No More Trial & Error: Optimizing 3D Printing with Physics + AI

In this episode of Materials Minute, Taylor Sparks, Editor-in-Chief at Integrating Materials and Manufacturing Innovation, breaks down groundbreaking research on improving 3D printing (Fused Filament Fabrication) by combining physics-based modeling with surrogate-assisted optimization.

Researchers from Northwestern University, Washington State University, and MIT developed a GPU-accelerated thermal model to predict cooling rates and optimize print quality without costly trial-and-error. Using Gaussian Process Regression (GPR), they explored nozzle temperature, print speed, bed temperature, and layer height to find optimal settings for PLA prints.

Timestamps:
00:00 — Introduction & why 3D printing parameters matter

00:45 — Physics-based GPU thermal model & cooling rate metric

01:16 — Gaussian Process Regression surrogate model explained

02:15 — Comparing optimized results vs random search

If you want to dive deeper, check out the full paper in Integrating Materials and Manufacturing Innovation.
https://link.springer.com/article/10....

#3DPrinting #MaterialsScience #BayesianOptimization #GPUComputing #AdditiveManufacturing #MaterialsMinute