Design of ultra-high-strength Al-Zn-Mg-Cu alloys via Machine Learning

Yingbo Zhang*, Jiaheng Li, Hui Chen

School of Materials Science and Engineering, Southwest Jiaotong University, Chengdu 610031

ABSTRACT: Al-Zn-Mg-Cu (7xxx-series) alloys have been vastly used as structural parts in the aerospace industry and have shown increasingly appreciable potentials in transportation applications primarily due to their superior physico-mechanic properties and machinability. Noteworthily, recent decades have witnessed the rapid development of some emerging engineering materials (e.g., magnesium alloys, titanium alloys and composites), which challenge 7xxx-series alloys in a broad spectrum. Nowadays, further performance advancement is in dire need for 7xxx-series alloys to get more opportunities in new applications and to persistently remain highly competitive in their dominant fields. Mechanical properties leap without a big sacrifice of production costs is a long-standing and intractable pursuit in the structural materials science community. Here, we realize the pursuit on Al-Zn-Mg-Cu (7xxx-series) alloys through composition optimization. We successfully optimized out a desired 7xxx-series alloy using machine learning with only 4 iterations. The optimized alloy exhibits a spectacular ultimate tensile strength of 952 MPa and an acceptable elongation of 6.3% after an economical and productive preparation process. Microstructural analyses of the developed alloy are conducted systematically. We also propose three possible strategies for further optimization of 7xxx-series alloys towards strength. Our study demonstrates the high efficiency of machine learning method in searching for complex multi-component alloys, and first reveals the nano-sized Al8Cu4Y network structure in T6-treated 7xxx-series alloys. As a result, the optimized alloy has been incorporated into the candidate materials for the mass productions of some critical parts in high-speed trains.

Keywords: machine learning; Al-Zn-Mg-Cu; ultra-high-strength