Abstract: With the rising demand for low-cost, recyclable, and high-performance materials, there are increasingly stringent requirements for low-alloyed titanium (Ti) components with excellent mechanical properties. In this study, assisted by a machine-learning-based design strategy, we fabricated a laser powder bed fusion low-alloyed high-performance Ti-O alloy. Ultimate tensile strength of 1005.9 MPa and an elongation of 20.5 % were achieved after three iterations of active learning. Improved mechanical property is attributed to the increased content of solid-soluted O, which reduces grain size and enhances 〈c + a〉 dislocation activities for a more pronounced dislocation strengthening effect. This research provides a robust framework for designing high-performance titanium alloys.

Key words: Machine learning, Laser powder bed fusion, Ti-O alloy, Multi-objective optimization, Strength-ductility trade-off