Aplicativo basado en analítica de datos para la optimización del mantenimiento predictivo de equipos de potencia en subestación eléctrica Sabanalarga (Atlántico)
No hay miniatura disponible
Fecha
2026
Autores
Toledo Buitrago, Lorena Isabel
Carrillo Colón, Yeiner Vanesa
Buitrago Beltran, Cristian David
Jalal Contreras, Luis Fernando
Título de la revista
ISSN de la revista
Título del volumen
Editor
Ediciones Universidad Simón Bolívar
Facultad de Ingenierías
Facultad de Ingenierías
Resumen
Este proyecto de investigación propone el diseño de una plataforma tecnológica basada en analítica de datos para optimizar el mantenimiento predictivo de equipos de potencia en la subestación elevadora Sabanalarga, operada a 110 kV por TRANSELCA. La problemática identificada radica en que los procesos actuales de mantenimiento se fundamentan principalmente en inspecciones preventivas y correctivas, lo que limita la capacidad de anticipar fallas y afecta la confiabilidad y disponibilidad del sistema eléctrico. Por ende, el objetivo es diseñar un sistema que integre variables críticas de operación como temperatura, vibración y disrupciones mediante mecanismos de recopilación, procesamiento y análisis avanzado de datos. A partir de modelos de inteligencia artificial, la plataforma identifica patrones de desgaste y comportamientos anómalos, permitiendo predecir fallas con mayor precisión y fortalecer la toma de decisiones basada en evidencia.
La metodología adopta un enfoque mixto que combina el análisis cuantitativo de variables operativas con análisis cualitativo derivado de encuestas, entrevistas técnicas y validación experta con profesionales del sector eléctrico. Los resultados evidencian una aceptación favorable hacia la implementación de la plataforma y confirman que las variables estudiadas son determinantes para anticipar fallas térmicas, mecánicas y dieléctricas.
Se concluye que la integración de analítica de datos y conocimiento técnico especializado constituye una alternativa viable para mejorar la confiabilidad, eficiencia operativa y gestión del mantenimiento, facilitando la transición de un modelo reactivo a uno predictivo mediante alertas tempranas, visualización de datos y generación automatizada de informes.
This research project proposes the design of a technological platform based on data analytics to optimize the predictive maintenance of power equipment in the Sabanalarga lift substation, operated at 110 kV by TRANSELCA. The identified problem lies in the fact that current maintenance processes are based mainly on preventive and corrective inspections, which limits the ability to anticipate failures and affects the reliability and availability of the electrical system. Thus, the goal is to design a system that integrates critical operating variables such as temperature, vibration, and disruptions through advanced data collection, processing, and analysis mechanisms. Using artificial intelligence models, the platform identifies wear patterns and abnormal behaviors, allowing for more accurate failure predictions and strengthening evidence-based decision-making. The methodology adopts a mixed approach that combines quantitative analysis of operational variables with qualitative analysis derived from surveys, technical interviews and expert validation with professionals in the electrical sector. The results show a favorable acceptance towards the implementation of the platform and confirm that the variables studied are determinative to anticipate thermal, mechanical and dielectric failures. It is concluded that the integration of data analytics and specialized technical knowledge constitutes a viable alternative to improve reliability, operational efficiency and maintenance management, facilitating the transition from a reactive model to a predictive one through early warnings, data visualization and automated reporting.
This research project proposes the design of a technological platform based on data analytics to optimize the predictive maintenance of power equipment in the Sabanalarga lift substation, operated at 110 kV by TRANSELCA. The identified problem lies in the fact that current maintenance processes are based mainly on preventive and corrective inspections, which limits the ability to anticipate failures and affects the reliability and availability of the electrical system. Thus, the goal is to design a system that integrates critical operating variables such as temperature, vibration, and disruptions through advanced data collection, processing, and analysis mechanisms. Using artificial intelligence models, the platform identifies wear patterns and abnormal behaviors, allowing for more accurate failure predictions and strengthening evidence-based decision-making. The methodology adopts a mixed approach that combines quantitative analysis of operational variables with qualitative analysis derived from surveys, technical interviews and expert validation with professionals in the electrical sector. The results show a favorable acceptance towards the implementation of the platform and confirm that the variables studied are determinative to anticipate thermal, mechanical and dielectric failures. It is concluded that the integration of data analytics and specialized technical knowledge constitutes a viable alternative to improve reliability, operational efficiency and maintenance management, facilitating the transition from a reactive model to a predictive one through early warnings, data visualization and automated reporting.
Descripción
Palabras clave
Analítica de datos, Inteligencia artificial, Mantenimiento predictivo, Subestaciones eléctricas, Plataformas tecnológicas