The widespread adoption of metal additive manufacturing (AM) is hindered by the inherent complexity of multi-physical phenomena and material-specific behaviors, which challenge the reliable prediction of microstructure, mechanical properties, and defects.
This research establishes a quantitative understanding of the entire metal AM process chain by developing a suite of rigorously validated, physics-based closed-form analytical models. These encompass thermal distribution, microstructure evolution, mechanical property prediction, and defect formation, bridging the critical gap between microstructural evolution and macroscopic performance.
To address the industry challenge of parameter optimization, we pioneer a transformative hybrid methodology. By integrating these foundational physical models with a genetic algorithm optimization engine, the framework achieves orders-of-magnitude computational cost reduction and accelerated convergence. It enables both forward prediction of properties from given parameters and inverse optimization of process conditions to meet target specifications.
This work crystallizes a scientific framework for causality-aware manufacturing, merging first-principles physics with evolutionary optimization. It delivers a provably generalizable, physics-encoded paradigm for process-structure-property control, establishing a new archetype for closed-loop knowledge generation in industrial science.
Learning Objectives:
Upon completion, participants will be able to explain the causal relationship between key additive manufacturing process parameters and the evolution of critical microstructural features (e.g., grain size, texture).
Upon completion, participants will be able to demonstrate how the developed physics-based modeling framework integrated with genetic algorithms can be applied to optimize process parameters for target mechanical properties.
