Study on strengthening effects of Zr-Ti-Nb-O alloys via high throughput powder metallurgy and data-driven machine learning

Abstract

Multi-principal elements alloys (MPEAs) have been attracted extensive attention in industry due to their extraordinary properties. However, owning to their large degree of freedom in composition design, finding the principal influence factors of the material properties and further coordinating a desirable combination of conflicting properties is always a challenge. In this respect, we have developed a strategy on studying the MPEAs for its composition design in a certain mechanical property by high throughput preparation of powder metallurgy and by machine learning. We chose Zr-Ti-Nb-O alloys as target materials. To unveil key features that mainly influence the mechanical properties, models selection, features screening, and further features importance ordering were performed. The results indicate that the strength and plasticity are dominated by Λ parameter, difference of atomic radius, difference of shear modulus, etc. The prediction error for the strength and plasticity can reach to below 10% and 16%, respectively. According to analysis of the key features, a strength model is modified and used for evaluating the contributions of solid solution strengthening among principle and trace elements. The strategy proposed here will be applicable on element selections for a large variety of material property modulations in the MPEAs prepared by powder metallurgy.

Keywords

• Solid solution alloy
• Machine learning
• Features screening
• Strength
• Fracture strain