Evaluación de mantenimiento, reparación y overhaul (MRO) en operaciones aéreas de aeronaves comerciales narrow-body a través de modelado, visualización y análisis de datos de aerolíneas.
Loading...
Date
2025
Journal Title
Journal ISSN
Volume Title
Publisher
Universidad de Concepción
Abstract
En un escenario de creciente complejidad operacional y expansión del transporte aéreo comercial, la gestión de los procesos de mantenimiento, reparación y overhaul (MRO), impacta directamente la eficiencia y disponibilidad de las flotas. Frente a la problemática que los esfuerzos de mantenimiento siguen siendo mayoritariamente, generando interrupciones operativas y pérdidas económicas a las aerolíneas, el siguiente trabajo plantea como objetivo desarrollar una herramienta analítica que permita caracterizar el comportamiento operacional de la flota narrow-body de LATAM Airlines Chile, específicamente de los modelos Airbus A320 y A321, a partir de datos de fuentes abiertas. Lo anterior busca identificar comportamientos anómalos de forma temprana que puedan ser atribuidos a posibles eventos MRO, mediante la integración de análisis de datos, modelado y visualización interactiva. Para ello, se recopilan registros operacionales mediante la API (licencia Explorer) de Flightradar24, específicamente operaciones de los meses de marzo, abril y mayo de 2025 en respuesta a las limitaciones que impone la fuente. Los datos obtenidos son sometidos a un proceso de depuración para asegurar su integridad y trazabilidad, los que luego son analizados utilizando un modelo K-Means basado en dos variables: el tiempo de vuelo (Flight Time) y el tiempo de rotación en tierra de las aeronaves (Turnaround Time – TAT). El modelo permite clasificar los vuelos en distintas categorías operacionales, incluyendo eventos críticos compatibles con MRO, en base a expectativas operacionales predefinidas para las flotas. Los resultados se integran en un dashboard interactivo desarrollado en Streamlit (Python), que facilita la exploración dinámica de los datos mediante tres módulos: un panel general para monitorear el rendimiento y detectar eventos críticos; un análisis global que resalta rutas y aeronaves con alta incidencia anómala; y un análisis detallado por matrícula que describe patrones operacionales individuales. Los principales hallazgos a través de la herramienta incluye la detección de eventos críticos, como la detención de 28 días de la aeronave CC-BEF (Airbus A321) en el aeropuerto Rafael Hernández (Puerto Rico), una instalación con capacidad especializada para mantenimiento mayor, o el vuelo de la aeronave CC-BFI (Airbus A320) a Sao Carlos para una intervención MRO. La validación con expertos de LATAM Airlines Chile confirmó que esta detención corresponde a un mantenimiento programado. Asimismo, se identificaron rutas como Calama–Santiago con alta recurrencia de vuelos anómalos que, tras el contraste con conocimiento experto, se atribuyen principalmente a factores operativos, más que a mantenimiento no programado. En definitiva, la herramienta transforma datos operacionales en visualizaciones que permiten detectar de forma temprana anomalías operacionales y eventos de mantenimiento. Esto facilita priorizar y planificar intervenciones con mayor eficiencia, optimizando la disponibilidad de la flota y reduciendo tiempos improductivos. Además, su interfaz modular apoya el análisis a nivel de flota e individual, y permite la integración futura de datos internos que superen limitaciones actuales, potenciando su precisión, confiabilidad y promoviendo decisiones basadas en evidencia.
