The development of copper alloys combining high strength with excellent electrical conductivity remains a critical challenge for advanced applications in electronics, communications, and energy systems. Traditional alloy design approaches, which often rely on empirical trial-and-error methods, are inefficient in navigating the complex trade-offs among multiple properties. In this study, an interpretable machine learning (ML) framework was established to design precipitation-strengthened Cu-Ni-based multicomponent alloys. By integrating regression models, including support vector regression, random forest, AdaBoost, and XGBoost, with SHAP-based feature-importance analysis, the key compositional factors affecting Vickers hardness (HV) and electrical conductivity (EC) were identified. The optimized XGBoost model exhibited high predictive accuracy for both properties, with R 2 values higher than 0.86. Coupled with the NSGA-II multi-objective optimization algorithm, Pareto-optimal compositions were efficiently obtained. Experimental validation confirmed that the designed alloys achieved a favorable strength-conductivity balance, with the C3N alloy reaching 308 HV and 30.6% IACS, and the C0N alloy reaching 256 HV and 35.0% IACS. Microstructural characterization indicates that the improved properties originate from the combined effects of precipitation behavior, grain refinement, and deformation-induced dislocation strengthening. This work demonstrates a robust and interpretable ML-assisted strategy for accelerating the development of high-performance Cu-based alloys and provides a transferable framework for other multicomponent alloy systems.
Xu et al. (Fri,) studied this question.