3D Printing Parameter Optimization Using Large Language Models (LLMS)

Optimization of slicing parameters in fused deposition modeling (FDM) usually takes place manually and needs a lot of expertise and trial and error. This paper examines the use of a large language model (LLM) to help optimize the contextual FDM printing parameters. A Workflow was developed where a radically deployed LLM analyzed the slicing profiles and calculated input values for the parameters, which were then applied to G-code. The method was tested with a standard print from 3DBenchy, so the original slicing setup was compared to the setup optimized by the LLM. Experimental results reveal 27.7 min printing time reduction in experimental results and a consumption of 7.75 meters in the usage of filament. Some of the adjustments made were increased travel and infill speeds and decreased extrusion flow. Surface roughness and dimensional accuracy exhibited low differences (0.25 and 0.52 um and 0.0% and 0.033%, respectively), which were within acceptable ranges. Though background scaled on one geometry and material configuration, the findings suggest that the tuning of the parameters that the LLM can enhance can also increase the level of efficiency in FDM processes without undermining the quality of the prints. DOI: https://doi.org/10.52783/pst.3414