In a context of increasing operational complexity and sustained growth in commercial air transport, the management of Maintenance, Repair, and Overhaul (MRO) efforts has a direct impact on fleet efficiency and availability. Given the persistent challenge that maintenance efforts remain largely reactive, causing operational interruptions and economic losses for Airlines, this study aims to develop an analytical tool to characterize the operational behavior of LATAM Airlines Chile's narrow-body fleet, specifically Airbus A320 and A321 models, using open-source data. The goal is to enable early identification of anomalous behaviors potentially associated with MRO events by integrating data analysis, unsupervised modeling, and interactive visualization. Operational records were collected via Flightradar24 API (Explorer license), focusing on flights conducted in March, April, and May 2025, due to access limitations inherent to the data source. The data underwent a cleaning process to ensure integrity and traceability. Subsequently, a K-Means clustering model was applied, based on two variables: Flight Time and Turnaround Time (TAT). This model enabled the classification of flights into operational categories, including critical events aligned with MRO patterns, using predefined expectations for each fleet type. The results are integrated into an interactive dashboard developed with Streamlit (Python), which enables dynamic data exploration through three modules: a general panel for monitoring fleet performance and detecting critical events; a global analysis highlighting routes and aircraft with high anomaly incidence; and a detailed analysis by tail number describing individual operational patterns. Key findings include the detection of critical events such as the 28-day ground time of aircraft CC-BEF (Airbus A321) at Rafael Hernández Airport (Puerto Rico), a facility specialized in heavy maintenance, and the flight of aircraft CC-BFI (Airbus A320) to São Carlos for an MRO intervention. Validation with LATAM Airlines Chile experts confirmed that this downtime corresponded to scheduled maintenance. Additionally, routes like Calama–Santiago exhibited a high recurrence of anomalous flights, which expert review attributed mainly to operational factors rather than unscheduled maintenance. Overall, the tool transforms operational data into visualizations that enable early detection of operational anomalies and maintenance events. This facilitates prioritizing and planning interventions more efficiently, optimizing fleet availability, and reducing unproductive downtime. Moreover, its modular interface supports both fleet-level and individual aircraft analysis and allows future integration of internal data to overcome current limitations, enhancing accuracy, reliability, and evidence-based decision-making.
In a context of increasing operational complexity and sustained growth in commercial air transport, the management of Maintenance, Repair, and Overhaul (MRO) efforts has a direct impact on fleet efficiency and availability. Given the persistent challenge that maintenance efforts remain largely reactive, causing operational interruptions and economic losses for Airlines, this study aims to develop an analytical tool to characterize the operational behavior of LATAM Airlines Chile's narrow-body fleet, specifically Airbus A320 and A321 models, using open-source data. The goal is to enable early identification of anomalous behaviors potentially associated with MRO events by integrating data analysis, unsupervised modeling, and interactive visualization. Operational records were collected via Flightradar24 API (Explorer license), focusing on flights conducted in March, April, and May 2025, due to access limitations inherent to the data source. The data underwent a cleaning process to ensure integrity and traceability. Subsequently, a K-Means clustering model was applied, based on two variables: Flight Time and Turnaround Time (TAT). This model enabled the classification of flights into operational categories, including critical events aligned with MRO patterns, using predefined expectations for each fleet type. The results are integrated into an interactive dashboard developed with Streamlit (Python), which enables dynamic data exploration through three modules: a general panel for monitoring fleet performance and detecting critical events; a global analysis highlighting routes and aircraft with high anomaly incidence; and a detailed analysis by tail number describing individual operational patterns. Key findings include the detection of critical events such as the 28-day ground time of aircraft CC-BEF (Airbus A321) at Rafael Hernández Airport (Puerto Rico), a facility specialized in heavy maintenance, and the flight of aircraft CC-BFI (Airbus A320) to São Carlos for an MRO intervention. Validation with LATAM Airlines Chile experts confirmed that this downtime corresponded to scheduled maintenance. Additionally, routes like Calama–Santiago exhibited a high recurrence of anomalous flights, which expert review attributed mainly to operational factors rather than unscheduled maintenance. Overall, the tool transforms operational data into visualizations that enable early detection of operational anomalies and maintenance events. This facilitates prioritizing and planning interventions more efficiently, optimizing fleet availability, and reducing unproductive downtime. Moreover, its modular interface supports both fleet-level and individual aircraft analysis and allows future integration of internal data to overcome current limitations, enhancing accuracy, reliability, and evidence-based decision-making.
Description
Tesis presentada para optar al título de Ingeniero/A Civil Aeroespacial.
Keywords
Aeronáutica comercial, Líneas aéreas, Procesamiento electrónico de datos