Modelo de mantenimiento basado en datos para un telar geotextil

Abstract

This project presents the development of a data-driven maintenance model applied to a geotextile loom, with the aim of improving maintenance management and reducing unplanned downtime. Additionally, the proposal integrates an IoT system for the acquisition, transmission, and storage of the loom’s operational data, including current consumption, fault records, and operator identification, using electronic devices, communication through the MQTT protocol, and a centralized database. Based on the collected information, a data preprocessing and analysis stage was carried out, enabling the design and implementation of a supervised machine learning model focused on preventive and predictive maintenance. This model identifies patterns associated with critical faults and generates early alerts that support decision-making during the loom’s operation. The results obtained demonstrate that the use of historical and real-time data facilitates the early detection of faults, improves equipment availability, and reduces dependence on manual records, thereby minimizing human errors. In this way, the project demonstrates the technical feasibility of integrating IoT and Machine Learning technologies into textile machinery, laying the groundwork for future predictive maintenance strategies in similar industrial environments.