AdditiveLLM leverages fine-tuned language models to predict potential defects in additive manufacturing processes before they occur, enabling proactive quality control.

• Accurately predicts critical defect regimes including Keyholing, Lack of Fusion, and Balling
• Evaluates different input formatting methods to optimize prediction performance
• Demonstrates how AI can transform manufacturing quality assurance through early defect detection
• Bridges the gap between language processing technology and engineering applications

This research offers significant value for manufacturing engineers by enabling preventive quality control, reducing waste, and improving production efficiency in 3D printing operations.

AdditiveLLM: Large Language Models Predict Defects in Additive Manufacturing