Autonomous process set-up with Bayesian optimization for high shape fidelity in hydrogel-based 3D printing

In extrusion-based 3D printing, particularly with challenging materials, optimizing printing parameters is crucial for improving print quality. Hydrogels, characterized by viscoelasticity and shear-thinning behavior, present significant challenges in extrusion control. To address these challenges, this study proposes a Bayesian optimization (BO) framework to autonomously optimize printing parameters—pre-loading, nozzle movement speed, and extrusion rate—and improve process efficiency with print stability. Due to the complexity of the extrusion process, experimental data were used instead of physical models or simulations. Preliminary experiments identified key defects such as infill void and tail leakage, which were quantified using image preprocessing and custom scoring methods. BO was applied to iteratively refine printing parameters through an autonomous setup, enabling rapid convergence toward the optimal settings. The system efficiently optimized parameters within 10 iterations across different sizes, shapes, and material concentrations. Furthermore, the optimized in-plane parameters were successfully extended to 3D structures using a layer-by-layer deposition approach. This framework provides a robust and efficient solution for improving the print quality in complex materials like hydrogels, ensuring high precision and overall process stability in extrusion-based 3D printing.