Machine learning assisted optimization design and microstructure-property evolution mechanisms of 600 °C high temperature burn resistant titanium alloys

• A machine learning and genetic algorithm integrated approach was employed to design a novel titanium alloy with superior oxidation resistance, demonstrating a 42% reduction in the parabolic oxidation rate constant compared to the conventional burn resistant Ti-alloys. • The designed alloy exhibits outstanding tensile strength at both room temperature and 600 °C, attributed to a dual-phase synergistic strengthening mechanism. • The alloy shows significantly enhanced burn-resistant performance compared to conventional titanium alloys, with superior anti-ignition capability at temperatures above 600 °C. • A universal alloy design framework was established, which is applicable to other titanium-based systems and validated by close agreement between simulation and experimental results. This study establishes a machine learning assisted optimization design framework for 600 °C high temperature burn resistant titanium alloys, integrating domain expertise and experimental data analysis. An high-precision predictive machine learning model(XGBoost) for oxidation resistance (R 2 = 0.98) combined with genetic algorithm performed global compositional optimization in the Ti-V-Cr-Al-Si-C-Mo-Nb system, minimizing the parabolic oxidation rate (kp) to 1.02 × 10 −6 mg 2 /cm 4 ·s. A representative composition of a burn resistant titanium alloy (TF600) capable of withstanding 600 °C was ultimately identified. This alloy exhibits well-balanced properties with a kp value as low as 1.02 × 10 −6 mg 2 /cm 4 ·s. After optimization of the solution and aging treatment process, the room temperature and 600 °C tensile strengths of TF600 reached 1177 MPa and 970 MPa respectively. Through microstructural characterization of the TF600 alloy and cellular automata simulations, the strengthening mechanism was elucidated as being dominated by the pinning effect of 50–150 nm lath α phase interfaces and dispersion strengthening by 20–80 nm Cr/V-rich Laves phases. The burn resistant mechanism was attributed primarily to the barrier effect of a mixed oxide film composed of V 2 O 5 and Cr 2 O 3 . This research provides a novel pathway for the machine learning assisted optimization and precise design of advanced high-temperature titanium alloys.