LLM-ADAM: A Generalizable LLM Agent Framework for Pre-Print Anomaly Detection in Additive Manufacturing

2026-05-04

AI Summary

This work addresses the challenge in additive manufacturing where users, often lacking domain expertise, frequently configure suboptimal slicing parameters, leading to thermal or geometric defects that existing methods struggle to detect in G-code prior to printing. To overcome this limitation, the authors propose a generalizable framework based on a multi-agent large language model (LLM) architecture, comprising Extractor-LLM, Reference-LLM, and Judge-LLM, which collaboratively perform structured G-code parsing, documentation-driven parameter alignment, and anomaly detection. Designed with modularity, the approach is agnostic to specific printers, materials, or underlying LLMs, thereby significantly enhancing its generalization capability. Evaluated on a test set of 200 samples, the method achieves an accuracy of 87.5%, substantially outperforming a single-LLM baseline (59.5%) and approaching the performance ceiling for most defect types.

Abstract

Additive manufacturing (AM) continues to transform modern manufacturing by enabling flexible, on-demand production of complex geometries across diverse industries. Fused filament fabrication (FFF) has extended AM to laboratories, classrooms, and small production environments, but this accessibility shifts process-planning responsibility to users who may lack manufacturing expertise. A syntactically valid slicer profile can still encode thermally or geometrically harmful settings, and subtle G-code edits can alter extrusion, cooling, or adhesion before a print begins. Pre-print G-code screening catches accidental or adversarial machine-program errors before material or machine time is wasted. This paper proposes LLM-ADAM as a generalizable LLM framework for pre-print anomaly detection in AM. The framework decomposes the task into three roles: Extractor-LLM maps a G-code file to a structured process-parameter schema; Reference-LLM converts printer and material documentation into aligned operating ranges; and Judge-LLM interprets a deterministic deviation table and G-code evidence to decide whether a part is non-defective or belongs to an anomaly class. Printers, materials, and LLM backbones are interchangeable test conditions, not fixed assumptions. We evaluate the framework on an N=200 FFF G-code corpus spanning two desktop printer families, two materials, and five classes including non-defective, under-extrusion, over-extrusion, warping, and stringing. The best framework configuration reaches 87.5% accuracy, compared with 59.5% for the strongest engineered single-LLM baseline. The results show that structured decomposition, rather than backbone strength alone, is the dominant source of improvement, with defect classes identified at or near ceiling for leading configurations while residual errors concentrate on conservative false alarms for non-defective samples.

Problem

• Additive Manufacturing
• G-code anomaly detection
• Pre-print screening
• Fused Filament Fabrication
• Process parameter validation

Innovation

• LLM agent framework
• pre-print anomaly detection
• additive manufacturing
• structured decomposition
• G-code analysis