illustrated through a case study of five truck fleets operating in an open-pit coal mine in northern Colombia. The methodology then prioritizes one fleet of nine tanker trucks to complete all proposed steps. The results obtained allow for strategic focus on highly critical equipment and definition of maintenance strategies, thereby improving fleet operational availability and reducing costs associated with failures. This approach highlights the importance of adapting maintenance strategies according to the criticality and actual behavior of assets, with a view to offering a practical perspective on the application of graphical analysis tools in industrial maintenance management.Publicación Transformación digital en la construcción : estrategia para disminuir la fragmentación de procesos en empresas colombianas del sector(Universidad EAFIT, 2025) Jiménez Suaza, Juan Camilo; Botero Botero, Luis FernandoPublicación Geological Origin and Epistemic Uncertainty in Rockfill Dams : A Comprehensive Statistical Assessment of Material Properties and Design Parameters(Universidad EAFIT, 2025) Cardenas Restrepo, Luis Felipe; Ruiz Restrepo, Daniel FelipePublicación Análisis de clasificación para la mejora en la gestión del reabastecimiento de inventario : estudio de caso(Universidad EAFIT, 2025) Madrid Ramírez, Juan Pablo; Ramírez Echeverri, Sergio AugustoPublicación Proposal of a business model for energy transition through the 2 use of emerging technology in Spanish industry(Universidad EAFIT, 2025) Gutiérrez González, Lorenzo; Restrepo Giraldo, John DairoPublicación Modelo de visión artificial basado en detección de bordes para la generación de plantillas de tatuajes(Universidad EAFIT, 2025) Martínez Berrio, Socrates; Hincapie Montoya, Edwin Mauricio; Hincapie Montoya, Edwin MauricioPublicación Diseño de los modelos de abastecimiento para elementos comprados en una empresa que ensambla carrocerías para buses(Universidad EAFIT, 2025) Valencia Ospina, David; Restrepo Suárez, RodrigoPublicación Towards Detection of Fruit Ripening and Moniliophthora Roreri in Cocoa Using a Capacitive Sensor(Universidad EAFIT, 2025) López-Duque, Ossman Antonio; Rendón Velásquez, David; NingunoPublicación Sistema multiagente basado en modelos de lenguaje para optimizar la selección y asignación de talento en outsourcing(Universidad EAFIT, 2025) Álvarez-Petro, Oscar David; Martínez Vargas, Juan DavidPublicación Value Proposition Design and Product Market Fit for Energy Solution in the Mining and Metals Industry in Spain(Universidad EAFIT, 2025) Giraldo-Gómez, Sebastián; Restrepo Giraldo, John DairoPublicación Metodología de rediseño de empaques para el mejoramiento de la usabilidad(Universidad EAFIT, 2025) Campuzano Medina, Santiago; Guarín Grisales, Álvaro de JesúsPublicación Implementación de una herramienta de información logística para mejorar la gestión de faltantes en la línea de ensamble de Renault Sofasa S.A.S.(Universidad EAFIT, 2025) Montoya López, Sofía; Murillo Rengifo, Andrés FelipePublicación Implementación de una arquitectura serverless para la validación jurídica y clasificación UNSPSC en procesos de contratación pública Colombiana(Universidad EAFIT, 2025) Flórez Vasco, John Faber; Lalinde Pulido, Juan GuillermoPublicación Estabilización de una arcilla de alta compresibilidad con escoria de refinería en la región caribe de Colombia(Universidad EAFIT, 2025) Tamayo Ricardo, Rubén Darío; Vidal Valencia, JuliánPublicación Data-Driven ML for the Prediction of Syngas Composition and End-Use Application from residual biomass gasification(Universidad EAFIT, 2025) González Velandia, Laura Catalina; García-Freites, Samira; Sanjuan Mejía, Marco; Acosta Villamil, David; Aristizábal Castrilló, Adriana; Botero Vega, María Luisa; Builes Toro, Santiago; Vega Botero, Maria Luisa; Toro Builes, Santiago; Promigas S.A. E.S.P Transmetano S.A.