Machine Learning in Gel-Based Additive Manufacturing: From Material Design to Process Optimization.

Area of Science:

• Additive Manufacturing
• Materials Science
• Artificial Intelligence

Background:

• Gel-based additive manufacturing (GAM) traditionally relies on trial-and-error for material design and process optimization.
• Existing methods face limitations in predicting gel properties and ensuring consistent printability.
• Accelerated material discovery and process control are crucial for advancing GAM applications.

Purpose of the Study:

• To provide a comprehensive review of machine learning (ML) applications in GAM.
• To explore ML's role in gel formulation, printability prediction, and real-time process control.
• To identify current challenges and future directions for ML in GAM.

Main Methods:

• Review of recent literature on ML algorithms (e.g., neural networks, random forests, support vector machines) applied to GAM.
• Analysis of ML's capability in modeling gel properties (rheology, elasticity, swelling, viscoelasticity) using compositional and processing data.
• Examination of data-driven formulation and closed-loop robotics advancements.

Main Results:

• ML enables accurate modeling of gel properties from diverse datasets.
• Data-driven approaches and robotics are transitioning GAM towards autonomous material discovery.
• Significant progress has been made in predictive printability and real-time process adjustments.

Conclusions:

• ML integration significantly enhances material design and process optimization in GAM.
• Addressing data sparsity, model robustness, and system integration are key challenges.
• Future work should focus on multimodal sensing, generative design, and automated experimentation for broader applications in tissue engineering, biomedical devices, and sustainable materials.