Prediction and design of high hardness high entropy alloy through machine learning

• Published in: Materials & design Vol. 235; p. 112454
• Main Authors: Ren, Wei, Zhang, Yi-Fan, Wang, Wei-Li, Ding, Shu-Jian, Li, Nan
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.11.2023; Elsevier
• Subjects: High-entropy alloy; Intelligent optimization algorithm; Interpretable machine learning; Machine learning; Solid solution strengthening; Interpretable machine learning; Solid solution strengthening; Intelligent optimization algorithm; High-entropy alloy; Machine learning
• ISSN: 0264-1275; 1873-4197
• DOI: 10.1016/j.matdes.2023.112454

[Display omitted] •Particle swarm algorithm is used to inversely optimize the alloy composition according to target performance.•An optimized feature set including 3 features is obtained by GA, interpretable ML and solid-solution strengthening.•A general design framework is summarized for prediction and composition optimization of various HEA performances. Two data-driven machine learning (ML) models were proposed for the hardness prediction of high-entropy alloys (HEA) and the composition optimization of high hardness HEAs, respectively. The hardness prediction model combined interpretable ML methods with solid solution strengthening theory, and the R2 and RMSE values of 0.9716 and 39.2525 were respectively achieved under the leave-one-out validation method. The optimization model adopted an intelligent optimization algorithm to design the optimized elemental molar ratios of high hardness HEAs and was experimentally verified. A general design framework was summarized for prediction and composition optimization of various HEA performances.