Machine learning assisted design of aluminum-lithium alloy with high specific modulus and specific strength
Huiyu Li, Xiwu Li, Yanan Li, Wei Xiao, Kai Wen, Zhihui Li, Yongan Zhang, Baiqing Xiong
Abstract
Advanced aluminum-lithium alloys are the key structural materials urgently needed for the development of light-weighted aircraft in the aerospace field. In this study, we employ a machine learning approach accompanied by domain knowledge to realize the accelerated design of aluminum-lithium alloy with high specific modulus and specific strength by identifying an optimal combination of key features through a three-step feature filtering of datasets containing 145 alloys. The maximum specific modulus in the experimental alloys that screened from the predicted results increases by 4 % compared with the maximum specific modulus in the comparative dataset. The specific modulus of the designed alloy with the best comprehensive performance increased by 12.6 % compared with the widely used 2195-T8 alloy while maintaining a similar specific strength. Machine learning shows appealing feasibility and reliability in the field of materials design.