Accelerated composition-process-properties design of precipitation-strengthened copper alloys using machine learning based on Bayesian optimization

Designing new alloys with high performance is challenging due to the large search space for composition and process parameters. We propose an alloy design strategy based on machine learning algorithms for navigating the enormous search space. Specifically, feature engineering was applied to screen the major features, and a three-step alloy design strategy was employed to extract the required composition. The material design strategy for the multi-performance optimization of Cu-Ni-Si alloy through Bayesian optimization was proposed. This work provides novel insights into the comprehensive properties of Cu-Ni-Si alloys using machine learning with small data, with potential applicability to other materials systems. A machine learning strategy combining feature engineering with three-step alloy design is used to identify the optimal composition space for Cu-Ni-Si alloys with high strength, electrical conductivity, and elongation.