Integração de ia e NB-IOT no monitoramento de equipamentos industriais para controle de qualidade

Abstract

This work presents the development of an innovative predictive monitoring system for industrial inkjet printers, specifically designed to detect printing failures like "nozzle missing" in large-format K-JET machines (3.2 meters). The system leverages Artificial Intelligence (AI) in an Intelbras IM1 camera to identify anomalies, generating real-time UDP data. A robust NB-IoT (Narrowband-IoT) device, the HTM32, captures these network data and transmits critical alerts via a 5G cellular network to the Telegram messaging application. This approach ensures continuous and reliable communication, even if local internet infrastructure fails. The objective is to enhance preventive maintenance strategies, minimize material waste, and reduce downtime in industrial printing environments. Two main development lines are explored: first, the implementation of an efficient and highly accurate AI-based vision system for anomaly detection; second, the integration of NB-IoT on a 5G network for resilient real-time alerting. The performance of the developed system is validated through tests replicating real operational scenarios, from data acquisition at the printer to the mobile user interface for notifications. The results achieved, demonstrating the system's effectiveness in providing rapid and reliable alerts, are presented and discussed.