Interpretable machine learning prediction of mechanical properties of AlCoCrFeNi/graphene composites

Abstract

This study combines molecular dynamics (MD) simulations and machine learning (ML) methods to predict the mechanical properties of AlCoCrFeNi/graphene. The effects of the number of graphene layers, Al concentration and temperature on the mechanical properties of the materials were explored, and it was found that the number of graphene layers had a positive effect on the mechanical properties, while the opposite was true for Al concentration and temperature. Next, nine ML models were used to predict the mechanical properties, of which the CatBoost model performed best on the test set of Young's modulus (E). On the test set of tensile strength (TS), the XGBoost model had the best performance. Then the shapley additive interpretation (SHAP) method was used to analyze the characteristic contribution of the XGBoost model, and the validation results confirmed that the method was feasible and provided effective guidance for the design of high-performance high-entropy alloy composites.