Machine learning-based comprehensive prediction model for L12 phase-strengthened Fe–Co–Ni-based high-entropy alloys
Document Type
Journal Article
Publication Title
Acta Metallurgica Sinica (English Letters)
Volume
37
First Page
1858
Last Page
1874
Publisher
Springer
School
Centre for Advanced Materials and Manufacturing / School of Engineering
Funders
National Natural Science Foundation of China (52161011, 52373236) / Natural Science Foundation of Guangxi Province (2023GXNSFDA026046) / Guangxi Science and Technology Project (Guike AB24010247) / Central Guiding Local Science and Technology Development Fund Projects (Guike ZY23055005) / Scientific Research and Technology Development Program of Guilin (20220110-3) / Scientifc Research and Technology Development Program of Nanning Jiang-nan District (20230715-02) / Guangxi Key Laboratory of Superhard Material (2022-K-001) / Guangxi Key Laboratory of Information Materials (231003-Z, 231013-Z, 231033-K) / Engineering Research Center of Electronic Information Materials and Devices, Ministry of Education (EIMD-AB202009) / Major Research Plan of the National Natural Science Foundation of China (92166112) / Innovation Project of GUET Graduate Education (2022YCXS200) / Projects of MOE Key Lab of Disaster Forecast and Control in Engineering in Jinan University (20200904006) / Guangdong Province International Science and Technology Cooperation Project (2023A0505050103) / Open Project Program of Wuhan National Laboratory for Optoelectronics (2021WNLOKF010)
Abstract
L12 phase-strengthened Fe–Co–Ni-based high-entropy alloys (HEAs) have attracted considerable attention due to their excellent mechanical properties. Improving the properties of HEAs through conventional experimental methods is costly. Therefore, a new method is needed to predict the properties of alloys quickly and accurately. In this study, a comprehensive prediction model for L12 phase-strengthened Fe–Co–Ni-based HEAs was developed. The existence of the L12 phase in the HEAs was first predicted. A link was then established between the microstructure (L12 phase volume fraction) and properties (hardness) of HEAs, and comprehensive prediction was performed. Finally, two mutually exclusive properties (strength and plasticity) of HEAs were coupled and co-optimized. The Shapley additive explained algorithm was also used to interpret the contribution of each model feature to the comprehensive properties of HEAs. The vast compositional and process search space of HEAs was progressively screened in three stages by applying different prediction models. Finally, four HEAs were screened from hundreds of thousands of possible candidate groups, and the prediction results were verified by experiments. In this work, L12 phase-strengthened Fe–Co–Ni-based HEAs with high strength and plasticity were successfully designed. The new method presented herein has a great cost advantage over traditional experimental methods. It is also expected to be applied in the design of HEAs with various excellent properties or to explore the potential factors affecting the microstructure/properties of alloys.
DOI
10.1007/s40195-024-01774-1
Access Rights
subscription content
Comments
Li, X., Wang, C., Zhang, L., Zhou, S., Huang, J., Gao, M., ... & Tan, Z. (2024). Machine learning-based comprehensive prediction model for L12 phase-strengthened Fe–Co–Ni-based high-entropy alloys. Acta Metallurgica Sinica (English Letters), 37, 1858-1874. https://doi.org/10.1007/s40195-024-01774-1