Predicting the Properties of the Refractory High-Entropy Alloys for Additive Manufacturing-Based Fabrication and Mechatronic Applications

The findings of the last years regarding the deep and machine learning algorithms provided the substantial impetus to the discovery of heretofore unknown data. The future of mechanical/electromechanical engineering lies in the design of smart materials and technologies. In this paper, the problem of material design in particular for applications in mechatronics industry and additive manufacturing-based production has been considered. Developed high-entropy alloys could outreach limited properties of conventionally used steels, ceramics and superalloys. The thermal and mechanical properties of refractory metals-based high-entropy alloys has been studied using a complex of analytical algorithms (linear, random forest and gradient boosting regression). The highest accuracy has been achieved by applying the gradient boosting model (above 91%). Performed calculations allowed to verify the properties of different alloys, hence simplify their further selection for the manufacturing. From the developed ranking of overall properties make the TiNbHfTaW, CrNbHfTaW and VNbHfTaW alloys demonstrated the best results for being used in applications for mechanical and electromechanical engineering.