3D-Printed Parts Design and Quality Control based on Data Analytics — 56a — Lucas Eastlund, Matthew Percival, with advisor Lin Guo, South Dakota School of Mines and Technology

In this study, we apply Machine Learning (ML) algorithms to optimize and understand the correlations between various processing parameters and property variables in 3D printing. While previous research has often focused on individual parameters, our work simultaneously explores the entire 3D printing process — encompassing design, machine settings, material, environment, and inspection — to assess their combined impact on product quality.

Using the MakerBot Replicator+, we printed two customizable, everyday-use parts. We varied key processing parameters such as extruder temperature, printing speed, infill density, infill pattern, layer height, and wall layers. We then evaluated the resulting tensile strength, stiffness, and surface roughness of the printed parts.

Our analysis included multivariable regression, achieving an R-square value of 0.902, indicating strong predictive accuracy. We identified infill density as the most significant factor influencing tensile strength, with infill pattern, printing speed, and wall layers also contributing. Conversely, extruder temperature and layer height were found to have minimal impact.

Our findings suggest that ML techniques can effectively optimize 3D printing parameters to enhance product quality. The expected outcome of this project is the development of a quality control model that leverages real-time data to detect and predict quality issues, provide solutions, and manage emergent properties throughout the design and manufacturing process of 3D-printed parts.

South Dakota School of Mines & Technology
Lin Guo