Integrating sensors and Machine Learning: A smart monitoring system prototype for quality assurance in additive manufacturing for the aerospace industry

Ensuring quality of aerospace components produced via additive manufacturing is challenging because conventional methods are often reactive and insufficient to detect and prevent defects in real time, not addressing the complexity and precision required. The main aim of this study is to develop an innovative smart monitoring system which integrates an electronic nose (e-nose), a thermal camera, and convolutional neural networks (CNN) to detect anomalies and defects during the production process of components designed for aerospace purposes. The adopted methodology involves combining chemical and thermal data collected by the e-nose and thermal camera, to be analyzed using a CNN. The system is designed for anomaly detection and enables real-time corrective actions, optimizing manufacturing parameters. The CNN’s iterative learning capabilities ensure adaptive and improved monitoring over time. Results demonstrate that this integrated multi-sensor approach has the potential to enhance significantly anomaly detection accuracy, to reduce defects and material waste, and ensure compliance with aerospace quality standards. The prototype originality lies in the synergistic integration of advanced monitoring technologies with machine learning for AM processes, providing a proactive solution to defect prevention. Practical implications include increased production efficiency, reduced costs, and improved sustainability, as well as potential scalability to other high-stakes industries requiring rigorous quality assurance.