Page 237 - Read Online

Page 2 of 29 Liu et al. J. Mater. Inf. 2026, 6, 18

200,000 candidate compositions of near-α titanium alloys (Ti-Al-V-Mo-Zr-Sn system). A breakthrough
combination of 1,600 MPa dynamic compressive strength and 26% ductility at a strain rate of 3,000 s is
-1
achieved, which is superior to that of the TA15 alloy, as confirmed by a constitutive equation model. This
framework successfully designs novel near α-titanium alloys with strength-ductility synergy through knowledge-
driven feature engineering and multi-objective optimization algorithms, establishing a new paradigm for the
intelligent design of titanium alloys under extreme conditions.

INTRODUCTION
Titanium alloys, characterized by low density, high specific strength, excellent ductility, and corrosion
resistance, are regarded as ideal structural materials under impact loading [1-5] . However, due to their low
thermal conductivity, titanium alloys are prone to thermoplastic instability under high-speed impact,
exhibiting high adiabatic sensitivity and a tendency to form adiabatic shear bands (ASBs). The non-uniform
deformation and stress concentration between these bands and the matrix accelerate crack propagation,
ultimately leading to material failure [6-12] . Improving the strength-ductility balance of titanium alloys has been
considered a potential strategy to improve stress-strain behavior, reduce adiabatic shear sensitivity, and
enhance impact resistance . Near α-titanium alloys, which benefit from multi-scale synergistic energy
[4]
dissipation mechanisms and high-temperature stability, have been regarded as ideal impact-resistant
materials [13-16] . However, current research lacks comprehensive and systematic development of near
α-titanium alloys, particularly for high-speed impact applications . Therefore, it is essential to develop
[17]
efficient methods for producing high-performance near-α titanium alloys suitable for extreme conditions.

The design of composition and microstructure is critical for balancing the strength and toughness of
titanium alloys under extreme impact conditions. Traditional theoretical strategies, such as those grounded
in molybdenum equivalence and d-electron theory (Bo-Md method), have successfully guided the
development of various high-performance alloys [18-20] . However, these approaches rely heavily on costly
“trial-and-error” processes and frequently suffer from limited predictive accuracy when dealing with
complex multi-element interactions due to the scarcity of comprehensive statistical data [21,22] .

To overcome these limitations, data-driven machine learning (ML) has emerged as a transformative
approach for mapping complex nonlinear relationships among composition, processing, and
performance [23-35] . Recent studies have demonstrated that ensemble learning frameworks, such as support
vector machine (SVM) and random forest (RF), can significantly enhance prediction accuracy for
mechanical properties and fatigue life, particularly when physical features such as electron work function
(EWF) are incorporated [36-42] . Despite these advancements, current ML applications often focus on
single-target prediction or operate as “black boxes” lacking physical interpretability. Furthermore, effectively
integrating domain knowledge with multi-objective evolutionary algorithms to systematically resolve the
trade-off between strength and ductility remains a significant challenge [43-49] .

In this work, a multi-objective genetic optimization algorithm framework based on the nondominated
sorting genetic algorithm II (NSGA-II) is proposed to efficiently design near-α titanium alloys with excellent
strength-ductility synergy and high resistance to adiabatic shear under high-strain-rate conditions. By
employing a knowledge-enabled feature engineering approach, the key parameters influencing the strength
and ductility of titanium alloys are identified. The optimal ML model is selected through comparative
evaluation and integrated into the multi-objective optimization framework. Combined with constitutive
modeling and phase diagram calculations, this approach enables the design of near α-titanium alloys that
meet target performance requirements. The proposed method can rapidly generate and screen titanium alloy
compositions that simultaneously exhibit high strength and ductility under extreme strain conditions,

providing an innovative pathway for the accelerated design of titanium alloys for extreme environments.