Machine Learning-Guided Optimization of Additively Manufactured TiB₂-Modified Aluminum Alloys.

Area of Science:

• Materials Science
• Mechanical Engineering
• Computational Science

Background:

• Additive manufacturing (AM) provides design freedom for aluminum alloys, but process optimization is complex.
• Interdependencies between processing parameters and mechanical properties hinder alloy development.

Purpose of the Study:

• To develop an integrated machine learning (ML) and genetic algorithm (GA) framework for optimizing laser powder bed fusion (LPBF) of aluminum alloys.
• To accelerate the discovery of optimal processing conditions for high-performance aluminum alloys.

Main Methods:

• Established a high-accuracy Light Gradient Boosting Machine (LGBM) model to predict mechanical properties from composition and process parameters.
• Employed a GA-guided optimization to identify optimal LPBF processing conditions.
• Conducted microstructural analysis to understand strengthening mechanisms.

Main Results:

• The ML-GA framework identified optimal parameters yielding a tensile strength of 505 MPa and 8.5% fracture elongation.
• Achieved superior mechanical performance compared to existing data.
• Microstructural analysis revealed in situ-formed TiB2 nanoparticles promoting ultrafine equiaxed grains.

Conclusions:

• ML-driven optimization effectively navigates the high-dimensional parameter space in AM.
• The developed workflow accelerates the design of high-performance, customized aluminum alloys.
• Strengthening mechanisms include grain boundary and dislocation strengthening enhanced by refined microstructure and nanoparticles.