Integrating Machine Learning into Design and Process Intelligence: Enabling the Future of Additive Manufacturing

Additive Manufacturing (AM) continues to reshape the boundaries of engineering design and production by enabling the fabrication of geometrically complex, highly customized, and functionally integrated components. To fully capitalize on these capabilities, engineering workflows must evolve—redefining not only how products are conceived but also how process behavior is understood and controlled.

This research seminar presents selected contributions from Prof. Zhao’s coherent research program, which integrates Design for Additive Manufacturing (DfAM), computational modeling, and machine learning (ML) to support intelligent and manufacturable engineering solutions. A central focus of this work is a structured DfAM methodology that addresses three core questions: what to redesign, how to redesign, and whether the new design can be manufactured. This framework supports the development of bio-inspired multifunctional structures, part consolidation methods, and ML-assisted manufacturability assessment.

Building on this foundation, the research further incorporates ML to enhance both early-stage design exploration and in-situ process monitoring. This includes efforts to extract design knowledge, predict manufacturability, and detect defects through data-driven and multi-modal learning techniques. In parallel, the work addresses broader challenges in the application of ML to AM—such as data scarcity, imbalance, domain heterogeneity, and deployment across variable conditions—to bridge the gap between model development and real-world implementation.

By bringing together insights from design theory, machine learning, and manufacturing process monitoring, this research seminar explores pathways toward the development of intelligent, adaptive, and scalable AM systems.