Applying AI in Supporting Additive Manufacturing Machine Maintenance

Abstract Based on recent Artificial Intelligence advancements, this paper addresses implementing an AI-based maintenance strategy within a controlled production environment. The architecture employs real-time machine monitoring. Data is first processed locally for early fault detection and later cloud-based to support predictive maintenance functions. A Large Language Model trained with domain-specific knowledge provides operators with basic instructions to manage repetitive faults. The architecture is implemented and tested as a source of experimental data.